Abstract
Purpose
To explore the association between infantile colic and adolescent migraine.
Method
In a randomized general population sample, families expecting their first child were prospectively followed for infantile colic and adolescent migraine.
Results
Colic was diagnosed in 160 (13%) of 1267 infants until the age of 3 months. Migraine was ascertained in 129 (16%) of 787 adolescents at age 18 years. History of infantile colic was identified in 96 (12%) of 787 adolescents and no such history in 658 (88%) of 787 adolescents. Migraine was present in 22 (23%)/96 adolescents who had a history of infantile colic, but in only 74 (11%)/658 ones who had no such history. Of the 22 adolescents, 14 (64%) had migraine without aura and eight (36%) had migraine with aura. Infants with colic had an almost three-fold risk (risk ratio 2.8, 95% confidence interval 1.2–6.5) for adolescent migraine without aura, but no increased risk for migraine with aura (0.8, 0.3–2.2).
Discussion and conclusions
Infantile colic seems to be associated with an increased risk for migraine without aura, but not for migraine with aura. Whether infantile colic per se is a type of infantile migraine or an antecedent of future migraine remains to be answered by further research.
Introduction
Since the 1880s, gastrointestinal symptoms such as cyclic vomiting (1) and other periodic syndromes have been associated with migraine (2,3). The syndromes are considered comparable to migraine without headache (4,5). Cullen and MacDonald (6) concluded that, although sometimes presenting with typical adult migraine, childhood migraine is frequently associated with one or more symptoms, including gastrointestinal symptoms, dizzy spells, periodic attacks of fever or periodic attacks of limb pain or stiffness. Katerji and Painter (7) reported on a firstborn child who, after an uneventful pregnancy, at the age of 2 weeks, began to have episodes of irritability, inconsolable crying, face scratching, back arching, trunk twisting and extending legs, most episodes being accompanied by vomiting. The episodes varied between initially being weekly to later being daily in frequency and between 1.5 to 4 hours in duration; the episodes were relieved by cyproheptadine. The authors suggested the case as a form of very early-onset childhood migraine, based in part on the strong family history of migraine. No follow-up data were provided. Guidetti et al. (2) found ‘hyper-reactivity’ to be a common symptom and considered it as a warning sign, as many children with infantile hyper-reactivity later suffered from migraine.
Infantile colic, formerly also referred to as ‘paroxysmal fussing’, is described as ‘unexplained paroxysms of irritability, fussing, or crying which may develop into agonized screaming’(8). Recently, infantile colic has been reported to be associated with later migraine (9,11). All the previous studies are retrospective case–control studies, which start from migraine and go back through the patient's disease history. We undertook a prospective follow-up study of a fixed cohort so as to show the association between a history of infantile colic and future migraine.
Method
The present study is a sub-study of the Finnish Family Competence (FFC) study. The FFC study was designed to assess the competence of Finnish families to observe their first-born children's physical and neuropsychological growth and development, signs of developmental disorders, morbidity and behaviour including infantile colic as well as paroxysmal/episodic disorders like periodic syndromes, migraine and other headaches and epilepsy. The FFC study population was sampled from a general population resident in a geographically defined area, a region in south-western Finland. The subject collection was based on stratified randomized cluster sampling. The area was stratified into the southern and northern parts, mainly due to somewhat different socio-economic circumstances. The clusters consisted of municipalities of the sampled health authority areas. Randomization was carried out by selecting by lot 11 of the total of 35 health authority areas, weighted according to each stratum. All the 67 maternity health care clinics and 72 well-baby clinics of these 11 health authority areas participated in the study.
Families expecting their first child in 1986 and having their first visit, on the average at 10th week of pregnancy, at a maternity health care clinic were recruited to the study population. Virtually all (>99%) Finnish mothers use public maternity health care clinic services and subsequently give birth in medical units providing care at specialist level (12,13). Maternity health care nurses offered 1582 consecutive women the opportunity to participate, 1443 of whom gave informed consent and the remaining 139 (9%) declined. The parents were questioned as to their socio-economic circumstances from childhood to the present, overall health, lifestyle, motion habits, long-term illnesses and long-term medications. The study subjects were married or cohabiting couples (n = 1407, 98%) or single women (n = 36, 2%). The occupational distribution of those who declined was non-significantly different from that of the participants (14). There were 1294 deliveries, of which three were stillbirths, eight children died in early infancy and five children moved abroad. Furthermore, seven women who delivered twins and one woman who delivered triplets were excluded. Data on colic were defective in three infants. Thus, 1267 infants were eligible for the study. Data on infantile colic were available on 1171 (92%) of 1267 families for follow-up.
When the children reached 3 months, the parent (or parents) answered a question about the presence and intensity of colic pain, defined previously (14) on a segmented 5-point numeric version of a visual analogue scale with rating from a score of 1 (hardly any pain) to a score of 5 (severe pain) (14). To double-check the severity of colic, the health nurse then discussed the colic and its treatment with the parents and made her final severity rating. This was because, in contrast to the experienced public health care unit nurses, the parents of first-born infants were obviously inexperienced in their assessments. The overall agreement between the parental and the nurse's assessment of the colic pain was moderate (weighted kappa (95% confidence interval (CI)) = 0.40 (0.37–0.43)). Of the nurses' assessments of the severity of pain, 77% were in concordance with the mothers, while over 50% of the 325 children (29% of 1130) whose mothers reported severe pain were considered as moderate by the nurses. The nurses assessed 160 children (13% of 1267) as having severe colic pain, i.e. a score of 5 on the visual analogue scale. The nurses' assessments are consistent with the 13% prevalence of colic previously reported in a Finnish population (15). Thus, in the present study, infantile colic was defined as the nurses' assessment of severe colic in otherwise healthy and well-fed infants, who had no other causes that would account for crying. No formal inter-rater tests were made among the nurses. However, the nurses had regular meetings with the FFC researchers about the study including, among other things, child development, behaviour and short-term and long-term illnesses. Furthermore, the nurses had opportunities whenever necessary to consult expert paediatricians and/or child neurologists familiar with the FFC study and its objectives.
The families have been followed up with questionnaires eight times after the child was 3 months old. Each time, the parents were asked about the child's chronic illnesses and the Childhood Behavior Checklist (CBCL) (16,17) questionnaire was included at ages 3, 12, 15 and 18 years. During the last contact at the age of 18, the adolescents completed a semi-structured questionnaire about their overall health, including two pages of questions on all the items of the International Classification of Headache Disorders (ICHD) criteria for headache and migraine symptoms (18). Migraine was defined according to the ICHD criteria (18). All those who reported either migraine with aura or both migraine with and without aura were considered as having migraine with aura in the analyses. A total of 787 adolescents (62%) participated in the study at 18 years of age and comprise the present study population. Patients with a history of infantile colic were compared with those who had no such history. Dropout analysis showed a history of infantile colic in 96 (12%) of 787 participants and in 64 (13%) of 480 dropouts. The difference is non-significant (p = 0.60). The study design has been previously described in detail (19).
Factors analysed as potential determinants of colic included paternal or maternal contemporary migraine, other headaches, backaches, allergic diseases and socio-economic status. A short temper was used as a temperamental characteristic of the children, assessed by the item ‘temper’ in the Child Behavior Checklist for ages 2–3 by Achenbach et al. (16). The score ‘often’ indicates a short temper, while ‘sometimes’ and ‘never’ indicate a calmer personality.
Factors analysed as potential determinants of migraine at 18 years of age included somatic long-term comorbidity, such as: allergic diseases, bronchial asthma, coeliakia, diabetes mellitus, epilepsy and thyroid diseases checked on six occasions throughout the follow-up; concurrent somatic symptoms: extracephalic pain experience at least weekly in the back, buttocks, chest or extremities. Psychopathology was assessed by concurrent Youth Self-Report (YSR) (20) scores and the long-term effects using the CBCL/4-18 (17) scores at ages 3, 12 and 15 years. The highest deciles in CBCL or YSR scores were considered as high, while a high score for the long-term effect required a high score at least once in the three measurements. In addition, at the age of 18 a low sense of coherence (21) was defined as a score below the median value 64. Scores 4 and 5 on a 5-point likert scale from 1 = ‘very good’ to 5 = ‘very bad’ for the question ‘Have you slept well during the preceding 3 months?’ were rated as prevalent sleep disturbances.
Statistical analysis
The data are given as n (%). Modified Poisson regression models were used to study the association between childhood colic and migraine at 18 years of age and the associations of covariates with both of them. Due to the well-known difference in the prevalence of migraine between females and males in the late teenage years (22) and the associations of sex with many of the covariates (23), sex was included in all the models to control for a potential confounding effect. Finally, all covariates that were at least borderline (p < 0.1) significantly associated with either migraine (the response variable) or colic (the main predictor) were included in a multivariable model. The results are given as risk ratios (RR) with 95% CI. A p-value of < 0.05 (two-tailed) was considered statistically significant. Statistical computations were done using the SAS System for Windows, release 9.4 (SAS Institute, Cary, NC, USA).
Ethics
The baseline and follow-up study designs, including the present one, were approved by the Institutional Review Board, the First Ethics Committee of the Turku University Hospital.
Results
At the age of 3 months, colic was diagnosed in 160 (13%) of the 1267 infants (48% female). Short-tempered children had a two-fold increase in the risk of colic (RR (95% CI) 2.5 (1.1–3.7)) compared to their peers with a calmer personality. No association with colic was found in the paternal or maternal characteristics including contemporary migraine, other headaches, backache, allergic diseases and socio-economic status.
Among the 787 adolescents who participated in the follow-up study at age 18 years, migraine was diagnosed in 16%, i.e. in 21% of girls and in 10% of boys. The participants had migraine with aura in 57% of the cases and without aura in 43%. A history of infantile colic was identified in 96 (12%) of the 787 adolescents (in 11% of girls and 13% boys). Migraine preceded by a history of colic was diagnosed in 22 (23%) of the 96 patients in comparison with 107 (15%) of the 691 who had no history of colic.
Distributions of 18-year-old Finnish adolescents according to migraine without aura and with aura compared to adolescents with no migraine (18).
CBCL: Childhood Behavior Checklist; YSR: Youth Self-Report; SOC: sense of coherence.
Predictors to migraine without and with aura symptoms.
CI: confidence interval; CBCL: Childhood Behavior Checklist; YSR: Youth Self-Report; SOC: sense of coherence.
Risk ratios (RR) from modified Poisson regression for defined migraine type compared to no migraine (18).
Discussion and conclusions
In our population-based, prospective 18-year follow-up cohort, a history of infantile colic proved to be significantly associated with adolescent migraine without aura, but not with migraine with aura. The finding is in line with previous reports suggesting periodic syndromes as a type of migraine without aura (4,5). In our patients, the 13% prevalence rate of infantile colic is of the same order as the rates of 8–14% previously reported (10,24). The present prevalence rate of adolescent migraine is comparable with previous reports on patients of the same age group (25,26). The prevalence of migraine with aura was 57% in our participants. This rate is comparable with that of the 58% obtained in another population cohort (n = 3580) from the same geographic area (27).
The definition of infantile colic was based on both the parents' assessment of the severity of the child's pain and their discussion with their named public nurse about the frequency, severity and duration of single episodes of crying. The nurse classified the severity of crying. In the 1980s, the nurses of the public maternity and child welfare clinics made up the most stable staff in Finnish health care. Continuity of work over many years enabled good familiarity with and knowledge of the social circumstances and needs of the population and families of the catchment area. Thus, the nurses were experts on their target families in health matters.
Colic is reportedly associated with gastrointestinal problems, such as gallbladder hypocontractibility (28), low plasma cholecystokinin levels (29), Helicobacter pylori infection (30) and preterm and small for gestational age (31), but not with neurovegetative disturbances like abnormalities in sleep structure (31) or dysbalance of the autonomic nervous system (32). Predisposing psychosocial family factors were, however, associated with colic (33). Infantile colic is a self-limiting disorder, but long-term follow-up studies revealed more often than expected sleep disorders and temper tantrums, recurrent abdominal pain, allergic manifestations and aggressiveness (14,34). The authors, however, make no mention of migraine or other headaches at a later age (14,34).
Two retrospective studies reported migraine following a history of colic. Twenty-nine children aged 7–12 years with migraine (72% without aura and 28% with aura) and a similar number of non-migrainous controls were compared for a history of infantile colic (9). Fifteen (52%) children with migraine and six (20%) controls had a history of colic. Distribution as regards migraine with aura and without aura was not given. The probability of having a history of colic was four-fold in patients compared with controls. In contrast to our study, parental migraine was found to be associated with the occurrence of colic (9,10). In a multi-centre, case–control study of 208 children aged between 6-18 years with new-onset migraine, and 471 controls of the same age, seen at an emergency department for minor trauma during the same time interval, were compared for a history of infantile colic (11). A history of infantile colic was 6.6-fold in children with migraine compared with the controls and the history was somewhat more common in children without aura (odds ratio (OR) 7.0) than in those with aura (OR 5.7). The more common association with migraine without aura is in agreement with our observation and again suggests different origins of the two types of migraine (18,35).
A limitation of the present study might be the definition of infantile colic. There is, however, no generally accepted definition, even though the Wessel criteria (8) and their modifications are often used. The strengths of the study include random sampling from the general population, subsequent good external generalizability and a prospective time directionality. Currently, there are no similar prospective population-based follow-up studies.
In conclusion, it can be said that infantile colic is associated with increased risk of future migraine without aura, but not with migraine with aura. Further prospective studies are needed to explore the association and nature of colic as a possible separate type of early childhood migraine.
Clinical implications
Abdominal pains, including colic, are not uncommon in infancy. An agreement on the definition of infantile colic is needed. In children with infantile colic, migraine without aura is significantly more common than would be expected. The mechanisms are still unknown. Colicky infants need special attention for the risk of future migraine.
Footnotes
Funding
This work received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest
None declared.