To the Editor: In the study conducted by Ren-Rong Wu, M.D., et al. (1), published in the March 2008 issue of the Journal, the effect of dose was successfully controlled by assigning the two study groups (olanzapine plus metformin group and olanzapine plus placebo group) to the same dosages. However, although the groups’ baseline and endpoint scores on the Scale for the Assessment of Positive Symptoms and Scale for the Assessment of Negative Symptoms did not differ, it might have been necessary to adjust the olanzapine dosage based on the patients’ clinical presentations. The difference in the Treatment Emergent Symptom Scale scores during the follow-up visits was not reported, which exposed patients to two medications with potentially serious adverse reactions without clinical evaluation.
Two inclusion criteria were 1) no antipsychotic treatment for at least 3 months before enrollment in the study—not antipsychotic naive—and 2) first psychotic episode of schizophrenia. However, the tables in the article indicated that the mean duration of illness was 6.8 years for the metformin group and 7.6 years for the placebo group.
Reporting the completer analysis eliminated the randomization effect. Simultaneously, in the intent-to-treat analysis, the authors used the last observation carried forward method. Although widely accepted, this method remains biased (2). Last observation carried forward assumes that responses following dropout remain constant at the last observed value. In this instance, special attention should be given, since the longer a patient is treated with olanzapine, the more weight the patient is likely to gain. Dr. Wu et al. demonstrated this in their study when the difference in weight gain diverged after 8 weeks of follow-up. In relying on this supposedly more conservative statistical method, the effect of metformin may have been overestimated.
Metformin is known to attenuate weight gain by reducing caloric intake (3). Accordingly, reduced caloric intake is a potential mediator in the association between metformin use and weight loss but is also an independent cause of weight loss. In the study conducted by Dr. Wu et al., there were no data pertaining to caloric intake difference between the two study groups at baseline or follow-up. Monitoring caloric intake but not necessarily controlling for it may help better explain this association.
Although no serious adverse events were reported in the active treatment group, one must consider that metformin is stamped with a black box warning as a result of the feared, albeit debated (4), lactic acidosis. A thorough risk-benefit assessment should be undertaken when considering the use of metformin to prevent the metabolic derangement associated with olanzapine as opposed to resorting to a more benign atypical antipsychotic agent. Justifying the use of this combination in antipsychotic-naive patients with a first psychotic episode seems to be a more challenging task.
1.Wu RR, Zhao JP, Guo XF, He YQ, Fang MS, Guo WB, Chen JD, Li LH: Metformin addition attenuates olanzapine-induced weight gain in drug-naive first-episode schizophrenia patients: a double-blind, placebo-controlled study. Am J Psychiatry 2008; 165: 352–358
2.Ware JH: Interpreting incomplete data in studies of diet and weight loss. N Engl J Med 2003; 348:2136–2137
3.Paolisso G, Amatao L, Eccellente R: Effect of metformin on food intake in obese subjects. Eur J Clin Invest 1998; 28:441–446
4.Salpeter S, Greyber E, Pasternak G, Salpeter E: Risk of fatal and nonfatal lactic acidosis with metformin use in type 2 diabetes mellitus. Cochrane Database Syst Rev 2006; CD002967
The author reports no competing interests.
This letter (doi: 10.1176/appi.ajp.2008.08030432) was accepted for publication in May 2008.