Merck Announces Preliminary Analyses of Off-Drug Extension of APPROVe Study

In the off-drug follow-up period for patients in theAPPROVe study, there was not a statistically significant difference inthe risk of confirmed thrombotic cardiovascular events in patients whohad previously taken VIOXX compared to those who had previously takenplacebo, according to preliminary analyses announced today by thestudy sponsor, Merck & Co., Inc. This prespecified analysis includedpatients regardless of when they discontinued study therapy.Furthermore, in the one-year off-drug follow-up period for patientswho completed approximately three years of therapy in the APPROVestudy, there was not a statistically significant difference in therisk of confirmed thrombotic cardiovascular events in patients who hadpreviously taken VIOXX compared to those who had previously takenplacebo. In these analyses, the data were insufficient to concludethat there was an increased relative risk of confirmed thromboticcardiovascular events following discontinuation of therapy. In theprespecified primary analysis of each patient's four-year data (thatcombined data from the on-drug period and the off-drug periodregardless of when patients discontinued study therapy) the differencein the risk of confirmed thrombotic cardiovascular events betweengroups initially observed in the on-drug period of the study remainedstatistically significant.

In the four-year data, there was an increased relative risk ofconfirmed heart attacks in the VIOXX group compared to the placebogroup and an increased relative risk of confirmed ischemic strokes inthe VIOXX group compared to the placebo group. Mortality was similarbetween the VIOXX and placebo groups in the four-year data. Thesepreliminary analyses have been shared with regulatory agencies.

"Our preliminary analyses of the off-drug period did notdemonstrate a statistically significant increased risk of confirmedcardiovascular thrombotic events after patients in the APPROVe studystopped taking VIOXX," said Peter S. Kim, Ph.D., president of MerckResearch Laboratories. "The limited data in the APPROVe study onstroke have to be interpreted in the context of the extensive data wehave previously published, which consistently showed no increased riskof strokes in patients taking VIOXX."

The Company reported the results of the three-year APPROVe study(the "base study") in September 2004. As previously reported, in thebase study, there was an increased relative risk for confirmedthrombotic cardiovascular events, such as heart attack and stroke,beginning after 18 months of treatment in the patients taking VIOXXcompared to those taking placebo. Given the questions raised by thedata in the base study and the perceptions regarding alternativetherapies available, the Company decided to voluntarily withdraw VIOXXworldwide at that time. In accordance with the APPROVe clinical studyprotocol, the Company announced that it planned to follow patients forone year after they came off treatment. The results announced todayare the preliminary safety analyses including this follow-up period.

A copy of the summary report of confirmed thromboticcardiovascular events from the APPROVe off-drug extension, which wasshared with regulatory agencies, is attached and available byvisiting: The VIOXX Information Center on www.merck.com.

About the APPROVe Study

APPROVe (Adenomatous Polyp Prevention on VIOXX) was amulti-center, randomized, placebo-controlled, double-blind studydesigned to evaluate the efficacy of 156 weeks (three years) oftreatment with VIOXX 25 mg in preventing recurrence of colorectalpolyps in patients with a history of colorectal adenomas. The one-yearoff-drug extension of APPROVe addressed recurrence of polyps,thrombotic cardiovascular events and mortality. Recurrence of polypswill be addressed in a future publication.

About Merck

Merck & Co., Inc. is a global research-driven pharmaceuticalcompany dedicated to putting patients first. Established in 1891,Merck currently discovers, develops, manufactures and markets vaccinesand medicines to address unmet medical needs. The Company devotesextensive efforts to increase access to medicines through far-reachingprograms that not only donate Merck medicines but help deliver them tothe people who need them. Merck also publishes unbiased healthinformation as a not-for-profit service. For more information, visitwww.merck.com.

Forward-Looking Statement

This press release (including the attachment) contains"forward-looking statements" as that term is defined in the PrivateSecurities Litigation Reform Act of 1995. These statements are basedon management's current expectations and involve risks anduncertainties, which

may cause results to differ materially from those set forth in thestatements. The forward-looking statements may include statementsregarding product development, product potential or financialperformance. No forward-looking statement can be guaranteed, andactual results may differ materially from those projected. Merckundertakes no obligation to publicly update any forward-lookingstatement, whether as a result of new information, future events, orotherwise. Forward-looking statements in this press release should beevaluated together with the many uncertainties that affect Merck'sbusiness, particularly those mentioned in the cautionary statements inItem 1 of Merck's Form 10-K for the year ended Dec. 31, 2005, and inits periodic reports on Form 10-Q and Form 8-K, which the Companyincorporates by reference.

APPROVe Off-Drug Extension

Preliminary Analyses of Thrombotic Cardiovascular Safety

APPROVe was designed to evaluate the efficacy of rofecoxib 25 mgin preventing recurrence of colorectal polyps in patients with ahistory of colorectal adenomas. As previously reported in the 156-weekbase study, there was an increased relative risk for confirmedthrombotic cardiovascular events beginning after 18 months oftreatment in the patients taking rofecoxib compared to those takingplacebo. The APPROVe protocol called for a one-year off-drug extensionto address recurrence of polyps, confirmed thrombotic cardiovascularevents and mortality. This report will provide data on confirmedthrombotic cardiovascular events and mortality from the off-drugextension. Recurrence of polyps will be addressed in a separatereport.

Confirmed Thrombotic Cardiovascular Events and Mortality

Statistical Approach to the Data

The primary analysis for the on-drug base study included data forall 2587(1) patients through day 14 post-discontinuation of studytherapy.

The primary analysis for the off-drug extension used an intentionto treat (ITT) approach to data analysis and thus included data bothon study therapy and post-discontinuation.

The secondary analysis for the off-drug extension considered onlythe data from each patient's off-drug period (that is, data startingon day 15 post-discontinuation of study therapy) and did not considerevents during the on-drug period.

Additional analyses evaluated data from the off-drug period forsubgroups of patients who had different lengths of on-drug exposurebefore discontinuing from the base study. These analyses explored therelative risks of events after stopping therapy in patients with up to6 months, 18 months, and approximately 3 years on-drug experience.

Three approaches were taken to the primary and secondary analysesbased on different censoring rules. There were 2 prespecified rules:1) data censored at Week 210, the time at which a patient would havefinished the base study plus one year follow-up, and 2) data censoredat an October 31, 2005 cut-off date, which would represent anapproximately 1- year follow-up after the last patient's final visitin the base study. This provided data in some patients out as far asweek 299. A third approach included all information before and afterthe October 31, 2005 analysis cut-off date that was available in-houseas of March 15, 2006(2). This provided data in some patients out asfar as week 306.

The primary endpoint was confirmed thrombotic cardiovascularevents. The Anti-Platelet Trialists' Collaboration (APTC) endpoint wassecondary.

In general, the results using the 3 censoring approaches weresimilar. The report will focus on confirmed thrombotic cardiovascularevent data through Week 210 and highlight differences using the otherapproaches and other endpoints.

ITT Analysis, Confirmed Thrombotic Cardiovascular Events

Among the 2,587 patients enrolled in the APPROVe study, confirmedthrombotic cardiovascular event data for the period beginning on day15 after discontinuation of study therapy were available for 2,178.

The APPROVe off-drug extension provided additional data to whathad been obtained in the base study. A total of 74(3) patients hadconfirmed thrombotic cardiovascular events in the base study. Therewere 39 additional patients with a first confirmed thromboticcardiovascular event in data through week 210 and a further 9 patientswith a first confirmed thrombotic cardiovascular event in data throughOctober 31, 2005. There were 5 additional patients with firstconfirmed thrombotic cardiovascular events after the October 31, 2005cut-off date.

The relative risk of patients having confirmed thromboticcardiovascular events in the 156-week on-drug base study was 1.92 (95%CI 1.19, 3.11, p=0.008). The difference remained significant afterincluding data from the off-drug follow-up period. In the ITT analysisthrough week 210 the relative risk was 1.74 (95% CI 1.19, 2.55,p=0.004).

In the 156-week on-drug base study, the relative risk of patientshaving confirmed thrombotic cardiovascular events was not constantover time: 1.18 (95% CI 0.64, 2.15) for the first 18 months and 4.45(95% CI 1.77, 13.32) for the period beyond month 18. In the ITTanalysis up to Week 210, the relative risk of confirmed thromboticcardiovascular events for the first 18 months was 1.27 (95% CI 0.71,2.25). The relative risk beyond month 18 was 2.22 (95% CI 1.32, 3.73).In comparing these datasets, the relative risks for the first 18months are similar. However, in the period beyond month 18, therelative risk for rofecoxib compared to placebo in the ITT analysis of210-week data is approximately half what had been observed in theon-drug analysis of the 156-week data.

Between-group differences in myocardial infarction were observedboth in the base study and in ITT analyses of the base study plusoff-drug follow-up data. In the ITT analysis up to Week 210, therewere 31 patients with myocardial infarction in the rofecoxib group and15 in the placebo group (p=0.017). Between-group differences inischemic cerebrovascular accidents were observed in the ITT analyses.In the ITT analysis up to Week 210, there were 17 patients withischemic cerebrovascular accidents in the rofecoxib group and 6 in theplacebo group (p=0.024).

Off-Drug Extension data, Confirmed Thrombotic CardiovascularEvents

The off-drug extension secondary analyses considered only the datafrom each patient's off-drug period (that is, data starting on day 15post-discontinuation of study therapy) whether or not the patient hadan event in the on-drug period. This was considered a conservativeapproach to the analysis. There were 44 patients with confirmedthrombotic cardiovascular events in patients with follow-up datathrough week 210. Of these, 5 had a first confirmed thromboticcardiovascular event in the base study (4 in the rofecoxib group and 1in the placebo group). There were an additional 9 patients withconfirmed thrombotic cardiovascular events in patients with follow-updata through October 31, 2005.

There were 28 patients through week 210 with confirmed thromboticcardiovascular events in the group previously on rofecoxib compared to16 in the group previously on placebo, a difference that was notstatistically significant (relative risk 1.64, 95% CI 0.89, 3.04,p=0.115).

The numeric difference between groups in confirmed thromboticcardiovascular events observed during the off-drug period was mainlydue to patients who experienced a confirmed ischemic cerebrovascularaccident. In data through week 210, there were 7 patients in the grouppreviously on rofecoxib and 0 in the group previously on placebo withconfirmed ischemic cerebrovascular accident (p=0.022). In data throughOctober 31, 2005, there was 1 additional patient in the grouppreviously on rofecoxib and 0 in the group previously on placebo withconfirmed ischemic cerebrovascular accident (p=0.010); in data beyondthe October 31, 2005 cut-off date, there were 2 additional patients inthe group previously on placebo with confirmed ischemiccerebrovascular accident. Including all available data, there were atotal of 8 patients in the group previously on rofecoxib and 2 in thegroup previously on placebo with confirmed ischemic cerebrovascularaccident (p=0.134).

Analyses of off-drug data in patients with different duration oftreatment in the base study

An analysis of the off-drug extension was performed for thosepatients who completed the on-drug base study(4). This analysis thusexplored the relative risks of events after stopping therapy inpatients with the longest on-drug experience.

Among the 2,587 patients enrolled in the APPROVe study there were1857 patients who completed the on-drug base study. Off-drug extensiondata were available for 1721 (93%) patients. There were 15 patientsthrough week 210 with confirmed thrombotic cardiovascular events inthe group previously on rofecoxib compared to 9 in the grouppreviously on placebo (RR=1.85, 95% CI 0.81, 4.22, p=0.146). This isnotably less than the relative risk for patients in the on-drug basestudy during the interval from month 19 through 36 (RR=4.45, 95% CI1.77, 13.32).

Analyses for the off-drug extension were also performed for thosepatients who completed: 1) 6 months or less of the on-drug base study,and 2) 18 months or less of the on-drug base study. These analysesthus explored the relative risks of events after stopping therapy inpatients with shorter on-drug experience.

Off-drug extension data were available for 99 patients in therofecoxib group who completed 6 months or less of the on-drug basestudy and 53 patients in the placebo group. There were 6 patientsthrough week 210 with confirmed thrombotic cardiovascular events inthe group previously on rofecoxib compared to 3 in the grouppreviously on placebo (RR=1.10, 95% CI 0.23, 6.77).

Off-drug extension data were available for 174 patients in therofecoxib group who completed 18 months or less of the on-drug basestudy and 118 patients in the placebo group. There were 8 patientsthrough week 210 with confirmed thrombotic cardiovascular events inthe group previously on rofecoxib compared to 5 in the grouppreviously on placebo (RR=1.04, 95% CI 0.34, 3.19).

Mortality

Among the 2,587 patients enrolled in the APPROVe study, mortalitydata for the period beginning on day 15 after discontinuation of studytherapy were available for 2,448. Mortality rates were similar betweengroups in all analyses performed.

Conclusions

-- For the first 18 months, the relative risk of confirmed thrombotic cardiovascular events for rofecoxib compared to placebo in the ITT analysis was similar to that observed in the on-drug base study.

-- Beyond 18 months, the relative risk of confirmed thrombotic cardiovascular events for rofecoxib compared to placebo in the ITT analysis was approximately half of that observed in the on-drug base study. This lowering of the relative risk was mostly due to the off-drug follow-up data observed during the period beyond month 36 and to the off-drug follow-up data in patients who prematurely discontinued study therapy.

-- There was no statistically significant increased relative risk of confirmed thrombotic cardiovascular events for rofecoxib compared to placebo in off-drug follow-up data in all patients regardless of when they discontinued therapy as well as in those who completed 150 weeks of on-drug treatment in the base study. In these analyses, data are insufficient to conclude that there was an increased relative risk following discontinuation of therapy.

-- Mortality was similar between groups.(1) One patient, AN 91495, previously thought to have been randomized to rofecoxib 50 mg, had in fact been randomized to placebo. Thispatient discontinued therapy after 37 days without having had an AE and without colonoscopy.(2) Follow-up data remain outstanding from 1 investigator site that had randomized 17 patients.(3) MRL learned of thrombotic events in 2 patients in the studysubsequent to finalization of the APPROVe clinical study report on15-Mar-2005. AN 91133, rofecoxib group, had confirmed myocardialinfarction on relday 684. AN 90052, placebo group, had confirmedperipheral artery occlusion on relday 784.(4) Patients were considered to have completed the base study if theywere on drug for at least 150 weeks.