J Neurol DOI 10.1007/s00415-015-7704-9
ORIGINAL COMMUNICATION
Reduced duration of breastfeeding is associated with a higher risk of multiple sclerosis in both Italian and Norwegian adult males: the EnvIMS study Giammario Ragnedda1,2 • Stefania Leoni1,2 • Maria Parpinel3 • Ilaria Casetta4 • Trond Riise5 • Kjell-Morten Myhr6,7 • Christina Wolfson8,9 • Maura Pugliatti1,2,5,10
Received: 2 December 2014 / Revised: 25 February 2015 / Accepted: 7 March 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Breastfeeding for at least 4 months has been found to be associated with a reduced risk of immunemediated diseases including multiple sclerosis (MS). Using data from a large multinational case–control study (EnvIMS), the association between MS and breastfeeding was investigated in two distinct populations. A questionnaire (EnvIMS-Q) which included a section on feeding during the first year of life was administered to MS cases and to age and sex frequency-matched controls from Italy and Norway. Logistic regression was used to estimate the odds ratio (ORs) and 95 % confidence intervals (95 % CIs) as a measure of the association between MS and exposure to prolonged breastfeeding (4 months or more, used as the reference category), vs. no breastfeeding or breastfeeding for less than 4 months (reduced exposure). Education, smoking habits, smoking in mother’s pregnancy, and other types of milk used in infant feeding were included as covariates. A total of 547 cases and 1039 controls in Italy, and 737 cases and 1335 controls in Norway were studied. The distribution of prolonged (reference) breastfeeding
differed between the Norwegian (65.4 %) and the Italian (48.9 %) study participants. A significant association between MS and reduced/no exposure to breastfeeding was found overall for Italy (ORadj = 1.37; 95 % CI 1.09, 1.73), but not for Norway (ORadj = 1.14; 95 % CI 0.92, 1.40). However, only in men, significant associations were observed for both populations (ORItaly = 2.33; 95 % CI 1.50, 3.65, ORNorway = 2.13; 95 % CI 1.37, 3.30). Reduced exposure to breastfeeding in males was found to be associated with increased risk of MS in Italy and in Norway.
& Maura Pugliatti [email protected]
6
The Norwegian Multiple Sclerosis Registry and Biobank, Department of Neurology, Haukeland University Hospital, Bergen, Norway
7
Department of Clinical Medicine, KG Jebsen Centre for MS-Research, University of Bergen, Norway Haukeland University Hospital, Bergen, Norway
8
Research Institute of the McGill University Health Centre, Montreal, QC, Canada
9
Department of Epidemiology and Biostatistics and Occupational Health, McGill University, Montreal, QC, Canada
10
Department of Medicine, Faculty of Medicine, McGill University, Montreal, Canada
1
2
3
4
5
Department of Biomedical Sciences, University of Sassari, Sassari, Italy Department of Clinical and Experimental Medicine, University of Sassari, Viale San Pietro 10, 07100 Sassari, Italy Unit of Hygiene and Epidemiology, Department of Medical and Biological Sciences, University of Udine, Udine, Italy Department of Biomedical and Specialty Surgical Sciences, Section of Clinical Neurology, University of Ferrara, Ferrara, Italy
Keywords Multiple sclerosis Breastfeeding Case–control studies Risk factors Epidemiology
Introduction Multiple sclerosis (MS) is the most common inflammatory disease of the central nervous system affecting young adults [1]. Despite the growing acceptance of a genetic
Department of Global Public Health and Primary Care, University of Bergen, Bergen, Norway
123
J Neurol
susceptibility for MS, the increasing incidence observed during the last 50 years in Western countries cannot be explained by changes in genetic risk factors, emphasizing the importance of environmental factors related to lifestyle [2–6]. Similar incidence trends have been observed for other immune-mediated diseases such as type-1 diabetes (T1D), inflammatory bowel diseases (IBD) and asthma [7]. Breastfeeding has previously been reported to influence the risk of T1D, and recently, patients developing diabetes have been shown to have a shorter average duration of breastfeeding (3.3 vs. 4.6 months) and to have been exposed to cow’s milk earlier, compared to unaffected siblings [8, 9]. Breastfeeding has also been shown a protective role for other autoimmune diseases like IBD, atopic dermatitis, and asthma [10–12]. Breastfeeding for at least 4 months has also recently been reported to protect against the risk of developing MS, but the impact of shorter duration/no breastfeeding in determining the risk for MS is still debated [13]. We therefore investigated the role of breastfeeding in determining MS risk by examining both country specific and pooled associations using data from a large multinational case–control study (the EnvIMS Study).
Methods Study design The study design and methodology have been reported elsewhere [14]. Briefly, this study builds upon an international multicentre case–control study of environmental risk factors in MS (the EnvIMS Study) which aims to examine possible associations between MS status and past exposures such as infections, sunlight and lifestyle factors, by means of a shared standard methodology in populations differing in disease risk and possibly in the distribution of some environmental factors [15, 16]. A postal self-administered questionnaire (the EnvIMS-Q) was designed in English, translated into the language of the participating countries and tested for reliability, crosscultural validity and perceived difficulty of its use, to record self-reported age-specific exposures in individuals with MS as well as in non-MS controls [14]. EnvIMS-Q is divided into sections representing domains of the most interesting exposures in MS etiological research: sun exposure, diet, intake of vitamin supplements, medical history focusing on childhood infections and immunemediated comorbidity, cigarette smoking habits, physical activity, occupational factors, and reproductive factors for female participants. The EnvIMS study received ethics approval at each study site [14].
123
Study population The Italian and Norwegian EnvIMS population samples have been included for this study. The questionnaire was sent to cases and controls aged 18 years or older and return of the EnvIMS-Q questionnaire was considered evidence of informed consent. All cases diagnosed according to the McDonald criteria and with a clinical onset within 10 years prior to data collection were selected from populationbased MS databases [14, 17]. Four times as many age and sex frequency-matched controls were randomly selected from general population health administrative databases of each region under study. EnvIMS-Q was sent to 1368 MS patients and 4728 controls in Norway, and 1692 cases and 6414 controls in Italy. The response rates were 69.7 and 36.3 % among cases and controls in Norway, and 42.8 and 20.9 %, respectively, in Italy. Exposure The primary exposure of interest in this study was ‘feeding in the first year of life’ and was assessed with one specific question in the Diet section in both the Norwegian and the Italian EnvIMS-Q [14]. In Norway, participants were asked ‘‘Were you breastfed?’’, with the following response options: ‘‘No’’, ‘‘Don’t know’’ or ‘‘Yes’’. If the response was ‘‘Yes’’, then they were asked ‘‘For how many months?’’ with the following response options: ‘‘1–3 months’’, ‘‘4–6 months’’, ‘‘7–9 months’’, ‘‘for 10 months or longer’’. In Italy, participants were able to provide detail, i.e., whether they received only ‘‘Breastfeeding’’, ‘‘Artificial milk (powdered milk)’’, ‘‘Other sugary liquids (chamomile, herb infusions, etc.)’’, ‘‘Other milks (goat, sheep, cow, donkey)’’, or in combination. For each of these options, participants were asked to provide information on the duration (number of months). For each of the indicated durations (1–3, 4–6, 7–9, and 10? months), they could choose the ‘‘I don’t know’’ option. Despite differences in the question design between the Norwegian and Italian questionnaires, it was possible to harmonize the responses into two risk categories for both populations. Based on existing literature as well as previous work on MS, we defined the following three categories: (1) ‘‘no breastfeeding or breastfeeding up to 3 months’’ (i.e., no/reduced breastfeeding), (2) ‘‘unknown’’ including those who ticked off ‘‘don’t know’’ (n = 428) and those who did not respond to one or more of the breastfeeding items (n = 720) and (3) ‘‘breastfeeding for 4 months and longer’’ (i.e., prolonged) (reference) [8, 9, 11–13]. In any of these periods, breastfeeding could have occurred exclusively or in combination with intake of other types of milk or liquids. Since it was not possible to distinguish exclusive from non-exclusive
J Neurol
breastfeeding in Norway, breastfeeding was treated as ‘non-exclusive’ for Italy as well. Covariates Other covariates were included in the analyses as possible confounders: Sex, participant education (none, primary, junior high, secondary school degree, and university degree), smoking habits (never-smoker, ever-smoker), smoking in mother’s pregnancy (yes, don’t know/missing, no), other types of milk (artificial and animal, only for the Italian sub-analysis). Whether or not the participants received help from parents in completing the questionnaires was also recorded. Statistical analysis Frequencies (percent) were reported for categorical variables and mean [standard deviation (SD)] for continuous variables. We used logistic regression to estimate odds ratios (ORs) and 95 % confidence intervals (95 % CIs) as the measure of the association between MS and past exposure to prolonged breastfeeding. In addition to breastfeeding, the model also included sex, participants’ education, smoking habits and smoking in mother’s pregnancy as potential confounders. For the purposes of matching cases and controls on ‘age at onset’, and to ensure that the exposure opportunities for cases and controls were frequency-matched (e.g., participant’s education, smoking habit, help from parents in filling out the questionnaire), each control was assigned an index age corresponding to the age at onset of a case frequencymatched by year of birth [18]. To test whether there was effect modification due to country (Italy vs. Norway) and sex, we included the appropriate interaction terms in the model. The Statistical Package for the Social Sciences (SPSS) version 19 for Windows and OSX (SPSS Inc., IBM, Somers, New York, USA) was used for statistical analysis.
Results A total of 1586 (88.4 %) Italian participants and 2072 (80.9 %) Norwegian participants provided information about breastfeeding. Individuals who did not provide any kind of information regarding the question about feeding in the first year of life, and those who ticked only ‘I don’t know’ options were excluded from the analysis. The demographic and main characteristics of the study population and of the individuals who did not provide information about breastfeeding (‘non-responders’) are given in Table 1.
In the Norwegian study population, 127 (14.5 %) cases and 202 (13.0 %) controls, respectively, ticked the ‘I don’t know’ option; this option was not conceived for the 3-month periods of breastfeeding in the Norwegian EnvIMS-Q. In the Italian study population, the ‘I don’t know’ option was ticked in 35 (46.1 %) and 56 (37.6 %) cases and controls, respectively, within the stratum of no/ reduced breastfeeding; in 41 (53.9 %) and 93 (62.4 %) cases and controls, respectively, within the stratum of prolonged breastfeeding. A higher proportion of participants, breastfed for 4 months and longer in the overall EnvIMS Norwegian sample (65.4 %) as compared to the Italian sample (48.9 %). In the pooled data, 55.5 % of cases reported prolonged breastfeeding as compared to 59.7 % of controls. When stratifying for area (Norway vs. Italy), no association was found between MS status and no/reduced breastfeeding in Norway (OR = 1.13; 95 %CI 0.91, 1.40), whereas a statistically significant association was found for Italy (OR = 1.37; 95 %CI 1.09, 1.73). However, the difference in estimate was not statistically significant between the two countries (p = 0.64). In the Italian dataset, these measures of association remained significant after adjustment for the addition of artificial and animal milk (data not shown). After stratification by sex, a strong association was found in both Norwegian and Italian men, but not in women (Table 2). This difference between men and women was statistically significant (p = 0.0003). The results did not change even after including those who ticked ‘I don’t know’ option in the analysis, nor when adjusting for sun exposure at age 0–5 years nor obesity (data not shown).
Discussion Breastfeeding has been reported to influence the risk of developing immune-mediated diseases, but to date, it has been inadequately addressed in MS [8–13, 19, 20]. In this study, we have shown that not being breastfed or being breastfed for 3 months or less, might contribute to an increasing risk for adult MS and that this association may be sex- and population-specific. This association has previously been investigated, yielding conflicting results [19, 20]. Recently, Conradi et al. [13] reported on the possible protective role of prolonged breastfeeding (defined as 4 months or more) and the risk for MS. In their case–control study, conducted on 245 MS patients and 296 controls, ‘no breastfeeding’ was used as the reference category. We question the choice of the reference value, since the complete absence of exposure to
123
123 538 (73.0) 36.4 (9.7) (35.7, 37.1) 7.1 (2.7) (6.9, 7.3)
(95 % CIs)
Sex, females (%)
Age at disease onset, years, mean (SD)b
(95 % CIs)
Disease duration, years, mean (SD)
(95 % CIs)
470 (63.8)
C4 months: N (%)
332 (45.7)
480 (66.8)
Ever-smoker: N (%)
c
b
50.0 (10.5)
34 (18.7)
28 (15.4)
46 (25.3)
99 (54.4)
135 (77.1)
40 (22.9)
56 (31.1)
N/A
50 (27.8) 74 (41.1)
N/A
N/A
(7.4, 8.2)
7.8 (2.5)
(40.6, 43.7)
42.2 (10.7)
104 (57.1)
(48.5, 51.5)
49.6 (10.9)
27 (8.8)
38 (12.3)
78 (25.3)
181 (58.8)
174 (58.0)
126 (42.0)
127 (42.6)
N/A
47 (15.8) 124 (41.6)
N/A
N/A
N/A
N/A
226 (73.4)
(48.4, 50.8)
For the Norwegian study population, Secondary degree and University degree are aggregate
Age at disease onset was known for N = 918 (99.9 %) Norwegian cases and for N = 585 (93.9 %) Italian cases
Age at time of study
331 (24.8)
219 (16.4)
358 (48.6)
157 (22.2)
Received help to complete the EnvIMS-Q: N (%)
Yes: N (%)
155 (11.9)
477 (67.4) 74 (10.5)
No: N (%)
925 (71.2)
669 (51.1)
641 (48.9)
Don’t know: N (%)
Cigarette smoking during mother’s pregnancy
239 (33.2)
Never-smoker: N (%)
724 (54.8)
N/A
N/A
Secondary degree: N (%)
University degree: N (%)c
885 (66.3)
450 (33.7)
N/A
N/A
968 (72.5)
139 (10.5) 458 (34.7)
Cigarette smoking habits
a
45.1 (10.7) (44.6, 45.7)
96 (13.2) 299 (41.1)
None/elementary degree: N (%) Junior high degree: N (%)
Education
267 (36.2)
0–3 months: N (%)
Breastfeeding
43.5 (10.1) (42.8, 44.2)
Age at time of study, years, mean (SD)a
251 (45.9)
26 (5.0)
25 (4.8)
466 (90.1)
281 (51.8)
261 (48.2)
89 (16.6)
260 (48.5)
18 (3.4) 169 (31.5)
242 (44.2)
305 (55.8)
(5.3, 5.8)
5.6 (2.7)
(31.7, 33.4)
32.6 (10.1)
359 (65.6)
(37.5, 39.2)
38.3 (10.3)
Cases (547)
323 (31.1)
68 (7.0)
52 (5.4)
851 (87.6)
420 (40.7)
611 (59.3)
230 (22.8)
456 (45.2)
32 (3.2) 291 (28.8)
533 (51.3)
506 (48.7)
N/A
N/A
724 (69.7)
(38.4, 39.7)
39.1 (10.7)
Controls (1039)
Cases (182)
Cases (737)
Controls (308)
Known info on breastfeeding
Missing info on breastfeeding
Known info on breastfeeding Controls (1335)
Italy
Norway
Table 1 Demographic and clinical characteristics of the study participants
26 (34.2)
3 (4.2)
1 (1.4)
67 (94.4)
47 (62.7)
28 (37.3)
14 (18.7)
25 (33.3)
5 (6.7) 31 (41.3)
N/A
N/A
(5.3, 6.5)
5.9 (2.7)
(33.2, 37.7)
35.4 (9.6)
55 (72.4)
(39.4, 43.9)
41.7 (9.7)
Cases (76)
25 (18.8)
5 (4.0)
5 (4.0)
114 (91.9)
70 (53.8)
60 (46.2)
18 (14.0)
49 (38.0)
8 (6.2) 54 (41.9)
N/A
N/A
N/A
N/A
102 (76.7)
(41.3, 45.0)
43.1 (10.9)
Controls (133)
Missing info on breastfeeding
J Neurol
J Neurol Table 2 Crude and adjusted odds ratios (95 %CIs) for the association between MS and duration of breastfeeding Crude OR (95 % CIs)
ORadj (95 % CIs)
C4 months
1.0
1.0
0–3 months
1.12 (0.93, 1.35)
1.14a (0.92, 1.40)
Men (N = 566)
1.80 (1.22, 2.67)
2.13b (1.37, 3.30)
Women (N = 1506)
0.97 (0.78, 1.20)
0.95b (0.75, 1.21)
C4 months 0–3 months
1.0 1.33 (1.08, 1.64)
1.0 1.37a (1.09, 1.73)
Men (N = 503)
1.84 (1.28, 2.65)
2.33b (1.50, 3.65)
Women (N = 1083)
1.14 (0.89, 1.48)
1.09b (0.82, 1.46)
Breastfeeding Norway
Italy
OR Odds ratio a
Adjusted for sex, education (none, primary, junior high, secondary school degree, university degree), cigarette smoking habits (never smoker, ever smoker), cigarette smoking during mother’s pregnancy (yes, no)
b
Adjusted for education (none, primary, junior high, secondary school degree, university degree), cigarette smoking habits (never smoker, ever smoker), cigarette smoking during mother’s pregnancy (yes, no)
breastfeeding in the first weeks of life is not the reference attitude in the general population [21, 22]. Furthermore and most importantly, a newborn’s nutrition at 4 months and older often implies nutrients other than breast milk (e.g., infant formula) which may very likely confound an association between ‘prolonged breastfeeding’ and reduced risk for MS. Whether breastfeeding was ‘exclusive’ or not, was not reported by Conradi et al. To minimize such confounding, breastfeeding for at least 4 months was used as the reference category in our study. Furthermore, breastfeeding up to 3 months can more reliably be assumed to be exclusive. This assumption is in line with evidence showing that the prevalence of breastfeeding, and especially exclusive breastfeeding, has been shown to decline after the first trimester [21, 22]. Overall, our findings are consistent with the possible role of breastfeeding in the MS biological mechanisms speculated by Conradi et al., pointing to reduced duration of breastfeeding as a potential risk factor for MS, at least in men. A protective effect of breastfeeding in the pathogenesis of different immune-mediated diseases, including MS, could be related to components with antimicrobiotic activities, such as immunoglobulins, lysozyme, fatty acids, oligosaccharides (human milk oligosaccharides, HMO), polyamines, lactoferrin, and other glycoproteins and peptides [23]. These factors may play a fundamental role in promoting the development of immune system, protection from toxic agents and pathogens, and modulating the
composition of intestinal microbiota in infants [24]. Up to 200 different types of HMOs have been discovered [25]. Besides defending from potential pathogens, HMOs are important probiotics, i.e., non-digestible substances which selectively stimulate the growth and activity of one or more colon bacterial species conferring the host benefits [26]. As prebiotics, HMOs stimulate the growth of beneficial bacteria, e.g., the Bifidobacteria, predominantly found in the stool of breastfed children, and the Bacteroides, featuring a fundamental role in the degradation of dietary fibers in adults [25, 27]. These two bacterial populations have recently been found to help prevent autoimmune reactions [28]. Besides the HMOs, proteic components of breast milk also appear to have prebiotic activity, contributing to the development of an optimal bacterial flora [29]. Protein levels in breast milk are well known to be most elevated at the beginning of breastfeeding, progressively declining after 3–4 months [29]. These findings highlight the importance of breastfeeding during the very first months of life on the development of the immune system and how its even partial deficit may contribute to determining autoimmune disorders. In the stratified analysis by country, no association between breastfeeding and MS risk was found in Norwegians. Interestingly, a higher proportion of prolonged breastfeeding (64 % among cases and 66 % among controls) was observed in this population, as compared to the Italians (44 and 51 %). This is in line with the reported breastfeeding behaviors of Scandinavians i.e., between 58 and 80 % breastfeed for at least 6-months [22]. Despite a lower proportion of cases as compared to controls exposed to prolonged breastfeeding in Norway, the association to MS did not differ significantly among Norwegian participants. One possible explanation is due to cultural factors ensuring a satisfactory breast milk intake in the whole population. After stratification by sex, a strong association was found among men but not among women for the adjusted pooled data, and separately among both Norwegians and Italians. This finding is difficult to explain. Interestingly, however, a number of studies in pediatric populations have revealed a differential impact of breastfeeding on disease outcome (e.g., respiratory tract infection) by infant sex [30–32]. Breastfeeding protection against acute respiratory infections has been hypothesized not to be conferred by passive transfer of humoral immunity or anti-infective properties of soluble molecules, which should not differ by infant sex [33]. Sex-specific differences in T-helper immune response interacting with breastfeeding may in fact affect infants’ susceptibility to infections [33]. Limitations of our study may be related to the case– control design, i.e., the study groups are selected by the presence or absence of the outcome (MS), with the amount
123
J Neurol
and frequency of exposure (breastfeeding) determined by recall, many years after exposure possibly introducing bias. 1586 (88.4 %) Italian participants and 2072 (80.9 %) Norwegian participants provided information about breastfeeding. An attempt to minimize recall bias was made by running an analysis including only participants with information from close relatives. The estimates remained essentially the same. With regard to the non-responders, slight differences were seen as compared individuals who provided information on breastfeeding (Table 1). A significant inverse association was found between the duration of breastfeeding and participants’ age at study time, for both the Norwegian and the Italian study population. However, since no difference in the mean age was observed between cases and controls in those ‘non-responders’, a cohort effect or any other factor related to age cannot explain the difference in risk. The similar patterns of association between breastfeeding and risk for MS obtained for the two populations under study by sex, minimizes the possibility that such association measures reflect consistent bias, and rather support a universal sex-specific biologic effect on the risk. Acknowledgments The study was supported by grants from The Italian MS Society/Foundation (Fondazione Italiana Sclerosi Multipla, FISM, grants n. 2007/R/14, and n. 2008/R/19 to M. Pugliatti), The Western Norway Regional Health Authority (Helse Vest) Norway (grants n. 911421/2008 to M. Pugliatti and n. 911474/2009 to K-M Myhr), The University of Bergen, Norway (2007 to T. Riise), and The Norwegian MS Society (2011 to T. Riise). The authors wish to acknowledge Kristin Wesnes and Kjetil Bjørnevik (Department of Neurology, Haukeland University Hospital, Bergen, Norway; Department of Global Public Health and Primary Care, University of Bergen, Bergen, Norway); Bin Zhu and Sandra Magalhaes (Research Institute of the McGill University Health Centre, Montreal Canada), for data cleaning.
4.
5.
6. 7. 8.
9.
10.
11.
12.
13.
14.
15.
Conflicts of interest Giammario Ragnedda, Stefania Leoni, Maria Parpinel, Ilaria Casetta, Trond Riise, Kjell-Morten Myhr, Christina Wolfson report no competing interests in relation to the study. 16. Ethical standard Maura Pugliatti is a member of the Italian MS Foundation Social, Clinical and Behavioural Scientific Committee since 2012. 17.
References 18. 1. Ramagopalan SV, Dobson R, Meier UC, Giovannoni G (2010) Multiple sclerosis: risk factors, prodromes, and potential causal pathways. Lancet Neurol 9:727–739 2. International Multiple Sclerosis Genetics Consortium, Wellcome Trust Case Control Consortium (2011) Genetic risk and a primary role for cell-mediated immune mechanisms in multiple sclerosis. Nature 476:214–219 3. International Multiple Sclerosis Genetics Consortium (IMSGC) (2013) Analysis of immune-related loci identifies 48 new
123
19. 20.
21.
susceptibility variants for multiple sclerosis. Nat Genet 45:1353–1360 Munger KL, Zhang SM, O’Reilly E, Herna´n MA, Olek MJ, Willett WC, Ascherio A (2004) Vitamin D intake and incidence of multiple sclerosis. Neurology 62:60–65 Simon KC, van der Mei IAF, Munger KL, Ponsonby A, Dickinson J, Dwyer T, Sundstro¨m P, Ascherio A (2010) Combined effects of smoking, anti-EBNA antibodies, and HLADRB1*1501 on multiple sclerosis risk. Neurology 74:1365–1371 Riise T, Nortvedt MW, Ascherio A (2003) Smoking is a risk factor for multiple sclerosis. Neurology 61:1122–1124 Bach JF (2002) The effect of infections on susceptibility to autoimmune and allergic diseases. N Engl J Med 347:911–920 Rosenbauer J, Herzig P, Kaiser P, Giani G (2007) Early nutrition and risk of type 1 diabetes mellitus. A nationwide case-control study in preschool children. Exp Clin Endocrinol Diabetes 115:502–508 Alves JG, Figueiroa JN, Meneses J, Alves GV (2012) Breastfeeding protects against type 1 diabetes mellitus: a case-sibling study. Breastfeed Med 7:25–28 Klement E, Cohen RV, Boxman J, Joseph A, Reif S (2004) Breastfeeding and risk of inflammatory bowel disease: a systematic review with meta-analysis. Am J Clin Nutr 80:1342–1352 Gdalevich M, Mimouni D, David M, Mimouni M (2001) Breastfeeding and the onset of atopic dermatitis in childhood: a systematic review and meta-analysis of prospective studies. J Am Acad Dermatol 45:520–527 Gdalevich M, Mimouni D, Mimouni M (2001) Breast-feeding and the risk of bronchial asthma in childhood: a systematic review with meta-analysis of prospective studies. J Pediatr 139:261–266 Conradi S, Malzahn U, Paul F, Quill S, Harms L, Then Bergh F, Ditzenbach A, Georgi T, Heuschmann P, Rosche B (2013) Breastfeeding is associated with lower risk for multiple sclerosis. Mult Scler 19:553–558 Pugliatti M, Casetta I, Drulovic J, Granieri E, Holmøy T, Kampman MT, Landtblom AM, Lauer K, Myhr KM, Parpinel M, Pekmezovic T, Riise T, Zhu B, Wolfson C (2012) A questionnaire for multinational case-control studies of environmental risk factors in multiple sclerosis (EnvIMS-Q). Acta Neurol Scand Suppl 95:43–50 Lossius A, Riise T, Pugliatti M, Bjørnevik K, Casetta I, Drulovic J, Granieri E, Kampman MT, Landtblom AM, Lauer K, Magalhaes S, Myhr KM, Pekmezovic T, Wesnes K, Wolfson C, Holmøy T (2014) Season of infectious mononucleosis and risk of multiple sclerosis at different latitudes; the EnvIMS Study. Mult Scler 20:669–674 Bjørnevik K, Riise T, Casetta I, Drulovic J, Granieri E, Holmøy Kampman MT, Landtblom AM, Lauer K, Lossius A, Magalhaes S, Myhr KM, Pekmezovic T, Wesnes K, Wolfson C, Pugliatti M (2014) Sun exposure and multiple sclerosis risk in Norway and Italy: the EnvIMS study. Mult Scler 20:1042–1049 Polman CH, Reingold SC, Edan G (2005) Diagnostic criteria for multiple sclerosis: 2005 revisions to the ‘‘McDonald Criteria’’. Ann Neurol 58:840–846 Lundy E (2012) The Effect of Assigning Different Index Dates for Control Exposure Measurement on Odds Ratio Estimates in Department of Mathematics and Statistics, McGill University Spencely M, Dick G (1982) Breast-feeding and multiple sclerosis. Neuroepidemiology 1:216–222 Pisacane A, Impagliazzo N, Russo M, Valiani R, Mandarini A, Florio C, Vivo P (1994) Breast-feeding and multiple sclerosis. BMJ 308:1411–1412 Santorelli G, Petherick E, Waiblinger D, Cabieses Fairley L (2013) Ethnic differences in the initiation and duration of breast
J Neurol
22.
23.
24.
25.
26.
feeding—results from the Born in Bradford Birth Cohort Study. Paediatr Perinat Epidemiol 27:388–392 Ho¨rnell A, Lagstro¨m H, Lande B, Thorsdottir I (2013) Breastfeeding, introduction of other foods and effects on health: a systematic literature review for the 5th Nordic Nutrition Recommendations. Food Nutr Res 57:20823 Fernandez L, Langa S, Martı´n V, Maldonado A, Jime´nez E, Martı´n R, Rodrı´guez JM (2013) The human milk microbiota: origin and potential roles in health and diseases. Pharmacol Res 69:1–10 Jeong K, Nguyen V, Kim J (2012) Human milk oligosaccharides: the novel modulator of intestinal microbiota. BMB Report 45:433–441 Marcobal A, Barboza M, Froehlich JW, Block DE, German JB, Lebrilla CB, Mills DA (2010) Consumption of human milk oligosaccharides by gut-related microbes. J Agric Food Chem 58:5334–5340 Sherman PM, Cabana M, Gibson GR, Koletzko BV, Neu J, Veereman-Wauters G, Ziegler EE, Walker WA (2009) Potential roles and clinical utility of prebiotics in newborns infant and children: proceedings from a global prebiotic summit meeting, New York City, June 27–28, 2008. J Pedriatr 155:S61–S70
27. Marcobal A, Sonneburg JL (2012) Human milk oligosaccharide consumption by intestinal microbiota. Clin Microbiol Infect 18(suppl 4):12–15 28. Atarashi K, Honda K (2011) Microbiota in autoimmunity and tolerance. Curr Opin Immunol 23:761–768 29. Lo¨nnerdal B (2003) Nutritional and physiologic significance of human milk proteins. Am J Clin Nutr 77(suppl):1537S–1543S 30. Wright AL, Holberg CJ, Martinez FD, Morgan WJ, Taussig LM (1989) Breast feeding and lower respiratory tract illness in the first year of life. Group Health Medical Associates. BMJ 299:946–949 31. Sinha A, Madden J, Ross-Degnan D, Soumerai S, Platt R (2003) Reduced risk of neonatal respiratory infections among breastfed girls but not boys. Pediatrics 112:303–307 32. Libster R, Bugna Hortoneda J, Laham FR, Casellas JM, Israele V, Polack NR, Delgado MF, Klein MI, Polack FP (2009) Breastfeeding prevents severe disease in full term female infants with acute respiratory infection. Pediatr Infect Dis J 28:131–134 33. Klein MI, Bergel E, Gibbons L, Coviello S, Bauer G, Benitez A, Serra ME, Delgado MF, Melendi GA, Rodrı´guez S, Kleeberger SR, Polack FP (2008) Differential gender response to respiratory infections and to the protective effect of breast milk in preterm infants. Pediatrics 121:1510–1516
123
ORIGINAL COMMUNICATION
Reduced duration of breastfeeding is associated with a higher risk of multiple sclerosis in both Italian and Norwegian adult males: the EnvIMS study Giammario Ragnedda1,2 • Stefania Leoni1,2 • Maria Parpinel3 • Ilaria Casetta4 • Trond Riise5 • Kjell-Morten Myhr6,7 • Christina Wolfson8,9 • Maura Pugliatti1,2,5,10
Received: 2 December 2014 / Revised: 25 February 2015 / Accepted: 7 March 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Breastfeeding for at least 4 months has been found to be associated with a reduced risk of immunemediated diseases including multiple sclerosis (MS). Using data from a large multinational case–control study (EnvIMS), the association between MS and breastfeeding was investigated in two distinct populations. A questionnaire (EnvIMS-Q) which included a section on feeding during the first year of life was administered to MS cases and to age and sex frequency-matched controls from Italy and Norway. Logistic regression was used to estimate the odds ratio (ORs) and 95 % confidence intervals (95 % CIs) as a measure of the association between MS and exposure to prolonged breastfeeding (4 months or more, used as the reference category), vs. no breastfeeding or breastfeeding for less than 4 months (reduced exposure). Education, smoking habits, smoking in mother’s pregnancy, and other types of milk used in infant feeding were included as covariates. A total of 547 cases and 1039 controls in Italy, and 737 cases and 1335 controls in Norway were studied. The distribution of prolonged (reference) breastfeeding
differed between the Norwegian (65.4 %) and the Italian (48.9 %) study participants. A significant association between MS and reduced/no exposure to breastfeeding was found overall for Italy (ORadj = 1.37; 95 % CI 1.09, 1.73), but not for Norway (ORadj = 1.14; 95 % CI 0.92, 1.40). However, only in men, significant associations were observed for both populations (ORItaly = 2.33; 95 % CI 1.50, 3.65, ORNorway = 2.13; 95 % CI 1.37, 3.30). Reduced exposure to breastfeeding in males was found to be associated with increased risk of MS in Italy and in Norway.
& Maura Pugliatti [email protected]
6
The Norwegian Multiple Sclerosis Registry and Biobank, Department of Neurology, Haukeland University Hospital, Bergen, Norway
7
Department of Clinical Medicine, KG Jebsen Centre for MS-Research, University of Bergen, Norway Haukeland University Hospital, Bergen, Norway
8
Research Institute of the McGill University Health Centre, Montreal, QC, Canada
9
Department of Epidemiology and Biostatistics and Occupational Health, McGill University, Montreal, QC, Canada
10
Department of Medicine, Faculty of Medicine, McGill University, Montreal, Canada
1
2
3
4
5
Department of Biomedical Sciences, University of Sassari, Sassari, Italy Department of Clinical and Experimental Medicine, University of Sassari, Viale San Pietro 10, 07100 Sassari, Italy Unit of Hygiene and Epidemiology, Department of Medical and Biological Sciences, University of Udine, Udine, Italy Department of Biomedical and Specialty Surgical Sciences, Section of Clinical Neurology, University of Ferrara, Ferrara, Italy
Keywords Multiple sclerosis Breastfeeding Case–control studies Risk factors Epidemiology
Introduction Multiple sclerosis (MS) is the most common inflammatory disease of the central nervous system affecting young adults [1]. Despite the growing acceptance of a genetic
Department of Global Public Health and Primary Care, University of Bergen, Bergen, Norway
123
J Neurol
susceptibility for MS, the increasing incidence observed during the last 50 years in Western countries cannot be explained by changes in genetic risk factors, emphasizing the importance of environmental factors related to lifestyle [2–6]. Similar incidence trends have been observed for other immune-mediated diseases such as type-1 diabetes (T1D), inflammatory bowel diseases (IBD) and asthma [7]. Breastfeeding has previously been reported to influence the risk of T1D, and recently, patients developing diabetes have been shown to have a shorter average duration of breastfeeding (3.3 vs. 4.6 months) and to have been exposed to cow’s milk earlier, compared to unaffected siblings [8, 9]. Breastfeeding has also been shown a protective role for other autoimmune diseases like IBD, atopic dermatitis, and asthma [10–12]. Breastfeeding for at least 4 months has also recently been reported to protect against the risk of developing MS, but the impact of shorter duration/no breastfeeding in determining the risk for MS is still debated [13]. We therefore investigated the role of breastfeeding in determining MS risk by examining both country specific and pooled associations using data from a large multinational case–control study (the EnvIMS Study).
Methods Study design The study design and methodology have been reported elsewhere [14]. Briefly, this study builds upon an international multicentre case–control study of environmental risk factors in MS (the EnvIMS Study) which aims to examine possible associations between MS status and past exposures such as infections, sunlight and lifestyle factors, by means of a shared standard methodology in populations differing in disease risk and possibly in the distribution of some environmental factors [15, 16]. A postal self-administered questionnaire (the EnvIMS-Q) was designed in English, translated into the language of the participating countries and tested for reliability, crosscultural validity and perceived difficulty of its use, to record self-reported age-specific exposures in individuals with MS as well as in non-MS controls [14]. EnvIMS-Q is divided into sections representing domains of the most interesting exposures in MS etiological research: sun exposure, diet, intake of vitamin supplements, medical history focusing on childhood infections and immunemediated comorbidity, cigarette smoking habits, physical activity, occupational factors, and reproductive factors for female participants. The EnvIMS study received ethics approval at each study site [14].
123
Study population The Italian and Norwegian EnvIMS population samples have been included for this study. The questionnaire was sent to cases and controls aged 18 years or older and return of the EnvIMS-Q questionnaire was considered evidence of informed consent. All cases diagnosed according to the McDonald criteria and with a clinical onset within 10 years prior to data collection were selected from populationbased MS databases [14, 17]. Four times as many age and sex frequency-matched controls were randomly selected from general population health administrative databases of each region under study. EnvIMS-Q was sent to 1368 MS patients and 4728 controls in Norway, and 1692 cases and 6414 controls in Italy. The response rates were 69.7 and 36.3 % among cases and controls in Norway, and 42.8 and 20.9 %, respectively, in Italy. Exposure The primary exposure of interest in this study was ‘feeding in the first year of life’ and was assessed with one specific question in the Diet section in both the Norwegian and the Italian EnvIMS-Q [14]. In Norway, participants were asked ‘‘Were you breastfed?’’, with the following response options: ‘‘No’’, ‘‘Don’t know’’ or ‘‘Yes’’. If the response was ‘‘Yes’’, then they were asked ‘‘For how many months?’’ with the following response options: ‘‘1–3 months’’, ‘‘4–6 months’’, ‘‘7–9 months’’, ‘‘for 10 months or longer’’. In Italy, participants were able to provide detail, i.e., whether they received only ‘‘Breastfeeding’’, ‘‘Artificial milk (powdered milk)’’, ‘‘Other sugary liquids (chamomile, herb infusions, etc.)’’, ‘‘Other milks (goat, sheep, cow, donkey)’’, or in combination. For each of these options, participants were asked to provide information on the duration (number of months). For each of the indicated durations (1–3, 4–6, 7–9, and 10? months), they could choose the ‘‘I don’t know’’ option. Despite differences in the question design between the Norwegian and Italian questionnaires, it was possible to harmonize the responses into two risk categories for both populations. Based on existing literature as well as previous work on MS, we defined the following three categories: (1) ‘‘no breastfeeding or breastfeeding up to 3 months’’ (i.e., no/reduced breastfeeding), (2) ‘‘unknown’’ including those who ticked off ‘‘don’t know’’ (n = 428) and those who did not respond to one or more of the breastfeeding items (n = 720) and (3) ‘‘breastfeeding for 4 months and longer’’ (i.e., prolonged) (reference) [8, 9, 11–13]. In any of these periods, breastfeeding could have occurred exclusively or in combination with intake of other types of milk or liquids. Since it was not possible to distinguish exclusive from non-exclusive
J Neurol
breastfeeding in Norway, breastfeeding was treated as ‘non-exclusive’ for Italy as well. Covariates Other covariates were included in the analyses as possible confounders: Sex, participant education (none, primary, junior high, secondary school degree, and university degree), smoking habits (never-smoker, ever-smoker), smoking in mother’s pregnancy (yes, don’t know/missing, no), other types of milk (artificial and animal, only for the Italian sub-analysis). Whether or not the participants received help from parents in completing the questionnaires was also recorded. Statistical analysis Frequencies (percent) were reported for categorical variables and mean [standard deviation (SD)] for continuous variables. We used logistic regression to estimate odds ratios (ORs) and 95 % confidence intervals (95 % CIs) as the measure of the association between MS and past exposure to prolonged breastfeeding. In addition to breastfeeding, the model also included sex, participants’ education, smoking habits and smoking in mother’s pregnancy as potential confounders. For the purposes of matching cases and controls on ‘age at onset’, and to ensure that the exposure opportunities for cases and controls were frequency-matched (e.g., participant’s education, smoking habit, help from parents in filling out the questionnaire), each control was assigned an index age corresponding to the age at onset of a case frequencymatched by year of birth [18]. To test whether there was effect modification due to country (Italy vs. Norway) and sex, we included the appropriate interaction terms in the model. The Statistical Package for the Social Sciences (SPSS) version 19 for Windows and OSX (SPSS Inc., IBM, Somers, New York, USA) was used for statistical analysis.
Results A total of 1586 (88.4 %) Italian participants and 2072 (80.9 %) Norwegian participants provided information about breastfeeding. Individuals who did not provide any kind of information regarding the question about feeding in the first year of life, and those who ticked only ‘I don’t know’ options were excluded from the analysis. The demographic and main characteristics of the study population and of the individuals who did not provide information about breastfeeding (‘non-responders’) are given in Table 1.
In the Norwegian study population, 127 (14.5 %) cases and 202 (13.0 %) controls, respectively, ticked the ‘I don’t know’ option; this option was not conceived for the 3-month periods of breastfeeding in the Norwegian EnvIMS-Q. In the Italian study population, the ‘I don’t know’ option was ticked in 35 (46.1 %) and 56 (37.6 %) cases and controls, respectively, within the stratum of no/ reduced breastfeeding; in 41 (53.9 %) and 93 (62.4 %) cases and controls, respectively, within the stratum of prolonged breastfeeding. A higher proportion of participants, breastfed for 4 months and longer in the overall EnvIMS Norwegian sample (65.4 %) as compared to the Italian sample (48.9 %). In the pooled data, 55.5 % of cases reported prolonged breastfeeding as compared to 59.7 % of controls. When stratifying for area (Norway vs. Italy), no association was found between MS status and no/reduced breastfeeding in Norway (OR = 1.13; 95 %CI 0.91, 1.40), whereas a statistically significant association was found for Italy (OR = 1.37; 95 %CI 1.09, 1.73). However, the difference in estimate was not statistically significant between the two countries (p = 0.64). In the Italian dataset, these measures of association remained significant after adjustment for the addition of artificial and animal milk (data not shown). After stratification by sex, a strong association was found in both Norwegian and Italian men, but not in women (Table 2). This difference between men and women was statistically significant (p = 0.0003). The results did not change even after including those who ticked ‘I don’t know’ option in the analysis, nor when adjusting for sun exposure at age 0–5 years nor obesity (data not shown).
Discussion Breastfeeding has been reported to influence the risk of developing immune-mediated diseases, but to date, it has been inadequately addressed in MS [8–13, 19, 20]. In this study, we have shown that not being breastfed or being breastfed for 3 months or less, might contribute to an increasing risk for adult MS and that this association may be sex- and population-specific. This association has previously been investigated, yielding conflicting results [19, 20]. Recently, Conradi et al. [13] reported on the possible protective role of prolonged breastfeeding (defined as 4 months or more) and the risk for MS. In their case–control study, conducted on 245 MS patients and 296 controls, ‘no breastfeeding’ was used as the reference category. We question the choice of the reference value, since the complete absence of exposure to
123
123 538 (73.0) 36.4 (9.7) (35.7, 37.1) 7.1 (2.7) (6.9, 7.3)
(95 % CIs)
Sex, females (%)
Age at disease onset, years, mean (SD)b
(95 % CIs)
Disease duration, years, mean (SD)
(95 % CIs)
470 (63.8)
C4 months: N (%)
332 (45.7)
480 (66.8)
Ever-smoker: N (%)
c
b
50.0 (10.5)
34 (18.7)
28 (15.4)
46 (25.3)
99 (54.4)
135 (77.1)
40 (22.9)
56 (31.1)
N/A
50 (27.8) 74 (41.1)
N/A
N/A
(7.4, 8.2)
7.8 (2.5)
(40.6, 43.7)
42.2 (10.7)
104 (57.1)
(48.5, 51.5)
49.6 (10.9)
27 (8.8)
38 (12.3)
78 (25.3)
181 (58.8)
174 (58.0)
126 (42.0)
127 (42.6)
N/A
47 (15.8) 124 (41.6)
N/A
N/A
N/A
N/A
226 (73.4)
(48.4, 50.8)
For the Norwegian study population, Secondary degree and University degree are aggregate
Age at disease onset was known for N = 918 (99.9 %) Norwegian cases and for N = 585 (93.9 %) Italian cases
Age at time of study
331 (24.8)
219 (16.4)
358 (48.6)
157 (22.2)
Received help to complete the EnvIMS-Q: N (%)
Yes: N (%)
155 (11.9)
477 (67.4) 74 (10.5)
No: N (%)
925 (71.2)
669 (51.1)
641 (48.9)
Don’t know: N (%)
Cigarette smoking during mother’s pregnancy
239 (33.2)
Never-smoker: N (%)
724 (54.8)
N/A
N/A
Secondary degree: N (%)
University degree: N (%)c
885 (66.3)
450 (33.7)
N/A
N/A
968 (72.5)
139 (10.5) 458 (34.7)
Cigarette smoking habits
a
45.1 (10.7) (44.6, 45.7)
96 (13.2) 299 (41.1)
None/elementary degree: N (%) Junior high degree: N (%)
Education
267 (36.2)
0–3 months: N (%)
Breastfeeding
43.5 (10.1) (42.8, 44.2)
Age at time of study, years, mean (SD)a
251 (45.9)
26 (5.0)
25 (4.8)
466 (90.1)
281 (51.8)
261 (48.2)
89 (16.6)
260 (48.5)
18 (3.4) 169 (31.5)
242 (44.2)
305 (55.8)
(5.3, 5.8)
5.6 (2.7)
(31.7, 33.4)
32.6 (10.1)
359 (65.6)
(37.5, 39.2)
38.3 (10.3)
Cases (547)
323 (31.1)
68 (7.0)
52 (5.4)
851 (87.6)
420 (40.7)
611 (59.3)
230 (22.8)
456 (45.2)
32 (3.2) 291 (28.8)
533 (51.3)
506 (48.7)
N/A
N/A
724 (69.7)
(38.4, 39.7)
39.1 (10.7)
Controls (1039)
Cases (182)
Cases (737)
Controls (308)
Known info on breastfeeding
Missing info on breastfeeding
Known info on breastfeeding Controls (1335)
Italy
Norway
Table 1 Demographic and clinical characteristics of the study participants
26 (34.2)
3 (4.2)
1 (1.4)
67 (94.4)
47 (62.7)
28 (37.3)
14 (18.7)
25 (33.3)
5 (6.7) 31 (41.3)
N/A
N/A
(5.3, 6.5)
5.9 (2.7)
(33.2, 37.7)
35.4 (9.6)
55 (72.4)
(39.4, 43.9)
41.7 (9.7)
Cases (76)
25 (18.8)
5 (4.0)
5 (4.0)
114 (91.9)
70 (53.8)
60 (46.2)
18 (14.0)
49 (38.0)
8 (6.2) 54 (41.9)
N/A
N/A
N/A
N/A
102 (76.7)
(41.3, 45.0)
43.1 (10.9)
Controls (133)
Missing info on breastfeeding
J Neurol
J Neurol Table 2 Crude and adjusted odds ratios (95 %CIs) for the association between MS and duration of breastfeeding Crude OR (95 % CIs)
ORadj (95 % CIs)
C4 months
1.0
1.0
0–3 months
1.12 (0.93, 1.35)
1.14a (0.92, 1.40)
Men (N = 566)
1.80 (1.22, 2.67)
2.13b (1.37, 3.30)
Women (N = 1506)
0.97 (0.78, 1.20)
0.95b (0.75, 1.21)
C4 months 0–3 months
1.0 1.33 (1.08, 1.64)
1.0 1.37a (1.09, 1.73)
Men (N = 503)
1.84 (1.28, 2.65)
2.33b (1.50, 3.65)
Women (N = 1083)
1.14 (0.89, 1.48)
1.09b (0.82, 1.46)
Breastfeeding Norway
Italy
OR Odds ratio a
Adjusted for sex, education (none, primary, junior high, secondary school degree, university degree), cigarette smoking habits (never smoker, ever smoker), cigarette smoking during mother’s pregnancy (yes, no)
b
Adjusted for education (none, primary, junior high, secondary school degree, university degree), cigarette smoking habits (never smoker, ever smoker), cigarette smoking during mother’s pregnancy (yes, no)
breastfeeding in the first weeks of life is not the reference attitude in the general population [21, 22]. Furthermore and most importantly, a newborn’s nutrition at 4 months and older often implies nutrients other than breast milk (e.g., infant formula) which may very likely confound an association between ‘prolonged breastfeeding’ and reduced risk for MS. Whether breastfeeding was ‘exclusive’ or not, was not reported by Conradi et al. To minimize such confounding, breastfeeding for at least 4 months was used as the reference category in our study. Furthermore, breastfeeding up to 3 months can more reliably be assumed to be exclusive. This assumption is in line with evidence showing that the prevalence of breastfeeding, and especially exclusive breastfeeding, has been shown to decline after the first trimester [21, 22]. Overall, our findings are consistent with the possible role of breastfeeding in the MS biological mechanisms speculated by Conradi et al., pointing to reduced duration of breastfeeding as a potential risk factor for MS, at least in men. A protective effect of breastfeeding in the pathogenesis of different immune-mediated diseases, including MS, could be related to components with antimicrobiotic activities, such as immunoglobulins, lysozyme, fatty acids, oligosaccharides (human milk oligosaccharides, HMO), polyamines, lactoferrin, and other glycoproteins and peptides [23]. These factors may play a fundamental role in promoting the development of immune system, protection from toxic agents and pathogens, and modulating the
composition of intestinal microbiota in infants [24]. Up to 200 different types of HMOs have been discovered [25]. Besides defending from potential pathogens, HMOs are important probiotics, i.e., non-digestible substances which selectively stimulate the growth and activity of one or more colon bacterial species conferring the host benefits [26]. As prebiotics, HMOs stimulate the growth of beneficial bacteria, e.g., the Bifidobacteria, predominantly found in the stool of breastfed children, and the Bacteroides, featuring a fundamental role in the degradation of dietary fibers in adults [25, 27]. These two bacterial populations have recently been found to help prevent autoimmune reactions [28]. Besides the HMOs, proteic components of breast milk also appear to have prebiotic activity, contributing to the development of an optimal bacterial flora [29]. Protein levels in breast milk are well known to be most elevated at the beginning of breastfeeding, progressively declining after 3–4 months [29]. These findings highlight the importance of breastfeeding during the very first months of life on the development of the immune system and how its even partial deficit may contribute to determining autoimmune disorders. In the stratified analysis by country, no association between breastfeeding and MS risk was found in Norwegians. Interestingly, a higher proportion of prolonged breastfeeding (64 % among cases and 66 % among controls) was observed in this population, as compared to the Italians (44 and 51 %). This is in line with the reported breastfeeding behaviors of Scandinavians i.e., between 58 and 80 % breastfeed for at least 6-months [22]. Despite a lower proportion of cases as compared to controls exposed to prolonged breastfeeding in Norway, the association to MS did not differ significantly among Norwegian participants. One possible explanation is due to cultural factors ensuring a satisfactory breast milk intake in the whole population. After stratification by sex, a strong association was found among men but not among women for the adjusted pooled data, and separately among both Norwegians and Italians. This finding is difficult to explain. Interestingly, however, a number of studies in pediatric populations have revealed a differential impact of breastfeeding on disease outcome (e.g., respiratory tract infection) by infant sex [30–32]. Breastfeeding protection against acute respiratory infections has been hypothesized not to be conferred by passive transfer of humoral immunity or anti-infective properties of soluble molecules, which should not differ by infant sex [33]. Sex-specific differences in T-helper immune response interacting with breastfeeding may in fact affect infants’ susceptibility to infections [33]. Limitations of our study may be related to the case– control design, i.e., the study groups are selected by the presence or absence of the outcome (MS), with the amount
123
J Neurol
and frequency of exposure (breastfeeding) determined by recall, many years after exposure possibly introducing bias. 1586 (88.4 %) Italian participants and 2072 (80.9 %) Norwegian participants provided information about breastfeeding. An attempt to minimize recall bias was made by running an analysis including only participants with information from close relatives. The estimates remained essentially the same. With regard to the non-responders, slight differences were seen as compared individuals who provided information on breastfeeding (Table 1). A significant inverse association was found between the duration of breastfeeding and participants’ age at study time, for both the Norwegian and the Italian study population. However, since no difference in the mean age was observed between cases and controls in those ‘non-responders’, a cohort effect or any other factor related to age cannot explain the difference in risk. The similar patterns of association between breastfeeding and risk for MS obtained for the two populations under study by sex, minimizes the possibility that such association measures reflect consistent bias, and rather support a universal sex-specific biologic effect on the risk. Acknowledgments The study was supported by grants from The Italian MS Society/Foundation (Fondazione Italiana Sclerosi Multipla, FISM, grants n. 2007/R/14, and n. 2008/R/19 to M. Pugliatti), The Western Norway Regional Health Authority (Helse Vest) Norway (grants n. 911421/2008 to M. Pugliatti and n. 911474/2009 to K-M Myhr), The University of Bergen, Norway (2007 to T. Riise), and The Norwegian MS Society (2011 to T. Riise). The authors wish to acknowledge Kristin Wesnes and Kjetil Bjørnevik (Department of Neurology, Haukeland University Hospital, Bergen, Norway; Department of Global Public Health and Primary Care, University of Bergen, Bergen, Norway); Bin Zhu and Sandra Magalhaes (Research Institute of the McGill University Health Centre, Montreal Canada), for data cleaning.
4.
5.
6. 7. 8.
9.
10.
11.
12.
13.
14.
15.
Conflicts of interest Giammario Ragnedda, Stefania Leoni, Maria Parpinel, Ilaria Casetta, Trond Riise, Kjell-Morten Myhr, Christina Wolfson report no competing interests in relation to the study. 16. Ethical standard Maura Pugliatti is a member of the Italian MS Foundation Social, Clinical and Behavioural Scientific Committee since 2012. 17.
References 18. 1. Ramagopalan SV, Dobson R, Meier UC, Giovannoni G (2010) Multiple sclerosis: risk factors, prodromes, and potential causal pathways. Lancet Neurol 9:727–739 2. International Multiple Sclerosis Genetics Consortium, Wellcome Trust Case Control Consortium (2011) Genetic risk and a primary role for cell-mediated immune mechanisms in multiple sclerosis. Nature 476:214–219 3. International Multiple Sclerosis Genetics Consortium (IMSGC) (2013) Analysis of immune-related loci identifies 48 new
123
19. 20.
21.
susceptibility variants for multiple sclerosis. Nat Genet 45:1353–1360 Munger KL, Zhang SM, O’Reilly E, Herna´n MA, Olek MJ, Willett WC, Ascherio A (2004) Vitamin D intake and incidence of multiple sclerosis. Neurology 62:60–65 Simon KC, van der Mei IAF, Munger KL, Ponsonby A, Dickinson J, Dwyer T, Sundstro¨m P, Ascherio A (2010) Combined effects of smoking, anti-EBNA antibodies, and HLADRB1*1501 on multiple sclerosis risk. Neurology 74:1365–1371 Riise T, Nortvedt MW, Ascherio A (2003) Smoking is a risk factor for multiple sclerosis. Neurology 61:1122–1124 Bach JF (2002) The effect of infections on susceptibility to autoimmune and allergic diseases. N Engl J Med 347:911–920 Rosenbauer J, Herzig P, Kaiser P, Giani G (2007) Early nutrition and risk of type 1 diabetes mellitus. A nationwide case-control study in preschool children. Exp Clin Endocrinol Diabetes 115:502–508 Alves JG, Figueiroa JN, Meneses J, Alves GV (2012) Breastfeeding protects against type 1 diabetes mellitus: a case-sibling study. Breastfeed Med 7:25–28 Klement E, Cohen RV, Boxman J, Joseph A, Reif S (2004) Breastfeeding and risk of inflammatory bowel disease: a systematic review with meta-analysis. Am J Clin Nutr 80:1342–1352 Gdalevich M, Mimouni D, David M, Mimouni M (2001) Breastfeeding and the onset of atopic dermatitis in childhood: a systematic review and meta-analysis of prospective studies. J Am Acad Dermatol 45:520–527 Gdalevich M, Mimouni D, Mimouni M (2001) Breast-feeding and the risk of bronchial asthma in childhood: a systematic review with meta-analysis of prospective studies. J Pediatr 139:261–266 Conradi S, Malzahn U, Paul F, Quill S, Harms L, Then Bergh F, Ditzenbach A, Georgi T, Heuschmann P, Rosche B (2013) Breastfeeding is associated with lower risk for multiple sclerosis. Mult Scler 19:553–558 Pugliatti M, Casetta I, Drulovic J, Granieri E, Holmøy T, Kampman MT, Landtblom AM, Lauer K, Myhr KM, Parpinel M, Pekmezovic T, Riise T, Zhu B, Wolfson C (2012) A questionnaire for multinational case-control studies of environmental risk factors in multiple sclerosis (EnvIMS-Q). Acta Neurol Scand Suppl 95:43–50 Lossius A, Riise T, Pugliatti M, Bjørnevik K, Casetta I, Drulovic J, Granieri E, Kampman MT, Landtblom AM, Lauer K, Magalhaes S, Myhr KM, Pekmezovic T, Wesnes K, Wolfson C, Holmøy T (2014) Season of infectious mononucleosis and risk of multiple sclerosis at different latitudes; the EnvIMS Study. Mult Scler 20:669–674 Bjørnevik K, Riise T, Casetta I, Drulovic J, Granieri E, Holmøy Kampman MT, Landtblom AM, Lauer K, Lossius A, Magalhaes S, Myhr KM, Pekmezovic T, Wesnes K, Wolfson C, Pugliatti M (2014) Sun exposure and multiple sclerosis risk in Norway and Italy: the EnvIMS study. Mult Scler 20:1042–1049 Polman CH, Reingold SC, Edan G (2005) Diagnostic criteria for multiple sclerosis: 2005 revisions to the ‘‘McDonald Criteria’’. Ann Neurol 58:840–846 Lundy E (2012) The Effect of Assigning Different Index Dates for Control Exposure Measurement on Odds Ratio Estimates in Department of Mathematics and Statistics, McGill University Spencely M, Dick G (1982) Breast-feeding and multiple sclerosis. Neuroepidemiology 1:216–222 Pisacane A, Impagliazzo N, Russo M, Valiani R, Mandarini A, Florio C, Vivo P (1994) Breast-feeding and multiple sclerosis. BMJ 308:1411–1412 Santorelli G, Petherick E, Waiblinger D, Cabieses Fairley L (2013) Ethnic differences in the initiation and duration of breast
J Neurol
22.
23.
24.
25.
26.
feeding—results from the Born in Bradford Birth Cohort Study. Paediatr Perinat Epidemiol 27:388–392 Ho¨rnell A, Lagstro¨m H, Lande B, Thorsdottir I (2013) Breastfeeding, introduction of other foods and effects on health: a systematic literature review for the 5th Nordic Nutrition Recommendations. Food Nutr Res 57:20823 Fernandez L, Langa S, Martı´n V, Maldonado A, Jime´nez E, Martı´n R, Rodrı´guez JM (2013) The human milk microbiota: origin and potential roles in health and diseases. Pharmacol Res 69:1–10 Jeong K, Nguyen V, Kim J (2012) Human milk oligosaccharides: the novel modulator of intestinal microbiota. BMB Report 45:433–441 Marcobal A, Barboza M, Froehlich JW, Block DE, German JB, Lebrilla CB, Mills DA (2010) Consumption of human milk oligosaccharides by gut-related microbes. J Agric Food Chem 58:5334–5340 Sherman PM, Cabana M, Gibson GR, Koletzko BV, Neu J, Veereman-Wauters G, Ziegler EE, Walker WA (2009) Potential roles and clinical utility of prebiotics in newborns infant and children: proceedings from a global prebiotic summit meeting, New York City, June 27–28, 2008. J Pedriatr 155:S61–S70
27. Marcobal A, Sonneburg JL (2012) Human milk oligosaccharide consumption by intestinal microbiota. Clin Microbiol Infect 18(suppl 4):12–15 28. Atarashi K, Honda K (2011) Microbiota in autoimmunity and tolerance. Curr Opin Immunol 23:761–768 29. Lo¨nnerdal B (2003) Nutritional and physiologic significance of human milk proteins. Am J Clin Nutr 77(suppl):1537S–1543S 30. Wright AL, Holberg CJ, Martinez FD, Morgan WJ, Taussig LM (1989) Breast feeding and lower respiratory tract illness in the first year of life. Group Health Medical Associates. BMJ 299:946–949 31. Sinha A, Madden J, Ross-Degnan D, Soumerai S, Platt R (2003) Reduced risk of neonatal respiratory infections among breastfed girls but not boys. Pediatrics 112:303–307 32. Libster R, Bugna Hortoneda J, Laham FR, Casellas JM, Israele V, Polack NR, Delgado MF, Klein MI, Polack FP (2009) Breastfeeding prevents severe disease in full term female infants with acute respiratory infection. Pediatr Infect Dis J 28:131–134 33. Klein MI, Bergel E, Gibbons L, Coviello S, Bauer G, Benitez A, Serra ME, Delgado MF, Melendi GA, Rodrı´guez S, Kleeberger SR, Polack FP (2008) Differential gender response to respiratory infections and to the protective effect of breast milk in preterm infants. Pediatrics 121:1510–1516
123
