Heartbeat: The gut microbiota and heart failure

Catherine M Otto, Kazem Rahimi

Patients with acute heart failure are athigh risk of adverse clinical outcomes.Although elevated levels of trimethyla-mine N-oxide (TMAO), a gut-derivedmetabolite, have been shown to be inde-pendently associated with cardiovasculardisease and poor outcomes in stable heartfailure patients, their role in patients withacute heart failure is not well understood.In this issue of Heart, Suzuki and collea-gues (see page 841) investigate this ques-tion further and show that the addition ofTMAO to established multivariable riskmodels led to better stratification of acuteheart failure patients’ risk of in-hospitalmortality (figure 1). However, the associ-ation was attenuated when renal functionwas added to the models.

In an accompanying editorial, Tang andWilson (see page 813) provide an over-view of the role of TMAO as a marker ofrisk in other clinical settings and commentthat the complementary findings from thisnew analysis provide important insights asto how gut microbia may contribute todisease progression in HF. They speculatethat TMAO may in fact have a causal rolein HF deterioration. To this end, theyconclude that “further studies are war-ranted to determine if circulating TMAOlevels can serve as a guidance to tailor

dietary modulation or therapeutic inter-diction in either acute or chronic HF”.Multimodality imaging provides sophis-

ticated measures of myocardial perfusion,sympathetic denervation and myocardialscarring, each of which has been proposedto predict the risk of ventricular arrhyth-mias in patients with ischemic heartdisease. Rijnierse and colleagues (see page832) performed a detailed study in 52patients with left ventricular (LV) systolicdysfunction (ejection fraction 35%) dueto ischemic heart disease, with multimod-ality imaging at baseline followed by elec-trophysiological study (EPS) afterplacement of an implantable cardioverter-defibrillator (ICD) (figure 2). Hyperaemicmyocardial blood flow (MBF) measuredby positron emission tomography was thesingle best independent predictor or an

inducible sustained ventricular arrhythmiaon EPS. Interestingly, predictive value ofhyperemic MBF was not improved byaddition other risk markers includingsympathetic denervation size, extent ofmyocardial scar, perfusion defect size, LVvolumes or ejection fraction.

Commenting on this study in an edito-rial, Kaufmann and Stehli (see page 815)express surprise that data from the currentstudy did not support the value of previ-ously identified risk predictors, such asextent of scar tissue, and they suggest thatfurther follow-up on this study cohortwill be of interest because ventriculararrhythmias on EPS “may not be a perfectsurrogate marker for sudden cardiac deathand ICD discharges”. In addition, theycaution that “so far there is no solid evi-dence from prospective randomised

Figure 2 Examples of positron emission tomography (PET) and late gadoliniumenhanced-cardiovascular MRI (LGE)-CMR results in two patients with electrophysiological study(EPS)-positive (A) and EPS-negative (B) result. (A) A patient with LGE scar in the inferolateral wallwith corresponding PET perfusion and innervation defects, and significant innervation-perfusionmismatch as depicted in red. Quantitative PET revealed a severely impaired hyperaemicmyocardial blood flow (MBF). (B) A patient with LGE scar in the anteroseptal wall withcorresponding PET perfusion and innervation defects, and limited area of innervation-perfusionmismatch (red area). Quantitative PET revealed a relatively preserved hyperaemic MBF.

Figure 1 Forest plots to show binary logisticORs (dots) and 95% CIs (horizontal bars) forcardiac risk factors for multivariate models forthe individual components of in-hospitalmortality prediction algorithms of ADHERE(triangles), OPTIMIZE-HF (circles) andGWTG-HF (squares) registries with the additionof TMAO. BP, blood pressure; BUN, blood ureanitrogen; COPD, chronic obstructive pulmonarydisease; EF, ejection fraction; PH, past history;TMAO, trimethylamine N-oxide; *p<0.05;**p<0.0005.

Correspondence to Professor Catherine M Otto,Division of Cardiology, University of Washington,Seattle, WA 98195, USA; cmotto@u.washington.edu

Heart June 2016 Vol 102 No 11 811

Heartbeat on M

ay 1, 2020 by guest. Protected by copyright.

http://heart.bmj.com

/H

eart: first published as 10.1136/heartjnl-2016-309848 on 12 May 2016. D

ownloaded from

clinical trial data to support that any ofthese tests alone or in combination shouldbe used to efficiently guide ICD therapyand therefore none of the major societieshave specifically endorsed any of theseapproaches”.

The diagnosis of coronary disease atfirst presentation of suspected anginaremains challenging, leading to costlydiagnostic testing in many patients with alow likelihood of disease. Sekhri and col-leagues (see page 869) assessed the valueof simple clinical measures for predictionof coronary deaths over 10 years offollow-up in a multicenter cohort of 8762patients with suspected angina (table 1).By including pulse rate, smoking status,diabetes and ECG findings, the proposedprognostic model was superior to thestandard Diamond-Forrester model whichincludes only age, sex and symptoms(figure 3). Readers may want to try theonline Prognosis In Suspected Angina(PISA) calculator provided by theauthors at this website https://www.sealedenvelope.com/trials/pisa/

The Education in Heart article in thisissue (see page 882) reviews antithrombo-tic therapy in medically managed patientswith a non-ST elevation acute coronarysyndromes. The primary learning objec-tives for this article include the ability toexplain the mechanisms of action, poten-tial side effects, and current guidelinesrelevant to the pharmacological agentsused in treatment of this condition(figure 4).This issue’s Image Challenge (see page

881) shows an unusual echocardiographic

finding that is important to know about –be sure to look at the online videos, usingthe loop feature in your video viewer.

To cite Otto CM, Rahimi K. Heart 2016;102:811–812.

Heart 2016;102:811–812.doi:10.1136/heartjnl-2016-309848

Table 1 Multivariable predictors ofcoronary death (n=8762, 233 coronarydeaths)

Variable HR (95% CI) p Value

Age (per 10 years) 2.33 (2.05 to 2.65) <0.0001SexFemale 1 <0.0001Male 1.91 (1.45 to 2.52)

Character symptomsAtypical 1 0.0043Typical 1.59 (1.20 to 2.12)Non-cardiac 1.05 (0.67 to 1.63)Pulse rate (per10 bpm)

1.22 (1.13 to 1.33) <0.0001

Current smokerNo 1 0.0016Yes 1.64 (1.21 to 2.23)

Diabetes (y/n)No 1 <0.0001Yes 1.99 (1.46 to 2.70)

ECG normalNormal 1 <0.0001Abnormal 1.96 (1.49 to 2.59)

Harrell’s C=0.83.Adding hospital to the above model (non-Newham vsNewham) HR: 1.01 (95% CI 0.77 to 1.32), p=0.29.

Figure 3 Kaplan–Meier cumulative coronary mortality by quarters of risk for the full prognosticmodel (based on table 1). There were 5, 14, 41, and 173 coronary deaths in risk groups 1 (lowestrisk quarter) to 4 (highest risk quarter), respectively.

Figure 4 A decision algorithm of antithrombotic therapy in medically managed patients withNSTE-ACS. NSAID, non-steroidal anti-inflammatory drug; NSTE-ACS, non-ST-segment elevationacute coronary syndromes; NSTEMI, non-ST-segment elevation myocardial infarction; TIA,transient ischaemic attack; UFH, unfractionated heparin.

812 Heart June 2016 Vol 102 No 11

Heartbeat on M

ay 1, 2020 by guest. Protected by copyright.

http://heart.bmj.com

/H

eart: first published as 10.1136/heartjnl-2016-309848 on 12 May 2016. D

ownloaded from