Epidemiology of chronic kidney disease in KwaZulu-Natal: evaluation of risk factors, complications and diagnostic methods.

Abstract

Chronic kidney disease (CKD) is associated with increased morbidity and mortality as well as costly renal replacement therapy. The aim was to determine risk factors and complications that contribute to morbidity as well as a suitable diagnostic detection method for CKD. Methods Observational studies were done at 2 hospitals. To assess risk factors, 283 patients were included at the Durban site, and sub-studies undertaken within this sample for CKD complications, while 302 patients were studied at the Empangeni outreach site. To evaluate predictive performance of estimated glomerular filtration rate (eGFR), data from 148 patients were analysed. A further 76 patients were recruited, to develop an African equation. Cockcroft Gault, Modified Diet in Renal Disease and CKD Epidemiology Collaboration eGFR equations were compared with technicium-99-mdiethylenetriaminepentaacetic acid (99mTc-DTPA)-GFR as the gold standard. Body composition was assessed by anthropometry and dual energy X-ray absorptiometry. Data were analysed with STATA. Results The commonest CKD risk factors were hypertension (75%), diabetes (29%) and human immunodeficiency virus (HIV) infection (24%), with HIV commoner at the outreach site (28.5% vs 19.8%). Over 80% of females and ∼60% males were overweight/obese overall; however, clinical cardiovascular disease was commoner in Durban (28% vs 5%). Complications were observed in early CKD; prevalence increased as eGFR declined from ≥90 ml/min/1.73m2 to <30 ml/min/1.73m2: hyperuricaemia increased from 17% to 74%, metabolic acidosis (11.6% to 72.7%), anaemia (2.9% to 69.7%), hyperphosphataemia (10.1% to 48.5%), all p<0.001, respectively, and hypocalcaemia from 1.5% to 18.2% (p=0.003). Lower GFR levels were also associated with lower serum albumin levels, and lower whole body as well as regional lean mass and fat mass in males. A further observation at GFR <30 ml/min/1.73m2 was that eGFR underestimated 99mTc-DTPA-GFR in African patients. Prediction of 99mTc DTPA-GFR was also poor at GFR levels ≥60 ml/min/1.73m2, with eGFR overestimating 99mTc-DTPA-GFR. An eGFR equation developed in African patients resulted in significantly better GFR prediction and showed the lowest bias, highest precision as well as accuracy. Conclusion Efforts are needed to enable non-nephrologists to manage CKD risk factors and complications. Prediction of GFR may be substantially improved by using an equation developed in Africans.