Comparison of Multiple Epidemiologic Methods for Risk Benefit Assessment of Anti- rheumatic Drugs

[Guest guest]

Comparison of Multiple Epidemiologic Methods for Risk Benefit Assessment of

Anti- rheumatic Drugs

Category:  30 Health services research

Grant W Cannon1, Holden2, Juhaeri Juhaeri2, Wanju Dai2

1VA and University of Utah, Salt Lake City, UT;2Aventis Pharmaceuticals,

Bridgewater, NJ

Presentation Number: 150

Poster Board Number: 150

Keywords: Leflunomide, Risk benefit assessment, Epidemiologic methods

INTRODUCTION: Five methods - a meta-analysis, relative-value adjusted

number-needed-to-treat (RV-NNT), relative-value adjusted minimal clinical

efficacy (RV-MCE), proportional reporting ratio (PRR), and a cohort study -

were used to evaluate the benefit-risk profile of leflunomide (LEF) in

comparison to methotrexate (MTX) and other disease modifying antirheumatic

drugs (DMARDs) in rheumatoid arthritis (RA) patients. The purpose of this

study was to determine if consistent results would be seen in benefit-risk

assessment with these different methods.

METHODS: Meta-analysis, RV-NNT, and RV-MCE used data from large phase III

randomized clinical trials comparing LEF, MTX and sulfasalazine (SSZ).

Relative value is the value of avoiding an AE (or AEs) relative to avoiding

the disease of interest and is determined by the patient. The PRR examined

the reporting rate of specific AEs to FDA for LEF in comparison to other

DMARDs (including gold, MTX, SSZ, hydroxycholoroquine, penicillamine, and

azathioprine). The cohort analysis gave AE rates according to exposure to

LEF, MTX, other DMARDs (including tumor necrosis factor blockers), DMARD

combinations, and nonsteroidal anti-inflammatory drugs (NSAIDs) in over

40,000 RA enrolled from 9/1998 to 12/2000 in a large managed care database.

RESULTS:

The meta-analysis demonstrated a serious AE rate in year one: LEF = 14.3%,

MTX = 17.1%, SSZ 18.8% and drug related serious AE rate LEF =7.5%, MTX =

8.2%, and SSZ = 13.5%. RV-NNT is defined as NNT divided by RV-adjusted NNH

(the number of patients needed to harm, or to treat to achieve one AE). If

this ratio is less than 1, the new treatment should be considered. The

RV-NNT compared the NNT to achieve an ACR20 response to the relative-value

adjusted NNH for six AEs (hypertension, pain, infection, CV events, abnormal

liver tests, skin reactions). In the analysis, these rates (ACR20 and AE)

for LEF were compared with MTX. NNT/NNH for LEF versus MTX was 15/108 for

severe RA and 15/27 for mild RA. The RV-MCE is a modification of the RV-NNT

with adjustment for potential clinical improvement and potential AEs. Over

all levels of assumptions, the benefit-risk ratio for LEF was comparable or

better than MTX. The PRR of multiple AEs found no large difference with LEF

and other DMARDs. Specifically, for hepatic fibrosis PRR = 0.78 and for

cirrhosis PRR = 1.19. In the cohort study, AE rates per 100 person-years

(PY) were LEF=12.2, MTX = 18.9, Other DMARDs=18.89, NSAIDs alone = 40.37,

COX-2 alone = 33.78, LEF + MTX=5.32, LEF+other DMARD = 7.4, and MTX+Other

DMARD = 8.55.

CONCLUSION: These studies have used five different approaches to do

quantitative benefit-risk analysis and compare different drugs. These

analyses have demonstrated a benefit-risk profile for LEF similar to other

DMARDs currently employed in the treatment of RA.