Adherence to Creatinine Prescribing Thresholds in a Cohort of Patients with Type 2 Diabetes Mellitus on Metformin – a Retrospective Study

Presentation Number: SUN 285
Date of Presentation: April 2nd, 2017

Akuffo Quarde*1, Alan Scott Sacerdote2 and Gul Bahtiyar2
1Woodhull Medical & mental Health Center, Brooklyn, NY, 2SUNY Downstate Medical Center, Brooklyn, NY


Metformin (MET) has been recommended by many societies as the first line treatment for Type 2 Diabetes Mellitus (T2DM)¹. The United States Food and Drug Administration (US FDA) previously recommended avoiding metformin in the setting of serum Creatinine (Cr) thresholds above 1.4mg/dL and 1.5mg/dL in women and men respectively.

A retrospective cohort of T2DM patients attending outpatient practices in a community based hospital who had received at least a month’s prescription of MET within a year of their last documented serum Cr were identified in a registry. 1008 charts were reviewed prior to April 2016 - 843 charts met inclusion criteria for analysis. The appropriateness of MET therapy was determined based on the previous US FDA prescribing guideline. For patients who received MET inappropriately, at least two subsequent clinic visits were reviewed to document appropriate discontinuation of therapy or lack thereof.

The mean age of subjects was 58.3 years. Male and female subjects represented 40.9% and 59.1% respectively. The mean BMI of our cohort was 31.1kg/m². Based on the previous US FDA guideline, 52 subjects representing 6.2% of the cohort received a MET prescription inappropriately. All but 2 subjects (0.2%) had their metformin prescription discontinued on subsequent clinic visits. Of the subjects who had received MET inappropriately, we estimated the percentage of patients who would qualify for MET under the recent revised product labeling. 91% and 57% of female and male subjects respectively had their GFR being >45ml/min/m².

We compared primary care physicians and nurse practitioners in terms of the incidence of inappropriate MET prescriptions. A Z test for proportions of two groups, using a two-tailed test for difference with 0.05 level of significance was carried out. The estimate of difference between groups was -0.0035 (95% CI for difference of -0.039, 0.032) with a Fischer’s exact test P value = 0.871. We therefore failed to reject the null hypothesis of no difference between the two groups.

Our estimated incidence of inappropriate MET use based on previous US FDA prescribing guidelines, is close to estimates from other smaller retrospective studies. It seems there is a strong adherence to the previous US FDA guidelines across a diverse prescriber population based on our study. With the new guidelines advocating for the use of estimated GFR instead of an absolute serum Cr value in deciding when to initiate and continue MET therapy, prescribers would need to change a long-standing prescribing tendency.

Clinicians should continue to be aware of the possible risk of MET associated lactic acidosis, without necessarily halting the use of a medication with demonstrable benefits in T2DM.


Nothing to Disclose: AQ, ASS, GB