*Article* **Plasma Oxalate as a Predictor of Kidney Function Decline in a Primary Hyperoxaluria Cohort**

**Ronak Jagdeep Shah 1, Lisa E. Vaughan 2, Felicity T. Enders 2, Dawn S. Milliner 1,3 and John C. Lieske 1,4,\***


Received: 1 April 2020; Accepted: 16 May 2020; Published: 20 May 2020

**Abstract:** This retrospective analysis investigated plasma oxalate (POx) as a potential predictor of end-stage kidney disease (ESKD) among primary hyperoxaluria (PH) patients. PH patients with type 1, 2, and 3, age 2 or older, were identified in the Rare Kidney Stone Consortium (RKSC) PH Registry. Since POx increased with falling estimated glomerular filtration rate (eGFR), patients were stratified by chronic kidney disease (CKD) subgroups (stages 1, 2, 3a, and 3b). POx values were categorized into quartiles for analysis. Hazard ratios (HRs) and 95% confidence intervals (95% CIs) for risk of ESKD were estimated using the Cox proportional hazards model with a time-dependent covariate. There were 118 patients in the CKD1 group (nine ESKD events during follow-up), 135 in the CKD 2 (29 events), 72 in CKD3a (34 events), and 45 patients in CKD 3b (31 events). During follow-up, POx Q4 was a significant predictor of ESKD compared to Q1 across CKD2 (HR 14.2, 95% CI 1.8–115), 3a (HR 13.7, 95% CI 3.0–62), and 3b stages (HR 5.2, 95% CI 1.1–25), *p* < 0.05 for all. Within each POx quartile, the ESKD rate was higher in Q4 compared to Q1–Q3. In conclusion, among patients with PH, higher POx concentration was a risk factor for ESKD, particularly in advanced CKD stages.

**Keywords:** plasma oxalate; primary hyperoxaluria; estimated glomerular filtration rate; chronic kidney disease; Urine Oxalate; end-stage renal disease

### **1. Introduction**

Primary hyperoxaluria (PH) is a rare inherited autosomal recessive genetic disease caused by defects in genes that encode proteins important for glyoxylate metabolism [1]. Currently, three distinct forms are known—PH1 results from mutations in the enzyme alanine-glyoxylate aminotransferase (AGT) which is encoded the *AGXT* gene, PH2 is caused by a deficiency of the glyoxylate reductase/hydroxypyruvate reductase (GRHPR) enzyme encoded by *GRHPR*, and PH3 occurs when the mitochondrial enzyme, 4-hydroxy-2-oxoglutarate aldolase (HOGA) is deficient due to mutations in the *HOGA1* gene. Based upon current numbers in the Rare Kidney Stone Consortium (RKSC) PH registry, approximately 70% of diagnosed patients are PH1, 10% are PH2, 10% PH3, and 10% do not have an identified genetic cause [2].

The metabolic consequence of each of these enzyme deficiencies is a marked increase in hepatic oxalate production. Since oxalate cannot be metabolized by humans, the excess released into the plasma must be excreted by the kidneys, with less than 10% eliminated through the gastronintestinal tract. Calcium oxalate stones and nephrocalcinosis can result from high urinary oxalate (UOx) excretion; the latter can be associated with interstitial inflammation and fibrosis and may contribute to progressive

chronic kidney disease (CKD) and end-stage kidney disease (ESKD) [3]. Once patients approach ESKD, excess oxalate can no longer be eliminated by the kidneys, and it accumulates in the body, potentially leading to systemic oxalosis. Among PH patients, there is wide variability in clinical course, with some progressing to ESKD in early childhood, while other PH patients retain kidney function into their fifth or sixth decade. Prediction of long-term outcomes using biomarkers is an important tool for clinical management, particularly now that novel treatment strategies with the potential to reduce hepatic oxalate production in PH are ready for clinical trials [4]. Patients with PH typically excrete >0.7 mmol/1.73 m2/day [5], and we previously reported that higher UOx predicts future ESKD risk within the PH patient group [2]. We also recently reported that plasma oxalate (POx) concentration correlates with UOx excretion [6].

Since 24 h urine collections can be difficult to obtain, especially on a repeated basis or in younger children, a blood biomarker that predicts UOx and other clinical features of PH could be clinically valuable. Furthermore, in patients with advanced CKD, UOx may no longer reflect systemic oxalate burden; in such cases, POx may represent a more accurate biomarker [7]. Therefore, we examined data in the RKSC PH registry in order to determine whether POx represents a viable biomarker that predicts the future loss of kidney function among patients with confirmed PH and at varying CKD stages.

#### **2. Results**

#### *2.1. Baseline Characteristics*

There were 227 patients who met the criteria for this study (Figure 1). During follow-up, ESKD developed in nine of the 118 patients (7.6%) in the CKD 1 group, 29 of 135 (21.5%) of the CKD 2 patients, 34 of 72 (47.2%) in the CKD3a, and 31 of 45 (68.8%) in the CKD 3b. There was one death in the CKD 1 group, nine deaths each in the CKD 2 and 3a groups, and 10 deaths in the CKD 3b group. The proportion of patients with PH1 increased by CKD stage (Table 1), representing 56.8% with CKD1, 73.3% with CKD2, 86.1% with CKD3a, and 84.4% with CKD3b. Due to the analysis plan, the median age at PH diagnosis also differed according to which patients experienced each CKD stage, from 5.4 years (CKD1) to 16.0 years (CKD3b). Median follow-up time was 5.3, 8.8, 6.6, and 1.8 years for CKD stages 1–3b, respectively.

**Figure 1.** Flowchart of inclusion criteria for analysis cohort. From a total of 545 PH1 patients in the Registry, 227 were eligible for this analysis.


**Table 1.** Clinical characteristics of patients with primary hyperoxaluria who did not have ESKD at or before diagnosis.

Continuous variables are expressed as median with 25th, 75th percentiles. *n*, number; PH, primary hyperoxaluria; PH1, primary hyperoxaluria type 1; PH2, primary hyperoxaluria type 2; PH3, primary hyperoxaluria type 3; y, years.

Baseline POx increased by CKD stage from (3.1 (2.1, 5.7) μmol/L in CKD1 (*n* = 38) to 14.4 (10.5, 20.0) μmol/L in CKD3b (*n* = 17) (Table 2). The results were similar within the PH1 subset, increasing from 3.9 [2.4, 6.8] μmol/L in CKD1 (*n* = 24) to 14.9 (11.6, 21.5) μmol/L in CKD3b (*n* = 16). The numbers were not sufficient for a similar sub-analysis in PH2 and PH3. The risk of incident ESKD was higher in patients with PH1 compared to PH2 and PH3 in CKD2 and CKD3b; the results were similar albeit non-significant in CKD1 (HR 7.45; 95% CI 0.92–60.2; *p* = 0.06) and CKD3a (HR 5.74; 95% CI 0.78–42.1; *p* = 0.085) (Table 3).


**Table 2.** Baseline and follow-up POx quartiles, by CKD stage.

### *2.2. POx and ESKD*

When treated as a continuous predictor, higher baseline POx values were significantly associated with a higher risk of ESKD in CKD2 (HR 1.17; 95% CI 1.01–1.35; *p* = 0.033), CKD3a (HR 1.29; 95% CI 1.09–1.53; *p* = 0.004) and CKD3b (HR 1.24; 95% CI 1.08–1.42; *p* = 0.003) (Table 3). Baseline POx values in Q4 compared to POx in Q1 were also associated with a higher risk of ESKD in CKD3a (HR 13.88; 95% CI 1.41–137; *p* = 0.024) and CKD3b (HR 42.1; 95% CI 3.29–539; *p* = 0.004) (Table 3).

During follow-up (Table 4), POx was significantly associated with ESKD risk across all CKD stages: CKD1 (HR 1.12; 95% CI 1.02–1.24, *p* = 0.018), CKD2 (HR 1.17; 95% CI 1.08–1.25; *p* < 0.001), CKD3a (HR 1.19; 95% CI 1.11–1.27; *p* < 0.001), and CKD3b (HR 1.12; 95% CI 1.04–1.21; *p* = 0.003). When POx was considered by quartile, Q4 was a significant predictor of ESKD compared to Q1 across

the CKD stages as well: CKD 2 (HR 14.2; 95% CI 1.76–115; *p* = 0.013), CKD3a (HR 13.7; 95% CI 3.02–62; *p* < 0.001), and CKD3b (HR 5.19; 95% CI 1.10–24.5; *p* = 0.038). Within each POx quartile, the ESKD rate was higher for the later CKD stages. Within each CKD stage, the ESKD rate was also higher in the fourth POx quartile compared to the first three (Table S1 and Figure 2). Thus, the greatest ESKD rate was for the CKD 3b subjects in the highest POx quartile (Figure 2).

**Figure 2.** ESKD rate by POx quartile during follow-up by CKD stage. ESKD rates were estimated for each CKD stage group by dividing individual patient follow-up times into intervals based on the time between the POx measures or last follow-up. Person-time and ESKD events were summed within POx quartiles with the rate = [100 × (events/person-time)]; error bars represent 95% CI of the ESKD rate (see Supplemental Table S1 for numerical values). ESKD rates were similar for the lower three quartiles (Q) but increased for the highest POx quartile across CKD stages 2–3b (Q4 vs. Q1 HR 14.21; 95% CI 1.76–114.7; \* *p* < 0.05 for CKD stage 2, HR 13.66; 95% CI 3.02–61.91; \*\* *p* < 0.001 for CKD stage 3a, and HR 5.19; 95% CI 1.10–24.5; \* *p* < 0.05 for CKD stage 3b).



PH1, primary hyperoxaluria type 1; *n*: number available for analysis; E: = ESKD events; 95% CI, 95% confidence interval. *p*-values in bold denote significance at the 0.05 level. Harrell's c index is provided. The proportional hazards assumption was met for all models with reported HRs. † Not estimable, sample size, and no. of events too small for a variable with four levels.



Not estimable, sample size, and no. events, too small for a variable with four levels. *p*-values in bold denote significance at the 0.05 level. Harrell's c index is provided. *n* Follow-up intervals;E=ESKDevents.
