Skip to main page content

• HOME
• CURRENT ISSUE
• PAST ISSUES
• MOBILE
• ABOUT
• ACP
• SUBSCRIPTIONS
• E-MAIL ALERTS
• SUBMIT

RSS Feeds

1. Understanding Results of Aspirin dose and Plavix with Risk of Bleeding presented by CHARISHMA trial

   • Azhar Supariwala, MD
   • Brenda Matti, MD

   St.Luke's-Roosevelt Hospital Center, Columbia University College of Physicians and Surgeons,New York

   This letter is in reference to the recent article by the CHARISHMA investigators “Aspirin to Prevent Cardiovascular Disease: The Association of ASA Dose and Plavix With Thrombosis and Bleeding” Annals of Internal Medicine, March 2009, Vol. 150.

   It was non-randomized study. Baseline characteristics were very different in the 3 groups with more than 20 variables in the model that may skew results. The univariate analyses shows that the risk of severe or life threatening bleeding with an absolute risk reduction (ARR) of 0.9 % in people who took plavix and aspirin (ASA) >100mg. The number need to harm (NNH) is 111 (1/ARR) for a follow-up period of 2 ½ years.

   	Plavix & Aspirin <_100mg column="column" a="a" th="th">	Plavix & Aspirin >100mg (%)Column B	ARR (%) Column C [B-A]	NNH for ASA >100mg Column D = 1/C	NNH for ASA <_100mg column="column" e="1/C</th">
   Incidence of severe or life threatening bleeding. Univariate	1.7	2.6	0.9	111	---
   Incidence of severe or life threatening bleeding. Multivariate adjusted for >20 factors HR=1.3 (95% CI 0.83 – 2.04) p not significant >0.05	1.7	2.21	0.51	196	---
   Upper Limit of HR 2	1.7	3.4	1.7	58	---
   Lower limit of HR 0.83	1.7	1.4	-0.3	---	333

   In individuals with plavix and ASA > 100 mg, the multivariate analyses of this post-hoc retrospective case-control study with significantly different baseline characteristics adjusted for more than 20 covariates is statistically not significant. A point estimate of 1.3 for HR with a large CI (95% CI; 0.83 – 2.04). Incidence of severe or life threatening bleeding being 1.7 in group with ASA <_100 mg="mg" _30="_30" of="of" hr="hr" makes="makes" the="the" arr="arr" _0.51.="_0.51." this="this" nnh="nnh" _196="_196" _1="_1" not="not" ignoring="ignoring" large="large" ci="ci" and="and" upper="upper" limit="limit" that="that" is="is" _2.04="_2.04" _1.7="_1.7" _58="_58" for="for" approx="approx" _2="_2" _="_" years.="years." investigators="investigators" mention="mention" many="many" people="people" take="take" asa="asa" hence="hence" a="a" small="small" risk="risk" effect="effect" size.="size." population="population" size="size" on="on" plavix="plavix" smaller="smaller" than="than" aspirin="aspirin" only.="only." other="other" hand="hand" it="it" shows="shows" _10="_10" reduced="reduced" harm="harm" in="in" individuals="individuals" taking="taking" _100mg="_100mg" vs.="vs." _0.90="_0.90" _95="_95" _0.61="_0.61" _1.6="_1.6" statistically="statistically" significant="significant" but="but" stressed="stressed" upon.="upon." suggests="suggests" did="did" have="have" dose-response-effect="dose-response-effect" relationship.="relationship." similarly="similarly" group="group" without="without" has="has" _29="_29" severe="severe" or="or" life="life" threatening="threatening" bleeding="bleeding">100mg ASA group has 20% reduced risk as compared to <_100 mg="mg" aspirin="aspirin" statistically="statistically" not="not" significant.="significant." p="p">

   If we see the point estimate HR of 1.3 (95% CI, 0.83 – 2.04) with this large CI, we are 95% confident that the results lie between this CI (even after accepting the case-control study and >20 factors adjustments). Where it lies, we do not know. If we do a similar study, the chance of having reduced risk for bleeding with ASA >100 mg (let’s say HR 0.9) is the same as the chance of increased risk (let’s say HR 1.3. If the authors proposed a null hypothesis that “ There is no significant difference in the risk of severe or life threatening bleeding in individuals taking plavix with ASA <100mg and ASA >100mg,” we will accept the null hypothesis with these results.

   With the results of this study, we conclude that the possibility of harm provided by ASA in dose of >100 mg with plavix is the same as possibility of reduced harm. A randomize study is needed to make the estimate more precise. Finally, considering both the safety and efficacy, ASA of 100 mg faired better than both <_100 mg="mg" and="and">100 mg in individuals with and without plavix with statistically not significant p value. The CURRENT-OASIS 7 (Clopidogrel Optimal Loading Dose Usage to Reduce Recurrent Events and Optimal Antiplatelet Strategy for Interventions) will give concrete results.

   Conflict of Interest:

   None declared

   Submit response
   Published July 28, 2009