OBJECTIVE: To study the ability of lipid variables to predict incident coronary artery disease (CAD) events in patients with type 1 diabetes at different stages of nephropathy. RESEARCH DESIGN AND METHODS: Patients (n = 3,520) with type 1 diabetes and available lipid profiles participating in the Finnish Diabetic Nephropathy Study (FinnDiane) were included in the study. During a follow-up period of 10.2 years (8.6-12.0), 310 patients suffered an incident CAD event. RESULTS: Apolipoprotein B (ApoB)/ApoA-I ratio was the strongest predictor of CAD in normoalbuminuric patients (hazard ratio 1.43 [95% CI 1.17-1.76] per one SD increase), and ApoB was the strongest in macroalbuminuric patients (1.47 [1.19-1.81]). Similar results were seen when patients were stratified by sex or glycemic control. LDL cholesterol was a poor predictor of CAD in women, normoalbuminuric patients, and patients with HbA 1c below the median (8.3%, 67 mmol/L). The current recommended triglyceride cutoff of 1.7 mmol/L failed to predict CAD in normoalbuminuric patients, whereas the cohort median 0.94 mmol/L predicted incident CAD events. CONCLUSIONS: In patients with type 1 diabetes, the predictive ability of the lipid variables differed substantially depending on the patient's sex, renal status, and glycemic control. In normoalbuminuric patients, the ratios of atherogenic and antiatherogenic lipoproteins and lipids were the strongest predictors of an incident CAD event, whereas in macroalbuminuric patients, no added benefit was gained from the ratios. Current treatment recommendations may need to be revised to capture residual CAD risk in patients with type 1 diabetes.