The way that type 2 diabetics respond to initial treatment could predict the long term effectiveness of their treatment regimen…
There are a variety of treatment options for type 2 diabetics, with no clear reason why some drugs work better than others in some individuals. That is why diabetes treatment guidelines focus on custom regimens for individual patients based on a variety of factors, including disease progression, risk of hypoglycemia, and life expectancy. To help practitioners predict the long term efficacy of initial drug choices for diabetic patients, the short term or initial response that a patient experiences may indicate whether that patient will remain responsive to that drug. This suggestion has proven effective in other areas of practice, including psychiatry (with depression, schizophrenia, bipolar disorder) and biotherapy (with plaque psoriasis, rheumatoid arthritis). There is not a lot of information available that indicates that practitioners can use this model in the treatment of diabetes but if practitioners are afforded the opportunity to appropriately calculate a patient’s response to medication, there is an increased chance for improved outcomes due to better treatment choices and/or modifications and decreased time on substandard therapy. This study examined how predictive early response to treatment is to consequent treatment response using three frequently prescribed medications: metformin, a sulfonylurea, and insulin glargine. If participants responded early, at 12 weeks, it was expected that they would also respond at 24 and 52 weeks.
Data from previous randomized trials were used to examine patient treatment response, which included 2,269 patients. Patients being treated with metformin and sulfonylurea were evaluated at 12, 24, and 52 weeks. Patients being treated with insulin glargine were evaluated at 12 and 24 weeks. Early treatment response was identified as a 1% or more decrease in hemoglobin A1C or a hemoglobin A1C less than 7%. There were 4 predictive parameters identified: sensitivity (% of consequent responders accurately identified), specificity (% of consequent non-responders accurately identified), positive predictive value (% of consequent responders among early responders), and negative predictive value (% of consequent non-responders among early non-responders).
The sensitivity, specificity, positive predictive value, and negative predictive value measures for improved patient hemoglobin A1C were as follows: metformin – at 24 weeks [0.83, 0.81, 0.44, 0.96] and at 52 weeks [0.73, 0.84, 0.56, 0.92]; sulfonylurea – at 24 weeks [0.79, 0.94, 0.71, 0.96] and at 52 weeks [0.45, 0.94, 0.74, 0.82]; insulin glargine – at 24 weeks [0.67, 0.89, 0.65, 0.90]. There was a high correlation between measured hemoglobin A1C levels and change in A1C levels from baseline at 12 weeks versus 24 weeks and 12 weeks versus 52 weeks. At 12 weeks versus 24 weeks, the correlation coefficient for measured hemoglobin A1C levels ranged from 0.76-0.87, and the correlation coefficient for change in hemoglobin A1C levels ranged from 0.84-0.89. At 12 weeks versus 52 weeks, the correlation coefficient for measured hemoglobin A1C levels ranged from 0.72-0.73, and the correlation coefficient for change in hemoglobin A1C levels ranged from 0.76-0.80.
Negative predictive measures were consistently high, demonstrating that if a patient does not respond within 12 weeks of initiating treatment, there is an 89-98% chance of treatment failure if that treatment is continued. Based on the correlation coefficients, this study supports the idea that long term treatment response can be accurately predicted based on a patient’s short term or initial treatment response.
- After starting a patient on a treatment regimen, treatment effectiveness should be evaluated as early as 12 weeks later.
- Early response to treatment can be an indication of long term treatment effectiveness.
- If a patient is a non-responder after 12 weeks, there may be other causal factors (i.e. noncompliance).
Fu H, Cao D, Boye K S, et al. “Early Glycemic Response Predicts Achievement of Subsequent Treatment Targets in the Treatment of Type 2 Diabetes: A Post Hoc Analysis.” Diabetes Therapy (2015): n. pag. Web. 7 July 2015.