Influence of duration of breastfeeding on manifestation of migraine syndrome and on age at which symptoms occur was examined by analyzes of anamnesis data about nutrition of 24011 children age 3 to 16, residing in Vojvodina (northern province of Serbia). The conclusions were induced by comparing data for children without recurrent headache, for children with recurrent non-migraine headache (18,83%) and with migraine headache (8,63%). Children without headache were breastfed on an average 4 months and 27 days, and children with non-migraine headache 5 months and 3.9 days. Children with migraine headache had the shortest period of breastfeeding – on an average 4 months and 10.8 days. Children with non-aura migraine were significantly shorter breastfed (3 months and 27 days) than children with other migraine syndromes (4 months 5.1 days), particularly than children with aura migraine (5 months 18.9 days).
Predisponing factors of migraine syndrome are risk factors for evincing the migraine syndrome, but they don’t cause it directly and don’t have clear causal relationship with migraine attack itself (1, 2, 3, 4). Along with basal genetically and constitutionally determined characteristics of children, factors of perinatal and psychomotor development, nutrition, social and economic milieu, environmental factors were observed and their connection with migraine syndrome (5, 6, 7). Characteristics of children which would be cognizable in the period before migraine syndrome evinces would enable prevention of migraine syndrome by adequate ultra early prevention. Breastfeeding can bee important (8, 9, 10).
Duration of breastfeeding has, to this day not described and not known effect on onset of migraine syndromes. Objective was to determine influence of breastfeeding duration on migraine syndrome in children, various types of migraine syndrome and age when first symptoms occur.
This research was carried out in the territory of Vojvodina, Serbia’s northern province, which has a population of 2,031,992, according to the last census (in 2002). During the study, which lasted from 1988 to 2004, each year 2000 questionnaires were given to the participants, who were drawn from 23 preschools and 42 grade schools in 9 cities in Vojvodina (Novi Sad, Subotica, Kikinda, Zrenjanin, Vrsac, Bela Crkva, Melenci, Futog and Temerin). Children were selected according to their month and year of birth, and the first 3 letters of their first name and surname by a multistage, stratified, clustered sampling procedure. This ensured that children could not enter the study twice during the long research period. In total, 30,636 children aged 3–17 years were surveyed (15,202 girls and 15,434 boys).
The subjects and/or their parents were asked to fill out a questionnaire in their places of residence. Questionnaires were distributed to children and/or their parents, selected by random sampling. The semistructured questionnaire, which was specially developed for this study by the author, was designed according to the International Headache Society criteria. It was a screening questionnaire, which was completed by children aged 15–17 years and by parents of younger subjects. It included 3 sections: (1) items about the child’s sociodemographic characteristics and his/her family and school; (2) items about the child’s nursing and development and (3) items about headaches.
The questionnaire was developed in 3 phases. In the first, semistructured interviews with pediatricians, researchers and nurses were organized to select relevant domains. The domains for the section about headaches were selected based on the International Classification of Headache Disorders – II criteria. More than 150 possible items were identified. Precise, comprehensive and appropriate items were included in the first form. The possible responses were open-ended options or categorical judgments. In the second phase, the questionnaire was pretested in semistructured interviews on a small group of children who either did or did not suffer from headaches (16 families were included).
This phase aimed to evaluate the face and content validity of the questionnaire. Additionally, the sensitivity was evaluated by correlating the data from the questionnaire and the medical records of the children who had headaches. This phase resulted in a revised version, which was evaluated only on healthy children. Fifty children and adolescents completed the questionnaire twice in 3 weeks. The nonresponse rate, response distributions, graphical response presentation (response inconsistency) and questionnaire burdens (time to complete, formatting, etc.) were analyzed. A number of items were modified or eliminated and the final form included 93 items which required 20 min to complete.
This time only data about breastfeeding were analyzed.
The inclusion criteria were: age 3–17 years, signed informed consent from parents of children aged 3–15 years, and from parents and children aged 15–17 years. The exclusion criterion was a previous diagnosis of a disease that has headache as a symptom.
The accuracy of the questionnaire used in this survey was based on classification and diagnostic criteria for headache disorders, cranial neuralgias and facial pain, and the revised International Headache Society criteria (11). The median age of the study subjects was 9 years and 2.5 months (range 3–17 years, standard deviation (SD) 3.44). The large number of children surveyed allowed for definite conclusions.
The study was conducted in 2 phases: completion of the questionnaire and, for those with RH, a face-to-face interview. Based on data gathered by the questionnaire, children who had RH underwent an extended interview and neurological examination.
The accuracy of the questionnaire used in this survey was based on International Headache Society criteria. Using the society’s classification codes, migraine was accepted as 1.1–1.7, migraine with aura was 1.2.2–1.2.6, migraine without aura was 1.1 and other migraine syndromes were 1.3–1.7. RH was accepted as all headache types that appeared 1 3 times per month, without separating them due to specific characteristics. The characteristics of migraine attacks according to the International Classification of Headache Disorders II were used for the diagnosis, and to analyze the data and prepare them for statistical analysis. All types of RH (idiopathic or cryptogenic RH) that were not migraine
Of the questionnaires, 3,364 (10%) may have been biased, 4.5% due to possible double interview and 5.5% due to missing data in the questionnaires. Separate data according to RH were analyzed for children who had completely answered the questionnaire.Only 24011 had complete data about breastfeading.
The Hi 2 test, Levin test and ANOVA were used as statistical methods. A significance level of 5% was used (p <0.05). Duration of breastfeeding was tested by analysis of variance, multiple level test and Scheffe test. All statistical analyses were performed with SPSS 15.0 (SPSS Inc., Chicago, Ill., USA).
There was no significant difference by age, sex, nationality and educational level of parents. Recurrent headaches of non-migraine type were defined in 18,83%, and migraine in 8,63% of children in explored group. In migraine group, migraine with aura had 25,32% of children, migraine without aura 67,50% of children and other migraine syndromes had 7,17% of children.
Children with migraine headaches had the shortest period of breastfeeding, 4 months and 10,8 days in average. Children without headaches were breastfed 4 months and 27 days, and children with non-migraine headaches 5 months and 3,9 days. (Table 1)
Table 1. Type of headaches related to duration of breastfeeding.
In the interval from 0 to 13 months in children with aura migraine, from 0 to 16 months in children with non-aura migraine and from 0 to 15 months in children with migraine equivalents, with statistic significance degree less than 0.0001, Scheffe test confirmed that children with non-aura migraine were significantly shorter breastfed than children with other migraine syndroms, and particularly shorter than children with aura migraine. Children with non-aura migraine were breastfed average 3 months 27 days, children with other migraine syndroms 4 months 5.1 days, and children with aura migraine 5 months 18.9 days. This result shows completely new, ultraearly trigger factor of migraine syndrome – duration of breastfeeding and requires further detailed investigation of breastfeeding influence on migraine syndrome. (Table 2) Comparing the duration of breastfeeding and age of headache onset related to duration of breastfeeding and type of headache, resulted in positive correlation ratio for headaches in general and nonmigraine headaches.
Table 2. Breastfeeding in children with migraine syndrome.
There is also positive correlation between non-migraine headaches and duration of breastfeeding. Children with non-migraine headaches were breastfed average 5 months 4.8 days and headaches occur in average age of 6 years 2.8 months, which shows positive Pearson correlation ratio 0.18 with statistic significance of correlation ratio 0.001. Presence of positive Pearson correlation ratio excludes direct influence of duration of breastfeeding on age of onset of non-migraine headaches.
Children with migraine headaches were breastfed average 4 months 10.8 days and migraine headaches occur in average age of 5 years 2.4 months, which shows negative Pearson correlation ratio -0,07, with statistic significance of correlation ratio 0.01. Negative pearson correlation ratio clearly shows the reciprocal influence of duration of breastfeeding on migraine syndrome onset, and earlier onset of migraine syndrome in children who were breastfed for shorter time, respectively (p=0,01). Children with aura migraine were breastfed average 5 months 19.3 days, and migraine headache with aura occurs in average age of 4 years 9.9 months, which shows negative Pearson correlation ratio -0,01. Negative Pearson correlation ratio for duration of breastfeeding and onset of aura migraine clearly shows the statistically significant effect of duration of breastfeeding on age of onset of aura migraine. Aura migraine also occurs earlier in children who were shorter breastfed. Same relation with negative Pearson correlation ratio -0,01 also appears for non aura migraine. Children with non-aura migraine were breastfed average 3 months 27 days, and non-aura migraine headache has its onset in average age of 5 years 6.5 months. Negative Pearson correlation ratio confirms clearly the influence of duration of breastfeeding on age of onset of non-aura migraine. Non-aura migraine occurs at earlier age in shorter breastfed children. (p=0,01). In other migraine syndromes and migraine equivalents, there was positive Pearson correlation ratio, so there was no direct effect of duration of breastfeeding on onset of other migraine syndromes and migraine equivalents.
From these results, we can clearly conclude that duration of breastfeeding has direct influence not only on migraine syndrome onset in general, but on age of onset of first symptoms of migraine syndrome. Migraine syndrome occurs at earlier age in shorter breastfed children. The most significant influence of duration of breastfeeding on onset of migraine syndrome is on n migraine without aure. Defining of duration of breastfeeding as an early predisponing factor of migraine syndrome in children gives the possibility for ultraearly prevention of migraine syndrome, especially in children with positive heredity for migraine syndrome.
- Bille B. (1997) A 40 year follow of school children with migraine. Cephalalgia June; 17940: 488-491.
- Vahlquist B, Hackzell G. (1949) Migraine of early onset of 31 cases in which the disease first appeared between 1 and 4 years of age. Acta pediatr Scand;18:622-636.
- Rufo M, Rodrigez C et al. (1996) The presentation form of headache in children, Rev Neurol Mar;24(127):268-272.
- Kne¾eviæ-Poganèev M. Specific features of migraine syndrome in children. The Journal of Headache and Pain.(2006)7:206-10.
- Kne¾eviæ-Poganèev M. Migraine syndrome in children Todra. Beograd 2003.
- Hockaday JM. (1988) Headache in children. British Journal of Hospital Medicine;27:41-48.
- Barlow CF. (1984) Headache and migraine in childhood. Philadelphia: Oxford Blackwell Scientific Publishers.
- Sillanpaa M, Anttila P. (1983) Changes in the prevalence of migraine and other headaches during the first seven school years. Headache; 23:15-19.. 1962; 51(136):1-151.
- Egger J, Wilson J, Carter CM et al. (1983) Is migraine food allergy? Lancet; 2(8355):856-858.
- Merikangas KR, Whitaker AE, Isler H, Angst J. (1994) The Zurich Study XXIII. Epidemiology of headache syndromes in the Zurich cohort of young adults. Em Arch Psychiatry Clin Neurosci; 244(3):14-24.
- Headache Classification Subcommittee of the International Headache Society. The International Classification of Headache Disorders: 2nd edition. Cephalalgia.2004;24 Suppl 1:9-160.
- Olesen J. The International Classification of Headache Disorders:2nd edition (ICHD-II) Rev Neurol (Paris).2005 Jul;16(6-7);689-91.