Study Shows Increased Risk of Mental Disorders in Childhood Cancer Survivors and Potential Increased Risk in Young Siblings  

Get Permission

Survivors of childhood cancer are at risk for long-term adverse physical and mental effects, but little is known about the effects of illness in siblings of these patients. In a study reported in The Lancet Oncology, Lasse Wegener Lund, MD, of the Danish Cancer Society Research Centre and Paediatrics and Adolescent Medicine, Juliane Marie Centre, in Copenhagen, and colleagues found that cancer survivors were at long-term risk of mental disorders and that increased risk may also be present in siblings who were young at the time of diagnosis of the patient’s illness.1

Study Details

The study involved assessment of hospital contacts for mental disorders in a population-based cohort of 7,085 Danish children treated for cancer between 1975 and 2010 and in 13,105 of their siblings using data from the Danish Psychiatric Central Research Registry. For the survivor cohort, families with two or more children with cancer were excluded from the analysis, due to anticipated high levels of psychosocial stress.

The sibling and survivor cohorts were compared with two population-based cohorts that did not include childhood cancer survivors or siblings of survivors. For each child treated for cancer, 20 control comparators born on the same day and without cancer at the date of diagnosis of the survivor were identified (n = 144,700), as were siblings of these comparators (n = 260,400).

First hospital inpatient contacts for mental disorders from 1975 to 1994 and both inpatient and outpatient contacts from 1995 to 2011 were included in the analysis. All diagnoses of mental disorders were categorized into two main groups: psychotic disorders, consisting of the subcategories of organic psychosis, schizophrenia and other psychoses, and bipolar depression; and nonpsychotic disorders, including other affective disorders (mainly unipolar depression), anxiety, personality disorder, and two groups defined as neurodevelopmental and other nonpsychotic organic disorders and emotional and behavioral disorders.

For the cancer survivors, age at time of diagnosis was 0 to 4 years in 31%, 5 to 9 years in 18%, 10 to 14 years in 19%, and 15 to 19 years in 32%. The most common diagnoses were CNS tumors (25%), leukemia (24%), and lymphoma (13%).

Overall Risk

Follow-up was conducted for up to 37 years, with a median of 8.8 years for survivors and 17.3 years for comparators. On multivariate analysis adjusting for calendar period (1975–1980 as reference) and hospital contact for mental disorder in a parent or sibling 5 years before cancer diagnosis, male survivors had a 40% increased risk (hazard ratio [HR] = 1.40, 95% confidence interval [CI] = 1.24–1.58) and female survivors had a 20% increased risk (HR = 1.20, 95% CI = 1.05–1.37) for hospital contact for any psychiatric disorder compared with population-based comparators. For both sexes combined, the hazard ratio was 1.38 (95% CI = 1.26–1.51). In a sensitivity analysis of only inpatient contacts, the hazard ratio was 1.51 (95% CI = 1.24–1.84) for male survivors and 1.36 (95% CI = 1.08–1.72) for female survivors. Increased risks were found in survivors of central nervous system (CNS) tumors, hematologic malignancies, and solid tumors.

No increased risk of hospital contact was found for brothers (HR = 0.99, 95% CI = 0.91–1.08) or sisters (HR = 1.01, 95% CI = 0.93–1.10) of survivors overall or in specific psychiatric subtypes.

Risk of Particular Disorders

Male survivors were at significantly increased risk of all psychotic disorders (HR = 1.39, 95% CI = 1.07–1.81), all nonpsychotic disorders (HR = 1.45, 95% CI = 1.28–1.64), unipolar depression (HR = 1.55, 95% CI = 1.19–2.01), neurodevelopmental and other nonpsychotic organic disorders (HR = 1.77, 95% CI = 1.44–2.17), and emotional and behavioral disorders (HR = 1.50, 95% CI = 1.27–1.76). There were too few cases (≤ 30) of organic psychosis, bipolar depression, and anxiety in male survivors to be included in the multivariate analysis; risk for organic psychosis was significantly increased on univariate analysis.

Female survivors were at significantly increased risk of all nonpsychotic disorders (HR = 1.18, 95% CI = 1.03–1.36), neurodevelopmental and other nonpsychotic organic disorders (HR = 2.29, 95% CI = 1.69–3.09), and emotional and behavioral disorders (HR = 1.21, 95% CI = 1.02–1.45). There were too few cases of organic psychosis, schizophrenia and other psychoses, and bipolar depression to be included in multivariate analysis; risk of organic psychosis was significantly increased on univariate analysis.

Effect of Age at Diagnosis

Among survivors, risk was significantly increased among males in all age groups (HRs = 1.31–1.71); risk was greater among those aged 0 to 9 years (HRs = 1.66–1.71) compared with those aged 10 to 19 years (HRs = 1.31–1.44), although risk did not differ significantly across age groups (P = .31). Among female survivors, risk was higher with younger age (P = .043 across age groups), with risk significantly increased in those diagnosed at 0 to 4 years (HR = 1.67) and nonsignificantly increased at ages 5 to 14 (HRs = 1.27–1.28). There were no significant trends in risk by number of years since cancer diagnosis in males (P = .84) or females (P = .56)

Among siblings, the effect of age at the time of their sibling’s diagnosis was significant for both brothers (P = .0045) and sisters (P < .0001), with risk in both brothers and sisters being greater when they were 0 to 9 years of age at the time of diagnosis (HRs = 1.12–1.24, significant for brothers who were aged 5 to 9 years) than when they were older (HRs = 0.79–0.97); both brothers (HR = 0.79) and sisters (HR = 0.83) had risk significantly below that of the general population when they were aged > 15 years at the time of diagnosis.

The greatest risk was among sisters born after their sibling’s diagnosis (HR = 1.59, 95% CI = 1.30–1.95). Young age at diagnosis was associated with increased risk for behavioral and emotional disorders in both brothers (P < .0001 for trend) and sisters (P = .0018) and for schizophrenia (P = .0034), unipolar depression (P < .0001), and anxiety (P = .011) in sisters. The risk for hospital contact for a mental disorder increased with longer time since sibling’s cancer diagnosis in sisters (P = .0023) but not in brothers (P = .28).

The investigators concluded, “Childhood cancer survivors should be followed up for mental late effects, especially those diagnosed in young age. Further, clinicians should also be aware that siblings who were young at the time of cancer diagnosis might be at increased risk for mental health disorders.” ■

Disclosure: The study authors reported no potential conflicts of interest.


1. Lund LW, Winther JF, Dalton SO, et al: Hospital contact for mental disorders in survivors of childhood cancer and their siblings in Denmark: A population-based cohort study. Lancet Oncol 14:971-980, 2013.

Related Articles

COMMENTARY: Psychosocial Health in Long-Term Survivors of Childhood Cancer

The study by Lund and colleagues discussed in this issue of The ASCO Post reinforces concerns about psychosocial health in long-term survivors of childhood cancer. Taking advantage of national registries, these investigators demonstrated that, when compared to the general population, survivors are...