Internalizing Symptoms in Adolescents With Sickle Cell Disease

Abstract

Objective

Sickle cell disease (SCD) is a genetic blood disorder that may affect patients’ mood and behavior. However, measuring the prevalence of internalizing symptoms (anxiety and depression) in patients with SCD has been elusive. We assessed internalizing symptoms in adolescents with SCD to evaluate prevalence and to test whether neurocognitive performance and frequency of pain-related episodes were associated with internalizing concerns.

Methods

One hundred eighty-five patients (57% HbSS/HbSß⁰-thalassemia, 43% HbSC/HbSß⁺-thalassemia), ages 12–18 years, received a neuropsychological evaluation as a part of a larger cohort study. Internalizing symptoms were measured using the Behavior Assessment System for Children, Second or Third Edition. Scores on the depression and anxiety scales were compared to normative values using Wilcoxon signed rank test. Spearman correlations examined associations between neurocognitive performances and internalizing symptoms. Robust multivariable regression models measured associations between internalizing symptoms and age, sex, sickle genotype, total hemoglobin, fetal hemoglobin, socioeconomic status, and frequency of pain episodes.

Results

Parent- and self-reported ratings of internalizing symptoms were not elevated compared to normative expectations. Overall, 1.8% and 6.3% of the sample displayed clinically elevated symptoms of anxiety and depression based on self-report, respectively. There were no associations between internalizing symptoms and neurocognitive performance (all p > .05). In multivariable analyses, the frequency of pain episodes was positively associated with self-reported anxiety (p = .006) and parent-reported depressive symptoms (p = .017).

Conclusions

Adolescents with SCD do not report elevated internalizing symptoms compared to normative expectations. Further research is needed to examine the trajectory of internalizing symptoms and the bidirectional relationship between pain and psychosocial functioning in SCD.

Introduction

Sickle cell disease (SCD) is an autosomal recessive disorder that distorts the shape of hemoglobin (Hb) molecules, which transport oxygen throughout the body (Kato et al., 2018). Inadequate oxygen delivery to tissue and organ systems causes a cascade of pathophysiological symptoms including anemia, infection, pain crises, stroke, tissue damage, organ failure, and premature death (Ware et al., 2017). In the United States, it is estimated that approximately 100,000 individuals are living with SCD (Centers for Disease Control and Prevention, 2019). Black individuals experience the highest prevalence rates when compared to White or Hispanic individuals in the United States (Ojodu et al., 2014). Patients with SCD experience greater rates of poverty and fewer protective socioeconomic factors when compared with the U.S. Black population (Ghafuri et al., 2020; Tewari et al., 2015).

SCD is associated with reduced health-related quality of life (HRQoL), including worse physical, psychosocial, emotional, social, and school functioning compared to healthy peers (Dale et al., 2011). Disease severity has been associated with worse HRQoL (Dampier et al., 2010). Recurring vaso-occlusive pain episodes, which are frequent and debilitating, can last hours to days and often result in patients seeking medical care and hospitalization. In addition to effects on HRQoL, SCD is associated with a stigma due to disease burden and premature mortality rates (Grosse et al., 2011; Wakefield et al., 2017). Despite these complications, most children and adolescents diagnosed with SCD are highly resilient and report age-normative HRQoL (Badawy, Thompson, Lai, et al., 2017).

Internalizing symptoms broadly refer to symptoms of anxiety and sadness (Achenbach & Edelbrock, 1981). The internalizing dimension was initially based on clinical classification systems but was later validated empirically as a distinct dimension of behavior and has been used for several decades in research (Woodman et al., 2016). It is important to identify and diagnose internalizing symptoms in patients with SCD as they negatively impact HRQoL (Graves et al., 2016), increase the risk for pain-related functional disability (Palermo et al., 2008), and are associated with poorer medication adherence (Badawy, Thompson, Penedo, et al., 2017).

Identifying the prevalence of internalizing symptoms in patients with SCD has been elusive, with estimates spanning 2–56% for depression (Benton et al., 2007; Jerrell et al., 2011; Jonassaint et al., 2016) and 8–41% for anxiety (Reader et al., 2020; Valrie et al., 2020). A national review of over 21,000 inpatient discharge records for children with sickle cell anemia found that 6% of patients were discharged with a psychiatric disorder diagnosis, including anxiety and depression (Myrvik et al., 2012). The varied prevalence estimates across studies are likely due, in part, to sample age differences. The clinical presentation, risk factors, and prevalence of internalizing symptoms differ markedly across development, particularly between children and adolescents (Costello et al., 2011; Rao & Chen, 2009). Therefore, it is important to characterize internalizing symptoms based on the developmental stage of the population.

Prior research has been limited by biased samples and inconsistent measurement techniques Many of the studies assessing rates of internalizing symptoms in SCD have selected samples based on prior clinical presentation (Goldstein-Leever et al., 2018; Simon et al., 2009). Self-reported internalizing symptoms in SCD, particularly mood, have received limited attention in the literature. Further complicating efforts to identify the prevalence and severity of internalizing symptoms, Grant et al. (2000) observed that estimates of depression in adults with SCD differed dramatically based on measurement technique. They found that 26% of patients met criteria for depression when interviewed compared to 40% meeting criteria based on responses to a self-report measure. Because the prevalence of internalizing symptoms is highly dependent on the measurement technique, using multiple metrics (e.g., parent- and self-report) likely provides a better estimate, and is more consistent with clinical practice.

Although there are some inconsistencies, studies have observed that internalizing symptoms in children and adolescents with SCD are not associated with demographic characteristics (Sehlo & Kamfar, 2015) or disease factors, measured as a composite (Ezenwosu et al., 2021) or laboratory values (Wallen et al., 2014). Intrapersonal and socio-ecological factors (e.g., family and social support) have been identified as having more influence on symptoms of depression and anxiety (Burlew et al., 2000; Sehlo & Kamfar, 2015). Some studies have suggested that greater frequency of acute painful episodes is associated with elevated internalizing symptoms in a bidirectional manner or fashion (Cepeda et al., 1997; Goldstein-Leever et al., 2018; Reader et al., 2020).

Children with SCD are at-risk for global cognitive deficits, as well as weaknesses in areas of verbal comprehension, nonverbal reasoning, working memory, executive functioning, and processing speed (Heitzer, Cohen, et al., 2022; Heitzer et al., 2021). These weaknesses are exacerbated by cerebrovascular events (Debaun & Kirkham, 2016; King et al., 2014; Steen et al., 2003). Internalizing symptoms, particularly depression, are associated with cognitive weaknesses in children and adults. A large meta-analysis found that cognitive performances in areas of executive functioning (e.g., inhibition and working memory) and processing speed were associated with major depression (Snyder, 2013). For children with SCD, poor working memory skills were associated with parental reports of depressive symptoms, whereas stronger verbal comprehension skills were associated with improved coping with disease-related stress (Prussien et al., 2018). Broadly, anxiety is associated with disruptions in executive functioning and working memory (Moran, 2016; Robinson et al., 2013). To date, no studies have evaluated the relationship between self-reported symptoms of internalizing symptoms and cognitive performances in youth with SCD.

Our first aim was to examine the prevalence of internalizing symptoms among an unbiased sample (all patients with SCD monitored for the surveillance regardless of disease severity) of adolescents with SCD using both parent- and self-report measures. We hypothesized that symptoms of anxiety and depression would be elevated based on both parent- and self-report ratings. Our second aim was to examine the relationship between internalizing symptoms and neurocognitive performance in adolescents with SCD. We hypothesized that greater internalizing symptoms would be negatively associated with performance on neurocognitive measures of working memory, processing speed, and verbal reasoning. Finally, we sought to examine the relationship between demographic and disease factors with internalizing symptoms. We hypothesized that the severity of internalizing symptoms would be associated with greater frequency of acute painful episodes, but there would be no relationship between internalizing symptoms with disease severity (sickle genotype, Hb, HbF) or treatment history (hydroxyurea exposure).

Methods

This research was approved by the Institutional Review Board (IRB) at St. Jude Children’s Research Hospital (Memphis, USA). The legal guardian of each participant gave written informed consent and adolescents gave assent according to the requirements of the IRB.

Participants

All participants of the Sickle Cell Clinical Research Intervention Program (SCCRIP) study, ages 12–18 years, were eligible for this study. The design of SCCRIP has been described previously (Hankins et al., 2018). Briefly, SCCRIP is a longitudinal lifetime cohort study that collects retrospective and prospective data on clinical, neurocognitive, psychosocial, geospatial, and health outcomes of children, adolescents, and adults with SCD. Neuropsychological assessments are performed approximately every 3–4 years on all participants, irrespective of disease symptoms. We included SCCRIP participants from two developmental stages: early adolescence (ages 12–13 years) and late adolescence (ages 16–17 years) at the time of their assessment. Of the 483 eligible participants, 185 received a neuropsychological assessment. Forty-six participants received both assessments as part of longitudinal monitoring, but only the late adolescence assessment was included in analyses. In total, we included assessments from 78 early adolescents and 107 late adolescents. There were no demographic or clinical differences among those who completed the internalizing and neurocognitive assessments and those who did not (Supplemental Table 1).

Demographic, Medical, and Treatment Variables

Demographic, medical, and treatment variables were abstracted from the SCCRIP database. The Social Vulnerability Index (SVI) (Cutter et al., 2003; Flanagan et al., 2018) was used to classify individuals based on social vulnerabilities at the neighborhood level (e.g., poverty, education, housing data); a higher percentile score indicates higher social vulnerability. Hydroxyurea was administered to participants with HbSS/HbSβ⁰-thalassemia according to NHLBI guidelines (Yawn et al., 2014). For participants with HbSC/HbSβ⁺-thalassemia, hydroxyurea initiation was guided by the frequency of acute disease complications (Dehury et al., 2015; Luchtman‐Jones et al., 2016). Hematologic indices including total Hb and fetal hemoglobin (HbF) were performed on the day of neurocognitive testing or were the average value of measurements within 3 months prior to testing. The number of acute pain episodes within the last 2 years prior to the Behavior Assessment System for Children (BASC) measurements was collected. To quantify the number of acute pain episodes, a list of all acute hospital visits from the health record was pulled. Then, a second review was done by research assistants who go over each discharge diagnosis and confirm the data that came from electronic health record or reassign the discharge diagnosis. If a research assistant is in doubt, they ask a physician for clarification. A classification of acute pain visit is provided if there is an acute vaso-occlusive painful event that required medical attention and administration of analgesics (opioids or non-opioids) by a medical facility was provided.

Internalizing Symptom Measures

The BASC, Second or Third Edition, parent- and self-report scales measured internalizing symptoms (Kamphaus & Reynolds, 2007; Reynolds & Kamphaus, 2015). The BASC is a broad-band questionnaire used to evaluate the behavior and self-perceptions of children and young adults aged 2–25 years. The adolescent version of the BASC is designed for individuals aged 12–21 years and has a nationally based normative sample. Both adolescents and their parents completed the BASC scale. The normative samples of the BASC-2 and BASC-3 parent- and self-report (ages 12–18 years) measures contain approximately 14–16% of individuals who identify as Black, broadly consistent with the U.S. population (Kamphaus & Reynolds, 2007; Reynolds & Kamphaus, 2015). The anxiety scale measures the tendency to be nervous, fearful, or worried. Specific items assess ease of becoming stressed, fearfulness, and concerns about mistakes. The depression scale measures feelings of unhappiness, sadness, and loss of interest in activities. Specific items assess ease of becoming upset, having negative thoughts, and appearing lonely. Raw scores of endorsed items are compared to normative values (gender based) and T scores with a mean of 50 and a standard deviation (SD) of 10 are calculated for all scales. A T score of ≥70 on the BASC is considered to be clinically significant.

Neurocognitive Measures

The Wechsler Abbreviated Scale of Intelligence—Second Edition (WASI-II) (Wechsler, 2011) provided an estimated Full-Scale IQ (4-subtest), Verbal Comprehension Index (verbal reasoning), and Perceptual Reasoning Index (perceptual reasoning). Depending on the participant’s age, the Wechsler Adult Intelligence Scale, Fourth Edition (WAIS-IV) (Wechsler, 2014a) or Wechsler Intelligence Scale for Children, Fourth Edition (WISC-IV) (Wechsler, 2014b), Digit Span (working memory) and Coding (processing speed) subtests were used. Standard scores with a mean of 100 and an SD of 15 were calculated.

Our group has previously presented the neurocognitive outcome data from SCCRIP, including the WASI-II, WAIS-IV, and WISC-IV (see Heitzer et al., 2021; Heitzer, Longoria, et al., 2022; Partanen et al., 2020; Saulsberry‐Abate et al., 2021; Trpchevska et al., 2022). The primary outcomes (BASC anxiety and depression scales) presented in this study have not previously been published.

Statistical Analyses

Summary statistics including mean and SD for continuous variables and count and proportion for categorical variables were reported. They were compared using two-sample t-test or Wilcoxon rank sum test for continuous variables or Fisher’s exact test for categorical variables between early and late adolescences. Data normality was checked using Shapiro–wilk test. Scores on the depression and anxiety scales were compared to normative values using Wilcoxon signed rank or one-sample t-test. Cohen’s D was calculated as an effect size to measure the difference from normative values. Spearman correlation test was used to examine the correlations between two continuous variables. Univariate and multivariable robust regressions were used to assess the associations of internalizing symptoms with age, sex, disease severity (genotype, total Hb, HbF), hydroxyurea treatment history, socioeconomic status, neurocognitive performance, and frequency of pain episodes. Robust regression may be more appropriate in the presence of outliers and/or residuals that have a distribution that is heavier in the tails than a normal distribution (Field & Wilcox, 2017). The use of robust M-estimation, as implemented in the robust package in R (lmRob function), allowed for the relaxation of some of the assumptions of ordinary least squares and attenuated the influence of outliers. For BASC parent-report, the outcome data on 5% of patients were not available. For BASC self-report, the outcome data on ≤40% of patients were not available. The distributions of SCD genotype, gender, hydroxyurea treatment, age at evaluation, social vulnerability, total Hb, HbF, and pain episodes (past 2 years) were similar between patients with and without BASC self-report data. Therefore, we used the complete case analysis for all statistical analysis and no imputation of missing values was performed. All p-values were two-sided and <.05 were considered significant. SAS version 9.4 and R-4.1.2 were used to conduct all statistical analysis. The strength of the correlation coefficient (ρ) was classified as negligible (<|0.30|), weak (|0.30|–|0.50|), moderate (|0.50|–|0.70|), and strong (≥|0.70|) (Hinkle et al., 2003).

Results

Demographic and Clinical Characteristics

Demographic and clinical characteristics by age group (early and late adolescence) are displayed in Table I. Ratings of internalizing symptoms were completed for a total of 185 patients. On average, patients were 15.2 (SD = 2.0) years of age at evaluation and roughly half the sample was male (51.4%). Approximately 57% of the sample was diagnosed with a severe SCD genotype (SS/SB⁰-thalassemia) and 53% had a history of hydroxyurea treatment. The average SVI (M = 66.9, SD = 25.5) fell in the very high range, demonstrating that most of the sample lived in neighborhoods with low socioeconomic status. As expected, levels of total Hb were higher (p = .006) in the late adolescence group (M = 10.6, SD = 1.9) compared to early adolescence (M = 9.8, SD = 1.7).

Table I.

Demographic and Clinical Characteristics by Age Group

Variable	Total	Early adolescence	Late adolescence	p-Value
	(N = 185)	(N = 78)	(N = 107)
	Frequency (%)	Frequency (%)	Frequency (%)
SCD genotype				.487
SC/SB+/other	79 (42.7%)	31 (39.7%)	48 (44.9%)
SS/SB0	106 (57.3%)	47 (60.3%)	59 (55.1%)
Gender				.541
Female	90 (48.6%)	40 (51.3%)	50 (46.7%)
Male	95 (51.4%)	38 (48.7%)	57 (53.3%)
Hydroxyurea treatmenta				.489
No	87 (47.0%)	39 (50.0%)	48 (44.9%)
Yes	98 (53.0%)	39 (50.0%)	59 (55.1%)
	Mean (SD)	Mean (SD)	Mean (SD)
Age at evaluation (years)	15.2 (2.0)	13.0 (0.4)	16.8 (0.6)	<.001
Social vulnerability indexb	66.9 (25.5)	64.5 (26.9)	68.6 (24.4)	.366
Total hemoglobin (g/dl)	10.2 (1.8)	9.8 (1.7)	10.6 (1.9)	.006
Fetal hemoglobin (%)	9.9 (8.8) (N = 180)	10.5 (8.6) (N = 77)	9.5 (8.9) (N = 103)	.216
Pain episodes (# in the past 2 years)	1.0 (2.4)	0.7 (1.1)	1.2 (3.0)	.353
Verbal Comprehensionc	88.9 (13.4) (N = 159)	90.2 (15.1) (N = 69)	87.9 (11.9) (N = 90)	.288
Digit Spand	8.5 (2.7) (N = 179)	8.7 (3.0) (N = 77)	8.3 (2.5) (N = 102)	.321
Codingd	7.9 (2.6) (N = 181)	7.9 (2.9) (N = 76)	7.9 (2.4) (N = 105)	.807

Note. N = sample size; SCD = sickle cell disease; SD = standard deviation. Verbal Comprehension, Digit Span, and Coding are measured by the age-appropriate Wechsler Intelligence Scales. Categorical comparisons are based on the chi-square test and continuous comparisons are based on the Wilcoxon signed rank test. Bold denotes p < .05.

^aHistory of hydroxyurea treatment.

^bMeasured by the Social Vulnerability Index. Classifies individuals based on social vulnerabilities at the neighborhood level (e.g., poverty, education, housing data); a higher percentile (0–100) score indicates higher social vulnerability.

^cCalculated standard score with a mean of 100 and SD of 15.

^dCalculated scaled score with a mean of 10 and SD of 3.

Internalizing Symptoms Compared to Normative Expectations

Comparison of parent- and self-report ratings of internalizing symptoms to normative expectations is displayed in Table II. Data were available for 175 patients with parent-reported outcomes and 112 patients with self-reported outcomes (111 had both). There was no difference in demographic or medical characteristics between patients with and without self-report ratings (Supplemental Table 2). Parent ratings of both anxiety (p < .001) and depression (p = .012) symptoms fell below normative expectations indicating that parents rated their children as having fewer internalizing symptoms than same-age children from a national normative sample. Self-report ratings of anxiety and depression symptoms did not differ from normative expectations (p > .20). Figure 1 displays the distribution of scores for ratings of anxiety and depression based on parent- and self-report measures.

Table II.

Internalizing Symptoms Compared to Normative Expectations

Internalizing symptoms	N	Sample, mean (T score)	Sample, SD	Normative, mean	Cohen’s D	p-Value
Anxiety—parent-report	175	44.7	10.6	50	0.497	<.0001
Depression—parent-report	175	48.3	8.83	50	0.191	.012
Anxiety—self-report	111	48.9	10.0	50	0.108	.257
Depression—self-report	112	48.8	11.1	50	0.112	.237

Note. N = sample size; SD = standard deviation; T score = standardized score with a mean of 50 and standard deviation of 10. Internalizing symptom scores are based on the anxiety and depression scales from the Behavior Assessment System for Children, Second or Third Edition. Higher scores are indicative of greater symptoms. Gender-based normative data were used for all participants. p-Values are based on one-sample t-test. Bold denotes p < .05.

Figure 1.

Histograms of internalizing symptom scores. Internalizing symptom scores based on the Anxiety and Depression scales from the Behavior Assessment System for Children, Second or Third Edition. Higher scores are indicative of greater symptoms. Gender-based normative data were used for all participants. T scores have a mean of 50 and SD of 10.

Associations Between Parent- and Self-Report Ratings

There were significant, yet negligible correlations between parent- and self-report ratings of internalizing symptoms. Correlations were broadly consistent for measures of anxiety (ρ = 0.251, p = .011) and depression (ρ = 0.258, p = .0089). Overall, adolescents self-reported anxiety ratings were elevated compared to parent-report (p = .013), but there was no significant difference between self- and parent-reported ratings of depression (p = .422) (see Supplemental Table 3 for details).

Internalizing Symptoms Across Age Groups

Table III displays ratings of anxiety and depression across early adolescence (ages 12–13 years) and late adolescence (ages 16–17 years) groups. Based on parent-report, 2.9% and 1.7% of the sample reached the cutoff for clinical significance (T ≥ 70) for the anxiety and depression scales, respectively. For self-report ratings, 1.8% of the sample reached the clinical cutoff for anxiety and 6.3% of the sample reached clinical significance for depression. Notably, all seven cases with clinical levels of depression (based on self-report) fell in the late adolescence group (10.8%; p = .021). The mean scores also differed between the early and late adolescence groups on the self-reported depression scale (p = .031). No other group differences were observed on the anxiety and depression scales for clinically elevated symptoms or standardized scores (all p > .05).

Table III.

Internalizing Symptoms by Age Group

Variable	Total	Early adolescence	Late adolescence
Clinical cutoff (T ≥ 70)	Frequency (%)	Frequency (%)	Frequency (%)	p-Value
Anxiety—parent-report				1.00
No	170 (97.1%)	73 (97.3%)	97 (97.0%)
Yes	5 (2.9%)	2 (2.7%)	3 (3.0%)
Depression—parent-report				.259
No	172 (98.3%)	76 (100.0%)	96 (97.0%)
Yes	3 (1.7%)	0 (0.0%)	3 (3.0%)
Anxiety—self-report				.507
No	109 (98.2%)	47 (100.0%)	62 (96.9%)
Yes	2 (1.8%)	0 (0.0%)	2 (3.1%)
Depression—self-report				.021
No	105 (93.8%)	47 (100.0%)	58 (89.2%)
Yes	7 (6.3%)	0 (0.0%)	7 (10.8%)

Standardized T score	Mean (SD)	Mean (SD)	Mean (SD)	p-Value
Anxiety—parent-report	44.7 (10.6)	45.4 (11.1)	44.2 (10.2)	.495
Depression—parent-report	48.3 (8.8)	47.9 (8.1)	48.6 (9.4)	.866
Anxiety—self-report	48.9 (10.0)	48.9 (9.2)	48.9 (10.7)	.777
Depression—self-report	48.8 (11.1)	46.0 (8.3)	50.8 (12.5)	.031

Note. N = sample size; SCD = sickle cell disease; SD = standard deviation. Internalizing symptom scores are based on the anxiety and depression scales from the Behavior Assessment System for Children, Second or Third Edition. Higher scores are indicative of greater symptoms. Gender-based normative data were used for all participants. Categorical comparisons are based on Fisher’s exact test and continuous comparisons are based on the Wilcoxon signed rank test. Bold denotes p < .05.

Internalizing Symptoms and Neurocognitive Functioning

Table IV displays associations between internalizing symptoms and neurocognitive functioning. There were no associations between parent- or self-report ratings of internalizing symptoms and any neurocognitive outcome (p > .05).

Table IV.

Associations between Internalizing Symptoms and Neurocognitive Performance

Internalizing symptoms	Neurocognitive	Spearman’s correlation
		ρ	p-Value
Anxiety—parent-report	VCI	0.114	.16
	DST	−0.046	.55
	Coding	0.010	.90
Depression—parent-report	VCI	−0.038	.65
	DST	−0.055	.47
	Coding	−0.040	.61
Anxiety—self-report	VCI	0.089	.39
	DST	−0.15	.12
	Coding	−0.008	.94
Depression—self-report	VCI	0.091	.38
	DST	−0.012	.90
	Coding	−0.103	.28

Note. DST = Digit Span Total; VCI = Verbal Comprehension Index. Internalizing symptom scores are based on the anxiety and depression scales from the Behavior Assessment System for Children, Second or Third Edition. Higher scores are indicative of greater symptoms. Gender-based normative data were used for all participants. Performance on the Verbal Comprehension Index is based on the Wechsler Abbreviated Scale of Intelligence—Second Edition. Performance on Digit Span and Coding is based on the Wechsler Intelligence Scale for Children—Fourth or Fifth Edition.

Predictors of Internalizing Symptoms

Robust regression models of internalizing symptoms including genotype, sex, age at evaluation, hydroxyurea treatment history, Hb, HbF, SVI, and the number of pain episodes are displayed in Table V. Male sex (p = .038), HbSS/HbSβ⁰-thalassemia genotype (p = .030), and hydroxyurea treatment history (p = .030) were associated with greater parent-reported anxiety symptoms. The number of pain episodes was positively associated with self-reported anxiety (p = .006) and parent-reported depressive symptoms (p = .017). No medical, treatment, or demographic variables were associated with self-reported depressive symptoms (all p > .05).

Table V.

Predictors of Internalizing Symptoms

Outcome	Predictors	Parent-report (N = 175)			Self-report (N = 112)
		β	SE β	p-Value	β	SE β	p-Value
Anxiety	(Intercept)	52.5	8.71	<.0001	58.3	14.8	.0002
	Genotypea	−6.44	2.94	.0301	−2.84	5.38	.598
	Sexb	3.48	1.66	.0375	−1.77	2.67	.509
	Age at evaluation	−0.153	0.423	.719	0.352	0.693	.612
	Hydroxyurea treatmentc	4.56	2.08	.0301	0.234	3.28	.943
	Total hemoglobin (g/dl)	−0.731	0.685	.288	−0.883	1.19	.461
	Fetal hemoglobin (%)	0.0111	0.127	.931	−0.229	0.232	.328
	Social vulnerabilityd	−0.00151	0.0311	.961	−0.0429	0.0532	.422
	Pain episodese	0.294	0.32	.361	2.47	0.877	.00579
Depression	(Intercept)	43.3	8.04	<.0001	43.6	7.76	<.0001
	Genotype	−1.59	2.73	.56	−1.42	2.9	.627
	Sex	2.28	1.57	.148	−0.612	1.54	.692
	Age at evaluation	−0.0043	0.397	.991	0.653	0.384	.0923
	Hydroxyurea treatment	0.773	1.98	.697	1.74	1.91	.366
	Total hemoglobin (g/dl)	−0.0353	0.648	.957	−0.594	0.628	.347
	Fetal hemoglobin (%)	0.0852	0.116	.465	−0.202	0.129	.122
	Social vulnerability	0.0221	0.029	.447	−0.00921	0.0311	.767
	Pain episodes	0.717	0.297	.0168	0.209	0.278	.455

Note. N = sample size; SE = standard error. Robust multiple regression models of internalizing symptoms. Internalizing symptom scores based on the anxiety and depression scales from the Behavior Assessment System for Children, Second or Third Edition. Higher scores are indicative of greater symptoms. Gender-based normative data were used for all participants. p-Values <.05 were in bold.

^aGenotypes split into two groups: SC/SB⁺/other (reference group) vs. SS/SB⁰.

^bFemale used as the reference group.

^cHistory of hydroxyurea treatment. No treatment was the reference group.

^dMeasured by the Social Vulnerability Index. Classifies individuals based on social vulnerabilities at the neighborhood level (e.g., poverty, education, housing data); a higher percentile (0–100) score indicates higher social vulnerability.

^eNumber of pain episodes within the past 2 years.

Discussion

Internalizing symptoms can negatively impact HRQoL and exacerbate clinical symptoms in SCD. To address these concerns, it is important to understand the prevalence and modifiers of internalizing symptoms based on both parent- and self-report measures. In contrast to our hypotheses, symptoms of anxiety and depression were not elevated in a sample of adolescents with SCD based on either parent- or self-report. Surprisingly, parents rated their adolescents with SCD as having fewer depression and anxiety symptoms than would be expected based on a large normative sample of adolescents from the general population in the United States, suggesting positive coping and resiliency among this population. There were minimal associations between measures of neurocognitive performance and ratings of internalizing symptoms. Consistent with our hypothesis, frequency of pain episodes was positively associated with symptoms of anxiety and depression.

Due to debilitating physical symptoms, SCD is considered a highly stressful and burdensome experience for adolescents diagnosed with the condition. These adolescents experience frequent hospitalizations and procedures, episodes of pain, and increased fatigue. Furthermore, there are often disruptions to age-appropriate activities (e.g., participating in sports and school attendance) and an increase in uncontrollable negative life events (e.g., bullying, romantic break-up, parent conflict) experienced starting in adolescence (Barakat et al., 2008; Hankin, 2006; Sil et al., 2016). Based on these symptoms, associated complications, and developmental changes, it is realistic to expect an increased frequency of internalizing difficulties among these adolescents. Yet, our study does not suggest that there are elevations in internalizing symptoms compared to normative expectations on the BASC rating scale. These results are consistent with some studies demonstrating no elevations in internalizing symptoms among children and adolescents with SCD (Cepeda et al., 1997; Noll et al., 2007; Valrie et al., 2020; Yang, 1994), but in contrast with others that did (Barbarin et al., 1994; Morgan & Jackson, 1986). Although we did not observe elevations in internalizing symptoms when comparing average standardized scores, there were seven cases (6.3% of sample) in the older adolescence group that reached the clinical cutoff (approximately 98th percentile) for depression based on self-report. This suggests that there is a subset of patients with SCD with severe depressive concerns that require monitoring and treatment.

The finding that internalizing symptoms in adolescents with SCD are consistent with normative expectations is likely due to multiple factors. Although there are many features of SCD that potentially predispose children and adolescents to internalizing symptoms, it is important to recognize that most patients are well adjusted. Our sample of adolescents with SCD appear to demonstrate emotional resiliency in coping with SCD. Consistent observations are seen in other pediatric populations with medical conditions. For example, studies of pediatric patients diagnosed with cancer demonstrate that these patients display internalizing symptoms that are broadly consistent with healthy peers or normative data (Phipps, 2005, 2007). We assessed patients at a single timepoint rather than evaluating lifetime prevalence of internalizing symptoms. It is possible that patients experienced these difficulties in the past, which was not captured in the current data. Consistent with the SCD population, our sample was composed entirely of patients who identify as Black. In the United States, healthy Black individuals endorse fewer symptoms of depression and anxiety than White or Hispanic individuals (Williams et al., 2007). Because the normative sample of the BASC (Kamphaus & Reynolds, 2007; Reynolds & Kamphaus, 2015) only contains a small percentage (14–16%) of Black individuals, it is possible that we underestimated the presence of internalizing symptoms in this population. Lastly, studies that documented higher rates of internalizing symptoms among patients with SCD (Barbarin et al., 1994; Morgan & Jackson, 1986) were conducted prior to the widespread use of hydroxyurea. Although we did not observe a significant relationship between hydroxyurea use and internalizing symptoms, the standard of care for patients with SCD in the United States has significantly improved in the past decade as evidenced by improved laboratory markers of disease severity (refer to Heitzer, Longoria, et al., 2022). These changes in medical care are potentially leading to improved psychosocial functioning.

No associations between parent- or self-report ratings of internalizing symptoms and neurocognitive performance were observed. The absence of these associations is likely due in part to the low rates of internalizing symptoms and lack of variability across the sample. Prior work demonstrating associations between mood and anxiety symptoms with cognitive outcomes were broadly based on samples with more significant internalizing concerns (Moran, 2016; Snyder, 2013). A recent study among youth with SCD (Prussien et al., 2018) demonstrated a negative relationship between parent-rated depressive symptoms and working memory performance but not verbal comprehension. Overall, their sample (n = 44) demonstrated greater depressive symptoms than observed in our study potentially leading to the observed associations. Although our findings do not suggest an association between internalizing symptoms and neurocognitive performance among adolescents with SCD, future research with larger samples and including adults with SCD are warranted to further explore these relationships.

Consistent with findings in healthy youth (Achenbach et al., 1987), we found low concordance between parent- and self-report ratings of internalizing symptoms among adolescents with SCD (i.e., there were small correlations between parent and adolescent ratings of anxiety and depression symptoms). Adolescents perceive internalizing concerns different from their caregivers and may display divergent behaviors depending on the setting. A thorough assessment of internalizing symptoms in adolescents must include both parent- and self-report. Joint clinical interviews and discussion with parent(s) and the patient may improve concordance. Future research should examine if greater concordance among parent- and self-report ratings reflects positive parent–child communication.

All patients with clinically elevated symptoms of depression were in the late adolescence group, and there was a significant difference in depressive symptoms between the early and late adolescence groups in univariate analyses. Yet, in multivariable analyses, there was no effect of age on any internalizing outcome. Prior studies suggest an increase in internalizing symptoms as patients with SCD transition to adulthood (Adam et al., 2017; Wallen et al., 2014). An increase in symptoms in adulthood may stem from worsening disease status, difficulties managing medical care and daily responsibilities, fewer available resources for social support, or increased functional impairment. Further longitudinal research is needed to understand the trajectory of internalizing symptoms in SCD. Notably, depressive symptoms are associated with poorer HRQoL and higher healthcare costs in adulthood (Adam et al., 2017), although the directionality of this relationship is unclear. Increased psychosocial monitoring and support during the transition from pediatric to adult care may limit the development of significant internalizing symptoms in adult patients with SCD.

Multivariable analyses controlling for disease and demographic factors demonstrated that the frequency of acute painful episodes requiring hospitalization was positively associated with parent-reported symptoms of depression and self-reported anxiety. In the general population, depression, anxiety, and pain are intricately linked (Dufton et al., 2009; IsHak et al., 2018). Similar findings have been reported in patients with SCD (Reader et al., 2020; Toumi et al., 2018). These studies have hypothesized that there is a bidirectional link between physical and psychological sequelae in SCD (Toumi et al., 2018). Because patients with frequent pain episodes are more likely to present with internalizing symptoms, behavioral interventions that can target both symptom domains and underlying somatic concerns are likely to be beneficial. Trials of cognitive behavioral therapy in SCD have demonstrated promise in reducing the affective components of pain (Anie & Green, 2015) and the ability to cope with pain (Thomas et al., 1999). Acceptance and commitment therapy (Masuda et al., 2011) and eHealth applications (Palermo et al., 2018) have also shown promise in managing pain in SCD. Further research to understand the association between pain and internalizing symptoms in SCD is needed and may lead to tailored treatments for this population. Due to the elevated rates of internalizing symptoms among patients with frequent painful episodes, greater monitoring and screening for psychosocial difficulties among these patients is warranted.

Study strengths include a large sample, parent- and self-report of internalizing symptoms, and comprehensive systematic (not symptom-based) assessment of medical and treatment factors. Yet, limitations do exist. The primary limitation of the study is the lack of a sibling or demographically matched control group. Normative data for the BASC only include a small percentage of Black individuals, potentially leading to an underestimation of the extent of internalizing symptoms among our sample. The collection and dissemination of race-based normative data for standardized psychosocial measures may allow for more meaningful comparisons; however, caution is warranted due to the potential for over-pathologizing behaviors among specific racial/ethnic groups. The current data set is cross-sectional, limiting our ability to examine the trajectory of internalizing symptoms across the lifespan. Most patients did not receive neuroimaging. Thus, we could not examine the potential associations between neurological factors (e.g., silent infarcts) and internalizing symptoms. Measurement of acute pain episodes only included the most severe events requiring hospitalization to allow for reliable measurement. Yet, most episodes of pain in SCD are managed in the home environment. There may have been confounding variables that influenced patients’ tendency to utilize healthcare services for pain. Future studies should utilize multiple pain metrics to fully capture the influence of pain on internalizing symptoms. Many participants did not complete self-report ratings of internalizing symptoms due to time demands during the appointments. This led to a smaller sample of self-reported outcomes; however, the group with self-report ratings appeared to be representative of the larger sample. Measurement of internalizing symptoms was limited to a single checklist completed by adolescents and their parents. Inclusion of teacher ratings, diagnostic interviews, and/or symptom specific measures would have complemented the collected data and led to a more comprehensive view of the internalizing symptoms experienced by adolescents with SCD. Future research examining internalizing symptoms in younger cohorts with SCD is needed to determine whether findings differ across developmental stages.

To conclude, in this study, adolescents with SCD and their caregivers did not endorse elevated levels of internalizing symptoms compared to normative expectations. However, a substantial subset of older adolescents (10.8%) displayed clinically elevated symptoms of depression, likely requiring intervention (e.g., cognitive behavioral therapy). Patients with greater frequency of acute painful episodes demonstrated more internalizing concerns. Further research examining the relationship between pain and internalizing symptoms in SCD is needed to determine optimal intervention methods. Internalizing concerns may increase as patients transition from pediatric to adult care and longitudinal research examining this trajectory is needed.

Supplementary Data

Supplementary data can be found at: https://academic.oup.com/jpepsy.

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.