Specific genetic syndromes play a role in the etiology of only a very small proportion of hematopoietic malignancies in children.1, 2 The syndromes include familial neoplastic syndromes, inherited immunodeficiency and bone marrow failure syndromes.1 The potential role of inherited susceptibility in childhood acute leukemia (AL) and lymphoma has yet to be elucidated. Large population-based registry-linked studies have investigated the familial aggregation of Hodgkin's lymphoma (HL), non-Hodgkin's lymphoma (NHL) or leukemia.3–7 In children, few population-based registry-linked studies have been published,8–13 most of them relating to first-degree relatives.8–12 The estimated relative risks for a positive family history of cancer ranged from 0.8 to 1.0 for childhood leukemia8, 10–13 and from 1.0 to 1.8 for childhood lymphoma taken as a whole.8, 10–12 A number of case–control studies have addressed the association between history of cancer in first and second degree relatives of children with leukemia and observed OR from 1.1 to 1.6.14–18 Only 1 case–control study has been published for childhood HL and NHL.15 Overall the number of published studies in children is weak, particularly for lymphomas. The aim of the present study was to investigate the association between childhood AL, HL and NHL and a family history of cancer in the first- and second-degree relatives of the index children.
The role of a family history of cancer in the etiology of childhood hematopoietic malignancies was investigated using the data from the ESCALE study. ESCALE, a population-based case–control study, was carried out in France over the period, 2003–2004. A total of 773 cases of acute leukemia (AL), 130 of Hodgkin's lymphoma (HL), 163 of non-Hodgkin's lymphoma (NHL) and 1,681 population-based controls were included. The controls were randomly selected from the French population and were frequency matched with the cases on age and gender. Cancer history in first- and second-degree relatives was reported by the mothers in a structured telephone questionnaire that was the same for the cases and controls. Odds ratios (ORs) were estimated using an unconditional regression model taking into account the stratification variables and potential confounders. A family history of cancer was associated with an increased risk of HL (OR = 1.5 [1.0–2.2]) and NHL (OR = 1.8 [1.3–2.5]), but not AL (OR = 1.0 [0.9–1.2]). The ORs were higher when at least 2 relatives had a history of cancer or when 1 case occurred before age 46 years. Only HL was significantly associated with a family history of hematopoietic malignancies (OR = 2.0 [1.0–3.8]), mainly because of a significant association with a history of HL (OR = 5.4 [1.3–22]). In conclusion, the study findings support the hypothesis of familial susceptibility to childhood lymphoma, but do not suggest familial susceptibility to childhood AL. © 2007 Wiley-Liss, Inc.
Patients and methods
The ESCALE study was a French national population-based case–control study conducted in 2003 and 2004 to investigate the role of infectious, environmental and genetic factors in 4 childhood neoplastic diseases. The present paper focuses on AL, HL and NHL.
Case and control ascertainment
The cases were identified directly by the investigators assigned to each French pediatric oncology hospital department, with the support of the French National Registry of Childhood Blood Malignancies.19 For the cases to be eligible, leukemia or lymphoma was to have been newly diagnosed between January 1, 2003, and December 31, 2004. The cases were also required to be aged less than 15 years and resident in France at the time of diagnosis. Cases who had been adopted, or whose biological mother had died, or whose mother did not speak French or whose mother presented with a psychiatric disorder were not eligible. For ethical reasons, the mothers of children who had died or who were receiving hospital palliative care were not eligible. Out of the 1,321 cases (945 AL, 169 HL, 207 NHL) of childhood hematopoietic malignancies identified during the study period, 1,188 (851 AL, 151 HL, 186 NHL) cases were eligible. The reasons for noneligibility consisted in the child's death (34 AL, 3 HL, 7 NHL), hospital palliative care (7 AL), biological mother's death (10 AL, 3 NHL), non-French-speaking mother (29 AL, 11 HL, 8 NHL) or mother with serious psychiatric disorders (14 AL, 4 HL, 3 NHL).
The participation rates were 91, 86 and 88% for AL, HL and NHL, respectively.
The controls were randomly selected from the French population using a quota sampling method. A first sample of 60,000 addresses representative of the French population in terms of region and degree of urbanization was constituted from the French national telephone directory (plus randomly generated unlisted numbers). The quotas were designed to make the controls similar to all the cases of all types of cancer in term of age and gender, using the French National Registry of Childhood Blood Malignancies19 and the Regional Childhood cancer Registries20 as reference. Additional quotas were used to ensure that the control group was also representative of the French general population in terms of the number of children aged under 15 years living in the household, conditionally on age, based on the 1999 population census. Like the cases, the controls were not to have been adopted and were to have a biological mother who could be interviewed (alive, not presenting with a serious psychiatric disorder and French- speaking).
Out of the 50,217 phone numbers dialed, 22,584 did not connect to a home number, 24,410 were ineligible and for 862, the respondent hung up before eligibility could be checked. For the 2,361 remaining numbers, there were 679 refusals to participate. Thus, 1,682 mothers were interviewed (71.2%). One control with a history of neuroblastoma was excluded and 1,681 children were included as controls.
The telephone interviews with the cases' and controls' biological mothers were carried out by the same trained interviewers using structured questionnaires. The cases' mothers were interviewed at least 2 months after the diagnosis. The telephone questionnaire elicited information on demographic and socioeconomic characteristics, country of birth of each grandparent, parental occupational history, childhood environment and personal and familial medical history. The first- (parents and siblings) and second-degree (grandparents, uncles, aunts, half-brothers and half-sisters, but not the nephews and nieces of index children) genealogy was elicited and, for each relative, the mother was asked whether he or she had ever had a diagnosis of leukemia, Hodgkin's disease, lymphoma, myeloma or a solid tumor in any of the following disease sites: oral cavity or ENT (ear, nose or throat); lung; esophagus or stomach; liver; colon, rectum or anus; breast; thyroid; skin; melanoma; bone; kidney; bladder; uterus or ovary; prostate or brain. An open question targeted other or unclearly specified disease sites. The date of, or age at, diagnosis was also elicited.
Separate analyses by childhood hematopoietic malignancy and leukemia subtype (i.e., acute lymphoblastic leukemia [ALL] and acute myeloblastic leukemia [AML]) were carried out. Unconditional logistic regression was used to estimate the odds ratios (OR) and 95% confidence intervals (CI), with adjustment for the stratification variables, age and gender. The analyses were restricted to children aged 5 years or more for HLs and 2 years or more for NHLs because of the small numbers of cases in the lowest age groups. Family history was analyzed for all family members and separately for first- and second-degree relatives. The malignancies in relatives were considered by disease site, by group of disease sites and as a whole. Further analyses investigated the role of the relative's relationship to the case–control status (paternal or maternal relative), the age at onset of the earliest cancer in the family and the total number of relatives with a cancer in the family. Investigation for and taking account of potential confounding by the age and number of relatives were conducted. The relative's age was his/her age at the time of interview or death. The analyses were repeated after excluding the children with Down's syndrome. A sensitivity analysis was also conducted to evaluate the extent of a potential survival bias by allocating a positive family history of cancer to cases excluded because of death or hospital palliative care. Additional analyses by hematopoietic malignancy (AL, HL, NHL) and leukemia subtype (ALL, AML) were carried out using the polytomous logistic regression model. Additional analyses were also carried out by HL and NHL subtypes. The SAS® software package (version 9, Cary, NC) was used for all the analyses. All p values were two-tailed.
The cases consisted in 1,066 children with an incident hematopoietic malignancy: 773 cases of AL (654 ALL, 101 AML, and 18 unspecified or biphenotypic AL), 130 of HL, of whom 128 aged more than 4 years (19 mixed cell, 79 nodular sclerosis, 17 lymphocyte predominant, 1 lymphocyte depleted and 12 not other specified HL) and 163 of NHL, of whom 162 aged more than 1 year (73 Burkitt, 20 anaplastic large cell, 37 B-cell lymphoblastic and 27 T-cell lymphoblastic and 5 unspecified NHL).
Case and control comparability
The distribution of the cases and controls by the quota variable combining age and gender is shown in Table I. The cases and controls were similar with respect to that variable for the study as a whole, but not for each hematopoietic malignancy. In particular, the lymphoma cases were significantly older than the controls. All the strata contained more than 1 control per case for adjustment, with more controls per lymphoma case in the youngest strata. The controls' parents were slightly more educated and had a higher professional status than the parents of the leukemia and HL cases (Table II). There was no significant difference between the cases and controls in terms of the mothers' or grandparents' ages and the number of family members, after controlling for child age and gender. Consanguinity within first- and second-degree relatives was observed in none of the cases or controls.
|Age (year)||Controls (n = 1681) (%)||Acute leukemia||Hodgkin's lymphoma (n = 130)||Non-Hodgkin's lymphoma (n = 163)|
|ALL (n = 654)||AML (n = 101)||All (n = 773)|
|0–1||201 (12)||35||17||54 (7)||–||1 (1)|
|2||79 (5)||50||5||55 (7)||–||2 (1)|
|3||87 (5)||57||2||60 (8)||1 (1)||9 (6)|
|4||89 (5)||42||4||48 (6)||–||11 (7)|
|5–6||126 (7)||51||6||61 (8)||7 (5)||24 (15)|
|7–8||96 (6)||31||5||37 (5)||11 (8)||18 (11)|
|9–11||137 (8)||49||9||59 (8)||22 (17)||23 (14)|
|12–14||117 (7)||35||6||45 (6)||21 (16)||26 (16)|
|0–1||168 (10)||44||19||63 (8)||–||–|
|2||74 (4)||45||3||48 (6)||–||2 (1)|
|3||79 (5)||47||2||50 (6)||–||5 (3)|
|4||56 (3)||35||3||38 (5)||1 (1)||5 (3)|
|5–6||102 (6)||57||5||62 (8)||1 (1)||7 (4)|
|7–8||67 (4)||33||5||39 (5)||2 (2)||9 (6)|
|9–11||88 (5)||27||7||35 (5)||10 (8)||8 (5)|
|12–14||115 (7)||16||3||19 (2)||54 (42)||13 (8)|
|Acute leukemia||Hodgkin's lymphoma||Non-hodgkin's lymphoma|
|Cases (n = 773)||Controls (n = 1681)||OR||95% CI||Cases (n = 128)||Controls (n = 848)||OR||95% CI||Cases (n = 162)||Controls (n = 1312)||OR||95% CI|
|Maternal educational level|
|Paternal educational level||**||**|
|Intellectual and Scientific jobs, Managers and inter-Mediate professions||282||713||1.0||–||38||360||1.0||–||66||556||1.0||–|
|Administrative and Sales workers||226||477||1.2||[1.0–1.5]||33||221||1.5||[0.9–2.5]||41||356||1.0||[0.7–1.5]|
|Factory and agricltural workers, unemployed||168||274||1.5||[1.2–1.9]||28||134||2.1||[1.2–3.7]||25||220||1.0||[0.6–1.7]|
|Maternal age at interview (year)|
|Mean (sd)||36 (6)||36 (6)||41 (6)||39 (5)||39 (6)||37 (5)|
|Grand-parents' mean age (year)|
|Mean (sd)||63 (7)||63 (7)||68 (7)||66 (7)||66 (7)||65 (7)|
|Number of family members|
|16 and more||148||279||1.1||[0.8–1.4]||35||161||1.9||[1.0–3.7]||31||228||1.2||[0.7–2.1]|
|Mean (sd)||12 (4)||12 (4)||14 (4)||13 (4)||13 (3)||12 (4)|
Family history of cancer
The ORs associated with a positive family history of cancer in first-degree relatives, second-degree relatives and first- and second-degree relatives are shown in Table III. A family history of cancer was positively associated with childhood HL and NHL (HL:OR = 1.5 [1.0–2.2]; NHL:OR = 1.8 [1.3–2.5]). No association with AL was observed (AL:OR = 1.0 [0.9–1.2]). A positive family history of cancer was reported for 42% of the AL, 61% of the HL and 61% of the NHL cases and 42% of the controls. For each childhood hematopoietic malignancy, the ORs were higher when at least 2 relatives had had a cancer or when at least 1 family member had experienced a malignancy before age 46 years. The ORs associated with a history of malignancy in a paternal relative were higher than those for a history of malignancy in a maternal relative for AL, but not for either type of lymphoma.
|ALL||AML||All leukemia||Hodgkin's lymphoma||Non-Hodgking's lymphoma|
|Cases (n = 654)||OR||95% CI||Cases (n = 101)||OR||95% CI||Cases (n = 773)||OR||95% CI||Controls (n = 1681)||Cases (n = 128)||Controls (n = 848)||OR||95% CI||Cases (n = 162)||Controls n = 1312||OR||95% CI|
|Family history of cancer|
|First-and second-degree relatives|
|Number of relatives with cancer|
|2 and more||88||1.5||[1.1–2.0]**||12||1.1||[0.6–2.2]||103||1.4||[1.1–1.9]**||165||31||100||2.5||[1.4–4.2]***||29||135||2.2||[1.4–3.6]**|
|Earliest cancer onset age in family|
|No cancer in family||374||1.0||–||62||1.0||–||444||1.0||–||968||50||438||1.0||–||62||726||1.0||–|
|Side of family with cancer|
|No cancer in family||374||1.0||–||62||1.0||–||444||1.0||–||968||50||438||1.0||–||62||726||1.0||–|
|Only parternal relatives||131||1.2||[0.9–1.5]||12||0.7||[0.4–1.3]||150||1.2||[0.9–1.5]||290||29||166||1.3||[0.8–2.2]||38||238||1.7||[1.1–2.6]*|
|Only maternal relatives||87||0.7||[0.6–1.0]*||20||1.1||[0.6–1.8]||108||0.8||[0.6–1.0]||305||27||171||1.2||[0.7–2.0]||43||251||1.8||[1.2–2.8]**|
Table IV shows the results by disease site and group of sites. A family history of hematopoietic malignancy was only associated with HL (AL:OR = 1.0 [0.6–1.5]; HL:OR = 2.0 [1.0–3,8]; NHL:OR = 1.0 [0.5–2.1]). HL was statistically significantly associated with a family history of lymphoma (OR = 3.2 [1.2–8.0]) and, more particularly, with a history of HL (OR = 5.4 [1.3–22]). There were significant associations between childhood AL and a family history of uterine or ovarian cancer (OR = 1.6 [1.1–2.4]). Family history of cancer of the colon, rectum or anus was statistically significantly associated with both childhood Hodgkin's (OR = 2.4 [1.0–5.4]) and non-Hodgkin's (OR = 2.4 [1.2–4.7]) lymphomas. NHL was also statistically significantly associated with a family history of testicular (OR = 6.3 [1.4–29]), uterine/ovarian (OR = 2.6 [1.4–4.8]) and pancreas cancer (OR = 3.4 [1.0–11]).
|Cancer in relatives||ALL||AML||All leukemia||Hodgkin's lymphoma||Non-Hodgkin's lymphoma|
|Cases (n = 654)||OR||95% CI||Cases (n = 101)||OR||95% CI||Cases (n = 773)||OR||95% CI||Controls (n = 1.681)||Cases (n = 128)||Controls (n = 848)||OR||95% CI||Cases(n = 162)||Controls (n = 1312)||OR||95% CI|
|Oral cavity, ENT||36||1.4||[0.9–2.2]||5||1.2||[0.5–3.2]||42||1.4||[1.0–2.0]||67||8||34||1.6||[0.7–3.7]||11||52||1.8||[0.9–3.6]|
|Colon, rectum, anus||23||1.1||[0.7–1.8]||3||0.9||[0.3–2.8]||26||1.0||[0.6–1.7]||56||9||30||2.4||[1.0–5.4]*||12||44||2.4||[1.0–3.8]*|
|Oral cavity, ENT||36||1.4||[0.9–2.2]||5||1.2||[0.5–3.2]||42||1.4||[1.0–2.0]||67||8||34||1.6||[0.7–3.7]||11||52||1.8||[0.9–3.6]|
|Colon, rectum, anus||23||1.1||[0.7–1.8]||3||0.9||[0.3–2.8]||26||1.0||[0.6–1.7]||56||9||30||2.4||[1.0–5.4]*||12||44||2.4||[1.2–4.7]*|
|Other or NOS||9||1.6||[0.7–3.7]||1||1.2||[0.2–9.7]||10||1.5||[0.7–3.5]||15||2||10||1.6||[0.3–8.0]||–||13||–||–|
|Other or NOS||8||0.9||[0.4–2.2]||1||1||[0.1–7.3]||9||0.9||[0.4–2.1]||19||1||13||0.6||[0.1–4.6]||1||18||0.4||[0.1–2.9]|
|Other or NOS||22||0.9||[0.5–1.5]||1||0.2||[0.0–1.8]||23||0.8||[0.5–1.5]||65||10||35||2.2||[1.0–4.8]*||9||52||1.3||[0.6–2.7]|
The analyses by lymphoma subtypes were based on small numbers. A family history of cancer was significantly associated with anaplastic large cell lymphoma (OR = 4.4 [1.4–13.3]) and T-cell lymphoblastic lymphoma (OR = 3.0 [1.3–6.9]), while ORs of 1.5 (95% CI: [0.9–2.4]) and 1.4 (95% CI: [0.7–2.7]) were observed for Burkitt's lymphoma and B-cell lymphoblastic lymphoma, respectively. With respect to HL, ORs were of 2.1 (95% CI: [0.8–5.5]), 1.2 (95% CI: [0.7–1.9]) and 1.8 (95% CI: [0.7–5.0]) for mixed cellularity, nodular sclerosis and lymphocyte predominant subtypes, respectively. The lymphocyte predominant subtype was significantly related to family history of Hematopoietic malignancy (OR = 5.1 [1.6–16.5]).
The results were unchanged after exclusion of the children with Down's syndrome. The use of polytomous logistic regression models for ALL and AML, for AL and NHL (restricted to age greater than or equal to 2 years) and for AL, HL and NHL (restricted to age greater than or equal to 5 years) did not alter the results. With regard to potential confounders, no variable was statistically significantly associated with both a childhood hematopoietic malignancy and a family history of cancer, after controlling for child age and gender. Neither parental professional category nor educational level was associated with a family history of cancer in the control group. The results were also unchanged when the analyses were adjusted for familial structure (mother's and father's ages, mean ages of grandparents and of uncles and aunts, number of relatives), socioeconomic status (professional category, educational level, degree of urbanization of the place of residence), breast-feeding and early infections, which have been previously shown to be related to leukemia in the literature and were also related in this study.21 The results were unchanged after exclusion of the children who had at least 1 grandparent born in Africa or Asia.
A positive association with a family history of cancer was observed for NHL and HL, but not for AL. Only HL was associated with a positive family history of hematopoietic malignancy, mainly due to the high OR for family history of HL. Significant positive associations were also observed with a family history of solid tumor in the following disease sites: colon/rectum/anus with Hodgkin's and NHL; uterine/ovarian with NHL and AL and testicular and pancreas with NHL. The numbers were too small to allow the analysis of leukemia or lymphoma aggregation separately for the first and second degrees.
Genetic susceptibility could be reflected by an increased familial incidence and earlier onset of malignancies. Indeed, for lymphoma, the ORs were higher when at least 2 relatives had presented with cancer or when the age at diagnosis of a single case was less than 46 years. This tendency was also observed, at a lesser extent, for leukemia; a weak genetic susceptibility to leukemia cannot be then excluded with regard to our results.
With an α error of 5%, the statistical power of the study to evidence an OR of 1.5, which was the order of magnitude of the relationship between childhood AL and family history of cancer in 3 previous case–control studies,15–17 was greater than 99%.
The relatives were quite young: the mean age of the parents at the time of the interview was about 40 years; that of the grandparents about 65 years. A lack of statistical power for specific types of cancer with a high age of onset is therefore probable. The positive results for cancers with lower ages of onset were less censored.
The cases were identified through the data collection system of the French National Registry of Childhood Blood Malignancies, making case selection at the identification stage unlikely. Cases who had died or were receiving palliative care (41 AL, 3 HL and 7 NHL) were not eligible. However, the absence of a relationship between leukemia and family history of cancer is unlikely to result from a survival bias. The severity of leukemia is not known to be related to family history, and, in the present study, the leukemia cases who died during the year following the telephone interview had an even less frequently positive family history of cancer than the survivors (35% vs. 42%). All of the 41 ineligible leukemia cases who had died or were receiving palliative care would have had to have a positive family history of cancer in order for cases selection to have masked a slight association, with a corresponding OR of 1.2 (95% CI: [1.0–1.4]), instead of the observed one of 1.0 (95% CI: [0.9–1.2]). None of the lymphoma cases eligible during the study period had a known ataxia telangiectasia or any other genetic syndrome reported in the French National Registry of Childhood Blood Malignancies. Therefore, a selection by survival of cases related to these familial rare diseases is unlikely.
The controls were randomly selected from the overall population. The national telephone directory was used as the basis for random selection. Selection of controls with listed numbers only was avoided by randomly generating unlisted numbers. There were no differences between the cases and controls with regard to gender or age (considering all types of cancer), or between the controls and overall population with regard to birth order. The quota recruitment process was thus successful. The refusals to take part could have been related to the parental socioeconomic status or educational level that appeared higher among controls than among HL and, to a lesser extent, leukemia cases. The control mothers educational levels were very similar to those of the French population, but the control fathers' educational levels were higher.22–24 However, parental socioeconomic status and educational level were not associated with a family history of cancer in the control group and the results were unchanged after adjustment for these variables. We collected no direct information on ethnicity, which might be related to the risk of childhood cancer and to history of cancer in relatives. However, excluding the children whose any grandparent was born in Africa or Asia had no influence on the results.
Nondifferential misclassification bias was probable since the cancer history data only consisted in the mothers' reports. However, bias was limited by the use of specific closed questions and by the fact that the mothers were asked about her and the father's close family, i.e. children, siblings and parents. Adults with cancer have been shown to report their family history of cancer in first degree relatives quite accurately.25–33 Bondy et al. studied the accuracy of data on family history of cancer obtained by interviewing the mothers of children and adolescents with sarcoma34 and reported positive predictive values (PPV) of 88 and 71% for first- and second-degree relatives of children, respectively. For Perillat et al.,17 the PPV of mothers' reports on first-degree relatives of children with leukemia was 100%. The sensitivity and specificity of those 2 studies could not be determined since the negative reports were not analyzed.
Differential recall between cases and controls cannot be excluded. Studies on adults have reported no or few differential errors between cancer cases and their controls with respect to cancer in relatives.27–31 To our knowledge, no study of accuracy with respect to children has been conducted.
Misclassification may be more marked for paternal than maternal family history of cancer, given that mothers were interviewed. However, the ORs associated with the paternal and maternal histories were very similar for HL and NHL. This may suggest that the recall bias was weak for lymphomas. Differences were observed only for leukemia, but there would appear to be no obvious reason for a recall bias for leukemia and not for lymphoma.
Five previous case–control studies have estimated the association between childhood leukemia and family history of cancer.14–18 One study only addressed ALL18 and a second targeted children aged less than 19 months.14 A positive association was found in all the studies. Three reported significant cancer ORs of 1.5 (95% CI: [1.0–2.1]),15 1.5 (95% CI: [1.0–2.2])17 and 1.6 (95% CI: [1.2–2.1])16; 2 reported significant ORs for a family history of hematopoietic malignancy of 2.7 (95% CI: [1.1–6.9])17 and 2.1 (95% CI: [1.2–3.6]).18 Six population-based studies on childhood leukaemia linked Cancer Registry data with a family database.8–13 One of these studies only addressed childhood ALL9 and another, AML.13 With the exception of the study by Hasle and Olsen,13 all the registry studies considered first-degree relatives only. None of these studies detected a positive association between childhood leukemia and family history of cancer, with observed standardized incidence rate ratio from 0.8 to 1.0. Three of the registry studies8, 12, 13 did not consider cancer that occurred in relatives before the index child's birth in order to avoid survival bias. Consequently they may have failed to detect early-onset malignancies and may have underestimated associations, if an earlier onset age was particularly associated with familial cancers.
Fewer studies have addressed the association between family history of cancer and childhood lymphoma.8, 10–12, 15, 35 One of the previous case–control studies reported quite similar ORs for childhood leukemia, HL and NHL (LA:OR = 1.5 [1.0–2.1]; LH:OR = 1.4 [0.7–2.9]; LNH:OR = 1.8 [1.0–3.3]).15 Four of the previous registry studies8, 10–12 also addressed childhood lymphoma, without distinguishing between Hodgkin's and NHL. The association with family history of cancer was always stronger for lymphoma than for leukemia, with observed standardized incidence rate ratio ranging from 1.0 to 1.8. Le Bihan et al.35 did not observe an increased frequency of cancer in the relatives of children with NHL in a familial study.
With regard to associations with specific types of malignancy in relatives, the literature is relatively heterogeneous and chance may explain some of the associations. Significant positive associations between childhood AL and a family history of cancer in the following disease sites was observed: oral cavity16; esophagus or stomach,17 colon,16, 17 uterus or ovary16; breast18; brain16; melanoma17; testis9; and kidney.12 Childhood lymphoma, taken as a whole, was shown to be significantly associated with a family history of any lymphoma10, 11 and brain tumor.12
In conclusion, our results support the hypothesis of a familial susceptibility to childhood lymphoma but not to childhood leukemia. Besides genetic factors, they may also reflect the existence of environmental or infectious risk factors shared by the family members.
We are grateful to Mrs. Marie-Hélène Da Silva and Dr. Christophe Steffen (INSERM U754), who coordinated the recruitment of the cases; Mrs. Aurélie Goubin and the staff of the French National Registry of Childhood Blood Malignancies, who contributed to case detection and verification; Mrs. Sabine Mélèze and Mrs. Marie-Anne Noel (Institut CSA), who coordinated the selection of the controls and the interviews; and Mrs. Catherine Tricoche (Callson) and the team of interviewers, who interviewed the cases and the controls.
|Principal investigator||Degree||Hospital||City (France)|
|André Baruchel||MD||Saint-Louis / Robert Debré||Paris|
|Claire Berger||MD||Centre Hospitalier Universitaire||Saint-Etienne|
|Christophe Bergeron||MD||Centre Léon Bérard||Lyon|
|Jean-Louis Bernard||MD||La Timone||Marseille|
|Pierre Bordigoni||MD||Centre Hospitalier Universitaire||Nancy|
|Patrick Boutard||MD||Centre Hospitalier Régional Universitaire||Caen|
|Gérard Couillault||MD||Hôpital d'Enfants||Dijon|
|Lionel De Lumley||MD||Centre Hospitalier Régional Universitaire||Limoges|
|Anne-Sophie Defachelles||MD||Centre Oscar Lambret||Lille|
|Alain Fischer||MD||Hôpital des Enfants Malades||Paris|
|Virginie Gandemer||MD||Centre Hospitalier Universitaire – Hôpital Sud||Rennes|
|Olivier Hartmann||MD||Institut Gustave Roussy||Villejuif|
|Jean-Pierre Lamagnere||MD||Centre Gatien de Clocheville||Tours|
|Françoise Lapierre||MD||Centre Hospitalier Universitaire Jean Bernard||Poitiers|
|Patrick Lutz||MD||Hôpital de Hautepierre||Strasbourg|
|Geneviève Margueritte||MD||Arnaud de Villeneuve||Montpellier|
|Françoise Mechinaud||MD||Hôpital Mère et Enfants||Nantes|
|Gérard Michel||MD||La Timone||Marseille|
|Frédéric Millot||MD||Centre Hospitalier Universitaire Jean Bernard||Poitiers|
|Martine Münzer||MD||American Memorial Hospital||Reims|
|Brigitte Nelken||MD||Jeanne de Flandre||Lille|
|Hélène Pacquement||MD||Institut Curie||Paris|
|Brigitte Pautard||MD||Centre Hospitalier Universitaire||Amiens|
|Yves Perel||MD||Pellegrin Tripode||Bordeaux|
|Alain Pierre-Kahn||MD||Enfants Malades||Paris|
|Emmanuel Plouvier||MD||Centre Hospitalier Régional||Besançon|
|Xavier Rialland||MD||Centre Hospitalier Universitaire||Angers|
|Alain Robert||MD||Hôpital des Enfants||Toulouse|
|Hervé Rubie||MD||Hôpital des Enfants||Toulouse|
|Christine Soler||MD||Fondation Lenval||Nice|
|Danièle Sommelet||MD||Centre Hospitalier Universitaire||Nancy|
|Jean-Pierre Vannier||MD||Charles Nicolle||Rouen|