Study: Many non-oncologic surrogate markers lack meta-analyses linking them to outcomes

Over half of the non-oncologic surrogate markers listed in the US Food and Drug Administration’s (FDA) Adult Surrogate Endpoint Table do not have a meta-analysis of clinical trials evaluating the relationship between the treatment effect associated with the marker and a clinical outcome, according to a recent study published in JAMA.

Joshua Wallach, of the department of epidemiology at the Rollins School of Public Health, Emory University in Atlanta, and colleagues performed a review of meta-analyses, systematic reviews, and pooled analyses of clinical trials where surrogate markers were used as primary endpoints in non-oncologic chronic diseases, identifying 37 surrogate markers from FDA’s Adult Surrogate Endpoint Table across 21 chronic diseases up to March 2023.

While surrogate markers in the study were used as primary endpoints for drugs in clinical trials for diseases under both accelerated approval and traditional approval pathways, the researchers identified 17 surrogate markers (77%) that were used for drugs being considered for traditional approval. Surrogate markers in the study classified by the FDA as appropriate for accelerated approval included markers for Alzheimer's disease (3 studies) and primary glomerular disease (1 study). For traditional approval pathways, the surrogate markers included hypercholesterolemia (11 studies), hypertension (10 studies), type 2 diabetes (9 studies), chronic obstructive pulmonary disease (6 studies), osteoporosis (6 studies), hypertriglyceridemia (3 studies), chronic kidney disease (2 studies), gout (2 studies), HIV (2 studies), hyperphosphatemia (1 study), pulmonary fibrosis (1 study), and secondary hyperparathyroidism (1 study).

Wallach and colleagues found that 22 of 37 surrogate markers (59%) used as primary endpoints for chronic diseases did not have a meta-analysis linking the treatment effect measured by the surrogate marker and a clinical outcome. For 15 surrogate markers (41%), there was at least one meta-analysis, and a total of 54 unique meta-analyses across these markers. Within these 54 meta-analyses, there were 200 associations linked to 109 unique surrogate marker–clinical outcome pairs, and of these, 59 pairs (54%) had a reported coefficient of determination, with 10 of 59 (17%) having high-strength evidence and 50 pairs (46%) showing “slopes, effect estimates, results from meta-regression analyses, or another metric of association or prediction.”

“For drugs receiving accelerated approval, the expectation is that they fill an unmet medical need. Therefore, we may be more willing to accept the uncertainty introduced by a surrogate marker (given that confirmatory studies are conducted in the postmarket setting),” Wallach said in an interview. “However, for drugs receiving traditional approval, one would expect that there is already evidence supporting the association between the treatment effect on the surrogate marker and the treatment effect on a specific clinical outcome.

FDA could do more to highlight ‘validated’ surrogates

While the validity of surrogate markers like progression-free survival and response rate have been questioned in oncologic clinical trials, there has not been as much research into how this has impacted clinical trials involving non-oncologic drugs, Joseph Ross, professor of medicine and public health at Yale School of Medicine and senior author, said in an interview with Regulatory Focus.

Wallach and Ross both noted that FDA could provide additional information on what evidence is required to validate a surrogate marker. FDA’s requirements for including and removing surrogate markers from their Adult Surrogate Endpoint Table are not known, but FDA does require evidence from meta-analyses as a minimum requirement to use surrogate markers in traditional approval. “There is currently no gold standard, and there are opportunities for the FDA to clarify their criteria for use of surrogate endpoints for chronic diseases,” Wallach said.

“As a step to start, the agency could publicly report and routinely update an accompanying summary of evidence supporting marker surrogacy for clinical benefit, or clarify when no studies are available,” Ross said. “Although individual studies may be the only source of evidence for surrogate markers classified by FDA as being appropriate for accelerated approval, FDA could reconsider whether to allow surrogate markers with low-strength evidence from multiple meta-analyses to be used as primary end points in clinical trials supporting traditional drug approval.”

Together with other prior research in oncology, the study raises “concerns about the validity and reliability of the clinical trial evidence demonstrating product efficacy when the trials used surrogate markers as primary efficacy end points,” Ross said. “Some surrogate markers are valid and strong proxies, such as FEV₁ for COPD and proteinuria for primary glomerular disease, but many other surrogate markers lack the same high-strength evidence. More research is needed to validate these surrogate markers if our regulators continue to allow them for product approval trials, or if there is information beyond the published literature that suggests that they are better validated, that information needs to be made public.”

JAMA Wallach et al.

Editor’s note: Ross is a deputy editor at JAMA but was not involved with editorial decisions concerning the article.

Study: Many non-oncologic surrogate markers lack meta-analyses linking them to outcomes

Related topics

Recommended content

Welcome to the new RAPS Digital Experience