Abstract
We conducted a self-administered questionnaire to investigate ice-cream headache in school adolescents aged 13-15 in Taiwan. The target population was 8789 students in 6 public junior high schools. A total of 8359 students completed the questionnaire (response rate 95.1%). The prevalence of ice-cream headache was 40.6%. It was significantly higher in boys than in girls, and increased with grade. Students with migraine had a higher frequency of ice-cream headache compared with the students without migraine (55.2% vs. 39.6%, P < 0.0001). The prevalence of ice-cream headache increased among students with more migrainous features. Approximately one third of students decreased their intake of ice cream, or abstained completely, especially the younger students. Our study suggests icecream headache is very common in Taiwanese adolescents, and it is more common in students who experienced migraine.
Introduction
Ice cream consumption is a common cause of headache that has been known for a long time. The lifetime prevalence of cold stimulus headache – those occurring during external exposure to cold and during ingestion of cold food or drink – was 15% in a cross-sectional survey of an adult community population (1). Drummond and Lance (2) reported that 189 (36.7%) of 530 patients attending a headache clinic had experienced ice-cream headache. The relationship between migraine and ice-cream headache is controversial. Raskin and Knittle (3) found that ice-cream headache occurred much more commonly in migraine sufferers; however, Bird et al. (4) found ice-cream headache to be more common in nonmigraineurs.
To examine the prevalence and characteristics of ice-cream headache and its relationship to migraine, we conducted a survey by questionnaire among students aged 13–15 years using a population-based sample. This study is also part of an adolescent migraine survey study (5).
Subject and methods
Sampling
The study subjects were a convenient sample of public junior high school students. In Taiwan, education is obligatory from ages 6–15. More than 90% of teenagers aged 13–15 attend public junior high schools, and the remaining teenagers attend private schools. The study was conducted in two stages. The first stage of the headache survey included four junior high schools that were randomly chosen from northern, central, southern and eastern Taiwan. All of the students in these four schools were the targeted population (n = 4436). The second stage included 4167 students in 2 schools in Taipei (northern Taiwan). The target population was equal to 1% of the adolescents aged 13–15 years in Taiwan.
Questionnaire
During the first stage of the study we used a self-administered 25-item migraine questionnaire designed according to International Headache Society (HIS) criteria. The details of the migraine questionnaire are described elsewhere (5). In brief, the questionnaire included basic demographic data, headache profiles, aura manifestations, painkiller usage, school absence, ice cream headache, recurrent abdominal pain, cola-drinking habit and menstruation in girls. The question for ice-cream headache was, ‘Do you develop headaches when you eat ice cream, ice pole, sorbet or other crushed ice desserts?’ We did not define the frequency of ice cream headache. To make a diagnosis of migraine, the headache features had to fulfil all of the IHS diagnostic criteria from B to D for migraine (IHS code 1.1), except that the criteria A, i.e. ‘at least five attacks’ was not mandatory in this study (6). ‘Non-migraine headache’ was diagnosed if headaches did not fulfil the above modified criteria. Subjects who had both migraine and nonmigraine headaches were classified as migraineurs. Three neurologists, who were blind to the questionnaire results, interviewed 217 sampled students with a semistructured questionnaire and reached a headache diagnosis at a consensus meeting according to the modified IHS criteria described above. The sensitivity of the headache questionnaire used for diagnosing migraine was 74.1%, with a specificity of 80.4%.
In the second stage, we inquired about the details of the ice-cream headache including the frequency, location, duration, and intensity, and if they avoided ice cream due to headache.
Survey procedures
Letters describing our study aims and methodology were mailed to the principals of the sample schools for their approval. All the students in these sampled schools answered a self-administered questionnaire.
Statistical analysis
The prevalence of ice-cream headache was reported as the number of cases per 100 persons with 95% confidence intervals (CI). Student's t-test and χ2 tests were used for comparison when appropriate. The χ2 test for trend analysis was used to analyse the trend among different age groups and numbers of migraine symptoms. Pearson correlation test was used to test the relationship between the location of ice-cream headaches and other headaches. A P-value of < 0.05 was considered statistically significant.
Results
The target population in the first stage consisted of 4436 students (2216 Male, 2220 Female) in 4 public junior high schools. A total of 4239 questionnaires (2096 M, 2143 F) were returned for a response rate of 95.6%. We excluded the students who had no answer for the item regarding ice-cream headache. The resulting final sample size consisted of 4192 students. The target population in the second stage was 4353 (2393 M, 1960 F) in 2 public junior high schools in Taipei. A total of 4241 questionnaires (2321 M and 1920 F) were returned for a response rate of 95.1%. After excluding the incomplete samples, the final samples size was 4167 (2276 M and 1891 F). The overall response rate of the two stages was 95.1%.
Prevalence of ice-cream headache
Overall, 3394 students (40.6%) reported they had ice-cream headache. Table 1 shows the age- and sex-specific prevalence (%) of ice-cream headache. Boys had a significantly higher prevalence of ice-cream headache than girls (43.4% vs. 37.5%, χ2 = 30.1, d.f. = 1, P < 0.0001). The prevalence increased with grade (χ2 for linear trend = 13.2, d.f. = 2, P= 0.001). This trend was also found in each gender.
The grade- and sex-specific prevalence of ice-cream headache
95% CI, 95% confidence interval.
Ice-cream headache and migraine
A total of 526 (6.2%) students (284 F/242 M) had migraine according to IHS criteria. Ice-cream headache was reported by 55.2% of migraineurs; 45.5% in students with nonmigraine headache; and 29.1% in students without any headache (χ2 for trend = 249.8, d.f. = 2, P < 0.0001).
We grouped the students according to the six migrainous symptoms (unilateral, throbbing, aggravation by physical activity, severe intensity, nausea/vomiting, photophobia and phonophobia). Figure 1 shows that the prevalence of ice-cream headache increased with an increasing number of migrainous symptoms (χ2 for linear trend = 346.3, d.f. = 6, P < 0.001). A similar pattern was found in both genders.
The prevalence of ice-cream headache and the number of migraine symptoms (unilateral, throbbing, aggravation by physical activity, severe intensity, nausea/vomiting, photophobia, and phonophobia) and gender. ▪ Total; □ Male; Female.
Characteristics of ice cream headache
In the second stage of the study, 1744 (41.9%) of the students reported ice-cream headache. Table 2 shows the frequency, location, duration, and intensity of ice-cream headache in these 1744 students. Most students had infrequent attacks, and reported that the duration of pain was very short and the intensity mild. The most common locations of pain were the bilateral temporal and bilateral frontal regions. The students rarely abstained from ice cream or other icy foods because of headache. The intensity and duration of ice-cream headache were not different among the students who abstained, decreased, or continued to eat ice cream. The younger students were more likely to decrease or abstain from ice cream because of headache (46.6% in grade 7, 33.9% in grade 8, 33.7% in grade 9; χ2 = 23.8, d.f. = 2, P < 0.001).
The frequency, location, duration, intensity, and abstinence of ice-cream headache
The students with migraine reported more severe intensity of ice-cream headache than the nonmigraine students (χ2 for linear trend = 11.4, d.f. = 2, P= 0.003). Nevertheless, the most intense headaches even in students with migraine were mild. Students with a history of head injury were more likely to have ice-cream headache than students without a history of injury (47.3% vs. 39.2%; χ2 = 24.5, d.f. = 1, P < 0.0001). The odds ratio was higher in girls with a history of head injury (1.5, 95% CI, 1.2–1.9) than in boys with a history of head injury (1.3, 95% CI, 1.1–1.5). The frequency of headache other than ice-cream headache was higher in ice-cream headache students (1.93 times per month vs. 2.92 times per month, F = 41.2, P < 0.001). Among the 1084 students with ice-cream headache and other headache, 577 (53.2%) reported the same location for both. The percentage of the same location between ice-cream headache and other headache did not differ between the patients with migraine and nonmigraine headache (48.8% vs. 51.2%, P= 0.294). The location of ice-cream headache correlated to the location of other headaches (frontal, r = 0.47; whole head, r = 0.23; occipital, r = 0.57; vertex, r = 0.35 and temporal region, r = 0.37, all P < 0.001).
Discussion
Our study confirms that ice cream headache is a common phenomenon among adolescents, and that most ice-cream headaches are brief and mild. Our study also supports the theory that ice-cream headache is more common among migraine patients.
We are not sure why the correlation between migraine and ice-cream headache were different in the previous two studies (3, 4). The discrepancy might be due to differences in the study populations. Raskin and Knittle (3) conducted their study in 108 non-neurological hospitalized patients, 59 of them had migraine (37 F and 22 M, mean age 45 years) and 49 of them were nonheadache control population (23 F and 26 M, mean age 49 years). During interview, ice cream headache was reported in 93% of migraine sufferers and 31% in controls. The study population of Bird et al. (4) consisted of 70 patients (65 F and 5 M, mean age 41 years) from a migraine clinic and 50 medical and dental student volunteers (19 F and 30 M, mean age 20 years). In contrast to Raskin & Kittle (3) and our studies, 27% of the migraine patients and 40% of the students reported previous ice cream headaches. Following palatal application or a swallow of ice cream, 17% of the migraine patients and 46 of the students developed headache. The patients with migraine were treated in the migraine clinic, and the volunteers were self-selected. Thus, selection bias cannot be ruled out in the study by Bird et al. (4). The other potential biases in this study are the temperature of ice cream and the medication used by migraine patients. The ice cream used for the students was colder than used for the migraine patients (−26°C vs. − 15°C). We are not sure if the lower temperature might provoke ice cream headache more easily. The migraine patients came from the migraine clinic, so most of them might be under migraine treatment. It is possible antimigraine drugs could modulate the development of ice-cream headache. It deserves further study to clarify it.
Raskin has suggested that ice cream headache may represent a model of migraine (7). Sleigh observed decrease of the middle-cerebral-arterial flow velocities in a subject during the ice-cream headache developed (8). The progression of vasoconstriction to vasodilation in cranial arteries in the pathogenesis of migraine might also apply to the ice cream headache.
Consistent with Drummond and Lance (2), we found that susceptibility to ice-cream headache increased in direct proportion to the number of migrainous symptoms associated with the patient's other headache. It implies that the susceptibility to ice-cream headache increases toward the migraine end of the headache spectrum and they might share a common pathogenesis.
About half of the students who had both ice cream headache and other headache reported the location of the ice-cream headache and the habitual site of headache were the same. It suggests that segments of the central pain pathways remain hyperexcitable between spontaneous attacks. The previous head injury history also associated with a higher prevalence of ice-cream headache. It is also possible the central pain pathway became hyperexcitable after trauma. Nevertheless, we don't know why the prevalence of ice cream headache was greater in girls than boys with a history of head injury.
It is interesting that boys had a higher incidence of ice-cream headache than girls. It is possible that boys ate larger quantities of colder ice cream and ate faster than girls. In Bird's study (4), rapid ingestion of crushed ice slurry was particularly likely to provoke a headache. In Taiwan, the most popular icy dessert in summer is crushed ice with varieties of fruits and beans. It might easily provoke headache. No need for abstinence from ice cream was suggested (9) but nearly a half of the grade 7 students tried to decrease their ice cream consumption, or abstain. As they grow up, they might learn it is a benign condition and the percentage of avoidance decreased.
Our study was conducted in the community and the sample size was large, minimizing the possibility of a sampling bias. Nevertheless, some methodological issues should be addressed here. First, the migraine diagnosis was obtained by questionnaire rather than physician diagnosis. Nevertheless, the sensitivity and specificity of this questionnaire are appropriate for a field study (5). Second, recall bias could not be excluded in this survey. The lifetime prevalence of ice-cream headache was 40.6% among our participants. Although it is lower than what was reported by Canadian adolescents (10), it is higher than another study (1). Most of the characteristics of ice-cream headache that the students reported were consistent with the other reports. In addition, since we studied the lifetime prevalence of ice-cream headache, the prevalence increased as age accordingly as we found. Thus, we expect the bias is small. Third, we are not sure if these results apply to the older population. Further study is required for confirmation.