Cancer Causes Control (2015) 26:1003–1017 DOI 10.1007/s10552-015-0593-5
ORIGINAL PAPER
Parental smoking, maternal alcohol, coffee and tea consumption during pregnancy, and childhood acute leukemia: the ESTELLE study L. Orsi1,2 • J. Rudant1,2,3 • R. Ajrouche1,2 • G. Leverger4,5 • A. Baruchel6,7 B. Nelken8 • M. Pasquet9 • G. Michel10 • Y. Bertrand11 • S. Ducassou12 • V. Gandemer13 • P. Lutz14 • L. Saumet15 • P. Moreau16 • D. Hemon1,2 • J. Clavel1,2,3
•
Received: 16 January 2015 / Accepted: 22 April 2015 / Published online: 9 May 2015 Ó Springer International Publishing Switzerland 2015
Abstract Purpose To investigate the role of parental smoking during pre-conception and pregnancy, maternal beverage consumption (alcohol, coffee and tea) during pregnancy and their possible interactions, in the etiology of childhood acute leukemia (CL). Methods The ESTELLE study included 747 cases of CL [636 cases of acute lymphoblastic leukemia (ALL) and 100 cases of acute myeloblastic leukemia (AML)] diagnosed in France in 2010–2011 and 1,421 population controls frequency-matched with the cases on age and gender. Data were obtained from structured telephone questionnaires administered to the mothers. The odds ratios (OR) and their 95 % confidence intervals were estimated using
Electronic supplementary material The online version of this article (doi:10.1007/s10552-015-0593-5) contains supplementary material, which is available to authorized users. & L. Orsi
[email protected] 1
2
3
INSERM U1153 Epidemiology and Biostatistics Sorbonne Paris Cite´ Center (CRESS), Epidemiology of Childhood and Adolescent Cancers Team (EPICEA), Villejuif, France
unconditional logistic regression models adjusted for potential confounders. Results AML, but not ALL, was non-significantly associated with alcohol drinking during pregnancy [OR = 1.3 (0.8–2.0)] with a significant positive dose–response trend (p-trend = 0.02). Pre-conception paternal smoking was significantly associated with ALL [OR = 1.2 (1.1–1.5)] and AML [OR = 1.5 (1.0–2.3)]. CL was not associated with maternal smoking [OR = 1.0 (0.8–1.2)], or maternal coffee [OR = 0.9 (0.8–1.1)] or tea drinking [OR = 0.9 (0.8–1.1)] during pregnancy. However, a high consumption of coffee ([2 cups/day) was significantly associated with ALL [OR = 1.3 (1.0–1.8)]. Conclusions The findings constitute additional evidence that maternal alcohol drinking during pregnancy may be involved in AML, and that paternal smoking before pregnancy may be a risk factor for CL. The role of maternal coffee drinking in CL remains unclear and should be investigated 9
Hoˆpital des Enfants, Toulouse, France
10
AP-HM, Hoˆpital la Timone, Marseille, France
11
Institut d’He´matologie et d’Oncologie Pe´diatrique, Lyon, France
12
Paris-Descartes University, UMRS-1153, Epidemiology and Biostatistics Sorbonne Paris Cite´ Center (CRESS), Paris, France
Hoˆpital Pellegrin Tripode, Bordeaux, France
13
CHU Hoˆpital Sud, Rennes, France
14
Hoˆpital de Hautepierre, Strasbourg, France
RNHE - National Registry of Childhood Hematopoietic Malignancies, Villejuif, France
15
Hoˆpital Arnaud de Villeneuve, Montpellier, France
16
Hoˆpital Me`re-Enfant, CHU-Nantes, Nantes, France
4
AP-HP, Hoˆpital Armand Trousseau, Paris, France
5
Universite´ Paris 6 Pierre et Marie Curie, Paris, France
6
AP-HP, Hoˆpital Robert Debre´, Paris, France
7
Universite´ Paris 7, Paris, France
8
CHRU, Hoˆpital Jeanne de Flandre, Lille, France
123
1004
further in consortium analyses and in large birth cohort studies with exposure assessment more contemporaneous with the exposure, before the occurrence of the disease. Keywords Childhood acute leukemia Parental smoking Coffee Alcohol Abbreviations ALL Acute lymphoblastic leukemia AML Acute myeloblastic leukemia BCP-ALL B cell precursor acute lymphoblastic leukemia CL Childhood acute leukemia NRCH National registry of childhood hematopoietic malignancies
Introduction Among childhood cancers, childhood acute leukemia (CL) is the most frequent cancer with about 450 newly diagnosed cases each year in France [1]. Maternal tobacco smoking and beverage consumption during pregnancy are relatively prevalent factors which may expose the fetus, after placental barrier crossing, to known human carcinogens such as benzene, polycyclic aromatic hydrocarbons and aromatic amines for tobacco [2] and ethanol for alcohol [3]. Although being classified in the group 3 by the IARC [4], caffeine may act as an inhibitor of DNA topoisomerase II and induce abnormalities of chromosome 11q23, similar to those induced by epipodophyllotoxins [5]. Moreover, smoking has been shown to affect spermatogenesis by altering sperm morphology, motility and concentration and to increase damage to sperm DNA [6]. These findings constituted the rationale for investigating those parental exposures as potential risk factors for CL. Overall, most of the case–control studies conducted in the last 30 years did not show any significant association between maternal smoking during the pre-conception period or during pregnancy and acute lymphoblastic leukemia (ALL) or acute myeloblastic leukemia (AML) [7–24]. However paternal smoking during pre-conception has been suggested as a risk factor for CL and the association needs to be confirmed. Eleven studies out of sixteen [10–23, 25, 26] reported a positive association between ALL and preconception paternal smoking, and six of the associations were significant [10–12, 14, 20, 21]. For AML, pre-conception paternal smoking was positively associated in five [12, 17, 20, 22, 24] out of eleven [10–13, 15, 17–20, 22, 24] studies.
123
Cancer Causes Control (2015) 26:1003–1017
The potential association between maternal alcohol drinking during pregnancy and ALL remains unclear: significant positive associations were evidenced in five studies [11, 16, 18–20], while three studies reported significant inverse associations [27–29]. For AML, the literature suggests a positive association (cf. [30] for a review): all the published studies [11, 16, 19, 20, 27, 31, 32] but one [18] reported a positive association with OR ranging from 1.3 to 2.6. Fewer studies have addressed the potential association between CL and maternal caffeinated beverage consumption during pregnancy. For ALL, the results for maternal coffee consumption during pregnancy are heterogeneous: two studies [16, 33] evidenced significant positive associations, while the other studies did not report any association [5, 18, 34]. With regard to AML, all the published case–control studies [5, 16, 18, 33] reported OR greater than 1. Interestingly, three studies, including two led by the authors’ team, reported more marked associations between maternal coffee drinking during pregnancy and ALL or AML in children born to non-smoking mothers [18, 33, 34]. There are relatively few published studies that have addressed maternal tea consumption during pregnancy [5, 16, 33, 34]. The studies suggest that there is no association with CL. The present study, using data generated by the French large-scale population-based case–control study ESTELLE [35], investigated the roles of parental smoking, maternal beverage (alcohol, coffee and tea) consumption during pregnancy, which are common habits, and explored their potential joint effect, in the etiology of CL.
Materials and methods The ESTELLE study is a nation-wide population-based case–control study that was conducted in 2010 and 2011 to investigate the role of infectious, environmental, and genetic factors in five childhood neoplastic diseases (acute leukemia, lymphoma, neuroblastoma, hepatoblastoma and brain tumor). The present paper focuses on acute leukemia. Case and control selection The cases were directly identified by the investigators of the National Registry of Childhood Hematopoietic Malignancies (NRCH), in all the pediatric oncology units in France, in 2010 and 2011. The eligible CL cases were children diagnosed with CL in compliance with the NRCH registration criteria. The cases were required to be \15 years old and resident in mainland France at the time of diagnosis. Children who had been adopted (n = 5), or whose biological mother had died (n = 4), was absent
Cancer Causes Control (2015) 26:1003–1017
(n = 1), did not speak French (n = 31), had a serious social problem (n = 7) or psychiatric disorder (n = 5), or could not be interviewed for ethical reasons because the child was in palliative care or had died (n = 8) were not eligible. Information on the leukemia subtype was obtained subsequently from the NRCH. The population controls were children free from cancer selected in France in 2010 and 2011, using a quota sampling method. Forty successive sets of 5,000–25,000 allocable telephone numbers were randomly generated and dialled over the 2 years of the subject recruitment period. Of the 312,022 numbers which were necessary to complete control sampling, 34,983 resulted in a contact with a household (Supplemental Figure 1). Quotas were used to obtain, overall, at least one control per case for each year of age*sex*type of cancer strata, based on the expected numbers derived from the national registry of childhood cancers. Controls aged \1 year were overrepresented to gain power in that age category. The quotas also ensured that the control group had the same distribution as the general population of children aged \15 years living in the households, conditionally on age. In the same way as for the cases, the children who had been adopted, or whose biological mother had died or did not speak French were not eligible as controls. In all, 747 CL cases out of the 801 eligible cases (participation rate: 93 %) and 1,421 controls out of the 1,662 eligible controls (participation rate: 86 %) were included in the analysis. Data collection and exposure definitions The mothers of the cases and controls were interviewed by the same trained interviewers using structured questionnaires with computer-assisted telephone interviewing (CATI). The questionnaire elicited information on demographic and socioeconomic characteristics, childhood environment and lifestyle, familial and personal medical history, and maternal and paternal lifestyle characteristics. In addition, a subset of the children’s fathers [n = 247, 178 control fathers and 69 CL case fathers (53 ALL case fathers, 13 AML case fathers)] completed a standardized self-administered questionnaire which elicited information on the fathers’ own socioeconomic characteristics, familial medical history and lifestyle characteristics such as tobacco smoking and alcohol drinking. The present study focused on parental smoking and maternal alcohol, coffee and tea consumption. With regard to tobacco consumption, mothers were asked if they had ever smoked, even occasionally, during the 3-month period preceding the conception and also during the index pregnancy. If so, they were asked to state their average consumption of cigarettes for the pre-conception period and for each trimester of pregnancy. Mothers were also asked
1005
whether the father of the index child had ever smoked and, if so, his age at the start of smoking and the year of quitting, if applicable. Mothers were also asked to quantify the average daily consumption of cigarettes in the 3-month period preceding the conception of the index child and during the pregnancy. With regard to beverages, the mothers were asked if they had consumed alcohol (wine, beer, cider, spirits), coffee or tea (without any distinction between caffeinated/ decaffeinated coffee) during pregnancy, and during the first trimester of pregnancy. If so, they were asked to quantify their daily, weekly or monthly intakes. For tobacco exposures, dichotomous variables were generated for each parent and for the two periods of interest (pre-conception period and during pregnancy). The subjects having smoked at least one cigarette, even occasionally, were classified as ever a smoker. The number of cigarettes smoked was categorized using 4-class categorical variables, with cut-offs defined a priori: 0; \1–4 cigarettes/day; 5–10 cigarettes/day; 11 or more cigarettes/day for maternal smoking; 0; \1–9 cigarettes/day; 10–15 cigarettes/day; 16 or more cigarettes/day for paternal smoking. Analyses were repeated restricting cigarette smoking to consumption of at least one cigarette per day, and considering smokers of \1 cigarette/day in the reference category. Maternal alcohol consumption during pregnancy was defined as any alcohol consumption, including occasional consumption. The quantity of alcohol consumed was categorized using four classes with cut-offs defined a priori (0;\1 glass/week; 1–2 glasses/week; [2 glasses/week). Maternal consumption of coffee or tea during pregnancy was defined as maternal consumption of at least one cup of coffee or cup of tea per week. The daily consumption of coffee and tea was categorized using categorical variables with cut-offs defined from the controls and corresponding to the median and third quartile of the distribution of coffee consumption (0, B1 cup/day; [1–2 cups/day; [2 cups/day) and to the first quartile and the second tertile of tea consumption (0, \1 cup/day, 1 cup/day, [1 cup/day). Analyses were also reiterated after including occasional consumers of coffee or tea in the exposed category. Statistical analysis All the analyses were performed with SAS v9.3 (Cary, NC). The odds ratios (OR) and their 95 % confidence intervals (95 % CI) were estimated using unconditional logistic regression models for AL overall, for the subtypes of AL (ALL and AML) and for the cytogenetic groups of B-cell precursor ALL [BCP ALL], and using non-ordinal polytomous logistic regression for the subtypes of ALL (BCP ALL, mature B-cell or Burkitt ALL, T-cell ALL). All the models included the child’s age at diagnosis or
123
1006
interview (\2; 2; 3; 4; 5–6; 7–8; 9–12; C12), sex (boy vs girl), mother’s level of education (incomplete secondary; completed secondary; completed tertiary), child’s birth order (at least second-born vs first-born) as categorical variables and mother’s age at child’s birth as a continuous variable. Adjustments for other sociodemographic characteristics (parental socioeconomic level, rural/urban status at the place of residence), breastfeeding and European status were implemented in additional analyses. For each exposure of interest, quantitative variables were explored (number of cigarettes per day, number of glasses of alcohol per week, number of cups of coffee per day, number of cups of tea per day). The tests for trend were computed from the categorical variables. The subjects in each class of the categorical variables were assigned the median value of that class. First, the deviation from linearity was tested by a likelihood ratio test, comparing the model with the newly generated quantitative variable to the full model with the categorical variable. If linearity was not rejected, the p value of the trend was determined by testing the slope of the quantitative variable using a Wald-test. Stratified analyses were also conducted to investigate the association between CL and each of the exposures of interest by the strata of (a) the other exposures and (b) sociodemographic characteristics. Tests of interaction were performed by fitting the interaction-term in the logistic model.
Results Overall, 747 incident CL cases (636 ALL, 100 AML, and 11 unspecified or biphenotypic CL) and 1,421 controls were included in the analysis. The ALL included 503 B-cell precursor ALL (BCP-ALL), 95 T-cell ALL and 26 Burkitt’s ALL (Supplementary Table 1).
Case–control comparability The distribution of the cases and controls by age, sociodemographic characteristics and potential confounders is shown in Table 1. As expected, children \1 year old were overrepresented among the controls. Each age*sex stratum comprised at least one control per case (more than 2 in the first two age groups). The cases and controls were similarly distributed with respect to maternal educational level, parental professional category and European status. However, the cases lived in urban areas more often than the controls, particularly the AML cases, and the ALL case mothers were significantly younger than the control mothers. Birth order and breastfeeding were both significantly inversely associated with ALL.
123
Cancer Causes Control (2015) 26:1003–1017
Association between the exposures of interest among the controls Among the controls (supplementary Table 2), parental smoking at the birth of the index child was reported by the youngest mothers more often, and less often by those having the highest educational level or belonging to the highest socio-economic category. In contrast, maternal self-reporting of alcohol, coffee or tea consumption during pregnancy increased with maternal age. The mothers who were more educated reported alcohol or tea consumption more often than the less educated mothers. Self-reported maternal smoking was positively associated with maternal consumption of coffee and inversely associated with maternal consumption of tea. Lastly, self-reported alcohol consumption during pregnancy was significantly more frequent among the control mothers who also reported consuming coffee or tea during pregnancy.
Parental smoking Table 2 shows the associations between parental smoking and CL for ALL and AML. Cigarette smoking during pregnancy and during the 3-month period preceding pregnancy was reported by 20 and 37 % of control mothers, respectively. Among case mothers, the prevalences were 21 and 39 %, respectively. There was no association between maternal smoking and CL or the ALL or AML subtypes, irrespective of the period considered [OR = 1.0 (0.8–1.2) and 1.1 (0.9–1.3), respectively, Table 2]. No positive trend with the number of cigarettes smoked per day was evidenced. No association was evidenced when exposures were considered by trimester of pregnancy; the ORs were close to or slightly \1.0. The case mothers reported that the index child’s father had smoked significantly more often than the control mothers did, for both periods of interest [ORs = 1.3 (1.1–1.6), 1.3 (1.1–1.6)]. The associations were more marked for AML [OR = 1.5 (1.0–2.3) for pre-conception paternal smoking; OR = 1.7 (1.1–2.5) for paternal smoking during pregnancy] than for ALL [OR = 1.2 (1.0–1.5) for pre-conception paternal smoking, OR = 1.3 (1.0–1.5) for paternal smoking during pregnancy]. For ALL, the ORs were higher for smoking \10 cigarettes daily than for the highest consumption; no significant trend was evidenced. For AML, significant trends were evidenced for both periods (p trend = 0.03 and 0.02, respectively), with ORs of close to 2.0 for smoking more than 15 cigarettes daily. No joint effect of paternal and maternal smoking was detected. Similar results were observed when the analyses were performed after including occasional tobacco consumption of \1 cigarette per day in the reference category. The
122 287 229 222 91 274
1 2–3
4–5
6–9
10–11
12–14
0
No
Urban ([100,000 people.)
455 506 313
25–29
30–34
C35
714
Completed tertiary
942 295
High
Medium
Parental socioeconomic status
1
308
Completed secondary
Missing
398
Incomplete secondary
Maternal education
147
\25
Maternal age at index child’s birth (years)
7
525
Intermediate (5,000 –100,000 people)
Missing
577 312
Rural (\5,000 people)
Place of residence at diagnosis
1,421
Yes
Landline numbers
196
0
Age at diagnosis/interview (years)
21
66
50
22
28
22
36
32
10
37
22
41
0
100
19
6
16
16
9 20
14
157
479
1
367
156
223
134
247
255
111
3
306
159
279
164
583
87
68
171
133
75 179
34
21
64
49
21
30
18
33
34
15
41
21
38
22
78
12
9
23
18
10 24
5
%
1.0
1.0
ns
0.9
0.9
1.0
ns
0.5
0.6
0.7
1.0
**
1.3
1.1
1.0
ns
***
OR
n
n
%
All CL (n = 747)
Controls (n = 1,421)
(0.8–1.3)
Ref.
(0.7–1.1)
(0.7–1.2)
Ref.
(0.4–0.7)
(0.5–0.8)
(0.5–1.0)
Ref.
(1.0–1.6)
(0.9–1.4)
Ref.
95 % CI
Table 1 Distribution of age, socio-demographic characteristics and potential confounders by case/control status
19
140
402
1
309
138
188
105
214
217
100
2
252
133
249
134
502
62
57
147
124
63 164
n
22
63
49
22
30
17
34
34
16
40
21
39
21
79
10
9
23
19
10 26
3
%
ALL (n = 636)
1.1
1.0
ns
0.9
0.9
1.0
ns
0.4
0.6
0.7
1.0
***
1.2
1.0
1.0
ns
***
OR
(0.9–1.4)
Ref.
(0.7–1.1)
(0.7–1.2)
Ref.
(0.3–0.6)
(0.4–0.8)
(0.5–0.9)
Ref.
(1.0–1.5)
(0.8–1.4)
Ref.
95 % CI
15
70
54
17
29
28
29
34
9
1
47
24
28
26
74
22
9
20
9
12 15
13
n
15
70
54
17
29
28
29
34
9
48
24
28
26
74
22
9
20
9
12 15
13
%
0.7
1.0
ns
1.1
0.8
1.0
ns
1.6
1.0
1.3
1.0
ns
1.8
1.6
1.0
*
ns
OR
AML (n = 100)
(0.4–1.2)
Ref.
(0.7–1.7)
(0.4–1.4)
Ref.
(0.7–3.6)
(0.4–2.1)
(0.6–2.8)
Ref.
(1.1–3.0)
(0.9–2.9)
Ref.
95 % CI
Cancer Causes Control (2015) 26:1003–1017 1007
123
123 496 331
541 880
2
C3 Breastfeeding in the first year of life
No
Yes
62
38
23
35
42
12
88
13
418
329
133
253
361
7
110
630
111
56
44
18
34
48
15
85
15
0.8
1.0
*
0.6
0.8
1.0
**
1.3
1.0
ns
1.2
OR
(0.7–1.0)
Ref.
(0.5–0.8)
(0.7–1.0)
Ref.
(1.0–1.7)
Ref.
(0.9–1.5)
95 % CI 94
351
285
109
212
315
7
110
519
n
55
45
17
33
50
17
83
15
%
ALL (n = 636)
0.8
1.0
*
0.6
0.8
1.0
***
1.3
1.0
ns
1.2
OR
(0.7–1.0)
Ref.
(0.5–0.8)
(0.6–0.9)
Ref.
(1.0–1.7)
Ref
(0.9–1.6)
95 % CI
63
37
21
39
40
18
82
15
n
ns p C 0.05; * 0.01 B p \ 0.05; ** 0.001 B p \ 0.01; *** p \ 0.001
63
37
21
39
40
18
82
15
%
1.1
1.0
ns
1.0
1.2
1.0
ns
1.6
1.0
ns
1.1
OR
AML (n = 100)
OR and 95 % CI were estimated by unconditional logistic regression models adjusted for age (\2; 2; 3; 4; 5–6; 7–8; 9–12; C12) and sex (boy versus girl)
594
1
Birth order
11
175
No
Missing
1,235
184
%
n
n
%
All CL (n = 747)
Controls (n = 1,421)
Yes
Both mother and father born in Europe
Low
Table 1 continued
(0.7–1.7)
Ref.
(0.6–1.7)
(08–1.9)
Ref.
(0.9–2.7)
Ref
(0.6–1.9)
95 % CI
1008 Cancer Causes Control (2015) 26:1003–1017
Cancer Causes Control (2015) 26:1003–1017
1009
Table 2 Association between CL and self-reported parental smoking Controls (n = 1,421)
All CL (n = 747)
ALL (n = 636)
AML (n = 100)
n
n
OR
95 % CI
n
OR
95 % CI
n
OR
95 % CI
Pre-conception maternal smoking (cigarettes/day) Never
897
450
1.0
Ref.
381
1.0
Ref.
64
1.0
Ref.
Ever
519
294
1.1
(0.9–1.3)
252
1.1
(0.9–1.3)
36
1.0
(0.7–1.5)
\5
118
64
1.0
(0.7–1.4)
56
1.0
(0.7–1.5)
8
1.0
(0.5–2.1)
5–10
245
133
1.0
(0.8–1.3)
113
1.0
(0.8–1.4)
16
0.9
(0.5–1.7)
C11
156
97
1.1
(0.9–1.5)
83
1.1
(0.8–1.5)
12
1.1
(0.6–2.2)
p-trend = 0.39
p-trend = 0.47
p-trend = 0.83
Maternal smoking during pregnancy (cigarettes/day) Never
1,129
588
1.0
Ref.
499
1.0
Ref.
82
1.0
Ref.
Ever
290
158
1.0
(0.8–1.2)
136
1.0
(0.8–1.3)
18
0.9
(0.5–1.5)
\5
146
79
1.0
(0.7–1.3)
68
1.0
(0.7–1.3)
10
1.0
(0.5–1.9)
5–10
115
64
1.0
(0.7–1.4)
55
1.0
(0.7–1.5)
6
0.7
(0.3–1.7)
C11
29
15
1.0
(0.5–1.9)
13
1.0
(0.5–2.0)
2
1.0
(0.2–4.1)
p-trend = 0.98
p-trend = 0.95
p-trend = 0.66
By trimester of pregnancy No smoking during pregnancy
1,129
588
1.0
Ref.
499
1.0
Ref.
82
1.0
Ref.
Ever smoke during first trimester
279
154
1.0
(0.8–1.3)
132
1.0
(0.8–1.3)
18
0.9
(0.5–1.6)
Ever smoke during 2nd trimester
244
129
1.0
(0.8–1.2)
112
1.0
(0.8–1.3)
14
0.8
(0.4–1.5)
Ever smoke during 3rd trimester
239
130
1.0
(0.8–1.3)
113
1.0
(0.8–1.3)
14
0.8
(0.5–1.5)
Pre-conception paternal smoking (cigarettes/day) Never
806
371
1.0
Ref.
319
1.0
Ref.
48
1.0
Ref.
Ever \10
585 125
355 82
1.3 1.4
(1.1–1.6) (1.1–2.0)
297 71
1.2 1.4
(1.0–1.5) (1.0–2.0)
51 11
1.5 1.5
(1.0–2.3) (0.8–3.0)
10–15
229
122
1.1
(0.9–1.5)
105
1.1
(0.9–1.5)
15
1.2
(0.6–2.1)
C16
231
151
1.4
(1.1–1.7)
121
1.2
(0.9–1.6)
25
1.9
(1.1–3.2)
p-trend = 0.02
p-trend = 0.13
p-trend = 0.03
Paternal smoking during pregnancy (cigarettes/day) Never
837
384
1.0
Ref.
331
1.0
Ref.
49
1.0
Ref.
Ever
553
345
1.3
(1.1–1.6)
287
1.3
(1.0-1.5)
51
1.7
(1.1–2.5)
\10
118
82
1.5
(1.1–2.1)
70
1.5
(1.1–2.1)
12
1.8
(0.9–3.5)
10–15
222
119
1.1
(0.9-1.5)
102
1.1
(0.9–1.5)
15
1.3
(0.7–2.3)
C16
213
144
1.4
(1.1–1.8)
115
1.3
(1.0-1.7)
24
2.0
(1.2–3.4)
p-trend = 0.01
p-trend = 0.09
p-trend = 0.02
Joint effect of paternal and maternal smoking Pre-conception Neither mother nor father
629
289
1.0
Ref.
249
1.0
Ref.
38
1.0
Only mother
177
80
0.9
(0.7-1.3)
68
0.9
(0.7–1.3)
10
0.9
Ref. (0.5–1.9)
Only father Both
252 330
148 207
1.3 1.3
(1.0–1.6) (1.0–1.6)
120 177
1.2 1.3
(0.9–1.6) (1.0–1.6)
25 26
1.6 1.4
(1.0–2.8) (0.8–2.3) Ref.
During pregnancy Neither mother nor father
749
350
1.0
Ref.
301
1.0
Ref.
47
1.0
Only mother
88
34
0.8
(0.5–1.2)
30
0.8
(0.5–1.3)
2
0.4
(0.1–1.5)
Only father
359
226
1.3
(1.1–1.6)
186
1.3
(1.0–1.6)
35
1.6
(1.0–2.5)
Both
194
119
1.3
(1.0–1.6)
101
1.2
(0.9–1.6)
16
1.4
(0.8–2.6)
Odds Ratios (OR) and 95 % Confidence Intervals (95 % CI) were estimated by unconditional logistic regression including age (\2; 2; 3; 4; 5–6; 7–8; 9–12; C12), sex (boy vs girl), mother’s age at child’s birth (continuous), mother’s education (incomplete secondary; completed secondary; completed tertiary) and birth order (at least second-born vs first-born)
123
1010
agreement between the exposures generated through maternal reports and those generated through paternal selfreports was excellent and similar for the CL cases and controls. For ever/never pre-conception paternal smoking, among the 240 subjects available for the agreement study, 9 were discordant [percentage agreement (pa) = 96 %, kappa (j) = 0.92 (0.88–0.97)]: 2 fathers who reported smoking \10 cigarettes/day were reported to be nonsmokers by the mothers and seven fathers who reported being non-smokers were reported to be smokers of \10 cigarettes/day by the mothers. The agreement was also excellent for ever/never paternal smoking during pregnancy [pa = 95 %, j = 0.91 (0.85–0.96)] and the quantity of cigarettes smoked pre-conception and during pregnancy [pa = 86 %, j = 0.83 (0.77–0.89); pa = 85 %, j = 0.80 (0.74–0.87), respectively]. The analyses performed using exposure definitions based on paternal self-reporting led to similar results, although the numbers were small, particularly for AML.
Maternal coffee and tea drinking during pregnancy (Table 3) Consumption of coffee and tea during pregnancy was reported by 55 and 39 % of the control mothers and by 53 and 37 % of the case mothers. Coffee and tea drinking during pregnancy was not associated with CL for overall consumption, with ORs close to or slightly\1.0. However, a high consumption of coffee ([2 cups/day) was significantly associated with ALL [OR = 1.3 (1.0–1.8)]. No positive association between coffee or tea consumption and AML was observed. Maternal smoking had no influence on the associations.
Maternal alcohol drinking during pregnancy (Table 4) Consumption of any alcohol during pregnancy was reported by 25 % of the control mothers and by 23 % of the case mothers (Table 4). ALL was not associated with alcohol drinking during pregnancy [OR = 1.0 (0.8–1.2)]. For AML a significant positive dose–response trend was evidenced (p-trend = 0.02) and the OR for overall consumption was 1.3 (0.8–2.0). Similar results were observed for alcohol consumption during the first trimester of pregnancy. No interaction between maternal alcohol drinking and coffee consumption during pregnancy was evidenced. However, the risk of ALL was non-significantly increased among children born to mothers who smoked and drank alcohol during pregnancy [OR = 1.4 (0.9–2.1], pinteraction = 0.004).
123
Cancer Causes Control (2015) 26:1003–1017
Sub-group analysis Table 5 shows the association, by subtype of ALL and cytogenetic group, for B-cell precursor ALL (BCP-ALL). No association was detected for T-cell ALL. In contrast, although the numbers were small, Burkitt’s ALL was significantly associated with paternal smoking [OR = 2.5 (1.0–6.0) for pre-conception smoking and 2.3 (1.0–5.4), for smoking during pregnancy], and positively, but not significantly, with maternal smoking. For BCP-ALL, the associations with paternal smoking were similar to those observed for ALL taken together, and stronger for the subgroup with moderate hyperdiploidy. No association was evidenced for ETV6-RUNX1 positive or MLL positive BCP-ALL. Stratification, adjustments and sensitivity analyses Analyses stratified on the child’s age at diagnosis/interview, rural/urban status of the place of residence, maternal age at birth, and parental socioeconomic status yielded similar results. Additional adjustments on rural/urban status at the place of residence, parental socioeconomic status, parental education, breastfeeding and Caucasian status, and exclusion of the cases with a cell phone only also yielded similar results. Mutual adjustment on maternal cigarette smoking, paternal cigarette smoking, maternal alcohol consumption and maternal coffee consumption did not change the results.
Discussion The main results of the present study consist in the positive associations between childhood AML and maternal alcohol drinking during pregnancy, with the ORs increasing with the quantity of alcohol consumed, and paternal pre-conception smoking. Paternal smoking was also associated with ALL, and particularly PBC-ALL with moderate hyperdiploidy. Maternal consumption of coffee and tea during pregnancy was not associated with ALL or AML. The study benefited from the reliable detection system of the French National Registry of Childhood Hematopoietic Malignancies, which has a high degree of exhaustiveness. The participation rate of the eligible case mothers was very high (93 %). Therefore, selection bias related to non-inclusion is unlikely to explain the results. Selection by survival may have occurred, since the mothers of cases who had died were not interviewed. However, only 8 children were not eligible because they had died or were in palliative care, and to the authors’ knowledge, the exposures under study are not expected to impact survival.
Cancer Causes Control (2015) 26:1003–1017
1011
Table 3 Association between CL and self-reported maternal consumption of caffeinated beverages during pregnancy Controls (n = 1,421)
All CL (n = 747)
ALL (n = 636)
n
n
n
OR
95 % CI
OR
AML (n = 100) 95 % CI
n
OR
95 % CI
Coffee during pregnancy (cups/day) Never/occasionally
633
351
1.0
Ref.
298
1.0
Ref.
46
1.0
Ref.
Regular (C1 cup/week)
786
389
0.9
(0.8–1.1)
334
1.0
(0.8–1.2)
51
0.9
(0.6–1.3)
B1 cup/day
366
161
0.8
(0.7–1.0]
135
0.8
(0.6–1.1)
25
0.9
(0.6–1.5)
[1 cup/day to 2 cups/day
211
108
1.0
(0.7–1.3)
88
1.0
(0.7–1.3)
18
1.1
(0.6–1.9)
[2 cups/day
209
8
0.5
(0.2–1.1)
120
1.1
(0.9–1.5)
111
1.3
(1.0–1.7)
Coffee during first trimester of pregnancy Never/occasionally 671
365
1.0
Ref.
310
1.0
Ref.
48
1.0
Ref.
Regular (C1 cup/week)
748
375
1.0
(0.8–1.2)
322
1.0
(0.8–1.2)
49
0.9
(0.6–1.3)
B1 cup/day
359
156
0.8
(0.7–1.0)
131
0.8
(0.6-1.1)
24
0.9
(0.6–1.5)
[1 cup/day to 2 cups/day
187
98
1.0
(0.8–1.3)
80
1.0
(0.7–1.4)
16
1.1
(0.6-2.0)
[2 cups/day
202
121
1.2
(0.9–1.6)
111
1.3
(1.0–1.8)
9
0.6
(0.3–1.2)
Tea during pregnancy Never/occasionally
862
470
1.0
Ref.
406
1.0
Ref.
59
1.0
Ref.
Regular (C1 cup/week)
553
272
0.9
(0.8–1.1)
226
0.9
(0.7–1.1)
40
1.0
(0.7–1.6)
\1 cup/day
134
66
0.9
(0.6–1.2)
56
0.9
(0.6–1.2)
8
0.9
(0.4–1.9)
1 cup/day
234
106
0.8
(0.7–1.1)
86
0.8
(0.6–1.1)
19
1.2
(0.7–2.1)
[1 cup/day
185
100
1.0
(0.8–1.4)
84
1.0
(0.8–1.4)
13
1.0
(0.5–1.8)
Tea during first trimester of pregnancy Never/occasionally
845
468
1.0
Ref.
403
1.0
Ref.
61
1.0
Ref.
Regular (C1 cup/week)
570
274
0.9
(0.7–1.1)
229
0.9
(0.7–1.1)
38
0.9
(0.6-1.4)
\1 cup/day 1 cup/day
127 220
62 101
0.9 0.9
(0.6–1.2) (0.7–1.1)
54 83
0.9 0.8
(0.6–1.3) (0.6–1.1)
6 17
0.6 1.1
(0.3–1.5) (0.6–1.9)
[1 cup/day
178
91
1.0
(0.7-1.3)
77
1.0
(0.7–1.3)
11
0.8
(0.4–1.6)
Joint effect of maternal coffee drinking and tea drinking during pregnancy No exposure
344
197
1.0
Ref.
170
1.0
Ref.
26
1.0
Ref.
Coffee drinking only
517
270
1.0
(0.8–1.2)
234
1.0
(0.8–1.3)
32
0.7
(0.4–1.3)
Tea drinking only
286
152
1.0
(0.8–1.3)
126
1.0
(0.7–1.3)
20
0.8
(0.5–1.5)
Both
267
118
0.8
(0.6–1.1)
99
0.8
(0.6–1.1)
19
0.9
(0.5–1.6)
Joint effect of maternal coffee drinking and cigarette smoking during pregnancy No exposure
555
302
1.0
Ref.
259
1.0
Ref.
39
1.0
Ref.
Coffee drinking only
574
282
0.9
(0.8–1.2)
237
1.0
(0.8–1.2)
42
1.0
(0.6–1.6)
Cigarette smoking only Both
77
49
1.0
(0.7-1.6)
39
1.0
(0.6–1.5)
7
1.4
(0.6–3.3)
212
107
0.9
(0.7–1.2)
97
1.0
(0.7–1.3)
9
0.6
(0.3–1.3)
Odds Ratios (OR) and 95 % Confidence Intervals (95 % CI) were estimated by unconditional logistic regression including age (\2; 2; 3; 4; 5–6; 7-8; 9–12; C12), sex (boy vs girl), mother’s age at child’s birth (continuous), mother’s education (incomplete secondary; completed secondary; completed tertiary) and birth order (at least second-born vs first-born)
Control selection was based on random generation of listed and unlisted telephone numbers. Households with no landline numbers were not accessible for control selection, whereas the cases with a cell phone only (22 % of the cases) were included, resulting in a potential for selection bias. However, excluding the cases with a cell phone only did not change the results. Although the participation rate for the controls was high (86 %), refusal to take part or inability to contact eligible households may have been linked to
socioeconomic characteristics such as parental socioeconomic status or educational level, factors known to be associated with tobacco smoking and alcohol drinking [36, 37]. Compared to figures provided by French national perinatal surveys (ENP) carried out in 1998, 2003 and 2010 [38], the control mothers were slightly more educated than the women in the source population in 1998 and 2003 (Supplementary Table 3), which might be indicative of higher refusal to participate among less educated mothers and have led to underestimation of
123
1012
Cancer Causes Control (2015) 26:1003–1017
Table 4 Association between CL and self-reported maternal alcohol consumption during pregnancy Controls (n = 1,421)
All CL (n = 747)
ALL (n = 636)
n
n
n
OR
95 % CI
OR
AML (n = 100) 95 % CI
n
OR
95 % CI
Alcohol during pregnancy (glass/week) Never
1,069
574
1.0
Ref.
495
1.0
Ref.
69
1.0
Ref.
Ever
348
168
1.0
(0.8–1.2)
136
1.0
(0.8–1.2)
31
1.3
(0.8–2.0)
\1
232
109
1.0
(0.7–1.2)
92
1.0
(0.7–1.3)
17
1.1
(0.6–1.8)
1–2
77
39
1.0
(0.7–1.5)
32
1.0
(0.6–1.5)
7
1.3
(0.6–3.0)
[2
36
18
1.1
(0.6–2.0)
10
0.7
(0.4–1.5)
7
2.8
(1.2–6.7)
p-trend = 0.70
p-trend = 0.38
p-trend = 0.02
Kind of alcohol Wine No Yes
1,189
633
1.0
Ref.
229
110
1.0
(0.8–1.3)
546
1.0
Ref.
77
1.0
Ref.
86
1.0
(0.7–1.3)
23
1.4
(0.9–2.3)
1,271
680
1.0
Ref.
583
1.0
Ref.
86
1.0
Ref.
145
64
0.9
(0.7–1.2)
50
0.8
(0.6–1.2)
14
1.3
(0.7–2.4)
1,321
696
96
49
1.0
Ref.
96
1.0
Ref.
91
1.0
Ref.
1.0
(0.7–1.5)
40
1.0
(0.7–1.5)
9
1.3
(0.6–2.7)
553
1.0
Ref.
80
1.0
Ref. (0.7–2.2)
Beer/Cider No Yes Spirits No Yes
Alcohol during first trimester of pregnancy (glass/week) Never
1,176
643
1.0
Ref.
Ever
206
87
0.9
(0.7–1.1)
68
0.8
(0.6–1.1)
19
1.3
\1
103
36
0.7
(0.5–1.1)
31
0.8
(0.5–1.2)
5
0.7
(0.3–1.7)
1–2 [2
68 34
34 17
0.9 1.1
(0.6–1.4) (0.6–2.0)
27 10
0.9 0.7
(0.6–1.4) (0.4–1.5)
7 7
1.4 3.0
(0.6–3.2) (1.2-7.1)
p-trend = 0.80
p-trend = 0.38
p-trend = 0.01
Joint effect of maternal alcohol drinking and cigarette smoking during pregnancy No exposure
855
470
1.0
Ref.
408
1.0
Ref.
56
1.0
Ref.
Alcohol drinking only
272
115
0.9
(0.7–1.1)
88
0.8
(0.6–1.0)
26
1.3
(0.8–2.2)
Smoking only
213
104
0.8
(0.6–1.1)
87
0.8
(0.6–1.1)
13
0.9
(0.5–1.8)
76
53
1.3
(0.9–1.9)
48
1.4
(0.9–2.1)
5
1.0
(0.4–2.5)
Both
Joint effect of maternal alcohol drinking and coffee drinking during pregnancy No exposure
504
295
1.0
Ref.
253
1.0
Ref.
35
1.0
Ref.
Alcohol drinking only
128
55
0.8
(0.6–1.2)
44
0.8
(0.5-1.2)
11
1.1
(0.5–2.2)
Coffee drinking only
564
273
0.9
(0.7–1.1)
239
0.9
(0.7–1.1)
31
0.7
(0.5–1.2)
Both
220
113
1.0
(0.7–1.3)
92
1.0
(0.7–1.3)
20
1.2
(0.7–2.1)
Odds Ratios (OR) and 95 % Confidence Intervals (95 % CI) were estimated by unconditional logistic regression including age (\2; 2; 3; 4; 5–6; 7–8; 9–12; C12), sex (boy vs girl), mother’s age at child’s birth (continuous), mother’s education (incomplete secondary; completed secondary; completed tertiary) and birth order (at least second-born vs first-born)
maternal alcohol consumption prevalence among the controls. However, the associations were stable across the strata of parental education and parental socioeconomic status. Moreover, with regard to the ENP [38], the control mothers were very similar to the source population in terms of tobacco smoking and alcohol drinking during pregnancy (Supplementary Table 3). Data collection was carefully conducted in order to minimize differential misclassifications. The interviews
123
were conducted under similar conditions for the case and control mothers, during the same period, with a standardized questionnaire, by the same interviewers, who had been specially trained for the study. Moreover, neither the subjects nor the interviewers were aware of the hypotheses under study. Non-differential errors cannot be excluded since exposures are based on the mother’s recall and may have occurred up to 15 years before the interview. This is likely to have biased estimates toward the null hypothesis,
0.7 1.2 1.1
Maternal smoking during pregnancy
Paternal smoking during pre–conception
Paternal smoking during pregnancy
0.9
Regular tea drinking (C1 cup/w)
0.6 0.9
Beer/cider
Liquor
1.1 1.0 1.2 1.3
Paternal smoking during pre-conception
Paternal smoking during pregnancy
OR
Maternal smoking during pregnancy
Regular tea drinking (C1 cup/w) 1.0 1.1 0.8
Any
Wine Beer/Cider
Alcohol drinking
1.0 0.9
Regular coffee drinking (C1 cup/w)
Maternal drinking during pregnancy
(0.5–1.9)
(0.3–1.1)
(0.7–1.7)
(0.6–1.4)
2
(0.8–1.5) (0.5–1.2)
(0.8–1.3)
(0.7–1.2)
(0.8–1.2)
(1.0–1.6)
(1.0–1.5)
(0.8–1.3)
(0.9–1.4)
95 % CI
Negative (n = 465)
MLL
Maternal smoking during pre-conception
Parental smoking
1.0
Wine
(0.7–1.3)
(0.7–1.3)
(0.8–1.6)
(0.8–1.6)
(0.5–1.1)
(0.6–1.1)
BCP-ALL (n = 503)
0.9
Any
Alcohol drinking
1.0
Regular coffee drinking (C1 cup/w)
Maternal drinking during pregnancy
0.8
Maternal smoking during pre-conception
Parental smoking
0.8
1.3
2.4
1.8
1.1
1.5
1.7
1.6
1.8
1.5
1.3 1.7
1.4
0.5
1.1
1.5
1.7
1.5
1.1
OR2
(0.4–4.0) (0.6–5.2)
(0.6–3.6)
(0.2–1.3)
(0.5–2.5)
(0.7–3.2)
(0.8–3.6)
(0.6–3.8)
(0.5–2.4)
95 % CI
Positive (n = 29)
(0.3–2.4)
(0.6–2.9)
(1.3–4.3)
(1.0–3.0)
(0.7–1.8)
(0.9–2.4)
(1.1–2.9)
(1.0–2.7)
(1.1–3.2)
(0.9–2.5)
95 % CI
OR2
OR2 95 % CI
47–50 chr (n = 71)
None (n = 189)
Numerical abnormalities
BCP-ALL (n=503)
(0.5–2.2)
(0.6–1.8)
(0.6–1.7)
(0.7–1.6)
(0.6–1.2)
(0.7–1.4)
(1.1–2.1)
(1.0–2.0)
(0.7–1.6)
(0.9–1.8)
95 % CI
1.0 0.8
1.0
0.9
1.0
1.3
1.2
1.0
1.1
OR1
(0.8–1.4) (0.6–1.2)
(0.8–1.3)
(0.7–1.1)
(0.8–1.2)
(1.0–1.6)
(1.0–1.6)
(0.8–1.3)
(0.9–1.3)
95 % CI
Total BCP-ALL
1.1
1.0
1.0
1.0
0.9
1.0
1.5
1.4
1.0
1.3
OR2
[50 chr (n = 163)
(0.7–1.8)
(0.5–1.3)
(0.9–1.7)
(0.9–1.5)
(0.7–1.2)
(0.8–1.3)
(1.1–1.8)
(1.0–1.7)
(0.7–1.3)
(0.8–1.4)
95 % CI
0.9 0.9
0.8
0.9
1.0
2.3
2.5
1.7
1.5
OR1
(0.2–3.1) (0.2–3.8)
(0.3–2.3)
(0.4–2.2)
(0.5–2.4)
(1.0–5.4)
(1.0–6.0)
(0.7–4.1)
(0.7–3.4)
95 % CI
Burkitt ALL (n = 26)
1.1
0.8
1.2
1.2
0.9
1.0
1.4
1.3
1.0
1.1
OR2
Negative (n = 349)
ETV6-RUNX1
(0.3–2.1)
(0.5–2.0)
(0.4–1.6)
(0.4–1.3)
(0.8–1.8)
(0.6–1.3)
(0.6–1.4)
(0.7–1.5)
(0.5–1.5)
(0.5–1.3)
95 % CI
0.6 1.0
0.8
0.8
0.8
1.1
1.0
0.9
1.1
OR1
(0.3–1.2) (0.5–2.1)
(0.5–1.4)
(0.5–1.3)
(0.5–1.2)
(0.7–1.7)
(0.7–1.6)
(0.6–1.6)
(0.7–1.7)
95 % CI
T ALL (n = 95)
0.7
1.0
0.8
0.7
1.2
0.8
0.9
1.0
0.9
0.8
OR2
Positive (n = 119)
Table 5 Association between ALL and self-reported parental consumption of tobacco, maternal consumption of coffee, tea and alcohol, by ALL subtype and cytogenetic group
Cancer Causes Control (2015) 26:1003–1017 1013
123
Odds Ratios (OR) and 95 % Confidence Intervals (95 % CI) were estimated by unconditional logistic regression including age (\2; 2; 3; 4; 5–6; 7–8; 9–12; C12), sex (boy vs girl), mother’s education (incomplete secondary; completed secondary; completed tertiary), birth order (at least second-born vs. first-born) and mother’s age at child’s birth (continuous)
2
1.1 Liquor
Odds Ratios (OR) and 95 % Confidence Intervals (95 % CI) were estimated by polytomous logistic regression including age (\2; 2; 3; 4; 5–6; 7–8; 9–12; C12), sex (boy vs girl), mother’s education (incomplete secondary; completed secondary; completed tertiary), birth order (at least second-born vs. first-born) and mother’s age at child’s birth (continuous)
1
0.8 (0.2–4.6) 1.0 (0.7–1.7) 1.1 (0.1–6.1) 0.8
OR1 95 % CI 95 % CI OR2 95 % CI OR
2
Negative (n = 465)
Positive (n = 29)
OR1
Total BCP-ALL MLL
(0.7–1.7)
95 % CI
OR1
T ALL (n = 95) Burkitt ALL (n = 26) BCP-ALL (n = 503) Table 5 continued
123
(0.3–1.9)
Cancer Causes Control (2015) 26:1003–1017 95 % CI
1014
and decreased statistical power. This probably applies less to the mothers of the youngest children. However, the analyses stratified on the child’s age at diagnosis/interview generated similar results, and, in particular, no association with maternal smoking was evidenced. Moreover, using a more specific definition of tobacco exposure (C1 cigarette/day), which reduced non-differential misclassification, did not change the results. Stopping tobacco and alcohol consumption during pregnancy is widely recommended in the context of public health programmes. Reporting smoking and alcohol drinking during pregnancy may be accompanied by guilt, possibly differential, particularly for the highest consumptions. A true association might be masked if case mothers under-reported their tobacco consumption more often than control mothers. Thus, with the hypothesis of a true association with OR = 1.2 between maternal smoking during pregnancy and AL, and a true smoking prevalence of 21 % among control mothers, case mothers would have under-reported tobacco smoking fourfold more often than control mothers, which seems unlikely. Similarly, differential over-reporting of paternal tobacco consumption by case mothers might have biased the association observed with paternal smoking. Nevertheless, the agreement study conducted on a subsample of 247 participants in the ESTELLE study showed that the agreement between maternal interview reports and paternal self-reports was excellent and similar for the cases and controls. Control fathers were also similar to the source population in the same age group in terms of tobacco smoking, when compared to estimates from the 2005 and 2010 French Health Barometer surveys [39] (Supplementary Table 3). Thus, underreporting of paternal smoking by control mothers also seems unlikely to explain the association. It is also unlikely that differential reporting of maternal alcohol consumption during pregnancy would occur specifically for AML, and with a significant dose–response relationship. Potential confounding by factors related to child age and sex were controlled by adjusting for those variables. Additional adjustments for other socio-demographic factors (urban/rural status, parental socioeconomic status) did not change the results. None of the exposures of interest was a confounder of the others, since mutual adjustments did not change the results. In line with the results of the ESTELLE study, previous studies [11, 16, 19, 20, 27, 31, 32] included in the metaanalysis by Latino-Martel et al. [30] suggested that maternal alcohol drinking during pregnancy might be associated with AML [meta-OR = 1.6 (1.1–2.2)], but not with ALL [meta-OR = 1.1 (0.9–1.3)]. Latino-Martel et al. also observed that the increase in AML risk might be restricted to children aged 0 to 4 years, which was not the case in the present study.
Cancer Causes Control (2015) 26:1003–1017
Although the results of this study were indicative of an association between ALL and the highest level of coffee intake, the study did not clearly replicate the positive association between CL and maternal coffee consumption that was observed in two previous case–control studies led by the authors’ team [16, 33]. An Australian study did not show any positive association with ALL [34], while two other studies reported a non-significant positive association with the highest coffee consumption [5, 18]. These inconsistent results call for clarification by further studies and pooled analyses in large consortia. In this study, pre-conception paternal smoking was associated with both ALL [OR = 1.2 (1.0–1.5)] and AML [OR = 1.5 (1.0–2.3)\. Consistently with the present study, a recent meta-analysis [40], which included most of the pertinent published studies [8–15, 17–20, 25, 26], reported pooled estimates of 1.3 (1.1–1.5) and 1.2 (1.1–1.4) for ALL and paternal smoking during pre-conception and during pregnancy. As in the ESTELLE study, AML was significantly positively associated with pre-conception paternal smoking in the French population-based ESCALE case–control study [20] [OR = 1.5 (1.0–2.3)], with a significant dose–response trend, but not in the French ELECTRE case–control study [18] [OR = 0.9 (0.5–1.6)]. The American NCCLS study [17] reported a significant positive association between AML and pre-conception paternal smoking [OR = 3.8 (1.0–14.7)], but the estimate was only based on 46 cases. A recent update of the NCCLS study [22], in which case recruitment was extended until 2008, showed a non-significant association with prenatal (i.e., 3 months before conception and/or during pregnancy) paternal smoking [OR = 1.4 (0.8–2.2)]. In the SETIL study [24], pre-conception paternal smoking was significantly positively associated with AML [OR = 1.8 (1.0–3.2] for smokers of more than 10 cigarettes per day). In this study, no association between maternal smoking during pregnancy and ALL or AML was observed. Similarly, the results of most of the European [7, 10, 14– 16, 18, 20, 23, 25, 32, 41], North-American [11, 13, 17, 19, 22, 31, 42, 43] and Australian [21] case–control studies were also not indicative of an association between maternal smoking and ALL or AML. This was clearly indicated in a recent meta-analysis [44] that reported ORs of 1.0 (0.9–1.1) for ALL and 1.0 (0.9–1.1) for AML with respect to maternal smoking during pregnancy. The results for paternal smoking were strengthened for BCP-ALL with moderate hyperdiploidy and an association with maternal alcohol drinking was evidenced for BCP-ALL with moderate hyperdiploidy. Only two studies have reported the results for prenatal paternal smoking by cytogenetic group, with inconclusive results. In the NCCLS study [22], prenatal paternal smoking was not associated with high hyperdiploidy BCP-ALL [OR = 1.0
1015
(0.6–1.4)], while Milne et al. [21] reported a positive association [OR = 1.5 (0.9–2.3)]. To the authors’ knowledge, investigation for an association with maternal alcohol drinking by cytogenetic group had not yet been conducted. In conclusion, the findings of this study constitute additional evidence that maternal alcohol drinking during pregnancy may be involved in AML, and that paternal smoking before pregnancy may be a risk factor for CL. The role of maternal coffee drinking in CL remains unclear and should be investigated further in consortium analyses and in large birth cohort studies with exposure assessment more contemporaneous with the exposure, before the occurrence of the disease. Acknowledgments The authors are grateful to: Noureddine Balegroune, Sofie`ne Ben Salha and the team of clinical research associates who contributed to the recruitment of the cases; Laure Faure and the staff of the French National Registry of Childhood Hematopoietic Malignancies, who contributed to case detection and verification; Christophe David and the team of interviewers (Institut IPSOS), who recruited the controls and interviewed the cases and controls and Elsa Charles for her technical assistance The authors would like to thank all of the Socie´te´ Franc¸aise de lutte contre les Cancers de l’Enfant et de l’Adolescent (SFCE) principal investigators: Andre´ Baruchel (Hoˆpital Saint-Louis/Hoˆpital Robert Debre´, Paris), Claire Berger (Centre Hospitalier Universitaire, Saint-Etienne), Christophe Bergeron (Centre Le´on Be´rard, Lyon), Ge´rard Michel (Hoˆpital La Timone, Marseille), Yves Bertrand (Institut d’He´matologie et d’Oncologie Pe´diatrique, Lyon), Pascal Chastagner (Centre Hospitalier Universitaire, Nancy), Patrick Boutard (Centre Hospitalier Re´gional Universitaire, Caen), Ge´rard Couillault (Hoˆpital d’Enfants, Dijon), Christophe Piguet (Centre Hospitalier Re´gional Universitaire, Limoges), Anne-Sophie Defachelles (Centre Oscar Lambret, Lille), Franc¸ois Demeocq (Hoˆpital Hoˆtel-Dieu, Clermont-Ferrand), Alain Fischer (Hoˆpital des Enfants Malades, Paris), Virginie Gandemer (Centre Hospitalier Universitaire– Hoˆpital Sud, Rennes), Dominique Valteau-Couanet (Institut Gustave Roussy, Villejuif), Philippe Colombat (Centre Gatien de Clocheville, Tours), Frederic Millot (Centre Hospitalier Universitaire Jean Bernard, Poitiers), Guy Leverger (Hoˆpital Armand-Trousseau, Paris), Patrick Lutz (Hoˆpital de Hautepierre, Strasbourg), Nicolas Sirvent (Hoˆpital Arnaud de Villeneuve, Montpellier), Xavier Rialland (Hoˆpital Me`re et Enfants, Nantes), Martine Mu¨nzer (American Memorial Hospital, Reims), Brigitte Nelken (Hoˆpital Jeanne de Flandre, Lille), Franc¸ois Doz (Institut Curie, Paris), Brigitte Pautard (Centre Hospitalier Universitaire, Amiens), Yves Perel (Hoˆpital Pellegrin Tripode, Bordeaux), Alain Pierre-Kahn (Hoˆpital Enfants Malades, Paris), Emmanuel Plouvier (Centre Hospitalier Re´gional, Besanc¸on), Xavier Rialland (Centre Hospitalier Universitaire, Angers), Alain Robert (Hoˆpital des Enfants, Toulouse), Herve´ Rubie (Hoˆpital des Enfants, Toulouse), Nicolas Sirvent (L’Archet, Nice), Marilyne Poiree (Fondation Lenval, Nice), Jean-Pierre Vannier (Hoˆpital Charles Nicolle, Rouen), Dominique Plantaz (Centre Hospitalier Universitaire, Grenoble), Philippe Lemoine (Hoˆpital Morvan, Brest) and Christian Sainte Rose (Centre Hospitalier Universitaire Necker, Paris). Funding This work was supported by grants from the Ligue Nationale Contre le Cancer (LNCC), the Agence Nationale de Se´curite´ Sanitaire de l’alimentation, de l’Environnement et du Travail (PNREST Anses, Cancer TMOI AVIESAN, 2013/1/248), the Institut
123
1016
Cancer Causes Control (2015) 26:1003–1017
National du Cancer (INCa), the association Enfants et Sante´ and the Agence Nationale de la Recherche (ANR-10-COHO-0009). Ethical Standards The study protocol complied with the French regulations relating to databases and ethics and the pertinent approvals (CNIL, French Data Protection Authority, no. 90828 and CPP Idf IV, Comitee for Personal Protection, no. 2008/12NICB, respectively) were obtained. Conflict of interest of interest.
The authors declare that they have no conflict
References 1. Lacour B, Guyot-Goubin A, Guissou S, Bellec S, Desandes E, Clavel J (2010) Incidence of childhood cancer in France: national children cancer registries, 2000–2004. Eur J Cancer Prev 19:173–181 2. IARC (2004) Tobacco smoke and involuntary smoking. IARC Monogr Eval Carcinog Risks Hum 83:1–1438 3. IARC (2010) Alcohol consumption and ethyl carbamate. IARC Monogr Eval Carcinog Risks Hum 96:3–1383 4. IARC (1991) Coffee, tea, mate, methylxanthines and methylglyoxal. IARC working group on the evaluation of carcinogenic risks to humans. Lyon, 27 February to 6 March 1990. IARC Monogr Eval Carcinog Risks Hum 51:1–513 5. Ross JA, Potter JD, Reaman GH, Pendergrass TW, Robison LL (1996) Maternal exposure to potential inhibitors of DNA topoisomerase II and infant leukemia (United States): a report from the children’s cancer group. Cancer Causes Control 7:581–590 6. Zenzes MT (2000) Smoking and reproduction: gene damage to human gametes and embryos. Hum Reprod Update 6:122–131 7. van Steensel-Moll HA, Valkenburg HA, Vandenbroucke JP, van Zanen GE (1985) Are maternal fertility problems related to childhood leukaemia? Int J Epidemiol 14:555–559 8. Magnani C, Pastore G, Luzzatto L, Terracini B (1990) Parental occupation and other environmental factors in the etiology of leukemias and non-Hodgkin’s lymphomas in childhood: a case– control study. Tumori 76:413–419 9. John EM, Savitz DA, Sandler DP (1991) Prenatal exposure to parents’ smoking and childhood cancer. Am J Epidemiol 133:123–132 10. Sorahan T, Lancashire R, Prior P, Peck I, Stewart A (1995) Childhood cancer and parental use of alcohol and tobacco. Ann Epidemiol 5:354–359 11. Shu XO, Ross JA, Pendergrass TW, Reaman GH, Lampkin B, Robison LL (1996) Parental alcohol consumption, cigarette smoking, and risk of infant leukemia: a childrens cancer group study. J Natl Cancer Inst 88:24–31 12. Ji BT, Shu XO, Linet MS et al (1997) Paternal cigarette smoking and the risk of childhood cancer among offspring of nonsmoking mothers. J Natl Cancer Inst 89:238–244 13. Brondum J, Shu XO, Steinbuch M, Severson RK, Potter JD, Robison LL (1999) Parental cigarette smoking and the risk of acute leukemia in children. Cancer 85:1380–1388 14. Sorahan T, McKinney PA, Mann JR et al (2001) Childhood cancer and parental use of tobacco: findings from the inter-regional epidemiological study of childhood cancer (IRESCC). Br J Cancer 84:141–146 15. Pang D, McNally R, Birch JM (2003) Parental smoking and childhood cancer: results from the United Kingdom Childhood Cancer Study. Br J Cancer 88:373–381 16. Menegaux F, Steffen C, Bellec S et al (2005) Maternal coffee and alcohol consumption during pregnancy, parental smoking and risk of childhood acute leukaemia. Cancer Detect Prev 29:487–493
123
17. Chang JS, Selvin S, Metayer C, Crouse V, Golembesky A, Buffler PA (2006) Parental smoking and the risk of childhood leukemia. Am J Epidemiol 163:1091–1100 18. Menegaux F, Ripert M, Hemon D, Clavel J (2007) Maternal alcohol and coffee drinking, parental smoking and childhood leukaemia: a French population-based case–control study. Paediatr Perinat Epidemiol 21:293–299 19. MacArthur AC, McBride ML, Spinelli JJ, Tamaro S, Gallagher RP, Theriault G (2008) Risk of childhood leukemia associated with parental smoking and alcohol consumption prior to conception and during pregnancy: the cross-Canada childhood leukemia study. Cancer Causes Control 19:283–295 20. Rudant J, Menegaux F, Leverger G et al (2008) Childhood hematopoietic malignancies and parental use of tobacco and alcohol: the ESCALE study (SFCE). Cancer Causes Control 19:1277–1290 21. Milne E, Greenop KR, Scott RJ et al (2012) Parental prenatal smoking and risk of childhood acute lymphoblastic leukemia. Am J Epidemiol 175:43–53 22. Metayer C, Zhang L, Wiemels JL et al (2013) Tobacco smoke exposure and the risk of childhood acute lymphoblastic and myeloid leukemias by cytogenetic subtype. Cancer Epidemiol Biomark Prev 22:1600–1611 23. Farioli A, Legittimo P, Mattioli S et al (2014) Tobacco smoke and risk of childhood acute lymphoblastic leukemia: findings from the SETIL case–control study. Cancer Causes Control 25:683–692 24. Mattioli S, Farioli A, Legittimo P et al (2014) Tobacco smoke and risk of childhood acute non-lymphocytic leukemia: findings from the SETIL study. PLoS One 9:e111028 25. Schuz J, Kaatsch P, Kaletsch U, Meinert R, Michaelis J (1999) Association of childhood cancer with factors related to pregnancy and birth. Int J Epidemiol 28:631–639 26. Lee KM, Ward MH, Han S et al (2009) Paternal smoking, genetic polymorphisms in CYP1A1 and childhood leukemia risk. Leuk Res 33:250–258 27. Alexander FE, Patheal SL, Biondi A et al (2001) Transplacental chemical exposure and risk of infant leukemia with MLL gene fusion. Cancer Res 61:2542–2546 28. Infante-Rivard C, Krajinovic M, Labuda D, Sinnett D (2002) Childhood acute lymphoblastic leukemia associated with parental alcohol consumption and polymorphisms of carcinogen-metabolizing genes. Epidemiology 13:277–281 29. Milne E, Greenop KR, Scott RJ et al (2013) Parental alcohol consumption and risk of childhood acute lymphoblastic leukemia and brain tumors. Cancer Causes Control 24:391–402 30. Latino-Martel P, Chan DS, Druesne-Pecollo N, Barrandon E, Hercberg S, Norat T (2010) Maternal alcohol consumption during pregnancy and risk of childhood leukemia: systematic review and meta-analysis. Cancer Epidemiol Biomark Prev 19:1238–1260 31. Severson RK, Buckley JD, Woods WG, Benjamin D, Robison LL (1993) Cigarette smoking and alcohol consumption by parents of children with acute myeloid leukemia: an analysis within morphological subgroups–a report from the Childrens Cancer Group. Cancer Epidemiol Biomark Prev 2:433–439 32. van Duijn CM, van Steensel-Moll HA, Coebergh JW, van Zanen GE (1994) Risk factors for childhood acute non-lymphocytic leukemia: an association with maternal alcohol consumption during pregnancy? Cancer Epidemiol Biomark Prev 3:457–460 33. Bonaventure A, Rudant J, Goujon-Bellec S et al (2013) Childhood acute leukemia, maternal beverage intake during pregnancy, and metabolic polymorphisms. Cancer Causes Control 24:783–793 34. Milne E, Royle JA, Bennett LC et al (2011) Maternal consumption of coffee and tea during pregnancy and risk of childhood ALL: results from an Australian case–control study. Cancer Causes Control 22:207–218 35. Ajrouche R, Rudant J, Orsi L et al.(2014) Maternal reproductive history, fertility treatments and folic acid supplementation in the
Cancer Causes Control (2015) 26:1003–1017
36.
37.
38.
39.
40.
risk of childhood acute leukemia: the ESTELLE Study. Cancer Causes Control Peretti-Watel P, Constance J, Seror V, Beck F (2009) Cigarettes and social differentiation in France: is tobacco use increasingly concentrated among the poor? Addiction 104:1718–1728 Beck F, Legleye S, Maillochon F, de Peretti G (2009) Gender issue in the analysis of addictive behaviours observed in the Barome`tre sante´ study. France 2005:90–93 Blondel B, Lelong N, Kermarrec M, Goffinet F (2012) Trends in perinatal health in France between 1995 and 2010: results from the National Perinatal Surveys. J Gynecol Obstet Biol Reprod (Paris) 41:151–166 Beck F, Guignard R, Richard JB, Wilquin JL, Peretti-Watel P (2009) Increasing trends in smoking in France: main results of the French Health Barometer. France 2010:230–233 Liu R, Zhang L, McHale CM, Hammond SK (2011) Paternal smoking and risk of childhood acute lymphoblastic leukemia: systematic review and meta-analysis. J Oncol 2011:854584
1017 41. Cnattingius S, Zack M, Ekbom A, Gunnarskog J, Linet M, Adami HO (1995) Prenatal and neonatal risk factors for childhood myeloid leukemia. Cancer Epidemiol Biomark Prev 4:441–445 42. Infante-Rivard C, Krajinovic M, Labuda D, Sinnett D (2000) Parental smoking, CYP1A1 genetic polymorphisms and childhood leukemia (Quebec, Canada). Cancer Causes Control 11:547–553 43. Okcu MF, Goodman KJ, Carozza SE et al (2002) Birth weight, ethnicity, and occurrence of cancer in children: a populationbased, incident case–control study in the State of Texas, USA. Cancer Causes Control 13:595–602 44. Klimentopoulou A, Antonopoulos CN, Papadopoulou C et al (2012) Maternal smoking during pregnancy and risk for childhood leukemia: a nationwide case–control study in Greece and meta-analysis. Pediatr Blood Cancer 58:344–351
123