Does Published Evidence Reflect the Whole Story?

Source: Turner EH, et al. N Engl J Med. 2008;358:252-260.

Currently, when industry applies to the FDA for new drug licensing, they must register all trials performed, whether published or not. The evidence base readily available to clinicians often does not contain all of these trials; hence concern has been raised that non-published trials might have outcomes less favorable to sponsor interests, thereby providing a more rosy portrait of efficacy in the published literature than in the total evidence base.

Turner et al chose to consider clinical trials of FDA-approved antidepressants. They examined 74 studies that had been registered with the FDA. Among these clinical trials, half were positive, and all but 1 of these was published. Of the remaining 37 negative trials (negative or of questionable results as assessed by FDA reviewers), 22 were not published. Some of the negative trials (n = 11) that were published portrayed outcomes in an overly optimistic light according to the assessment of the authors.

"Publication bias" is the term used to account for the variability in likelihood that a study will achieve published status due to some particular characteristic, rather than on the merits of its scientific integrity alone. Trials with a positive outcome (Drug A is better than Drug B for Disease X) are more likely to be published than negative trials (Drug A failed to improve the outcome of disease X); a failed trial (Drug A did not achieve its primary endpoint, but secondary endpoints are favorably hypothesis-generating) has historically fared similarly.

Overall, the authors suggest that the efficacy of antidepressants as demonstrated in published clinical trials is more robust than would be concluded on the basis of the entirety of clinical trial experience.

Vitamin E levels and Physical Decline with Aging

Source: Bartali B, et al. JAMA. 2008;299(3):308-315.

Most of us anticipate that our older years will be associated with a decline in physical function. The graying of America is associated with a decline to the level of disability for some, which reduces quality of life and taxes the health care system. The SPPB (Short Physical Performance Battery) is a tool that quantifies the progression of functional decline.

Using a suburban population of seniors (>age 65) from Florence, Italy, investigators followed 698 subjects for 3 years. In addition to monitoring with the SPPB, the micronutrients vitamin B6, vitamin B12, vitamin D, vitamin E, and folate were measured.

After adjustment for confounders, multivariate analysis found a significant relationship between lower vitamin E levels and physical decline, but none of the other micronutrients demonstrated a meaningful relationship after multivariate analysis.

Based upon this data, the authors conclude that a relationship between lower vitamin E levels and physical decline has been demonstrated. Whether the relationship between vitamin E and physical function is causally related is not established by this longitudinal observational trial. Even if vitamin E is someday determined to affect physical decline, the clinical trials to date have been discouraging about other cardiovascular outcomes in persons supplemented with vitamin E.

Effort, Efficacy, and Effectiveness

Source: Viera AJ, et al. J Clin Hypertens. 2008;10;105-111.

The vicissitudes of enabling lifestyle changes to effect reductions in blood pressure are sufficiently well accepted that clinicians may be tempted to sidestep this phase of management in preference for the more routinely effective pharmacotherapies. Recall that the distinction between "efficacy" and "effectiveness," in evidence-based-medicine parlance, is that the former refers to outcomes seen in a controlled clinical trial, and the latter to outcomes seen/anticipated in the hands of the typical community clinician. Even though the efficacy of lifestyle interventions is well established, clinicians may wonder whether it is worth the effort to try and achieve the effectiveness that might be attained in the typical practice setting.

The Behavior Risk Factor Surveillance System, administered under the auspices of the CDC, has indicated that 90% of hypertensive adults reported receiving advice about lifestyle modification for blood pressure control (n = >27,000).

When comparing those who did vs did not recall receiving clinician advice about behavior modification, those who recalled advice from their clinician were over 1.6 times more likely to have modified their behavior. In other words, our efforts do impact effectiveness.

Comprehensive Diabetes Care: The Whole Enchilada

Source: Goede P, et al. New Engl J Med. 2008;358:580-591.

The greatest challenge for any potentially mortal disorder is to prove that therapeutic interventions reduce not only disease-specific mortality, but also all-cause mortality. For instance, PSA screening is associated with reduced prostate-cancer related mortality, but has never been convincingly shown to reduce all-cause mortality, leaving in question whether PSA screening results in some sort of mortality "trade-off," with no net benefit.

The Steno-2 Study enrolled type 2 diabetics with microalbuminuria (n=160) in 1993. Study subjects were randomized to intensive therapy (tight control of glucose, cholesterol, triglycerides, blood pressure, and microalbuminuria) or usual diabetic care. The study concluded after 8 years, at which point both groups and their clinicians in were encouraged to employ intensive therapy post-trial.

In 2006 (five years after formal study closure), Gaede et al examined death from any cause comparing the two original study groups. The hazard ratio for all-cause mortality was 0.54 in the intensive treatment group (24 deaths vs 40 deaths), which included a 57% lesser chance of cardiovascular death, and a 59% lesser incidence of cardiovascular events.

Part of the intensive regimen included low-dose ASA and ACE (or ARB) for all, regardless of blood pressure. Although glucocentric focus in diabetes produces microvascular benefits, dealing with the "whole enchilada"-BP, lipids, the renin-angiotensin system, and platelets-increases the payoff for the patient.

The ENDORSE Study: Preventable Hospital Deaths

Source: Cohen AT, et al. Lancet. 2008;371;387-394.

Pulmonary embolism (PUL-E) is generally accepted as the most common preventable cause of in-hospital deaths in developed countries. As many as 10% of in-hospital deaths are attributable to PUL-E, hence prevention of venous thromboembolism is an epidemiologically compelling mandate. ENDORSE (Epidemiologic International Day for the Evaluation of Patients at Risk for Venous Thromboembolism in the Acute Hospital Care Setting) is a chart audit of at-risk hospitalized patients to ascertain the rate of appropriate VTE thromboprophylaxis (TPX). Hospitals from 32 countries (including the US, UK, Germany and France) provided data on 68,183 hospital patients, categorically divided into surgical (45%) and medical (55%). American College of Chest Physicians (ACCP) guidelines were used to assess risk status and appropriateness of TPX.

Considering the entire study population, slightly over half (58.5%) of high-risk surgical patients, and less than half (39.5%) of the high-risk medical patients received appropriate TPX. Germany and Switzerland achieved distinctively higher levels of appropriate overall TPX (85% and 80%, respectively). With a few exceptions, the tendency to effectively prophylax medical patients less often than surgical patients was seen in all nations.

Reduced mortality through TPX for VTE is readily achievable, yet there remains an important gap between our potential to utilize VTE TPX and current practice worldwide.

Risks of Stopping Clopidogrel

Source: Ho PM, et al. JAMA. 2008;299(5):532-539.

In addition to aspirin (ASA), clopidogrel (CPG) has demonstrated class 1A level evidence of its efficacy in acute coronary syndromes (ACS), and is typically continued for 1 year post-discharge. In ACS subjects it has been previously observed that at the time of cessation of antiplatelet therapies (ASA and heparin, for example), the risk of adverse cardiovascular (CV) events increases, suggesting a transient hyperthrombotic state, since this signal of increased events is not sustained. The authors sought to determine whether a CPG rebound effect occurs, as would be manifest by a cluster of CV events shortly after CPG discontinuation.

A cohort of VA hospital patients (n=3,137) post-discharge for ACS comprised the subjects for this retrospective cohort study. The study subjects were equally distributed between medically treated ACS and PCI treated patients. The incidence of MI and all-cause mortality was examined in the period 0-90 days, 91-180 days, and 181-270 days post CPG cessation.

Overall, there was a statistically significant increased risk of adverse CV events (1.82 relative risk) when comparing the immediate 0-90 day interval with the rest of the followup period. Both medically treated and PCI patients had similar outcomes.

This is the first study of its kind to specifically examine risk in the immediate CPG-cessation period. If confirmed, clinicians may have to rethink either the duration of CPG therapy, the discontinuation scheme (eg, down-titration, supplementary antiplatelet therapy), or both.