Evaluation of the Performance of Predictive Formulae in the Assessment of Glomerular Filteration Rate in Patients with Sickle Cell Disease

Abstract

Background: The magnitude of sickle cell nephropathy is high. Creatinine clearance is perhaps the best means to determine the severity of renal impairment in resource limited economies. With the various sources of errors inherent in the process of collecting  timed urine for clearance studies, it is imperative to find alternative means of assessing the status of renal function. Predictive formulae have not been studied in Nigerians with sickle cell disease (SCD). Objective: To evaluate five commonly used predictive formulae to estimate glomerular filtration rate (GFR) in patients with sickle cell disease in the out-patient setting. Also, to determine if any predictive formulae can be used in place of the measured value. Methods: Consecutive SCD patients, who attended the haematology outpatient clinic of the Obafemi Awolowo University Teaching Hospitals Complex (OAUTHC), were recruited into the study over a seven-month period. Those on medications that can interfere with renal function were excluded. Demographic details were collected and recorded. Blood and 24-hour urine samples were analyzed for serum and urinary creatinine, and creatinine clearance. The latter was compared with glomerular filtration rates estimated from five predictive formulae. Results: One hundred SCD patients (including 79 HbSS and 21 HbSC patients) were studied. The mean age was 26.2 ± 7.4 years and 54% were females. The highest agreement was between measured GFR and Cockcroft-Gault (CG) estimates (k = 0.50), followed by Mawer (k = 0.49), Hull (k = 0.26), Gates (k = 0.21) and MDRD (k = 0.02). Using the Bland-Altman technique, the Hull (mean = -15 ± 55) and MDRD (mean = -48 ± 61) formulae significantly underestimated GFR while the Mawer (mean = 8 ± 39) and CG (mean = -10 ± 39) formulae overestimated GFR. The Gates formula (mean = 0.6 ± 54) showed no difference with measured GFR. Conclusion: If we rely on the serum creatinine and the predictive formulae that are commonly in use, the status of renal function many patients with sickle cell disease will be inappropriately classified. However the CG formula may be used with the understanding of its limitation.