Insulin or Sulfonylureas to Supplement Metformin in Patients with Diabetes?

Summary and Comment |
June 10, 2014

Insulin or Sulfonylureas to Supplement Metformin in Patients with Diabetes?

  1. Thomas L. Schwenk, MD

Insulin plus metformin was associated with higher all-cause mortality in a retrospective study.

  1. Thomas L. Schwenk, MD

Metformin and lifestyle modifications are the mainstay of initial treatment for patients with type 2 diabetes. The choice of a second medication for patients who do not achieve full control with this regimen is unclear. Investigators conducted a retrospective cohort study in the Veterans Affairs healthcare system and identified patients who had initiated metformin and then received either insulin or a sulfonylurea to achieve better glycemic control between 2001 and 2008. About 12,000 sulfonylurea recipients and about 2500 insulin recipients were matched closely for many demographic and clinical covariates through propensity scores. Median age was 60, and median glycosylated hemoglobin (HbA1c) level was 8.1% in both groups.

The combined endpoint of acute myocardial infarction and stroke occurred in a similar proportion of each group (10.2 per 1000 person-years for insulin vs. 11.9 per 1000 person-years for sulfonylureas), but all-cause mortality through 2009 was 44% more common in those receiving insulin — a significant difference.


Recent studies of patients with type 2 diabetes have suggested that monotherapy with sulfonylureas is associated with worse cardiovascular outcomes than monotherapy with metformin (NEJM JW Gen Med Dec 6 2012 and NEJM JW Gen Med Jun 18 2013); in contrast, in the current study, researchers examined add-on therapy with sulfonylurea or insulin in patients who were already receiving metformin. The authors expected patients who were receiving insulin to fare better because their glycemic control would be tighter, but the results showed the opposite. Clinical recommendations based on these results would be premature, because even high-quality propensity score matching might not account for all confounding factors, such as perceived disease severity, patient and physician preference for treatment approach, and adherence to medications.

Editor Disclosures at Time of Publication

  • Disclosures for Thomas L. Schwenk, MD at time of publication Editorial boards UpToDate


Reader Comments (13)

* * Physician, Internal Medicine, health services kerala

need evaluation for new generation insulin

esunz Physician, Family Medicine/General Practice, Indian Health Service

Unfortunately, given the way Journal Watch has headlined this study, it will be prematurely interpreted as reason to stop metformin and switch to insulin alone where insulin seems indicated. Most would be adding insulin where glycemic control is worsening - and common sense would dictate that not adding insulin would lead to worse outcomes still.

Mongjam MeghachandraSingh Physician, Preventive Medicine, Department of Community Medicine, Maulana Azad Medical College, New Delhi

The present study results should be taken cautiously since there are several other factors which may influence mortality among those receiving either insulin or sulfonylureas for treatment of Type 2 diabetes (T2DM). There are also conflicting reports in contrast to the present study results. We need to find out the other existing comorbidities, risk factors for other morbidities in addition to T2DM. In general practice, insulin is preferred at last as an additional medication when metformin alone is unable to control blood sugar levels. In such cases, addition of sulfonylureas is routinely taken as a first choice and insulin as a second choice. Opinions may differ, use of oral hypoglycemics rather than insulin injection is rather patient friendly in terms of intake route and also undesirable sore in the injection sites of insulin. A lot more well planned studies need to corroborate the present findings.



Janet Physician, Family Medicine/General Practice, Solo Practiceat the Healing Arts Center

Insulin can fuel cancer growth as they prefer the glucose pathway (that's why we do PET scans to detect them) and also need insulin like growth factor to grow and metastasize. Also since lipid particle size is shifted toward more dense atherogenic particles by inflammation (which glucose and insulin both produce) then it stands to reason if we want our type 2 diabetics to be healthy, then we should not keep recommending that they eat carbs that require much insulin to process. Alzheimer's disease is also now being termed "Type 3 Diabetes" (of the brain) due to insulin resistance. They all do better on a slightly ketogenic diet using low inflammatory heart healthy fats and lean protein or wild salmon for fuel along with ample green leafy vegetables and other vegetables that grow above the ground (this doesn't include grains such as corn, wheat, rice etc, or sugars). This is all more what we are designed to eat and it's only been the last century that our diet has become overloaded with quick releasing carbs coupled with a more sedentary lifestyle- the confluence of which leads to most of the chronic disease states (along with the increased contaminants from the environment). A good step by step reference for people wanting to achieve better health by eating a human appropriate diet is "Diet Evolution" by Steven Gundry MD (I am not related to him nor know him, but his recommendations have helped all of my patients who followed it accurately).

William Wiggins Medical Student

Can someone explain to me why they make these big notices if they don´t expect us to take it seriously? I assume they should continue studies, and stay tight lipped until they have something to announce that can be used in general practice.


Need to read the original study: What was the treated mean A1c with both therapies? I assume the A1c mean cited as 8.1% was at entrance to study.

ROBERT DUNN Physician, Emergency Medicine, Royal Adelaide Hospital

Please report absolute values instead of relative values (or at least state which you are using). The mortality difference between the groups was about 2%, not 44%. The overuse of relative values in the medical literature distorts the truth.

* * Physician, Endocrinology

It should be considered premature if decisions are taken based on this study that was obviously biased against insulin-use. LOOK at the wide disparity in the study populations. What statistical tool was used to correct for this very powerful confounding factor?
Also, the age, mainly elderly patients would likely have more hypoglycaemic episodes with insulin use..

Tony Kelpie MRCGP Physician, Family Medicine/General Practice, Southampton England
Competing Interests: I am a Type 1 diabetic

Hypoglycaemia is one element which may contribute to any excess of mortality . It is interesting that the authors expected superior results with Insulin but found the reverse. Also interesting that longstanding concerns re sulphonylureas may also apply to Insulin use in T2DM. Raises questions about the 'put them all on Insulin' approach that some have favoured

JOSE DE JESUS GARCIA MENDOZA Physician, Internal Medicine, issste hospital

Of course, this result is impressive, but certainly I should use insulin plus metformin in sicker patients or at least in those with more difficult glycemic control. So it is time to wait for a deep review or individual analysis.

Pooria.hamedani Resident, Endocrinology, Hospital


KOLLAR JOZEF Other, Cardiology, Slovakia

Several scientist count hyperinsulinemia to risk factors for atherthrombosis. I accept result of this study as credible. Study support risk role of Insulin.

Your Comment

(will not be published)

Filtered HTML

  • Allowed HTML tags: <a> <em> <strong> <cite> <blockquote> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Do you have any conflict of interest to disclose?
This question is for testing whether you are a human visitor and to prevent automated spam submissions.

Vertical Tabs

* Required

Reader comments are intended to encourage lively discussion of clinical topics with your peers in the medical community. We ask that you keep your remarks to a reasonable length, and we reserve the right to withhold publication of remarks that do not meet this standard.

PRIVACY: We will not use your email address, submitted for a comment, for any other purpose nor sell, rent, or share your e-mail address with any third parties. Please see our Privacy Policy.