Clinical Assessment of Late Health Outcomes in Survivors of Wilms Tumor

The SJLIFE study, initiated in 2007, is a single center retrospective cohort study with prospective clinical assessments of 5 year survivors of childhood cancer treated at St. Jude Children’s Research Hospital (SJCRH) from 1962 to 2012.¹⁰  This analysis includes participants diagnosed with Wilms tumor who returned for an on-campus health evaluation (n = 280). Additionally, a community control group (n = 625) was recruited for comparison (Supplemental Table 3). The study protocol was approved by the SJCRH institutional review board, and informed consent was obtained before participation.

Clinical encounters included a history and physical examination; a fasting laboratory battery; questionnaires detailing demographics, medical history, health habits, and quality of life; and standardized neurocognitive and physical function testing.¹⁰  Therapeutic exposures abstracted from the medical record included cumulative chemotherapy doses, radiation fields or doses, surgeries, and major medical events during and after cancer treatment. Participants consented to medical record release for validation and grading of health conditions diagnosed before SJLIFE evaluations.

One hundred sixty eight chronic health conditions were assessed and graded according to a modified version of the National Cancer Institute’s Common Terminology Criteria for Adverse Events v4.03¹¹  (grade 1 = asymptomatic or mild, 2 = moderate requiring minimal intervention, 3 = severe or disabling, 4 = life threatening). Prevalence of health conditions and psychological outcomes was assessed.

Standardized neurocognitive assessments (Supplemental Table 4), performed by certified examiners, supervised by a board-certified neuropsychologist, measured performance in 4 domains: attention,¹²  executive function,^13–15  memory,¹⁶  and processing speed.¹⁵  Age-adjusted z-scores were used for grading; ≥1 and <2 standard deviations below the population mean were considered mild (grade 1), ≥2 and <3 SD below the population mean moderate (grade 2), and ≥3 SD below the population mean severe (grade 3). Participants unable to complete a particular test were excluded from analysis for that domain.

Performance testing (Supplemental Table 5) was conducted by exercise specialists to ascertain physical function in 6 domains: aerobic capacity,¹⁷  mobility,¹⁸  strength,^19,20  endurance,²⁰  flexibility,^21,22  and balance.²³  Survivors performing >1.5 SDs below age- and sex-matched z-scores for controls were considered impaired in that domain. Participants unable to perform a particular test were excluded from analysis for that domain.

Two-sample t tests were used to compare continuous variables. $χ2$ or Fisher’s exact test, as appropriate, were used for categorical variables between participants and eligible nonparticipants and between survivors and controls.

The prevalence of physical function impairments was compared to the standard population using a 1 sample t test. Modified Poisson regression, adjusted for age, sex, and race, was used to compare the prevalence of Common Terminology Criteria for Adverse Events graded conditions, subsequent neoplasms, neurocognitive outcomes, and physical function between survivors and controls. Smoking was added to models for pulmonary outcomes.

The standardized incidence ratio (SIRs) of secondary neoplasms was calculated using age-, sex-, and calendar year–specific cancer incidence from the National Cancer Institute’s Surveillance, Epidemiology, and End Results program. Nonmelanoma skin cancers are not registered in the Surveillance, Epidemiology, and End Results database and were not included in the SIR calculation.

The mean cumulative count was calculated by individual chronic conditions after accounting for clinically relevant chronicity and recurrence measures. The overall cumulative burden of grade 1 to 4 and 2 to 4 conditions was calculated after summing the individual mean cumulative counts. The age-specific cumulative burden and 95% confidence intervals (CI) were estimated by summing chronic health conditions occurring per individual using the bootstrap percentile method. Statistical testing was considered significant at P < .05. All analyses were completed by using SAS 9.4 and R.

Of 445 eligible Wilms tumor survivors, 280 (63%) returned for evaluation. Sixty-nine (15.5%) agreed but had not yet returned, 16 (3.6%) participated by survey only, 46 (10.3%) declined, 9 (2.0%) could not be located, and 25 (5.6%) were pending recruitment (Supplemental Fig 4).

Participants were more likely to be non-Hispanic White (73.2% vs 62.4%, P = .02), older at evaluation (median 30.5 vs 26.0 years, P = .01), and farther from cancer diagnosis (26.0 vs 22.0 years, P < .01) than nonparticipants (Supplemental Table 3). Demographic and treatment-related characteristics of study participants are shown in Table 1. Nearly all (92.1%) were diagnosed before age 6 years (median 3.0 years, range: 0.0–15.0), with a median time from diagnosis of 26.0 years (6.0–54.0). Compared to controls, survivors were similar in age (median 30.5 [9.0–58.0] vs 31.0 [12.0–70.0] years, P = .32), but more racially diverse (26.8% non-White vs 17.6%, P < .01). Survivors (≥25 years old) were less likely to work full time (P < .01), be married (P < .01), or complete higher education (P < .01) than community controls.

All Wilms survivors underwent surgery, with the majority being unilateral nephrectomy (96.4%). The most common chemotherapeutic exposures included vincristine (99.3%) and dactinomycin (97.9%), and two-thirds (67%) were treated with doxorubicin. Nearly 60% received abdominal (whole and/or hemiabdomen) radiation (median dose 1200 [880–6120] centigray) and 25% chest radiation (median dose 1200 [900–4400] centigray). Twenty-six survivors (9%) experienced relapse (Table 1).

By age 40 years, survivors averaged 12.7 (95% confidence interval [CI] 11.7–13.8) grade 1 to 4 and 7.5 (95% CI 6.7–8.2) grade 2 to 4 conditions, compared to 4.2 (95% CI 3.9–4.6) and 2.3 (95% CI 2.1–2.5), respectively, among controls (Fig 1). Conditions of the cardiovascular, pulmonary, endocrine, and reproductive systems contributed the largest excess burden (Supplemental Tables 6 and 7).

Nearly all (99.6%) Wilms tumor survivors had at least 1 grade 1 to 4 condition and 91.8% a grade 2 to 4 condition. The most prevalent conditions, adjusted for sex, current age, and race, are reported in Table 2. Over 50% of survivors had grade 1 to 4 hypertension and 19.2% were on medication or met criteria for treatment initiation (grade 2–4) compared to 42.6% and 11.2%, respectively, among controls (P < .01). Grade 2 to 4 cardiomyopathy was identified in 6.5% of survivors vs 2.6% of controls (P = .01). One-fifth of survivors had a total cholesterol >200 mg/dL or triglycerides >150 mg/dL (grade 1–4), with 1.8% and 6.3%, respectively, exceeding 300 mg/dL (≥grade 2) and/or already on lipid-lowering therapy. Excess grade 2 to 4 obstructive (11.7% vs 2.9%, P < .01), restrictive (9.6% vs 0.2%, P < .01), and diffusion (10.4% vs 0.3%, P < .01) pulmonary impairments were observed in survivors compared to controls. Adjusting for smoking status, pulmonary diffusion defects were associated with doxorubicin (respiratory rate [RR]: 3.9, 95% CI: 1.2–12.6) and restrictive deficits with chest radiation (RR: 12.3, 95% CI: 1.8–86.4).

Survivors also experienced a higher rate of abnormal glucose metabolism (22.4% vs 17.5% grade 1–4 and 8.0% vs 2.7% grade 2–4, P < .01), and among survivors this was associated with abdominal radiation (RR: 4.8, CI: 1.7–13.7). This was not associated with overweight or obesity (52.5% vs 58.7%, P = .13) in survivors compared to controls, as might be expected in the general population. In fact, 5.7% of survivors were underweight at clinical evaluation compared to 2.2% of controls (P < .01).

Survivors experienced more kidney dysfunction than controls (10.0% vs 0.6% grade 1–4; 7.2% vs 0.6% grade 2–4, P < .01). Among adult (≥18 years old) survivors (n = 235) the median estimated glomerular filtration rate was 96.5 mL/min/1.73 m² (range: 3.7–146.7).

Over 8% of survivors experienced grade 2 to 4 peripheral neuropathy, compared to 2.6% of controls (P < .01). Females experienced excess ovarian failure before age 40 (9.3% vs 0.6%, P < .01), whereas 1.8% of males had evidence of Leydig cell insufficiency.

A mean of 27.7 years from diagnosis, 24 (8.6%) Wilms survivors developed 36 subsequent neoplasms (SN), one-half occurring in previous radiation fields. Nonmelanoma skin cancers were most common (n = 22), followed by gastrointestinal (n = 3), thyroid (n = 2), breast (n = 2), sarcomas (n = 5), and melanomas (n = 2). Survivors had nearly a fourfold increased risk (SIR 3.8, CI: 2.1–6.3) for SN compared to the general population. Although suggestive, an association with radiation exposure did not reach statistical significance (RR: 3.4, CI: 0.9–13.4, P = .08).

Among 246 survivors who completed neurocognitive testing, 52.1% had grades 1 to 3 and 19.7% grades 2 to 3 executive function impairment (Fig 2), compared to 45.0% and 11.6%, respectively, among controls (P < .01). Processing speed (19.7% vs 8.4%) and memory (20.9% vs 9.9%) similarly showed a higher prevalence of moderate to severe impairments (grade 2–3) compared to controls (P < .01). No differences in attention were identified. Associations with treatment exposures were not identified. Non-White survivors were noted to have double the risk of neurocognitive impairment across all domains compared to non-Hispanic White survivors (Supplemental Table 8). In addition to excess neurocognitive impairment, anxiety (24.8% vs 15.9% [P < .01]) and/or depression (25% vs 13.1% [P < .01]) were identified more often in survivors than controls.

Two hundred seventy survivors completed physical function testing and impairments exceeded expected rates across all domains (Fig 3). Excess impairments were identified in aerobic function, mobility, strength, endurance, and flexibility (P < .05). No associations with treatment exposures or level of physical activity were identified. However, having a grade 2 to 4 cardiac condition was associated with impaired 6-minute walk (RR: 2.2,95% CI: 1.3–3.8), grip strength (RR: 2.5, 95% CI: 1.1–5.3), and knee extension at 60°/second (RR: 1.6, 95% CI: 1.1–2.5) and 300°/second (RR: 2.0, 95% CI: 1.4–3.0), respectively (Supplemental Table 10).

Multimodal treatment strategies have advanced cure rates for children diagnosed with Wilms tumor. However, treatment exposures have significant impacts on long-term health. Our systematic, clinical evaluations identified nearly a threefold increased health burden, largely driven by endocrine and cardiopulmonary conditions. Additionally, we found clinically assessed impairments in neurocognitive and physical function that can impact health and quality of life for this population. In our analysis, nearly all (99.6%) had at least 1 chronic health condition, with most (92%) requiring clinical intervention (grade 2–4). Survivors had an almost fourfold increased risk (SIR: 3.8, CI: 2.1–6.3) for subsequent neoplasms. This excess health burden was accompanied by a higher-than-expected prevalence of mental health disorders, limited neurocognitive performance, and physical function impairments. These factors may contribute to impaired social attainment and lower rates of employment, marriage, and educational attainment in this population. Attention to these unique risk profiles offer opportunities for early detection and treatment of these health consequences to preserve quality of life.

Wilms treatment regimens have historically been associated with an elevated risk of cardiovascular disease,^24,25  the leading causes of nonmalignant late mortality in survivors of childhood cancer, as well as in the general population.²⁶  This analysis identified excess cardiomyopathy in survivors compared with controls, which is anticipated after cardiotoxic exposures (anthracyclines and chest radiation). Additionally, survivors in our cohort experienced excess cardiovascular risk factors (hypertension, hyperlipidemia, diabetes)²⁷  and nearly one-third (29.6%) reported a smoking history. When criteria from the American Heart Association and American College of Cardiology²⁸  for hypertension (Supplemental Table 11) are applied to participant data, 26.6% of survivors meet criteria for a diagnosis of hypertension, with an additional 22% meeting criteria for prehypertension at a median of 31 years of age. Similarly, using American Diabetes Association criteria for diabetes²⁹  (Supplemental Table 11), 33.2% of our survivors met criteria for prediabetes and 7.7% diabetes, compared with 19.8% and 4.5%, respectively, in controls. Modification of these risk factors can play a major role in preserving survivor health. In a CCSS study, the presence of hypertension significantly augmented the risk of developing heart failure (RR: 19.4, 95% CI: 11.4–33.1) and premature coronary artery disease (RR: 6.1, 95% CI: 3.4–11.2) among survivors exposed to chest radiation compared to those without hypertension.³⁰  Lifestyle modification (smoking, diet, etc) and management of chronic conditions may have significant health implications for Wilms tumor survivors.

The increased prevalence of heart disease and cardiovascular risk factors has implications for kidney function.³¹  This study includes survivors of bilateral and syndromic Wilms tumor, which may partially account for the higher prevalence of chronic kidney disease in our study compared to previous studies of unilateral, nonsyndromic Wilms tumor survivors.³²  Two-thirds of survivors with renal impairment in our study had comorbid hypertension, diabetes, or both. In three-fourths of these cases, the onset of diabetes and/or hypertension was documented before reduced kidney function. Hyperfiltration injury after nephrectomy may also be a contributing factor. Animal studies involving unilateral nephrectomy and five-sixths infarction of the contralateral kidney showed an increase in single nephron glomerular filtration rates, morphologic changes, and hypertension, suggesting injury from increased filtration.³³  Lacking longitudinal data from the time of nephrectomy, our cross-sectional analysis may have missed the hyperfiltration period. Studies are needed to elucidate the association of hyperfiltration and adverse kidney function among Wilms tumor survivors.

The identification of a significant burden of SNs in this population is consistent with previous studies of kidney tumors³⁴  and highlights the role of cancer screening and surveillance in survivors exposed to radiation, according to the Children’s Oncology Group Long-Term Follow-up Guidelines.³⁵  These survivors should undergo a thorough annual skin exam, and those with chest and/or abdominal radiation are at increased risk of breast^36,37  and colon³⁸  cancer, respectively. Early initiation of screening may lead to detection at earlier stages, additional treatment options, and opportunities to improve outcomes.

In addition to increased risk for cardiometabolic, kidney and neoplastic late effects, long-term survivors of Wilms tumor are at increased risk of infertility. Nearly 10% of female survivors experienced premature ovarian insufficiency, likely because of abdominal radiation exposing the ovaries.^7,39–42  None of the females exposed to hemiabdominal radiation experienced premature ovarian insufficiency. Previous reports have also identified structural uterine abnormalities,^43,44  and risks for premature deliveries^45–47  among female Wilms survivors exposed to these radiation fields.

Our study uniquely included assessments of mental health and neurocognitive function in this population not exposed to central nervous system (CNS)–directed therapies. Current literature is mixed on the prevalence of mental health conditions in survivors of childhood cancer.^5,48–50  Although some studies have reported that depression and anxiety at similar or even reduced rates compared to the general population, our analysis suggests Wilms survivors experience an increased mental health burden. Nearly one-quarter reported symptoms consistent with depression and anxiety. An increased prevalence of moderate to severe impairments of executive function, processing speed, and memory was also observed. Neurocognitive impairment has been described in a smaller sample (n = 158) of this same population and similarly did not identify treatment-related associations but did note associations with chronic health conditions.⁵¹  Our analysis expands the previous study by grading neurocognitive findings to quantify specific impairments more explicitly. In the absence of traditional treatment-related risk factors, such as CNS-directed therapies, we performed analyses investigating the possible impact of the presence of cardiopulmonary disease on neurocognitive function. Impairments in processing speed were associated with the presence of grade 2 to 4 cardiovascular disease. However, associations with other neurocognitive domains were not observed. Traditionally, only survivors exposed to CNS-directed therapies have been considered at risk for neurocognitive impairments; however, the overall burden and duration of chronic health conditions may also contribute. In an analysis from the CCSS, path analysis demonstrated direct effects of cardiopulmonary (RR: 1.3, 95% CI: 1.1–1.4) and endocrine (RR: 1.1, 95% CI: 1.0–1.3) conditions on impaired task efficiency.⁵²  Research is needed to determine if early interventions may alter the trajectory of cognitive function in this aging population.

Across all 4 domains, survivors of non-White race experienced a relative risk of neurocognitive impairment two- to threefold that of White survivors (Supplemental Table 8), suggesting a differential impact by race. This finding persisted when comparing non-White survivors to non-White controls alone (Supplemental Table 9). Further research, including social determinants of health and the neurodevelopmental impact of age at treatment, is needed.

Studies of Wilms tumor survivors in the CCSS and British CCSS noted increased physical limitations assessed by self-report.^5,8  Using physical performance testing, we observed excess impairments in aerobic function, mobility, strength, endurance, and flexibility. Previous studies including smaller numbers of Wilms survivors have not observed these impairments.^53,54  Reports from SJLIFE have suggested that exposure to abdominopelvic radiotherapy may cause body composition changes that affect aerobic function, strength, and endurance.⁵⁵  In our study, there were no clear associations between these outcomes and treatment exposures. However, having a cardiac condition was associated with impairment in the aerobic function, strength, and endurance domains, which may have implications for physical and mental health. To our knowledge this is the largest and most comprehensive physical function assessment in Wilms survivors to date, suggesting this population may benefit from a formalized physical training intervention.

Despite systematic health assessments performed in SJLIFE, our study has some limitations. Thirty-seven percent of eligible survivors were not included in the analysis (Supplemental Fig 4) and differed from participants by age and race. Participation required returning to SJCRH, and some survivors who declined participation may have been excluded because of inability to travel, potentially under- or overestimating the prevalence of conditions. Our findings may not be representative of younger survivors more recently diagnosed because of the limited recruitment to date of these individuals. Despite uniquely characterizing neurocognitive and physical performance, not all participants were able to complete this testing resulting in potential for bias.

Survivors of Wilms tumor experience an excess burden of chronic health conditions threatening quality of life and longevity. Neurocognitive and physical function impairments were more prevalent than anticipated, despite a lack of CNS-directed treatment. Focused health maintenance and surveillance for these unique health risks may provide opportunities for early intervention, with the goal of reducing morbidity and preservation of long-term health.

CCSS

Childhood Cancer Survivor Study

CNS

central nervous system

CI

confidence interval

CTCAE

Common Terminology Criteria for Adverse Events

RR

respiratory rate

SIR

standardized incidence ratio

SJCRH

St. Jude Children’s Research Hospital

SJLIFE

St. Jude Lifetime

SN

subsequent neoplasms