Adding Metformin to Insulin May Not Reduce Mortality
By Laurie Barclay, MD
Medscape Medical News
April 23, 2012 — The combination of metformin and insulin was no better than insulin alone for reducing all-cause or cardiovascular mortality in patients with type 2 diabetes, according to the results of a systemic review and meta-analysis published online April 19 in the British Medical Journal.
"I doubt this will change how we currently practice nor affect any professional consensus recommendations," Brian K. Irons, PharmD, told Medscape Medical News in an email interview when asked for independent comment.
In fact, Dr. Irons explains, the study actually provides “additional data to support the combination of both metformin and insulin, as the combination can limit the weight gain associated with insulin use." Dr. Irons is division head of ambulatory care and associate professor of pharmacy practice at the School of Pharmacy, Texas Tech University Health Sciences Center in Lubbock. "Metformin may still provide some benefit in limiting gluconeogenesis and to a lesser extent improve insulin sensitivity, which may help attenuate the amount of insulin that is needed for a given patient, particularly in those who are overweight or obese."
The goal of this systematic review of randomized clinical trials, which used meta-analyses and trial sequential analyses, was to compare the benefits and harms of metformin and insulin vs insulin alone in patients with type 2 diabetes. Bianca Hemmingsen, PhD student at the Copenhagen Trial Unit, Centre for Clinical Intervention Research at Rigshospitalet, Copenhagen University Hospital in Denmark, and colleagues searched the literature for pertinent randomized, controlled trials.
This search included the Cochrane Library, MEDLINE, Embase, Science Citation Index Expanded, Latin American Caribbean Health Sciences Literature, and CIANHL until March 2011. Other sources included abstracts presented at the American Diabetes Association and European Association for the Study of Diabetes, bibliographies of identified trials, and the US Food and Drug Administration Web site. The reviewers also contacted investigators of pertinent trials and pharmaceutical companies.
Inclusion criteria for studies in the meta-analysis were randomized clinical trials comparing metformin and insulin vs insulin alone, with or without placebo, in patients older than age 18 years with type 2 diabetes. The intervention period had to be 12 weeks or longer. Publication language or status, predefined and reported outcomes, and antidiabetic interventions used before enrollment were not reasons for exclusion.
Of 26 randomized trials identified, enrolling a total of 2286 participants, 23 trials enrolling a total of 2117 participants had usable data. Risk of bias was high for all trials, and there were few data for patient-relevant outcomes.
Compared with insulin alone, metformin and insulin were not associated with any significant difference in all-cause mortality (relative risk [RR], 1.30; 95% confidence interval [CI], 0.57 - 2.99) or cardiovascular mortality (RR, 1.70; 95% CI, 0.35 - 8.30). Trial sequential analyses revealed that additional trials were needed to allow reliable conclusions concerning these outcomes.
Severe hypoglycemia occurred significantly more frequently with metformin and insulin than with insulin alone, according to a fixed-effect model but not a random-effects model (RR, 2.83; 95% CI, 1.17 - 6.86). Compared with insulin alone, metformin and insulin were associated with reduced hemoglobin A1c, weight gain, and insulin dose in a random-effects model. Trial sequential analyses showed sufficient evidence for lowering of hemoglobin A1c by 0.5 percentage point, weight gain by 1 kg, and insulin dose by 5 U/day.
"There was no evidence or even a trend towards improved all-cause mortality or cardiovascular mortality with metformin and insulin, compared with insulin alone in type 2 diabetes," the study authors write. "Data were limited by the severe lack of data reported by trials for patient relevant outcomes and by poor bias control."
Strengths and Limitations
"This is the first meta-analysis that I know of that directly compares the combination of metformin with insulin compared to insulin alone," Dr. Irons said. He notes that it is admirable that the primary outcomes of the meta-analysis were focused on all-cause and cardiovascular mortality.
However, “[t]he individual trials were not focused on mortality (cardiovascular or all-cause), and as such, the total number of outcomes between groups is quite small and limits the ability for adequate comparison," Dr. Irons said. "Due to its limitations, [the study] cannot answer the question of 'is there still a potential cardiovascular or all-cause mortality benefit of metformin when insulin is added to it?'"
Other limitations noted by Dr. Irons included small size and relatively short duration of most of the trials. Only 2 of the evaluated studies lasted longer than 1 year, and most were less than 6 months in duration. As is typical for meta-analyses, determination of outcomes and inclusion/exclusion criteria often differed among the individual trials.
"Metformin's cardiovascular mortality benefit, from a prospective clinical trial standpoint, has only minimal support and has not been shown to be more effective in this capacity when compared to other diabetes agents," Dr. Irons concluded. "If you truly want to answer the question: is a cardiovascular mortality benefit of metformin retained with the addition of insulin, then large, prospective studies adequately powered to detect this difference are needed."
The Copenhagen Trial Unit funded this study. All study authors declare that: the study received funding from the Copenhagen Insulin and Metformin Therapy Trial Group. Four of the authors have reported equity in Novo Nordisk A/S; 2 have received fees from Novo Nordisk A/S for speech making; and one was employed at Steno Diabetes Centre, Gentofte, Denmark, when the systematic review began. Another author is employed at Steno Diabetes Center, which is an academic institution owned by Novo Nordisk A/S. After the initial draft of the present manuscript, 1 of the authors became employed at Boehringer Ingelheim in Germany. Dr. Irons has disclosed no relevant financial relationships.
BMJ. Published online April 19, 2012.
