Sage Journals: Discover world-class research

Abstract

Objective

To assess the relative risk, predictive value and population attributable risk fraction of pre-school episodic syndromes for later migraine in primary school age children.

Methods

This retrospective cohort study used health insurance data on 55,035 children born in 2006 with no diagnosis of migraine up to the age of 5 years. The relative risk, probability and population attributable risk fraction of migraine prompting a physician visit at the age of 6–10 years in children with episodic syndromes included in the International Classification of Headache Disorders (benign paroxysmal torticollis, benign paroxysmal vertigo, cyclic vomiting syndrome, recurrent abdominal symptoms and abdominal migraine) and those not included in the International Classification of Headache Disorders (pavor nocturnus, somnabulism and bruxism) diagnosed up to the age of 5 years were determined.

Results

The period prevalence of individual episodic syndromes ranged between 0.01% and 1.40%. For episodic syndromes included in the International Classification of Headache Disorders (recurrent abdominal symptoms and abdominal migraine) and for the episodic syndromes not included in the International Classification of Headache Disorders (somnambulism), the risk for later migraine was increased by factors of 2.08, 21.87 and 3.93, respectively. The proportion of risk for migraine in primary school children explained by any episodic syndromes included in the International Classification of Headache Disorders was 2.18% and for any episodic syndromes not included in the International Classification of Headache Disorders it was 0.59%.

Conclusion

Several pre-school episodic syndromes are risk factors for migraine in primary school age children. The fraction of migraine in primary school age children explained by prior episodic syndromes, however, is below 3%. A probability to develop primary school age migraine above 50% was only observed for abdominal migraine.

Keywords

Prevalence relative risk population attributable risk

Introduction

Episodic syndromes in pre-school age children may be both a precursor of migraine or may be observed concomitantly as a comorbidity in children with migraine (1 –4). Besides cyclic vomiting syndrome (CVS), abdominal migraine, benign paroxysmal torticollis, benign paroxysmal vertigo (1 –4), night terror (pavor nocturnus) (5), somnambulism, (6) bruxism (7) and recurrent gastrointestinal disturbance are discussed as episodic syndromes (1). Some episodic syndromes in infancy have specific International Classification of Diseases (ICD) codes, such as abdominal migraine and pavor nocturnus, but for most of the episodic syndromes there are no specific codes in ICD-10. Recurrent gastrointestinal disturbance, not classified as abdominal migraine or CVS, may also be a precursor of later migraine in children (4).

There are limited data on the temporal sequence of episodic syndromes in pre-school age children and later onset of migraine in children (5,8 –10). Migraine may start in primary school age children (11) and its early manifestation at primary school age may set the track for later migraine and headaches (12). Associations of pre-school episodic syndromes with migraine at primary school age may only be considered as a potential risk factor if the pre-school episodic syndromes precede later migraine. Long-term health care data allow the assessment of the temporal sequence between physician consultations for episodic syndromes at pre-school age and later consultations for migraine.

Regarding the temporal sequence of episodic syndromes and later migraine, two issues are of interest. (A) Strength of the association – how much higher is the risk of later migraine prompting a physician visit in primary school age children with episodic syndromes compared to the risk of children without episodic syndromes? (B) What is the risk for later migraine in children with an episodic syndrome at pre-school age? We addressed these questions in health care data collected between 2006 and 2015.

Methods

Data source and available data

People living in Germany, who have a job contract, have a compulsory statutory health insurance, which also covers family members. Only people with an annual income exceeding €57,600 (status 2017) are exempt and have the opportunity to choose between private and statutory health insurance (13). About 90% of the German population has statutory health insurance (14).

In this study, we had access to the BARMER dataset covering 2006 to 2015. With about 8.6 million health insurance holders in 2015, BARMER is the second largest statutory health insurance in Germany, serving about 11% of the German population (15,16). For all statutory health insurances, the billing for physician treatment is handled by regional billing agencies (Kassenärztliche Vereinigungen). On a 3-monthly basis, physicians claim their treatment fees from the billing agencies. Claims for treatment fees require the date of the physician visit and ICD code for the diagnosis related to the visit.

The billing agencies charge the statutory health insurance (e.g. BARMER) based on the number of visits and the diagnosis. For each insured individual, BARMER thus collects prospectively information on physician contacts and diagnoses. The BARMER data warehouse has generated a dataset from 2006 to 2015, which was accessible for the analysis underlying this study. The data warehouse dataset provides a uniform patient identifier and patients’ year of birth and sex. All data were collected prospectively. We analysed these data as a retrospective cohort study in 2017.

Cohorts compared

We used data for children born in 2006 with complete follow-up until 2015 (n = 55,347). Physician visits related to episodic syndromes up to 2010 defined the cohorts compared. Patients with no episodic syndromes were the control group; patients with episodic syndromes were the exposed group. In order to assure temporal sequence we excluded patients with a diagnosis of migraine up to 2010 (except of those with abdominal migraine). Physician visits for migraine in 2011 through 2015 were the outcome of interest between the exposed and control groups (Figure 1).

Figure 1.

Flow chart showing the study population of children with/without episodic syndromes emerging during pre-school age and follow up for cases with migraine prompting a physician visit in primary school age^a

Case definitions

For all analysis, we used ICD-10 codes. We defined ‘pre-school’ as the age from birth to ≤5 years (2006–2010) and ‘primary school age’ from >5 to 10 years (2011–2015).

For episodic syndromes included in the International Classification of Headache Disorders (ICHD), we followed the ICHD-3 beta diagnostic criteria (1) regarding the minimum number of episodes by estimating the number of episodes by the number of physician visits. As many episodic syndromes do not have a specific ICD-10 code, the most fitting and most commonly used codes by paediatricians were considered for each episodic syndrome. For benign paroxysmal torticollis (ICD M43.6 and F45.98), the required number of physician visits was ≥2 (ICHD-3 says recurrent attacks); for CVS (ICD R11), the number was ≥5 visits (ICHD-3 requires ≥5 episodes); in a sensitivity analysis, a less strict case definition for CVS requiring ≥2 visits was used; for recurrent gastrointestinal disturbance (ICD K30, K58, K59 and R10), the number was ≥5 visits (according to ICHD-3).

For benign paroxysmal vertigo (ICD H81.1, R42, F45.8, H81.4 and H81.3) and abdominal migraine (ICD 43.8), ≥5 episodes are required in the ICHD-3. Since we only have information on the number of visits and not on the number of episodes, we felt that after at least two visits with a good explanation of the natural cause of these conditions, parents would no longer see a physician after each episode. We therefore required only ≥2 visits in accordance with the ICHD-3 recommendation for benign paroxysmal torticollis.

Pavor nocturnus (ICD F51.4), somnambulism (ICD F51.3) and bruxism (ICD F45.8), which have a specific ICD-10 code, are just mentioned in the ICHD; they are not included as episodic syndromes and do not have diagnostic criteria. All of these conditions may appear threatening to parents at first occurrence, prompting a physician visit. After an explanation of their benign character and cause, however, there is no need to see a physician following each episode. Therefore, we required only ≥1 visit for these conditions.

For the outcome assessment, we defined migraine in primary school age children by any ICD G43 codes.

Statistical analyses

For each specific episodic syndrome, we compared children with this specific episodic syndrome to children without any episodic syndrome with respect to diagnosis of migraine at primary school age. We assessed the period prevalence and absolute number of children seeing a physician due to the respective episodic syndrome (according to our case definition) in 2006 through 2010. We estimated the relative risk for migraine at primary school age in children with pre-school episodic syndromes with reference to children without episodic syndromes with a 95% confidence interval (CI). The population attributable risk fraction (PARF) of each episodic syndrome and all episodic syndromes combined for the risk for migraine in primary school age children was estimated using a model based approach (17). We additionally assessed the probability for developing migraine at primary school age for each of the episodic syndromes in pre-school age children (proportion of children with primary school age migraine of all children with a certain episodic syndrome at pre-school age); 95% CIs were calculated for these proportions.

Statistics were performed with the statistical software R, version 3.4.3 (R Foundation for Statistical Computing, Vienna, Austria; https://www.r-project.org/).

Ethics

The study was approved by the ethics committee of the medical faculty of the Ludwig-Maximilians-University, Munich (17-744 UE).

Results

In the BARMER dataset we identified n = 55,347 children born in 2006 with complete follow-up of a 10-year period; n = 312 (0.65%) of these had a migraine diagnosis (excluding abdominal migraine) before 2011. These had to be excluded since we attempted to estimate the temporal sequence of episodic syndromes and migraine, leaving n = 55,035 children for the analysis (Figure 1).

The period prevalence for any episodic syndrome in pre-school age children was 4.91% (2.99% for episodic syndromes included in the ICHD and 2.22% for episodic syndromes not included in the ICHD). For individual episodic syndromes, the period prevalence in pre-school age children was in the range of 0.01% to 1.40%. The number of children visiting a physician due to migraine at primary school age was n = 1844 (3.35%) in total with n = 1709 (3.11%) for the absence of any episodic syndrome. The flow chart (Figure 1) depicts the number of children with a physician contact for migraine at primary school age by episodic syndrome.

For several episodic syndromes the relative risk for later migraine in primary school age children was significantly increased: of the episodic syndromes included in the ICHD, recurrent gastrointestinal symptoms had a relative risk of 2.08, and abdominal migraine a relative risk of 21.87; for episodic syndromes not included in the ICHD, somnambulism in pre-school age children was significantly associated with primary school migraine with a relative risk of 3.93 (Table 1). The proportion at risk for migraine in primary school children explained by episodic syndromes was 2.77% in total (sum of the individual PARF estimates), 2.18% for the episodic syndromes included in the ICHD, and 0.59% for the episodic syndromes not included in the ICHD. The probability for developing migraine at primary school age in children exposed to the individual risk factors was in the range 3.95–12.82%, except for abdominal migraine with a risk of 71.43% (Table 1). For children with abdominal migraine, the risk for persistent abdominal migraine was 28.57%.

Table 1.

Pre-school (≤5 years) episodic syndromes and migraine in primary school age children.^a

Episodic syndrome prompting a physician visit ≤5 years old		Migraine diagnosis at 6–10 years old
Type (no. of physician visits)	Period prevalence in % (n)	Proportion of children developing migraine, % [95% CI] (n)	PARF %	Relative risk by episodic syndrome [95% CI]
Episodic syndrome included in the ICHD
Benign paroxysmal torticollis (≥2)	0.83 (456)	3.95 [2.16–5.73] (18)	0.02	1.21 [0.64–1.91]
Benign paroxysmal vertigo (≥2)	0.43 (239)	5.02 [2.25–8.61] (12)	0.12	1.54 [0.98–2.67]
Cyclic vomiting syndrome (≥5)	0.99 (546)	6.04 [4.05–8.38] (33)	0.56	1.85 [0.95–2.58]
Recurrent abdominal symptoms (≥5)	1.20 (663)	6.79 [4.87–8.70] (45)	1.10	2.08 [1.00–2.77]
Abdominal migraine (≥2)	0.01 (7)	71.43 [37.96–100.00] (5)	0.36	21.87 [3.6–35.03]
Any episodic syndrome included in the ICHD	2.99 (1649)	5.64 [4.53–6.75] (93)	2.18	1.73 [1.41–2.12]
Episodic syndrome not included in the ICHD
Pavor nocturnus (≥1)	1.40 (770)	4.81 [3.29–6.32] (37)	0.00	1.47 [0.70–2.02]
Somnabulism (≥1)	0.07 (39)	12.82 [2.33–23.31] (5)	0.38	3.93 [2.20–8.91]
Bruxism (≥1)	0.77 (426)	4.69 [2.69–6.70] (20)	0.21	1.44 [0.79–2.21]
Any episodic syndrome not reported in the ICHD	2.22 (1220)	5.25 [3.99–6.50] (64)	0.59	1.61 [1.26–2.05]
Any episodic syndrome	4.91 (2701)	5.00 [4.18–5.82] (135)	2.77	1.53 [1.29–1.82]

Prevalence, proportion of children developing migraine at primary school age, relative risk for migraine at primary school age (comparator: children without the respective episodic syndrome) and PARF.

ICHD: International Classification of Headache Disorders; PARF: population attributable risk fraction.

In a sensitivity analysis, we attempted to find out whether the failure to observe a significant association between CVS and migraine might be related to the strict case definition requiring at least five visits. We therefore assessed whether a case definition requiring fewer physician visits would yield a significant association. Based on 10,983 children with CVS of pre-school age with at least two physician visits (20% of all included children), we observed a significant relative risk for migraine at primary school age of 1.46 (95% CI: 1.32–1.66).

Discussion

These data confirm an increased relative risk for migraine in primary school children related to any episodic syndrome and specifically to somnambulism, recurrent abdominal symptoms and abdominal migraine. For benign paroxysmal torticollis, benign paroxysmal vertigo, CVS, pavor nocturnus and bruxism, the point estimate suggested a small to moderate risk; the respective 95% CIs included one (no effect), however, precluding definite conclusions. The contribution of all episodic syndromes combined for the risk of migraine in pre-school children was below 3%. The probability for developing migraine at primary school age for exposed children was high (>50%) for abdominal migraine but considerably lower for somnambulism and recurrent abdominal symptoms.

Recent reviews have included infantile colic, CVS, abdominal migraine, benign paroxysmal torticollis and benign paroxysmal vertigo as episodic syndromes associated with migraine (2,3). Unfortunately we could not assess the established role of infantile colic for later migraine (18) in the BARMER dataset. Infantile colics typically are self-limited at 6 months of age. The BARMER data only provide the year of birth for each child (not the exact month) and therefore case definitions were too imprecise to estimate the association of infantile colics and later migraine.

For benign paroxysmal torticollis an association with migraine has been reported (19,20). It is unclear, however, whether benign paroxysmal torticollis precedes later migraine in children. Benign paroxysmal torticollis is the rarest episodic syndrome. Therefore, it might be surprising that we had follow-up data on more than 400 children. The reason for this high prevalence might be the lack of a specific ICD code for benign paroxysmal torticollis. We identified cases using ICD M43.6 and F45.98, with which, however, muscular torticollis cases might also be included. Therefore, the risk ratio of 1.21 (95% CI: 0.64–1.91) observed with our case definition must be interpreted with caution.

A retrospective series of 100 children with benign paroxysmal vertigo reported an association with migraine (21). Based on a prospective follow-up of 239 cases with at least two physician contacts for benign paroxysmal vertigo in pre-school age children, we observed a risk ratio of 1.54 (95% CI: 0.98–2.67). Since the lower margin of the 95% CI was only slightly below 1.0, these findings may hint at a moderately increased risk for early-onset migraine at primary school age in children with benign paroxysmal vertigo of pre-school age.

Abdominal migraine was the strongest risk factor and had the highest probability for later migraine in primary school age children. The respective 95% CIs were wide, however, since there were only a few cases of abdominal migraine diagnosed in pre-school children. This reflects the later mean age of onset of abdominal migraine at about 7 years (22,23).

CVS is a condition belonging to functional abdominal symptoms in the paediatric population (24). Cyclic vomiting typically starts at the age of 5 years and has been linked to migraine (25). Based on 546 children with CVS of pre-school age with more than four physician visits, we observed a relative risk for migraine at primary school age of almost 2.0, with a lower border of the 95% CI only slightly below 1.0. In a sensitivity analysis, we attempted to find out whether the failure to observe a significant association between CVS and migraine might be related to the strict case definition requiring at least five visits. With a case definition requiring only two physician visits, a significant risk ratio of 1.46 (95% CI: 1.32–1.66) was observed. The prevalence of CVS with at least two physician visits, however, was much higher (19.96% compared to 0.99%, with the case definition requiring at least five visits), which might point to the inclusion of children with vomiting in the course of infections, which are very common at this young age, but are definitely not episodic syndromes. These findings may therefore hint at a moderately increased risk for later migraine in primary school age children with cyclic vomiting at pre-school age.

Recurrent gastrointestinal disturbances have also been included in episodic syndromes linked to migraine (26). We chose a broad spectrum of ICD codes for identification of recurrent gastrointestinal disturbances (K30, K58, K59 and R10). In order to make sure that these were indeed repetitive episodic symptoms we required a minimal number of at least five related physician visits (4) for the case definition. There was a significant association related to later migraine at primary school age, but it was considerably less strong than for abdominal migraine. Some of these cases might possibly be undiagnosed cases of abdominal migraine.

The role of night terror (pavor nocturnus) for later migraine was suggested from a cross-sectional study in adolescents (5). Night terrors are common in pre-school age children but may not necessarily prompt a physician visit. The condition is usually self-limiting within 1–2 years (27). Our prospectively collected data with follow-up of 770 children with pavor nocturnus and a relative risk of 1.47 (95% CI: 0.70–2.02) cannot rule out a temporal sequence from night terrors at pre-school age to migraine at primary school age.

An association between somnambulism and migraine was reported more than 30 years ago (28). A subsequent retrospective study suggested that somnambulism in late infancy is a precursor of later migraine (6). We confirmed the temporal sequence based on prospectively collected data.

The data clearly show an overall effect of episodic syndromes on later migraine at primary school age. Episodic syndromes may thus be a risk factor for later migraine rather than only an indicator of a potentially shared pathophysiology. The fraction of migraine cases in primary school age children explained by episodic syndromes, however, is below 3%, indicating that few cases of migraine at primary school age could be attributed to episodic syndromes at pre-school age.

Episodic syndromes at pre-school age may not only be a precursor of migraine at primary school age. Being a precursor, indeed, might not necessarily imply causality, but rather a shared pathophysiology. The concept of shared pathophysiology is supported by the observation of later episodic syndromes at primary school age in some children with migraine at pre-school age. The concept of shared pathophysiology would be compatible with episodic syndromes preceding migraine or occurring later on in the cause of migraine. The unique potential of these data is to show the temporal sequence.

Strengths and limitations

Using complete, prospectively collected health care data over a 10-year period we could assesses the temporal sequence of episodic syndromes and later migraine in primary school age children, without the common bias related to losses to follow-up in prospective cohort studies. A further strength is the number of cases with episodic syndromes with follow-up, allowing the detection of even small relative risks.

There are, however, a number of limitations of our data. For many episodic syndromes (e.g. benign paroxysmal torticollis) no specific ICD-10 codes exist. We therefore had to use a more generalized code (e.g. M43.6 (‘torticollis’) and F45.98 (‘psychogenic torticollis’)) and focus on the episodic character of paroxysmal torticollis by requiring ≥2 physician visits because of torticollis for the data analysis. We had no opportunity to validate the diagnoses for the episodic syndromes according to the ICHD-3 beta criteria because of data protection concerns. By using the number of physician visits to define the minimal number of episodes, we can be certain about the attainment of the minimum required number of episodes that were severe enough to prompt a physician visit. Random misclassification of the exposure therefore is likely. Since random misclassification accounts for an underestimation of strength of the association, the true relative risks might be even higher.

Prediction of later migraine at primary school age by pre-school episodic syndromes was poor except for abdominal migraine. We definitely miss incident cases of migraine beyond primary school age, limiting the validity of the data for prediction of later migraine. The prevalence of migraine increases substantially in adolescence, so many of these children may as yet go on to develop migraine in adolescence. Prediction may thus improve with longer follow-up of the cohort since some of the children with early episodic syndromes might acquire migraine during adolescence. The prevalence of migraine increases substantially in adolescence and thus the a priori probability of migraine. Follow-up of the cohorts is planned to allow for assessment of the probability to develop migraine throughout childhood and adolescence. Hence, reassuring parents that the probability for later migraine at primary school age in children with pre-school episodic syndrome is low based on these data is only generalizable to reassuring them that the risk in ages 6–10 is low – not necessarily that the risk EVER is low.

A further limitation might arise from non-optimal age ranges for assessment of episodic syndromes. We applied the same age range to identify all episodic syndromes at pre-school age although the mean onset of episodic syndromes described in the literature varies widely (e.g. benign paroxysmal torticollis between 2 and 8 months, abdominal migraine at 7 years) (22,23).

Conclusion

Several episodic syndromes are a risk factor for later migraine. The fraction of migraine cases at primary school age in children explained by prior episodic syndromes, however, is below 3%. A predictive value for primary school age migraine above 50% was only observed for abdominal migraine.

Public health relevance

Episodic syndromes have been linked to migraine, but a temporal sequence has not yet been established in prospective data.

Several episodic syndromes are risk factors for later migraine.

Despite being risk factors, the risk of migraine at primary school age in most exposed children is low, except for abdominal migraine.

The proportion of later migraine explained by pre-school episodic syndrome is lower than 3%.

Footnotes

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Headache Classification Committee of the International Headache Society (IHS). The International Classification of Headache Disorders, 3rd edition (beta version). Cephalalgia 2013; 33: 629–808.

Lebron

Vasconcellos

. The episodic syndromes that maybe associated with migraines. Semin Pediatr Neurol 2016; 23: 6–10.

Rothner

Parikh

. Migraine variants or episodic syndromes that may be associated with migraine and other unusual pediatric headache syndromes. Headache 2016; 56: 206–214.

Winner

. Abdominal migraine. Semin Pediatr Neurol 2016; 23: 11–13.

Fialho

Pinho

Lin

et al.

Sleep terrors antecedent is common in adolescents with migraine. Arq Neuropsiquiatr 2013; 71: 83–86.

Giroud

d’Athis

Guard

et al.

[Migraine and somnambulism. A survey of 122 migraine patients]. Rev Neurol 1986; 142: 42–46.

Masuko

Villa

Pradella-Hallinan

et al.

Prevalence of bruxism in children with episodic migraine – A case-control study with polysomnography. BMC Res Notes 2014; 7: 298–298.

Dooley

Augustine

Gordon

et al.

Alice in Wonderland and other migraine associated phenomena-evolution over 30 years after headache diagnosis. Pediatr Neurol 2014; 51: 321–323.

Krams

Echenne

Leydet

et al.

Benign paroxysmal vertigo of childhood: Long-term outcome. Cephalalgia 2011; 31: 439–443.

10.

Lindskog

Odkvist

Noaksson

et al.

Benign paroxysmal vertigo in childhood: A long-term follow-up. Headache 1999; 39: 33–37.

11.

Abu-Arafeh

Razak

Sivaraman

et al.

Prevalence of headache and migraine in children and adolescents: A systematic review of population-based studies. Dev Med Child Neurol 2010; 52: 1088–1097.

12.

Antonaci

Voiticovschi-Iosob

Di Stefano

et al.

The evolution of headache from childhood to adulthood: A review of the literature. J Headache Pain 2014; 15: 15–15.

13.

Bundesregierung Deutschland. Beitragsbemessungsgrenzen der Kranken- und Rentenversicherung 2017, https://www.bundesregierung.de/Content/DE/Artikel/2016/10/2016-10-12-bemessunggrenzen-sozialversicherung.html [updated 12.10.2016] (accessed 15 December 2017).

14.

Hein

Koster

Bollschweiler

et al.

Prevalence of inflammatory bowel disease: Estimates for 2010 and trends in Germany from a large insurance-based regional cohort. Scand J Gastroenterol 2014; 49: 1325–1335.

15.

BARMER GEK. Kerndaten der BARMER GEK, Stand 01.06.2015, https://www.barmer.de/presse/infothek/daten-und-fakten/kerndaten-42164 (accessed 30 July 2018).

16.

Statistisches Bundesamt (Destatis). Bevölkerung: Deutschland, 31.12.2012 bis 31.12.15, Altersjahre, https://www-genesis.destatis.de/genesis/online;jsessionid=23ED8210212D537241EDCF787C10D009.tomcat_GO_2_1?operation=previous&levelindex=2&levelid=1505294029304&step=22017 (accessed 15 December 2016).

17.

Ruckinger

von Kries

Toschke

. An illustration of and programs estimating attributable fractions in large scale surveys considering multiple risk factors. BMC Med Res Methodol 2009; 9: 7–7.

18.

Gelfand

. Infant colic. Semin Pediatr Neurol 2016; 23: 79–82.

19.

Rosman

Douglass

Sharif

et al.

The neurology of benign paroxysmal torticollis of infancy: Report of 10 new cases and review of the literature. J Child Neurol 2009; 24: 155–160.

20.

Drigo

Carli

Laverda

. Benign paroxysmal torticollis of infancy. Brain Dev 2000; 22: 169–172.

21.

Batu

Anlar

Topcu

et al.

Vertigo in childhood: A retrospective series of 100 children. Eur J Paediatr Neurol 2015; 19: 226–232.

22.

Cuvellier

Lepine

. Childhood periodic syndromes. Pediatr Neurol 2010; 42: 1–11.

23.

Spiri

Rinaldi

Titomanlio

. Pediatric migraine and episodic syndromes that may be associated with migraine. Ital J Pediatr 2014; 40: 92–92.

24.

Lewis

Palsson

Whitehead

et al.

Prevalence of functional gastrointestinal disorders in children and adolescents. J Pediatr 2016; 177: 39–43.e3.

25.

Gelfand

. Episodic syndromes that may be associated with migraine: A.K.A. “the Childhood Periodic Syndromes”. Headache 2015; 55: 1358–1364.

26.

Winner

. Childhood periodic syndromes and migraine. Curr Pain Headache Rep 2005; 9: 197–201.

27.

Kotagal

. Parasomnias of childhood. Curr Opin Pediatr 2008; 20: 659–665.

28.

Barabas

Ferrari

Matthews

. Childhood migraine and somnambulism. Neurology 1983; 33: 948–949.

Do pre-school episodic syndromes predict migraine in primary school children? A retrospective cohort study on health care data

Abstract

Objective

Methods

Results

Conclusion

Keywords

Introduction

Methods

Data source and available data

Cohorts compared

Case definitions

Statistical analyses

Ethics

Results

Discussion

Strengths and limitations

Conclusion

Public health relevance

Footnotes

Declaration of conflicting interests

Funding

References