Nephropathy Progression in African Americans With a Family History of ESKD: Implications for Clinical Trials in APOL1-Associated Nephropathy

To the Editor:

Individuals with APOL1 high-risk genotypes possess 2 renal risk variants (RRVs) and generally have faster declines in eGFR than those with low-risk genotypes (0/1 RRV). However, genotypes may not predict outcomes in a given patient.^1,2 Before the APOL1 discovery, family history of ESKD was a recognized CKD risk factor in African Americans.³

The Natural History of APOL1-Associated Nephropathy Study (NHAANS) followed up 833 African American first-degree relatives of nondiabetic ESKD probands from 505 families to monitor disease course; 22% had high-risk APOL1 genotypes.^4,5 Up to 3 first-degree relatives of an African American index case with ESKD potentially due to APOL1 were recruited and their rates of change in eGFR, UACR, serum creatinine, and cystatin C were determined. Most participants were unaware of whether they had CKD; those with ESKD were not eligible. Preference was given to siblings and children ± 15 years of the age at onset of ESKD in the family proband. Additional recruitment criteria and statistical methods are in Item S1. Wake Forest School of Medicine’s Institutional Review Board (#00009019) approved NHAANS; participants provided written informed consent. After baseline visits, APOL1 low-risk participants were re-evaluated about 2 years later; high-risk participants were re-evaluated twice more over 3 years.

Table 1 displays baseline data by APOL1 genotype. Low-risk individuals had 2 visits over a median of 2.4 (IQR, 2.3-2.6) years; high-risk participants had 3 visits over a median of 3.4 (IQR, 2.8-3.6) years. Unadjusted analyses comparing high- to low-risk participants showed earlier age at ESKD in family probands (P = 0.008) and nominally higher baseline UACR in high-risk participants, though the latter was not statistically significant (P = 0.06). Table 2 displays changes in clinical variables between the baseline and final visits by APOL1 genotype. Although unadjusted analyses revealed UACR, BMI, and systolic BP increased more in high-risk participants, analyses relied on the robustness of the Wilcoxon test against mild deviation from independence and did not consider age at study (relative to age at ESKD in family probands) or follow-up duration.

Table 1.

Baseline Demographic and Clinical Data in NHAANS Participants

	Full Sample		0/1 APOL1 Risk Variants		2 APOL1 Risk Variants
Variables	N	Value	n	Value	n	Value	P
African ancestry proportion	789	0.8 [0.8, 0.9]	613	0.8 [0.8, 0.9]	176	0.8 [0.8, 0.9]	0.4
Age, y	833	49.3 [39.3, 58.0]	650	49.5 [39.6, 58.3]	183	48.6 [38.9, 56.5]	0.1
Proband age at ESKD onset, y	833	49.0 [39.0, 57.0]	650	50.0 [40.0, 57.0]	183	44.0 [36.0, 56.5]	0.008
Female sex	833	60.3%	650	59.2%	183	63.9%	0.3
BMI, kg/m2	826	31.2 [26.4, 36.9]	644	31.4 [26.6, 37.0]	182	30.7 [25.4, 36.4]	0.3
ACEi/ARB use	833	9.4%	650	10.2%	183	6.6%	0.1
Cystatin C, mg/L	828	0.8 [0.7, 0.9]	645	0.8 [0.7, 0.9]	183	0.8 [0.6, 0.9]	0.8
eGFRcr, mL/min/1.73 m2	833	96.2 [77.2, 112.3]	650	96.6 [78.4, 112.2]	183	95.0 [74.9, 112.2]	0.6
eGFRcr-cys, mL/min/1.73 m2	828	102.2 [84.1, 117.2]	645	102.8 [85.0, 117.1]	183	100.3 [82.4, 117.1]	0.8
eGFRcys, mL/min/1.73 m2	828	106.7 [85.8, 119.3]	645	106.8 [86.8, 118.9]	183	105.4 [84.8, 121.0]	0.8
Kidney diseasea	832	23.2%	649	22.7%	183	25.1%	0.5
Stage 3a CKD (eGFR 45-60)	833	6.2%	650	6.2%	183	6.6%	0.8
Stage 3b CKD (eGFR 30-45)	833	2.2%	650	2.3%	183	1.6%	0.6
Stage 4 CKD (eGFR 15-30)	833	0.7%	650	0.8%	183	0.5%	0.8
Stage 5 CKD (eGFR < 15)	833	0.2%	650	0.2%	183	0.5%	0.3
UACR, mg/g	832	5.3 [2.7, 16.6]	649	5.2 [2.6, 16.3]	183	6.8 [3.0, 18.4]	0.06
UACR > 30 mg/g	832	13.1%	649	12.9%	183	13.7%	0.8
UACR > 300 mg/g	832	4.9%	649	4.3%	183	7.1%	0.1
Serum glucose, mg/dL	828	88.0 [80.0, 99.0]	645	88 [80.0, 99.0]	183	86 [79.0, 96.0]	0.1
Diabetes	794	20.3%	617	21.1%	177	17.5%	0.3
Diastolic BP, mm Hg	823	85.0 [78.0, 95.5]	641	85.0 [77.5, 95.5]	182	85.2 [79.0, 96.5]	0.7
Systolic BP, mm Hg	823	126.5 [115.0, 143.0]	641	126.5 [115.5, 144.0]	182	126.2 [114.0, 140.5]	0.3
Hypertensionb	823	42.0%	641	42.4%	182	40.7%	0.7

Abbreviations: ACEi/ARB, angiotensin-converting enzyme inhibitor/angiotensin receptor blocker use; APOL1, apolipoprotein L1 gene; BMI, body mass index; BP, blood pressure; CKD, chronic kidney disease; CKD-EPI, CKD Epidemiology Collaboration; eGFR, estimated glomerular filtration rate; eGFR_cr, eGFR calculated using creatinine-based CKD-EPI equation; eGFR_cr-cys, eGFR calculated using creatinine- and cystatin C–based CKD-EPI equation; eGFR_cys, eGFR calculated using cystatin C–based CKD-EPI equation; ESKD, end-stage kidney disease; UACR, urinary albumin-creatinine ratio.

^aDefined as eGFR_cr < 60 mL/min/1.73 m² and/or UACR > 30 mg/g.

^bDefined as prior diagnosis, use of antihypertensive medications, or study BP > 140/90 mm Hg.

Table 2.

Changes in Clinical and Laboratory Parameters in NHAANS Participants Between First and Last Visit

	Full Sample		0/1 APOL1 Risk Variants		2 APOL1 Risk Variants
Variable Assessed for Change	N	Value	n	Value	n	Value	P
Systolic BP, mm Hg	650	3.0 [−7.5, 14.0]	489	2.0 [−8.5, 13.0]	161	6.0 [−3.5, 15.5]	0.01
Diastolic BP, mm Hg	650	0.0 [−8.0, 8.0]	489	0.0 [−8.0, 70]	161	1.5 [−9.0, 9.0]	0.2
BMI, kg/m2	653	0.2 [−1.5, 2.0]	490	0.03 [−1.7, 1.8]	163	0.4 [−0.9, 2.3]	0.03
Serum glucose, mg/dL	600	3.0 [−7.0, 14.0]	490	3.0 [−8.0, 14.0]	110	4.5 [−4.0, 16.0]	0.05
UACR, mg/g	664	1.0 [−2.2, 8.5]	495	0.7 [−2.6, 7.5]	169	2.2 [−1.1, 13.7]	0.03
Cystatin C, mg/L	651	0.0 [−0.1, 0.1]	490	0.02 [−0.05, 0.1]	161	−0.01 [−0.08, 0.05]	<0.001
eGFRcr, mL/min/1.73 m2	656	1.3 [−6.8, 10.2]	495	1.2 [−7.1, 9.7]	161	1.4 [−5.9, 10.5]	0.6
eGFRcr-cys, mL/min/1.73 m2	657	−3.0 [−9.6, 3.0]	490	−3.0 [ −9.7, 3.0]	167	−2.9 [−9.3, 2.8]	0.7
eGFRcys, mL/min/1.73 m2	657	0.0 [0.0, 0.1]	490	0.02 [−0.05, 0.1]	167	0.02 [−0.03, 0.09]	0.9

Note: Values given as median [interquartile range]. Median time between the baseline and final visit was 2.4 [2.3, 2.6] years in the APOL1 0/1 risk variant carriers and 3.4 [2.8, 3.6] years in the APOL1 2 risk variant carriers.

Table S1 displays longitudinal mixed-model results considering age, sex, baseline eGFR_cr and UACR, time between visits, and APOL1 genotype for effect on eGFR change. APOL1 genotypes were not significant predictors of change in eGFR_cr, serum creatinine, serum cystatin C, or UACR. Computing eGFR_cys or eGFR_cr-cys yielded similar results. Baseline eGFR and follow-up time predicted change in eGFR; baseline UACR and APOL1 genotype lacked significant effects. Baseline cystatin C was the only parameter associated with change in cystatin C. Baseline serum creatinine and UACR were the only parameters associated with change in serum creatinine, and change in UACR was associated only with baseline UACR. Median 2.5-year changes in UACR, serum creatinine, cystatin C, and eGFR fluctuated regardless of APOL1 genotype; increasing UACR and reduction in eGFR were infrequent.

These results are important for designing clinical trials to assess the effectiveness of APOL1-based therapies. For example, an antisense oligonucleotide directed against APOL1 may prevent nephropathy in African ancestry populations.⁶ It is imperative that clinical end points accurately reflect treatment efficacy in studies. In patients with focal segmental glomerulosclerosis and nephrotic syndrome, proteinuria reduction predicts slower rates of eGFR decline.⁷ However, the majority of African Americans with APOL1-associated nephropathy have low levels of proteinuria, with protein excretion typically <2 g/d (often normal).^8,9 Novel biomarkers of kidney disease are currently under study, including tumor necrosis factor receptor 1 and 2 and kidney injury molecule 1.¹⁰ It is unclear whether they predict risk for CKD or reflect ongoing kidney injury less responsive to therapy in low-proteinuria patients.

NHAANS contributes to understanding the risk for developing CKD in healthy carriers of APOL1 RRVs; most participants lacked nephropathy at baseline. Considering conventional risk factors (baseline eGFR and UACR) and dynamic variables (interval between measurements), analyses revealed similar eGFR and UACR trajectories in healthy persons with 0/1 or 2 APOL1 RRVs. Small changes in eGFR and UACR were observed in the full cohort, with large interindividual variation in the direction of change (Figs S1 and S2).

A limitation of NHAANS was the inability to incorporate exposure to environmental protective or risk factors that may occur before, during, or after the observation periods. Stored biosamples from participants at each study visit will permit future analyses of these types. Participation in NHAANS may have led to earlier CKD diagnoses with alterations in treatment. Thus, the observed course may not fully reflect the natural history. Relatively short follow-up may have precluded identification of an association between carriage of APOL1 RRVs and CKD. As knowledge of the pathogenesis of APOL1-associated nephropathy increases, future clinical studies should commit to long-term follow-up and include dynamic environmental risk factor assessments.

In conclusion, NHAANS demonstrated wide variability in change in eGFR and UACR over a median 2.4- to 3.4-year follow-up in African Americans at risk for nephropathy based on family history of nondiabetic ESKD. APOL1 genotypes were not predictive of eGFR change during the study; other measures may have utility. Follow-up durations need to be considered when developing clinical trials that aim to slow nephropathy progression in African Americans with subnephrotic proteinuria.