Behavioral Health and Adult Milestones in Young Adults With Perinatal HIV Infection or Exposure

Children with PHIV and PHEU children 9 to 16 years old were recruited into the Child and Adolescent Self-Awareness and Health (CASAH) study from 4 NYC medical centers from 2003 to 2008.²⁵ Inclusion criteria were (1) perinatal HIV exposure, (2) cognitive capacity to complete an interview (excluding severe intellectual deficits), (3) residing with a caregiver who could provide legal consent, and (4) being able to speak English or Spanish. Participants and caregivers completed an assessment battery at enrollment and an 18-month follow-up (FU). The CASAH study received additional funding to add yearly interviews (CASAH 2: 2008–2013; CASAH 3: 2013–2018). In this analysis, we use data from the sixth interview (CASAH 3; FU 5), which was conducted from 2014 to 2017. Caregivers participated in the CASAH and CASAH 2 studies; only young adults (YAs) participated in the CASAH 3 study. At FU 5, a 2- to 3-hour interview consisting of validated measures was administered by trained research assistants. Measures affected by social desirability were assessed by using an audio computer-assisted self-interview (ACASI).

This study was approved by the Columbia University–New York State Psychiatric Institute Institutional Review Board. At FU 5, YAs provided written consent for interviews and collection of data from their medical providers.

Data were collected at each interview, including from caregivers at enrollment and YAs at FU 5.

Psychiatric functioning was assessed with the Diagnostic Interview Schedule for Children, a structured diagnostic instrument that asks about symptoms of common psychiatric diagnoses as defined by the American Psychiatric Association’s Diagnostic and Statistical Manual of Mental Disorders system.^26,–28 We examined specific diagnoses within the categories of anxiety, mood, and disruptive behavior disorders and substance use disorders (SUDs).

Neurocognitive executive function was assessed with the Trail Making Test (TMT) and Weschler Adult Intelligence Scale Digit Span subtest.^29,30 The TMT is a timed, 2-part measure that requires participants to use visual search, scanning, mental flexibility, and executive function to sequentially connect encircled numbers (TMT A) and encircled numbers and letters (TMT B). The time, in seconds, required to complete TMT A and TMT B is reported. The Weschler Adult Intelligence Scale Digit Span is a 3-part measure of working memory; scores from 3 tests are converted to a Digit Span standard score.

Participants answered questions via ACASI about (1) lifetime and recent (past 3 months) oral, vaginal, or anal sex; (2) recent condomless vaginal or anal sex; (3) recent types of sex partners; and (4) 30-day use of cigarettes, alcohol, and marijuana. Sexual behavior questions came from the Adolescent Sexual Behavior Assessment^31,32; cigarette, alcohol, and marijuana use came from the Monitoring the Future study.^33,34 

Using the Adolescent Sexual Behavior Assessment and via ACASI, participants reported pregnancy history and pregnancy outcomes, contraception use, and sexually transmitted infections (STIs) in the past year.

Participants reported on education, employment, incarceration, housing, romantic relationships, current living situation (including financing), and history of homelessness.

HIV RNA viral load (VL) and CD4⁺ cell counts from the 12 months before FU 5 and current ART regimen were obtained from participants’ health care providers.

Descriptive data (means, SDs, and frequencies) were presented for the total sample and separately by HIV status. Outcomes for PHIVYAs and PHEUYAs were compared by using χ² tests for categorical variables and t tests for continuous variables. For days hospitalized, a nonparametric test was used (Mann–Whitney) because of a skewed distribution. The CD4⁺ and VL values obtained closest to the interview were selected, with values dichotomized as CD4⁺ count <100 and >500 cells per mm³ and VL <50, <200, and >1000 copies per mL. ART regimens were categorized by drug class, number of pills, and frequency of administration. We declared findings to be statistically significant if the corresponding P values were ≤.05. Data were analyzed by using IBM SPSS Statistics 23 (IBM SPSS Statistics, IBM Corporation).

At enrollment, participants included 340 youth, 9 to 16 years old, and their caregivers (206 PHIV and 134 PHEU; Table 1). At FU 5, 248 (151 PHIV and 97 PHEU) YAs completed interviews (Supplemental Fig 1); 1 PHEUYA was excluded from analysis because of HIV seroconversion before FU 5. At enrollment, participants were on average 12.6 years old (SD 2.25), 51% were of female sex, 65% were African American and/or black, and 42% were Latino. Average household income was $27 718, with PHEU youth coming from significantly lower-income homes compared with participants with PHIV ($24 090 vs $30 111; P = .009). Most participants had a female primary caregiver (88%), but PHEU patients were twice as likely to be living with a birth parent (P < .001), and thus, their caregivers were more likely to be living with HIV (P < .001).

At FU 5, participant ages ranged from 18 to 28 years, with PHIVYAs being older than PHEUYAs (mean age 22.8 vs 21.7 years; P = .002). Between enrollment and FU 5, participants completed a mean of 5 interviews (range 1–6). Eleven deaths were reported, 10 among those with PHIV (4 attributed to HIV, 1 to systemic lupus erythematosus, and 5 unknown) and 1 PHEU patient (from an accident). There were no significant differences in enrollment characteristics (eg, age, race, sex, income, HIV status, and caregiver type) among YAs retained versus not retained for FU 5. However, YAs without a psychiatric disorder in both groups were more likely to be lost to FU, although the difference was only significant for PHEUYAs (data not shown).

At FU 5, significantly more PHIVYAs received housing assistance (P = .002), public assistance (P < .001), food stamps (P < .001), and disability payments (P < .001) compared with PHEUYAs. Most had health insurance, with more PHIVYAs receiving Medicaid (P = .008) and fewer PHIVYAs than PHEUYAs being uninsured (P = .002). Significantly more PHIVYAs than PHEUYAs reported deaths of birth parents (61% vs 35%; P < .001) or other caregivers (eg, grandparents; 25% vs 13%; P = .033).

Laboratory data were available for 126 of 151 PHIVYAs. Mean CD4⁺ was 491 cells per mm³ (range 2–1430 cells per mm³); 44% had a CD4⁺ of >500 cells per mm³, 20% had a CD4⁺ of <200 cells per mm³, and 15% had a CD4⁺ of <100 cells per mm³ (Table 2). Approximately two-thirds (64%) were virally suppressed to <200 copies per mL (range undetectable to >2 000 000 copies per mL; mean 2.4 log₁₀ copies per mL; SD 1.66). ART regimen data were available for 132 of 151 PHIVYAs, including 8 who were not prescribed ART. Among 124 who were prescribed ART, 67% received a 1-drug– or 2-drug–class regimen, and 31% received a 3-class combination. Fifty-eight percent received a regimen that included a boosted protease inhibitor (bPI), and half (50%) received an integrase inhibitor. Most PHIVYAs (87%) were on once-daily regimens, but only 33% took a single fixed-dose combination pill; 34% were taking 4 or more pills daily.

Asthma and allergies were the most prevalent health conditions, with no significant HIV status differences seen (Table 2). In the past year, 19% of participants received inpatient medical care (22% PHIV; 14% PHEU), with PHIVYAs averaging 3.6 hospital days (range 1–120 days) compared with <1 day reported by PHEUYAs. Significantly more PHIVYAs than PHEUYAs (1) were prescribed psychotropic medication treatments (16% vs 6%; P = .022) and/or (2) received any nonpsychopharmacologic mental health services (42% vs 20%; P < .001). Few (2 per group) reported substance use treatment.

There were few HIV status differences in psychiatric disorder rates; 27% of all participants met criteria for any non-SUD psychiatric disorder (Table 3). Anxiety disorders (22%; particularly generalized anxiety disorder) and mood disorders (11%; particularly major depression) were the most prevalent.

There were no significant HIV status differences for substance use or SUD (Table 3). Overall, 28% reported cigarette use, 63% reported alcohol use, and 47% reported marijuana use in the past 30 days. Nearly one-quarter (24%) reported daily marijuana use, and 35% reported being drunk in the past 30 days. High rates of YAs met SUD criteria (28%), with marijuana (20%) and alcohol (15%) use disorders being the most prevalent. Finally, 10% of participants met criteria for both a psychiatric disorder and SUD (data not shown).

Neurocognitive function differed by HIV status (Table 3). They required a significantly longer time to complete TMT A (28.3 vs 23.2 seconds; P < .001) and obtained significantly lower Digit Span standard scores (7.6 vs 8.6; P = .009); there were no HIV group differences on TMT B.

We found no significant HIV status differences in sexual behaviors (Table 4). The majority had initiated sex (93%); in the past 3 months, 63% had vaginal, 12% had anal, and 41% had condomless sex. In the past 3 months, 60% reported sex with opposite-sex partners only, 12% had at least 1 same-sex partner, and 28% reported no partnered sex.

Among women, 41% reported lifetime pregnancy, 23% had a pregnancy end in a live birth, and 20% were currently living with their child (Table 4). Among men, 38% reported lifetime pregnancy in a partner, 18% reported a pregnancy ending in a live birth, and 6% were living with a child. There were few HIV group differences in contraceptive use and STIs. Almost all sexually active participants reported lifetime (91%) and past-year (76%) use of condoms. Male condoms were the most common birth control method used in the past year (70%); other methods included withdrawal (49%), having oral instead of vaginal sex (21%), and oral contraceptives (19%). Methods used varied by HIV status, with more PHEUYAs reporting withdrawal (P = .026) and oral contraceptive use (P = .042).

There were no statistically significant HIV status differences in educational achievement or employment status (Table 5). Sixty-seven percent of YAs had graduated high school (HS), obtained a general equivalency diploma, or were on track to graduate HS, and 19% attended college. Less than half were working (42%), but 62% were either working or in school. Twelve percent of YAs reported ever being incarcerated, 2% in the past year, with no HIV status differences.

Almost all participants had ever been in a romantic relationship (93%), with nearly half (48%) being currently in one. The vast majority (94%) were single and never married.

Significant HIV status differences were found in living situations: more PHEUYAs were living rent free with family or friends compared with PHIVYAs (68% vs 38%; P < .001), whereas PHIVYAs were more likely to be renting a room or apartment or contributing rent (54% PHIV vs 26% PHEU; P < .001), likely because of HIV-related subsidies and/or housing assistance. Overall, 16% of the cohort had ever been homeless, with no HIV status differences.

Using data from CASAH, 1 of the first and largest behavioral health studies in the United States that includes a highly representative cohort of PHIVYAs from NYC and a comparison cohort of PHEUYAs from similar backgrounds, we report on long-term behavioral, psychosocial, and biomedical outcomes among individuals who were born at the height of the perinatal HIV epidemic in United States. We found that despite the challenges of living with a chronic, stigmatized illness, behavioral health and psychosocial outcomes among PHIVYAs appear to be comparable to those of PHEUYAs who grew up in similar circumstances. As they reach adulthood, many PHIVYAs are achieving normal adult milestones, completing school, securing employment, engaging in relationships, and starting families. At the same time, both PHIVYAs and PHEUYAs have high rates of psychiatric disorders and SUDs, incarceration, homelessness, and high-risk sexual behaviors at rates that are similar to those of YAs living in poor, urban communities throughout the United States.^35,–37 

The biomedical health outcomes of PHIVYAs were mixed. Ten PHIVYAs were known to have died since enrollment, with only 1 PHEUYA death occurring. Furthermore, more than one-third of the PHIVYAs with available data had viremia (>200 copies per mL), and 15% were severely immunocompromised, putting them at high risk for disease progression and HIV transmission to sexual partners. An analysis of youth with PHIV in the United States revealed that the proportion of time spent viremic and with a low CD4⁺ cell count is substantially higher among adolescents with PHIV and PHIVYAs compared with 7- to 12-year-old children with PHIV.⁸ Low rates of VS (45.9%) have been reported among PHIVYAs on ART at US adolescent centers.¹³ In NYC, VS rates among PHIVYAs were lower compared with nonperinatally HIV–infected YAs (61% vs 72%).¹⁰ In our cohort, among those on treatment, most received potent regimens with bPIs and/or integrase inhibitors given once daily. Nonetheless, we know that many PHIVYAs have had periods of inadequate adherence,^14,38,39 harbor a multidrug-resistant virus,⁴⁰ and are subject to a variety of risk factors, including high rates of psychiatric disorders^23,41,42 that put them at continued risk for suboptimal adherence, viremia, and poor health outcomes.

Unique to our study, we used a well-validated diagnostic psychiatric interview, the Diagnostic Interview Schedule for Children, to evaluate psychiatric disorders, whereas researchers in most studies use symptom checklists. We found that 27% of YAs had a psychiatric disorder, with no significant group differences seen. Rates of psychiatric disorders were similar to or slightly greater than in cohorts of uninfected YAs in the United States.^43,–45 Importantly, rates of depression are similar to those reported for African American and/or black, Latino, and low-income YAs (6%–10%) in the United States.⁴⁶ Rates of anxiety disorders are also consistent with those reported among college-aged individuals.⁴⁷ That being said, given that PHIVYAs were more likely to have received mental health treatment, including psychotropic medications, it is possible that they may be at greater risk, but treatment has improved their mental health to the level of the PHEUYAs who have less access to mental health services.

Both groups reported high rates of cigarette, alcohol, and marijuana use, and 28% met diagnostic criteria for SUD, which is notably higher than what has been reported from a national study of YAs, in which 21% of white, 17% of Hispanic, and 16% of black 18- to 24-year-old respondents met criteria for drug or alcohol dependence or abuse.^48,49 Although there were no group differences in substance use generally, given that most PHIVYAs were engaged in long-term comprehensive HIV care and had higher rates of mental health treatment, it is of concern that rates were not lower.^50,51 

CASAH participants performed below average on tests of executive function and planning ability, with lower performance being among PHIVYAs than among PHEUYAs. Differences in executive function between patients with PHIV and PHEU patients have been attributed to the cumulative impact of late ART initiation with suboptimal regimens, periods of incomplete adherence and sustained viremia, and both the direct and indirect impact of HIV on the central nervous system.^52,53 Compromised neurocognitive function may exacerbate difficulties with transitioning to adulthood. Cognitive remediation interventions may be helpful, as has been demonstrated in adults with other brain insults.⁵⁴ 

Almost all participants reported having engaged in sexual intercourse, 41% of women reported a pregnancy, and 38% of men reported a partner’s pregnancy, with no notable HIV group differences seen. Pregnancy outcomes were similar (miscarriage, abortion, and live birth) for women with PHIV and PHEU women. Researchers in several cohort studies have reported on pregnancies among PHIVYAs, but we are the first to demonstrate that pregnancies among PHIVYA women were no greater than among PHEU women.^49,55,–57 Among partners of young men with PHIV, a higher proportion of pregnancies resulted in the birth of a child, and more PHIVYA men (7 of 16; 44%) compared with PHEUYA men (0 of 4; 0%) were living with their child. It is possible that surviving childhood with HIV motivates a desire to be involved with one’s children and that financial subsidies provided to PHIVYAs in NYC enable more child support.

Achievement of adult milestones was also mixed, with few differences by HIV status seen. Almost all reported romantic relationships, with few being married, which corresponds with national data for this age group.⁴⁸ Two-thirds had received an HS diploma or the equivalent, but only 19% were in college.⁵⁸ High rates of incarceration (12%) and homelessness (16%) were reported, reflecting the challenges faced by many YAs in poor, urban communities.⁵⁹ 

There are several limitations of our study. Our findings may not be broadly generalizable because we only included NYC participants. However, we believe that the participants are representative of PHIVYAs and PHEUYAs because NYC has been an epicenter for the US perinatal HIV epidemic, participants were recruited from 4 large HIV centers, recruitment criteria were broad, and the enrollment of eligible subjects was high (77%).²³ In addition, demographics from CASAH participants correspond with those of several national studies of populations from across the United States, in which the majority of youth with PHIV and PHEU come from inner-city families whose lives were impacted by poverty and substance use, with similar sociodemographic milieus and family contexts.^18,40,60 Another limitation of this analysis is that CD4⁺ and VL data were missing for 17% and ART regimen data were missing for 13% of participants with PHIV. Additionally, we do not have data on drug resistance. Also, there was some participant attrition from enrollment, although 75% were retained for 10 to 14 years, and we found few differences on key enrollment characteristics between those who did and did not complete FU 5, thus limiting attrition bias. Counter to expectation, those YAs without a psychiatric diagnosis were less likely to continue in both groups, but the effect was only significant among PHEU patients.²³ This finding, in combination with PHIVYAs having higher rates of treatment, may have masked a mental health vulnerability of PHIVYAs. Finally, the absence of an HIV-unexposed cohort in this study limits our ability to fully discern the impact of HIV on the health outcomes of PHEUYAs and PHIVYAs.

CASAH represents 1 of the few longitudinal studies in which researchers assess health, behavioral health, and psychosocial outcomes of PHIVYAs. Many are successfully achieving YA milestones and positive health outcomes despite inherent socioeconomic and psychosocial challenges. However, high rates of psychiatric disorders, SUDs, and cognitive problems may place many PHIVYAs at risk for poor long-term outcomes. Although we found few group differences, the range of psychosocial problems in the context of PHIV can lead to devastating consequences. Given the staggering numbers of PHIVYAs worldwide, our findings reveal a critical need for integrated mental health, substance use, and health services for HIV-affected populations.⁶¹ With our findings, we also draw attention to the growing population of PHEUYAs that warrants similarly urgent opportunities for prevention and treatment.

ACASI

audio computer-assisted self-interview

ART

antiretroviral treatment

bPI

boosted protease inhibitor

CASAH

Child and Adolescent Self-Awareness and Health Study

FU

follow-up

HS

high school

NYC

New York City

PHEU

perinatally HIV–exposed but uninfected

PHEUYA

perinatally HIV–exposed but uninfected young adult

PHIV

perinatally acquired HIV infection

PHIVYA

young adult living with perinatally acquired HIV infection

STI

sexually transmitted infection

SUD

substance use disorder

TMT

Trail Making Test

VL

viral load

VS

viral suppression

YA

young adult