Why Some Diabetes Patients Fail to Respond to Incretin Therapies: Q&A with Dr. Elena Christofides

GLP-1 medications might be most well-known as weight-loss options, but they were originally developed as a treatment for type 2 diabetes. For many patients, the drugs remain a highly effective treatment for the condition.

However, GLP-1s are not perfect and don’t work for every patient. In April of this year, Pharmaceutical Executive spoke with Mark Bagnall, CEO of Phenomix, discussed various reasons why a significant number of patients go off the medication within the first year of using it. His discussion focused on how side effects can be severe and incapitating, while the results can be inconsistent.

Pharmaceutical Executive recently spoke with Dr. Elena Christofides, scientific/medical advisor for Sparrow Pharmaceuticals, about a different issue some GLP-1 users face. Specifically, patients with high cortisol levels face unique challenges that GLP-1s may not be able to tackle.

Pharmaceutical Executive: How do elevated cortisol levels impact type II diabetes treatments?
Dr. Elena Christofides: The incretin class has produced genuinely remarkable medications — significant blockbusters in both diabetes and obesity — but the reality is they don't always work for everyone. Patients are not always achieving the responses that clinical trial data would lead us to expect. That raises an important question: why aren't people reaching their goals, or getting as much response as the average or mean would suggest?

Cortisol is a significant biological agent that controls virtually every system in the body. If a patient has cortisol dysrhythmia, they are not going to achieve the same degree of response in their diabetes or obesity management that the published clinical averages would predict. And epidemiologically, this is proving to affect a meaningful portion of the population — not a small subset.

The mechanism is direct: cortisol disrupts glucose signaling, insulin signaling, hormone integration, and fuel partitioning. If a patient has a cortisol problem, the drugs we assume to be most effective in managing diabetes and obesity will be less effective — and in many cases, ineffective. That makes identifying cortisol dysfunction a critical step in ensuring that the patients being treated with any given medication are actually part of the population that will respond to it. It also has significant implications for clinical trial design: if cortisol dysregulation is not accounted for in patient selection, the efficacy data that results will not reflect what these medications can actually achieve in appropriately selected patients.

PE: What are practical testing options for identifying elevated cortisol?
Christofides: There are numerous ways to test for cortisol, but it is worth being clear about which tests are and are not appropriate. The least effective approach — and one that should never be used for this purpose — is a random cortisol and ACTH value. For reasons I can't fully explain, this still persists in clinical practice, perhaps because we routinely use random testing for other hormones. But the cortisol system cannot be assessed that way. It is too episodic.

The tests that were designed to evaluate cortisol have important limitations worth understanding. The 24-hour urine free cortisol was developed over 50 years ago specifically to identify pituitary disease, which means it has limited utility in adrenal disease. The salivary cortisol test was developed in the 1990s to detect recurrent pituitary disease in patients who had already undergone surgery — making it highly sensitive, but not well-suited for population screening. It can flag people with circadian disruption who don't actually have disease, which significantly limits its specificity.

The only true dynamic test of the HPA axis is the overnight dexamethasone suppression test, or ODST. This test was designed specifically to assess the responsiveness of the cortisol system — because a healthy HPA axis should respond to external steroids. When you administer dexamethasone, you should be able to suppress endogenous cortisol production. That is the hallmark of a functioning system. If you cannot suppress endogenous cortisol with exogenous steroids, you have identified a priori a dysfunctional HPA axis.

Realistically, the ODST is the only test that matters for screening purposes — the only one that tells you whether you are dealing with an HPA axis problem in the first place.

PE: How might cortisol-targeted approached be paired with GLP-1 medications?
Christofides: Diabetes is a complex disease, and when treating it we operate from assumptions about how a patient will respond to prescribed medications based on clinical trial averages. When a patient doesn't respond as expected, that is the signal to think about alternative diagnoses that could be negatively influencing their glucose control or undermining the effectiveness of their medications.

That is the point at which we begin screening for cortisol disruption. By identifying it and adding cortisol-regulating medications to the regimen, we can address the true underlying pathology — the pathology that may have contributed to the patient's diabetes in the first place, that may be driving their hypoglycemia, or that is at minimum significantly limiting the effectiveness of the medications we've already prescribed. The goal is to remove the barrier that is preventing the patient from achieving the outcomes we would expect from that class of medications.

PE: How can doctors be better educated about cortisol issues?
Christofides: We have been doing a significant amount of medical education — for physicians, nurse practitioners, PAs, and even patients — and the most important shift we need to make is getting providers to simply consider that cortisol may be part of the problem.

There is still a degree of blind spot in the medical community around this. Many providers believe that Cushing's syndrome or hypercortisolemia is too rare to be relevant to their practice, so it never enters the conversation. We need to change that starting point — with trainees, with new graduates, and with the providers delivering the bulk of diabetes care. The epidemiologic data shows this is not a small patient population. Cortisol dysregulation is more prevalent than the medical community currently assumes, and that assumption needs to be updated.

The practical implication is this: when a patient doesn't respond to their medications, the first thought should not be that the patient is non-adherent. The first thought should be that something has been missed — that there may be another diagnosis contributing to that patient's lack of response. Historically, the medical world has defaulted to a non-adherence assumption when medications don't work as expected. That framework needs to change. A patient who isn't getting to goal deserves a diagnostic workup, not a behavioral assumption.