Comment on “Is Addition of Vasodilators to Loop Diuretics of Value in the Care of Hospitalized Acute Heart Failure Patients? Real-World Evidence from a Retrospective Analysis of a Large United States Hospital Database”

Abstract

We read with great interest the manuscript by Gradman et al, 1 where increased costs but equivocal outcomes were found in a cohort of acute heart failure (AHF) patients treated with intravenous (IV) diuretic plus vasodilator compared with a matched group who received IV diuretic alone. The authors conclude that their results “extend findings from earlier prospective studies questioning the value of IV vasodilators in addition to IV loop diuretics in the management of hospitalized patients with AHF.” It should be noted that the study findings were based on a retrospective analysis of aggregate data abstracted from the Thomson Reuters Marketscan Hospital Drug Database. As the authors acknowledge, this dataset is limited in that available information and important basic contributors to clinical decisionmakinginthesettingofAHF,suchasbloodpressure and cardiac biomarker results, are missing. Despite attempts toaccountfortheseandotherpotentialunobservedfactorsby excluding patients with a diagnosis of myocardial infarction (MI)oracutecoronarysyndrome(ACS),andusingamatched sample for comparison based on propensity score, the likelihood of significant residual confounding exists, and bias, especiallybyindication,ishigh.Asevidenceofthelatter,patientswhoreceivedvasodilators(particularlythosewhowere given nitrates) were far more likely to undergo cardiac catheterization, a procedure that is used primarily for the evaluation and management of suspected coronary artery disease (CAD). Although those patients may not have been specifically coded as having MI or ACS and therefore were not eliminated by the study’s selection process, it is clear that, atsomepointduringtheirin-hospitalevaluation,thepotential for CAD was considered. In such cases, vasodilators (especially nitrates) may have been given for that specific reason and not for treatment of AHF. By knowing only that vasodilators were given, but not why, the potential for a reverse causal relationship is very real. Moreover, and perhaps of even greater consequence, is that the source data do not provide any information on when vasodilators were administered relative to the patient’s acute presentation. It may well be that patients received them later in their course. Indeed, data from the ADHERE registry show that the benefits of vasodilator therapy are time dependent, with better clinical (and financial) outcomes when administered early and absolutely no mortality benefit if initiated O24 hours after presentation. 2 That the lack of ability to determine time of vasodilator administration is a critical weakness of the Gradman et al analysis is further supported by a recent systematic review that included 35 IV vasodilator studies where drug was administered either in the prehospital setting or in the emergency department. 3 Although that review did not find evidence of associated mortality reduction, symptom improvement and safety were consistently seen, suggesting that vasodilators do have a potentially important role to play in the management of AHF. Several large, prospective, randomized controlled trials of IV vasodilators that are adequately powered to evaluate mortality as an end point are currently underway. 4 Pending their results, deviation from current recommendations that include early IV vasodilators as part of standard care in nonhypotensive patients 5 based on this retrospective analysis of a limited dataset seems to be ill advised. Disclosures