American Academy of Pediatrics

TABLE 1

Adverse Effects of Phototherapy in Newborn Infants

Citation Country	Methods	Participants or Inclusion Criteria	Exclusion Criteria	Interventions	Outcomes	Results	Risk of Bias	Notes or Conclusions
Asthma and Allergies
Aspberg 2007 Sweden	Retrospective population-based registry study	14 803 children born between 1987 and 1999 and recorded in the Swedish Hospital Discharge Register as hospitalized with asthma between 2 y of age and 2001 were compared to all children in the Swedish Medical Birth Registry (n = 1 386 029) for those same years.	Children younger than 2 y of age at the time of hospital admission for asthma were excluded because of uncertainty of asthma diagnosis at <2 y.	Multiple maternal and infant characteristics including phototherapy exposure were studied.	Hospitalization for childhood asthma after age 2 y.	Phototherapy-associated odds ratio (Mantel-Haenszel) for asthma requiring hospitalization (OR: 1.27, 95% CI: 1.08–1.5) after excluding infants with risk factors (except for icterus or jaundice) associated with both asthma and phototherapy (measured confounders).	Unable to adjust for bilirubin levels when estimating associations between phototherapy and asthma. Evaluated multiple risk factors (multiple comparisons).	Hyperbilirubinemia and phototherapy were both associated with increased odds of a childhood asthma hospitalization after age 2 y. However, it was impossible to distinguish the phototherapy-associated effect on asthma from that of jaundice.
Aspberg 2010 Sweden		Perinatal data for singleton children prescribed antiasthmatic medication (n = 61 256) were compared to corresponding data for all singleton children born in Sweden from January 1990 to June 2003 and surviving through June 2005 (n = 1 338 319). Data from 3 large Swedish health registers: Swedish Medical Birth Register, Swedish Prescribed Drug Register, and Swedish Hospital Discharge Register.		Numerous antenatal and neonatal variables including icterus and phototherapy.	Asthma as defined by singleton children who received an asthma medication.	Association of phototherapy with: All childhood asthma (OR 1.35; 95% CI: 1.3–1.4). Asthma stated in prescription (OR 1.42; 95% CI: 1.3–1.5). Asthma on systemic corticosteroids (OR 1.47; 95% CI: 1.2–1.8). A diagnosis of icterus (jaundice) was also associated with asthma (OR 1.45, 95% CI: 1.4–1.5).	Adjusted for birth year, maternal age, parity, smoking during pregnancy, and years of involuntary childlessness. Evaluated multiple risk factors (multiple comparisons). There was no adjustment for degree of hyperbilirubinemia. Phototherapy duration and intensity could not be evaluated.	Phototherapy and jaundice were both associated with a diagnosis of asthma. However, it was impossible to distinguish the phototherapy-associated effect on asthma from that of jaundice.
Ku 2012 Taiwan	Retrospective cohort	11 321 children born 1 997–2000 and randomly selected from the Taiwanese National Health Insurance Research Database (NHIRD).	Subjects with questionable basic data (not precisely defined) were excluded.	Those with neonatal jaundice (exposed) were classified as the study group or icteric children. Those without (nonexposed) were classified as the control group or nonicteric children.	Asthma defined as: at least 4 outpatient or emergency department diagnoses of asthma between ages 1–10 y, or 1 asthma diagnosis during a hospital admission between ages 1–10 y.	Phototherapy-associated odds ratio (Mantel-Haenszel) after excluding children with 5 risk factors associated with asthma and phototherapy (confounders) (preterm or low birth wt, neonatal infections, other respiratory conditions, other birth conditions, and sex). (OR: 1.03; 95% CI 0.72–1.4; P = .958.) Jaundice ICD-9 diagnosis-associated odds of asthma after exclusion of children with the above risk factors (OR: 1.64; 95% CI 1.36–1.98; P < .001).	Unable to adjust for bilirubin concentrations when estimating associations between phototherapy and asthma.	Hyperbilirubinemia but not phototherapy was associated with increased odds of a childhood asthma diagnosis after age 1 y.
Sun 2013 Taiwan	Retrospective cohort	11 328 children randomly selected from the Taiwanese National Health Insurance Research Database (NHIRD) and born between 1997 and 2000.	Incomplete basic data, such as conflicting sex and uncertain birth date.	Main exposure of interest was neonatal jaundice as measured by ICD code. Phototherapy was also recorded as a proxy for bilirubin concentration >15 mg/dL, at which phototherapy charges were reimbursed.	Childhood allergic rhinitis between birth and 10-y of age. Diagnostic criteria for allergic rhinitis were at least 3 outpatient allergic rhinitis diagnoses, 1 during a hospitalization, or 1 in an emergency department.	Association between allergic rhinitis and: Phototherapy (RR 1.06; 95% CI: 0.89–1.27) (Risk difference 2.7; 95% CI: -4.8 to 10.1) Neonatal jaundice (aOR 1.46; 95% CI: 1.2–1.7).	Adjusted for the following confounders: preterm/low birth wt, neonatal infection, other respiratory conditions, other birth conditions, and sex. Unable to adjust for bilirubin concentration. Phototherapy duration and intensity could not be evaluated.	There was no association between phototherapy and childhood allergic rhinitis. Neonatal jaundice was associated with a childhood allergic rhinitis diagnosis before age 10 y.
Das 2015	Systematic review and meta-analysis	7 observational studies (n = 101 499 total participants) that included infants with hyperbilirubinemia and/or those receiving phototherapy in the neonatal period and who were followed up to 12 y. Search terms: [(newborn OR infant OR neonate*) AND (hyperbilirubinemia OR icterus OR jaundice) AND (phototherapy OR therapy OR treatment) AND (allergy OR atopy OR asthma OR allergic rhinitis OR rhinitis OR rhino-conjunctivitis OR hay fever OR atopic dermatitis OR allergic eczema)].	Children with outcome assessment at <1 y of age or those with chronic illness, congenital malformation, or kernicterus.	Neonatal hyperbilirubinemia (cutoff level not defined). Neonatal phototherapy.	Allergic rhinitis or conjunctivitis. Asthma. Eczema.	Hyperbilirubinemia: Asthma (OR: 4.26; 95% CI: 4.05–4.5). Allergic rhinitis (OR: 5.37; 95% CI: 4.16–6.92). Phototherapy: Asthma (OR: 3.81; 95% CI: 3.53–4.11). Allergic rhinitis (OR: 3.04; 95% CI: 2.13–4.32).	Fixed-effects logistic regression model was used for meta-analysis of the individual study results. Did not identify whether studies adjusted for bilirubin levels when estimating associations between phototherapy and allergy or asthma outcomes. Did not evaluate duration of phototherapy exposure. Only 2 of the 7 included studies (Aspberg 2007; Ku 2012) evaluated the effect of phototherapy.	Hyperbilirubinemia and phototherapy were both associated with increased odds of asthma and allergic rhinitis. However, it was impossible to separate the hyperbilirubinemia-associated effect from that of phototherapy. Risk differences and number needed to harm were not calculated.
Wei 2015 Taiwan	Nested case-control	27 693 patients within the Taiwanese National Health Insurance Research database (NHIRD) with a new diagnosis of neonatal jaundice from 2000 to 2007. For each child with neonatal jaundice, 2 children without neonatal jaundice from the same period were randomly selected and included in the nonneonatal jaundice cohort (n = 55 367) Patients with nonneonatal jaundice were frequency-matched with patients with neonatal jaundice according to sex, age, parental urbanization, parental occupation, and index date.	Children >1 mo of age, diagnosed with allergic disease before the index date of neonatal jaundice diagnosis, and those missing information on sex or age.	The main exposure of interest was neonatal jaundice as diagnosed by ICD code. Other diagnoses including phototherapy were examined within the cohort of infants with neonatal jaundice.	Allergic conjunctivitis, allergic rhinitis, atopic dermatitis, asthma, and urticaria as diagnosed by ICD code by the end of 2008	Association between phototherapy in infants with neonatal jaundice and: Allergic conjunctivitis (aHR 1.38; (95% CI: 1.3–1.5). Allergic rhinitis (aHR 1.37; 95% CI: 1.3–1.4) Asthma (aHR 1.21; 95% CI: 1.2–1.3) Atopic dermatitis (aHR 1.22; 95% CI: 1.1–1.3) Urticaria (aHR 1.12; 95% CI: 1.03–1.2). All 5 allergic disease were increased in the neonatal jaundice cases as compared to the group of matched nonjaundiced patients).	Analyses were adjusted for age, sex, comorbidities of fetal and newborn respiratory conditions, infections, prematurity, low birth wt, other birth conditions, and glucose-6-phosphate dehydrogenase deficiency. There was no adjustment for degree of hyperbilirubinemia. Phototherapy duration and intensity could not be evaluated.	Phototherapy and a diagnosis of neonatal jaundice were both associated with 5 childhood allergic diseases (allergic conjunctivitis, allergic rhinitis, asthma, allergic dermatitis, and urticaria). It was not possible to differentiate the effects of hyperbilirubinemia and effects of phototherapy on these allergic outcomes.
Egeberg 2016 Denmark	Retrospective cohort	All children (n = 673 614) born in the 10-y period between 1997- 2006 and recorded in the Danish National Patient Register.	As per inclusion criteria.	Neonatal blue light phototherapy, neonatal jaundice, birth gestation, birth wt as recorded in the Danish National Patient Register.	Primary end-point was the first occurrence of atopic dermatitis measured at first, second, and fifth birthdays as recorded in the Danish National Patient Register.	Association between atopic dermatitis and: Phototherapy at first birthday (aIRR 1.11; 95% CI: 0.92–1.35). Phototherapy at second birthday (aIRR 1.01; 95% CI: 0.88–1.15). Phototherapy at fifth birthday (aIRR 0.97; 95% CI: 0.88–1.08). Neonatal jaundice at first birthday (aIRR 1.13; 95% CI: 1.06–1.21). Neonatal jaundice at second birthday (aIRR 1.13; 95% CI: 1.08–1.18). Neonatal jaundice at fifth birthday (aIRR 1.12; 95% CI: 1.08–1.16).	Adjusted IRRs were mutually adjusted for age, sex, neonatal phototherapy, neonatal jaundice, birth wt, gestation, and season of birth. P values corrected by the Bonferroni method with corrected 2-tailed P value <0.008 considered significant. Children were followed from birth until either first occurrence of atopic dermatitis, migration, death, or (in 3 separate analyses) their first, second, or fifth birthday.	Although neonatal jaundice was linked to an increased rate of childhood atopic dermatitis, neonatal phototherapy was not associated with an atopic dermatitis diagnosis.
Kuzniewicz 2018 Northern California, USA	Retrospective cohort	109 212 newborns born during 2010-2014 at ≥35 wk’ gestation. 11 Northern California Hospitals (Kaiser Permanente [KP]) that employed universal TSB screening before discharge.	Infants who did not remain in the same hospital system during their birth admission, infants who did not remain in the KP health plan for at least 25 mo, and infants without TSB levels.	Phototherapy (defined as a recorded nursing phototherapy flow sheet or both a phototherapy procedure code and an order for phototherapy).	Asthma (diagnosed after 2 y of age: defined as 1) at least 2 asthma diagnoses from any outpatient or inpatient encounter separated by ≥30 d; and 2) at least 2 asthma medication prescriptions in a 12-mo period separated by ≥30 d).	Association of phototherapy with asthma: (multivariable Cox-model adjusted hazard ratio [aHR]: 1.01; 95% CI: 0.92–1.11). (Propensity-score aHR: 1.07; 95% CI: 0.96–1.20.) Relative to TSB levels of 3-5.9 mg/dL, infants with maximum TSB levels between 9-17 mg/dL were at a significantly increased hazard for asthma. However, there was not a significant increase in asthma associated with maximum TSB ≥18 mg/dL (no TSB dose-effect) (aHR: 1.04; 95% CI: 0.90–1.20).	Cox proportional hazard models used for primary multivariable analysis to account for differing follow-up times by infant. Among infants within 3 mg/dL of 2009 AAP phototherapy threshold, propensity scores for receiving phototherapy were used to adjust for TSB levels before, but not after, receipt of phototherapy. Unable to measure phototherapy duration. However, hospital phototherapy treatment was compared to home phototherapy and there was no difference in the outcome.	Phototherapy use was not associated with asthma, and phototherapy is unlikely to increase or decrease asthma risk. Modest levels of hyperbilirubinemia were associated with an increased risk of asthma, but this association was not observed at higher bilirubin levels.
Tham 2018 Singapore	Prospective cohort	Growing Up in Singapore Toward healthy Outcomes (GUSTO) study	Per inclusion criteria.	Neonatal phototherapy for neonatal hyperbilirubinemia during the birth hospitalization as recorded for the GUSTO study.	Diagnosis of eczema, wheezing or use of nebulizer or inhaler, and rhinitis by parental survey Atopic sensitization was assessed through skin prick testing.	Association of phototherapy with: Allergen sensitization by month 60 (aOR 1.1; 95% CI: 0.6–2.0) Eczema by month 60 (aOR 1.3; 95% CI: 0.7–2.3). Rhinitis by month 60 (aOR 1.0; 95% CI: 0.6–1.9). Early onset of wheeze and use of nebulizer or inhaler by month 60 (aOR 0.6; 95% CI: 0.3–1.2). No associations were found between bilirubin levels and these measured outcomes.	Adjusted for sex, ethnicity, breastfeeding levels, family history of allergy, gestational age, maternal education levels and mode of delivery. Parental surveys at risk for recall bias.	No association was found between either phototherapy or bilirubin level and allergic sensitization, eczema, rhinitis, wheezing or nebulizer or inhaler use.
Kuniyoshi 2021	Systematic review and meta-analysis	The systematic review included 19 observational studies (cohort, case-control, cross-sectional) that evaluated the association of neonatal jaundice and neonatal phototherapy with childhood allergic diseases. 14 studies were included in the meta-analysis for jaundice and 8 studies in the meta-analysis for phototherapy.	Case-reports, case-series, systematic reviews, and meta-analyses were excluded. 5 studies were excluded from meta-analysis because no effect-estimate was reported.	Neonatal jaundice and phototherapy. For studies that stratified bilirubin levels, results with TSB >15 mg/dL were used.	Childhood-onset (≤19-y) allergic diseases (asthma, atopic dermatitis, allergic rhinitis, or food allergies).	Association of phototherapy with: Asthma (OR 1.24; 95% CI: 1.1–1.4). Atopic dermatitis (OR 1.31; 95% CI: 1.2–1.4). Allergic rhinitis (OR 1.38; 95% CI: 0.9-2.0). Association of jaundice with: Asthma (OR 1.46; 95% CI: 1.4–1.5). Atopic dermatitis (OR 1.3; 95% CI: 1.1–1.6). Allergic rhinitis (OR 3.01; 95% CI: 0.9–10.3).	Random-effects models weighted by inverse variance estimates were used for meta-analysis. Did not evaluate intensity or duration of phototherapy exposure.	Jaundice and phototherapy were both associated with increased odds of asthma and atopic dermatitis. Jaundice, but not phototherapy, was associated with allergic rhinitis. No studies were found that quantitatively assessed the association between either jaundice or phototherapy and food allergies. It was impossible to separate the jaundice-associated effect from that of phototherapy.
Autism Spectrum Disorder
Wu 2016 Northern California, USA	Retrospective Cohort Study	457 855 infants born at ≥35 wk’ gestation who were recorded in the Kaiser Permanente of Northern California (KPNC) Database (1995–2011).	Infants who died during the birth hospitalization, were transferred out of the KPNC system, or were followed <60 d. Infants with at least 2 inpatient or outpatient physician ICD diagnoses of: Trisomy 21, other genetic disorders, or congenital anomalies diagnosed at <15 d.	In-hospital or home phototherapy.	ASD diagnosis either by ASD evaluation center, a clinical ASD specialist outside the ASD center (ie, psychiatrist, psychologist, or developmental pediatrician), or by a general pediatrician. Hyperbilirubinemia was also not associated with ASD after adjustment for confounders.	Phototherapy (aHR: 1.10; 95% CI: 0.98–1.24) was not associated with ASD in the primary analysis (Cox proportional hazards model). Propensity-adjusted sensitivity analyses also revealed no association between phototherapy and ASD (aHR: 1.09; 95% CI: 0.95–1.24). This model examined the effect of phototherapy among only infants who had a TSB level within 3 mg/dL of the 2004 AAP phototherapy threshold. TSB ≥20 (aHR: 1.09; 95% CI: 0.89–1.35) was also not associated with ASD.	Propensity-analyses adjusted for bilirubin level before phototherapy. Unable to measure phototherapy duration or intensity.	After adjustment for the confounding effects of sociodemographic factors, neither phototherapy nor hyperbilirubinemia was a significant independent risk factor for ASD.
Breastfeeding and Familial Bonding
Kemper 1989 Yale-New Haven Hospital, New Haven, Connecticut, USA	Prospective observational study with matched controls	101 jaundiced infants born in 1987-1988 with a total serum bilirubin >11.99 mg/dL measured before discharge. 155 comparison subjects who were not visually jaundiced and who were seen by the same pediatrician as the matched, jaundiced infant.	Birth wt <2500 g or if infant spent >6 h in the ICU. Comparison group infants were excluded if a bilirubin level was obtained anytime up to 1 mo postnatal.	Diagnosis of jaundice (mean peak bilirubin in jaundiced infants was 16.2 mg/dL, and 55% received phototherapy).	Mothers were surveyed at 1 mo postnatal to assess (1) concerns about jaundice and phototherapy, (2) termination of breastfeeding within 1 mo postnatal, and (3) maternal behaviors consistent with the vulnerable child syndrome.	Of mothers whose infants received phototherapy, 70% thought its use implied that their child was moderately to seriously ill and 74% found its use upsetting. 53% of mothers whose infants received phototherapy vs 31% of those with jaundiced, non–phototherapy-treated infants had ever left their child with someone else for >1 h at 1 mo postnatal. 64% of those with breastfeeding interrupted in the hospital vs 36% without interruption had completely stopped breastfeeding at 1 mo (P < .05). 84% of mothers of jaundiced infants completed the 1-mo survey, as did 80% of mothers of comparison group infants.	Risk of recall bias (survey). Difficult to determine effect sizes for phototherapy since the main focus of the study was a jaundice diagnosis versus no jaundice diagnosis. Phototherapy specific data are presented as proportions for some outcomes. Although some baseline variables (potential confounders) were collected, these were not used to adjust estimates between jaundiced infants receiving phototherapy and jaundiced nontreated infants.	Mothers of infants who had a jaundice diagnosis were more concerned about their infant’s overall health at 1-mo postnatal. Maternal concerns were increased if their infant received in-hospital phototherapy. It should be noted that clinical practice (including increased phototherapy utilization) and parental counseling (on phototherapy and breastfeeding) may have changed over the 28 y since the study was published.
Schedle 1990 County Hospital (Kantonsspital), Aarau, Switzerland	Prospective cohort	107 healthy, term infants, 1986-1987.	Pregnancy complication, >10th percentile wt for gestation, congenital malformation, or neonatal illness.	Three group comparison: (1) 29 jaundiced infants with mean peak TSB of 15.85 mg/dL who received 2–3 d of phototherapy and partial mother-child separation, (2) 40 mildly jaundiced infants with mean peak TSB of 12.22 mg/dL who received neither phototherapy nor separation, and (3) a control group of 38 infants with no apparent jaundice. Measured baseline variables were similar in all groups except the phototherapy group had significantly more boys relative to the other 2 groups (X² = 0 .48; P = .009).	Denver Developmental Screening Test and Ainsworth Attachment Scores at 1 y of age.	There were no differences in Denver scores for social contact (X² = 0.1; P = .93), gross motor behavior (X² = 0.5; P = .77), and fine motor and adaptive performance (X² = 4.5; P = .10) at 1 y of age. No significant difference in the three study groups was found for mother-child attachment at 1 y of age (X² = 1.5; P = .83).	Were unable to completely differentiate effect of bilirubin and phototherapy since mean TSB was lower in the jaundiced non–phototherapy- treated group. Small-sample size.	Treatment with phototherapy as a newborn did not appear to affect neurodevelopment, as measured by the Denver Developmental Screening Test at 1-y of age and by maternal-infant attachment at 1 y of age.
Usatin 2010 Northern California, USA	Retrospective cohort	Infants born between 1995 and 2004 at gestation ≥36 wk with birth wt ≥2 kg. Patients were eligible if they (1) never had a TSB ≥12 mg/dL (n = 128 417) or (2) when they had a TSB ≥17 to <23 mg/dL (a range where phototherapy might be discretionary) as outpatients at 2 to 7 d postnatal and did (n = 1765) or did not (n = 6777) receive inpatient phototherapy after their initial birth admission. Kaiser Permanente of Northern California Database.	Newborns that received phototherapy during their initial birth admission, those with birth admission >48 h, or a conjugated or direct bilirubin ≥2 mg/dL in the first 30 postnatal days. Those that were readmitted to the hospital for any reason with a length of stay ≥96 h within the first 14 postnatal days. Those with <4 outpatient visits in the first year and/or no visits in the second year.	The 3 study comparison groups: 1) never had a TSB ≥12 mg/dL, 2) TSB ≥17 to <23 mg/dL without phototherapy, and 3) TSB ≥17 to <23 mg/dL with phototherapy.	Primary outcome variables: Total outpatient visits during postnatal days 15–364 and total numbers of various types of outpatient visits at ages (in days): 15–59, 60–119, 120–179, 180–364.	There were small increases in total year-1 visits for jaundiced infants without phototherapy (group 2) (adjusted incident rate ratio [aIRR]: 1.04; 95%: CI: 1.02–1.05) compared to infants with bilirubin <12 mg/dL (group 1). There were also slightly more total year-1 visits for jaundiced infants with phototherapy (group 3) (aIRR: 1.03; 95%: CI: 1.01–1.06) relative to those with jaundice and phototherapy (group 2). For both phototherapy and untreated groups of infants with bilirubin ≥17 to <23 mg/dL, increases in visits were greatest during days 15–59, appeared related to follow-up for the diagnosis of jaundice itself, and were not statistically significant in any other measured year-1 time periods. The adjusted increase in visits over group 1 (bilirubin <12 mg/dL) were 0.36 visits in year-1 by group 2 (jaundice and no phototherapy) and 0.73 visits by group 3 (jaundice and phototherapy). The average number of outpatient visits for the 8 most common sick-visit diagnostic codes did not vary by phototherapy use.	Poisson regression was used to account for potentially confounding variables including mother’s age, infant’s year, month, and hospital of birth, gestational age, sex, and race. TSB level that qualified for groups 2 or 3 were dichotomized at 17–19.9 and 20–22.9 mg/dL and used to adjust comparisons between group 2 and group 3 (phototherapy exposure). The authors noted that estimates should be interpreted in light of no correction for multiple comparisons.	Small increases in outpatient visits before age 1 were seen with both neonatal jaundice (+0.36) and phototherapy (+0.73) that most often occurred in the first 2 postnatal months. Phototherapy treatment and jaundice diagnoses were associated with only small increases in first-year outpatient visit rates, consistent with mild or infrequent contribution to the vulnerable child syndrome.
Waite 2016 United States of America	Retrospective cohort	4441 ever-breastfed infants born >35 wk’ gestation at ≥2200 g between May-December 2005 and recorded in the Centers for Disease Control and Prevention Infant Feeding Practices Study II (IFPS II). Infants who received treatments for jaundice other than phototherapy (eg, formula) were excluded.	Breastfeeds were never initiated or if mothers answered via survey that their infants received “other treatment for jaundice,” which included switching completely to formula feeds.	Phototherapy for the treatment of jaundice as indicated by parental report via survey at ∼3 wk postnatal. Parents first indicated that their child was ever jaundice and if so, they were asked whether treatment was required with phototherapy being 1 of the treatment options.	Primary outcome variables: Any breastfeeding at 1, 2, 4, 6, 9, and 12 mo postnatal based on maternal survey response. Secondary outcome variables: Exclusive breastfeeding at 1, 2, and 4 mo postnatal based on maternal survey response.	At 1-mo postnatal there was no increased odds of any breastfeeding (aOR: 1.14; 95% CI: 0.71–1.81) in phototherapy-exposed infants. No differences in any breastfeeding were noted at months 2, 4, 6, and 9 postnatal. At 12 mo postnatal, phototherapy exposure was associated with a reduced odds of any breastfeeding (aOR: 0.58; 95% CI:0.37–0.91). There was a reduced odds of exclusive breastfeeding in phototherapy-exposed infants (aOR: 0.69; 95% CI: 0.49–0.95) at 1-mo postnatal. The reduced odds of breastfeeding in phototherapy-exposed infants persisted at 2 and 4 mo postnatal.	Potential for recall bias due to determination of phototherapy exposure and breastfeeding history via postnatal maternal survey. The first postnatal survey at ∼3 wk asked the respondent whether the infant had jaundice at any time since birth. Mothers that answered that their baby had jaundice were then asked about phototherapy treatment. It is possible that some mothers may not have understood that jaundice and high bilirubin are synonymous terms.	Phototherapy treatment of hyperbilirubinemia may be associated with a decreased odds of exclusive breastfeeding during months 1–4 postnatal. Phototherapy was not associated with a change in the odds of any breastfeeding. Although there was a decrease in the odds of any breastfeeding at 12 mo postnatal, it seems unlikely that phototherapy was the cause given the one-year duration between phototherapy-exposure and change in breastfeeding habits. Unclear whether among jaundiced infants receiving treatment phototherapy has a more adverse effect on breastfeeding than other interventions.
Digitale 2021 Northern California, USA	Retrospective cohort	25 853 infants born at ≥35-wk’ gestation with TSB levels from 3 mg/dL below to 2.9 mg/dL above 2004 AAP phototherapy thresholds. Medical records at 16 Northern California Hospitals (Kaiser Permanente [KP]) from 2013-2017.	Infants with missing data on length of birth hospital stay, TSB <3 mg/dL, or TSB not measured were excluded. Infants whose first TSB higher than 3 mg/dL below phototherapy threshold was ≥3mg/dL above threshold excluded from primary analysis because phototherapy is no longer optional at those higher TSB concentrations.	Inpatient phototherapy during birth hospitalization defined as presence of phototherapy flow sheet in chart or both ICD procedure code and phototherapy order.	Exclusive breast mlk feeding via caregiver self-report at 2-mo well-child visit. Reporting feeding infant any breast milk at 2-mo well-child visit.	No association between phototherapy and exclusive breast milk feeding at 2-mo of age (aRR 0.99; 95% CI: 0.95–1.04). The adjusted estimate of the difference between observed breast milk feeding among those who received phototherapy (average treatment effect on the treated) (ATET) was -0.2% (95% CI: -2.0% to 1.5%) less than that estimated if those patients had not received phototherapy. Any breast milk feeding at 2-mo: aRR 1.02 (95% CI: 1.0–1.04) ATET for any breast milk feeding at 2-mo among those who received phototherapy: +1.6% (95% CI: 0.1%–3.1%)	Emulated a randomized trial by only including infants with a TSB level for which phototherapy was potentially indicated but not essential. Used a directed acyclic graph (DAG) to illustrate modeling assumptions. Adjusted for potential confounding variables using a modified Poisson regression model. Conducted sensitivity analyses by: (1) including infants excluded from primary analysis because of high initial TSB; (2) imputing missing data to include infants initially excluded from primary analysis; (3) analyzing infants who received formula before first TSB; and (4) analyzing infants who formula fed before first TSB. Estimates from sensitivity analyses were similar to primary analysis estimates.	Phototherapy during the birth hospitalization was not associated with reduced breast milk feeding at the 2-mo well-child visit. These results may not apply to hospital systems that do not actively support maternal lactation during photherapy. Therefore, lactation support should be encouraged.
Infant Behavior
Telzrow 1980 Boston Hospital for Women, Boston, Massachusetts, USA	Observational study with matched controls	10 low-risk term, nursery infants who subsequently received phototherapy for high bilirubin were compared to 10 similar infants who did not receive phototherapy. Sex, maternal breastfeeding intent, and rooming-in proportions were similar between groups.	Infants with intrapartum or postnatal abnormalities, including infants of Rh-positive mothers, were excluded.	Phototherapy at clinician discretion (high bilirubin was defined as >9 mg/dL before 48 h postnatal and >11 mg/dL before 72 h).	Brazelton Scale scoring at 3, 6, and 10 d postnatal. (Phototherapy was started at an average age of 76.5 h).	Visual orientation scores were lower in the phototherapy-treated group relative to the nontreated comparison group on postnatal days 3 (P < .01), 6 (P < .01), and 10 (P < .05). Motor scores were lower on day 6 (P < .05).	The authors were unable to separate phototherapy-specific effects from bilirubin-attributable or separation-attributable effects. It was not possible for the examiners to be blinded to treatment group Separate components (26 separate components and 6 cluster-group scores) were assessed within the Brazelton tests (multiple comparisons). Difficult to determine an estimate of effect-size. Decision to initiate phototherapy was at clinician discretion. Mothers were discharged at 3 d. Examinations at 6 d were performed for the phototherapy group within the hospital and for the comparison group at home. Interrater reliability coefficient was >0.85.	Infants that received phototherapy may have had short-term changes in visual orientation. Results may not apply to infants whose eyes are not covered during phototherapy or who are not separated from their parents.
Paludetto 1983 Newborn Nursery of Second School of Medicine; Naples, Italy	Prospective observational study with matched controls	30 term infants undergoing phototherapy for 6 h or more for jaundice (mean bilirubin, 13.3 mg/dL) and 30 comparison subjects matched for sex, gestation, Apgar score, obstetric history, and father’s profession. Enrolled on postnatal day 3.	Infants transferred to special care nursery or with any neonatal illness including Rh and ABO mismatch.	Phototherapy for 6 h or more for jaundice. (Treated with phototherapy in the nursery and the mothers fed them every 3 h versus rooming-in in the non–phototherapy-treated comparison group.)	Brazelton Scale scores at (1) postnatal day 3 (day of enrollment) (treated infants were under phototherapy an average of 24 h at first examination), (2) 24 h postenrollment, and 3) at 1 mo of age.	Visual orientation scores were lower in the phototherapy-treated group relative to the nontreated comparison group on postnatal days 3 (P < .005) and 4 (P < .01) and at 1 mo of age (P < .05).	The mean total bilirubin levels were lower in the comparison group (9.6 [range, 3.5–14.3]) relative to the phototherapy group (13.3 [range, 8.4–17.5]). The authors were unable to separate phototherapy-specific effects from bilirubin-attributable or separation-attributable effects. 26 separate components were assessed within the Brazelton tests (multiple comparisons). Difficult to determine an estimate of effect-size. Although there were general newborn unit phototherapy guidelines, clinicians were free to decide whether to start phototherapy. Only 12 matched pairs were available for 12-mo scoring follow-up. Brazelton scorers were unaware of treatment assignment, but might have noted different levels of visible skin jaundice in subjects.	Infants that received phototherapy may have had short-term changes in visual orientation. Results may not apply to infants whose eyes are not covered during phototherapy or who are not separated from their parents.
Cancer
Cnattingius 1995 (June) Sweden	Nested case-control study	613 cases of lymphatic leukemia were identified in the Swedish National Cancer Register and the Medical Birth Register between 1973-1989. 5 controls were matched to every case by sex and birth month and year.		93 total antenatal and perinatal exposures were examined including phototherapy and neonatal jaundice. Diagnoses and procedures determined by ICD codes.	Lymphatic (ie, acute, chronic, or unspecified lymphocytic) leukemia.	Association of childhood lymphatic leukemia with: Phototherapy (OR 1.0; 95% CI: 0.5–1.8). Physiologic jaundice (OR 1.2; 95% CI: 0.8–1.7).	Given number of studied exposures, the authors estimated 4 or 5 significant associations would be detected by chance alone.	Neither the use of phototherapy or neonatal jaundice was found to increase the odds of childhood lymphatic leukemia. The authors were unable to adjust for degree of hyperbilirubinemia or phototherapy intensity or duration.
Cnattingius 1995 (July) Sweden	Nested case-control study	98 cases of myeloid leukemia were identified in the Swedish National Cancer Register and the Medical Birth Register between 1973 and1989. 5 controls were matched to every case by sex and birth month and year.		Numerous antenatal and perinatal exposures were examined including phototherapy and neonatal jaundice. Diagnoses and procedures determined by ICD codes.	Myeloid (acute, chronic, or unspecified) leukemia.	Association of childhood myeloid leukemia with: Phototherapy (OR 7.5; 95% CI: 1.8–31.9). Physiologic jaundice (OR 2.5; 95% CI: 1.2–5.0). Reanalysis excluding cases with Trisomy 21: Phototherapy (OR 4.3; 95% CI: 0.9–21.9). Physiologic jaundice (OR 1.7; 95% CI: 0.8–4.0).	Given number of studied exposures, significant associations might be detected by chance alone.	Phototherapy and neonatal jaundice were associated with an increased odds of childhood myeloid leukemia. However, these associations were no longer statistically significant after excluding children with Trisomy 21 diagnoses. The authors were unable to adjust for degree of hyperbilirubinemia or phototherapy intensity or duration.
Olsen 1996 Denmark	Retrospective cohort	66 430 neonates discharged with a history of neonatal hyperbilirubinemia between 1977 and 1989 and recorded in the Danish Central Hospital Discharge Register.	Supplementary diagnoses of neonatal immaturity or hemolytic disease.	Diagnosis of hyperbilirubinemia by ICD code.	Diagnosis of cancer within the Danish Cancer Registry.	Among infants with hyperbilirubinemia diagnoses, 87 cancers were observed compared to 85 expected yielding standardized incidence [SIR] ratio=1.0; 95% CI: 0.8-1.3. Leukemia (SIR = 1.2; 95% CI: 0.8-1.7)	Phototherapy diagnoses were not recorded. It was estimated that 85% to 50% of infants in the study received phototherapy based on a random sample of 150 neonates with hyperbilirubinemia in Copenhagen in 1981. Average follow-up of 9.1 y (range 0–15 y) (= 499 502 person-years).	Hyperbilirubinemia was not associated with childhood cancer. The authors were not able to directly estimate the effect of phototherapy on cancer.
Berg 1997 Sweden	Case-control	30 cases of childhood malignant melanoma between 1973 and 1992 and 120 controls matched by same date of birth, same hospital, and same sex.	Per inclusion criteria.	Neonatal phototherapy as determined by Swedish Medical Birth Registry.	Childhood malignant melanoma as determined by Swedish Cancer Registry.	None of the infants diagnosed with malignant melanoma had received phototherapy compared to 11 of 120 controls (one-tailed P value as calculated in study = 0.08) (Odd ratio = 0; 95% CI: 0–1.2). Calculated 2-tailed P value = 0.156).	The average follow-up time was 18 y (range: 10-19). Phototherapy estimates were not adjusted for degree of hyperbilirubinemia or phototherapy intensity or duration.	No association was detected between phototherapy and childhood malignant melanoma.
Roman 1997 Southern England, UK	Case-control	177 cases with Leukemia or non-Hodgkin’s lymphoma and 354 age and sex-matched controls identified from the medical records of 3 hospitals.	Infants from multiple gestation pregnancies who died before birth hospital discharge, or with identified chromosomal anomalies or malformations.	Multiple antenatal and perinatal risk factors including phototherapy and neonatal jaundice, as extracted from hospital medical records.	Leukemia or non-Hodgkin’s lymphoma diagnosed between the age of 3 mo-30 y. Cases for children (0-14 y) extracted from the Childhood Cancer Research Group (1962–1992). Cases for young adults (15–29 y) extracted from Office of National Statistics (ONS) (1972-1987).	Association between any leukemia and: Phototherapy (OR 0.5; 95% CI: 0.1–2.3). Jaundice (OR 0.8; 95% CI: 0.5–1.5). Additionally, no associations were detected between jaundice or phototherapy and acute lymphocytic leukemia, acute myeloid leukemia, or non-Hodgkin’s lymphoma).	Phototherapy estimates were not adjusted for degree of hyperbilirubinemia.	Phototherapy and jaundice were not associated with leukemia or non-Hodgkin’s lymphoma in young adults.
Podvin 2006 Washington, USA	Case-control	595 cases born between 1980-2002 were identified from the Washington State Cancer Registry or Cancer Surveillance System of Western Washington. Controls (n = 5950) were randomly selected from birth certificate records of infants without leukemia and frequency of controls were matched to cases by year.	Per inclusion criteria.	Multiple antenatal and perinatal risk factors including phototherapy and neonatal jaundice. Jaundice diagnoses were documented on birth certificates before 1992. After 1987, ICD code-based phototherapy and neonatal jaundice diagnoses were available.	Childhood leukemia diagnosis (age<20).	Phototherapy (aOR 2.2; 95% CI: 1.0-4.9) Neonatal jaundice as documented by ICD code (aOR 1.4; 95% CI: 1.0–1.9) Neonatal jaundice as documented on birth certificate (aOR 2.1; 95% CI: 1.2–3.8).	Estimates were adjusted for maternal age, gestation, birth wt and race A secondary analysis that excluded infants with Trisomy 21 found similar results. Phototherapy estimates were not adjusted for degree of hyperbilirubinemia.	Phototherapy and neonatal jaundice were associated with childhood leukemia. The authors were unable to adjust for degree of hyperbilirubinemia or intensity or duration of phototherapy.
Brewster 2010 Grampian Region, Scotland, UK	Retrospective cohort	77518 singleton newborn infants born 1976–1990 that survived the neonatal period.	Neonatal deaths.	Received neonatal phototherapy versus no phototherapy.	Standardized incidence ratios for melanoma, basal cell, and squamous cell skin cancer.	Two cases of melanoma occurred in phototherapy-exposed versus 16 cases in unexposed. Standardized incidence ratio of 1.40 (95% CI, 0.17–5.04; P = .834) for melanoma. No cases of squamous cell or basal cell carcinoma of skin were observed in exposed persons.	Limited confounder control for age, sex, calendar period, and socioeconomic measure within a postal code. Bilirubin level unmeasured. Duration of phototherapy unmeasured. Median follow-up of 24 y.	No difference in skin cancer risk was detected in infants exposed to neonatal phototherapy.
Newman 2016 Northern California, USA	Retrospective cohort	449 621 newborn infants born 1995-2011 at ≥35 wk’ gestation who survived to discharge and were followed ≥60 d. Database from 15 Northern California Hospitals (Kaiser Permanente). 1995-2011.	Death or transferred to another hospital before initial hospital discharge. Cancer diagnosis before 60 postnatal days. <60 d of follow-up. Infants born at <35 wk’ gestation.	Received any neonatal phototherapy (in hospital, at home, or both) vs no phototherapy. Dose-response phototherapy variable (0- none, 1-home phototherapy only, 2-phototherapy in single admission, 3- phototherapy in multiple admits).	Cancer diagnoses between postnatal day 60 and the end of the follow-up period. Examined multiple categories of leukemia and other forms/sites of cancer.	Propensity-score adjusted hazard ratio for any cancer after phototherapy: (aHR: 1.0; 95% CI: 0.7–1.6). Lower limit of 95% CI for number needed to harm (NNH) for 10-y cancer risk ∼1100. For children with Down syndrome, lower limit of 95% CI for estimated NNH for leukemia = 23. Propensity-score adjusted hazard ratios for: Any leukemia: (aHR: 1.6; 95% CI: 0.8–3.5). Nonlymphocytic leukemia: (aHR: 1.9; 95% CI: 0.6–6.9). Liver cancer: (aHR 1.4; 95% CI: 0.2–12). In infants with ≥2 phototherapy admissions there was an association between phototherapy and acute myelogenous leukemia (AML) (aHR 8.5; 95% CI 1.9–38) (note: based on only 2 cases of AML with ≥2 readmissions for phototherapy).	Cox regression models including both traditional multivariable models and models incorporating propensity-score adjustment were used to adjust for potential, measured confounders, and duration of follow-up. Multivariable and propensity-score estimates were mutually consistent. Unable to measure total duration of phototherapy exposure. Used a directed acyclic graph (DAG) to illustrate modeling assumptions. Cancer diagnoses were confirmed by a medical record review.	In adjusted analyses, there was no statistically significant elevation in the risk for cancer. Unadjusted analyses did show an elevated risk. Given their elevated baseline risk for leukemia, children with trisomy 21 may be at an even higher risk for leukemia after phototherapy.
Wickremasinghe 2016 California, USA	Retrospective cohort	5 144 849 infants born at ≥35 wk’ gestation who survived >60 d. California Vital Statistics/Patient Discharge Dataset 1998-2007.	Death within 60-d of birth. Cancer diagnosis before 60 postnatal days. Infants born at <35 wk’ gestation.	Receipt of phototherapy during a hospitalization that began within the first 2 postnatal weeks as recorded by ICD-9 code for phototherapy (99.83, 99.83).	Hospital discharge diagnosis of cancer via ICD-9 code between 60-d postnatal and 1 y of age.	Propensity-score adjusted odds ratio for: All cancer: (aOR: 1.4; 95% CI: 1.1-1.9). Myeloid leukemia: (aOR: 2.6; 95% CI: 1.3–5.0). Renal cancer: (aOR: 2.5; 95% CI: 1.2–5.1). Found NNH of 10 638 in the general population and 1285 in children with trisomy 21.	Adjusted for the presence of a jaundice diagnosis, but unable to control for bilirubin concentration, which is also associated with a cancer diagnosis. Phototherapy duration was not measurable. Propensity-score adjustment to adjust for potential, measured confounders. 95% Cis were narrow showing precise estimates.	Adjusted analyses indicated that phototherapy may slightly increase the risk of cancer in the first postnatal year.
Auger 2019 Quebec, Canada	Retrospective cohort	786 998 infants born between 2006 and 2016. Dataset of discharge abstracts from all Quebec hospitals: “Maintenance and Use of Data for the Study of Hospital Clientele Registry of Quebec.”	Cancer diagnosis before age 2 mo or less than 2 mo of follow-up. Infants with missing gestational age or birth wt.	The exposure was defined as a 3-category variable: phototherapy, untreated jaundice, or no exposure Phototherapy at birth or during admissions in the first 28 postnatal days, as recorded by Canadian Classification of Health Interventions procedure codes (1.YZ.12.JA-DQ). Jaundice diagnosis recorded via ICD-10 codes (P58, P59).	Hospitalization for solid or hematopoietic childhood tumors between 2 mo and 11 y of age.	Propensity-score adjusted hazard ratios for: Any cancer: (Phototherapy versus no exposure: aHR: 1.16; 95% CI: 0.9–1.6). (Phototherapy versus untreated jaundice: aHR: 1.07; 95% CI: 0.8–1.5). (Untreated jaundice versus no exposure: aHR: 1.09; 95% CI: 0.9–1.3). Any cancer at 4-–1 y: (Phototherapy versus no exposure: aHR 2.21; 95% CI 1.5–3.3) (Phototherapy versus untreated jaundice: aHR 2.21; 95% CI 1.3–3.7). Late onset solid tumors, between age 4 and 11 y: (Phototherapy versus no exposure: aHR: 2.26; 95% CI: 1.3–3.8) (Phototherapy versus untreated jaundice: aHR: 2.21; 95% CI: 1.2–4.3). The cumulative incidence of childhood cancer was higher for infants with phototherapy (25.1 per 100 000 person-years) and untreated jaundice (23.0 per 100 000) compared to unexposed infants (21.6 per 100 000).	Unable to adjust for bilirubin concentration, which is also associated with a cancer diagnosis. Phototherapy duration was not measurable. Propensity-score adjustment to adjust for potential, measured confounders. Fine-Gray subdistribution hazard model to allow for estimation of probability of cancer over time (cumulative incidence function [CIF]), while accounting for death as a competing risk.	Neonatal phototherapy may be associated with a slightly increased risk of solid tumors in childhood. It was not possible to rule out increasing bilirubin concentration as a potential confounder, since a jaundice diagnosis was also associated with cancer, bilirubin concentrations were not available, and phototherapy would likely be initiated at higher bilirubin levels compared to untreated jaundice. Adjusted hazard ratios for any cancer and for solid tumor cancers, but not for hematopoietic cancers, significantly increased with age for infants exposed to phototherapy.
Seppala 2020 Finland	Case-control	2037 cases with childhood cancer diagnosed from Finnish Cancer Registry between 1996-2014. 5 sex and year-matched controls (n = 10 185) for each case from Finnish Medical Birth Registry.	8 cases and 82 controls due to missing birth wt or gestation data.	Multiple antenatal and perinatal risk factors. Prematurity associated variables were the main exposure variables of interest. Perinatal conditions and procedures, including phototherapy, were secondary exposures of interest.	Childhood cancer (age <20 y) as diagnosed from the Finnish Cancer Registry.	Association between childhood cancer and phototherapy: All newborns (aOR 1.11; 95% CI: 0.91-1.35). Term newborns (aOR 1.17; 95% CI: 0.91-1.49).	Analyses were adjusted for maternal age, maternal smoking, and number of pregnancies. Phototherapy estimates were not adjusted for degree of hyperbilirubinemia or phototherapy intensity or duration.	Phototherapy was not associated with childhood cancer.
Digitale 2021 Northern California, USA	Retrospective cohort	139 100 infants born at >35 wk’ gestation from 1995-2017 with qualifying bilirubin concentration 3 mg/dL below to 4.9 mg/dL above the 2004 AAP phototherapy threshold. Database from Northern California Hospitals (Kaiser Permanente). *Note: This is an extension of the Newman 2016 cohort with an additional 453 786 person-years of follow-up	Patients who died or were transferred to another hospital before initial hospital discharge . Cancer diagnosis before 60 postnatal days or only had a secondary cancer diagnosis. <60 d of follow-up. Infants born at <35 wk’ gestation.	Received any neonatal phototherapy during birth hospitalization or readmission versus no inpatient phototherapy.	Cancer diagnoses between postnatal day 60 and the end of the follow-up period. Required 2 total cancer diagnoses at different departments to reduce false-positive diagnoses. Time of cancer diagnosis defined as date of initial ICD cancer diagnosis. Examined multiple categories of hematopoietic cancers, solid tumors, and other cancers.	Propensity-score adjusted hazard ratio for any cancer afterphototherapy: (aHR: 1.13; 95% CI: 0.8–1.5). Propensity-score adjusted hazard ratios for: Any hematopoietic cancer: (aHR: 1.17; 95% CI: 0.7–1.8). Solid tumors: (aHR: 1.01; 95% CI: 0.7–1.6). Analyses specifically looking for an increase in cancer risk after age 4 y were also negative.	Cox regression models incorporated propensity-score adjustment to adjust for potential, measured confounders, (including bilirubin concentration) and duration of follow-up. Unable to measure total duration or intensity of phototherapy exposure.	Adjusted analyses, detected no significant elevation in the baseline risk for cancer.
Bugaiski-Shaked 2022 Soroka University Medical Center, Be’er-Sheva, Israel	Retrospective cohort	342 172 infants born at ≥32 wk’ gestation born 1998-2018. Perinatal database of obstetric and neonatal discharge summary data merged with hospitalization records from Soroka University Medical Center.	Neonates who died or were diagnosed with any malignancy before 30-d postnatal.	Phototherapy performed during the first 2 wk postnatal as defined by ICD-9 code for phototherapy diagnosis (Z9983).	First pediatric hospitalization with any diagnosis of neoplasm (malignant or benign).	Preterm birth and maternal age adjusted hazard ratios for: Childhood malignancies (aHR 1.89; 95% CI: 1.4–2.7) Benign tumors: (aHR: 1.27; 95% CI: 1.02–1.6) Childhood malignancies in term-infants: (aHR: 1.84; 95% CI: 1.2–2.7)	Unable to adjust for bilirubin concentration when estimating associations between phototherapy and neoplasms. Phototherapy intensity and duration were not measurable. Used Israeli Ministry of Health national cancer registry to cross-reference and verify child malignancy diagnoses that were in database. Median follow-up of 9.5 y (range 0–18 y).	Adjusted analyses indicated that phototherapy may slightly increase the risk of childhood malignancies and benign tumors cancer in the first postnatal year. It was not possible to rule out increasing bilirubin concentration as a potential confounder or to evaluate the impact of phototherapy intensity or duration on outcomes.
Insulin-Dependent Diabetes Mellitus
Dahlquist 2003 Sweden	Case-control	6487 cases born 1973–1997 who were recorded in the Swedish Childhood Diabetes Registry and also linked with the Swedish Medical Birth Registry. 2.8 million births for comparison.	Twins, infants born of diabetic mothers, and infants whose county of birth was unknown,	Phototherapy as recorded in database. Jaundice was also evaluated as an exposure.	A diagnosis of childhood diabetes as recorded in the Swedish Childhood Diabetes Registry.	Adjusted odds of type 1 diabetes in infants who had received phototherapy in 2 high-risk counties with high type I diabetes prevalence: (aOR 1.95; (95% CI: 1.2–3.2). Association between phototherapy and diabetes in all other counties: (aOR 1.06; 95% CI 0.8–1.4). Unadjusted association between therapy and diabetes, irrespective of diagnosis: (OR 3.79; (95% CI: 3.1– 4.6). Unadjusted association between jaundice diagnosis and diabetes: (OR 1.13; 95% CI 1.01–1.3).	Multivariable logistic regression analyses adjusted for birth year, preterm birth, newborn respiratory symptoms, and blood group immunization. Unable to adjust for bilirubin concentration or phototherapy duration or intensity.	Phototherapy was associated with type I diabetes in 2 Swedish counties with high diabetes prevalence, following adjustment for measured confounding variables. This association was not observed in all other Swedish counties in adjusted analyses. The analysis did not attempt to separate the jaundice-associated effect from that of phototherapy.
Newman 2016 Northern California, USA	Retrospective cohort	499 642 infants born at ≥35 wk’ gestation who were recorded in the Kaiser Permanente of Northern California (KPNC) Database (1995–2011).	Infants who died during birth hospitalization, who were transferred, and who were followed <60 d after birth.	In-hospital or home phototherapy.	Type-1 diabetes diagnosis defined as either ≥2 encounters (inpatient or outpatient) with a diagnosis of type 1 diabetes or 1 encounter and either ≥2 hemoglobin A1c levels ≥6.5% or a pharmacy record indicating insulin prescription. Diagnoses obtained from KPNC virtual data warehouse and diabetes registry.	Association between any phototherapy and type 1 diabetes: (aHR 1.06; 95% CI: 0.8 to 1.5). Adjusted incidence rate ratio (aIRR) for phototherapy:1.03 (95% CI: 0.8–1.4) corresponding to an estimated excess risk of 0.58 (95% CI, −4.7 to 5.9) per 100 000 person-years. Propensity-adjusted Cox model-derived association between phototherapy and type I diabetes: Restricted model that included only those with ≥1 TSB between −3 and +4.9 mg/dL from 2004 AAP phototherapy threshold (aHR 0.95; 95% CI, 0.60–1.5) Inclusive model of all subjects with dichotomous variable for any TSB level exceeding 2004 AAP (aHR 0.91; 95% CI: 0.7–1.3).	Propensity-analyses adjusted for bilirubin level before phototherapy. Unable to measure phototherapy duration or intensity.	There was no evidence of increased type 1 diabetes risk in children who had received phototherapy.
Melanocytic Nevi
Bauer 2004 Germany	Cross-sectional study	1812 white children aged 2–7 y attending day care centers in 2 German cities in 1998. Note: this very brief report describes a secondary analysis of a 1998 study (Wiecker et al Cancer. 2003; 97:628-638) designed to examine the association of sun exposure and nevus counts in parents with childhood development of melanocytic nevi.	Per inclusion criteria.	Neonatal phototherapy as determined by parent survey.	Total body nevus counts and pigmentary characteristics were assessed in all participating children (52.9% boys).	Multiple linear regression analysis did not identify an association between neonatal phototherapy and the number of melanocytic nevi in children (P = .21). The number of vacation days in the sun (P < .001) and an at-home outdoors activity score (P = .03) were associated with increased childhood melanocytic nevi.	Limited data are presented in this brief report. Thus, we are unable to calculate measures of effect size. A standardized interview of parents included questions of national origin, sun exposure history, sun protection, and history of neonatal jaundice. Risk of recall bias from the use of survey to determine phototherapy. Limited control for potential confounders at the time of phototherapy due to survey design. Unable to adjust for specific bilirubin concentrations in phototherapy-treated infants versus unexposed infants.	No association was found between phototherapy as reported via parent survey and childhood melanocytic nevus count at ages 2–7 y. Limitations include the potential for recall bias attributable to the survey design and an inability to adequately control for potential confounders at the time of phototherapy.
Matichard 2006 Necker Hospital, Paris, France	Case-control	Eighteen 8- to 9-y-old children who were born from 1994 to 1995 and treated with phototherapy for jaundice at birth and 40 age-matched schoolchildren who had not received neonatal phototherapy.	Per inclusion criteria.	Neonatal phototherapy for jaundice.	Total nevus count and nevus count by size in millimeters (mm) at ages 8–9 y.	Multivariate analysis using a variance model demonstrated an associated increase in nevi of size 2-5 mm (P < .001), but not an increase in total nevus count. Reported solar exposure was associated with an increase in the total nevus count (P = .02) and age (P = .02) and freckles (P = .03) were associated with an increase in 2–5 mm nevi.	A variance model was used to adjust for age, skin/eye/hair classifications, and survey-reported solar exposure. Effect sizes for multivariate analysis are not reported. Multiple comparisons. The evaluations were performed by the same dermatologist but at different locations: at the hospital for those receiving phototherapy and at a school for the nonexposed comparator group. Thus, the evaluation was not blinded.	An association was found at ages 8–9 between neonatal phototherapy exposure and an increased 2-5 mm nevus count. There was no association between phototherapy and total nevus count. Limitations include a nonblinded examination and the inability to control for confounders and bilirubin concentration in both groups at the time of potential phototherapy exposure shortly after birth.
Mahé 2008 France	Retrospective cohort	828 9-y-old children at 52 French primary schools in May-June 2007.	Children absent from school on the day of the skin examination.	Neonatal phototherapy for neonatal jaundice as determined by a standardized survey of parents and children.	Melanocytic nevus count as performed by 2 dermatologist-trained nurses.	No difference in the melanocytic nevus count was detected between students that reported phototherapy exposure versus unexposed students (mean± SD nevus count of 16.8±9.8 in exposed vs 16.7±10.5 in unexposed). Skin phototype, skin color, eye color, and number of sunburn episodes were associated with nevus count.	Statistical analysis was done by X² test. A standardized interview of parents included questions of sun exposure history, sun protection, and history of neonatal phototherapy. Risk of recall bias from the use of survey to determine phototherapy. Limited control for potential confounders at the time of phototherapy attributable to survey design. Unable to adjust for specific bilirubin concentrations in phototherapy-treated infants versus unexposed infants. Evaluation was not blinded.	No association was found between phototherapy as reported via parent/child survey and melanocytic nevus count at age 9 y. Limitations include the potential for recall bias because of the survey design, an unblinded analysis, and an inability to adequately control for potential confounders at the time of phototherapy.
Csoma 2011 The University of Szeged, Hungary	Cross-sectional twin study	58 pairs of twins (15 monozygotic pairs, 43 dizygotic) and one set of dizygotic triplets aged 3–30 y between January-April 2008 at The University of Szeged, Hungary. In each pair, one twin had received phototherapy for neonatal jaundice and the other had not.	Per inclusion criteria.	Neonatal blue light phototherapy as recorded in the neonatal medical chart. Icterus/jaundice, and prematurity as recorded in neonatal charts.	Melanocytic lesions as determined by skin examination. Uveal melanocytic lesions as determined by detailed ophthalmic examination.	In the multivariable linear regression analysis, neonatal blue light phototherapy was significantly associated with increased melanocytic nevi prevalence: (standardized coefficient: 0.158; 95% CI: 0.003–0.352; P = .47) (unstandardized β coefficient: 0.076). Age (standardized coefficient: 0.503; 95% CI: 0.051–0.101; P < .0001) and number of subtropical/tropical holidays (0.178; 95% CI: 0.012–0.332; P = .035) were also associated with increased skin lesion prevalence. Neonatal blue light phototherapy was associated with an increased odds of benign pigmented ocular lesions (aOR: 3.778; 95% CI: 1.694–8.423; P = .001).	Examining dermatologists and ophthalmologists were blinded to the subjects’ neonatal blue light phototherapy treatment history. Twin comparison controls for many antenatal and genetic variables present before potential phototherapy exposure. Were able to determine receipt of phototherapy exposure via chart review. Subjects or parents were surveyed on family skin lesion history, sun exposure and protection habits. These risk factors along with age and pigmentary traits were adjusted for in the multivariable analyses. Multivariable linear regression for skin lesions (natural log of skin lesions was normally distributed) and multivariable logistic regression for uveal lesions (not normally distributed).	This twin comparison study demonstrated an association between neonatal blue light phototherapy exposure and increased numbers of both melanocytic skin and uveal lesions.
Wintermeier 2014 Ludwig-Maximilian Hospital in Munich, Germany	Case-control	52 white, 5- to 6-y-old children who were previously exposed to neonatal blue light phototherapy after birth at Ludwig-Maximilian Hospital in Munich, Germany and 52 nonexposed children who were matched on age, gender, gestational age at birth, and skin phototype.	Per inclusion criteria.	Neonatal blue light phototherapy exposure.	Melanocytic nevus count at age 5–6 y by a single investigator. Café-au-lait macules, congenital nevi, and freckles were assessed separately.	The median nevus count in children with phototherapy was 17 compared to 18.5 in the matched nonexposed comparator group (P = .18) (Mann-Whitney U test). A higher count of café-au-lait macules (mean 0.5) were noted in the phototherapy exposed children relative to the nonexposed (mean 0.2) (P = .001). Skin phototype (P = .05), duration of sunny vacations (P = .02), and vacations in sunny areas (P = .03) were associated with increased melanocytic nevi.	Limited control for potential confounders at the time of phototherapy. Unable to adjust for specific bilirubin concentrations in phototherapy-treated infants versus unexposed infants. A history of family history of skin cancer, sun exposure, and sun protection was obtained by parental survey. Blinding of the examining dermatologist to phototherapy exposure was not noted.	No association was found at ages 5–6 between neonatal phototherapy exposure and melanocytic nevus count. Limitations include a nonblinded outcome examination and the inability to adjust for confounders and bilirubin concentration in both groups shortly after birth, the time of potential phototherapy exposure.
Lai 2016	Systematic review and meta-analysis	5 observational studies (n = 2921 total subjects) that evaluated the association of neonatal phototherapy with melanocytic nevi development. (Bauer 2004, Csoma 2011, Mahé 2008, Matichard 2006, Wintermeier 2014.) Search terms: (“photo-therapy”[Mesh] AND blue light*) AND (“nevus, pigmented”[Mesh] OR moles OR melanocytic nevi).	Studies were excluded if they were not cohort, case-control, or cross-sectional designs or if mean values and standard deviations of nevi counts were not included or calculable from the presented data. (4 reviews excluded and 1 excluded for insufficient information.)	Neonatal phototherapy for hyperbilirubinemia.	Melanocytic nevus count.	Those with previous phototherapy were not found to have a significantly higher mean melanocytic nevus count: (Weighted mean difference: 0.32; 95% CI: -0.67 to 1.31; P = .53).	DerSimonian and Laird random-effects models were used to calculate a weighted mean difference. There was both clinical heterogeneity and significant statistical heterogeneity between the 5 studies (I²=69.1%; P = .01). Visual inspection of the funnel plot could not exclude publication bias. Formal Egger’s test failed to provide strong evidence of a small-study effect (P = .09).	This meta-analysis of published studies did not demonstrate a significantly higher mean number of melanocytic nevi after exposure to neonatal phototherapy. However, there was a great degree of heterogeneity between the included studies.
Neurologic
Maimburg 2016 Denmark	Retrospective cohort	70 230 singleton children born between February 1998 and May 2003 as recorded in the Danish National Birth Cohort (DNBC).	Children with missing information on phototherapy status (n = 40), birth outcome (n = 9), and twins (n = 15).	Information on exposure to phototherapy for hyperbilirubinemia was obtained from a maternal questionnaire at 6 mo postnatal. When estimating the risk of epilepsy, children were followed from postnatal day 29 until onset of epilepsy, death, or end of the follow-up in August 2010. In the analyses for the risk of febrile seizure, children were followed from postnatal day 90 until onset of febrile seizure, death, or 5 y, whichever came first.	Information on epilepsy and febrile seizures were obtained from ICD-10 diagnoses within the Danish National Hospital Registry (DNHR).	Treatment with phototherapy for hyperbilirubinemia was associated with an increased hazard for developing epilepsy in early childhood (aHR: 1.66; 95% CI: 1.23–2.24). Phototherapy treatment was not associated with febrile seizures (aHR: 1.04; 95% CI: 0.86-1.27). The increased hazard of epilepsy was present for boys (aHR: 1.98; 95% CI: 1.40–2.78) but not for girls (aHR: 1.14; 95% CI: 0.64–2.02).	Cox proportional hazard regression models accounted for observed follow-up time of each subject and adjusted for potential, measured confounders (gender, birth gestation, socioeconomic status, maternal risk factors during pregnancy, and maternal age, parity, infertility treatment). Were unable to adjust for bilirubin concentrations at the time of potential or actual phototherapy exposure. Maternal survey response for phototherapy was validated via a medical record review of a random 100-patient sample of infants born at Aarhus University Hospital, which revealed 96% agreement.	Phototherapy for hyperbilirubinemia in newborn infants was associated with an increased risk of epilepsy for males in early childhood. No excess risk for febrile seizures was seen. A limitation was the inability to adjust for bilirubin concentration in both groups at the time of potential phototherapy exposure.
Newman 2018 Northern California, USA	Retrospective cohort	496 632 infants born at ≥35 wk’ gestation in 1995–2011 and followed ≥60 d. Kaiser Permanente of Northern California Database of 15 hospitals.	Exceeded 2009 AAP exchange transfusion thresholds (n = 1773) and those with seizure diagnoses at <60 d postnatal (n = 1237). Died during birth admission.	Dichotomous variable for any phototherapy within the hospital or at home.	Primary outcome: ≥1 encounter with a seizure diagnosis plus ≥1 AED prescription. Secondary outcomes: other combinations of ≥1 or ≥2 encounters with seizure or epilepsy diagnoses and ≥1 AED prescriptions.	Phototherapy-associated hazard ratio for Seizure (aHR: 1.22; 95% CI: 1.05–1.42; P = .009). Use of phototherapy increased during the study period from 2.4% in 1995% to 15.9% in 2011. More boys experienced both phototherapy and also seizures, but there was no effect modification (interaction) between sex and phototherapy exposure. Adjusted 10-y excess risks per 1000 infants associated with phototherapy were 2.4 (95% CI: 0.6–4.1) overall, 3.7 (95% CI: 1.2–6.1) for boys, and 0.8 (95% CI: -1.7 to 3.2) for girls.	Cox proportional hazard and Poisson regression models were used to adjust for TSB concentrations and other measured, potential confounders. Cox models accounted for observed follow-up time of each subject. Propensity-score adjusted model estimates were similar to Cox model estimates. Additional analyses showed that bilirubin concentrations of 20–24.9 mg/dL only became significantly associated with seizure-diagnoses in adjusted models when the phototherapy-exposure variable was removed. Did not have data on the duration or intensity of phototherapy.	Neonatal phototherapy treatment is associated with a slight risk in childhood seizures diagnosed after ≥60 d postnatal following adjustment for bilirubin measurements. Boys are more likely than girls to both receive phototherapy treatment of hyperbilirubinemia and to be diagnosed with childhood seizure.

Citation Country	Methods	Participants or Inclusion Criteria	Exclusion Criteria	Interventions	Outcomes	Results	Risk of Bias	Notes or Conclusions
Asthma and Allergies
Aspberg 2007 Sweden	Retrospective population-based registry study	14 803 children born between 1987 and 1999 and recorded in the Swedish Hospital Discharge Register as hospitalized with asthma between 2 y of age and 2001 were compared to all children in the Swedish Medical Birth Registry (n = 1 386 029) for those same years.	Children younger than 2 y of age at the time of hospital admission for asthma were excluded because of uncertainty of asthma diagnosis at <2 y.	Multiple maternal and infant characteristics including phototherapy exposure were studied.	Hospitalization for childhood asthma after age 2 y.	Phototherapy-associated odds ratio (Mantel-Haenszel) for asthma requiring hospitalization (OR: 1.27, 95% CI: 1.08–1.5) after excluding infants with risk factors (except for icterus or jaundice) associated with both asthma and phototherapy (measured confounders).	Unable to adjust for bilirubin levels when estimating associations between phototherapy and asthma. Evaluated multiple risk factors (multiple comparisons).	Hyperbilirubinemia and phototherapy were both associated with increased odds of a childhood asthma hospitalization after age 2 y. However, it was impossible to distinguish the phototherapy-associated effect on asthma from that of jaundice.
Aspberg 2010 Sweden		Perinatal data for singleton children prescribed antiasthmatic medication (n = 61 256) were compared to corresponding data for all singleton children born in Sweden from January 1990 to June 2003 and surviving through June 2005 (n = 1 338 319). Data from 3 large Swedish health registers: Swedish Medical Birth Register, Swedish Prescribed Drug Register, and Swedish Hospital Discharge Register.		Numerous antenatal and neonatal variables including icterus and phototherapy.	Asthma as defined by singleton children who received an asthma medication.	Association of phototherapy with: All childhood asthma (OR 1.35; 95% CI: 1.3–1.4). Asthma stated in prescription (OR 1.42; 95% CI: 1.3–1.5). Asthma on systemic corticosteroids (OR 1.47; 95% CI: 1.2–1.8). A diagnosis of icterus (jaundice) was also associated with asthma (OR 1.45, 95% CI: 1.4–1.5).	Adjusted for birth year, maternal age, parity, smoking during pregnancy, and years of involuntary childlessness. Evaluated multiple risk factors (multiple comparisons). There was no adjustment for degree of hyperbilirubinemia. Phototherapy duration and intensity could not be evaluated.	Phototherapy and jaundice were both associated with a diagnosis of asthma. However, it was impossible to distinguish the phototherapy-associated effect on asthma from that of jaundice.
Ku 2012 Taiwan	Retrospective cohort	11 321 children born 1 997–2000 and randomly selected from the Taiwanese National Health Insurance Research Database (NHIRD).	Subjects with questionable basic data (not precisely defined) were excluded.	Those with neonatal jaundice (exposed) were classified as the study group or icteric children. Those without (nonexposed) were classified as the control group or nonicteric children.	Asthma defined as: at least 4 outpatient or emergency department diagnoses of asthma between ages 1–10 y, or 1 asthma diagnosis during a hospital admission between ages 1–10 y.	Phototherapy-associated odds ratio (Mantel-Haenszel) after excluding children with 5 risk factors associated with asthma and phototherapy (confounders) (preterm or low birth wt, neonatal infections, other respiratory conditions, other birth conditions, and sex). (OR: 1.03; 95% CI 0.72–1.4; P = .958.) Jaundice ICD-9 diagnosis-associated odds of asthma after exclusion of children with the above risk factors (OR: 1.64; 95% CI 1.36–1.98; P < .001).	Unable to adjust for bilirubin concentrations when estimating associations between phototherapy and asthma.	Hyperbilirubinemia but not phototherapy was associated with increased odds of a childhood asthma diagnosis after age 1 y.
Sun 2013 Taiwan	Retrospective cohort	11 328 children randomly selected from the Taiwanese National Health Insurance Research Database (NHIRD) and born between 1997 and 2000.	Incomplete basic data, such as conflicting sex and uncertain birth date.	Main exposure of interest was neonatal jaundice as measured by ICD code. Phototherapy was also recorded as a proxy for bilirubin concentration >15 mg/dL, at which phototherapy charges were reimbursed.	Childhood allergic rhinitis between birth and 10-y of age. Diagnostic criteria for allergic rhinitis were at least 3 outpatient allergic rhinitis diagnoses, 1 during a hospitalization, or 1 in an emergency department.	Association between allergic rhinitis and: Phototherapy (RR 1.06; 95% CI: 0.89–1.27) (Risk difference 2.7; 95% CI: -4.8 to 10.1) Neonatal jaundice (aOR 1.46; 95% CI: 1.2–1.7).	Adjusted for the following confounders: preterm/low birth wt, neonatal infection, other respiratory conditions, other birth conditions, and sex. Unable to adjust for bilirubin concentration. Phototherapy duration and intensity could not be evaluated.	There was no association between phototherapy and childhood allergic rhinitis. Neonatal jaundice was associated with a childhood allergic rhinitis diagnosis before age 10 y.
Das 2015	Systematic review and meta-analysis	7 observational studies (n = 101 499 total participants) that included infants with hyperbilirubinemia and/or those receiving phototherapy in the neonatal period and who were followed up to 12 y. Search terms: [(newborn OR infant OR neonate*) AND (hyperbilirubinemia OR icterus OR jaundice) AND (phototherapy OR therapy OR treatment) AND (allergy OR atopy OR asthma OR allergic rhinitis OR rhinitis OR rhino-conjunctivitis OR hay fever OR atopic dermatitis OR allergic eczema)].	Children with outcome assessment at <1 y of age or those with chronic illness, congenital malformation, or kernicterus.	Neonatal hyperbilirubinemia (cutoff level not defined). Neonatal phototherapy.	Allergic rhinitis or conjunctivitis. Asthma. Eczema.	Hyperbilirubinemia: Asthma (OR: 4.26; 95% CI: 4.05–4.5). Allergic rhinitis (OR: 5.37; 95% CI: 4.16–6.92). Phototherapy: Asthma (OR: 3.81; 95% CI: 3.53–4.11). Allergic rhinitis (OR: 3.04; 95% CI: 2.13–4.32).	Fixed-effects logistic regression model was used for meta-analysis of the individual study results. Did not identify whether studies adjusted for bilirubin levels when estimating associations between phototherapy and allergy or asthma outcomes. Did not evaluate duration of phototherapy exposure. Only 2 of the 7 included studies (Aspberg 2007; Ku 2012) evaluated the effect of phototherapy.	Hyperbilirubinemia and phototherapy were both associated with increased odds of asthma and allergic rhinitis. However, it was impossible to separate the hyperbilirubinemia-associated effect from that of phototherapy. Risk differences and number needed to harm were not calculated.
Wei 2015 Taiwan	Nested case-control	27 693 patients within the Taiwanese National Health Insurance Research database (NHIRD) with a new diagnosis of neonatal jaundice from 2000 to 2007. For each child with neonatal jaundice, 2 children without neonatal jaundice from the same period were randomly selected and included in the nonneonatal jaundice cohort (n = 55 367) Patients with nonneonatal jaundice were frequency-matched with patients with neonatal jaundice according to sex, age, parental urbanization, parental occupation, and index date.	Children >1 mo of age, diagnosed with allergic disease before the index date of neonatal jaundice diagnosis, and those missing information on sex or age.	The main exposure of interest was neonatal jaundice as diagnosed by ICD code. Other diagnoses including phototherapy were examined within the cohort of infants with neonatal jaundice.	Allergic conjunctivitis, allergic rhinitis, atopic dermatitis, asthma, and urticaria as diagnosed by ICD code by the end of 2008	Association between phototherapy in infants with neonatal jaundice and: Allergic conjunctivitis (aHR 1.38; (95% CI: 1.3–1.5). Allergic rhinitis (aHR 1.37; 95% CI: 1.3–1.4) Asthma (aHR 1.21; 95% CI: 1.2–1.3) Atopic dermatitis (aHR 1.22; 95% CI: 1.1–1.3) Urticaria (aHR 1.12; 95% CI: 1.03–1.2). All 5 allergic disease were increased in the neonatal jaundice cases as compared to the group of matched nonjaundiced patients).	Analyses were adjusted for age, sex, comorbidities of fetal and newborn respiratory conditions, infections, prematurity, low birth wt, other birth conditions, and glucose-6-phosphate dehydrogenase deficiency. There was no adjustment for degree of hyperbilirubinemia. Phototherapy duration and intensity could not be evaluated.	Phototherapy and a diagnosis of neonatal jaundice were both associated with 5 childhood allergic diseases (allergic conjunctivitis, allergic rhinitis, asthma, allergic dermatitis, and urticaria). It was not possible to differentiate the effects of hyperbilirubinemia and effects of phototherapy on these allergic outcomes.
Egeberg 2016 Denmark	Retrospective cohort	All children (n = 673 614) born in the 10-y period between 1997- 2006 and recorded in the Danish National Patient Register.	As per inclusion criteria.	Neonatal blue light phototherapy, neonatal jaundice, birth gestation, birth wt as recorded in the Danish National Patient Register.	Primary end-point was the first occurrence of atopic dermatitis measured at first, second, and fifth birthdays as recorded in the Danish National Patient Register.	Association between atopic dermatitis and: Phototherapy at first birthday (aIRR 1.11; 95% CI: 0.92–1.35). Phototherapy at second birthday (aIRR 1.01; 95% CI: 0.88–1.15). Phototherapy at fifth birthday (aIRR 0.97; 95% CI: 0.88–1.08). Neonatal jaundice at first birthday (aIRR 1.13; 95% CI: 1.06–1.21). Neonatal jaundice at second birthday (aIRR 1.13; 95% CI: 1.08–1.18). Neonatal jaundice at fifth birthday (aIRR 1.12; 95% CI: 1.08–1.16).	Adjusted IRRs were mutually adjusted for age, sex, neonatal phototherapy, neonatal jaundice, birth wt, gestation, and season of birth. P values corrected by the Bonferroni method with corrected 2-tailed P value <0.008 considered significant. Children were followed from birth until either first occurrence of atopic dermatitis, migration, death, or (in 3 separate analyses) their first, second, or fifth birthday.	Although neonatal jaundice was linked to an increased rate of childhood atopic dermatitis, neonatal phototherapy was not associated with an atopic dermatitis diagnosis.
Kuzniewicz 2018 Northern California, USA	Retrospective cohort	109 212 newborns born during 2010-2014 at ≥35 wk’ gestation. 11 Northern California Hospitals (Kaiser Permanente [KP]) that employed universal TSB screening before discharge.	Infants who did not remain in the same hospital system during their birth admission, infants who did not remain in the KP health plan for at least 25 mo, and infants without TSB levels.	Phototherapy (defined as a recorded nursing phototherapy flow sheet or both a phototherapy procedure code and an order for phototherapy).	Asthma (diagnosed after 2 y of age: defined as 1) at least 2 asthma diagnoses from any outpatient or inpatient encounter separated by ≥30 d; and 2) at least 2 asthma medication prescriptions in a 12-mo period separated by ≥30 d).	Association of phototherapy with asthma: (multivariable Cox-model adjusted hazard ratio [aHR]: 1.01; 95% CI: 0.92–1.11). (Propensity-score aHR: 1.07; 95% CI: 0.96–1.20.) Relative to TSB levels of 3-5.9 mg/dL, infants with maximum TSB levels between 9-17 mg/dL were at a significantly increased hazard for asthma. However, there was not a significant increase in asthma associated with maximum TSB ≥18 mg/dL (no TSB dose-effect) (aHR: 1.04; 95% CI: 0.90–1.20).	Cox proportional hazard models used for primary multivariable analysis to account for differing follow-up times by infant. Among infants within 3 mg/dL of 2009 AAP phototherapy threshold, propensity scores for receiving phototherapy were used to adjust for TSB levels before, but not after, receipt of phototherapy. Unable to measure phototherapy duration. However, hospital phototherapy treatment was compared to home phototherapy and there was no difference in the outcome.	Phototherapy use was not associated with asthma, and phototherapy is unlikely to increase or decrease asthma risk. Modest levels of hyperbilirubinemia were associated with an increased risk of asthma, but this association was not observed at higher bilirubin levels.
Tham 2018 Singapore	Prospective cohort	Growing Up in Singapore Toward healthy Outcomes (GUSTO) study	Per inclusion criteria.	Neonatal phototherapy for neonatal hyperbilirubinemia during the birth hospitalization as recorded for the GUSTO study.	Diagnosis of eczema, wheezing or use of nebulizer or inhaler, and rhinitis by parental survey Atopic sensitization was assessed through skin prick testing.	Association of phototherapy with: Allergen sensitization by month 60 (aOR 1.1; 95% CI: 0.6–2.0) Eczema by month 60 (aOR 1.3; 95% CI: 0.7–2.3). Rhinitis by month 60 (aOR 1.0; 95% CI: 0.6–1.9). Early onset of wheeze and use of nebulizer or inhaler by month 60 (aOR 0.6; 95% CI: 0.3–1.2). No associations were found between bilirubin levels and these measured outcomes.	Adjusted for sex, ethnicity, breastfeeding levels, family history of allergy, gestational age, maternal education levels and mode of delivery. Parental surveys at risk for recall bias.	No association was found between either phototherapy or bilirubin level and allergic sensitization, eczema, rhinitis, wheezing or nebulizer or inhaler use.
Kuniyoshi 2021	Systematic review and meta-analysis	The systematic review included 19 observational studies (cohort, case-control, cross-sectional) that evaluated the association of neonatal jaundice and neonatal phototherapy with childhood allergic diseases. 14 studies were included in the meta-analysis for jaundice and 8 studies in the meta-analysis for phototherapy.	Case-reports, case-series, systematic reviews, and meta-analyses were excluded. 5 studies were excluded from meta-analysis because no effect-estimate was reported.	Neonatal jaundice and phototherapy. For studies that stratified bilirubin levels, results with TSB >15 mg/dL were used.	Childhood-onset (≤19-y) allergic diseases (asthma, atopic dermatitis, allergic rhinitis, or food allergies).	Association of phototherapy with: Asthma (OR 1.24; 95% CI: 1.1–1.4). Atopic dermatitis (OR 1.31; 95% CI: 1.2–1.4). Allergic rhinitis (OR 1.38; 95% CI: 0.9-2.0). Association of jaundice with: Asthma (OR 1.46; 95% CI: 1.4–1.5). Atopic dermatitis (OR 1.3; 95% CI: 1.1–1.6). Allergic rhinitis (OR 3.01; 95% CI: 0.9–10.3).	Random-effects models weighted by inverse variance estimates were used for meta-analysis. Did not evaluate intensity or duration of phototherapy exposure.	Jaundice and phototherapy were both associated with increased odds of asthma and atopic dermatitis. Jaundice, but not phototherapy, was associated with allergic rhinitis. No studies were found that quantitatively assessed the association between either jaundice or phototherapy and food allergies. It was impossible to separate the jaundice-associated effect from that of phototherapy.
Autism Spectrum Disorder
Wu 2016 Northern California, USA	Retrospective Cohort Study	457 855 infants born at ≥35 wk’ gestation who were recorded in the Kaiser Permanente of Northern California (KPNC) Database (1995–2011).	Infants who died during the birth hospitalization, were transferred out of the KPNC system, or were followed <60 d. Infants with at least 2 inpatient or outpatient physician ICD diagnoses of: Trisomy 21, other genetic disorders, or congenital anomalies diagnosed at <15 d.	In-hospital or home phototherapy.	ASD diagnosis either by ASD evaluation center, a clinical ASD specialist outside the ASD center (ie, psychiatrist, psychologist, or developmental pediatrician), or by a general pediatrician. Hyperbilirubinemia was also not associated with ASD after adjustment for confounders.	Phototherapy (aHR: 1.10; 95% CI: 0.98–1.24) was not associated with ASD in the primary analysis (Cox proportional hazards model). Propensity-adjusted sensitivity analyses also revealed no association between phototherapy and ASD (aHR: 1.09; 95% CI: 0.95–1.24). This model examined the effect of phototherapy among only infants who had a TSB level within 3 mg/dL of the 2004 AAP phototherapy threshold. TSB ≥20 (aHR: 1.09; 95% CI: 0.89–1.35) was also not associated with ASD.	Propensity-analyses adjusted for bilirubin level before phototherapy. Unable to measure phototherapy duration or intensity.	After adjustment for the confounding effects of sociodemographic factors, neither phototherapy nor hyperbilirubinemia was a significant independent risk factor for ASD.
Breastfeeding and Familial Bonding
Kemper 1989 Yale-New Haven Hospital, New Haven, Connecticut, USA	Prospective observational study with matched controls	101 jaundiced infants born in 1987-1988 with a total serum bilirubin >11.99 mg/dL measured before discharge. 155 comparison subjects who were not visually jaundiced and who were seen by the same pediatrician as the matched, jaundiced infant.	Birth wt <2500 g or if infant spent >6 h in the ICU. Comparison group infants were excluded if a bilirubin level was obtained anytime up to 1 mo postnatal.	Diagnosis of jaundice (mean peak bilirubin in jaundiced infants was 16.2 mg/dL, and 55% received phototherapy).	Mothers were surveyed at 1 mo postnatal to assess (1) concerns about jaundice and phototherapy, (2) termination of breastfeeding within 1 mo postnatal, and (3) maternal behaviors consistent with the vulnerable child syndrome.	Of mothers whose infants received phototherapy, 70% thought its use implied that their child was moderately to seriously ill and 74% found its use upsetting. 53% of mothers whose infants received phototherapy vs 31% of those with jaundiced, non–phototherapy-treated infants had ever left their child with someone else for >1 h at 1 mo postnatal. 64% of those with breastfeeding interrupted in the hospital vs 36% without interruption had completely stopped breastfeeding at 1 mo (P < .05). 84% of mothers of jaundiced infants completed the 1-mo survey, as did 80% of mothers of comparison group infants.	Risk of recall bias (survey). Difficult to determine effect sizes for phototherapy since the main focus of the study was a jaundice diagnosis versus no jaundice diagnosis. Phototherapy specific data are presented as proportions for some outcomes. Although some baseline variables (potential confounders) were collected, these were not used to adjust estimates between jaundiced infants receiving phototherapy and jaundiced nontreated infants.	Mothers of infants who had a jaundice diagnosis were more concerned about their infant’s overall health at 1-mo postnatal. Maternal concerns were increased if their infant received in-hospital phototherapy. It should be noted that clinical practice (including increased phototherapy utilization) and parental counseling (on phototherapy and breastfeeding) may have changed over the 28 y since the study was published.
Schedle 1990 County Hospital (Kantonsspital), Aarau, Switzerland	Prospective cohort	107 healthy, term infants, 1986-1987.	Pregnancy complication, >10th percentile wt for gestation, congenital malformation, or neonatal illness.	Three group comparison: (1) 29 jaundiced infants with mean peak TSB of 15.85 mg/dL who received 2–3 d of phototherapy and partial mother-child separation, (2) 40 mildly jaundiced infants with mean peak TSB of 12.22 mg/dL who received neither phototherapy nor separation, and (3) a control group of 38 infants with no apparent jaundice. Measured baseline variables were similar in all groups except the phototherapy group had significantly more boys relative to the other 2 groups (X² = 0 .48; P = .009).	Denver Developmental Screening Test and Ainsworth Attachment Scores at 1 y of age.	There were no differences in Denver scores for social contact (X² = 0.1; P = .93), gross motor behavior (X² = 0.5; P = .77), and fine motor and adaptive performance (X² = 4.5; P = .10) at 1 y of age. No significant difference in the three study groups was found for mother-child attachment at 1 y of age (X² = 1.5; P = .83).	Were unable to completely differentiate effect of bilirubin and phototherapy since mean TSB was lower in the jaundiced non–phototherapy- treated group. Small-sample size.	Treatment with phototherapy as a newborn did not appear to affect neurodevelopment, as measured by the Denver Developmental Screening Test at 1-y of age and by maternal-infant attachment at 1 y of age.
Usatin 2010 Northern California, USA	Retrospective cohort	Infants born between 1995 and 2004 at gestation ≥36 wk with birth wt ≥2 kg. Patients were eligible if they (1) never had a TSB ≥12 mg/dL (n = 128 417) or (2) when they had a TSB ≥17 to <23 mg/dL (a range where phototherapy might be discretionary) as outpatients at 2 to 7 d postnatal and did (n = 1765) or did not (n = 6777) receive inpatient phototherapy after their initial birth admission. Kaiser Permanente of Northern California Database.	Newborns that received phototherapy during their initial birth admission, those with birth admission >48 h, or a conjugated or direct bilirubin ≥2 mg/dL in the first 30 postnatal days. Those that were readmitted to the hospital for any reason with a length of stay ≥96 h within the first 14 postnatal days. Those with <4 outpatient visits in the first year and/or no visits in the second year.	The 3 study comparison groups: 1) never had a TSB ≥12 mg/dL, 2) TSB ≥17 to <23 mg/dL without phototherapy, and 3) TSB ≥17 to <23 mg/dL with phototherapy.	Primary outcome variables: Total outpatient visits during postnatal days 15–364 and total numbers of various types of outpatient visits at ages (in days): 15–59, 60–119, 120–179, 180–364.	There were small increases in total year-1 visits for jaundiced infants without phototherapy (group 2) (adjusted incident rate ratio [aIRR]: 1.04; 95%: CI: 1.02–1.05) compared to infants with bilirubin <12 mg/dL (group 1). There were also slightly more total year-1 visits for jaundiced infants with phototherapy (group 3) (aIRR: 1.03; 95%: CI: 1.01–1.06) relative to those with jaundice and phototherapy (group 2). For both phototherapy and untreated groups of infants with bilirubin ≥17 to <23 mg/dL, increases in visits were greatest during days 15–59, appeared related to follow-up for the diagnosis of jaundice itself, and were not statistically significant in any other measured year-1 time periods. The adjusted increase in visits over group 1 (bilirubin <12 mg/dL) were 0.36 visits in year-1 by group 2 (jaundice and no phototherapy) and 0.73 visits by group 3 (jaundice and phototherapy). The average number of outpatient visits for the 8 most common sick-visit diagnostic codes did not vary by phototherapy use.	Poisson regression was used to account for potentially confounding variables including mother’s age, infant’s year, month, and hospital of birth, gestational age, sex, and race. TSB level that qualified for groups 2 or 3 were dichotomized at 17–19.9 and 20–22.9 mg/dL and used to adjust comparisons between group 2 and group 3 (phototherapy exposure). The authors noted that estimates should be interpreted in light of no correction for multiple comparisons.	Small increases in outpatient visits before age 1 were seen with both neonatal jaundice (+0.36) and phototherapy (+0.73) that most often occurred in the first 2 postnatal months. Phototherapy treatment and jaundice diagnoses were associated with only small increases in first-year outpatient visit rates, consistent with mild or infrequent contribution to the vulnerable child syndrome.
Waite 2016 United States of America	Retrospective cohort	4441 ever-breastfed infants born >35 wk’ gestation at ≥2200 g between May-December 2005 and recorded in the Centers for Disease Control and Prevention Infant Feeding Practices Study II (IFPS II). Infants who received treatments for jaundice other than phototherapy (eg, formula) were excluded.	Breastfeeds were never initiated or if mothers answered via survey that their infants received “other treatment for jaundice,” which included switching completely to formula feeds.	Phototherapy for the treatment of jaundice as indicated by parental report via survey at ∼3 wk postnatal. Parents first indicated that their child was ever jaundice and if so, they were asked whether treatment was required with phototherapy being 1 of the treatment options.	Primary outcome variables: Any breastfeeding at 1, 2, 4, 6, 9, and 12 mo postnatal based on maternal survey response. Secondary outcome variables: Exclusive breastfeeding at 1, 2, and 4 mo postnatal based on maternal survey response.	At 1-mo postnatal there was no increased odds of any breastfeeding (aOR: 1.14; 95% CI: 0.71–1.81) in phototherapy-exposed infants. No differences in any breastfeeding were noted at months 2, 4, 6, and 9 postnatal. At 12 mo postnatal, phototherapy exposure was associated with a reduced odds of any breastfeeding (aOR: 0.58; 95% CI:0.37–0.91). There was a reduced odds of exclusive breastfeeding in phototherapy-exposed infants (aOR: 0.69; 95% CI: 0.49–0.95) at 1-mo postnatal. The reduced odds of breastfeeding in phototherapy-exposed infants persisted at 2 and 4 mo postnatal.	Potential for recall bias due to determination of phototherapy exposure and breastfeeding history via postnatal maternal survey. The first postnatal survey at ∼3 wk asked the respondent whether the infant had jaundice at any time since birth. Mothers that answered that their baby had jaundice were then asked about phototherapy treatment. It is possible that some mothers may not have understood that jaundice and high bilirubin are synonymous terms.	Phototherapy treatment of hyperbilirubinemia may be associated with a decreased odds of exclusive breastfeeding during months 1–4 postnatal. Phototherapy was not associated with a change in the odds of any breastfeeding. Although there was a decrease in the odds of any breastfeeding at 12 mo postnatal, it seems unlikely that phototherapy was the cause given the one-year duration between phototherapy-exposure and change in breastfeeding habits. Unclear whether among jaundiced infants receiving treatment phototherapy has a more adverse effect on breastfeeding than other interventions.
Digitale 2021 Northern California, USA	Retrospective cohort	25 853 infants born at ≥35-wk’ gestation with TSB levels from 3 mg/dL below to 2.9 mg/dL above 2004 AAP phototherapy thresholds. Medical records at 16 Northern California Hospitals (Kaiser Permanente [KP]) from 2013-2017.	Infants with missing data on length of birth hospital stay, TSB <3 mg/dL, or TSB not measured were excluded. Infants whose first TSB higher than 3 mg/dL below phototherapy threshold was ≥3mg/dL above threshold excluded from primary analysis because phototherapy is no longer optional at those higher TSB concentrations.	Inpatient phototherapy during birth hospitalization defined as presence of phototherapy flow sheet in chart or both ICD procedure code and phototherapy order.	Exclusive breast mlk feeding via caregiver self-report at 2-mo well-child visit. Reporting feeding infant any breast milk at 2-mo well-child visit.	No association between phototherapy and exclusive breast milk feeding at 2-mo of age (aRR 0.99; 95% CI: 0.95–1.04). The adjusted estimate of the difference between observed breast milk feeding among those who received phototherapy (average treatment effect on the treated) (ATET) was -0.2% (95% CI: -2.0% to 1.5%) less than that estimated if those patients had not received phototherapy. Any breast milk feeding at 2-mo: aRR 1.02 (95% CI: 1.0–1.04) ATET for any breast milk feeding at 2-mo among those who received phototherapy: +1.6% (95% CI: 0.1%–3.1%)	Emulated a randomized trial by only including infants with a TSB level for which phototherapy was potentially indicated but not essential. Used a directed acyclic graph (DAG) to illustrate modeling assumptions. Adjusted for potential confounding variables using a modified Poisson regression model. Conducted sensitivity analyses by: (1) including infants excluded from primary analysis because of high initial TSB; (2) imputing missing data to include infants initially excluded from primary analysis; (3) analyzing infants who received formula before first TSB; and (4) analyzing infants who formula fed before first TSB. Estimates from sensitivity analyses were similar to primary analysis estimates.	Phototherapy during the birth hospitalization was not associated with reduced breast milk feeding at the 2-mo well-child visit. These results may not apply to hospital systems that do not actively support maternal lactation during photherapy. Therefore, lactation support should be encouraged.
Infant Behavior
Telzrow 1980 Boston Hospital for Women, Boston, Massachusetts, USA	Observational study with matched controls	10 low-risk term, nursery infants who subsequently received phototherapy for high bilirubin were compared to 10 similar infants who did not receive phototherapy. Sex, maternal breastfeeding intent, and rooming-in proportions were similar between groups.	Infants with intrapartum or postnatal abnormalities, including infants of Rh-positive mothers, were excluded.	Phototherapy at clinician discretion (high bilirubin was defined as >9 mg/dL before 48 h postnatal and >11 mg/dL before 72 h).	Brazelton Scale scoring at 3, 6, and 10 d postnatal. (Phototherapy was started at an average age of 76.5 h).	Visual orientation scores were lower in the phototherapy-treated group relative to the nontreated comparison group on postnatal days 3 (P < .01), 6 (P < .01), and 10 (P < .05). Motor scores were lower on day 6 (P < .05).	The authors were unable to separate phototherapy-specific effects from bilirubin-attributable or separation-attributable effects. It was not possible for the examiners to be blinded to treatment group Separate components (26 separate components and 6 cluster-group scores) were assessed within the Brazelton tests (multiple comparisons). Difficult to determine an estimate of effect-size. Decision to initiate phototherapy was at clinician discretion. Mothers were discharged at 3 d. Examinations at 6 d were performed for the phototherapy group within the hospital and for the comparison group at home. Interrater reliability coefficient was >0.85.	Infants that received phototherapy may have had short-term changes in visual orientation. Results may not apply to infants whose eyes are not covered during phototherapy or who are not separated from their parents.
Paludetto 1983 Newborn Nursery of Second School of Medicine; Naples, Italy	Prospective observational study with matched controls	30 term infants undergoing phototherapy for 6 h or more for jaundice (mean bilirubin, 13.3 mg/dL) and 30 comparison subjects matched for sex, gestation, Apgar score, obstetric history, and father’s profession. Enrolled on postnatal day 3.	Infants transferred to special care nursery or with any neonatal illness including Rh and ABO mismatch.	Phototherapy for 6 h or more for jaundice. (Treated with phototherapy in the nursery and the mothers fed them every 3 h versus rooming-in in the non–phototherapy-treated comparison group.)	Brazelton Scale scores at (1) postnatal day 3 (day of enrollment) (treated infants were under phototherapy an average of 24 h at first examination), (2) 24 h postenrollment, and 3) at 1 mo of age.	Visual orientation scores were lower in the phototherapy-treated group relative to the nontreated comparison group on postnatal days 3 (P < .005) and 4 (P < .01) and at 1 mo of age (P < .05).	The mean total bilirubin levels were lower in the comparison group (9.6 [range, 3.5–14.3]) relative to the phototherapy group (13.3 [range, 8.4–17.5]). The authors were unable to separate phototherapy-specific effects from bilirubin-attributable or separation-attributable effects. 26 separate components were assessed within the Brazelton tests (multiple comparisons). Difficult to determine an estimate of effect-size. Although there were general newborn unit phototherapy guidelines, clinicians were free to decide whether to start phototherapy. Only 12 matched pairs were available for 12-mo scoring follow-up. Brazelton scorers were unaware of treatment assignment, but might have noted different levels of visible skin jaundice in subjects.	Infants that received phototherapy may have had short-term changes in visual orientation. Results may not apply to infants whose eyes are not covered during phototherapy or who are not separated from their parents.
Cancer
Cnattingius 1995 (June) Sweden	Nested case-control study	613 cases of lymphatic leukemia were identified in the Swedish National Cancer Register and the Medical Birth Register between 1973-1989. 5 controls were matched to every case by sex and birth month and year.		93 total antenatal and perinatal exposures were examined including phototherapy and neonatal jaundice. Diagnoses and procedures determined by ICD codes.	Lymphatic (ie, acute, chronic, or unspecified lymphocytic) leukemia.	Association of childhood lymphatic leukemia with: Phototherapy (OR 1.0; 95% CI: 0.5–1.8). Physiologic jaundice (OR 1.2; 95% CI: 0.8–1.7).	Given number of studied exposures, the authors estimated 4 or 5 significant associations would be detected by chance alone.	Neither the use of phototherapy or neonatal jaundice was found to increase the odds of childhood lymphatic leukemia. The authors were unable to adjust for degree of hyperbilirubinemia or phototherapy intensity or duration.
Cnattingius 1995 (July) Sweden	Nested case-control study	98 cases of myeloid leukemia were identified in the Swedish National Cancer Register and the Medical Birth Register between 1973 and1989. 5 controls were matched to every case by sex and birth month and year.		Numerous antenatal and perinatal exposures were examined including phototherapy and neonatal jaundice. Diagnoses and procedures determined by ICD codes.	Myeloid (acute, chronic, or unspecified) leukemia.	Association of childhood myeloid leukemia with: Phototherapy (OR 7.5; 95% CI: 1.8–31.9). Physiologic jaundice (OR 2.5; 95% CI: 1.2–5.0). Reanalysis excluding cases with Trisomy 21: Phototherapy (OR 4.3; 95% CI: 0.9–21.9). Physiologic jaundice (OR 1.7; 95% CI: 0.8–4.0).	Given number of studied exposures, significant associations might be detected by chance alone.	Phototherapy and neonatal jaundice were associated with an increased odds of childhood myeloid leukemia. However, these associations were no longer statistically significant after excluding children with Trisomy 21 diagnoses. The authors were unable to adjust for degree of hyperbilirubinemia or phototherapy intensity or duration.
Olsen 1996 Denmark	Retrospective cohort	66 430 neonates discharged with a history of neonatal hyperbilirubinemia between 1977 and 1989 and recorded in the Danish Central Hospital Discharge Register.	Supplementary diagnoses of neonatal immaturity or hemolytic disease.	Diagnosis of hyperbilirubinemia by ICD code.	Diagnosis of cancer within the Danish Cancer Registry.	Among infants with hyperbilirubinemia diagnoses, 87 cancers were observed compared to 85 expected yielding standardized incidence [SIR] ratio=1.0; 95% CI: 0.8-1.3. Leukemia (SIR = 1.2; 95% CI: 0.8-1.7)	Phototherapy diagnoses were not recorded. It was estimated that 85% to 50% of infants in the study received phototherapy based on a random sample of 150 neonates with hyperbilirubinemia in Copenhagen in 1981. Average follow-up of 9.1 y (range 0–15 y) (= 499 502 person-years).	Hyperbilirubinemia was not associated with childhood cancer. The authors were not able to directly estimate the effect of phototherapy on cancer.
Berg 1997 Sweden	Case-control	30 cases of childhood malignant melanoma between 1973 and 1992 and 120 controls matched by same date of birth, same hospital, and same sex.	Per inclusion criteria.	Neonatal phototherapy as determined by Swedish Medical Birth Registry.	Childhood malignant melanoma as determined by Swedish Cancer Registry.	None of the infants diagnosed with malignant melanoma had received phototherapy compared to 11 of 120 controls (one-tailed P value as calculated in study = 0.08) (Odd ratio = 0; 95% CI: 0–1.2). Calculated 2-tailed P value = 0.156).	The average follow-up time was 18 y (range: 10-19). Phototherapy estimates were not adjusted for degree of hyperbilirubinemia or phototherapy intensity or duration.	No association was detected between phototherapy and childhood malignant melanoma.
Roman 1997 Southern England, UK	Case-control	177 cases with Leukemia or non-Hodgkin’s lymphoma and 354 age and sex-matched controls identified from the medical records of 3 hospitals.	Infants from multiple gestation pregnancies who died before birth hospital discharge, or with identified chromosomal anomalies or malformations.	Multiple antenatal and perinatal risk factors including phototherapy and neonatal jaundice, as extracted from hospital medical records.	Leukemia or non-Hodgkin’s lymphoma diagnosed between the age of 3 mo-30 y. Cases for children (0-14 y) extracted from the Childhood Cancer Research Group (1962–1992). Cases for young adults (15–29 y) extracted from Office of National Statistics (ONS) (1972-1987).	Association between any leukemia and: Phototherapy (OR 0.5; 95% CI: 0.1–2.3). Jaundice (OR 0.8; 95% CI: 0.5–1.5). Additionally, no associations were detected between jaundice or phototherapy and acute lymphocytic leukemia, acute myeloid leukemia, or non-Hodgkin’s lymphoma).	Phototherapy estimates were not adjusted for degree of hyperbilirubinemia.	Phototherapy and jaundice were not associated with leukemia or non-Hodgkin’s lymphoma in young adults.
Podvin 2006 Washington, USA	Case-control	595 cases born between 1980-2002 were identified from the Washington State Cancer Registry or Cancer Surveillance System of Western Washington. Controls (n = 5950) were randomly selected from birth certificate records of infants without leukemia and frequency of controls were matched to cases by year.	Per inclusion criteria.	Multiple antenatal and perinatal risk factors including phototherapy and neonatal jaundice. Jaundice diagnoses were documented on birth certificates before 1992. After 1987, ICD code-based phototherapy and neonatal jaundice diagnoses were available.	Childhood leukemia diagnosis (age<20).	Phototherapy (aOR 2.2; 95% CI: 1.0-4.9) Neonatal jaundice as documented by ICD code (aOR 1.4; 95% CI: 1.0–1.9) Neonatal jaundice as documented on birth certificate (aOR 2.1; 95% CI: 1.2–3.8).	Estimates were adjusted for maternal age, gestation, birth wt and race A secondary analysis that excluded infants with Trisomy 21 found similar results. Phototherapy estimates were not adjusted for degree of hyperbilirubinemia.	Phototherapy and neonatal jaundice were associated with childhood leukemia. The authors were unable to adjust for degree of hyperbilirubinemia or intensity or duration of phototherapy.
Brewster 2010 Grampian Region, Scotland, UK	Retrospective cohort	77518 singleton newborn infants born 1976–1990 that survived the neonatal period.	Neonatal deaths.	Received neonatal phototherapy versus no phototherapy.	Standardized incidence ratios for melanoma, basal cell, and squamous cell skin cancer.	Two cases of melanoma occurred in phototherapy-exposed versus 16 cases in unexposed. Standardized incidence ratio of 1.40 (95% CI, 0.17–5.04; P = .834) for melanoma. No cases of squamous cell or basal cell carcinoma of skin were observed in exposed persons.	Limited confounder control for age, sex, calendar period, and socioeconomic measure within a postal code. Bilirubin level unmeasured. Duration of phototherapy unmeasured. Median follow-up of 24 y.	No difference in skin cancer risk was detected in infants exposed to neonatal phototherapy.
Newman 2016 Northern California, USA	Retrospective cohort	449 621 newborn infants born 1995-2011 at ≥35 wk’ gestation who survived to discharge and were followed ≥60 d. Database from 15 Northern California Hospitals (Kaiser Permanente). 1995-2011.	Death or transferred to another hospital before initial hospital discharge. Cancer diagnosis before 60 postnatal days. <60 d of follow-up. Infants born at <35 wk’ gestation.	Received any neonatal phototherapy (in hospital, at home, or both) vs no phototherapy. Dose-response phototherapy variable (0- none, 1-home phototherapy only, 2-phototherapy in single admission, 3- phototherapy in multiple admits).	Cancer diagnoses between postnatal day 60 and the end of the follow-up period. Examined multiple categories of leukemia and other forms/sites of cancer.	Propensity-score adjusted hazard ratio for any cancer after phototherapy: (aHR: 1.0; 95% CI: 0.7–1.6). Lower limit of 95% CI for number needed to harm (NNH) for 10-y cancer risk ∼1100. For children with Down syndrome, lower limit of 95% CI for estimated NNH for leukemia = 23. Propensity-score adjusted hazard ratios for: Any leukemia: (aHR: 1.6; 95% CI: 0.8–3.5). Nonlymphocytic leukemia: (aHR: 1.9; 95% CI: 0.6–6.9). Liver cancer: (aHR 1.4; 95% CI: 0.2–12). In infants with ≥2 phototherapy admissions there was an association between phototherapy and acute myelogenous leukemia (AML) (aHR 8.5; 95% CI 1.9–38) (note: based on only 2 cases of AML with ≥2 readmissions for phototherapy).	Cox regression models including both traditional multivariable models and models incorporating propensity-score adjustment were used to adjust for potential, measured confounders, and duration of follow-up. Multivariable and propensity-score estimates were mutually consistent. Unable to measure total duration of phototherapy exposure. Used a directed acyclic graph (DAG) to illustrate modeling assumptions. Cancer diagnoses were confirmed by a medical record review.	In adjusted analyses, there was no statistically significant elevation in the risk for cancer. Unadjusted analyses did show an elevated risk. Given their elevated baseline risk for leukemia, children with trisomy 21 may be at an even higher risk for leukemia after phototherapy.
Wickremasinghe 2016 California, USA	Retrospective cohort	5 144 849 infants born at ≥35 wk’ gestation who survived >60 d. California Vital Statistics/Patient Discharge Dataset 1998-2007.	Death within 60-d of birth. Cancer diagnosis before 60 postnatal days. Infants born at <35 wk’ gestation.	Receipt of phototherapy during a hospitalization that began within the first 2 postnatal weeks as recorded by ICD-9 code for phototherapy (99.83, 99.83).	Hospital discharge diagnosis of cancer via ICD-9 code between 60-d postnatal and 1 y of age.	Propensity-score adjusted odds ratio for: All cancer: (aOR: 1.4; 95% CI: 1.1-1.9). Myeloid leukemia: (aOR: 2.6; 95% CI: 1.3–5.0). Renal cancer: (aOR: 2.5; 95% CI: 1.2–5.1). Found NNH of 10 638 in the general population and 1285 in children with trisomy 21.	Adjusted for the presence of a jaundice diagnosis, but unable to control for bilirubin concentration, which is also associated with a cancer diagnosis. Phototherapy duration was not measurable. Propensity-score adjustment to adjust for potential, measured confounders. 95% Cis were narrow showing precise estimates.	Adjusted analyses indicated that phototherapy may slightly increase the risk of cancer in the first postnatal year.
Auger 2019 Quebec, Canada	Retrospective cohort	786 998 infants born between 2006 and 2016. Dataset of discharge abstracts from all Quebec hospitals: “Maintenance and Use of Data for the Study of Hospital Clientele Registry of Quebec.”	Cancer diagnosis before age 2 mo or less than 2 mo of follow-up. Infants with missing gestational age or birth wt.	The exposure was defined as a 3-category variable: phototherapy, untreated jaundice, or no exposure Phototherapy at birth or during admissions in the first 28 postnatal days, as recorded by Canadian Classification of Health Interventions procedure codes (1.YZ.12.JA-DQ). Jaundice diagnosis recorded via ICD-10 codes (P58, P59).	Hospitalization for solid or hematopoietic childhood tumors between 2 mo and 11 y of age.	Propensity-score adjusted hazard ratios for: Any cancer: (Phototherapy versus no exposure: aHR: 1.16; 95% CI: 0.9–1.6). (Phototherapy versus untreated jaundice: aHR: 1.07; 95% CI: 0.8–1.5). (Untreated jaundice versus no exposure: aHR: 1.09; 95% CI: 0.9–1.3). Any cancer at 4-–1 y: (Phototherapy versus no exposure: aHR 2.21; 95% CI 1.5–3.3) (Phototherapy versus untreated jaundice: aHR 2.21; 95% CI 1.3–3.7). Late onset solid tumors, between age 4 and 11 y: (Phototherapy versus no exposure: aHR: 2.26; 95% CI: 1.3–3.8) (Phototherapy versus untreated jaundice: aHR: 2.21; 95% CI: 1.2–4.3). The cumulative incidence of childhood cancer was higher for infants with phototherapy (25.1 per 100 000 person-years) and untreated jaundice (23.0 per 100 000) compared to unexposed infants (21.6 per 100 000).	Unable to adjust for bilirubin concentration, which is also associated with a cancer diagnosis. Phototherapy duration was not measurable. Propensity-score adjustment to adjust for potential, measured confounders. Fine-Gray subdistribution hazard model to allow for estimation of probability of cancer over time (cumulative incidence function [CIF]), while accounting for death as a competing risk.	Neonatal phototherapy may be associated with a slightly increased risk of solid tumors in childhood. It was not possible to rule out increasing bilirubin concentration as a potential confounder, since a jaundice diagnosis was also associated with cancer, bilirubin concentrations were not available, and phototherapy would likely be initiated at higher bilirubin levels compared to untreated jaundice. Adjusted hazard ratios for any cancer and for solid tumor cancers, but not for hematopoietic cancers, significantly increased with age for infants exposed to phototherapy.
Seppala 2020 Finland	Case-control	2037 cases with childhood cancer diagnosed from Finnish Cancer Registry between 1996-2014. 5 sex and year-matched controls (n = 10 185) for each case from Finnish Medical Birth Registry.	8 cases and 82 controls due to missing birth wt or gestation data.	Multiple antenatal and perinatal risk factors. Prematurity associated variables were the main exposure variables of interest. Perinatal conditions and procedures, including phototherapy, were secondary exposures of interest.	Childhood cancer (age <20 y) as diagnosed from the Finnish Cancer Registry.	Association between childhood cancer and phototherapy: All newborns (aOR 1.11; 95% CI: 0.91-1.35). Term newborns (aOR 1.17; 95% CI: 0.91-1.49).	Analyses were adjusted for maternal age, maternal smoking, and number of pregnancies. Phototherapy estimates were not adjusted for degree of hyperbilirubinemia or phototherapy intensity or duration.	Phototherapy was not associated with childhood cancer.
Digitale 2021 Northern California, USA	Retrospective cohort	139 100 infants born at >35 wk’ gestation from 1995-2017 with qualifying bilirubin concentration 3 mg/dL below to 4.9 mg/dL above the 2004 AAP phototherapy threshold. Database from Northern California Hospitals (Kaiser Permanente). *Note: This is an extension of the Newman 2016 cohort with an additional 453 786 person-years of follow-up	Patients who died or were transferred to another hospital before initial hospital discharge . Cancer diagnosis before 60 postnatal days or only had a secondary cancer diagnosis. <60 d of follow-up. Infants born at <35 wk’ gestation.	Received any neonatal phototherapy during birth hospitalization or readmission versus no inpatient phototherapy.	Cancer diagnoses between postnatal day 60 and the end of the follow-up period. Required 2 total cancer diagnoses at different departments to reduce false-positive diagnoses. Time of cancer diagnosis defined as date of initial ICD cancer diagnosis. Examined multiple categories of hematopoietic cancers, solid tumors, and other cancers.	Propensity-score adjusted hazard ratio for any cancer afterphototherapy: (aHR: 1.13; 95% CI: 0.8–1.5). Propensity-score adjusted hazard ratios for: Any hematopoietic cancer: (aHR: 1.17; 95% CI: 0.7–1.8). Solid tumors: (aHR: 1.01; 95% CI: 0.7–1.6). Analyses specifically looking for an increase in cancer risk after age 4 y were also negative.	Cox regression models incorporated propensity-score adjustment to adjust for potential, measured confounders, (including bilirubin concentration) and duration of follow-up. Unable to measure total duration or intensity of phototherapy exposure.	Adjusted analyses, detected no significant elevation in the baseline risk for cancer.
Bugaiski-Shaked 2022 Soroka University Medical Center, Be’er-Sheva, Israel	Retrospective cohort	342 172 infants born at ≥32 wk’ gestation born 1998-2018. Perinatal database of obstetric and neonatal discharge summary data merged with hospitalization records from Soroka University Medical Center.	Neonates who died or were diagnosed with any malignancy before 30-d postnatal.	Phototherapy performed during the first 2 wk postnatal as defined by ICD-9 code for phototherapy diagnosis (Z9983).	First pediatric hospitalization with any diagnosis of neoplasm (malignant or benign).	Preterm birth and maternal age adjusted hazard ratios for: Childhood malignancies (aHR 1.89; 95% CI: 1.4–2.7) Benign tumors: (aHR: 1.27; 95% CI: 1.02–1.6) Childhood malignancies in term-infants: (aHR: 1.84; 95% CI: 1.2–2.7)	Unable to adjust for bilirubin concentration when estimating associations between phototherapy and neoplasms. Phototherapy intensity and duration were not measurable. Used Israeli Ministry of Health national cancer registry to cross-reference and verify child malignancy diagnoses that were in database. Median follow-up of 9.5 y (range 0–18 y).	Adjusted analyses indicated that phototherapy may slightly increase the risk of childhood malignancies and benign tumors cancer in the first postnatal year. It was not possible to rule out increasing bilirubin concentration as a potential confounder or to evaluate the impact of phototherapy intensity or duration on outcomes.
Insulin-Dependent Diabetes Mellitus
Dahlquist 2003 Sweden	Case-control	6487 cases born 1973–1997 who were recorded in the Swedish Childhood Diabetes Registry and also linked with the Swedish Medical Birth Registry. 2.8 million births for comparison.	Twins, infants born of diabetic mothers, and infants whose county of birth was unknown,	Phototherapy as recorded in database. Jaundice was also evaluated as an exposure.	A diagnosis of childhood diabetes as recorded in the Swedish Childhood Diabetes Registry.	Adjusted odds of type 1 diabetes in infants who had received phototherapy in 2 high-risk counties with high type I diabetes prevalence: (aOR 1.95; (95% CI: 1.2–3.2). Association between phototherapy and diabetes in all other counties: (aOR 1.06; 95% CI 0.8–1.4). Unadjusted association between therapy and diabetes, irrespective of diagnosis: (OR 3.79; (95% CI: 3.1– 4.6). Unadjusted association between jaundice diagnosis and diabetes: (OR 1.13; 95% CI 1.01–1.3).	Multivariable logistic regression analyses adjusted for birth year, preterm birth, newborn respiratory symptoms, and blood group immunization. Unable to adjust for bilirubin concentration or phototherapy duration or intensity.	Phototherapy was associated with type I diabetes in 2 Swedish counties with high diabetes prevalence, following adjustment for measured confounding variables. This association was not observed in all other Swedish counties in adjusted analyses. The analysis did not attempt to separate the jaundice-associated effect from that of phototherapy.
Newman 2016 Northern California, USA	Retrospective cohort	499 642 infants born at ≥35 wk’ gestation who were recorded in the Kaiser Permanente of Northern California (KPNC) Database (1995–2011).	Infants who died during birth hospitalization, who were transferred, and who were followed <60 d after birth.	In-hospital or home phototherapy.	Type-1 diabetes diagnosis defined as either ≥2 encounters (inpatient or outpatient) with a diagnosis of type 1 diabetes or 1 encounter and either ≥2 hemoglobin A1c levels ≥6.5% or a pharmacy record indicating insulin prescription. Diagnoses obtained from KPNC virtual data warehouse and diabetes registry.	Association between any phototherapy and type 1 diabetes: (aHR 1.06; 95% CI: 0.8 to 1.5). Adjusted incidence rate ratio (aIRR) for phototherapy:1.03 (95% CI: 0.8–1.4) corresponding to an estimated excess risk of 0.58 (95% CI, −4.7 to 5.9) per 100 000 person-years. Propensity-adjusted Cox model-derived association between phototherapy and type I diabetes: Restricted model that included only those with ≥1 TSB between −3 and +4.9 mg/dL from 2004 AAP phototherapy threshold (aHR 0.95; 95% CI, 0.60–1.5) Inclusive model of all subjects with dichotomous variable for any TSB level exceeding 2004 AAP (aHR 0.91; 95% CI: 0.7–1.3).	Propensity-analyses adjusted for bilirubin level before phototherapy. Unable to measure phototherapy duration or intensity.	There was no evidence of increased type 1 diabetes risk in children who had received phototherapy.
Melanocytic Nevi
Bauer 2004 Germany	Cross-sectional study	1812 white children aged 2–7 y attending day care centers in 2 German cities in 1998. Note: this very brief report describes a secondary analysis of a 1998 study (Wiecker et al Cancer. 2003; 97:628-638) designed to examine the association of sun exposure and nevus counts in parents with childhood development of melanocytic nevi.	Per inclusion criteria.	Neonatal phototherapy as determined by parent survey.	Total body nevus counts and pigmentary characteristics were assessed in all participating children (52.9% boys).	Multiple linear regression analysis did not identify an association between neonatal phototherapy and the number of melanocytic nevi in children (P = .21). The number of vacation days in the sun (P < .001) and an at-home outdoors activity score (P = .03) were associated with increased childhood melanocytic nevi.	Limited data are presented in this brief report. Thus, we are unable to calculate measures of effect size. A standardized interview of parents included questions of national origin, sun exposure history, sun protection, and history of neonatal jaundice. Risk of recall bias from the use of survey to determine phototherapy. Limited control for potential confounders at the time of phototherapy due to survey design. Unable to adjust for specific bilirubin concentrations in phototherapy-treated infants versus unexposed infants.	No association was found between phototherapy as reported via parent survey and childhood melanocytic nevus count at ages 2–7 y. Limitations include the potential for recall bias attributable to the survey design and an inability to adequately control for potential confounders at the time of phototherapy.
Matichard 2006 Necker Hospital, Paris, France	Case-control	Eighteen 8- to 9-y-old children who were born from 1994 to 1995 and treated with phototherapy for jaundice at birth and 40 age-matched schoolchildren who had not received neonatal phototherapy.	Per inclusion criteria.	Neonatal phototherapy for jaundice.	Total nevus count and nevus count by size in millimeters (mm) at ages 8–9 y.	Multivariate analysis using a variance model demonstrated an associated increase in nevi of size 2-5 mm (P < .001), but not an increase in total nevus count. Reported solar exposure was associated with an increase in the total nevus count (P = .02) and age (P = .02) and freckles (P = .03) were associated with an increase in 2–5 mm nevi.	A variance model was used to adjust for age, skin/eye/hair classifications, and survey-reported solar exposure. Effect sizes for multivariate analysis are not reported. Multiple comparisons. The evaluations were performed by the same dermatologist but at different locations: at the hospital for those receiving phototherapy and at a school for the nonexposed comparator group. Thus, the evaluation was not blinded.	An association was found at ages 8–9 between neonatal phototherapy exposure and an increased 2-5 mm nevus count. There was no association between phototherapy and total nevus count. Limitations include a nonblinded examination and the inability to control for confounders and bilirubin concentration in both groups at the time of potential phototherapy exposure shortly after birth.
Mahé 2008 France	Retrospective cohort	828 9-y-old children at 52 French primary schools in May-June 2007.	Children absent from school on the day of the skin examination.	Neonatal phototherapy for neonatal jaundice as determined by a standardized survey of parents and children.	Melanocytic nevus count as performed by 2 dermatologist-trained nurses.	No difference in the melanocytic nevus count was detected between students that reported phototherapy exposure versus unexposed students (mean± SD nevus count of 16.8±9.8 in exposed vs 16.7±10.5 in unexposed). Skin phototype, skin color, eye color, and number of sunburn episodes were associated with nevus count.	Statistical analysis was done by X² test. A standardized interview of parents included questions of sun exposure history, sun protection, and history of neonatal phototherapy. Risk of recall bias from the use of survey to determine phototherapy. Limited control for potential confounders at the time of phototherapy attributable to survey design. Unable to adjust for specific bilirubin concentrations in phototherapy-treated infants versus unexposed infants. Evaluation was not blinded.	No association was found between phototherapy as reported via parent/child survey and melanocytic nevus count at age 9 y. Limitations include the potential for recall bias because of the survey design, an unblinded analysis, and an inability to adequately control for potential confounders at the time of phototherapy.
Csoma 2011 The University of Szeged, Hungary	Cross-sectional twin study	58 pairs of twins (15 monozygotic pairs, 43 dizygotic) and one set of dizygotic triplets aged 3–30 y between January-April 2008 at The University of Szeged, Hungary. In each pair, one twin had received phototherapy for neonatal jaundice and the other had not.	Per inclusion criteria.	Neonatal blue light phototherapy as recorded in the neonatal medical chart. Icterus/jaundice, and prematurity as recorded in neonatal charts.	Melanocytic lesions as determined by skin examination. Uveal melanocytic lesions as determined by detailed ophthalmic examination.	In the multivariable linear regression analysis, neonatal blue light phototherapy was significantly associated with increased melanocytic nevi prevalence: (standardized coefficient: 0.158; 95% CI: 0.003–0.352; P = .47) (unstandardized β coefficient: 0.076). Age (standardized coefficient: 0.503; 95% CI: 0.051–0.101; P < .0001) and number of subtropical/tropical holidays (0.178; 95% CI: 0.012–0.332; P = .035) were also associated with increased skin lesion prevalence. Neonatal blue light phototherapy was associated with an increased odds of benign pigmented ocular lesions (aOR: 3.778; 95% CI: 1.694–8.423; P = .001).	Examining dermatologists and ophthalmologists were blinded to the subjects’ neonatal blue light phototherapy treatment history. Twin comparison controls for many antenatal and genetic variables present before potential phototherapy exposure. Were able to determine receipt of phototherapy exposure via chart review. Subjects or parents were surveyed on family skin lesion history, sun exposure and protection habits. These risk factors along with age and pigmentary traits were adjusted for in the multivariable analyses. Multivariable linear regression for skin lesions (natural log of skin lesions was normally distributed) and multivariable logistic regression for uveal lesions (not normally distributed).	This twin comparison study demonstrated an association between neonatal blue light phototherapy exposure and increased numbers of both melanocytic skin and uveal lesions.
Wintermeier 2014 Ludwig-Maximilian Hospital in Munich, Germany	Case-control	52 white, 5- to 6-y-old children who were previously exposed to neonatal blue light phototherapy after birth at Ludwig-Maximilian Hospital in Munich, Germany and 52 nonexposed children who were matched on age, gender, gestational age at birth, and skin phototype.	Per inclusion criteria.	Neonatal blue light phototherapy exposure.	Melanocytic nevus count at age 5–6 y by a single investigator. Café-au-lait macules, congenital nevi, and freckles were assessed separately.	The median nevus count in children with phototherapy was 17 compared to 18.5 in the matched nonexposed comparator group (P = .18) (Mann-Whitney U test). A higher count of café-au-lait macules (mean 0.5) were noted in the phototherapy exposed children relative to the nonexposed (mean 0.2) (P = .001). Skin phototype (P = .05), duration of sunny vacations (P = .02), and vacations in sunny areas (P = .03) were associated with increased melanocytic nevi.	Limited control for potential confounders at the time of phototherapy. Unable to adjust for specific bilirubin concentrations in phototherapy-treated infants versus unexposed infants. A history of family history of skin cancer, sun exposure, and sun protection was obtained by parental survey. Blinding of the examining dermatologist to phototherapy exposure was not noted.	No association was found at ages 5–6 between neonatal phototherapy exposure and melanocytic nevus count. Limitations include a nonblinded outcome examination and the inability to adjust for confounders and bilirubin concentration in both groups shortly after birth, the time of potential phototherapy exposure.
Lai 2016	Systematic review and meta-analysis	5 observational studies (n = 2921 total subjects) that evaluated the association of neonatal phototherapy with melanocytic nevi development. (Bauer 2004, Csoma 2011, Mahé 2008, Matichard 2006, Wintermeier 2014.) Search terms: (“photo-therapy”[Mesh] AND blue light*) AND (“nevus, pigmented”[Mesh] OR moles OR melanocytic nevi).	Studies were excluded if they were not cohort, case-control, or cross-sectional designs or if mean values and standard deviations of nevi counts were not included or calculable from the presented data. (4 reviews excluded and 1 excluded for insufficient information.)	Neonatal phototherapy for hyperbilirubinemia.	Melanocytic nevus count.	Those with previous phototherapy were not found to have a significantly higher mean melanocytic nevus count: (Weighted mean difference: 0.32; 95% CI: -0.67 to 1.31; P = .53).	DerSimonian and Laird random-effects models were used to calculate a weighted mean difference. There was both clinical heterogeneity and significant statistical heterogeneity between the 5 studies (I²=69.1%; P = .01). Visual inspection of the funnel plot could not exclude publication bias. Formal Egger’s test failed to provide strong evidence of a small-study effect (P = .09).	This meta-analysis of published studies did not demonstrate a significantly higher mean number of melanocytic nevi after exposure to neonatal phototherapy. However, there was a great degree of heterogeneity between the included studies.
Neurologic
Maimburg 2016 Denmark	Retrospective cohort	70 230 singleton children born between February 1998 and May 2003 as recorded in the Danish National Birth Cohort (DNBC).	Children with missing information on phototherapy status (n = 40), birth outcome (n = 9), and twins (n = 15).	Information on exposure to phototherapy for hyperbilirubinemia was obtained from a maternal questionnaire at 6 mo postnatal. When estimating the risk of epilepsy, children were followed from postnatal day 29 until onset of epilepsy, death, or end of the follow-up in August 2010. In the analyses for the risk of febrile seizure, children were followed from postnatal day 90 until onset of febrile seizure, death, or 5 y, whichever came first.	Information on epilepsy and febrile seizures were obtained from ICD-10 diagnoses within the Danish National Hospital Registry (DNHR).	Treatment with phototherapy for hyperbilirubinemia was associated with an increased hazard for developing epilepsy in early childhood (aHR: 1.66; 95% CI: 1.23–2.24). Phototherapy treatment was not associated with febrile seizures (aHR: 1.04; 95% CI: 0.86-1.27). The increased hazard of epilepsy was present for boys (aHR: 1.98; 95% CI: 1.40–2.78) but not for girls (aHR: 1.14; 95% CI: 0.64–2.02).	Cox proportional hazard regression models accounted for observed follow-up time of each subject and adjusted for potential, measured confounders (gender, birth gestation, socioeconomic status, maternal risk factors during pregnancy, and maternal age, parity, infertility treatment). Were unable to adjust for bilirubin concentrations at the time of potential or actual phototherapy exposure. Maternal survey response for phototherapy was validated via a medical record review of a random 100-patient sample of infants born at Aarhus University Hospital, which revealed 96% agreement.	Phototherapy for hyperbilirubinemia in newborn infants was associated with an increased risk of epilepsy for males in early childhood. No excess risk for febrile seizures was seen. A limitation was the inability to adjust for bilirubin concentration in both groups at the time of potential phototherapy exposure.
Newman 2018 Northern California, USA	Retrospective cohort	496 632 infants born at ≥35 wk’ gestation in 1995–2011 and followed ≥60 d. Kaiser Permanente of Northern California Database of 15 hospitals.	Exceeded 2009 AAP exchange transfusion thresholds (n = 1773) and those with seizure diagnoses at <60 d postnatal (n = 1237). Died during birth admission.	Dichotomous variable for any phototherapy within the hospital or at home.	Primary outcome: ≥1 encounter with a seizure diagnosis plus ≥1 AED prescription. Secondary outcomes: other combinations of ≥1 or ≥2 encounters with seizure or epilepsy diagnoses and ≥1 AED prescriptions.	Phototherapy-associated hazard ratio for Seizure (aHR: 1.22; 95% CI: 1.05–1.42; P = .009). Use of phototherapy increased during the study period from 2.4% in 1995% to 15.9% in 2011. More boys experienced both phototherapy and also seizures, but there was no effect modification (interaction) between sex and phototherapy exposure. Adjusted 10-y excess risks per 1000 infants associated with phototherapy were 2.4 (95% CI: 0.6–4.1) overall, 3.7 (95% CI: 1.2–6.1) for boys, and 0.8 (95% CI: -1.7 to 3.2) for girls.	Cox proportional hazard and Poisson regression models were used to adjust for TSB concentrations and other measured, potential confounders. Cox models accounted for observed follow-up time of each subject. Propensity-score adjusted model estimates were similar to Cox model estimates. Additional analyses showed that bilirubin concentrations of 20–24.9 mg/dL only became significantly associated with seizure-diagnoses in adjusted models when the phototherapy-exposure variable was removed. Did not have data on the duration or intensity of phototherapy.	Neonatal phototherapy treatment is associated with a slight risk in childhood seizures diagnosed after ≥60 d postnatal following adjustment for bilirubin measurements. Boys are more likely than girls to both receive phototherapy treatment of hyperbilirubinemia and to be diagnosed with childhood seizure.

aHR, adjusted hazard ratio; aRR, adjusted relative risk; AED, antiepileptic drug; ICD, International Classification of Diseases; aIRR, incidence rate ratio; TSB; total serum bilirubin; ASD, autism spectrum disorder; ATET, average treatment effect on the treated.

Question: What are the adverse clinically detectable effects of phototherapy in newborns?

Population: Neonates born at ≥35 weeks’ gestation with hyperbilirubinemia who are candidates for phototherapy

Intervention: Treatment with phototherapy

Comparator: Nontreatment

Outcome: Increased risk of any adverse clinical outcome that is important to neonatal patients, their families, and/or the clinicians treating them

Excluded: case reports/series, nonsystematic review articles, manuscripts focused on preterm infants born at <35 weeks’ gestation

Table 1 References:

1. Aspberg S, Dahlquist G, Kahan T, et al. Is neonatal phototherapy associated with an increased risk for hospitalized childhood bronchial asthma? Pediatr Allergy Immunol. 2007 Jun;18(4):313–319

2. Aspberg S, Dahlquist G, Kahan T, Källén B. Confirmed association between neonatal phototherapy or neonatal icterus and risk of childhood asthma. Pediatr Allergy Immunol. 2010 Jun;21(4 Pt 2):e733–739

3. Auger N, Laverdière C, Ayoub A, et al. Neonatal phototherapy and future risk of childhood cancer. Int J Cancer. 2019 Jan;145(8):2061–2069

4. Bauer J, Büttner P, Luther H, et al. Blue light phototherapy of neonatal jaundice does not increase the risk for melanocytic nevus development. Arch Dermatol. 2004 Apr;140(4):493–494

5. Berg P, Lindelöf B. Is phototherapy in neonates a risk factor for malignant melanoma development? Arch Pediatr Adolesc Med. 1997 Dec;151(12):1185–1187

6. Brewster DH, Tucker JS, Fleming M, et al. Risk of skin cancer after neonatal phototherapy: retrospective cohort study. Arch Dis Child. 2010 Oct;95(10):826–831

7. Bugaiski-Shaked A, Shany E, Mesner O, et al. Association between neonatal phototherapy exposure and childhood neoplasm. J Pediatr. 2022 Feb 1:S0022-3476(22)00077-4

8. Cnattingius S, Zack MM, Ekbom A, et al. Prenatal and neonatal risk factors for childhood lymphatic leukemia. J Natl Cancer Inst. 1995 Jun 21;87(12):908–914

9. Cnattingius S, Zack M, Ekbom A, et al. Prenatal and neonatal risk factors for childhood myeloid leukemia. Cancer Epidemiol Biomarkers Prev. 1995 Jul-Aug;4(5):441–445

10. Csoma Z, Tóth-Molnár E, Balogh K, et al. Neonatal blue light phototherapy and melanocytic nevi: a twin study. Pediatrics. 2011 Oct;128(4):e856–e864

11. Dahlquist G, Kallen B. Indications that phototherapy is a risk factor for insulin-dependent diabetes. Diabetes Care. 2003 Jan;26(1):247–248

12. Das RR, Naik SS. Neonatal hyperbilirubinemia and childhood allergic diseases: a systematic review. Pediatr Allergy Immunol. 2015 Feb;26(1):2–11

13. Digitale JC, Chang PW, Li SX, et al. The effect of hospital phototherapy on early breastmilk feeding. Paediatr Perinat Epidemiol. 2021 Nov;35(6):717–725

14. Digitale JC, Kim MO, Kuzniewicz MW, Newman TB. Update on phototherapy and childhood cancer in a northern california cohort. Pediatrics. 2021 Nov;148(5):e2021051033

15. Egeberg A, Andersen YM, Gislason G, Skov L, Thyssen JP. Neonatal risk factors of atopic dermatitis in Denmark - Results from a nationwide register-based study. Pediatr Allergy Immunol. 2016 Jun;27(4):368–374

16. Kemper K, Forsyth B, McCarthy P. Jaundice, terminating breast-feeding, and the vulnerable child. Pediatrics. 1989 Nov;84(5):773–778

17. Ku MS, Sun HL, Sheu JN, et al. Neonatal jaundice is a risk factor for childhood asthma: a retrospective cohort study. Pediatr Allergy Immunol. 2012 Nov;23(7):623–628

18. Kuniyoshi Y, Tsujimoto Y, Banno M, et al. Neonatal jaundice, phototherapy and childhood allergic diseases: An updated systematic review and meta-analysis. Pediatr Allergy Immunol. 2021 May;32(4):690–701

19. Kuzniewicz MW, Niki H, Walsh EM, et al. Hyperbilirubinemia, phototherapy, and childhood asthma. Pediatrics. 2018 Oct;142(4)

20. Lai YC, Yew YW. Neonatal blue light phototherapy and melanocytic nevus count in children: a systematic review and meta-analysis of observational studies. Pediatr Dermatol. 2016 Jan-Feb;33(1):62–68

21. Mahé E, Beauchet A, Aegerter P, Saiag P. Neonatal blue-light phototherapy does not increase nevus count in 9-y-old children. Pediatrics. 2009 May;123(5):e896-900

22. Maimburg RD, Olsen J, Sun Y. Neonatal hyperbilirubinemia and the risk of febrile seizures and childhood epilepsy. Epilepsy Res. 2016 Aug;124:67–72

23. Matichard E, Le Hénanff A, Sanders A, et al. Effect of neonatal phototherapy on melanocytic nevus count in children. Arch Dermatol. 2006 Dec;142(12):1599–1604

24. Newman TB, Wickremasinghe AC, Walsh EM, et al. Phototherapy and risk of type 1 diabetes. Pediatrics. 2016 Nov;138(5):e20160687

25. Newman TB, Wickremasinghe AC, Walsh EM, et al. Retrospective cohort study of phototherapy and childhood cancer in Northern California. Pediatrics. 2016 Jun;137(6)

26. Newman TB, Wu YW, Kuzniewicz MW, et al. Childhood seizures after phototherapy. Pediatrics. 2018 Oct;142(4)

27. Olsen JH, Hertz H, Kjaer SK, Bautz A, Mellemkjaer L, Boice JD Jr. Childhood leukemia following phototherapy for neonatal hyperbilirubinemia (Denmark). Cancer Causes Control. 1996 Jul;7(4):411–414

28. Podvin D, Kuehn CM, Mueller BA, Williams M. Maternal and birth characteristics in relation to childhood leukemia. Paediatr Perinat Epidemiol. 2006 Jul;20(4):312–322

29. Paludetto R, Mansi G, Rinaldi P, et al. The behavior of jaundiced infants treated with phototherapy. Early Hum Dev. 1983 Oct;8(3-4):259–267

30. Roman E, Ansell P, Bull D. Leukaemia and non-Hodgkin's lymphoma in children and young adults: are prenatal and neonatal factors important determinants of disease? Br J Cancer. 1997;76(3):406–415

31. Schedle A, Fricker HS. Impact of hyperbilirubinaemia and transient mother-child separation in the neonatal period on mother-child attachment in the first year of life. Eur J Pediatr. 1990 May;149(8):587–591

32. Seppälä LK, Vettenranta K, Leinonen MK, et al. Preterm birth, neonatal therapies and the risk of childhood cancer. Int J Cancer. 2021 May 1;148(9):2139–2147

33. Sun HL, Lue KH, Ku MS. Neonatal jaundice is a risk factor for childhood allergic rhinitis: a retrospective cohort study. Am J Rhinol Allergy. 2013 May–Jun;27(3):192–196

34. Telzrow RW, Snyder DM, Tronick E, et al. The behavior of jaundiced infants undergoing phototherapy. Dev Med Child Neurol. 1980 Jun;22(3):317–26

35. Tham EH, Loo EXL, Goh A, et al. Phototherapy for neonatal hyperbilirubinemia and childhood eczema, rhinitis and wheeze. Pediatr Neonatol. 2019 Feb;60(1):28–34

36. Usatin D, Liljestrand P, Kuzniewicz MW, et al. Effect of neonatal jaundice and phototherapy on the frequency of first-year outpatient visits. Pediatrics. 2010 Apr;125(4):729–734

37. Waite WM, Taylor JA. Phototherapy for the treatment of neonatal jaundice and breastfeeding duration and exclusivity. Breastfeed Med. 2016 May;11:180-185

38. Wei CC, Lin CL, Shen TC, Kao CH. Neonatal jaundice and risks of childhood allergic diseases: a population-based cohort study. Pediatr Res. 2015 Aug;78(2):223–230

39. Wickremasinghe AC, Kuzniewicz MW, Grimes BA, et al. Neonatal phototherapy and infantile cancer. Pediatrics. 2016 Jun;137(6)

40. Wintermeier K, von Poblotzki M, Genzel-Boroviczény O, et al. Neonatal blue light phototherapy increases café-au-lait macules in preschool children. Eur J Pediatr. 2014 Nov;173(11):1519–1525

41. Wu YW, Kuzniewicz MW, Croen L, et al. Risk of autism associated with hyperbilirubinemia and phototherapy. Pediatrics. 2016 Oct;138(4):e20161813

View Large

This Feature Is Available To Subscribers Only