This study evaluates the ICHD – II diagnostic criteria of the two most common primary headaches with reliable statistical parameters. The study is specific for children. The results are applicable in in- and outpatient paediatric practices.
There are variable data and limited statistical evaluation of the characteristics of migraine in the literature (6,13,16,18,19,21,22,25,26,30,34,36).
According to our results, the main pain quality in migraine was pulsating, significantly more common than in TTH. Pulsating quality had 65% Se and high Sp – 78% and AUC – 0.71. This supported the use of this item as diagnostic for migraine. Most authors also established high Se of this pain quality (6,13,16,19,25,34,36,37). Pulsating pain was reported in higher percentage (78–96%) by adult population (21,38–40) and the lower occurrence of this symptom in paediatric population could be explained by a child’s difficulties in describing the pain quality.
Unilateral location had high Sp – 88%, but low Se – 27%. Low Se of this item in children and adolescents was also established in other studies (6,16,18,19,25,37), in contrast with higher Se (53–74%) in adults (21,38–40). That is why unilateral location was confirmed as diagnostic item in criterion C for migraine in adults in ICHD – II (23). Regardless of its lower Se in paediatric age, unilateral location occurred significantly more commonly in migraine than in TTH in children and adolescents, and unilateral location was considered specific for migraine. When combining unilateral, bifrontal and bitemporal location as diagnostic item ‘migrainous location’, the Se increased to 76%, the Sp was 45% and AUC was higher than the one of unilateral location. Therefore, the changed variant of location for paediatric age as either unilateral or bifrontal/bitemporal location was a more reliable diagnostic item and our results confirmed the new variant of location in criteria of ICHD– II (23).
Several authors described high Se and Sp of the item moderate or severe intensity for migraine (6,30,33). In contrast with them, Gherpieli et al. revealed low Sp of this item (34). Turkdogan et al. pointed out that moderate intensity was the prominent one for both headache types migraine and TTH (19). According to our results, severe intensity prevailed significantly than moderate intensity in migraine, and it was highly specific for migraine.
In our study, moderate or severe intensity had high Se – 95%, Sp – 63% and AUC – 0.79. When focusing only on severe intensity the Sp increased – 84%, whereas Se remained high – 60% and AUC was close to the one of moderate or severe intensity – 0.72. Considering these results and in accordance with Pajaron and Wober-Bingol et al., we could recommend changing the content of the item of intensity as only severe intensity (30,41). In this way, moderate intensity would not overlap in the diagnostic criteria for migraine and TTH.
Pain aggravation by routine physical activity occurs in high percentage (86–96%) of migrainous adults and is considered as typical symptom for migraine (21,39,40). The role of this item in migraine diagnosis in children is controversial (22,26,30,35,42). In our study, this symptom occured significantly more commonly in migraine than in TTH. It had high Se, high Sp and AUC – 0.73, which confirmed this item as a diagnostic one. Thus, our results supported the opinion of Kroner- Herwig et al. and Smetana et al., who considered it as an important differential factor in contrast with Rossi et al. and Seshia et al. (22,26,35,42).
All associated symptoms in migraine are reported in higher percentage in adult migrainous population (82–89% for nausea, photophobia, phonophobia vs. 60–68% in our patients and 47–50% for vomiting vs. 32% in ours) (21,38,39). The lower Se was the reason for changing the criterion of associated symptoms in ICHD-II, so that photo- and phonophobia was replaced by photo- and/or phonophobia and in this way criterion D is fulfiled when at least one of the four associated symptoms occurs.
According to our results, nausea and photophobia were the associated symptoms, which had the best balance between Se and Sp, AUC above 0.70. With such AUC nausea and photophobia were proved as the most reliable diagnostic items of associated symptoms for migraine. Nausea was pointed out as a symptom with good Se and Sp by Wober-Bingol et al. as well, and they considered it as an important differential diagnostic feature between migraine and TTH (30).
Turkdogan et al. revealed low percentage of children with only one of the two symptoms – photophobia and phonophobia and suggested this item to be returned in its previous variant – both symptoms to be required simultaneously (19).
In our study, the previous item photophobia and phonophobia had Se 49%, Sp – 85% and AUC – 0.67, whereas the new version – photophobia and/or phonophobia had higher Se – 81%, Sp – 57% and higher AUC – 0.69. Our results corresponded with the data from other authors and confirmed the new content of this item in ICHD – II – photophobia and/or phonophobia (16,33).
Vertigo or lightheadedness is not included in the accepted diagnostic criteria. We found it significantly more often in migrainous patients, than in TTH patients, similar to other authors (19,25). When using multivariate logistic regression analysis vertigo or lightheadedness was not a significant factor. These results support the rationale for continuing the discussion of vertigo as an additional symptom of migraine.
According to our results of AUC, the best diagnostic items of pain characteristics for migraine were moderate or severe intensity or only severe intensity, pain aggravation by routine physical activity and pulsating quality (AUC – above 0.70). Photophobia and nausea were the most reliable diagnostic items among associated symptoms according to AUC.
Using the multivariate logistic regression analysis, pulsating pain was the diagnostic item, which increased the migraine risk most -seven-fold. The next significant items were aggravation of the pain by routine physical activity, moderate or severe intensity and nausea (they increased the migraine risk five-fold). Family history of migraine also increased the migraine risk two-fold, which supported the suggestion of Seshia, this item to be an additional one in criterion C (35).
There is limited data in the literature evaluating the symptoms, included in the diagnostic criteria for migraine of ICHD-II by multivariate analysis. Karli et al. established that nausea was the most important factor for the diagnosis of migraine – it increased the risk five-fold (similar to our result), followed by vomiting (18). According to their results, pulsating pain came third – it increased migraine risk 3.5-fold, in comparison with our analysis, where pulsating quality was the most important factor (18). Ozge et al. reported that unilateral pain had the highest OR in their logistic regression analysis – 22.6 when they used the neurologist’s diagnosis as the gold standard (6). The next symptoms with high OR in their analysis were vomiting and pulsating quality (6).
There are even less studies, which evaluate the clinical characteristics of TTH in children and adolescents, compared with such about migraine (18–20,30).
Gallai et al. demonstrated that pressing-tightening quality, bilateral location and mild or moderate intensity of the headache had higher Se for TTH (73.8%, 75.7%, 87.4%, respectively) than the corresponding criteria for migraine, but their Sp was lower, especially for bilateral location (36).
In our study, the most sensitive items for TTH – no vomiting, bilateral location (Se at least 95%) had low Sp – less than 30%.
Mild intensity had the highest Sp, but it had low Se – 43%. Combining mild and moderate intensity, the Se increased – 91%, the Sp remained above 50–52% and AUC was larger than AUC of mild intensity. Thus, the content of this item – mild or moderate intensity was ascertained. Our results were similar to other data (18–20,43). Gallai et al. even considered that mild-to-moderate intensity should be an obligatory criterion for TTH (36).
The symptom pressing-tightening quality was considered as typical for TTH and Se of this symptom in several studies was above 50% (26,43), as in adults (52–78%) (39,44,45). According to our results, only the pain quality of headache characteristics and associated symptoms differed significantly in children and adults (pressing-tightening pain was more common in adult population) (39,44,45). In our study pressing-tightening quality had high Sp, but low Se – 35% and this result was in accordance to Ozge’s data (6). Many patients could not determine the quality of the pain, as well as Seshia et Wolstein’s patients of paediatric age (35). That’s why non – pulsating quality had high Se and Sp, AUC – 0.70 and this fact confirmed the content of this item as pressing-tightening (non-pulsating) quality.
The item no pain aggravation by routine physical activity had high Se and Sp above 50% and our results supported most of the existing data in the literature (19,26,30,43).
Among associated symptoms, the most frequent one was phonophobia and our results were similar to the Ozge’s results (6) and to studies in adult population (39,44,45). However, phonophobia occurred significantly more often in migraine than in TTH.
The items of pain characteristics with AUC above 0.70, which confirmed their significance as diagnostic items, were: no aggravation by physical activity, mild or moderate intensity, non-pulsating quality. Considering the associated symptoms, the best diagnostic items according to AUC were no photophobia and no nausea.
Multivariate logistic regression analysis of the clinical characteristics of TTH showed that no photophobia, bilateral location and no nausea were the most significant items – they increased the chance for TTH approximately to the same extent – 14-fold. The next significant items were no aggravation by physical activity and non-pulsating pain.
No family history of migraine in patients with headache also increased the TTH risk three-fold, whereas no vertigo was not a significant item. That is why the lack of family history of migraine could be recommended as additional item in differentiating migraine without aura and episodic TTH with overlapping diagnostic criteria.
Using multivariate logistic regression analysis, Karli et al. established that bilateral location was the most important item for TTH (it increased the TTH risk 3.5-fold) (18). The next significant item in their study was mild or moderate intensity (two-fold risk increase) (18). However, the authors, in contrast with us, included in the analysis, the occurrence of associated symptoms, not the lack of them and thus it was impossible to evaluate the significance of items included in the diagnostic criteria, such as no nausea, no vomiting, no photophobia and no phonophobia.