Premium
A new model for 5‐year risk of cardiovascular disease in Type 1 diabetes; from the Swedish National Diabetes Register (NDR)
Author(s) -
Cederholm J.,
EegOlofsson K.,
Eliasson B.,
Zethelius B.,
Gudbjörnsdottir S.
Publication year - 2011
Publication title -
diabetic medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.474
H-Index - 145
eISSN - 1464-5491
pISSN - 0742-3071
DOI - 10.1111/j.1464-5491.2011.03342.x
Subject(s) - medicine , hazard ratio , diabetes mellitus , proportional hazards model , quartile , type 2 diabetes , blood pressure , disease , confidence interval , cardiology , endocrinology
Diabet. Med. 28, 1213–1220 (2011) Abstract Aims We assessed the association between risk factors and cardiovascular disease in an observational study of patients with Type 1 diabetes from the Swedish National Diabetes Register. Methods A derivation sample of 3661 patients, aged 30–65 years, 6.1% with previous cardiovascular disease, baseline 2002, and 197 cardiovascular disease events when followed for 5 years until 2007. A separate validation data set of 4484 patients, baseline 2003, 201 cardiovascular disease events when followed for 4 years. Results Adjusted hazard ratios at Cox regression for fatal/non‐fatal cardiovascular disease were: diabetes duration 2.76 (2.21–3.44); onset age 1.47 (1.21–1.78); log ratio total cholesterol:HDL cholesterol 1.26 (1.09–1.45); log HbA 1c 1.19 (1.03–1.38); log systolic blood pressure 1.17 (1.01–1.34) (1 sd increase in continuous variables); smoker 1.76 (1.27–2.46); macroalbuminuria (> 200 μg/min) 1.52 (1.10–2.10); previous cardiovascular disease 3.51 (2.54–4.84). All eight variables were used to elaborate a risk equation for 5‐year cardiovascular disease risk. Regarding calibration in the derivation data set, ratio predicted 5‐year risk (mean 5.4 ± 7.9%) to observed event rate was 1.0. Discrimination was sufficient, with C‐statistic 0.83, sensitivity and specificity 72 and 77%, respectively, for the top quartile of predicted risk. Similarly, calibration and discrimination were adequate in the validation data set: ratio of predicted 4‐year risk/observed rate 0.94, C‐statistic 0.80, sensitivity and specificity 62 and 77%, respectively, for the top quartile. Conclusions This 5‐year cardiovascular disease risk model from a large observational study of patients with Type 1 diabetes in routine care showed adequate calibration and discrimination and can be useful for clinical practice. It should also be tested in patients with Type 1 diabetes from other countries.