Consider The Evidence: Med/Peds Journal Roundup

January 3, 2010

Insulin v. Triple oral therapy for DM

Filed under: diabetes — medblog @ 12:21 pm

Diabetes Care 12/23/09 –   “Insulin is the most effective hypoglycemic agent in our treatment armamentarium and is now recommended by the American Diabetes Association (ADA) guidelines as the second agent added after metformin” (FTA). Despite this fact (or because some of us were unaware of this guideline), physicians often prefer to max out that second and third oral medication before turning to insulin. An article in Diabetes Care looks at a comparison between a three drug regimen and insulin + metformin for newly diagnosed DM.

The study compared 2 groups of 29 subjects that were newly diagnosed diabetics (not a very large study). All patients had a 3 month lead-in with insulin (novolog 70/30 bk, ln) and metformin. The lead-in ensured that all subjects had controlled blood sugar when starting the comparison phase, and all were exposed to insulin administration. Subjects were then randomized to receive metformin + glyburide + pioglitazone or to continue metformin and insulin. Groups were compared on A1C, Quality of Life, and treatment satisfaction.

They found that A1C was similar in both treatment groups. But I think the other endpoints were just as, if not more important to the relevance of the study.   Weight gain, quality of life, and compliance are major factors that sway physician and patient choices on when to start insulin.

The study showed that the weight difference was not significant in the 2 groups – however looking at the raw numbers weight increased by about 4kg in the insulin group and about 7kg in the oral group. This seems like a big difference to me to end up being statistically insignificant. The CI was 0.89 – 8.04kg so I really wonder if the study was adequately powered to look at this. QoL and patient satisfaction were similar in the 2 groups, however I wonder if the patients that really had a problem with insulin were lost during the lead-in period. No information was provided about drop out during the lead-in time and the study ends up being selective for those patients that tolerated a 3 month period on insulin and agreed to enter the study, so I think there is a selection bias going on here.

Overall I think the study shows for those patients with new onset DM that need initial insulin to get them under control, there is little point to switching them back to orals if they are going to require 3 medications. The more perplexing scenario, however, is those patients that the physician believes does not necessarily need insulin as initial therapy but for which insulin or triple oral therapy may be options. And this study does not really address that because all patients were put on insulin for 3 months prior to study onset.

Diabetes Care October 2009 vol. 32 no. 10 1789-1795


Blog at