Re-thinking metformin in the age of automated clinical decision supportDarren Triller | January 12, 2011
Advances in computer technology and integration are increasing the development and utilization of electronic “clinical decision support” (CDS) features, a term used to refer to automated “alerts” or guidance that prompt physicians and other healthcare providers to take certain pre-defined actions, particularly when prescribing medications. While these alerts have shown the ability to improve patient care in some settings and situations, the processes may potentially lead to overly simplistic patient care decisions and limit the ability of a prescriber to apply clinical judgement.
For example, the antidiabetic medication metformin is utilized with extreme caution in patients with chronic kidney disease and other comorbidities, largely due to the high risk of lactic acidosis experienced with its predecessor, phenformin. In fact, FDA labeling presently contraindicates use of metformin in individuals in patients with serum creatinine >1.5mg/dl (males, http://www.drugs.com/pro/glucophage.html). Such a “cut and dried” contraindication lends itself to integration into CDS processes, as laboratory and pharmacy databases can readily identify prescriptions for the medication in the presence of an elevated creatinine serum concentration.
However, recent clinical evidence suggests that the existing creatine-based threshold for contraindication of the drug is far too conservative, and that metformin may actually be beneficial in patients with more advanced stages of kidney disease. Several recently published manuscripts catalogue a growing body of evidence suggestive of clinical and survival benefits afforded by metformin among patients with kidney disease and other comorbidities. Also, they present evidence indicating that risk of lactic acidosis is far lower than that of phenformin, and may actually be no greater than that of other classes of oral hypoglycemics. Interested parties are encouraged to evaluate the following manuscripts on the subject:
- Lactic Acidosis Induced by Metformin (Lalau, Drug Safety, 2010: 33(9) 727-740)
- Metformin Use and Mortality Among Patients With Diabetes and Atherothrombosis (Roussel, Archives Internal Medicine, 2010: 170(21) 1892-1899)
- Review: Metformin: Potential Benefits and Use in Chronic Kidney Disease (Pilmore, Nephrology, 2010: 15(4) 412-418)
These papers and others suggest that metformin may actually be under-utilized in patients with CKD and other “contra-indications”, depriving them of potential clinical and mortality benefits. Considering the rapid progression towards utilization of electronic medical records and CDS applications, pharmacists and other healthcare providers are cautioned not to implement overly simplistic “alerts” for medications such as metformin. Overly intrusive prompts to prescribers and “hard stops” in the prescription and distributive processes may further reduce the utilization of metformin, and result in unintended patient harm through needless avoidance of an efficacious therapy option. [This phenomenon is visible in a recent study of CDS and warfarin use, where a "hard stop" alert following the prescription of an interacting antibiotic led to unintended delays in treatment. The study was stopped prematurely due to such unintended consequences.]
Instead, it may be wise to develop systems that:
- Document the education of prescribers and other professionals regarding use of specific high risk drug(s)
- Optimize the utilization of such agents (e.g. goals of 100% use where indicated and risk-bene ratio positive; 0% where truly contraindicated)
- Immediately identify changes in individual patient risk profiles (e.g. acute renal failure), alerting prescriber(s)
- Prompt immediate and evidence-based responses to real or possible adverse drug events on case by case basis
- Tracks drug utilization, and applies continuous quality improvement principals to the monitoring system.
The availability of electronic systems and integrated data sources provides opportunities for the development of CDS features relating to the use of high risk drugs. However, as is evident from the literature on metformin, contraindications to drugs are not always “cut and dried”. Clinical decision support systems will provide the greatest overall benefit when they optimize the utilization of indicated medications while supporting timely and thoughtful evidence-based responses to potential contraindications and adverse drug events on a case by case basis.