Predicting 6-year mortality risk in patients with type 2 diabetes

Abstract

Objective: The objective of this study was to create a tool that predicts the risk of mortality in patients with type 2 diabetes.

Research design and methods: This study was based on a cohort of 33,067 patients with type 2 diabetes identified in the Cleveland Clinic electronic health record (EHR) who were initially prescribed a single oral hypoglycemic agent between 1998 and 2006. Mortality was determined in the EHR and the Social Security Death Index. A Cox proportional hazards regression model was created using medication class and 20 other predictor variables chosen for their association with mortality. A prediction tool was created using the Cox model coefficients. The tool was internally validated using repeated, random subsets of the cohort, which were not used to create the prediction model.

Results: Follow-up in the cohort ranged from 1 day to 8.2 years (median 28.6 months), and 3,661 deaths were observed. The prediction tool had a concordance index (i.e., c statistic) of 0.752.

Conclusions: We successfully created a tool that accurately predicts mortality risk in patients with type 2 diabetes. The incorporation of medications into mortality predictions in patients with type 2 diabetes should improve treatment decisions.

Figures

Figure 1

Mortality curve over time.

Figure 2

Nomogram for predicting 6-year probability of survival. Instructions for using the nomogram are as follows. Estimate the patient's GFR from his or her most recent serum creatinine level. Locate the value of the patient's age according to baseline medication and GFR in A, draw a line straight upward to the Points axis to determine the number of points contributed by age. Repeat this process for the other variables in the model. Sum the points achieved for each predictor in A. Repeat this process in B. Sum the points obtained in both parts of the nomogram, and find this total on the Total Points axis at the bottom of B. Draw a straight line down from the total points axis to determine the probability of 6-year survival. An important point to note about nomograms is the U-shaped relationship. In this nomogram, for instance, the LDL cholesterol predictor variable has a U-shaped relationship with the probability of survival. This is presented in the nomogram by having the direct relationship on one side of the scale and the indirect relationship on the other side of the scale. LDL cholesterol levels from 150 to 0 are shown under the scale and have a direct relationship with survival, whereas LDL cholesterol values from 150 to 450 are shown on the top of the scale and have an indirect relationship with survival. In other words, a patient with an LDL cholesterol of exactly 150 has the highest probability of survival, and as the LDL cholesterol goes up or down from 150, the risk of mortality increases. An example of use of the nomogram is the following. A 50-year-old man with type 2 diabetes presents today for his first visit at Cleveland Clinic. The physician caring for the patient (Pt) wants to know the risk of mortality for this specific patient over the next 6 years if he or she prescribes a BIG. Here are the characteristics for this patient along with the calculation using the survival nomogram: age 50 years, taking BIG, GFR 60 ml/min (18 points); A1C 10.0% (3 points); BMI 35 kg/m2 (0 points); systolic blood pressure 140 mmHg (0 points); diastolic blood pressure 80 mmHg (3 points); HDL cholesterol 35 mg/dl (6 points); LDL cholesterol 100 mg/dl (1 point); triglycerides 200 mg/dl (1 point); male sex (3 points); Caucasian (7 points); no heart disease (0 points); no heart failure (0 points); no smoking (0 points); no insulin (0 points); no ACE/ARBs (5 points); not newly diabetic (10 points); aspirin, yes (0 points); no clopidogrel (0 points); and no lipid-lowering drugs (0 points). Total points = 57. Probability of 6-year survival ∼0.94. Nomogram calculator available online from http://www.clinicriskcalculators.org.

Figure 3

Validation of the survival prediction. Vertical bars represent the 95% CIs by quintile. The 45° line represents a perfect prediction.