<!DOCTYPE article PUBLIC "-//NLM//DTD JATS (Z39.96) Journal Publishing DTD v1.1 20151215//EN" "https://jats.nlm.nih.gov/publishing/1.1/JATS-journalpublishing1.dtd">
<article xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" article-type="other" dtd-version="1.0" specific-use="sps-1.3" xml:lang="en">
  <front>
    <journal-meta>
      <journal-id journal-id-type="nlm-ta">Arq Bras Cardiol</journal-id>
      <journal-id journal-id-type="publisher-id">abc</journal-id>
      <journal-title-group>
        <journal-title>Arquivos Brasileiros de Cardiologia</journal-title>
        <abbrev-journal-title abbrev-type="publisher">Arq. Bras. Cardiol.</abbrev-journal-title>
      </journal-title-group>
      <issn pub-type="ppub">0066-782X</issn>
      <issn pub-type="epub">1678-4170</issn>
      <publisher>
        <publisher-name>Sociedade Brasileira de Cardiologia - SBC</publisher-name>
      </publisher>
    </journal-meta>
    <article-meta>
      <article-id pub-id-type="publisher-id" specific-use="scielo-v3">9wBykJvH3XSfTjxV5Gjrvzr</article-id>
      <article-id pub-id-type="publisher-id" specific-use="scielo-v2">S0066-782X2016000300247</article-id>
      <article-id pub-id-type="doi">10.5935/abc.20160033</article-id>
      <article-categories>
        <subj-group subj-group-type="heading">
          <subject>Viewpoint</subject>
        </subj-group>
      </article-categories>
      <title-group>
        <article-title>Ezetimibe: Clinical and Scientific Meaning of the IMPROVE-IT Study</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <name>
            <surname>Correia</surname>
            <given-names>Luis C. L.</given-names>
          </name>
          <xref ref-type="aff" rid="aff1"/>
          <xref ref-type="corresp" rid="c1"/>
        </contrib>
        <aff id="aff1">
          <institution content-type="orgname">Escola Bahiana de Medicina e Sa&#250;de P&#250;blica</institution>
          <institution content-type="orgdiv1">Hospital S&#227;o Rafael</institution>
          <addr-line>
            <named-content content-type="city">Salvador</named-content>
            <named-content content-type="state">BA</named-content>
          </addr-line>
          <country country="BR">Brazil</country>
          <institution content-type="original">Escola Bahiana de Medicina e Sa&#250;de P&#250;blica; Hospital S&#227;o Rafael, Salvador, BA - Brazil</institution>
          <institution content-type="normalized">Escola Bahiana de Medicina e Sa&#250;de P&#250;blica</institution>
        </aff>
      </contrib-group>
      <author-notes>
        <corresp id="c1"><bold>Mailing Address: Luis Cl&#225;udio Lemos Correia</bold>, Escola Baiana de Medicina - P&#243;s-gradua&#231;&#227;o. Av. Princesa Leopoldina 19/402, Gra&#231;a. Postal Code 40150080, Salvador, BA - Brazil. E-mail: <email>lccorreia@cardiol.br</email>
				</corresp>
        <fn fn-type="con">
          <p>
            <bold>Author contributions</bold>
          </p>
          <p>Conception and design of the research, Acquisition of data, Analysis and interpretation of the data, Statistical analysis, Obtaining financing, Writing of the manuscript and Critical revision of the manuscript for intellectual content: Correia LCL.</p>
        </fn>
        <fn fn-type="conflict">
          <p>
            <bold>Potential Conflict of Interest</bold>
          </p>
          <p>No potential conflict of interest relevant to this article was reported.</p>
        </fn>
      </author-notes>
      <pub-date pub-type="epub-ppub">
        <month>3</month>
        <year>2016</year>
      </pub-date>
      <volume>106</volume>
      <issue>3</issue>
      <fpage>247</fpage>
      <lpage>249</lpage>
      <history>
        <date date-type="received">
          <day>20</day>
          <month>08</month>
          <year>2015</year>
        </date>
        <date date-type="rev-recd">
          <day>11</day>
          <month>11</month>
          <year>2015</year>
        </date>
        <date date-type="accepted">
          <day>23</day>
          <month>11</month>
          <year>2015</year>
        </date>
      </history>
      <permissions>
        <license license-type="open-access" xlink:href="http://creativecommons.org/licenses/by/4.0">
          <license-p>This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.</license-p>
        </license>
      </permissions>
      <kwd-group xml:lang="en">
        <title>Keywords</title>
        <kwd>Hypolipidemic Agents</kwd>
        <kwd>Ezetimibe</kwd>
        <kwd>Cholesterol</kwd>
        <kwd>Hydroxymethylglutaryl CoA- Redutases</kwd>
        <kwd>Drug Combination</kwd>
      </kwd-group>
    </article-meta>
  </front>
  <body>
    <sec>
      <title>Ezetimibe: Clinical and scientific meaning of the IMPROVE-IT study</title>
      <p>After years of anxious wait, in June of the current year, the results of the clinical trial IMPROVE-IT were published in the New England Journal of Medicine.<xref ref-type="bibr" rid="B1">1</xref> This was the first study to properly evaluate the clinical impact of adding ezetimibe to statin therapy. Until then, we had known that the combination of ezetimibe with statin enhanced the antihyperlipidemic effect of statins (surrogate endpoint). However, a positive surrogate marker does not necessarily have a guaranteed clinical impact on the therapy.<xref ref-type="bibr" rid="B2">2</xref> Therefore, the IMPROVE-IT would fill a gap in the literature concerning the effect of ezetimibe on clinical outcomes.</p>
      <p>The publication of the ENHANCE study in 2008 created uncertainty about the efficacy of the therapy with ezetimibe.<xref ref-type="bibr" rid="B3">3</xref> In this randomized, clinical trial including patients with familial hypercholesterolemia, the therapy with ezetimibe added to 80 mg of simvastatin led to greater reductions in LDL cholesterol levels as compared with simvastatin alone. However, the effect was not accompanied by a greater reduction in atherosclerosis, measured by the carotid intima-media thickness (surrogate endpoint). At this point, it became evident that the industry was precipitate in promoting the use of a drug before its effects were demonstrated by clinical outcomes.</p>
      <p>In the same year, the New England Journal of Medicine published an article showing how the marketing to physicians and consumers had a positive impact on sales of ezetimibe in the American market, despite the absence of clinical evidence, differently from what was observed in Canada.<xref ref-type="bibr" rid="B4">4</xref> The marketing was primarily aimed at substituting conventional therapeutic regimens with the combination of ezetimibe with statin at low doses. Nevertheless, in the lack of large studies to evaluate the impact on clinical outcomes, whether the substitution of low doses for high doses of statins (combined with ezetimibe) would lead to decreases in the pleiotropic effects of statins remained unknown.</p>
      <p>Thus, the results of the IMPROVE-IT study were awaited. They were first presented at the American Heart Association's scientific sessions in November, 2014. In contrast to what usually occurs to large clinical trials, the study was not published on the day of presentation, which raised speculations that disagreements between authors and editors caused such delay in the publication, although the authors stressed that they had not submitted the study to the journal before the presentation at the meeting.</p>
      <p>In this context, we will evaluate the dual meaning of this study, which demonstrated not only the efficacy of ezetimibe, but also only a slight decrease of the risk. This observation encourages us to have an intense debate over the relative risk reduction, the absolute risk reduction (ARR) and the number needed to treat (NNT) in the critical analysis of the scientific evidence.</p>
    </sec>
    <sec>
      <title>The Proof of Concept</title>
      <p>The IMPROVE-IT study demonstrated that 10 mg of ezetimibe plus 40 mg of simvastatin reduced the incidence of clinical outcomes as compared with 40 mg of simvastatin in patients who had been recently suffered an acute coronary syndrome and had a low LDL cholesterol level (&lt;125 mg/dL).</p>
      <p>The methodology of the study has a high degree of confidence: low risk of bias, since it was a large study involving 18,000 randomized patients (homogeneity of the groups prevents confounding effects); double-blind (preventing performance bias or measurement bias due to biased observation); intention-to-treat analysis; and low loss to follow-up. With respect to the risk of random error, the study had adequate statistical power, based on the analysis of primary outcome (and not of a secondary outcome), on the analysis of the entire sample (and not of subgroups), and on the absence of early stopping of the trial for benefit (truncation).</p>
      <p>Therefore, this study was the first demonstration of clinical protection with ezetimibe, by showing the efficacy of its addition to the therapy with statin therapy in increasing the antihyperlipidemic effect.</p>
    </sec>
    <sec>
      <title>The size of the effect</title>
      <p>At the same time that the IMPROVE-IT study demonstrated the efficacy of ezetimibe, it also showed that such effect was almost irrelevant. In fact, this was a positive study concerning the presence of the effect, but negative in relation to the size of the effect.</p>
      <p>My point of view may sound strange, since the NNT of the IMPROVE-IT was 50, which is considered of moderate relevance. For example, with this NNT found in the CURE study,<xref ref-type="bibr" rid="B5">5</xref> we started to combine clopidogrel with acetylsalicylic acid in patients with acute coronary syndromes. So why am I saying that the IMPROVE-IT is a negative study regarding the size of the effect? The answer is based on the relative risk reduction of only 6%.</p>
      <p>The (correct) notion that has been supported by the evidence-based medicine is that the relative risk reduction alone may lead to an overestimation of the benefits of certain treatment, which makes us to calculate the ARR and the NNT (100/ARR) for a better overview of the effect. Indeed, authors have more commonly used relative risk reductions than absolute reductions,<xref ref-type="bibr" rid="B6">6</xref> since beneficial treatments have relative reductions of 20%-40% whereas absolute reductions of approximately 2%-4%, which makes the former more attractive to use.</p>
      <p>The lower value of the ARR not only makes it a more reliable parameter, but also approximates it to the real impact. For example, the information that I have received 50% of an inheritance (relative value) has little meaning if I do not know the total value of the inheritance (absolute value). However, when the relative risk reduction is low, both the ARR and the NTT may overestimate the benefits of the treatment. The roles are reversed. Why?</p>
      <p>The relative risk reduction represents an intrinsic effect of the therapy, and tends to be constant, independently of the patient's baseline risk.<xref ref-type="bibr" rid="B7">7</xref> In contrast, the ARR is not an intrinsic property, and varies accordingly with the baseline risk. For the same relative reduction, the higher the baseline risk, the higher the absolute risk. Therefore, when the intrinsic effect of the therapy is minimal, the authors may exacerbate the results and make them seem relevant by applying them in a sample in which the incidence of the outcome is high.</p>
      <p>That was what occurred in the IMPROVE-IT study, which compared a good treatment (statin) with a mere improvement of such treatment (combination with ezetimibe) in patients with low cholesterol (&lt;125 mg/dL). The expected intrinsic effect could be nothing but minimal. Once aware of this, the authors have chosen patients with acute coronary syndromes (higher risk) and defined outcome as the combination of five components (death, infarction, stroke, hospitalization for angina and need of revascularization), in place of the three traditionally used components (death, infarction and stroke). Consequently, the incidence of the outcome was 34%, and a slight relative reduction of 6% would lead to a reasonable absolute reduction of 2% and a NNT of 50.</p>
      <p>It is usually argued that ARR and NNT are more representative parameters of relevance, since it involves both the relative and the absolute reduction (relative risk reduction x absolute risk = ARR). However, when a very low relative reduction is accompanied by a reasonable NNT, we need to be cautious. There are two possibilities:</p>
      <p>First, the NNT is artificially exacerbated by the study sample, composed by a subgroup of patients with high risk, or by an outcome composed of many items, some of them less relevant than others. Both of these occurred in the IMPROVE-IT study - a population with high incidence of the outcome (acute coronary syndromes) had been chosen in order to clinically validate the use of ezetimibe; and second, the outcome was a combination of death, infarction, stroke (hard outcomes), hospitalization for angina and need of revascularization (soft outcomes). Of the 170 outcomes prevented, only one death was prevented, and the others included infarction and hospitalization.</p>
      <p>In the second possibility, the NNT correctly reflects the magnitude of the effect, because the risk of the disease is too high, and even a small relative risk reduction has a relatively good impact. This is what happens with the angiotensin-converting-enzyme (ACE) inhibitor in heart failure, which promotes 16% of relative risk reduction.<xref ref-type="bibr" rid="B8">8</xref> However, this is a high-mortality disease, leading to a high NNT. Here, there is no selection of a specific subgroup or a combination of many outcomes. We are only speaking of death.</p>
      <p>In light of this, we can follow the following rule: if the relative risk reduction is reasonable (&gt;20%), we should calculate the NNT, which will indicate whether the intrinsic effect reflects a relevant absolute difference. Or, if the relative risk reduction is low, we may conclude that the impact to the treatment is also low, and this should be reserved to situations of high incidence of the outcome. Thus, the analysis of the size of the effect should take into account both the relative risk reduction (intrinsic effect) and the impact on the mean patient (NNT).</p>
    </sec>
    <sec>
      <title>Imprecision of the size of the effect</title>
      <p>Another aspect of the IMPROVE-IT is the imprecision of its estimates. We are mentioning not only a relative risk reduction of only 6%, but also a confidence interval of 1%-11%. In other words, we cannot rule out the possibility of a relative risk reduction as low as 1%. Considering this worse scenario, the NNT would be of 288. Although the IMPROVE-IT is a large study (n=18,000), it becomes inaccurate in describing such a small effect.</p>
    </sec>
    <sec>
      <title>Clinical implications</title>
      <p>The IMPROVE-IT study demonstrated that the prescription of ezetimibe for patients in use of statins and reasonable cholesterol levels (mean LDL = 90 mg/dL) has a minimal beneficial effect. Then, the prescription of this combination would be an exception, rather than a rule, limited to unusual situations in which a relative risk reduction of 6% would promote an absolute risk reduction of satisfactory magnitude of hard outcomes. The minimal effect presented by the study implies that the treatment may be used in a minority of patients. For this reason, we believe that the IMPROVE-IT study yielded a more negative than positive result.</p>
      <p>Considering the evidence-based medicine paradigm, the utilization of indirect evidence can be considered adequate. And this was the main contribution of the IMPROVE-IT, i.e., the application of the concept proved by the study in other types of patients, including, for example, those who remain with high cholesterol levels despite statin therapy at maximum doses. This was not the focus of the IMPROVE-IT, although its results indirectly indicate that the combination with ezetimibe may reduce the occurrence of events in these patients - if the effect was observed in patients with normal cholesterol levels, why it would not be expected in patients with high cholesterol levels? According to the meta-analysis graph depicted in the discussion session of the IMPROVE-IT, the higher the ARR of cholesterol level, the higher the relative risk reduction.</p>
      <p>Therefore, if we use the concept of the IMPROVE-IT in a population with higher cholesterol level, we may achieve a higher benefit; that would be a good example of using indirect evidence. This tends to be a more useful application, since we would be correcting a still elevated cholesterol level despite the use of statin. This is the case in which the use of indirect evidence seems to make more sense than the use of direct evidence.</p>
      <p>In addition, in the IMPROVE-IT study, ezetimibe was combined with a statin therapy of moderate potential (40 mg of simvastatin). Wouldn't it be better to substitute it with a more potent regimen, using high doses of statin, as was performed in the PROVE-IT study?<xref ref-type="bibr" rid="B9">9</xref> We do not have this answer. In case a more potent statin regimen was already being used, then the therapy could be enhanced by ezetimibe. This is another argument in favor of the indirect use of evidence, differently from what was described in the IMPROVE-IT study. However, this also seems to be an exception indication.</p>
      <p>Therefore, for the above-mentioned reasons, the clinical applicability of the IMPROVE-IT study is limited.</p>
    </sec>
    <sec>
      <title>The Real Value of the IMPROVE-IT</title>
      <p>The greatest value of the IMPROVE-IT study is its scientific message that corroborates the hypothesis that cholesterol is a cardiovascular risk factor. Statins reduce cholesterol levels and the cardiovascular risk, representing a criterion of reversibility and supporting this hypothesis. However, some critics argue that statins reduce cardiovascular risk by their pleiotropic effects rather than by reducing cholesterol levels. The IMPROVE-IT demonstrated that the additional reduction in cholesterol levels by a drug other than statin also reduce the cardiovascular risk, hence rejecting the null hypothesis of the relationship between cholesterol and cardiovascular risk.</p>
      <p>And, sorry for the pun, but what the IMPROVE-IT actually improves is our scientific maturity.</p>
    </sec>
  </body>
  <back>
    <fn-group>
      <fn fn-type="supported-by">
        <p>
          <bold>Sources of Funding</bold>
        </p>
        <p>There were no external funding sources for this study.</p>
      </fn>
      <fn fn-type="other">
        <p>
          <bold>Study Association</bold>
        </p>
        <p>This study is not associated with any thesis or dissertation work.</p>
      </fn>
    </fn-group>
    <ref-list>
      <ref id="B1">
        <label>1</label>
        <mixed-citation>1 Cannon CP, Blazing MA, Giugliano RP, McCagg A, White JA, Theroux P, et al. Ezetimibe added to statin therapy after acute coronary syndromes. <italic>N Engl J Med</italic>. 2015;372(25):2387-97.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Cannon</surname>
              <given-names>CP</given-names>
            </name>
            <name>
              <surname>Blazing</surname>
              <given-names>MA</given-names>
            </name>
            <name>
              <surname>Giugliano</surname>
              <given-names>RP</given-names>
            </name>
            <name>
              <surname>McCagg</surname>
              <given-names>A</given-names>
            </name>
            <name>
              <surname>White</surname>
              <given-names>JA</given-names>
            </name>
            <name>
              <surname>Theroux</surname>
              <given-names>P</given-names>
            </name>
            <etal/>
          </person-group>
          <article-title>Ezetimibe added to statin therapy after acute coronary syndromes</article-title>
          <source>N Engl J Med</source>
          <year>2015</year>
          <volume>372</volume>
          <issue>25</issue>
          <fpage>2387</fpage>
          <lpage>2397</lpage>
        </element-citation>
      </ref>
      <ref id="B2">
        <label>2</label>
        <mixed-citation>2 Weintraub WS, L&#252;scher TF, Pocock S. The perils of surrogate endpoints. Eur Heart J.2015; 36(33):2212-8.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Weintraub</surname>
              <given-names>WS</given-names>
            </name>
            <name>
              <surname>L&#252;scher</surname>
              <given-names>TF</given-names>
            </name>
            <name>
              <surname>Pocock</surname>
              <given-names>S</given-names>
            </name>
          </person-group>
          <article-title>The perils of surrogate endpoints</article-title>
          <source>Eur Heart J</source>
          <year>2015</year>
          <volume>36</volume>
          <issue>33</issue>
          <fpage>2212</fpage>
          <lpage>2218</lpage>
        </element-citation>
      </ref>
      <ref id="B3">
        <label>3</label>
        <mixed-citation>3 Kastelein JJP, Akdim F, Stroes ESG, Zwinderman AH, Bots ML, Stalenhoef AFH, et al. Simvastatin with or without ezetimibe in familial hypercholesterolemia. N Engl J Med.2008;358(14):1431-43.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Kastelein</surname>
              <given-names>JJP</given-names>
            </name>
            <name>
              <surname>Akdim</surname>
              <given-names>F</given-names>
            </name>
            <name>
              <surname>Stroes</surname>
              <given-names>ESG</given-names>
            </name>
            <name>
              <surname>Zwinderman</surname>
              <given-names>AH</given-names>
            </name>
            <name>
              <surname>Bots</surname>
              <given-names>ML</given-names>
            </name>
            <name>
              <surname>Stalenhoef</surname>
              <given-names>AFH</given-names>
            </name>
            <etal/>
          </person-group>
          <article-title>Simvastatin with or without ezetimibe in familial hypercholesterolemia</article-title>
          <source>N Engl J Med</source>
          <year>2008</year>
          <volume>358</volume>
          <issue>14</issue>
          <fpage>1431</fpage>
          <lpage>1443</lpage>
        </element-citation>
      </ref>
      <ref id="B4">
        <label>4</label>
        <mixed-citation>4 Jackevicius CA, Tu JV, Ross JS, Ko DT, Krumholz HM. Use of ezetimibe in the united states and canada. N Engl J Med. 2008;358(17):1819-28.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Jackevicius</surname>
              <given-names>CA</given-names>
            </name>
            <name>
              <surname>Tu</surname>
              <given-names>JV</given-names>
            </name>
            <name>
              <surname>Ross</surname>
              <given-names>JS</given-names>
            </name>
            <name>
              <surname>Ko</surname>
              <given-names>DT</given-names>
            </name>
            <name>
              <surname>Krumholz</surname>
              <given-names>HM</given-names>
            </name>
          </person-group>
          <article-title>Use of ezetimibe in the united states and canada</article-title>
          <source>N Engl J Med</source>
          <year>2008</year>
          <volume>358</volume>
          <issue>17</issue>
          <fpage>1819</fpage>
          <lpage>1828</lpage>
        </element-citation>
      </ref>
      <ref id="B5">
        <label>5</label>
        <mixed-citation>5 Yusuf S, Zhac F, Mehta SR, Chrolavicius S, Tognoni G, Fox KK, et al. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without st-segment elevation. N Engl J Med. 2001;345(7):494-502.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Yusuf</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Zhac</surname>
              <given-names>F</given-names>
            </name>
            <name>
              <surname>Mehta</surname>
              <given-names>SR</given-names>
            </name>
            <name>
              <surname>Chrolavicius</surname>
              <given-names>S</given-names>
            </name>
            <name>
              <surname>Tognoni</surname>
              <given-names>G</given-names>
            </name>
            <name>
              <surname>Fox</surname>
              <given-names>KK</given-names>
            </name>
            <etal/>
          </person-group>
          <article-title>Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without st-segment elevation</article-title>
          <source>N Engl J Med</source>
          <year>2001</year>
          <volume>345</volume>
          <issue>7</issue>
          <fpage>494</fpage>
          <lpage>502</lpage>
        </element-citation>
      </ref>
      <ref id="B6">
        <label>6</label>
        <mixed-citation>6 Nuovo J, Melnikow J, Chang D. Reporting number needed to treat and absolute risk reduction in randomized controlled trials. <italic>JAMA</italic>. 2002;287(21):2813-4.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Nuovo</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Melnikow</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>Chang</surname>
              <given-names>D</given-names>
            </name>
          </person-group>
          <article-title>Reporting number needed to treat and absolute risk reduction in randomized controlled trials</article-title>
          <source>JAMA</source>
          <year>2002</year>
          <volume>287</volume>
          <issue>21</issue>
          <fpage>2813</fpage>
          <lpage>2814</lpage>
        </element-citation>
      </ref>
      <ref id="B7">
        <label>7</label>
        <mixed-citation>7 Schmid CH, Lau J, McIntosh MW, Cappelleri JC. An empirical study of the effect of the control rate as a predictor of treatment efficacy in meta-analysis of clinical trials. <italic>Stat Med</italic>.1998;17(17):1923-42.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Schmid</surname>
              <given-names>CH</given-names>
            </name>
            <name>
              <surname>Lau</surname>
              <given-names>J</given-names>
            </name>
            <name>
              <surname>McIntosh</surname>
              <given-names>MW</given-names>
            </name>
            <name>
              <surname>Cappelleri</surname>
              <given-names>JC</given-names>
            </name>
          </person-group>
          <article-title>An empirical study of the effect of the control rate as a predictor of treatment efficacy in meta-analysis of clinical trials</article-title>
          <source>Stat Med</source>
          <year>1998</year>
          <volume>17</volume>
          <issue>17</issue>
          <fpage>1923</fpage>
          <lpage>1942</lpage>
        </element-citation>
      </ref>
      <ref id="B8">
        <label>8</label>
        <mixed-citation>8 Effect of enalapril on survival in patients with reduced left ventricular ejection fractions and congestive heart failure. The SOLVD Investigators. <italic>N Engl J Med. 1991;325(5):293-302</italic>.</mixed-citation>
        <element-citation publication-type="journal">
          <article-title>Effect of enalapril on survival in patients with reduced left ventricular ejection fractions and congestive heart failure.</article-title>
          <person-group person-group-type="author">
            <collab>The SOLVD Investigators</collab>
          </person-group>
          <source>N Engl J Med</source>
          <year>1991</year>
          <volume>325</volume>
          <issue>5</issue>
          <fpage>293</fpage>
          <lpage>302</lpage>
        </element-citation>
      </ref>
      <ref id="B9">
        <label>9</label>
        <mixed-citation>9 Cannon CP, Braunwald E, McCabe CH, Rader DJ, Rouleau JL, Belder R, et al. Intensive versus moderate lipid lowering with statins after acute coronary syndromes. N Engl J Med. 2004;350(15):1495-504.</mixed-citation>
        <element-citation publication-type="journal">
          <person-group person-group-type="author">
            <name>
              <surname>Cannon</surname>
              <given-names>CP</given-names>
            </name>
            <name>
              <surname>Braunwald</surname>
              <given-names>E</given-names>
            </name>
            <name>
              <surname>McCabe</surname>
              <given-names>CH</given-names>
            </name>
            <name>
              <surname>Rader</surname>
              <given-names>DJ</given-names>
            </name>
            <name>
              <surname>Rouleau</surname>
              <given-names>JL</given-names>
            </name>
            <name>
              <surname>Belder</surname>
              <given-names>R</given-names>
            </name>
            <etal/>
          </person-group>
          <article-title>Intensive versus moderate lipid lowering with statins after acute coronary syndromes</article-title>
          <source>N Engl J Med</source>
          <year>2004</year>
          <volume>350</volume>
          <issue>15</issue>
          <fpage>1495</fpage>
          <lpage>1504</lpage>
        </element-citation>
      </ref>
    </ref-list>
  </back>
  <sub-article article-type="translation" id="s1" xml:lang="pt">
    <front-stub>
      <article-categories>
        <subj-group subj-group-type="heading">
          <subject>Ponto de Vista</subject>
        </subj-group>
      </article-categories>
      <title-group>
        <article-title>Ezetimibe: Significado Cl&#237;nico e Cient&#237;fico do Estudo IMPROVE-IT</article-title>
      </title-group>
      <contrib-group>
        <contrib contrib-type="author">
          <name>
            <surname>Correia</surname>
            <given-names>Luis C. L.</given-names>
          </name>
          <xref ref-type="aff" rid="aff2"/>
          <xref ref-type="corresp" rid="c2"/>
        </contrib>
        <aff id="aff2">
          <institution content-type="original">Escola Bahiana de Medicina e Sa&#250;de P&#250;blica; Hospital S&#227;o Rafael, Salvador, BA - Brasil</institution>
          <country country="BR">Brazil</country>
        </aff>
      </contrib-group>
      <author-notes>
        <corresp id="c2"><bold>Correspond&#234;ncia: Luis Cl&#225;udio Lemos Correi</bold>a, Escola Baiana de Medicina - P&#243;s-gradua&#231;&#227;o. Av. Princesa Leopoldina 19/402, Gra&#231;a. CEP 40150080, Salvador, BA - Brasil. E-mail: <email>lccorreia@cardiol.br</email>
				</corresp>
        <fn fn-type="con">
          <p>
            <bold>Contribui&#231;&#227;o dos autores</bold>
          </p>
          <p>Concep&#231;&#227;o e desenho da pesquisa, Obten&#231;&#227;o de dados, An&#225;lise e interpreta&#231;&#227;o dos dados, An&#225;lise estat&#237;stica, Obten&#231;&#227;o de financiamento, Reda&#231;&#227;o do manuscrito e Revis&#227;o cr&#237;tica do manuscrito quanto ao conte&#250;do intelectual importante: Correia LCL</p>
        </fn>
        <fn fn-type="conflict">
          <p>
            <bold>Potencial conflito de interesse</bold>
          </p>
          <p>Declaro n&#227;o haver conflito de interesses pertinentes.</p>
        </fn>
      </author-notes>
      <kwd-group xml:lang="pt">
        <title>Palavras-chave</title>
        <kwd>Hipolipemiantes</kwd>
        <kwd>Ezetimibe</kwd>
        <kwd>Colesterol</kwd>
        <kwd>Hidroximetilglutaril-CoA Redutases</kwd>
        <kwd>Combina&#231;&#227;o de Medicamentos</kwd>
      </kwd-group>
    </front-stub>
    <body>
      <sec>
        <title>Ezetimibe: Significado Cl&#237;nico e Cient&#237;fico do Estudo IMPROVE-IT</title>
        <p>Ap&#243;s anos de ansiosa espera, em junho do presente ano, foi publicado no <italic>New England Journal of Medicine</italic> o resultado do ensaio cl&#237;nico IMPROVE-IT.<xref ref-type="bibr" rid="B1">1</xref> Este foi o primeiro trabalho adequado para testar o impacto cl&#237;nico de se associar ezetimibe &#224; terapia com estatina. At&#233; ent&#227;o, sab&#237;amos que tal associa&#231;&#227;o potencializava o efeito antilipid&#234;mico das estatinas (desfecho substituto). No entanto, um resultado favor&#225;vel em desfecho substituto n&#227;o garante um impacto cl&#237;nico da terapia.<xref ref-type="bibr" rid="B2">2</xref> Portanto, o IMPROVE-IT preencheria a lacuna da literatura no que se refere ao efeito do ezetimibe em desfechos cl&#237;nicos.</p>
        <p>Em 2008, a publica&#231;&#227;o do estudo ENHANCE trouxe d&#250;vidas quanto &#224; efic&#225;cia do esquema envolvendo o ezetimibe.<xref ref-type="bibr" rid="B3">3</xref> Naquele ensaio cl&#237;nico randomizado que incluiu pacientes com hipercolesterolemia familiar, o esquema que adicionava ezetimibe &#224; sinvastatina 80 mg promoveu maior redu&#231;&#227;o de LDL-colesterol quando comparado &#224; sinvastatina isolada. No entanto, este efeito n&#227;o foi acompanhado de maior redu&#231;&#227;o da aterosclerose medida pela espessura m&#233;dio-intimal de car&#243;tidas (desfecho substituto). Naquele momento, ficou patente a precipita&#231;&#227;o da ind&#250;stria farmac&#234;utica em promover a utiliza&#231;&#227;o de uma droga antes da demonstra&#231;&#227;o de seu efeito em desfechos cl&#237;nicos.</p>
        <p>No mesmo ano, o <italic>New England Journal of Medicine</italic> publicou um artigo que mostrava como o <italic>marketing</italic> direcionado para m&#233;dicos e leigos impactou em grande aumento das vendas do ezetimibe no mercado americano, a despeito da aus&#234;ncia de evid&#234;ncias cl&#237;nicas, diferente do ocorrido no Canad&#225;.<xref ref-type="bibr" rid="B4">4</xref> O <italic>marketing</italic> era prioritariamente voltado para a substitui&#231;&#227;o de esquemas tradicionais com estatina pela combina&#231;&#227;o de ezetimibe e estatina em baixa dose. Por&#233;m, na aus&#234;ncia de estudos suficientemente grandes para avaliar o impacto em desfechos cl&#237;nicos, permanecia a d&#250;vida se a substitui&#231;&#227;o de alta dose por baixa dose de estatina (associado &#224; ezetimibe) promoveria redu&#231;&#227;o dos te&#243;ricos efeitos pleiotr&#243;picos das estatinas.</p>
        <p>Esta contextualiza&#231;&#227;o justifica a expectativa quanto aos resultados do IMPROVE-IT, que foram inicialmente apresentados no congresso do <italic>American Heart Association</italic> em novembro de 2104. Fugindo do habitual, o estudo n&#227;o foi publicado no mesmo dia da apresenta&#231;&#227;o como ocorre com os grandes ensaios cl&#237;nicos. Houve especula&#231;&#245;es de que fortes discord&#226;ncias entre editores e autores atrasaram a publica&#231;&#227;o, enquanto os autores negavam que o estudo havia sido submetido &#224; revista antes da apresenta&#231;&#227;o no congresso.</p>
        <p>Diante deste pano de fundo, analisaremos o significado dualista deste estudo, que demonstra a efic&#225;cia do ezetimibe, mas ao mesmo tempo evidencia uma redu&#231;&#227;o de risco beirando a irrelev&#226;ncia. Esta observa&#231;&#227;o nos motiva a aprofundar a discuss&#227;o a respeito do valor da redu&#231;&#227;o relativa do risco, da redu&#231;&#227;o absoluta do risco (RAR) e do n&#250;mero necess&#225;rio a tratar (NNT) na an&#225;lise cr&#237;tica de evid&#234;ncias cient&#237;ficas.</p>
      </sec>
      <sec>
        <title>A Prova do Conceito</title>
        <p>O IMPROVE-IT demonstrou que a associa&#231;&#227;o de ezetimibe 10 mg e sinvastatina 40 mg reduziu a incid&#234;ncia de desfechos cl&#237;nicos quando comparado ao uso de sinvastatina 40 mg em pacientes que sofreram recentemente s&#237;ndrome coronariana aguda e cujo LDL-colesterol era baixo (&lt; 125 mg/dL).</p>
        <p>A metodologia cient&#237;fica empregada conferiu ao estudo um alto n&#237;vel de confiabilidade: baixo risco de vi&#233;s, por ser um grande ensaio cl&#237;nico com 18000 pacientes randomizados (homogeneidade dos grupos garante aus&#234;ncia de efeitos de confus&#227;o); duplo-cego (prevenindo vi&#233;s de desempenho ou vi&#233;s de aferi&#231;&#227;o por tendenciosidade de observa&#231;&#227;o); an&#225;lise por inten&#231;&#227;o de tratar; e baixa perda de seguimento. Quanto ao risco de erro aleat&#243;rio (acaso), o estudo teve adequado poder estat&#237;stico, com base na an&#225;lise de desfecho prim&#225;rio (em n&#227;o de desfecho secund&#225;rio), na an&#225;lise de toda a amostra (e n&#227;o de subgrupos), e na aus&#234;ncia de interrup&#231;&#227;o precoce por benef&#237;cio (truncamento).</p>
        <p>Portanto, este estudo foi a primeira demonstra&#231;&#227;o de prote&#231;&#227;o cl&#237;nica com ezetimibe, ao mostrar a efic&#225;cia de sua introdu&#231;&#227;o &#224; terapia com estatina no incremento antilipid&#234;mico.</p>
      </sec>
      <sec>
        <title>O Tamanho do Efeito</title>
        <p>Ao mesmo tempo em que o IMPROVE-IT demonstrou a exist&#234;ncia de efic&#225;cia, este estudo evidenciou tal efic&#225;cia era m&#237;nima, quase irrelevante. Na verdade, este &#233; um estudo positivo em rela&#231;&#227;o &#224; presen&#231;a do efeito, mas negativo em rela&#231;&#227;o ao tamanho do efeito.</p>
        <p>Esta minha coloca&#231;&#227;o pode gerar estranheza, visto que o NNT do IMPROVE-IT foi de 50, o que &#233; considerado de moderada relev&#226;ncia. Por exemplo, este &#233; o NNT que nos fez associar Clopidogrel ao AAS em pacientes com s&#237;ndromes coronarianas agudas, de acordo com o estudo CURE.<xref ref-type="bibr" rid="B5">5</xref> Ent&#227;o, por que estou dizendo que o IMPROVE-IT &#233; um estudo negativo em rela&#231;&#227;o ao tamanho do efeito? Esta minha afirma&#231;&#227;o &#233; baseada na redu&#231;&#227;o relativa do risco de apenas 6%.</p>
        <p>A no&#231;&#227;o (correta) que se tem passado no ensino da medicina baseada em evid&#234;ncias &#233; de que a redu&#231;&#227;o relativa do risco sozinha pode gerar uma ideia hiperinsuflada do benef&#237;cio de um tratamento, o que nos obriga a calcular a RAR e o NNT (100/RAR) para ter uma melhor no&#231;&#227;o da magnitude do efeito. De fato, &#233; mais comum autores utilizarem redu&#231;&#245;es relativas do risco em vez de redu&#231;&#245;es absolutas,<xref ref-type="bibr" rid="B6">6</xref> pois normalmente os tratamento ben&#233;ficos t&#234;m redu&#231;&#245;es relativas de 20% a 40% e redu&#231;&#245;es absolutas em torno de 2 a 4%. Isto faz com que usualmente seja mais sedutor falar em redu&#231;&#227;o relativa.</p>
        <p>O menor valor da redu&#231;&#227;o absoluta n&#227;o somente a torna mais confi&#225;vel, como tamb&#233;m a aproxima mais do impacto real. Por exemplo, dizer que ganhei 50% de uma heran&#231;a (relativo) n&#227;o tem muito significado se eu n&#227;o souber quanto &#233; a heran&#231;a (absoluto).</p>
        <p>No entanto, quando a redu&#231;&#227;o relativa do risco &#233; muito pequena, a RAR e o NNT de um estudo positivo podem superestimar o benef&#237;cio do tratamento. Os pap&#233;is se invertem. Por qu&#234;?</p>
        <p>A redu&#231;&#227;o relativa do risco representa o efeito intr&#237;nseco do tratamento, o que tende a ser constante, independente do risco basal do paciente.<xref ref-type="bibr" rid="B7">7</xref> J&#225; a redu&#231;&#227;o absoluta n&#227;o &#233; uma propriedade intr&#237;nseca, e varia de acordo com o risco basal do paciente. Quanto maior o risco basal, maior a redu&#231;&#227;o absoluta para uma mesma redu&#231;&#227;o relativa. Assim, quando o efeito intr&#237;nseco do tratamento &#233; m&#237;nimo, os autores podem inflar o resultado e torn&#225;-lo relevante ao aplic&#225;-lo em uma amostra de grande incid&#234;ncia do desfecho.</p>
        <p>Foi isto que ocorreu no IMPROVE-IT. O estudo comparou um tratamento bom (estatina) com um mero aprimoramento deste tratamento (associa&#231;&#227;o de ezetimibe) em pacientes com colesterol LDL baixo (&lt; 125 mg/dL). O efeito intr&#237;nseco do tratamento s&#243; poderia ser m&#237;nimo. J&#225; sabendo disso, os autores escolheram pacientes com s&#237;ndromes coronarianas agudas (maior risco) e definiram o desfecho como o combinado de cinco componentes (morte, infarto, AVC, internamento por angina e necessidade de revasculariza&#231;&#227;o), em lugar dos 3 componentes tradicionais (morte, infarto e AVC). Desta forma, a incid&#234;ncia do desfecho foi de 34%, fazendo com que uma m&#237;nima redu&#231;&#227;o relativa de 6% resultasse em uma razo&#225;vel redu&#231;&#227;o absoluta de 2%, gerando um NNT de 50.</p>
        <p>Normalmente, argumenta-se que RAR e NNT sejam medidas mais representativas de relev&#226;ncia, pois este englobam tanto o relativo quanto o absoluto (redu&#231;&#227;o relativa do risco x risco absoluto = RAR). Mas quando uma redu&#231;&#227;o relativa muito pequena &#233; acompanhada de um razo&#225;vel NNT, devemos ter cuidado. Duas s&#227;o as possibilidades:</p>
        <p>Primeiro, o NNT est&#225; artificialmente insuflado pela amostra do estudo, formada de um subgrupo de pacientes com maior risco, ou por um desfecho combinado de v&#225;rios componentes, alguns menos relevantes. Ambos os fen&#244;menos ocorreram com o IMPROVE-IT - primeiro, uma popula&#231;&#227;o de grande incid&#234;ncia de desfecho (s&#237;ndromes coronarianas agudas) foi escolhida para validar clinicamente o ezetimibe; e, segundo, o desfecho foi o combinado de morte, infarto, AVC (duros), internamento por angina e necessidade de revasculariza&#231;&#227;o (moles). Dentre os 170 desfechos prevenidos, apenas 1 morte foi prevenida, sendo o restante divididos entre infarto ou revasculariza&#231;&#227;o.</p>
        <p>Na segunda possibilidade, o NNT retrata corretamente a magnitude de efeito, pois o risco da doen&#231;a &#233; alto, fazendo com que uma pequena redu&#231;&#227;o relativa do risco tenha um bom impacto m&#233;dio. &#201; o que ocorre com uso de inibidor da enzima conversora da angiotensina (ECA) na insufici&#234;ncia card&#237;aca, que promove apenas 16% de redu&#231;&#227;o relativa.<xref ref-type="bibr" rid="B8">8</xref> Por&#233;m, esta &#233; uma doen&#231;a em geral de alta mortalidade, garantindo um NNT alto. Aqui, n&#227;o h&#225; a sele&#231;&#227;o de um subgrupo espec&#237;fico de risco, nem combina&#231;&#227;o de muitos desfechos. Estamos falando apenas de morte.</p>
        <p>Sendo assim, podemos seguir a seguinte regra: se a redu&#231;&#227;o relativa do risco for razo&#225;vel (&gt; 20%), devemos calcular o NNT, que nos dir&#225; se o efeito intr&#237;nseco se reflete em uma diferen&#231;a absoluta importante. Em segundo lugar, se a redu&#231;&#227;o relativa do risco for pequena, conclu&#237;mos que o impacto do tratamento &#233; pequeno, e esse deve ser reservado apenas para situa&#231;&#245;es de alta incid&#234;ncia do desfecho. Portanto, a an&#225;lise de tamanho do efeito deve levar em conta simultaneamente a redu&#231;&#227;o relativa do risco (efeito intr&#237;nseco) e o impacto no paciente m&#233;dio (NNT).</p>
      </sec>
      <sec>
        <title>Imprecis&#227;o do Tamanho do Efeito</title>
        <p>Outro aspecto do IMPROVE-IT &#233; a imprecis&#227;o de suas estimativas. Falamos em uma redu&#231;&#227;o relativa de 6%, por&#233;m calcularmos um amplo intervalo de confian&#231;a entre 1% e 11%, ou seja, n&#227;o podemos afastar a possibilidade de a redu&#231;&#227;o relativa ser t&#227;o pequena quanto 1%. Considerando este pior cen&#225;rio, o NNT seria de 288. Embora o IMPROVE-IT seja um estudo grande (N = 18.000), este se torna impreciso quando tenta retratar um efeito t&#227;o pequeno.</p>
      </sec>
      <sec>
        <title>Implica&#231;&#245;es Cl&#237;nicas</title>
        <p>O IMPROVE-IT demonstrou que a prescri&#231;&#227;o de ezetimibe em pacientes que est&#227;o em uso de estatinas, com colesterol bastante razo&#225;vel (LDL-m&#233;dio = 90 mg/dL), promove um efeito ben&#233;fico de m&#237;nima magnitude. Um efeito deste tipo tornar&#225; a prescri&#231;&#227;o desta combina&#231;&#227;o uma exce&#231;&#227;o, reservada para inusitadas situa&#231;&#245;es em que uma redu&#231;&#227;o relativa do risco de 6% promova uma redu&#231;&#227;o absoluta de magnitude satisfat&#243;ria em desfechos duros. O m&#237;nimo efeito apresentado pelo estudo implica o uso do tratamento para uma minoria de pacientes. Por este motivo, retratamos este como um estudo mais negativo do que positivo.</p>
        <p>Dentro do paradigma da medicina baseada em evid&#234;ncias podemos julgar como adequada a utiliza&#231;&#227;o de evid&#234;ncias indiretas. E essa &#233; a maior utilidade do IMPROVE-IT, ou seja, a aplica&#231;&#227;o do conceito comprovado no estudo em outros tipos de pacientes. Refiro-me, por exemplo, ao paciente que permanece com colesterol alto, a despeito de um tratamento com dose m&#225;xima de estatina. N&#227;o foi exatamente isso que foi avaliado no trabalho, mas os resultados do IMPROVE-IT indiretamente indicam que a associa&#231;&#227;o de ezetimibe pode promover redu&#231;&#227;o de eventos nestes pacientes. Pois, se esse efeito foi observado no paciente com colesterol normal, por que n&#227;o promoveria no paciente com colesterol alto? Conforme gr&#225;fico da an&#225;lise de meta-regress&#227;o trazido pela discuss&#227;o do IMPROVE-IT, quando maior a RAR de colesterol, maior a redu&#231;&#227;o relativa do risco.</p>
        <p>Isso indica que se usarmos o conceito do IMPROVE-IT em uma popula&#231;&#227;o com n&#237;vel colesterol mais elevado, poder&#237;amos obter um benef&#237;cio maior; &#233; um bom exemplo do uso de evid&#234;ncia indireta. Esta tende a ser uma aplica&#231;&#227;o mais &#250;til, pois estar&#237;amos corrigindo n&#237;veis ainda elevados de colesterol, a despeito do uso da estatina. &#201; o caso em que a aplica&#231;&#227;o indireta de uma evid&#234;ncia parece fazer mais sentido do que a evid&#234;ncia direta.</p>
        <p>Al&#233;m disso, no IMPROVE-IT o ezetimibe foi associado a um esquema de moderada pot&#234;ncia de estatina (sinvastatina 40 mg). N&#227;o seria melhor trocar por estatina mais potente, em alta dose, como foi feito no estudo PROVE-IT?<xref ref-type="bibr" rid="B9">9</xref> N&#227;o temos a resposta do que seria melhor. Caso um esquema mais potente de estatina j&#225; estivesse sendo usado, o tratamento poderia ser ent&#227;o aprimorado com o ezetimibe. Mais um argumento a favor do uso indireto da evid&#234;ncia em vez de sua aplica&#231;&#227;o como feito no IMPROVE-IT. Mesmo assim, esta indica&#231;&#227;o tamb&#233;m parece ser de exce&#231;&#227;o.</p>
        <p>Desta forma, pelos motivos aqui discutidos, este estudo tem aplicabilidade cl&#237;nica limitada.</p>
      </sec>
      <sec>
        <title>O Verdadeiro Valor do IMPROVE-IT</title>
        <p>O maior valor do IMPROVE-IT est&#225; em sua mensagem cient&#237;fica, que cristaliza a hip&#243;tese do colesterol como fator de risco cardiovascular. Estatinas promovem redu&#231;&#227;o de colesterol e reduzem risco cardiovascular, representando um crit&#233;rio de reversibilidade a favor dessa hip&#243;tese. No entanto, alguns cr&#237;ticos argumentam que a redu&#231;&#227;o de risco com estatina poderia decorrer de seus efeitos pleiotr&#243;picos, e n&#227;o da redu&#231;&#227;o do colesterol. O IMPROVE-IT mostrou que a redu&#231;&#227;o adicional do colesterol com uma droga diferente de estatina tamb&#233;m reduz risco, rejeitando a hip&#243;tese nula da rela&#231;&#227;o entre o colesterol e risco cardiovascular.</p>
        <p>Sem perder o trocadilho, o que o IMPROVE-IT melhora mesmo &#233; nossa maturidade cient&#237;fica.</p>
      </sec>
    </body>
    <back>
      <fn-group>
        <fn fn-type="supported-by">
          <p>
            <bold>Fontes de financiamento</bold>
          </p>
          <p>O presente estudo n&#227;o teve fontes de financiamento externas.</p>
        </fn>
        <fn fn-type="other">
          <p>
            <bold>Vincula&#231;&#227;o acad&#234;mica</bold>
          </p>
          <p>N&#227;o h&#225; vincula&#231;&#227;o deste estudo a programas de p&#243;s-gradua&#231;&#227;o.</p>
        </fn>
      </fn-group>
    </back>
  </sub-article>
</article>
