Metformin is a hypoglycemic agent that is used in the treatment of type 2 diabetes. It is considered the drug of choice in patients who are overweight with no contraindications to its use (1). Metformin has been associated with lactic acidosis. Although metformin-associated lactic acidosis (MALA) is rare, it is fatal in approximately half of all cases. MALA is precipitated by clinical conditions that cause substantial tissue hypoperfusion and hypoxia and thus are identified as precautions and contraindications to its use in package labeling (1, 2). Four studies have previously found that metformin is commonly prescribed, despite risk factors, with a range of 24.5–54% in outpatients and 64–73% in inpatients (3, 5–7). The purpose of our study was to determine the frequency of inappropriate prescribing of metformin to patients with precautions or contraindications at an academic health center and secondarily to compare the findings with previously published studies.
This retrospective study was approved by the institutional review board of the University of Michigan. We performed a chart review where 100 inpatients and 103 outpatients were randomly selected from 300 patients who received metformin between January and June of 2003. Patients’ medical records were used to collect demographic data, clinical characteristics, laboratory parameters, and information about contraindications and precautionary conditions. Risk factors for MALA were identified through ICD-9 codes in the clinical data repository, laboratory tests, and medical notes.
A power analysis estimated that a sample size of 78 patients in each group was sufficient to detect a 20% difference in the frequency of inappropriate prescribing compared with the frequency (73%) reported by Holstein et al. (α = 0.05, β = 0.8) (3). A sample size of 100 patients per group was selected for simplicity. Frequencies of inappropriate prescribing were compared with literature-reported rates using Fisher’s exact test (4). Inpatient results were compared with those of Calabrese et al. (5). Outpatient results were compared with those of Emslie-Smith et al. (6).
Of 103 outpatients, 3 (2.9%) had contraindications to metformin use. This is significantly lower than the 24.5% incidence of contraindications reported by Emslie-Smith et al. (P = 0.0001) (6). The contraindication present in all three of these patients was congestive heart failure requiring drug therapy.
Of 100 inpatients, 68 (68%) had contraindications or precautionary conditions to metformin use. This incidence of inappropriate prescribing is not significantly different (P = 0.2524) from the results of Calabrese et al., who found a similar incidence of 62% of inpatients receiving metformin. In our inpatient sample, the most commonly seen precautions were concomitant cationic drug use and failure to measure serum creatinine before restarting metformin after a surgical procedure. No cases of metabolic acidosis were documented. Approximately 25% of inpatients had more than one precautionary or contraindicating condition; however, there were no patients with more than three risk factors.
We have demonstrated that in our health system, prescribing patterns do not conform to the guidelines established by the manufacturer and the Food and Drug Administration. While the incidence of inappropriate prescribing for outpatients is lower than that reported in the literature, inappropriate use of metformin among inpatients is frequent and occurs at a rate similar to that reported in the literature. Frequent metformin use in patients with risk factors was not accompanied by lactic acidosis in our population. Reexamination of these precautionary and contraindicating conditions in light of extensive worldwide experience with metformin could be helpful in more precisely identifying patients who are likely to develop MALA. This would minimize unnecessary avoidance of the drug in patients with type 2 diabetes who could benefit from treatment.
