Progression of chronic kidney disease: an illness-death model approach

BACKGROUND Chronic kidney disease (CKD) is a major contributor to mortality in the general population. Understanding the factors that drive this process will help delay progression of CKD. The study aimed to estimate the risks of kidney failure and death prior to and after the development of kidney failure among patients with pre-existing CKD, and to identify potential prognostic factors. METHOD Data were obtained from patients with CKD from Ubon Ratchathani province, Thailand from 1997 to 2011. The probability of each transition (i.e., CKD➔death (T1), CKD➔kidney failure (T2), and kidney failure➔death (T3)) was estimated using a competing risk model. A parametric survival model with restricted cubic spline function was applied to assess prognostic factors. Illness-death models were constructed for the 3 transitions. Among 32,106 patients with CKD, 5576 (17.4%), 4768 (14.9%), and 3056 (9.5%) respectively moved through T1, T2, and T3. RESULTS Diabetics had 22.6%, 13.5%, and 60.7% higher risks of T1, T2, and T3 than non-diabetics respectively (p < 0.001). Hypertension increased risks of T2 and T3 by 8.7% (p = 0.01) and 27.2% (p < 0.001), whereas cardiovascular disease increased risk of T1 and T3 by 76% and 42.7%, respectively (p = 0.01). Increasing HDL by 10 units respectively decreased risk of T1 and T2 by 0.5% (p = 0.002) and 1.4% (p < 0.001). In addition, renin-angiotensin blockade decreased risk of T2 by 35% (p < 0.001). CONCLUSIONS Diabetes and cardiovascular disease are associated with increasing mortality among CKD patients both before and after the development of kidney failure while hypertension is associated with increasing mortality mainly following kidney failure. Diabetes and hypertension are associated with an elevated risk of kidney failure while elevated HDL levels and renin-angiotensin blockade appear protective.