BACKGROUND: As life expectancy in sickle cell disease (SCD) has continued to increase, chronic diseases are playing a bigger role in morbidity and mortality in this population. SCD nephropathy is a known risk factor for early mortality, and risk of death increases with worsening glomerular filtration rate (GFR) (Elmariah et al. 2014; Platt et al. 1994). Glomerular ischemia, hyperfiltration, and hyposthenuria characterize the renal dysfunction found in SCD, but the natural history of and risk factors contributing to progression of disease remain unclear. Potential markers of SCD nephropathy, such as anemia severity, hemolytic rate, blood pressure (BP), and albuminuria, have been studied with varied conclusions (Aoki and Saad 1995; Asnani et al. 2016; Elmariah et al. 2014; Falk et al. 1992; Guasch et al. 2006; Gurkan et al. 2010). Our recent work also identified two variant alleles (“G1” and “G2”) of the APOL1 gene as risk factors for SCD nephropathy (Ashley-Koch et al. 2011). We sought to evaluate clinical, laboratory, and genetic predictors of SCD nephropathy progression in a longitudinal SCD cohort.
METHODS: We performed a chart review of 280 adults (age ≥18 at enrollment) with SCD followed regularly at the Duke University Medical Center and enrolled in a previously well-described study (Elmariah et al. 2014). All outpatient laboratory data were collected using the data extraction tool DEDUCE (Duke Enterprise Data Unified Content Explorer), and descriptive data were collected at time of enrollment and at most recent outpatient follow-up. APOL1 variants were genotyped as previously described (Ashley-Koch et al. 2011). Estimated GFR was calculated using both the extended MDRD and the CKD-EPI equations (Levey et al. 1999; Levey et al. 2009). CKD and hyperfiltration were defined as GFR <60 and >130 mL/min/1.73 m2, respectively. Statistical analysis was performed using SAS (SAS Systems, Cary, NC).
RESULTS: Mean age at enrollment was 33.3 years, and mean baseline GFR was 148.6 mL/min/1.73 m2 by MDRD and 127 mL/min/1.73 m2 by CKD-EPI (see Table 1 for baseline characteristics). Results from the two equations correlated well for GFRs in the low to normal range but diverged for GFRs over ~150 mL/min/1.73 m2, with MDRD producing a wider range (Fig. 1). Given the high prevalence of hyperfiltration in our cohort, we chose to perform further analyses using CKD-EPI. Baseline GFR was associated with baseline proteinuria (p<0.0001), hemoglobinuria (p<0.0001), Hb level (p=0.03), platelet count (p<0.0001), reticulocyte count (p=0.03), bilirubin (p<0.0001), systolic BP (p=0.0003), diastolic BP (p<0.0001), and the APOL1 G1 allele (p=0.02). To characterize the pattern of progression of SCD nephropathy, we then identified 183 subjects with at least 5 years of follow-up and ≥3 outpatient GFR measurements in order to calculate a rate of GFR decline. There were no statistically significant differences in baseline characteristics for the two cohorts (Table 1). The mean rate of GFR decline was 2.5 ± 5.6 mL/min/1.73 m2 per year, higher than the 1.27 mL/min/1.73 m2 estimated annual GFR decline in African Americans in the general population (Young et al. 2016).
We then defined a composite outcome of SCD nephropathy as either rapid rate of GFR decline (>3 mL/min/1.73 m2 per year, (Peralta et al. 2013; Young et al. 2016)), or incident CKD during the follow-up period. Seventy-three subjects (40%) developed nephropathy as thus defined. Using multivariable logistic regression modeling, with baseline GFR as a covariate, we found that lower baseline Hb (p=0.008), baseline proteinuria (p=0.01), higher reticulocyte count (p=0.03), and higher LDH (p=0.02) were all positive predictors of nephropathy. Baseline hemoglobinuria, white blood cell count, platelet count, bilirubin, BP, BMI, and APOL1 variants did not predict nephropathy. Subjects with hyperfiltration physiology at baseline also did not have a higher incidence of nephropathy over 5 years of follow up.
DISCUSSION: Our results highlight the rapid rate of GFR decline and the high prevalence of hyperfiltration in adults with SCD, while not showing a correlation between the two. Proteinuria and markers of a higher hemolytic rate, among other measures, were most predictive of a rapid GFR decline. Additional investigation into the longitudinal trajectory of SCD nephropathy and potential interventions targeted to slowing the rate of GFR decline may improve outcomes in SCD.
No relevant conflicts of interest to declare.