Screening for neuropathy is complex and underperformed in practice. If neuropathy rarely occurs first, screening for it could be delayed until retinopathy or nephropathy are documented. To address this, we aimed to definitively determine the sequence of microvascular complications in the natural history of type 1 diabetes (T1D) . Using data available from the public NIDDK Repository (1983-2012) , independent of the DCCT/EDIC research group, we performed a multistate analysis using biostatistics techniques for complex trivariate processes of the 1441 participants with T1D. Retinopathy (‘Eye’, E) , nephropathy (‘Kidney’, K) , and neuropathy (‘Nerve’, N) were each screened repeatedly and defined by the early-stage phenotypes (Figure 1) . Our model accounted for intermittent ascertainment and differing screening schedules. At baseline, participants had mean age 27±7 years and duration 6±4 years. Markov model simulations started in the absence of complications (Ø) . Eye was the most common initial complication (49%) , followed by nerve (28%) , and kidney (22%) . Median time to first complication was 5.3 years. 4% had no complications through follow-up. While retinopathy is conclusively most common, the high frequency of other initial complications reinforces the current practice of unselected early screening for all three, including neuropathy.
L.Lovblom: None. L.Briollais: None. G.Tomlinson: None. B.A.Perkins: Advisory Panel; Abbott Diabetes, Insulet Corporation, Sanofi, Board Member; Novo Nordisk, Other Relationship; Abbott Diabetes, Insulet Corporation, Medtronic, Novo Nordisk, Research Support; BMO Bank of Montreal, Novo Nordisk.
The Canadian Institutes of Health Research and the public NIDDK Central Repository.