Abstract
Dear Sir We are pleased that Dr Steiner (1) shares our evaluation that the results of this trial (2) have important implications for clinical practice. However, we do not share Dr Steiner's opinion that the use of fixed-combination analgesics is still controversial. The evidence-based evaluations of the US Headache Consortium and the German Migraine and Headache Society considered in their most recent evidence-based recommendations the fixed combination of acetylsalicylic acid (ASA), paracetamol (PAR) and caffeine (CAF) as a treatment of first choice for the treatment of migraine (3) and for migraine and tension-type headache under self-medication (4). Our study verified that this combination of ASA, PAR and CAF can effectively and safely treat the pain and disability related to episodic headaches and that it is significantly superior to monotherapy and the combination of ASA and PAR. Recently, this has once again been confirmed by the latest Therapeutic Guidelines of the German Neurological Society (DGN) in cooperation with German Migraine and Headache Society (5, 6) and the guideline of the European Federation of Neurological Societies (7).
Dr Steiner's assumption that the side-effects of the constituents would summate has definitely not been confirmed by the use of this combination in Europe and the USA over decades. The monographs of the German regulatory authorities stated explicitly that combinations of ASA, PAR and CAF do not imply any increased risks of side-effects (8–10).
The remaining points raised by Dr Steiner are answered as follows:
Dr Steiner's allegation of selection bias in our study is not correct. The study included patients who usually self-treat their primary headache (migraine and episodic tension-type headache) successfully with over-the-counter (OTC) analgesics. At the time of the development of the study design nearly all ASA-containing effervescent tablets contained vitamin C. These preparations were clearly profiled by the manufacturers (in particular also by Bayer Health Care) for the treatment of common colds and were used by the patients accordingly, which becomes evident form the increase in the market share of effervescent tablets in periods with increased incidence of colds. Especially for this reason, the preferred use of effervescent tablets was an exclusion criterion in our study in order to avoid bias by indication (inclusion of patients with secondary headaches).
The assumption by Dr Steiner that ASA in a formulation as effervescent tablet would have advantages over solid tablets lacks scientific evidence. None of the three studies which used ASA effervescent tablets (11–13) for migraine provides any data in this respect. The only study investigating ASA solid vs. ASA effervescent tablets in headache patients showed no significant difference between the two formulations (14).
Dr Steiner's assertion that the effervescent formulation for ASA was the widely preferred formulation in Germany can hold true only on the basis of the sales figures due to the clearly higher prices for effervescent tablets. However, the sales figures are completely irrelevant for determining the preference of the patients for a certain formulation. Due to the different package sizes, not even the number of sold packages can be used in answering this question, but only the number of tablets (counting units) of the respective formulations or the overall quantity of the active substance used in this formulation can be compared. According to the ‘IMS off take database 2005’, among the 1158.5 million counting units of ASA sold by pharmacies in Germany in 1998—the year of development of the study design—only 325.6 million counting units were effervescent tablets, which is 28.1% [compared with 832.9 million counting units (71.9%)] of the solid formulation).
Dr Steiner questioned the calculation of the primary end-point of time to 50% pain relief by linear interpolation (not extrapolation as stated erroneously by Dr Steiner). We deliberately avoided asking the patients to document this time directly, or something vague such as the time to meaningful pain relief, as this incorporates some subjective judgement and would be a less rigorous end-point. Instead, we chose the widely used and validated instrument of pain intensity assessment on a visual analogue scale and we calculated the time to 50% pain relief from the data received. Linear interpolation is by far the most common method to interpolate between adjacent observation time points. It has the fewest assumptions compared with other models. There is no rationale for choosing a more complex modelling technique. The time course of the pain intensity was overall decreasing after treatment with active drugs, also on an individual patient level. The method of linear interpolation is used in most publications to display the time course of pain intensity. The corresponding curves observed in our studies and displayed in the figures coincide well with those reported in the literature on the same treatments.
We did not present responder rates, where, for example, response has been defined as pain reduction by 50%, as these data correlated very well with our primary end-point and would not have provided additional information. In fact, the triple combination was superior to all other treatments in the percentage of patients, with 50% pain reduction at all time points of pain intensity assessment after intake of the medication. For example, 48, 44, 40, 43, 32 and 21% of patients 1 h after taking the triple combination of ASA, PAR and CAF, the combination without caffeine, monotherapy ASA, monotherapy PAR, caffeine alone, and placebo and 77, 69, 71, 65, 53 and 47% 2 h after taking the corresponding medications had at least a 50% pain reduction.
The pharmaceutical quality of all supplies of study medication in this clinical trial met the specifications of the manufacturers, including the respective specifications for the in vitro release of the active substances. For all preparations examined in vivo, C max and T max were within the ranges for standard preparations.
Dr Tfelt-Hansen raises two issues in his letter which we would like to answer as follows.
He confirmed that treatment differences observed in this thoroughly planned, well-conducted and comprehensively analysed clinical trial convincingly show that the triple combination of ASA, PAR and CAF is significantly more effective than the combination without caffeine and more effective than either of these drugs administered as monotherapy, including ASA.
He questions to some extent the degree of superiority. Our analyses show that faster median time to pain relief was associated with a better rating of efficacy by the patients. Being on average at least 30 min earlier pain free after treatment with the triple combination compared with all other treatments tested was obviously perceived as meaningful by the patients. These findings are in agreement with surveys in patients taking OTC analgesics that a difference of 5 min would be sufficient to say that one drug begins to provide relief faster than another drug (15).
Medication overuse headache (MOH), the second issue addressed by Dr Tfelt-Hansen, is one of the central problems with which the therapy of headache conditions is faced. The most important fact is that all and any analgesics and migraine remedies (ergotamine, ergotamine derivatives and triptans)—independent of their active substances, whether as single preparations or in combinations, independent of their prescription or OTC status—may cause MOH if used too frequently.
Concerning the chapter on MOH in The Headaches, it should be pointed out that four of the six studies investigating the incidence of MOH with various drugs do not provide any data on the combination of ASA, PAR and CAF. In the study by Baumgartner et al. (16), all patients with drug-induced headache took caffeine-containing preparations (100%), but the latter also contained ergotamine (80%), barbiturates (62%), codeine (26%) and tranquilizers (26%). In order to draw qualitative conclusions for certain combination medications, these preparations would have had to be evaluated individually, and not their active substances. The study by Diener et al. (17), which compared 108 patients with intermittent migraine with 90 patients with drug-induced chronic daily headache, showed that with migraine patients the most frequent but one preparation among the 12 most frequently used was Thomapyrin, i.e. the combination of ASA, PAR and CAF, which, however, was not found among the 12 preparations used most frequently by patients with drug-induced chronic daily headache. This does not at all support any particular risk of a combination for the development of MOH.
Since no cohort studies, case–control studies or thoroughly planned cross-sectional trials are available on the quantitative differences of MOH risks specifically for OTC analgesics and new studies such as the Vaga study (18) have yielded new findings concerning caffeine that support the earlier findings obtained under very artificial conditions (19), differentiating recommendations on the specific combination of ASA, PAR and CAF compared with the classical OTC single analgesics do not seem justified. It is much more important to give the patients the rule of thumb of ‘not more often than on 10 days per month and not longer than on three consecutive days’ for any migraine and pain remedies in the acute therapy of headaches.
There is no doubt that our study shifts the benefit-to-risk ratio for the fixed combination of ASA, PAR and CAF even further towards benefit. Also, the therapeutic recommendations and therapy guidelines (3–7) already mentioned above recommend the combination of ASA, PAR and CAF uniformly with the respective best evaluation category (if done so, explicitly as ‘remedy of first choice’). As far as this specific combination is concerned, the advice given to patients should indeed change.