Longitudinal Study of Glomerular Hyperfiltration and Normalization of Estimated Glomerular Filtration in Adults with Sickle Cell Disease

Abstract

Glomerular hyperfiltration is common in sickle cell disease (SCD) and precedes proteinuria and declining kidney function. We evaluated hyperfiltration in SCD patients and its “normalization.” Routine visit data were collected retrospectively from adult SCD patients in a single center from 2004-2013. Baseline was defined as first available serum creatinine and hyperfiltration as eGFR >130mL/min/1.73m² for women and >140mL/min/1.73m² for men. Normalization of hyperfiltration was eGFR reduction to 90-130mL/min/1.73m² for women or 90-140mL/min/1.73m² for men. Among 292 patients, median age was 27 (interquartile range [IQR]:20.0–38.0), and 56.8% had baseline hyperfiltration. Baseline hyperfiltration was inversely associated with age (OR:0.86, 95%CI: 0.82–0.90; p<0.0001), male sex (OR:0.16, 95%CI: 0.07–0.41; p=0.0001), hemoglobin (OR:0.76, 95% CI 0.61–0.94; p=0.01), weight (OR:0.96, 95%CI: 0.93–0.99; p=0.004), and ACE-I/ARB use (OR:0.08, 95%CI: 0.01–0.75; p=0.03), and positively with hydroxyurea use (OR:2.99, 95%CI: 1.18–7.56; p=0.02). Of 89 hyperfiltration patients without baseline proteinuria, 10 (11.2%) developed new-onset proteinuria (median 1.05 years [IQR:0.63–2.09]). Normalization of hyperfiltration was less likely with higher baseline eGFR (HR:0.90, 95%CI: 0.86–0.95; p<0.0001) and more likely in males (HR:6.35, 95%CI:2.71–14.86, <0.0001). Hyperfiltration is common in adult SCD patients, particularly when younger. Decline to normal values is more likely in males, possibly representing kidney function loss rather than improvement in hyperfiltration.

INTRODUCTION

Sickle cell disease (SCD) leads to structural and functional abnormalities in the kidney.(1) Much like in diabetes, SCD patients manifest increased glomerular filtration rates (GFR), even when accurately measured.(2) While no definition is consistently accepted, hyperfiltration is often defined using GFR thresholds between 130-140 mL/min per 1.73m².(3, 4) Hyperfiltration is noted in childhood with SCD and is thought to be a harbinger of kidney dysfunction.(1, 5) Several studies demonstrate that hyperfiltration precedes albuminuria.(5-7) Albuminuria, a clinical measure of glomerular injury and early manifestation of chronic kidney disease (CKD), may result from glomerular hypertension, increased oxidative stress, hemolysis and endothelial dysfunction.(1, 8, 9) As SCD patients age, estimated GFR (eGFR) appears to return to “normal” (eGFR 90-120ml/min per 1.73m²) possibly representing actual loss of kidney function. Although there has been much focus on albuminuria as an early clinical manifestation of CKD in SCD,(1, 10) available data on risk factors for hyperfiltration and decrease to normal eGFR remain limited.

In this retrospective study of an adult SCD patient cohort, we evaluated baseline clinical and laboratory variables associated with hyperfiltration, the association of hyperfiltration with proteinuria, and “normalization” of hyperfiltration over time.

MATERIALS AND METHODS

Study Design and Population

This study design, as previously described,(11) included adult SCD patients (HbSS, HbSC, HbSβ°, HbSβ⁺, etc) at a single academic medical center followed from 2004-2013 with longitudinal data collected via retrospective medical record review. Patients were evaluated during routine clinic visits when not experiencing acute complications. Baseline was defined as first available serum creatinine measurement during follow-up. Only patients with at least two eGFR values were included in these analyses. We excluded patients on dialysis, with kidney transplant or with baseline eGFR <90 mL/min per 1.73m² asserting that this group likely had CKD.(11) Estimated GFR was calculated using the creatinine-based Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation.(12) Hyperfiltration was defined as eGFR >130mL/min per 1.73m² for women and >140mL/min per 1.73m² for men. (6, 7) Normalization of hyperfiltration was defined as subsequent eGFR decline to 90–130mL/min per 1.73m² for women or 90–140mL/min per 1.73m² for men. Proteinuria was assessed by semi-quantitative dipstick urinalysis and defined as absent if 0-to-trace and present if ≥1+. Hemoglobinuria was defined as dipstick urinalysis with blood ≥1+ with fewer than 5 red blood cells per high power field on microscopy. The Institutional Review Board at University of North Carolina approved the study with waiver of consent for de-identified data analysis.

Statistical Analysis

Variables of interest were summarized by median and interquartile ranges (IQR) if continuous or by counts and percentages if categorical. We used logistic regression analyses to evaluate variables associated with baseline hyperfiltration. Each variable was assessed individually, adjusting for age and sex, and then in multivariable analysis. Variables included white blood cell count (WBC), hemoglobin, reticulocyte count, fetal hemoglobin (HbF), ferritin, lactate dehydrogenase, bilirubin (total, direct and indirect), hemoglobinuria, urine specific gravity, proteinuria, systolic blood pressure (SBP), diastolic blood pressure (DBP), weight, height, diabetes history, history of stroke, history of acute chest syndrome history, history of avascular necrosis, history of leg ulcers, angiotensin converting enzyme inhibitor/angiotensin receptor blocker (ACE-I/ARB) therapy, hydroxyurea therapy and chronic red blood cell (RBC) transfusion. In multivariable analysis, variables associated with the outcome of interest with p-values <0.3 in individual analyses without an excess of missing data (<20% missing) were included in initial multivariable modeling in addition to age and sex. Variables were selected for retention using backward elimination. Individual and multivariable analyses were adjusted for baseline age and sex. Individual variables were also assessed for interaction with SCD severity (based on genotype), adjusting for age and sex with variables considered of potential statistical significance if p for interaction was ≤0.20. Analyses were then stratified by severe (HbSS/HbSβ° thalassemia) and mild (HBSC/HbSβ⁺ thalassemia) SCD genotypes.

For time from baseline to new onset proteinuria or normalization of hyperfiltration analyses, Kaplan-Meier estimates were used for estimation of survival (or event-free) probabilities. Median time to normalization was calculated as the 50^th percentile of survival probability based on Kaplan-Meier estimates. We evaluated baseline variables by Cox regression analysis for time-to-normalization from baseline, where normalization was defined as the first time after baseline that eGFR reduced to 90–130mL/min per 1.73m² for women or 90–140mL/min per 1.73m² for men. Only patients with baseline hyperfiltration were included in this analysis. We first examined each variable noted above as well as baseline eGFR individually, adjusting for age and sex. Multivariable analysis utilized backward elimination for variable selection. Initial multivariable modeling included variables with p-values <0.3 in individual analyses and without severe missing data (>20% missing). Variables were selected for retention using backward elimination. Interaction of variables with genotype severity were performed as above, and analyses were then stratified by severe and mild SCD genotypes. For continuous variables, hazard ratios (HRs) were calculated for each unit increase in that exposure. All analyses were conducted using SAS OnDemand for Academics, Copyright© [2020] SAS Institute Inc., Cary, NC, USA.

RESULTS

Cohort Characteristics

We included 292 SCD patients (HbSS=192 [65.8%], HbSC=57 [19.5%], HbSβ° thalassemia=18 [6.2%], HbSβ⁺ thalassemia=20 [6.9%], HbSE=1 [0.3%], HbSD=2 [0.7%], SHPFH=2 [0.7%]). Baseline cohort characteristics are shown in Table 1. Individuals were followed for median of 4.02 years (IQR: 1.69–7.19 years) with median of 6 (IQR: 3–11) available eGFR values. Prevalence of hyperfiltration was 56.8% and decreased with increasing age, ranging from 83.9% in patients between 18-20 years to 0% in patients >50 years (Figure 1). Baseline hyperfiltration was observed in 138 of 210 patients (65.7%) with HbSS/HbSβ° thalassemia compared with 25 of 77 patients (32.5%) with HbSC/HbSβ⁺ thalassemia.

Table 1:

Baseline Demographics, Clinical and Laboratory Variables of Cohort

	All Sickle Cell Disease		HbSS/HbSβ° Thalassemia		HbSC/HbSβ+ Thalassemia
Variables	Number	Median (IQR) or Number (%)	Number	Median (IQR) or Number (%)	Number	Median (IQR) or Number (%)
Age (years)	292	27.0 (20.0 – 38.0)	210	27.0 (20.0 – 36.0)	77	28.0 (19.0 – 41.0)
Sex (female)	292	160 (54.8)	210	115 (54.8)	77	43 (55.8)
Weight (Kg)	289	68.1 (59.0 – 80.2)	207	65.1 (57.0 – 75.5)	77	78.9 (66.8 – 95.3)
Height (cm)	120	170.4 (162.5 – 175.9)	85	170.3 (163.0 – 176.0)	32	169.5 (161.5- 173.5)
Systolic Blood Pressure (mm Hg)	292	120 (110 – 132)	210	118 (109. – 130)	77	127 (118 – 135)
Diastolic Blood Pressure (mm Hg)	291	71 (63 – 79)	210	69 (61 – 75)	76	78 (70 – 83)
White Blood Cell Count (109/L)	292	10.4 (7.9 – 12.8)	210	11.0 (8.5 – 13.0)	77	8.6 (6.6 – 10.5)
Hemoglobin (g/dL)	292	9.7 (8.4 – 11.1)	210	9.1 (8.0 – 10.2)	77	12.1 (10.8 – 13.0)
Reticulocyte Count (109/L)	272	202.5 (122.4 – 284.8)	198	224.7 (159.7 – 308.2)	70	118.1 (86.9 – 203.3)
Lactate Dehydrogenase (U/L)	236	841 (622.5 – 1138)	168	924.5 (745 – 1244)	64	597.5 (506.5 – 799.5)
Total Bilirubin (mg/dL)	265	1.9 (1.2 – 3.4)	192	2.4 (1.5 – 4.3)	68	1.0 (0.8 – 1.5)
Indirect Bilirubin (mg/dL)	63	2.0 (1.2 – 3.3)	47	2.3 (1.6 – 3.9)	15	1.1 (0.8 – 1.4)
Direct Bilirubin (mg/dL)	63	0.09 (0.09 – 0.10)	47	0.09 (0.09 – 0.10)	15	0.09 (0.09 – 0.20)
Ferritin (μg/L)	124	226.5 (67.5 – 738.0)	87	294 (122 – 1090)	34	115.5 (42.0 – 322.0)
Fetal Hemoglobin (%)	127	5.1 (2.0– 9.4)	101	6.2 (3.4 – 10.4)	24	1.2 (0.4 – 2.9)
Serum Creatinine (mg/dL)	292	0.70 (0.60 – 0.80)	210	0.68 (0.54 – 0.80)	77	0.80 (0.70 – 0.91)
Estimated Glomerular Filtration Rate (mL/min/1.73m2)	292	138.9 (121.7 – 153.1)	210	144.5 (128.7 – 158.1)	77	125.9 (112.7 – 138.8)
Blood Urea Nitrogen (mg/dL)	273	8.0 (6.0 – 10.0)	201	8.0 (6.0 – 9.0)	67	8.0 (7.0 – 10.0)
Urine Specific Gravity	127	1.01 (1.01 – 1.01)	110	1.01 (1.01 – 1.01)	41	1.01 (1.01 – 1.01)
Proteinuria (yes)	163	23 (14.1)	118	20 (17.0)	42	3 (7.1)
Hemoglobinuria (yes)	152	12 (7.9)	108	12 (11.1)	41	0 (0)
History of Acute Chest Syndrome	280	218 (81.4)	203	178 (87.7)	72	48 (66.7)
History of stroke	266	33 (12.4)	191	29 (15.2)	70	4 (5.7)
History of Leg Ulcers	240	30 (12.5)	176	29 (16.5)	60	1 (1.7)
History of Avascular Necrosis	203	84 (41.4)	143	61 (42.7)	56	22 (39.3)
History of Priapism	113	44 (38.9)	83	36 (43.4)	29	8 (27.6)
History of Diabetes	292	12 (4.1)	210	6 (2.9)	77	6 (7.8)
Chronic RBC Transfusion	292	10 (3.4)	210	9 (4.3)	77	1 (1.3)
Hydroxyurea Therapy	290	98 (33.8)	208	86 (41.4)	77	12 (15.6)
ACE-I/ARB Therapy	290	20 (6.9)	209	12 (5.7)	76	8 (10.5)

IQR = interquartile range

ACE-I = angiotensin-converting enzyme inhibitor; ARB = angiotensin receptor blocker

Figure 1.

Occurrence of glomerular hyperfiltration in sickle cell disease by age category.

Of 163 patients with baseline dipstick urinalysis, proteinuria was present in 23 (14.1%), with a prevalence of 12.8% in 102 patients with hyperfiltration compared with 16.4% in 61 patients without hyperfiltration. Of 89 patients with baseline hyperfiltration without proteinuria, 10 (11.2%) patients developed new onset proteinuria over a median of 1.05 years (IQR: 0.63–2.09 years). The probability of developing proteinuria in this group was 30% during follow-up. Of 51 patients with neither baseline hyperfiltration nor proteinuria, 4 (7.8%) developed new onset proteinuria over median of 2.28 years (IQR: 1.06–3.61 years). The probability of developing proteinuria was <20% during follow-up.

Baseline Hyperfiltration

In the overall sample, odds of hyperfiltration were lower with older age (odds ratio [OR]: 0.87, 95% confidence interval [CI] 0.85–0.90; p<0.0001) and male sex (OR: 0.43, 95% CI 0.26–0.68; p=0.0004) (Table 2). In age- and sex-adjusted analyses, baseline hyperfiltration was significantly less likely with higher baseline hemoglobin (OR: 0.73, 95% CI: 0.62–0.86; p=0.0001), higher weight (OR: 0.95, 95% CI: 0.93–0.97; p<0.0001), higher SBP (OR: 0.80, 95% CI: 0.65–0.98; p=0.03), higher DBP (OR: 0.77, 95% CI: 0.61–0.97; p=0.02), and use of ACE-I/ARBs (OR: 0.21, 95% CI: 0.05–0.92; p=0.04) (Table 2). Hyperfiltration positively associated with higher reticulocyte count (OR: 1.05, 95% CI: 1.02–1.08; p=0.002), total bilirubin (OR: 1.36, 95% CI: 1.13–1.63; p=0.0009), lactate dehydrogenase (OR: 1.01, 95% CI: 1.00–1.02, p=0.03), history of acute chest syndrome (OR: 2.52, 95% CI: 1.11–5.71; p=0.03) and hydroxyurea therapy (OR: 1.95, 95% CI: 1.02-3.73; p=0.04). In analyses for interaction, age (p=0.20), male sex (p=0.16), WBC count (p=0.028), reticulocyte count (p=0.15), hemoglobin F (p=0.096), ferritin (p=0.0054), LDH (p=0.0040), and history of avascular necrosis (p=0.13) were noted to have potential interactions with genotype severity.

Table 2:

Association of Baseline Laboratory and Clinical Variables with Hyperfiltration

	All Sickle Cell Disease		Test for Interaction†	HbSS/HbSβ°-Thalassemia		HbSC/HbSβ+-Thalassemia
Variable	OR (95% CI)	p	p	OR (95% CI)	p	OR (95% CI)	p
Age	0.87 (0.85 - 0.90)	<0.0001	0.20	0.85 (0.81 - 0.89)	<0.0001	0.89 (0.84 - 0.95)	0.0003
Male sex	0.43 (0.26 - 0.68)	0.0004	0.16	0.49 (0.27 - 0.87)	0.01	0.20 (0.07 - 0.61)	0.005
White blood cell count	1.08 (0.99 - 1.17)	0.07	0.028	0.95 (0.84 - 1.06)	0.3	1.17 (0.97 - 1.40)	0.1
Hemoglobin	0.73 (0.62- 0.86)	0.0001	0.29	1.03 (0.80 - 1.32)	0.8	0.80 (0.50 - 1.29)	0.4
Reticulocyte count*	1.05 (1.02 - 1.08)	0.002	0.15	1.00 (0.97 - 1.04)	0.9	1.07 (0.96 - 1.18)	0.2
Hemoglobin F	1.03 (0.96 -1.10)	0.5	0.096	0.99 (0.91 - 1.08)	0.9	0.64 (0.35 - 1.18)	0.1
Ferritin*	1.00 (1.00 - 1.01)	0.3	0.0054	1.00 (1.00 - 1.00)	0.7	1.08 (1.00 - 1.17)	0.04
Lactate dehydrogenase*	1.01 (1.00 - 1.02)	0.03	0.0040	1.00 (0.99 - 1.01)	0.5	1.04 (1.01 - 1.07)	0.02
Total bilirubin	1.36 (1.13 - 1.63)	0.0009	0.81	1.16 (0.94 - 1.43)	0.2	1.10 (0.48 - 2.56)	0.8
Direct bilirubin	0.51 (0.01 - 29.64)	0.7	0.88	2.19 (0.00 - 1000.00)	0.9	N/A
Indirect bilirubin	1.78 (0.98 - 3.25)	0.06	0.51	1.61 (0.67 - 3.87)	0.3	0.42 (0.02 - 7.59)	0.6
Hemoglobinuria	0.34 (0.10 - 1.63)	0.2	***	0.18 (0.04 - 0.83)	0.03	***
Proteinuria	0.76 (0.25 - 2.34)	0.6	0.41	0.41 (0.12 - 1.44)	0.2	1.71 (0.04 - 68.46)	0.8
Systolic blood pressure*	0.80 (0.65 – 0.98)	0.03	0.39	0.85 (0.66 - 1.10)	0.2	1.02 (0.67 - 1.56)	0.9
Diastolic blood pressure*	0.77 (0.61 - 0.97)	0.02	0.26	0.78 (0.60 - 1.02)	0.07	1.08 (0.67 - 1.73)	0.7
Weight	0.95 (0.93 - 0.97)	<0.0001	0.31	0.96 (0.93 - 0.99)	0.004	0.98 (0.95 - 1.01	0.2
Height	1.03 (0.97 - 1.11)	0.3	0.38	1.07 (0.96 - 1.19)	0.2	0.96 (0.84 - 1.09)	0.5
History of diabetes	0.83 (0.13 - 5.13)	0.8	0.97	2.33 (0.25 - 21.43)	0.5	N/A
History of stroke	0.83 (0.32 - 2.16)	0.7	0.39	0.46 (0.14 - 1.49)	0.2	1.42 (0.15 - 13.42)	0.8
History of acute chest syndrome	2.52 (1.11 - 5.71)	0.03	0.38	0.91 (0.23 - 3.59)	0.9	1.72 (0.46 - 6.50)	0.4
History of avascular necrosis	0.94 (0.43 - 2.05)	0.9	0.13	0.40 (0.14 - 1.17)	0.1	1.85 (0.33 - 10.45)	0.5
History of leg ulcers	0.87 (0.30 - 2.53)	0.8	0.99	0.53 (0.16 - 1.80)	0.3	N/A
ACEi/ARB therapy	0.21 (0.05 - 0.92)	0.04	0.91	0.19 (0.03 - 1.13)	0.07	0.26 (0.02 - 3.72)	0.3
Hydroxyurea therapy	1.95 (1.02 - 3.73)	0.04	0.69	1.33 (0.60 - 2.98)	0.5	1.78 (0.40 - 8.02)	0.5
Chronic RBC transfusion	0.53 (0.11 - 2.51)	0.4	0.98	0.29 (0.05 - 1.80)	0.2	N/A

ACE-I = angiotensin-converting enzyme inhibitor; ARB = angiotensin receptor blocker; RBC = red blood cell

For continuous variables Odds Ratios (OR) are for a one unit change.

All variables (except age and sex) adjusted for age and sex.

^*OR presented per 10 unit change

^***model did not converge given low number of events

^†Testing for interaction with genotype severity was performed for each variable of interest adjusting for baseline age and sex.

In HbSS/HbSβ° thalassemia patients, as in the overall cohort, older individuals (OR: 0.85, 95% CI: 0.81–0.89; p<0.0001) and male individuals (OR: 0.49, 95% CI: 0.27–0.87; p=0.01) were less likely to have baseline hyperfiltration. Hyperfiltration was significantly less likely with hemoglobinuria (OR: 0.18, 95% CI: 0.04–0.83; p=0.03) and higher weight (OR: 0.96, 95% CI: 0.93–0.99; p=0.004). In HbSC/HbSβ⁺ thalassemia patients, relationships between hyperfiltration and both age (OR: 0.89, 95% CI: 0.84–0.95; p=0.0003) and sex (OR: 0.20, 95% CI: 0.07–0.61; p=0.005) were consistent with the overall cohort. Hyperfiltration was significantly more likely with higher ferritin (OR: 1.08, 95% CI: 1.00–1.17; p=0.04) and higher lactate dehydrogenase (OR: 1.04, 95% CI: 1.01–1.07; p=0.02).

Normalization of Hyperfiltration

Of 166 patients (56.8%) with baseline hyperfiltration, 41 (24.7%) had eGFR declines to normal with median time of 8.81 years (Figure 2). In HbSS/HbSβ° thalassemia patients, 35 of 138 patients (25.4%) with baseline hyperfiltration had eGFR declines to normal with median time of 8.21 years. In HbSC/HbSβ⁺ thalassemia patients, 6 of 25 patients (24.0%) had eGFR declines to normal and the probability for normalization was below 30% during follow-up. Decline in eGFR to normal was significantly more likely in older SCD patients with any genotype (hazard ratio [HR]: 1.11, 95% CI: 1.07–1.16; p<0.0001) and in severe genotypes (HR: 1.14, 95% CI: 1.10–1.19; p<0.0001) (Table 3). In age- and sex-adjusted analysis of all SCD patients, eGFR decline to normal was significantly less likely with higher baseline eGFR (HR: 0.90, 95% CI: 0.86–0.94; p<0.0001), and more likely with higher weight (HR: 1.02, 95% CI: 1.00–1.04; p=0.04) and higher hemoglobin F (HR: 1.14, 95% CI: 1.02–1.28; p=0.02). In analyses for interaction, age (p=0.033), male sex (p=0.20), ferritin (p=0.0093), BUN (0.099), indirect bilirubin (p=0.024), SBP (p=0.10) and DBP (p=0.078), weight (p=0.14), and hydroxyurea use (p=0.10) were noted to have potential interaction with genotype severity.

Figure 2.

Time to normalization of estimated glomerular filtration rate among all sickle cell disease patients in the cohort.

Table 3:

Factors Associated with Time to Decline in Estimated GFR to Normal Range in Sickle Cell Patients with Hyperfiltration

	All Sickle Cell Disease		Test for Interaction†	HbSS/HbSβ°-Thalassemia		HbSC/HbSβ+- Thalassemia
Variable	HR (95% CI)	p	p	HR (95% CI)	p	HR (95% CI)	p
Age	1.11 (1.07 - 1.16)	<0.0001	0.033	1.14 (1.10 - 1.19)	<0.0001	1.01 (0.91 - 1.13)	0.8
Male sex	1.29 (0.69 - 2.42)	0.4	0.20	1.18 (0.60 - 2.33)	0.6	3.77 (0.68 - 20.81)	0.1
White blood cell count	0.96 (0.87 - 1.06)	0.4	0.95	0.97 (0.86 - 1.09)	0.6	1.02 (0.83 - 1.26)	0.8
Hemoglobin	0.95 (0.78 - 1.17)	0.6	0.78	0.92 (0.69 - 1.23)	0.6	0.89 (0.48 - 1.66)	0.7
Reticulocyte count	1.00 (1.00 - 1.00)	0.4	0.70	1.00 (1.00 - 1.00)	0.4	1.00 (0.99 - 1.01)	0.9
Hemoglobin F	1.14 (1.02 - 1.28)	0.02	0.74	1.18 (1.03 - 1.35)	0.02	***
Ferritin	1.00 (1.00 - 1.00)	0.1	0.0093	1.00 (1.00 - 1.00)	0.2	***
Baseline eGFR	0.90 (0.86 - 0.94)	<0.0001	0.34	0.90 (0.86 - 0.95)	<0.0001	0.79 (0.64 – 0.97)	0.03
Blood urea nitrogen	0.96 (0.83 - 1.12)	0.6	0.099	1.05 (0.88 - 1.25)	0.6	0.68 (0.42 - 1.09)	0.1
Lactate dehydrogenase	1.00 (1.00 - 1.00)	0.7	0.85	1.00 (1.00 - 1.00)	0.9	1.00 (1.00 - 1.01)	0.5
Total bilirubin	0.97 (0.82 - 1.16)	0.7	0.22	1.00 (0.84 - 1.20)	1.0	0.10 (0.00 - 3.62)	0.2
Direct bilirubin	2.79 (0.00 – 21429.29)	0.8	***	9.80 (0.003 – 28473.05)	0.6	***
Indirect bilirubin	0.46 (0.17 - 1.23)	0.1	0.024	0.61 (0.21 - 1.79)	0.4	***
Hemoglobinuria	3.26 (0.91 – 11.72)	0.07	***	5.47 (1.36 – 22.04)	0.02	***
Proteinuria	1.49 (0.47 - 4.70)	0.5	0.99	1.86 (0.54 – 6.53)	0.3	***
Systolic blood pressure	1.02 (1.00 - 1.04)	0.1	0.10	1.02 (0.99 - 1.04)	0.2	1.10 (0.97 - 1.24)	0.1
Diastolic blood pressure	1.01 (0.99 - 1.03)	0.3	0.078	1.01 (0.98 - 1.03)	0.7	1.10 (1.00 - 1.21)	0.06
Weight	1.02 (1.00 - 1.04)	0.04	0.14	1.01 (0.99 - 1.04)	0.3	1.04 (0.97 - 1.11)	0.2
Height	1.03 (0.95 - 1.12)	0.5	0.25	1.04 (0.94 - 1.15)	0.5	1.06 (0.86 - 1.30)	0.6
History of diabetes	2.32 (0.54 – 9.99)	0.3	***	2.13 (0.48 – 9.35)	0.3	***
History of stroke	0.56 (0.19 - 1.65)	0.3	0.99	0.49 (0.15 – 1.56)	0.2	***
History of acute chest syndrome	1.31 (0.51 – 3.37)	0.6	0.23	2.09 (0.60 – 7.32)	0.2	***
History of avascular necrosis	0.89 (0.40 – 1.96)	0.8	0.99	1.09 (0.44 – 2.73)	0.8	***
History of leg ulcers	1.22 (0.42 – 3.55)	0.7	***	1.17 (0.40 – 3.44)	0.8	***
ACEi/ARB therapy	1.58 (0.21 – 11.82)	0.7	0.99	2.25 (0.29 – 17.32)	0.4	***
Hydroxyurea therapy	1.46 (0.76 - 2.82)	0.3	0.10	1.00 (0.48 – 2.11)	1.0	3.18 (0.60 – 16.88)	0.2
Chronic RBC transfusion	1.22 (0.16 – 9.10)	0.8	***	1.63 (0.21 - 12.44)	0.6	***

ACE-I = angiotensin-converting enzyme inhibitor; ARB = angiotensin receptor blocker; RBC = red blood cell

For hemoglobin F, ferritin, direct and indirect bilirubin, all observations have the same outcome (no normalization) and thus the analysis could not be performed. For H/O leg ulcers and chronic RBC transfusion, their values were all equal to 0 and thus no results were produced.

For continuous variables Hazard Ratios (HR) are for a one unit change.

All variables (except age and sex) adjusted for age and sex.

^***model did not converge given low number of events

^†Testing for interaction with genotype severity was performed for each variable of interest adjusting for baseline age and sex.

In patients with severe genotypes, eGFR decline to normal was significantly less likely with higher baseline eGFR (HR: 0.90, 95% CI: 0.86–0.95; p<0.0001) and more likely with higher HbF (HR: 1.18, 95% CI: 1.03–1.35; p=0.02). In this group, hemoglobinuria was also associated with eGFR normalization (HR: 5.47, 95% CI: 1.36–22.04; p=0.02). In mild genotypes, higher baseline eGFR was also associated with lower hazards of normalization (HR: 0.79, 95% CI: 0.64–0.97; p=0.03).

Multivariable Analyses

Multivariable analysis was conducted to identify laboratory and clinical variables associated with baseline hyperfiltration while accounting for potential confounders. In all patients, hyperfiltration was significantly less likely with older age (OR: 0.86, 95% CI: 0.82–0.90; p<0.0001), male sex (OR: 0.16, 95% CI: 0.07–0.41; p=0.0001), higher weight (OR: 0.96, 95% CI: 0.93–0.99; p=0.004), higher hemoglobin (OR: 0.76, 95% CI: 0.61–0.94; p=0.01) and ACE-I/ARB use (OR: 0.08, 95% CI: 0.01–0.75; p=0.03), but significantly more likely with hydroxyurea use (OR: 2.99, 95% CI: 1.18–7.56; p=0.02) (Table 4). In severe genotypes, hyperfiltration was significantly less likely with older age (OR: 0.81, 95% CI 0.75–0.87; p<0.0001), male sex (OR: 0.13, 95% CI: 0.04–0.50; p=0.003), higher lactate dehydrogenase (OR: 0.998, 95% CI: 0.997–1.000; p=0.03), higher weight (OR: 0.94, 95% CI: 0.89–0.99; p=0.01) and ACE-I/ARB use (OR: 0.07, 95% CI: 0.01–0.82; p=0.03). In mild genotypes, no variables beyond age (OR: 0.93, 95% CI: 0.87–0.99; p=0.03) and male sex (OR: 0.08, 95% CI: 0.02–0.46; p=0.004) were associated with baseline hyperfiltration.

Table 4:

Multivariable Model for Hyperfiltration at Baseline in Adults with Sickle Cell Disease

	All Sickle Cell Disease		HbSS/HbSβ° Thalassemia		HbSC/HbSβ+ Thalassemia
Variable	Odds Ratio (95% CI)	p	Odds Ratio (95% CI)	p	Odds Ratio (95% CI)	p
Age	0.86 (0.82 – 0.90)	< 0.0001	0.81 (0.75 – 0.87)	<0.0001	0.93 (0.87 – 0.99)	0.03
Male Sex	0.16 (0.07 – 0.41)	0.0001	0.13 (0.04 – 0.50)	0.003	0.08 (0.02 – 0.46)	0.004
Hemoglobin	0.76 (0.61 – 0.94)	0.01
Weight	0.96 (0.93 – 0.99)	0.004	0.94 (0.89 – 0.99)	0.01
Hydroxyurea Therapy	2.99 (1.18 – 7.56)	0.02
Use of ACE-I/ARBs	0.08 (0.01 – 0.75)	0.03	0.07 (0.01 – 0.82)	0.03
Lactate Dehydrogenase			0.998 (0.997 – 1.000)	0.03

ACE-I = angiotensin-converting enzyme inhibitor; ARB = angiotensin receptor blocker

For continuous variables Odds Ratios (OR) are for a one unit change.

Results from multivariable logistic regression with backward elimination for variable selection. Variables initially assessed for inclusion in the model: white blood cell count, hemoglobin, reticulocyte count, lactate dehydrogenase, total bilirubin, hydroxyurea therapy, weight, systolic blood pressure, diastolic blood pressure, acute chest syndrome, and ACE inhibitor/ARB therapy. The model was adjusted for age and sex.

In multivariable analysis of the association of normalization of hyperfiltration with baseline laboratory and clinical variables in all SCD patients, normalization was significantly less likely with higher baseline eGFR (HR: 0.90, 95% CI: 0.86–0.95; p<0.0001) and more likely with male sex (HR: 6.35, 95% CI: 2.71–14.86; p<0.0001) (Table 5). In severe genotypes, normalization was significantly less likely with higher baseline eGFR (HR: 0.90, 95% CI: 0.86–0.95; p<0.0001) and more likely with male sex (HR: 6.91, 95% CI: 2.64–18.12; p<0.0001) and older age (HR: 1.12, 95% CI: 1.05–1.20; p=0.0009). In mild genotypes, normalization of hyperfiltration was also less likely with higher baseline eGFR (HR: 0.78, 95% CI: 0.64–0.98; p=0.03) and more likely with male sex (HR: 15.90, 95% CI: 1.30–194.03; p=0.03).

Table 5:

Multivariable Model for Time to Normalization of Hyperfiltration in Adults with Sickle Cell Disease

	All Sickle Cell Disease		HbSS/HbSβ° Thalassemia		HbSC/HbSβ+ Thalassemia
Variable	Hazard Ratio (95% CI)	p	Hazard Ratio (95% CI)	p	Hazard Ratio (95% CI)	p
Age	1.05 (0.99 – 1.11)	0.10	1.12 (1.05 – 1.19)	0.0009	0.90 (0.79 – 1.03)	0.13
Male sex	6.35 (2.71 – 14.86)	<0.0001	6.91 (2.64 – 18.12)	<0.0001	15.90 (1.30 – 194.03)	0.03
Baseline eGFR	0.90 (0.86 – 0.95)	<0.0001	0.90 (0.86 – 0.95)	<0.0001	0.78 (0.64 – 0.98)	0.03

For continuous variables Hazard Ratios (HR) are for a one unit change.

eGFR = estimated glomerular filtration rate.

Results from multivariable Cox regression with backward elimination for variable selection. Variables initially included for assessment in model were baseline eGFR, weight, history of stoke, diabetes, systolic blood pressure, and hydroxyurea therapy. The model was adjusted for age and sex.

DISCUSSION

Repetitive vascular occlusion is thought to cause ischemic injury and microinfarction in the renal medulla, ultimately producing reduced medullary blood flow.(1) Resulting medullary hypoxia produces localized prostaglandin release, vasodilation, increased renal blood flow and subsequent GFR increase. Hyperfiltration is common in SCD, present in most children with sickle cell anemia, and precedes albuminuria development.(5, 7, 13, 14) Hyperfiltration may begin in early childhood,(15, 16) with eGFR plateauing later in childhood and then decreasing in adulthood.(5, 7, 16-20) We confirm high prevalence of hyperfiltration in young adults as well as its inverse association with age. Hyperfiltration occured more commonly in severe SCD genotypes. Although the eGFR of severe genotype patients appears to take longer to decline to normal values than those with milder genotypes, only about 25% patients had eGFR declines to normal ranges over the follow-up period.

Proteinuria was more prevalent in patients without hyperfiltration than in patients with hyperfiltration. Among patients without baseline proteinuria, those with hyperfiltration were morely likely to develop proteinuria during follow-up and in less time. While not statistically significant, odds of proteinuria were decreased in severe genotype patients with baseline hyperfiltration in univariate analysis. Proteinuria was not significantly associated with hyperfiltration in multivariable analysis. Published results for the association of elevated eGFR and albuminuria are conflicting.(16, 17, 21-26) A recent longitudinal study of sickle cell anemia patients with cystatin C measured in childhood showed persistent albuminuria occurred earlier and more frequently in those with early hyperfiltration.(5) Conceivably, patients in our adult cohort with proteinuria, but without hyperfiltration, had elevated eGFR in childhood that preceded glomerular damage and eGFR decline.

Male sex was associated with lower odds of baseline hyperfiltration, consistent in both severe and mild genotypes. In contradistinction to prior data suggesting male SCD patients are more prone to kidney injury,(27, 28) our finding could reflect differences in eGFR trajectory in males compared to females. In a sickle cell mouse model, hyperfiltration was noted earlier in males, and subsequent eGFR loss occurred more slowly in females.(29) Female mice demonstrated differences in ET-1 expression and glomerular permeability that may partly explain these findings.(30) In pediatric SCD patients, the same investigators found older age was associated with lower eGFR in males but higher eGFR in females.(31) Observational studies in adults with sickle cell disease have demonstrated steeper eGFR loss over time in males.(32-34) As our population was restricted to adults at cohort entry, eGFR in males may have already been declining from uncaptured historical hyperfiltration. If so, one would expect baseline hyperfiltration to be more likely in females in whom eGFR was actually preserved. Among those with baseline hyperfiltration, male sex was associated with greater odds of normalization in all patients and in severe genotype patients, consistent with the suggestiion males are more likely to have rapid kidney function decline.

Patients with higher weight had lower odds of hyperfiltration, primarily in severe genotypes. Body mass index (BMI) data was only available for 120 of 292 patients; median BMI at baseline was 23 (IQR: 20.5–27.2) (data not shown). Obesity may lead to compensatory hyperfiltration to meet metabolic demands of increased body mass, possibly by hemodynamic changes at the afferent arteriole, and is a known risk factor for albuminuria, CKD and end-stage kidney disease (ESKD).(35-37) Substantial weight loss is associated with reduction in measured GFR, perhaps due to resolution of hyperfiltration.(38-40) The reason for lower likelihood of hyperfiltration with higher body weight in our cohort is uncertain but notably the range of BMI in our cohort did not approach the range of obesity. The association between BMI and various measures of CKD progression, except for ESKD, is U-shaped,(41) suggesting low body weight, perhaps reflecting poor overall health, is also a risk factor for progressive kidney disease.

Higher hemoglobin level was significantly associated with lower hyperfiltration odds in the overall cohort. Although some suggest improvements in hemolytic anemia may normalize hyperfiltration in SCD,(6) a large clinical trial in infants showed no significant GFR decline following hydroxyurea therapy despite improvements in anemia.(15) In multivariable analyses, increased lactate dehydrogenase level was associated with lower odds of baseline hyperfiltration in severe genotypes. This is somewhat inconsistent with a previous report that hemolysis contributes to pathophysiology of hyperfiltration.(6) The lower likelihood of hyperfiltration in those with increased hemolysis may reflect the phenomenon described above with historical hyperfiltration and subsequent loss of kidney function to the “normal” range.

Hydroxyurea therapy was associated with greater odds of hyperfiltration. This finding was surprising as hydroxyurea decreases hemolysis in SCD.(15, 42) Previous controlled trials did not show an influence of hydroxyurea on GFR in young children with sickle cell anemia.(15, 43) Hydroxyurea increases nitric oxide bioavailability,(44-46) which could increase renal blood flow and hyperfiltration. However, in SCD mouse models, greater nitric oxide bioavailability ameliorated kidney disease. While overall hydroxyurea use in this cohort was relatively low, our finding may also reflect confounding by indication, as patients with more severe disease and greater likelihood of hyperfiltration are more likely to receive hydroxyurea. Hydroxyurea was also associated with higher likelihood of eGFR normalization in the whole cohort and mild SCD genotypes. While possibly representing improvement in hyperfiltration following hydroxyurea treatment, this finding more likely suggests that receiving hydroxyurea had more severe disease and greater probability of losing kidney function and “normalizing” from hyperfiltration.

ACE-I/ARB therapy was associated with decreased odds of hyperfiltration in the entire cohort and in severe genotypes patients. ACE-I/ARBs decrease albuminuria in SCD-related glomerulopathy,(47, 48) and appear to slow eGFR decline.(49) Although renin-angiotensin-aldosterone system (RAAS) blockade has hemodynamic effects, the observed relationship between ACE-I/ARBs and hyperfiltration may be related to common use of these agents in albuminuric patients, combined with high risk of kidney function loss in these same patients.(11) More studies of RAAS blockade are required to evaluate their effect on hyperfiltration.

In the overall cohort and in severe genotypes, those with higher baseline eGFR were less likely to normalize eGFR over the observation period. These patients could eventually have eGFR decline if followed further over time. Because their eGFR is higher at the start of observation, they would require greater decline to reach defined normal range. Based on these findings, one could hypothesize those adults with lower baseline eGFR closer to normal range have passed a hyperfiltration peak when younger and are now in a phase of declining eGFR.

Our study provides valuable information from “real-life” SCD patients and identifies clinical and laboratory factors associated with hyperfiltration, its association with proteinuria and eGFR “normalization”. Our study is limited by its retrospective, observational nature. Evaluating hyperfiltration is dependent upon the timepoint at which these patients entered the study. If hyperfiltration is present at baseline, it is plausible eGFR may improve to normal values or patients may exhibit eGFR loss due to CKD progression, which would both be captured as “normalization”. Assessing which outcome is true is difficult unless GFR is accurately measured over time. As eGFR underestimates kidney function in patients with SCD, it is conceivable that current creatinine-based definitions of hyperfiltration are inaccurate.(50, 51) Another limitation is that of unavailable variables of potential interest. Both hyperuricemia and nocturnal hypertension have been identified as risk factors for declining eGFR in young SCD patients,(52, 53) however, these measures were not routinely assessed in this adult cohort. Missing data for some available variables of interest limited our ability to assess findings noted in age- and sex-adjusted analyses in our more comprehensive multivariable models. For example, hemoglobinuria was associated with baseline hyperfiltration in severe genotypes and hemoglobin F was associated with time-to-normalization in both the entire cohort and severe genotypes. Neither of these could be evaluated in multivariate models, limiting interpretation of these data. Our proteinuria assessment was based upon semi-quantitative urinary dipstick testing and may not have accurately captured albuminuria.

We confirm the high prevalence of glomerular hyperfiltration in adults with SCD as well as its association with age and severe SCD genotypes. Longitudinal studies starting in childhood, with systematic capture of variables of interest, are required to better define the natural history of hyperfiltration and factors associated with eGFR decline to normal values.