Relationship between liver function tests & cardiovascular risk factors in stage 3-5 pre-dialysis chronic kidney disease

Chronic kidney disease (CKD) is characterized by a progressive and an irreversible deterioration of the renal function which is associated with increased morbidity as well as mortality particularly from cardiovascular disease (CVD)1. CVD events increase rapidly through CKD stage 3 by 43 per cent and stage 5 by 343 per cent12. Although the mechanism underlying the increased CVD risk is not well established, the presence of multiple classical and uraemia-related risk factors in CKD patients leads to premature cardiovascular mortality3. A number of uraemia-related biomarkers are identified as predictors of cardiac outcomes, however, there is limited utility of these diagnostic modalities in patients with CKD23.

Over the past decade, liver function tests (LFTs) including gamma-glutamyltransferase (GGT), alanine aminotransferase (ALT), alkaline phosphatase (ALP) and aspartate aminotransferase (AST) have emerged as markers of CVD risk in both population-based studies and patients with coronary artery disease456 independent of their relationship with metabolic syndrome or non-alcoholic steatohepatitis (NASH). The association of LFTs with CVD risk, irrespective of metabolic syndrome or NASH, has been linked to various mechanisms including direct or indirect contribution of GGT activity within atherosclerotic lesions, vascular calcification, pro-inflammatory activities, endothelial dysfunction, oxidative stress and impaired haemostasis4,5,7. Although LFTs have been proposed as independent risk markers for all-cause and cardiovascular death in various patient populations, the association between LFTs and CVD risk has not been extensively evaluated in patients with CKD. In the present study, the relationship between LFTs and biochemical cardiovascular risk factors (CRFs) were evaluated in stage 3-5 pre-dialysis CKD patients.

Material & Methods

Adult patients with stage 3 to 5 CKD without dialysis treatment presenting to outpatient nephrology clinic for follow up care between January 2014 and December 2017 were retrospectively considered for the study at Dışkapı Yıldırım Beyazıt Education and Research Hospital, Ankara, Turkey. Patients with active viral hepatitis B or C diagnosed by serological tests including HBsAg, anti-HBs, IgM/IgG anti-HBc, anti-HCV, HCV RNA (n=1), NASH (n=5), autoimmune hepatitis (n=1), chronic liver disease (n=4), cirrhosis (n=2), patients with acute/chronic inflammatory/autoimmune diseases (n=8), patients with malign disease (n=2) or history of malignancy (n=2) and CKD patients due to kidney transplant dysfunction (n=9) were excluded. The diagnosis of the liver diseases was based on medical history and biochemical and radiological findings without a liver biopsy. Thus, a total of 246 patients were enrolled for the study. The study protocol was approved by the local Ethics Committee in Ankara, Turkey.

The demographic information of all patients was collected from patients’ charts. LFTs (ALT, AST and GGT) and biochemical inflammatory and cardiovascular risk parameters were also recorded. All demographic and biochemical parameters were entered into a database for further analysis. Glomerular filtration rate (GFR) was estimated using the CKD-EPI equation (eGFR). Patients were assigned to eGFR stages as follows: stage 3 (eGFR 30-59 ml/min per 1.73 m²), stage 4 (eGFR 15-29 ml/min per 1.73 m²) and stage 5 (eGFR <15 ml/min per 1.73 m²).

Statistical analysis: Kolmogorov–Smirnov test was employed to check for the distribution of variables. Continuous variables with normally distributed data were presented as mean ± standard deviation. Unequal distributed variables were expressed as medians (interquartile ranges: Q1-Q3). Categorical variables were expressed as numbers and percentages. Spearman’s correlation coefficient was used to evaluate the relationship between variables since some variables were unequally distributed. A partial correlation analysis was used to remove the effect of GFR on the association between variables. Significance was considered at a two-tailed value of P<0.05. SPSS software (Statistical Package for the Social Sciences, version 20.0; SPSS Inc., Chicago, IL, USA) was used for carrying out all the statistical analysis in this study.

Results & Discussion

A total of 246 patients were included in the study. Baseline clinical and biochemical characteristics are shown in Table I. A correlation analysis was performed to explore the association between LFTs and biochemical CRFs. ALT was positively correlated with GFR, serum albumin, triglyceride and 25(OH)D. AST was positively correlated with GFR, serum albumin, high-density lipoprotein cholesterol (HDL-C) and 25(OH)D. Both ALT and AST correlated negatively with CRP and intact parathyroid hormone (iPTH). GGT was positively correlated with GFR, CRP and triglyceride and negatively correlated with HDL-C. When the analysis was performed on diabetic patients only, ALT was positively correlated with GFR; AST was positively correlated with GFR and HDL-C and negatively correlated with iPTH; GGT had no correlation (Table II). Since ALT, AST and GGT were all positively correlated with GFR in the entire study population, we further analyzed to see whether biochemical cardiovascular risk parameters which correlated with LFTs, were also affected by GFR. In the correlation analysis, GFR was positively correlated with serum albumin, triglyceride and 25(OH)D and negatively correlated with CRP, iPTH and albuminuria in both the overall study population and in the diabetic group. In partial correlation analysis performed to remove the effect of GFR on parameters tested, no significant correlation was found between LFTs and cardiovascular risk parameters. Likewise, no significant correlation existed between liver functions and CRFs after controlled for body mass index (BMI), GFR and presence of diabetes (Table III).

Serum aminotransferase levels have been previously found to be close to the lower limit of the normal range in patients with chronic dialysis8. Although the reduction in aminotransferase values in pre-dialysis CKD patients was reported in small patient groups, only AST level was analyzed9. The findings of the present study confirm that both ALT and AST levels are depressed in pre-dialysis CKD patients. High levels of uraemic toxins leading to suppression of aminotransferase synthesis and release from the hepatocyte into the bloodstream have been suggested to have a role in low serum aminotransferase levels89. Besides being an independent predictive marker of CVD in general population, GGT has also been suggested as a strong, independent risk factor for all-cause and cardiovascular death in haemodialysis patients10. However, only two studies have evaluated the association of GGT levels with CVD risk in pre-dialysis CKD patients. The first study retrospectively found an independent association between GGT and mortality. GFR values were not different between patients with high or normal serum GGT levels11. The second study demonstrated a significant association between circulating GGT and endothelial dysfunction in CKD stage 3-5 patients. Serum GGT levels were found to be inversely related to GFR12. Contrary to this, in our study, GGT levels decreased in parallel with the severity of CKD. This inconsistency may be due to the fact that those studies included CKD patients with a narrow range of GFR and the proportion of CKD patients did not adequately represent the actual distribution of patients due to CKD stages. In the present study, distribution of patients by stage of CKD was more appropriate than previous reports which allows for generalization of the findings across stage 3-5 pre-dialysis CKD population.

In the present study, all biochemical CRFs in correlation with liver enzymes also correlated with GFR. This finding is not surprising since traditional and non-traditional risk factors such as inflammation, malnutrition, endothelial dysfunction, oxidative stress, vascular calcifications, vitamin D deficiency and abnormal calcium-phosphate metabolism deteriorate with progressive decline in renal function13. Albuminuria has also been found to be associated with the progression of kidney disease14. After negating the effect of GFR, liver enzymes were no longer found to be related to biochemical CRF. This finding suggests that LFTs may not reflect the actual CVD risk in pre-dialysis CKD patients. Rather, the potential link between LFTs and biochemical CRFs may be secondary to the effect of decreased GFR on both circulating markers of CRF and LFTs.

Diabetes mellitus, BMI and NASH (non-alcoholic steatohepatitis) have been implicated with increased serum levels of ALT and GGT, are also related to CVD risk in the general population15. This correlation may be extended to the findings of the present study. CKD patients with NASH in this study may have more elevated levels of liver enzymes than CKD patients without NASH, and this difference could have potentially influenced our analysis. Keeping this in mind, in order to reduce this effect, patients with NASH were excluded. Furthermore, all correlation analyses were repeated for our diabetic subjects separately, and the results were similar to those obtained for the entire study population. Interestingly, there was no association between BMI and liver enzymes in our study population. Although this finding is seemingly unexpected, we believe that this point is the novelty of this study. That is, diabetes, NASH, BMI and CVD risk factors are traditionally well established to associate with each other in the general population. However, with a progressive decrease in kidney function, such associations become indistinct, and the kidney function takes an active role in those associations.

Although the present study included a relatively small number of patients for analysis, the distribution of the patient population was representative for CKD stages. In addition, since stage 3-5 CKD patients typically had a substantially higher risk for CVD than patients with stage 1-2 CKD, and patients on haemodialysis are affected by many other factors for CVD risk which arise from haemodialysis treatment itself, in the present study, patients with stage 3-5 pre-dialysis CKD represent a relatively homogeneous CKD population for potentially increased CVD risk factors. Therefore, our findings, although, obtained from a relatively small number of patients, provide valuable information and may be generalized to stage 3-5 pre-dialysis CKD population and even though subgroup analysis could not be performed in each CKD stage. Since the retrospective cross-sectional design of our study was not amenable to observe the relation between CRFs, LFTs and GFR in a longitudinal manner, prospective studies with larger sample sizes are needed to confirm our findings.