Abstract
Dear Sir The recently published study of Scher et al. (1) used data from the Frequent Headache Epidemiology Study to compare medication use among people classified as having chronic daily headache (CDH, 180+ headache days per year) with medication use among people with episodic headache (2–104 headache days per year). Results indicated positive associations between opioids and butalbital compounds and CDH, whereas aspirin and ibuprofen were protective. For nonprescription caffeine-containing combination analgesics, adjusted odds ratios (ORs) were 1.4 (95% confidence interval [CI], 0.9–2.2) for current use and 1.6 (95% CI, 1.0–2.7) for past use.
Scher et al. (1) interpreted their findings cautiously because of methodological limitations, including lack of data on dose-response and possible recall bias, reverse causality and confounding by indication. Furthermore, analyses only adjusted for potential confounding by age, gender, primary headache type (migraine or nonmigraine), recall year and number of medications. These variables were selected as potential confounders because they were “plausibly associated with treatment patterns” (1). However, other risk factors for CDH have been identified (2,3) (e.g. lower socioeconomic status, comorbid conditions causing pain, obesity, sleep disorders), and residual confounding is possible if these uncontrolled factors were associated with use of specific medications.
Scher et al. (1) did not explicitly discuss their results for nonprescription caffeine-containing combination analgesics, nor did they conclude that such drugs are associated causally with CDH. However, given the recent assertion by Bigal and Lipton (4) that “caffeine-containing over-the-counter products increase risk of progression” of migraine headache from an episodic to a chronic occurrence pattern, further consideration of the results for this class of analgesics reported by Scher et al. (1) is warranted.
The weakly positive results for nonprescription caffeine-containing combination analgesics reported by Scher et al. (1), taken alone, do not confirm a causal association with CDH because of the acknowledged methodological limitations of their study. Furthermore, external support for such an association from other studies is lacking. Bigal et al. (5) evaluated the relation between use of combinations containing acetaminophen, aspirin and caffeine and other medications and progression from episodic to chronic migraine, using data from the American Migraine Prevalence and Prevention (AMPP) study. In that study, opiate and barbiturate compounds were statistically significantly and positively associated with transformed migraine, but there was no association with acetaminophen/aspirin/caffeine combinations (adjusted OR, 0.87; 95% CI, 0.64, 1.19). In an earlier study of CDH in the Frequent Headache Epidemiology Study population, Scher et al. (6) reported that, after adjustment for confounders, the OR for dietary or medicinal caffeine intake was 1.36 (95% CI, 0.9–2.0; p = 0.12, not statistically significant) for high current consumption, and was 1.50 (95% CI, 1.0–2.3; p = .05, statistically significant) for high past consumption. However, the validity of these results is dubious because the observed association was restricted to younger subjects (<40 years of age) and was present in women but not men. No clear reason for this lack of internal consistency was apparent. Also, the study did not report data specifically for caffeine-containing drugs, and the results were not independent of those reported more recently by Scher et al. (1).
In summary, the results for nonprescription caffeine-containing combination analgesics reported by Scher et al. (1) are statistically weak, may not be valid and are not supported consistently by other, independent studies. The currently available epidemiological research on the possible relation between nonprescription caffeine-containing combination analgesics and CDH or headache progression does not provide a sufficient basis for definitive causal inference.
Disclosure
Dr Delzell has been a consultant for companies that manufacture nonprescription combination analgesics. Dr. Haag has been a consultant for various pharmaceutical companies.