Abstract
Background
Headache attributed to ingestion or inhalation of a cold stimulus (HICS), colloquially called ice-cream headache, is a common form of a primary headache in adults and children. However, previous studies on adults are limited due to the small number of patients. Furthermore, most of the subjects in previous studies had a history of other primary headaches.
Methods
Biographic data, clinical criteria of HICS and prevalence of primary headache were collected by a standardized questionnaire. A total of 1213 questionnaires were distributed; the return rate was 51.9% (n = 629); 618 questionnaires could be analyzed.
Results
In a cohort of 618 people aged between 17–63 years (females: n = 426, 68.9%), the prevalence of HICS was 51.3% (317 out of 618). There was no difference between men and women (51.3% vs. 51.6%). The duration of HICS was shorter than 30 sec in 92.7%. In the HICS group, localization of the pain was occipital in 17%. Trigemino-autonomic symptoms occurred in 22%, and visual phenomena (e.g. flickering lights, spots or lines) were reported by 18% of the HICS group. The pain intensity, but not the prevalence of HICS, was higher when tension-type headache and migraine or both were present as co-morbid primary headaches (Numeric Rating Scale (NRS) 4.58 and 6.54, p = 0.006). There was no higher risk of participants with migraine getting HICS than for those who did not have migraine (odds ratio = 1.17, 95% confidence interval (CI) 0.75–1.83; p = 0.496).
Conclusion
The results of this study modified the current criteria for HICS in the ICHD-3 regarding duration and localization. In addition, accompanying symptoms in about one fifth of the participants are not mentioned in the ICHD-3. Neither migraine nor tension-type headache seems to be a risk factor for HICS. However, accompanying symptoms in HICS are more frequent in subjects with another primary headache than in those without such a headache.
Keywords
Background
Headache attributed to ingestion or inhalation of a cold stimulus (HICS) is referenced under bullet point 4.5.2 of the 3rd edition of the International Classification of Headache Disorders (ICHD-3) (1). HICS is also known as ‘ice-cream headache’ or colloquially as ‘brain-freeze’. It is most likely caused by eating ice cream, drinking slush ice drinks, or inhaling very cold gases and is characterized by a short-lasting pain sensation in the frontal or temporal region, which may be of high intensity (1). There are some studies that examine HICS in an experimental setting, in which HICS is provoked by either eating ice cream, drinking cold water or keeping an ice cube pressed to the palate (2–10). Fewer studies explore the characteristics of HICS using questionnaires (5,11–15). All questionnaires concerning epidemiological and clinical features of HICS were conducted in subjects of grammar school age (5,13,15). In the past, subjects for experimental studies were mainly recruited from a group of participants who suffered from another primary headache disorder (3,7). More recent experimental studies used randomly recruited volunteers (9,10).
This study examines the epidemiological and clinical features of HICS including history of other primary headache in a group of healthy adolescent and adult subjects. The data were acquired using an anonymous, standardized questionnaire.
Methods
We performed a cross-sectional study using an anonymous questionnaire with 87 items in two sections. Participants were students and staff members of the Martin-Luther-University Halle-Wittenberg. Questionnaires were distributed and collected in lectures and seminars. The period of data collection was 6 months in 2015.
All items of this questionnaire were designed to evaluate the criteria for HICS given by the IHS in 2013 (ICHD-3, beta version). However, there were no relevant changes in the diagnostic criteria of HICS from ICHD-3, beta to ICHD-3. No other questionnaire was used as a template. The questionnaires were distributed after a brief oral introduction in order to educate the participants about how to check the boxes so that the forms could be analyzed electronically. The forms were collected roughly two hours after being distributed, with the possibility of a collection after one week. This was necessary to make sure every participant had adequate time to carefully think about every item of the questionnaire.
All items were designed in a multiple-choice manner to simplify an electronic evaluation. In most cases the items were designed in a dichotomous way (“yes” or “no”) to avoid possible confusion.
The first section of the questionnaire contained items concerning the participant’s age, sex, preexisting headache disorders and medication habits. The second section contained items relating to the characteristics of HICS (duration, intensity, quality, localization), circumstances of occurrence of HICS and habits of consuming chilled beverages and food as well as behavior regarding the management of preventing a probable HICS attack. A full version of the questionnaire is attached in the supplemental material.
Exclusion criteria were a failure to state sex, age and an age <17 years. No limitations regarding other primary headache disorders and regular intake of analgesics were made.
The ethics committee of the Martin-Luther-University approved the conducting of this study. Even though the data collection was anonymous, participants were able to contact the Department of Neurology via an email address provided in case of further questions or problems.
Statistical analysis
The obtained data were analyzed using IBS SPPS 23. A p-value of < 0.05 had to be attained in the statistical tests used to be considered statistically significant. To produce a p-value, Student’s t-test, χ2 test, Fisher’s exact test, Kruskal-Wallis H test and one-way ANOVA were used appropriately on differing constellations of variables. Not every individual included in this study filled out every item of the questionnaire.
Results
In total, 1213 questionnaires were distributed. There were 629 collected questionnaires (response rate 51.9%). Eleven questionnaires were invalid because of missing key details (sex, age). Thus, the final sample size was 618, with 68.9% (n = 426) females. The age range was 17–63 years (see flowchart, Figure 1).
Flowchart of study population.
Prevalence of HICS and other additional headache disorders
In total, 317 (51.3%) participants experienced HICS at least once (67.8% females). The number of participants who did not have HICS (the non-HICS group) was 301, including 211 (70.1%) females. The mean age in the HICS group was 23.3 ± 6.5 years (SD) and 24.5 ± 7.9 years (SD) in the group without HICS (p = 0.046). The flowchart in Figure 1 gives an overview of the sample population. The analysis of four age-related groups (n = 311) showed a non-significant decrease of HICS frequency with age (Figure 2).
Age-related distribution of participants with and without HICS. The percentage of HICS was not significantly different between the four groups.
The prevalence of HICS in females was 50.5%, in males 53.0% (p = 0.541). The prevalence of migraine in all participants was 16% (99/618), that of tension-type headache (TTH) was 24% (150/618). The frequency of migraine in the HICS group was 15.5% (49/317) and in the non-HICS group 17.7% (50/301). This difference was not significant. The prevalence of HICS in the migraine group was 49.5% (49/99) and in the non-migraine group 51.6% (268/519). This was not significantly different (p = 0.55). The frequency of TTH in the HICS group was 24% (76/317); and in the non-HICS group 25% (74/301). The prevalence of HICS in the TTH group was 50.6% (76/150) and in the non-TTH group 49.4% (241/468). This was not significantly different (p = 0.553). The analysis of the odds ratio (OR) in patients with migraine and HICS and TTH and HICS was 1.17 (95% CI 0.75–1.83; p = 0.496) and 1.16 (95% CI 0.80–1.70; p = 0.438) respectively.
Clinical characteristics of HICS (summarized in Table 1 and Table 2)
Intensity: The mean headache intensity in the HICS group without any other coincidental primary headache disorder (n = 186/235) was 4.58 ± 2.2 (SD) (NRS 0–10). Participants with migraine (10/235) or TTH (26/235) stated a higher intensity (mean: 5.7 ± 2.058 (SD) and 5.27 ± 2.07 (SD)) than those without any other primary headache disorder. This difference, however, was not significant.
Localization: According to the methods, the participants could select multiple locations for HICS. The most commonly affected area was the forehead with 73.6% (n = 220) followed by the temporal region with 72.5% (n = 219). Some participants stated a localization of HICS near the upper nose (26.2%; n = 76), the vertex region (20.3%; n = 58), the occiput (16.9%; n = 49) and in the region of the eyes (12.8%; n = 37). An overview is given in Figure 2. HICS was reported as unilateral in 20.1% (n = 61) and as bilateral in 36.5% (n = 111).
Duration: All participants had to decide between the four specifications in the questionnaire regarding the duration of the HICS (shorter than 10 seconds, about 15–30 seconds, longer than 1 minute, longer than 10 minutes). In 46.8% (n = 108), HICS lasted less than 10 sec. A further 45.9% (n = 106) stated that their headache completely resolved after 15–30 sec, 4.3% (n = 10) described a duration of >1 minute and only 1.3% (n = 3) reported a duration of >10 minutes. In all, 1.7% (n = 4) were not able to specify the duration of their headache with the provided items. Participants with a preexisting other primary headache disorder did not suffer from a longer duration of HICS than those without (p = 0.575; Kruskal-Wallis H test).
Quality: The most common pain quality of HICS was stabbing (51.8%; n = 162), pulling (28.1%; n = 88), dull/pressing (9.6%; n = 30) and pulsatile (3.8%; n = 12). None of these pain qualities were significantly associated with either TTH or migraine (p = 0.278; Fisher’s exact test) as co-morbid primary headaches.
Characteristics of HICS.
Discussion
Frequency
In this study, the lifetime prevalence of HICS (51.3%) is higher than in the group examined by Fuh et al. (40.6%) (13), but lower than in the recent publication by Zierz et al. (62%) (15). In both previous studies, children of school age were examined, whereas in our study adolescents and adults have been enrolled. In this study, 73.2% of the participants reported that the occurrence or the intensity of HICS was reduced by slower velocity of ingestion of cold food or beverages. It may be assumed that adults have learned to prevent HICS by consuming cold food with more caution. Therefore, the higher prevalence of HICS in the study of Zierz et al. may be attributed to the younger age of the participants. Fuh et al. conducted their study in Asia (Taiwan); the study of Zierz et al. and this study were conducted in Europe (Germany). It is known that headache disorders show a variant prevalence between different countries and regions; for example, migraine (with a prevalence of 9% in Asia compared to 15% in Europe (23)). It remains open whether those geographical differences might also be relevant for HICS. Fuh et al. reported a higher prevalence of HICS in boys (13). In the study of Zierz et al. (15) and in the present study, there was no influence of sex on the prevalence of HICS. Analyzing four age-related groups, there were no significant differences between the age groups (Figure 3). However, participants from 30–63 tended to have a slightly, but not significantly, lower prevalence than in the other groups. It remains open whether this corresponds to our previous findings showing that parents and teachers have stated a lower prevalence of HICS than young students (15).
Localization of HICS (localizations with a percentage less than 10% are not shown in the graphic); (a) face frontal, (b) face lateral.
Pain intensity
The ICHD-3 classification does not state a characteristic intensity of HICS (1). A score of ≤5 on the NRS is regarded as a mild intensity, whereas 6–7 correspond to moderate intensity. Intensities ≥8 are considered as severe (19). In our study, HICS has a mean intensity of 4.6 (SD ± 2.2; NRS 0–10) which corresponds to a mild intensity consistent to the results of Fuh et al., who reported a mild intensity of HICS in 65.5% of the cases. The mean pain intensity reported by Olivera and Valença was slightly higher (5.5 ± 2.2; NRS 0–10) (9). Mean pain intensities in participants with coexisting primary headache disorders (migraine and TTH) was slightly higher than those without a coexistent primary headache disorder (migraine: 5.7 ± 2.058, TTH: 5.27 ± 2.07). The higher HICS pain intensity in patients with additional migraine has also been described by Raskin and Knittle (21), Fuh et al. (13) and Olivera and Valença (9). Participants with a preexisting primary headache may experience a higher level of pain due to central sensitization (20).
Localization
The most frequent sites of HICS found by this study were to those previously described (3,4,7,9,10,13,21,22). Thus, temporal and frontal localizations can be regarded as typical sites for the occurrence of HICS. Occipital localization is not mentioned in the ICHD-3 classification for HICS (1). However, this localization of HICS was reported in 17.5% (13), 16.9% (present study) and was also mentioned by De Olivera and Valença (9). Thus, the occipital localization should be included in the comments of the classification criteria of HICS.
Duration
According to the current classification (ICHD-3), HICS should resolves within 10 minutes. However, the median duration of HICS in an experimental setting using ice water was 10 sec (range: 2–36 sec) and 42 sec (range: 8–125 sec) in an ice cube protocol (10). Fuh et al. reported a duration <30 sec in 72.3% of cases (13). In our study, the duration of HICS was <30 sec in 94.3% of cases. This shorter time limit should be included in the comment section of the classification criteria. However, in 1.3% (n = 3) the duration was even longer than 10 minutes.
Pain quality
In this study, most participants (51.8%; n = 162) stated a stabbing pain quality during HICS. This is in line with the results of Mages et al. (64%) (10). In all, 3.8% of the participants reported a pulsatile pain quality in this study. As a result, at least the stabbing pain quality should be part of the classification criteria of HICS.
Accompanying symptoms
In this study, 22% of the participants reported trigemino-autonomic symptoms (lacrimation, rhinorrhea and conjunctival injection) and 18.5% visual phenomena during HICS. Mages et al. also described lacrimation during HICS (10). The susceptibility for accompanying symptoms during HICS was higher when the patient suffered from another primary headache. In the present study, visual phenomena were significantly more frequent in subjects with another primary headache disorder. The described visual symptoms also typically occur during visual migraine aura. However, it remains open whether these visual symptoms in HICS and in migraine aura are due to similar mechanisms. These accompanying symptoms during HICS are not mentioned by the ICHD-3 classification (1).
Additional headache disorder
Frequency and intensity of HICS in additional headache disorders. Pain intensity was determined using the NRS ranging from 0–10 (most severe). For comparison, relevant previous studies are also shown in this table. CI: confidence interval, TTH: tension type headache.
Comparison between the established ICHD-3 criteria of HICS and key findings in our study. TAS: trigemino-autonomic symptoms; TTH: tension-type headache; n.m.: not mentioned.
Limiting factors: The evaluation of primary co-morbid headache disorders was only based on a standardized questionnaire. This was not confirmed by a personal interview. The use of headache-specific medication, especially prophylactic treatment of migraine, was not asked about in the questionnaire. This may influence both prevalence and intensity of HICS and the appearance of accompanying symptoms.
Article highlights
In adults, HICS has a prevalence of 51.3%. Accompanying symptoms during HICS occurred in 53% of the participants (22% trigeminal-autonomic symptoms and 18.5% visual phenomena). Our data expand the ICHD-3 criteria for HICS. Neither migraine nor TTH are risk factors for HICS.
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.