Sickle Cell Trait, Estimated GFR, and Risk of Adverse Outcomes in CKD

To the Editor:

Recent evidence suggests that sickle cell trait (SCT), defined as the heterozygous inheritance of sickle hemoglobin, is associated with an increased risk of prevalent and incident chronic kidney disease (CKD).^1,2 However, there is limited evidence as to whether SCT is associated with an increased risk of other relevant outcomes, including cardiovascular disease (CVD), CKD progression, end-stage kidney disease (ESKD), and death among those with established CKD.¹ Determining risk for these outcomes is important since referrals for nephrology care, vascular access, and transplant are largely based on the probability of adverse outcomes. Another complicating aspect of the study of CKD and SCT is the concern that serum creatinine as a filtration marker may poorly reflect true GFR for patients with sickle cell anemia due to non-GFR determinants;³ however, the performance of creatinine in SCT is unknown. Among participants in the African American Study of Kidney Disease and Hypertension (AASK) study, a well-characterized cohort of individuals with hypertension-attributed CKD followed for a median of 10 years, we evaluated the accuracy and precision of GFR measured by ¹²⁵I-iothalamate (mGFR) compared to creatinine-based GFR estimates (eGFR) as well as whether SCT was associated with an increased risk of CVD, CKD progression, ESKD, and death.

The AASK study originated as a trial of 1,094 African Americans with hypertension-attributed CKD (baseline mGFR 20–65 ml/min/1.73 m²), randomized to intensive or standard blood pressure goals and three different antihypertensives; participants were subsequently invited to a 5-year cohort phase of the study.⁴ For the current study, individuals with hemoglobin C trait, hemoglobin SS, or hemoglobin SC genotype were excluded. Participants were genotyped for SCT using ABI TaqMan assays (Foster City, California).² Percent European ancestry was estimated using 140 ancestry informative markers and ANCESTRYMAP, imputing mean percent European ancestry for 6 participants with missing data.⁵ The Chronic Kidney Disease Epidemiology Collaboration (CKD-Epi) equation was used to calculate eGFR.⁶ To evaluate the relationship between mGFR estimates and eGFR in participants with and without SCT, bias (mGFR–eGFR), precision (interquartile range of bias), and accuracy (1 - P30, or the percentage of eGFR measures within 30% of the mGFR) were calculated, and 95% confidence intervals (CI) were estimated using bootstrapping. Kaplan-Meier curves and Cox proportional hazards models were used to calculate the risk of CVD (any revascularization procedure, cardiac death, myocardial infarction, stroke, and heart failure), CKD progression (doubling in serum creatinine or incident ESKD), ESKD alone, and death. Cox regressions were adjusted for age, sex, baseline GFR, and percent African ancestry. Statistical analyses were conducted using Stata Statistical Software: Release 15 (College Station, TX).

Among the 652 AASK participants with available genotyping, 65 (10%) were heterozygous for sickle hemoglobin. Baseline characteristics, including mean age and sex, were comparable between participants with and without SCT (Table 1). The APOL1 high-risk genotype was present in 22.4% of those without SCT and 23.1% of those with SCT. Baseline mean serum electrolytes, including sodium, potassium, and urate, as well as 24-hour urine sodium and potassium, were similar between participants with and without SCT. Among participants without SCT, mGFR (SD) was 47 (13) ml/min/1.73 m² and eGFR (SD) was 45 (13) ml/min/1.73 m², and among participants with SCT, mGFR (SD) was 45 (14) ml/min/1.73 m² and eGFR (SD) was 44 (13) ml/min/1.73 m². There was strong correlation between mGFR and eGFR in both individuals with and without SCT (r = 0.78 and r = 0.84, respectively) (Figure 1 A–B). The median bias (95% CI) was 3.1 (2.3, 3.9) ml/min/1.73 m² in participants without SCT and 0.3 (−1.3, 1.9) ml/min/1.73 m² in participants with SCT; precision (95% CI) was 10.0 (8.8, 11.1) and 7.9 (4.0, 11.9); and accuracy (95% CI) was 10.4 (8.0, 12.8) and 4.6 (−0.5, 9.8), respectively. With respect to the risk of adverse outcomes, there were 52 CVD events, 325 cases of CKD progression, 195 ESKD events, and 104 deaths. There were no differences by SCT status for risk of CVD, CKD progression, ESKD, or death in unadjusted analyses (Figure 1 C–F). In analyses adjusted for demographics and ancestry, the hazard ratio (95% CI) associated with SCT for CVD was 0.78 (0.28, 2.19), for CKD progression was 1.02 (0.71, 1.46), for ESKD was 0.97 (CI: 0.61, 1.55), and for death was 0.75 (0.37, 1.56).

Table 1.

Baseline Characteristics

Characteristic	Without SCT	With SCT	p-value
Demographics
N (% total sample)	587 (90.0)	65 (10.0)
Age (mean, SD)	54.2 (10.4)	53.8 (11.2)	0.820
Women, N (%)	233 (39.7)	29 (44.6)	0.442
APOL1 renal-risk variant, N (%)	130 (22.4)	15 (23.1)	0.187
Serum Electrolytes (mEq/L or mg/dL) (mean, SD)
Sodium	139.0 (2.5)	139.0 (2.5)	0.835
Potassium	4.2 (0.5)	4.2 (0.5)	0.602
Urate	8.2 (1.8)	8.2 (2.0)	0.823
Phosphate	3.5 (0.5)	3.4 (0.5)	0.443
Magnesium	2.1 (0.2)	2.0 (0.2)	0.876
Calcium	9.1 (0.4)	9.1 (0.4)	0.606
Urine Electrolytes (mEq/L/24 hrs) (mean, SD)
Sodium	3.7 (2.0)	3.7 (2.0)	0.988
Potassium	1.8 (0.9)	1.9 (0.9)	0.165
GFR markers (ml/min/1.73m2) (mean, SD)
Measured GFR (125I-iothalamate)	46.6 (13.0)	45.4 (12.7)	0.482
eGFR by creatinine	43.9 (14.1)	44.4 (13.5)	0.781

Note: All eGFR estimates were calculated using CKD-EPI

Abbreviations: GFR, glomerular filtration rate; eGFR, estimated glomerular filtration rate

Figure 1. Correlation between measured GFR and estimated GFR (ml/min/1.73m²) in those without sickle cell trait (A) versus those with sickle cell trait (B) and Time to events for the adverse outcomes of CKD progression (C), ESKD alone (D), CVD (E), and death (F).

Panels A-B depict scatter plots with lines of best fit. There was strong correlation between mGFR and eGFR in those without SCT (r = 0.78) and those with SCT (r= 0.84). In panels C-F, time to event for selected adverse outcomes by SCT status are depicted.

Note: log rank test p-values > 0.05 for all outcomes

Our study benefited from a well-characterized cohort with long-term follow-up for multiple CKD-related events as well as measurement of GFR using ¹²⁵I-iothalamate, the gold standard for GFR estimation. We found that creatinine-based estimates of GFR performed well compared to a direct measurement of GFR, suggesting that creatinine may be used effectively in estimating GFR among persons with SCT who already have established CKD. In addition, consistent with earlier studies, we found that the prevalence of SCT among participants in AASK was higher than published SCT prevalence rates in the general population, suggesting that there may be an increased risk of CKD development among people with SCT. On the other hand, there was no significant association between SCT and the outcomes of CVD, CKD progression, and death in this cohort of hypertension-attributed CKD, although these findings may be related to small sample size.