When a treatment appears to work in a clinical trial, one of the most practical questions is: how long does it actually take to see that effect? For people living with a form of heart failure closely linked to obesity, a 2026 systematic review and meta-analysis published in the Journal of Cardiac Failure tried to answer exactly that question for a class of peptide therapies known as incretins.
The analysis pooled data from four randomized controlled trials involving 4,149 participants. All of the participants had overweight or obesity and a type of heart failure where the heart's pumping function is relatively intact, described clinically as heart failure with mildly reduced or preserved ejection fraction. The researchers used a statistical method that recalculated outcomes day by day across the entire follow-up period, allowing them to identify the precise point at which benefits became statistically meaningful and stayed that way.
What incretin peptides are
Incretin peptides are compounds that mimic hormones naturally released in the gut after eating. They act on receptors throughout the body, including in the brain, pancreas, and heart. The two specific molecules studied in the included trials were semaglutide, a GLP-1 receptor agonist, and tirzepatide, which targets both GLP-1 and GIP receptors.
Both compounds have been studied extensively in the context of metabolic disease. More recently, researchers have turned their attention to how these peptides interact with the cardiovascular system, particularly in patients whose heart disease appears to be driven at least partly by excess adipose tissue. The current meta-analysis is one of the first to look not just at whether these therapies help, but when that help begins to appear in the data.
The obesity and heart failure connection
Heart failure with preserved ejection fraction is considered the most obesity-related form of heart failure. In this condition, the heart muscle pumps blood out normally in terms of fraction, but the heart has trouble relaxing and filling properly between beats. Excess body weight contributes to this problem through several pathways, including inflammation, increased blood volume, and mechanical pressure on the heart.
Worsening heart failure in this population typically means hospitalizations, urgent medical visits, or death from cardiovascular causes. Reducing the rate of these events is the central goal of most trials in this space, and it is the primary outcome the meta-analysis examined.
Trial design and methods
The research team searched three major medical databases and included only randomized controlled trials, meaning studies where participants were randomly assigned to either receive an incretin peptide or a placebo. All included trials required at least 52 weeks of follow-up, and the weighted average median follow-up across the four trials was just over 2.4 years.
To estimate the time to benefit, the researchers used a technique called reconstructed individual participant data. Because most published trials report summary statistics rather than raw participant records, the team used established mathematical methods to reconstruct what the data from individual participants would have looked like. They then ran their statistical tests repeatedly, each time cutting off the data at a different point in the follow-up timeline, to find the day when the results first became statistically significant and when they remained consistently significant from that point forward.
The certainty of evidence was graded using the GRADE framework, a widely used system in systematic reviews. Risk of bias across all four trials was rated as low.
Key findings from the pooled data
When looking at the combined outcome of worsening heart failure events or cardiovascular death, incretin-based therapies were associated with a hazard ratio of 0.59 compared to placebo. A hazard ratio below 1 indicates fewer events in the treatment group. The 95 percent confidence interval ranged from 0.45 to 0.78, and the certainty of evidence was rated as moderate.
The first day on which this result crossed the threshold for statistical significance was day 126, or approximately 4.1 months into treatment. Importantly, the results did not remain consistently above that threshold until day 162, about 5.3 months in. The distinction matters because statistical fluctuations can cause a result to briefly appear significant before settling back, so researchers typically look for the point of sustained significance rather than the first crossing.
For the narrower outcome of worsening heart failure events alone, excluding cardiovascular death, the effect was even more pronounced. The hazard ratio was 0.33, with a confidence interval of 0.20 to 0.54, and the certainty of evidence was rated as high. Sustained statistical significance for this outcome was reached after day 183, or approximately six months.
What the timing means for research
The six-month window for sustained benefit has potential implications for how future trials are designed and how long follow-up periods need to be. If trials are too short, they may not capture the window in which these effects become detectable and stable. The authors of the meta-analysis suggest that approximately six months represents the minimum follow-up duration researchers should consider when designing studies of incretin peptides in this patient population.
The finding also adds nuance to how effect sizes are interpreted. A treatment might appear to have a modest effect at a short follow-up point but show a stronger and more consistent signal once the six-month threshold is reached. The iterative day-by-day approach used here is not yet standard in cardiovascular research, but it offers a more granular picture than simply reporting results at the trial's end.
Limitations and context
The meta-analysis has several limitations the authors acknowledged. Because the analysis relied on reconstructed rather than actual individual participant data, there is inherent imprecision in the day-by-day calculations. The four included trials also differed in certain design features and patient characteristics, and pooling across them introduces the usual uncertainties of any meta-analytic approach.
The literature also does not yet address whether the benefits observed here apply uniformly across all subgroups within this patient population, such as people with different degrees of obesity, different ages, or different underlying cardiovascular histories. Future trials with longer follow-up and pre-specified subgroup analyses will be needed to clarify those questions.
As with all findings from observational pooling of trial data, the results describe associations measured in controlled research settings. They describe what researchers found in that specific population and should not be read as guidance for individual clinical decisions.
