Sometimes we endocrinologists do not really know the answer to some of the most fundamental questions of our daily encounters. It took nearly 70 years of insulin availability before we knew for certain that intensive therapy reduced end-organ disease in people with type 1 diabetes. Even with open-and-shut results, dilemmas continue, such as how much hypoglycemia is an acceptable trade-off to protect vital organs. The need to reduce this complication encouraged additional research, resulting in insulins with better predictability, though the learned result of aggressive glycemic management has influenced medical care to this day.
Since then, the more prevalent type 2 diabetes has taken center stage, with a dramatic rise in new diagnoses and devastating end-organ meltdowns that are only partially manageable. Although metformin was available for 25 years before US Food and Drug Administration approval, only afterwards did studies establish it as the agent of choice for introduction of hypoglycemic therapy, now largely accepted.
Subsequent to this, the diabetes toolbox has expanded to include basal and rapid insulins that do not appear in nature, thiazolidinediones (TZDs), GLP-1 receptor agonists, oral DPP-4 inhibitors that enhance availability of endogenous GLP-1, and most recently, the SGLT2 inhibitors. As choices have expanded, we've all faced the challenge of choosing the best sequence of medications for patients not adequately maintained on metformin monotherapy.
In the absence of real data, we've developed our preferences on the basis of personal criteria — from our own experience, the practices of trusted colleagues, the scripted pitches of the company representatives telling us why their drug is superior to the others; perhaps newer or more economical is better. What we've lacked were direct comparisons of metabolic control and ultimate outcomes. These assessments were recently published in the landmark study of diabetic medication selection, "Glycemia Reduction Approaches in Diabetes: A Comparative Effectiveness Study (GRADE)," reported in two articles, one on glycemic outcomes and the other on microvascular and cardiovascular outcomes, and an editorial in The New England Journal of Medicine.
The study design illustrates some of the challenges of navigating a medical environment where new options enter while others declare themselves with adverse effects. This 36-center effort recruited patients with type 2 diabetes who were diagnosed within the previous 10 years and had tolerated metformin. Their A1c levels measured between 6.8% and 8.5% after a run-in period of 2-3 months during which the metformin dose was increased to the maximum tolerated. Then a second agent was added, with the 5000 volunteers randomly assigned to receive either glargine, liraglutide, glimepiride, or sitagliptin. Each drug dose was gradually escalated to the effective and tolerated dose.
Because studies of this type require a long lead-in, TZDs were not included due to safety concerns. SGLT2 inhibitors posed a design challenge, having been introduced to the marketplace after the protocol was set, yet established as a helpful, if not essential medicine for people with some renal or cardiac conditions. They could not be withheld, but their use was left to the discretion of each patient's physician.
Protocols were established for recording as events an A1c of about 7.0%, above 7.5%, and intensification of observation should it rise above 9.0%. When two consecutive A1c values rose above target, basal insulin was introduced to those assigned to the other three agents and increased in those assigned to glargine.
In addition to monitoring glycemia, participants were assessed for markers of cardiac and renal risks, including blood pressure and lipid therapy. Some was left to the discretion of treating physicians. Many who were not already on an antihypertensive or statin on entry into the study had this added as the quarterly assessments proceeded, suggesting that overall care became more meticulous as a consequence of enrollment.
Similar to every other experienced prescriber, I had developed my own impression of which drugs always helped and which were overpromoted and undereffective in my hands. Prescribers in practice also have to take cost into account, along with sometimes having to convince patients that they will not find daily injections that intrusive. By contrast, research protocols have willing users who are unaffected by expense.
On following these people for 5 years, the need to intensify treatment occurred in the vast majority, ranging from 67% with insulin to 77% with sitagliptin. Among the sitagliptin cohort, treatment failure occurred within the first year 55% of the time, compared with about 40% with the other medications. Both injectables outperformed both oral agents, though neither injectable outperformed the other.
Assessing benefits to cardiovascular risk, liraglutide had a slight edge over the others using lab targets, not events, with the 5-year time frame probably too short for this to emerge. More weight loss followed DPP-4 use and GLP-1 therapy than with insulin or the sulfonylurea. Hypoglycemia was most frequent with glimepiride, though at 2%, not an overwhelming barrier for use of the least expensive option.
While the purpose of the project was to compare medication options, which it did, the results just as significantly reinforced what we already knew about type 2 diabetes as a disease. The cynics among us realize that if you have a lot of different reasonable options, expect none of them to be optimal. The results confirmed this. Even among those entering the phase-in period with the lowest A1c, the majority still could not maintain an A1c under 7% for the duration of the study, despite the addition of one and sometimes two agents. Those starting with higher baseline A1c values saw their second agent fail faster.
On the reassuring side, the non-White and Native American participants, whose diabetic outcomes often underperform, seemed to do as well as everyone else when provided medication and consistent professional reassessment. This information is likely to enable better diabetic care, even if the medicines themselves leave something to be desired.
As with all "definitive" studies, we are left with unanswered questions. Are these patients like our patients? For primary care practices they are, though we endocrinologists often have a different population. Most days our patients' A1c levels are well above the 8.5% maximum recruited for the study. Perhaps the choice of drug would matter more for our patients with more severe hyperglycemia. These participants got by on only one drug. Our patients are often already taking two or three, yet with ongoing hyperglycemia.
The study population had been diabetic for less than 10 years and was followed for another 5 years. What happens when the time frame changes from years to decades? Do people with longer duration of disease or less pancreatic reserve respond better to one agent preferentially over another?
As much as this important assessment adds to our knowledge and will influence how diabetes is managed in its early years, we never quite run out of things we wish we knew better. Just as new options were introduced from the 1990s onward, it is likely that the options will continue to expand as the studied patients survive from 10 years to 40 years with their illness — maybe even options that generate more enduring glycemic control.
Lead image: Zaretskaya/Dreamstime
Medscape Diabetes © 2022 WebMD, LLC
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
Cite this: Landmark Study Assesses Treatments for Type 2 Diabetes - Medscape - Nov 08, 2022.