MO428: Fibrosis Markers, But Not Circulating Autotaxin (ATX) or Lysophosphatidic Acid (LPA) Associate With Disease Progression in Diabetic Kidney Disease (DKD)

Abstract

Abstract BACKGROUND AND AIMS Diabetic kidney disease (DKD) remains the most common cause of the end-stage renal disease (ESRD) worldwide. Kidney fibrosis predicts functional decline and ESRD in many, but not all patients [1]. Autotaxin (ATX) is the enzyme converting lysophosphatidylcholine to Lysophosphatidic acid (LPA), which is a key factor in the development of fibrosis and inflammation. Preclinical studies have shown that inhibition of ATX may be beneficial in DKD [2, 3]. In this study, we assessed the relationship between systemic/circulating ATX, LPA and fibrosis markers of collagen type III degradation (C3M), type VI formation (PRO-C6) and connective tissue growth factor (CTGF) with kidney disease stage in a retrospective DKD cohort. METHOD We measured ATX, LPA 18:2, LPA 20:4, pPRO-C6 and pCTGF in plasma, and uPRO-C6, uC3M in urine in 203 patients (including DKD and control) at recruitment and at a follow-up (FU) visit (n = 108). After adjusting for missingness, skewness, highly correlated confounders and excluding patients on dialysis, the cohort resulted in n = 155 patients at baseline, n = 53 at FU and n = 144 for the time to adverse outcome. To assess the relationship between the markers of interest with CKD stage and estimated glomerular filtration rate (eGFR), Bayes ordinal logistic regression and linear regression were applied, respectively. A random forest (RF) algorithm was used also to predict both outcomes and identify the most relevant predictors. Cox regression analysis was used to assess the relationship between the markers at baseline and time to adverse outcomes, defined as a decline in eGFR >30% or death. A sensitivity analysis including relevant confounding covariates was applied to all regression models. Benjamini & Hochberg adjustment was used to account for multiplicity. RESULTS pPRO-C6 and uC3M levels showed consistently significant associations with CKD stage (Figure 1a-b) also after adjusting for confounding variables both at baseline and FU. Higher CKD stages or lower eGFR values were associated with higher pPRO-C6 and lower uC3M levels. RF (10-fold cross-validated mean AUC = 0.75) also selected these markers as the most relevant predictors for CKD stage (Figure 1c). Similar results were obtained when using eGFR. During a median follow-up time of 4.00 years, 14.58% of the patients had an adverse outcome. Both baseline pPRO-C6 [hazard ratio (HR) 2.68 (95% confidence interval, CI 1.47–4.88, P -adjusted < 0.05 per 2-fold ng/mL increase] (Figure 2) and uPRO-C6 levels [HR 1.43 (1.08–1.89, P -adjusted< 0.05) per 2-fold ng/mmol increase] were significantly associated with development of adverse outcomes, whereas only a trend was found for uC3M level [HR 0.50 (0.27–0.91, P -adjusted< 0.06) per 2-fold µg/mmol increase]. After adjusting for confounders, none of the markers showed a significant association with time to adverse outcome. CONCLUSION pPRO-C6 and uC3M showed consistent associations with CKD stage and pPRO-C6, uPRO-C6 and uC3M (trend) also an association with the risk of adverse outcomes, which is in alignment with previous studies [4]. In contrast, no associations were observed for circulating ATX or LPAs. However, results should be interpreted carefully due to the relatively low number of patients in this cohort and missing histology to verify fibrosis or local ATX/LPA action.