Matched cohort study of hospitalization in children who have siblings with cancer
Abstract
Background
Health outcomes of children in families affected by cancer are poorly understood. The authors assessed the risk of hospitalization in children who have a sibling with cancer.
Methods
This was a longitudinal cohort study in which 1600 children who had a sibling with cancer were matched to 32,000 children who had unaffected siblings in Quebec, Canada, from 2006 to 2020. The exposure of interest was having a sibling with cancer. Outcomes included hospitalization for pneumonia, asthma, fracture, and other morbidities any time after the sibling was diagnosed with cancer. The children were followed over time, and Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the impact of having a sibling with cancer on the risk of hospitalization before age 14 years, adjusted for patient characteristics.
Results
Children who had a sibling with cancer had an increased risk of hospitalization compared with unaffected children (HR, 1.15; 95% CI, 1.02-1.29). Conditions associated with a greater risk of hospitalization included pneumonia, hemangioma, other skin conditions, sleep apnea, and inflammatory bowel disease. The risk of hospitalization was greatest for children whose older sibling had cancer (HR, 1.16; 95% CI, 1.01-1.32) and for children whose sibling had hematopoietic cancer (HR, 1.22; 95% CI, 1.01-1.48).
Conclusions
Children who have a sibling with cancer are at risk of hospitalization for conditions such as pneumonia, inflammatory bowel disease, and other morbidities. Families affected by childhood cancer may benefit from additional support to facilitate care for all children in the family.
Lay Summary
- Little is known about the health of children who have a brother or sister with cancer.
- The authors studied the types of hospitalization experienced by children who have siblings with cancer.
- The results indicated that having a sibling with cancer increased the chance of being hospitalized for pneumonia and other conditions that could have been preventable.
- The results also indicated that children who had an older sibling with cancer or a sibling with blood cancer had a greater chance of being hospitalized.
- The findings highlight the importance of providing timely care for children in families affected by childhood cancer.
Introduction
Not much is known about the long-term outcomes of children who have siblings with cancer. In the United States, over 10,000 children younger than 15 years are diagnosed with cancer annually.1 Most children with cancer require complex care and have siblings who may be affected by disruptions in family function and priorities.2 Treatment of childhood cancer is time-consuming and may draw parental attention away from other children in the family.3, 4 Parents report that 1 of the most challenging caregiving responsibilities while managing childhood cancer is meeting the needs of healthy siblings.2 Yet there is a scarcity of studies on the health outcomes of children whose siblings have cancer.
Most studies focus on mental health outcomes of children in families affected by cancer. Elevated levels of distress and symptoms of posttraumatic stress, anxiety, and depression have all been reported in children who have a sibling with cancer.3 Studies have shown that having a sibling with cancer is associated with a higher frequency of inpatient and outpatient visits for mental health care compared with having unaffected siblings.5-7 Having a sibling with cancer is a considerable source of stress8 and may be associated with somatic symptoms.3, 9 However, data on the risk of acute or chronic morbidity are largely absent. Preventable morbidities, such as injuries and infections, may be harder to mitigate because of decreased parental monitoring and time to consult health care professionals.4, 9 Because there is a lack of data on morbidity after a sibling is diagnosed with cancer, we prepared a portrait of hospitalization for a representative sample of children who have siblings with cancer using longitudinal data up to age 14 years.
Materials and Methods
Study Design and Population
We conducted a matched cohort study of children born in hospitals of Quebec, Canada, between 2006 and 2019 who were registered in the Maintenance and Use of Data for the Study of Hospital Clientele data set. The data comprise discharge abstracts for all hospitals in the province of Quebec and capture nearly all children in the population because >98% are hospitalized at birth. Each child has a unique health insurance number that facilitates the identification of hospitalizations anywhere in the province. Children are linked with their mother and siblings in the data.
The analysis was restricted to families with at least 2 children. Following previous literature, we did not include 9 families in which more than 1 sibling had cancer.6 The data contain all childhood cancers in Quebec because the diagnosis, treatment, and management of tumors is provided in hospital. Validation studies have shown that the data have more coverage than the provincial cancer registry.10 Childhood cancers are coded with morphology and topography codes in the International Classification of Diseases for Oncology, third edition.11
Exposure
The exposed group comprised children who had a sibling with cancer but were themselves free of any malignancy. We identified children who had siblings with hematopoietic cancers (leukemia, lymphoma) or solid tumors (retinoblastoma, neuroblastoma, and malignant neoplasms of the central nervous system, kidneys, liver, lungs, bones, soft tissues, germ cells, and integument). We matched each exposed child to 20 children who did not have cancer and whose siblings also were free of cancer. We matched the children on sex, birth order (first, second, third, fourth or higher born), total number of siblings (2, 3, or ≥4), preterm birth (<37 or ≥37 weeks), and year of birth.
We followed the children over time to identify hospitalizations before age 14 years that occurred between 2006 and 2020. Data past age 14 years, or after 2020, were not available. We did not include children who died at birth or had invalid insurance numbers and could not be followed over time.
Outcome
The main outcome measure was hospitalization for any morbidity before age 14 years. We categorized hospitalizations according to chapters in the tenth revision of the International Classification of Diseases (see Supporting Table 1). We also identified specific conditions that required treatment or management, including otitis, upper respiratory infection, pneumonia, infectious enteritis, dermatitis and eczema, nephritis, appendicitis, dental caries, asthma, allergy, autoimmune disease, inflammatory bowel disease, diabetes and metabolic disorders, anemia, hemangioma, hernia, gastroesophageal reflux, sleep apnea and other sleep disorders, arthropathy, fracture, and other injuries.
Data Analysis
We examined characteristics of the cohort (numbers and percentages), and calculated hospitalization rates with 95% confidence intervals (CIs). By using stratified Cox models, we compared the risk of childhood hospitalization between children who had a sibling with cancer and the matched comparison group. We estimated hazard ratios (HRs) and 95% CIs for the effect of having a sibling with cancer on the risk of childhood hospitalization. We also adjusted the models for maternal age at birth (aged <25, 25-34, or ≥35 years), pregnancy morbidity (preexisting or gestational diabetes, obesity, dyslipidemia, essential or gestational hypertension, preeclampsia), and socioeconomic disadvantage (most deprived quintile of the population based on neighborhood levels of education, income, and employment).
The time scale was measured in days. We began follow-up on the date of the sibling's cancer diagnosis. For children who were born only after the sibling developed cancer, we began follow-up on their date of birth. Follow-up ended at the time of first hospitalization for each morbidity or at the end of the study in 2020. We censored children who were never hospitalized during follow-up.
In secondary analyses, we considered whether the relative age of the child (younger vs older than the sibling with cancer) modified the results. We also determined whether the type of cancer (hematopoietic vs solid tumor) influenced the findings. In sensitivity analyses, we stratified the analysis based on whether siblings with cancer died during follow-up to determine whether the risk of hospitalization was affected by survival of the sibling. We evaluated whether children had an elevated risk of hospitalization even before the sibling was diagnosed with cancer by starting follow-up at birth rather than at the time of cancer diagnosis.
We used SAS version 9.4 (SAS Institute Inc) to analyze the data. Because the data were de-identified and informed consent was not required, we obtained an ethics waiver from the University of Montreal Hospital Center.
Results
There were 33,600 children in the study, including 1600 who had a sibling with cancer matched to 32,000 children who had unaffected siblings. The distribution of matched characteristics, including sex, total number of siblings, birth order, preterm birth, and year of birth, was similar between exposed and unexposed children (Table 1). The distribution of unmatched characteristics, including maternal age, pregnancy morbidity, and socioeconomic disadvantage, did not vary noticeably between exposed and unexposed children.
| Characteristic | No. of Children (%) | |
|---|---|---|
| Children Whose Siblings Have Cancer | Children With Unaffected Siblings | |
| Sexa | ||
| Male | 809 (50.6) | 16,180 (50.6) |
| Female | 791 (49.4) | 15,820 (49.4) |
| Total no. of siblingsa | ||
| 2 | 637 (39.8) | 12,740 (39.8) |
| 3 | 583 (36.4) | 11,660 (36.4) |
| ≥4 | 380 (23.8) | 7600 (23.8) |
| Birth ordera | ||
| Firstborn | 449 (28.1) | 8980 (28.1) |
| Second born | 658 (41.1) | 13,160 (41.1) |
| Third born | 324 (20.3) | 6480 (20.3) |
| Fourth born or higher | 169 (10.6) | 3380 (10.6) |
| Preterm birth, <37 wka | ||
| Yes | 99 (6.2) | 1980 (6.2) |
| No | 1501 (93.8) | 30,020 (93.8) |
| Year of birtha | ||
| 2006-2010 | 623 (38.9) | 12,241 (38.3) |
| 2011-2015 | 691 (43.2) | 13,696 (42.8) |
| 2016-2019 | 286 (17.9) | 6063 (18.9) |
| Maternal age at birth, y | ||
| <25 | 259 (16.2) | 4954 (15.5) |
| 25-34 | 1052 (65.8) | 21,979 (68.7) |
| ≥35 | 289 (18.1) | 5067 (15.8) |
| Pregnancy morbidityb | ||
| Yes | 241 (15.1) | 4457 (13.9) |
| No | 1359 (84.9) | 27,543 (86.1) |
| Socioeconomic disadvantage | ||
| Yes | 345 (21.6) | 6784 (21.2) |
| No | 1190 (74.4) | 23,964 (74.9) |
| Total | 1600 (100.0) | 32,000 (100.0) |
- a Children are matched on sex, total number of siblings, birth order, preterm birth, and year of birth.
- b Pregnancy morbidity includes preexisting or gestational diabetes, obesity, dyslipidemia, essential or gestational hypertension, and preeclampsia.
Children who had a sibling with cancer had a higher risk of hospitalization before age 14 years compared with children who had unaffected siblings (Table 2). Having a sibling with cancer was associated with 1.15 times the risk of any hospitalization in childhood (95% CI, 1.02-1.29). The risk of hospitalization was significantly greater for nervous system (HR, 1.57; 95% CI, 1.05-2.35), respiratory (HR, 1.18; 95% CI, 1.01-1.40), digestive (HR, 1.38; 95% CI, 1.09-1.74), and skin (HR, 1.60; 95% CI, 1.13-2.28) disorders.
| Type of Hospitalization | No. of Children Hospitalized | Hospitalization Rate per 1000 Person-Years (95% CI) | HR (95% CI)a | ||
|---|---|---|---|---|---|
| Children Whose Siblings Have Cancer | Children With Unaffected Siblings | Children Whose Siblings Have Cancer | Children With Unaffected Siblings | ||
| Any hospitalization | 295 | 5279 | 50.7 (45.2-56.8) | 44.7 (43.5-45.9) | 1.15 (1.02-1.29) |
| Infectious disease | 72 | 1205 | 10.7 (8.48-13.5) | 8.97 (8.47-9.49) | 1.19 (0.94-1.52) |
| Benign neoplasm | 9 | 112 | 1.29 (0.67-2.47) | 0.80 (0.67-0.97) | 1.64 (0.83-3.24) |
| Blood disorder | 33 | 497 | 4.79 (3.40-6.73) | 3.61 (3.31-3.94) | 1.35 (0.95-1.92) |
| Endocrine disorder | 37 | 661 | 5.37 (3.89-7.41) | 4.83 (4.47-5.21) | 1.13 (0.81-1.58) |
| Mental and behavioral | 15 | 211 | 2.15 (1.29-3.56) | 1.52 (1.33-1.74) | 1.41 (0.83-2.39) |
| Nervous system | 26 | 324 | 3.74 (2.55-5.50) | 2.34 (2.10-2.61) | 1.57 (1.05-2.35) |
| Eye and adnexa | 18 | 245 | 2.59 (1.63-4.11) | 1.77 (1.56-2.00) | 1.44 (0.89-2.32) |
| Ear and mastoid process | 112 | 2012 | 17.1 (14.2-20.6) | 15.3 (14.7-16.0) | 1.14 (0.94-1.38) |
| Circulatory system | 8 | 186 | 1.14 (0.57-2.28) | 1.34 (1.16-1.55) | 0.88 (0.43-1.78) |
| Respiratory system | 152 | 2642 | 23.8 (20.3-27.9) | 20.5 (19.8-21.3) | 1.18 (1.01-1.40) |
| Digestive system | 74 | 1074 | 10.9 (8.68-13.7) | 7.92 (7.46-8.41) | 1.38 (1.09-1.74) |
| Skin disease | 34 | 431 | 4.93 (3.52-6.89) | 3.13 (2.84-3.44) | 1.60 (1.13-2.28) |
| Musculoskeletal system | 12 | 212 | 1.72 (0.98-3.02) | 1.53 (1.33-1.75) | 1.10 (0.61-1.97) |
| Genitourinary system | 29 | 544 | 4.18 (2.90-6.01) | 3.96 (3.64-4.31) | 1.07 (0.74-1.56) |
| Ill defined conditions | 66 | 1155 | 9.73 (7.64-12.4) | 8.56 (8.08-9.06) | 1.16 (0.90-1.48) |
| Injury | 50 | 768 | 7.28 (5.52-9.61) | 5.61 (5.23-6.02) | 1.30 (0.97-1.73) |
| Other | 76 | 1012 | 11.3 (9.04-14.2) | 7.45 (7.01-7.93) | 1.53 (1.21-1.94) |
- Abbreviations: CI, confidence interval; HR, hazard ratio.
- a HRs are for children whose siblings have cancer versus children with unaffected siblings, adjusted for maternal age, pregnancy morbidity, and socioeconomic disadvantage.
The risk of hospitalization was greater for specific morbidities (Table 3). Having a sibling with cancer was associated with 1.70 times the risk of hospitalization for pneumonia (95% CI, 1.27-2.28) and 1.83 times the risk of hospitalization for sleep apnea (95% CI, 1.12-2.97). Risks were additionally elevated for hemangioma treatment (HR, 3.19; 95% CI, 1.34-7.59), complications of dermatitis (HR, 1.97; 95% CI, 1.26-3.09), and inflammatory bowel disease (HR, 2.85; 95% CI, 1.10-7.35).
| Morbidity | No. of Children Hospitalized | Hospitalization Rate per 10,000 Person-Years (95% CI) | HR (95% CI)a | ||
|---|---|---|---|---|---|
| Children Whose Siblings Have Cancer | Children With Unaffected Siblings | Children Whose Siblings Have Cancer | Children With Unaffected Siblings | ||
| Otitis | 109 | 1962 | 166.1 (137.7-200.4) | 149.4 (143.0-156.2) | 1.13 (0.93-1.38) |
| Upper respiratory infection | 43 | 785 | 62.9 (46.7-84.8) | 57.7 (53.8-61.8) | 1.12 (0.82-1.52) |
| Pneumonia | 49 | 588 | 71.8 (54.3-95.0) | 42.9 (39.6-46.5) | 1.70 (1.27-2.28) |
| Infectious enteritis | 19 | 319 | 27.3 (17.4-42.8) | 23.1 (20.7-25.8) | 1.21 (0.76-1.92) |
| Dermatitis and eczema | 21 | 222 | 30.2 (19.7-46.3) | 16.0 (14.0-18.3) | 1.97 (1.26-3.09) |
| Nephritis | 15 | 229 | 21.5 (13.0-35.7) | 16.5 (14.5-18.8) | 1.28 (0.76-2.17) |
| Appendicitis | 7 | 112 | 9.97 (4.76-20.9) | 8.04 (6.68-9.67) | 1.25 (0.58-2.68) |
| Dental caries | 29 | 429 | 41.8 (29.1-60.2) | 31.1 (28.3-34.1) | 1.32 (0.90-1.92) |
| Asthma | 23 | 369 | 33.1 (22.0-49.8) | 26.7 (24.1-29.6) | 1.26 (0.82-1.92) |
| Allergy | 127 | 2224 | 195.5 (164.3-232.7) | 170.7 (163.7-177.9) | 1.16 (0.97-1.39) |
| Autoimmune disease | 6 | 84 | 8.56 (3.84-19.0) | 6.03 (4.87-7.47) | 1.41 (0.61-3.25) |
| Inflammatory bowel disease | 5 | 33 | 7.13 (2.97-17.1) | 2.37 (1.68-3.33) | 2.85 (1.10-7.35) |
| Diabetes and metabolic disorders | 8 | 126 | 11.4 (5.70-22.8) | 9.06 (7.60-10.8) | 1.27 (0.62-2.59) |
| Anemia | 19 | 288 | 27.3 (17.4-42.8) | 20.8 (18.5-23.4) | 1.33 (0.83-2.11) |
| Hemangioma | 6 | 38 | 8.56 (3.85-19.1) | 2.79 (1.98-3.75) | 3.19 (1.34-7.59) |
| Hernia | 8 | 165 | 11.4 (5.71-22.8) | 11.9 (10.2-13.8) | 0.94 (0.46-1.91) |
| Gastroesophageal reflux | 11 | 129 | 15.7 (8.71-28.4) | 9.28 (7.81-11.0) | 1.80 (0.97-3.35) |
| Sleep apnea | 18 | 199 | 25.8 (16.3-41.0) | 14.3 (12.5-16.5) | 1.83 (1.12-2.97) |
| Arthropathy | 6 | 50 | 8.57 (3.85-19.1) | 3.59 (2.72-4.73) | 2.32 (0.99-5.43) |
| Fracture | 13 | 234 | 18.6 (10.8-32.0) | 16.9 (14.8-19.2) | 1.11 (0.64-1.95) |
| Other injuriesb | 5 | 103 | 7.13 (2.97-17.1) | 7.40 (6.72-8.15) | 1.00 (0.41-2.46) |
- Abbreviations: CI, confidence interval; HR, hazard ratio.
- a HRs are for children who have a sibling with cancer versus children with unaffected siblings, adjusted for maternal age, pregnancy morbidity, and socioeconomic disadvantage.
- b Other injuries include burns, foreign body, poisoning, and drowning.
The relative age of the sibling with cancer modified the risk of hospitalization (Table 4). Compared with unaffected children, having an older sibling with cancer was associated with 1.16 times the risk of hospitalization (95% CI, 1.02-1.32), whereas having a younger sibling with cancer was not significantly associated with the risk of hospitalization (HR, 1.10; 95% CI, 0.84-1.44). Children whose older sibling had cancer had a higher risk of hospitalization for respiratory (HR, 1.27; 95% CI, 1.06-1.52), digestive (HR, 1.53; 95% CI, 1.16-2.02), and skin (HR, 1.55; 95% CI, 1.05-2.30) disorders.
| Type of Hospitalization | Younger Sibling With Cancer | Older Sibling With Cancer | ||||
|---|---|---|---|---|---|---|
| Hospitalization Rate per 1000 Person-Years | HR (95% CI)a | Hospitalization Rate per 1000 Person-Years | HR (95% CI)a | |||
| Exposed, N = 661 | Unexposed, N = 13,220 | Exposed, N = 939 | Unexposed, N = 18,780 | |||
| Any hospitalization | 22.5 | 20.7 | 1.10 (0.84-1.44) | 71.1 | 61.7 | 1.16 (1.02-1.32) |
| Infectious disease | 4.63 | 3.83 | 1.19 (0.66-2.13) | 14.5 | 12.2 | 1.19 (0.92-1.55) |
| Benign neoplasm | 0.38 | 0.63 | 0.65 (0.09-4.76) | 1.84 | 0.91 | 2.06 (0.99-4.27) |
| Blood disorder | 2.69 | 1.68 | 1.68 (0.77-3.64) | 6.06 | 4.80 | 1.29 (0.87-1.92) |
| Endocrine disorder | 2.29 | 2.28 | 0.99 (0.44-2.26) | 7.26 | 6.41 | 1.16 (0.81-1.67) |
| Mental and behavioral | 2.68 | 1.64 | 1.64 (0.76-3.55) | 1.83 | 1.45 | 1.28 (0.62-2.62) |
| Nervous system | 3.06 | 1.64 | 1.82 (0.88-3.76) | 4.15 | 2.77 | 1.48 (0.92-2.40) |
| Eye and adnexa | 1.90 | 1.08 | 1.74 (0.69-4.36) | 3.00 | 2.18 | 1.36 (0.77-2.39) |
| Ear and mastoid process | 6.26 | 5.97 | 1.05 (0.64-1.74) | 24.1 | 21.4 | 1.15 (0.94-1.42) |
| Circulatory system | 1.14 | 0.99 | 1.15 (0.36-3.68) | 1.15 | 1.55 | 0.76 (0.31-1.86) |
| Respiratory system | 8.22 | 9.94 | 0.84 (0.54-1.30) | 34.2 | 27.4 | 1.27 (1.06-1.52) |
| Digestive system | 7.34 | 6.89 | 1.04 (0.66-1.66) | 13.1 | 8.55 | 1.53 (1.16-2.02) |
| Skin disease | 2.69 | 1.50 | 1.89 (0.87-4.11) | 6.28 | 4.12 | 1.55 (1.05-2.30) |
| Musculoskeletal system | 1.14 | 1.35 | 0.86 (0.27-2.74) | 2.07 | 1.63 | 1.22 (0.62-2.41) |
| Genitourinary system | 4.21 | 2.47 | 1.72 (0.93-3.19) | 4.16 | 4.88 | 0.86 (0.54-1.39) |
| Ill defined conditions | 3.86 | 4.15 | 0.93 (0.49-1.76) | 13.3 | 11.3 | 1.21 (0.92-1.58) |
| Injury | 6.19 | 4.23 | 1.44 (0.86-2.39) | 7.94 | 6.46 | 1.24 (0.88-1.75) |
| Other | 7.01 | 4.03 | 1.74 (1.07-2.81) | 14.0 | 9.58 | 1.48 (1.13-1.94) |
- Abbreviations: CI, confidence interval; HR, hazard ratio.
- a HRs are for children who have a sibling with cancer versus children with unaffected siblings, adjusted for maternal age, pregnancy morbidity, and socioeconomic disadvantage.
When we examined the type of neoplasm, having a sibling with hematopoietic cancer was associated with a somewhat greater risk of hospitalization (Table 5). Compared with unaffected children, children who had siblings with hematopoietic cancer had 1.22 times the risk of hospitalization (95% CI, 1.01-1.48). In contrast, children who had siblings with solid tumors had a less pronounced risk of hospitalization (HR, 1.11; 95% CI, 0.95-1.28). Children whose siblings had hematopoietic cancer were at risk of nervous system disorders (HR, 2.17; 95% CI, 1.16-4.07) and skin disorders (HR, 1.77; 95% CI, 1.00-3.13), whereas children whose siblings had solid tumors were at risk of benign neoplasms (HR, 2.24; 95% CI, 1.07-4.67).
| Type of Hospitalization | Sibling With Hematopoietic Cancer | Sibling With Solid Tumor | ||||
|---|---|---|---|---|---|---|
| Hospitalization Rate per 1000 Person-Years | HR (95% CI)a | Hospitalization Rate per 1000 Person-Years | HR (95% CI)a | |||
| Exposed, N = 564 | Unexposed, N = 11,280 | Exposed, N = 1050 | Unexposed, N = 21,000 | |||
| Any hospitalization | 54.1 | 44.1 | 1.22 (1.01-1.48) | 49.1 | 45.3 | 1.11 (0.95-1.28) |
| Infectious disease | 12.2 | 8.45 | 1.40 (0.97-2.04) | 9.89 | 9.35 | 1.08 (0.79-1.47) |
| Benign neoplasm | 0.39 | 0.75 | 0.52 (0.07-3.81) | 1.78 | 0.82 | 2.24 (1.07-4.67) |
| Blood disorder | 5.97 | 3.61 | 1.69 (0.99-2.86) | 4.27 | 3.62 | 1.20 (0.75-1.90) |
| Endocrine disorder | 7.60 | 4.94 | 1.52 (0.95-2.42) | 4.03 | 4.80 | 0.87 (0.54-1.40) |
| Mental and behavioral | 2.75 | 1.46 | 1.81 (0.83-3.93) | 1.77 | 1.57 | 1.15 (0.56-2.35) |
| Nervous system | 4.35 | 2.10 | 2.17 (1.16-4.07) | 3.57 | 2.51 | 1.42 (0.85-2.36) |
| Eye and adnexa | 1.57 | 1.60 | 0.96 (0.35-2.63) | 3.13 | 1.84 | 1.69 (0.98-2.92) |
| Ear and mastoid process | 15.8 | 15.4 | 1.03 (0.74-1.43) | 17.6 | 15.5 | 1.16 (0.92-1.47) |
| Circulatory system | 1.96 | 1.48 | 1.34 (0.54-3.31) | 0.66 | 1.27 | 0.54 (0.17-1.70) |
| Respiratory system | 22.3 | 20.7 | 1.10 (0.83-1.46) | 24.5 | 20.5 | 1.22 (0.99-1.49) |
| Digestive system | 10.1 | 8.07 | 1.23 (0.82-1.84) | 11.2 | 7.90 | 1.44 (1.07-1.93) |
| Skin disease | 5.16 | 2.95 | 1.77 (1.00-3.13) | 4.95 | 3.20 | 1.59 (1.03-2.45) |
| Musculoskeletal system | 1.96 | 1.58 | 1.36 (0.55-3.37) | 1.56 | 1.49 | 0.98 (0.46-2.11) |
| Genitourinary system | 4.74 | 4.22 | 1.12 (0.63-2.01) | 4.02 | 3.82 | 1.08 (0.67-1.74) |
| Ill defined conditions | 10.2 | 8.10 | 1.24 (0.83-1.86) | 9.56 | 8.88 | 1.12 (0.82-1.53) |
| Injury | 7.19 | 5.56 | 1.30 (0.81-2.10) | 7.44 | 5.62 | 1.33 (0.93-1.89) |
| Other | 14.1 | 7.06 | 1.93 (1.36-2.76) | 9.61 | 7.73 | 1.27 (0.93-1.74) |
- Abbreviations: CI, confidence interval; HR, hazard ratio.
- a HRs are for children who have a sibling with cancer versus children with unaffected siblings, adjusted for maternal age, pregnancy morbidity, and socioeconomic disadvantage.
In sensitivity analyses, children whose siblings died from cancer were not at increased risk of hospitalization (HR, 0.70; 95% CI, 0.45-1.08) compared with children who had unaffected siblings. However, children whose siblings survived cancer had an increased risk of hospitalization (HR, 1.21; 95% CI, 1.07-1.36). When we began follow-up at birth rather than at the time of the sibling's cancer diagnosis, the risk of hospitalization remained elevated although somewhat attenuated. Children who had a sibling with cancer had 1.08 times the risk of hospitalization (95% CI, 1.00-1.17) regardless of when the sibling was diagnosed with cancer.
Discussion
In this matched cohort study of 33,600 children with follow-up extending to age 14 years, having a sibling with cancer was associated with a 15% greater risk of hospitalization compared with having unaffected siblings. Children who had a sibling with cancer had an elevated risk of hospitalization for pneumonia, skin complications, hemangioma, inflammatory bowel disease, and sleep apnea. Children whose older sibling had cancer and children who had a sibling with hematopoietic cancer had the greatest risk of hospitalization. The findings suggest that children in families affected by childhood cancer have an elevated risk of hospitalization, some of which may be avoidable. Clinicians and caregivers should be alert to the excess risk of morbidity and anticipate strategies to facilitate timely health care for children living in families affected by child cancer.
Previous research of children who have siblings with cancer has focused primarily on psychosocial morbidity.3 A systematic review of over 100 studies published before 2016 found no major difference in levels of anxiety, depression, or emotional and behavioral function between children who have siblings with cancer and children who do not.3 However, studies rarely used a longitudinal design and frequently had nonrepresentative comparison groups.3 The few existing longitudinal cohort studies on this topic found conflicting evidence that having a sibling with cancer increased the risk of mental health care use.5-7 In our study, children whose siblings had cancer did not have a statistically significant risk of mental or behavioral hospitalization before age 14 years compared with matched controls, although we only analyzed events that were serious enough to require hospitalization.
Less is known about whether having a sibling with cancer affects physical morbidity. Cross-sectional studies suggest that somatic symptoms, such as headache, bowel problems, and sleeping problems, may be common in children whose siblings have cancer.9, 12 Observational studies suggest that morbidity may be elevated for all children in families experiencing cancer.13 However, the findings mostly stem from data focused on the child cancer survivor.3, 13 There is usually little focus on other children in the family. In 1 study, survivors of childhood acute lymphoblastic leukemia were found to have a similar risk of hospitalization as other children in the family but a greater risk compared with matched population controls.13 Digestive, musculoskeletal, and circulatory disorders were some of the more common causes of hospitalization in children whose siblings had leukemia.13
Stress from having a sibling with cancer may compound the risk of diseases over the life course.14 Stressful childhood environments can disturb the structure and function of the immune system and other organs during sensitive developmental periods.15 Childhood stress is associated with dysregulation in innate immunity and promotion of a proinflammatory state.16 A study of 169 families found that chronic stress was associated with greater natural killer cell activity, as well as respiratory tract infections, skin ailments, and nonspecific abdominal complaints in children.17 In our study, having a sibling with cancer was associated with several immune-related conditions, including pneumonia, dermatitis and eczema, and inflammatory bowel disease. There was also an increased risk of sleep apnea, a condition sensitive to immune risk factors like adenotonsillar hypertrophy.18 Therefore, stress-related effects on the immune system may contribute substantially to morbidity in children who have siblings with cancer.
Several aspects of our findings point to caregiving challenges as another potential contributor to morbidity in children whose siblings have cancer. Conditions like pneumonia and complications of inflammatory bowel disease or dermatitis and eczema were frequent in our data. Although early ambulatory care for upper respiratory infections or skin problems ultimately may prevent severe complications requiring hospitalization,19 1 study found that parents underestimate the number and intensity of physical symptoms in children whose siblings have cancer.9 Ambulatory care may be underused by parents who are more focused on the health care needs of the child with cancer than those of other children in the family.2, 3, 9 Interestingly, death of the sibling with cancer was associated with a lower risk of hospitalization in our analysis. A potential explanation could be that parents find it easier to provide care to other children after the passing of the sibling.20
Caregiving challenges may also explain why having an older sibling with cancer was associated with a higher risk of morbidity. Younger children depend on parents to detect and treat symptoms of acute illnesses at an early stage to avoid disease progression. In our data, younger children who had an older sibling with cancer had elevated risks of hospitalization for respiratory, digestive, and skin complications. Young children with these conditions may develop complications more quickly and require more timely intervention to avoid hospitalization.19 Parents who are preoccupied by an older child with cancer may have less time and ability to manage health issues in younger children.2 In contrast, older children may be able to advocate for their needs should symptoms progress or require medical attention.4, 21, 22
The type of cancer appeared to modify the risk of morbidity because children whose siblings had hematopoietic cancer had a greater risk of hospitalization than children whose siblings had solid tumors. Hematopoietic cancer may be associated with treatment-related stressors that affect other children in the family. Hematopoietic tumors frequently require systemic chemotherapy and stem cell transplantation, whereas solid tumors tend to be treated locally with surgery or radiation.23 Stem cell transplantation procedures can extend over several weeks and require close attention to prevent infections in the first 6 months posttransplantation.23, 24 Treatment for hematopoietic cancer often lasts years to maintain remission and includes inpatient care, prolonging parental time away from home, and diminished attention to the needs of other children.4, 25 Studies have shown that parental stress may be elevated up to 18 months posttransplantation.24
Some of the excess risk of hospitalization could be explained by an underlying familial predisposition to morbidity or a lower threshold for admission because of parental anxiety or physician precaution. Several childhood cancers are caused by genetic syndromes or alleles that could predispose family members to immune dysfunction.26, 27 In our data, children whose siblings had cancer were more frequently hospitalized for treatment of hemangioma, a benign vascular tumor that clusters with malignant cancer in families.28 For some children, the risk of hospitalization was elevated even before the sibling was diagnosed with cancer. However, genetic pathways likely only account for a small fraction of cases because genetic syndromes are rare,29 and we found associations with a range of disorders that are not easily explained by genetic pathways alone. The excess risk of hospitalization when the older sibling had cancer raises the possibility of a lower threshold for admission stemming from parental anxiety, better awareness of care, or physician precaution.30 However, increased care seeking may also lead to earlier treatment that prevents hospitalization.
This study had several strengths, including the use of matched longitudinal data comprising confirmed morbidities, which required hospitalization, rather than self-reported disorders, which may be inaccurately diagnosed. We used validated outcomes, not symptoms or mild conditions that were treated outside of hospital, which may be more susceptible to bias. However, there are limitations. We captured all childhood cancers during the study, but the number of cases precluded us from analyzing different cancer types separately. We matched children on key characteristics, but there may be residual confounding. We did not have information on ethnicity, parental occupation, childcare, type or duration of cancer treatment, paternity, or genetic syndromes that predispose families to cancer. Future studies should consider the effect of these factors on the risk of hospitalization. Because we used an administrative data set, coding errors may have led to misclassification and attenuated associations toward the null. The data reflect a jurisdiction with universal access to health care, and generalizability to other settings is to be determined.
Conclusion
In this matched cohort study of nearly 34,000 children, having a sibling with cancer was associated with an increased risk of hospitalization compared with having unaffected siblings. Some of the strongest risk was observed for conditions in which hospitalization was potentially avoidable, such as pneumonia, dermatitis, and inflammatory bowel disease. The risk of hospitalization was higher when the sibling was older or had hematopoietic cancer. The findings underscore the notion that children whose siblings have cancer are at higher risk of morbidity after the cancer diagnosis. Because the overall risk of hospitalization was modest, more data will be needed to determine whether facilitating access to early ambulatory care may have a clinical impact on morbidity in families experiencing childhood cancer. Closer clinical follow-up of children whose siblings have cancer may be justified, in addition to supporting psychosocial health as a standard of care.
Funding Support
This research was funded by the Canadian Institutes of Health Research (PCC-170244) and the Fonds de Recherche du Quebec-Sante (NA 296785; SM 284477).
Conflict of Interest Disclosures
The authors made no disclosures.
Author Contributions
Nathalie Auger: Conceived and designed the study, interpreted the results, drafted the initial article, and reviewed and revised the article. Sophie Marcoux, Philippe Bégin, Antoine Lewin, and Thuy Mai Luu: Conceived and designed the study, interpreted the results, and critically revised the article for important intellectual content. Ga Eun Lee: Conceived and designed the study, analyzed the data, interpreted the results, and drafted the initial article. Jessica Healy-Profitós: Conceived and designed the study, interpreted the results, and drafted the initial article. All authors approved the final version as submitted and agree to be accountable for all aspects of the work.
