Bollettino epidemiologico nazionale

New Anti-Inflammatory Agents in General Medicine

Giuseppe Traversa (1), Clara Bianchi (1), Francesca Pisetsky (1), and Mariangela Rossi (2)

(1) Laboratory of Epidemiology and Biostatistics, ISS

(2) Regional Health Authority of Umbria

During the summer of 2000, the new non steroidal anti-inflammatory drugs (NSAIDs), celecoxib and rofecoxib (subsequently referred to as COXIB) were introduced in the Italian market. These drugs act through selective inhibition of cyclo-oxygenase-2. As with other NSAIDs, COXIB are available through the National Health Service, although the cost of an average daily dose is nearly twice as high (1.4 versus 0.8 Euros).

In clinical trials, COXIB showed fewer gastrointestinal side effects than traditional NSAIDS 1,2). In particular, the COXIBs are thought to offer an advantage in the treatment of patients who require high-dose treatment over long periods (in the two trials cited, the average duration of therapy was 6-9 months). An additional potential advantage is lower co-prescription of gastroprotective drugs, especially among chronic users.

The objective of the present study is to describe the characteristics of COXIB use in the central Italian region of Umbria during the first 12 months following their introduction in Italy. Specifically, the analysis was focused on incident use of COXIB and traditional NSAIDS, with a comparison of the concomitant use of gastroprotective agents, duration of use, and the possible shift to COXIBs among chronic NSAID users.

From the regional pharmaceutical monitoring system of Umbria, which covers approximately 835,000 patients, persons on treatment with NSAIDS and gastroprotective agents between January 2000 and June 2001 were identified and their characteristics determined. Incident users were defined as those who received their first prescription of either COXIBs or NSAIDs after July 2000 (and who therefore had not been prescribed these agents for at least six months). Patients were classified by COXIB and NSAID status based on the first prescription received in the 12 months following July 2000.

For the analysis of subsequent prescriptions to the incident use, a period of current use was defined as the sum of the daily defined doses (DDD; the number of doses necessary to cover a day of therapy in an adult for each medication) over a standard 14 day period. In order to ensure a period of observation of at least a month, only those patients who received a prescription between July 2000 and May 2001 were included. Age, sex, and prior prescriptions of gastroprotective agents (used to identify those at greater risk of gastroduodenal lesions) were considered as possible counfounding variables.

In Italy, COXIB and NSAIDS represent 26% of all drug doses and slightly less than 40% of all pharmaceutical expenditures in Italy (Table 1). An estimation of the prevalence can be calculated using the population of a specific region. In Umbria, more than 200,000 of the 835,000 residents received at least one anti-inflammatory prescription over a one-year period (41,833 of whom received COXIB and 187,377 traditional NSAIDS). A total of 15.3% of the population were incident users (2.7% COXIB and 12.6% for other NSAIDS). The incident users of COXIBs were older than those receiving NSAIDs and were more likely to be female and to have received prescriptions for gastroprotective agents in the previous 6 months (Table 2).

The number of patients receiving concomitant prescriptions for gastroprotective agents was 5.0% for the COXIB and 5.8% for the traditional NSAIDS. After adjustment for confounding factors, the incident users of COXIB demonstrate a 25% lower probability of receiving a concomitant gastroprotective drugs than those using other NSAIDs (OR 0.75; 95% confidence intervals 0.7-0.8). However, limiting the analysis to those gastroprotective agents that have been shown to be protective against damage caused by NSAIDs (acid-pump inhibitors, prostaglandin, and anti-H2), the difference in the probability of co-prescription with COXIB and traditional NSAIDs (2.9% versus 2.5%) is no longer observed when possible confounders are taken into account.

Only an estimated 10% of incident users received a new prescription of one of the NSAIDs during the study period (10.9% for traditional NSAIDs and 9.2% for the COXIB). Among the users of traditional NSAIDs that continued treatment, 90.2% received the same drug, while for the COXIBs, only 50% did so.

The introduction of the COXIB resulted in a notable increase in costs for the National Health Service. In the first year in which they were available, the COXIBs represented 89% of the expenditures for NSAIDs. In Umbria, where their use was similar to that observed nationally, the profile of users of COXIBs and NSAIDs did not differ substantially. In both cases, use was predominantly acute and there was only a slight difference in concomitant use of gastroprotective drugs. The fact that only half of those who received a first COXIB prescription were apparently continued on the drug (50% versus 90% for traditional NSAIDs) may represent a lower efficacy of these drugs within the context of general medical practice and merits further study.

Although the higher daily cost of COXIB could be justified if they provided greater safety among chronic users and a potential savings in the concomitant use of gastroprotective agents, their current use in Italy does not justify the difference in cost of the two agents.

Comment

Nicola Magrini

CEVAS (Cednter for the Evaluation of the Effectiveness of Health Services, Modena

Traversa’s work represents an excellent example of a study to evaluate the patterns and appropriateness of the use of a new category of NSAIDs, the COXIB, and presents an interesting research hypothesis regarding their real clinical efficacy.

The results of the study are very interesting and in some ways unexpected, even though they have also been confirmed using a second Italian data base maintained by the province of Modena (Emilia-Romagna). The prescription of gastroprotective agents was similar between COXIB and traditional NSAID users even though such use is not routinely recommended for COXIB users and goes against the hypothesis that the costs of the COXIB would be counterbalanced by lesser prescription of gastroprotective agents. An aspect that merits further analysis is whether the patients treated with COXIB are individuals at greater risk of gastric pathology who were not considered candidates for traditional NSAIDs, but whose physicians prescribed a trial of COXIB with concomitant gastroprotective agents given as a precautionary measure. To further investigate this issue, it would be interesting to evaluate whether patients given COXIBs had a higher prevalence of gastric problems or of having been prescribed anti-ulcer drugs in previous years (the present study looked at use of gastroprotective agents only for the six-month period prior to receiving NSAIDs).

An additional finding that emerges from the study is the greater apparent use of NSAIDs and COXIBs for acute problems in Italy compared with the United States, where they are used predominantly in chronic conditions. Additionally, Traversa’s study shows that 90% of those who have received a traditional NSAID received a second prescription, while only half of those receiving COXIB did so. This greater tendency to change analgesic (pain represents the main indicator for use) could be a non-optimal response to COXIBs compared with traditional NSAIDs within the context of clinical practice and merits further evaluation in randomized studies conducted in such settings.

Finally, it is worthwhile reading the completed analysis (as described in the protocol) of the VIGOR (3) and CLASS (4) studies recently conducted by the United States Food and Drug Administration that have shown only a minor difference between traditional NSAIDS and COXIB with respect to gastric damage (after 13 months of treatment with respect to the 6 months reported in published studies) and a higher incidence of infarcts in the VIGOR study in patients treated with rofecoxib. It may also be of interest for health authorities in Italy to read the severe warnings that go beyond those issued by the manufacturers of these drugs.

References

1.Bombardier C, Laine L, Reicin A, et al. N Engl J Med 2000; 343: 1520-8.

2.Silverstein FE, Faich G, Goldstein JL, et al. JAMA 2000; 284:1247-55.

3.FDA. Vioxx (rofecoxib):
 www.fda.gov/ohrms/dockets/ac/01/briefing/3677b2_03_med.doc

4.FDA. Celebrex (celecoxib): www.fda/gov/ohrms/dockets/ac/01/briefing/3677b1.htm

5.Warning Rofecoxib: www.fda.gov/foi/warning_letters/g1751d.pdf

6.Warning Celecoxib: www.fda.gov/foi/warning_letters/m5097n.pdf