Relaxin-2

Neither primary endpoint was met. By day 180, 285 of 3,274 people on serelaxin (8.7%) and 290 of 3,271 on placebo (8.9%) had died from cardiovascular causes (hazard ratio 0.98, 95% CI 0.83 to 1.15, P=0.77). Worsening heart failure by day 5 occurred in 227 people on serelaxin (6.9%) and 252 on placebo (7.7%) (hazard ratio 0.89, 95% CI 0.75 to 1.07, P=0.19). Deaths from any cause by day 180 were 11.2% versus 11.9% (hazard ratio 0.94, 95% CI 0.81 to 1.08), and the journal abstract reports no significant differences in all-cause death, in cardiovascular death or rehospitalization, or in the length of the hospital stay. Secondary analyses of how quickly congestion signs improved gave hazard ratios between 1.00 and 1.08 across five signs, and do not change the failed primary results.

• Both co-primary endpoints were missed, so secondary results cannot establish a benefit.
• Participants were a selected group (hospitalized with congestion, raised natriuretic peptides, mild-to-moderate kidney impairment and systolic blood pressure of at least 125 mm Hg), so results may not apply to all acute heart failure.
• Of 6,600 people randomized, 55 were left out of the analysis (6,545 analyzed), as stated in the registry.
• The registry and the journal abstract give different P values for the same primary comparisons; the registry labels its P values one-sided.
• Only the journal abstract was reviewed, not the full paper, and the trial was funded by the drug's developer.

• P value for cardiovascular death through day 180: ClinicalTrials.gov results (labelled one-sided) reports P=0.3857; NEJM abstract reports P=0.77. The same comparison has a different P value in the two sources; the registry labels its P value one-sided, and the abstract does not state which test it used. The reason for the difference is not established from the sources reviewed.
• P value for worsening heart failure through day 5: ClinicalTrials.gov results (labelled one-sided) reports P=0.0968; NEJM abstract reports P=0.19. The same comparison has a different P value in the two sources; the registry labels its P value one-sided, and the abstract does not state which test it used. The reason for the difference is not established from the sources reviewed.

Kidney plasma flow rose 29% with serelaxin and 14% with placebo, a 13% relative increase with serelaxin (P=0.0386). GFR changes between 8 and 24 hours did not differ significantly. Filtration fraction rose 36% with serelaxin and 62% with placebo, a 16% relative decrease with serelaxin (P=0.0019). Systolic blood pressure moved in a similar way in both groups, and creatinine clearance showed no difference.

• A small Phase 2 mechanism study with a single 24-hour infusion; it measured kidney blood-flow markers, not symptoms, kidney outcomes or survival.
• Only 65 of the 118 enrolled patients were in the primary analysis, and all 22 patients from one site were excluded because of implausible measurements.
• Only the PubMed abstract was reviewed, and the registry posts no results, so numbers could not be cross-checked.
• The study enrolled patients with chronic heart failure and reduced ejection fraction; 89% of those analysed were men, so results may not apply to other patients.
• The authors' statement of 'beneficial renal hemodynamic effects' is an interpretation of surrogate measures and was not shown to translate into clinical benefit.

• Number of patients: ClinicalTrials.gov reports 118 participants enrolled; Circulation: Heart Failure abstract reports 65 patients in the primary analysis; all 22 patients from one site excluded before unblinding because of implausible measurements. The abstract accounts for 22 of the 53 people not in the primary analysis; the reasons for the rest are not given in the abstract. The reason for the difference is not established from the sources reviewed.