Abstract
Abstract. Serum creatinine, a surrogate for both renal function and homocysteine generation, is an important determinant of fasting plasma total homocysteine levels in stable renal transplant recipients. In this study, it is hypothesized that among stable renal transplant recipients with normal creatinine levels (i.e., ≤ 1.5 mg/dl), serum cystatin C, a more sensitive indicator of GFR, would better predict fasting total homocysteine levels compared with serum creatinine. Fasting plasma total homocysteine, folate, vitamin B12, and pyridoxal 5′-phosphate levels, along with serum cystatin C, creatinine, and albumin levels, were determined in 28 consecutive renal transplant recipients (mean age 47 ± 14 yr; 60.7% men) with stable allograft function, whose serum creatinine was ≤1.5 mg/dl. General linear modeling with analysis of covariance revealed that serum cystatin C was independently predictive (partial R = 0.494; P = 0.023) of fasting total homocysteine levels after adjustment for age, gender, vitamin status, albumin, and creatinine levels. In contrast, creatinine levels were not predictive of fasting total homocysteine levels in this model (P = 0.110) or an identical model that excluded cystatin C (P = 0.131). Serum cystatin C levels may reflect subtle decreases in renal function that independently predict fasting total homocysteine levels among stable renal transplant recipients with a normal serum creatinine.
Recently, we provided controlled evidence that clinically stable renal transplant recipients have an excess prevalence of mild hyperhomocysteinemia (1), which may contribute to their disproportionately high rates of arteriosclerotic outcomes (2). Creatinine, a surrogate for both renal function (i.e., GFR [3]) and stoichiometric generation of homocysteine coupled to creatine-creatinine synthesis (4), is an important determinant of total homocysteine (tHcy) levels in both renal transplant recipients (1,5) and the general population (6). It has been demonstrated, however, that the specific relationship between serum creatinine and GFR among renal transplant recipients is highly variable, being influenced by altered muscle mass and catabolic rate, as well as changes in tubular secretion of creatinine (7). Cystatin C is a nonglycosylated 13-kD basic protein produced at a stable rate by all investigated nucleated cells, whose serum concentration is primarily determined by the GFR (8). Consistent reports (9,10,11,12) now clearly indicate that serum cystatin C is superior to serum creatinine as an index of GFR in adults, including renal transplant recipients (12), particularly for the detection of early decreases in GFR (11,12). Using iohexol clearance data, it has previously been shown (13) that there is a strong, independent (inverse) association between GFR, but not serum creatinine, and fasting tHcy levels, which encompasses GFR that are only very mildly reduced (i.e., 50 to 75 ml · min-1). Accordingly, we hypothesized that among renal transplant recipients with normal creatinine levels (≤1.5 mg/dl), serum cystatin C, as a more sensitive indicator of GFR, would better predict fasting tHcy levels compared to serum creatinine. To test this hypothesis, fasting plasma tHcy, serum creatinine, and serum cystatin C were assessed, along with the other established determinants of tHcy levels (age [14], gender [14], B-vitamin status [14], and albumin [15]), in consecutive renal transplant recipients with clinically stable allograft function whose serum creatinine was ≤1.5 mg/dl.
Materials and Methods
The institutional review board at Rhode Island Hospital (Providence, RI) approved the study protocol, and all participants provided written informed consent. Study participants were 28 stable renal transplant patients (i.e., they were at least 6 mo posttransplantation with no clinical evidence of renal graft rejection), whose serum creatinine was ≤1.5 mg/dl. All participants lived in the Providence metropolitan area and were examined between October 1997 and May 1998. Information regarding prior vitamin supplement use was obtained by standardized interview, and subjects were either nonusers of any supplements containing folic acid, or had abstained from using such supplements for at least 6 wk by the time of their examination.
Overnight (10 to 14 h) fasting blood samples were collected from each participant. tHcy levels were determined by HPLC with fluorescence detection, and plasma pyridoxal 5′-phosphate levels were measured by radioenzymatic (tyrosine decarboxylase) assay, as reported earlier (1). Plasma folate and vitamin B12 levels were measured by radioassay (BioRad Quantaphase II, Hercules, CA). Serum creatinine (Jaffe method) and albumin (bromcresol method) levels were determined using standard techniques adapted for automated clinical chemistry laboratory analyzers. Serum cystatin C levels were determined by particle-enhanced immunoturbidimetry (11).
Statistical Analyses
All skewed continuous variables were (natural log) transformed to better approximate a normal distribution, and unadjusted correlations between continuous variables were assessed in a Pearson correlation matrix. General linear modeling with analysis of covariance was then performed to determine the independent association between potential predictor covariables and fasting tHcy levels. Immunosuppressive drug regimen data were not included in these models because there is no evidence that the agents currently utilized independently influence tHcy levels, after appropriate adjustment for renal function indices (1,5), or other established determinants of tHcy levels (5). Reported P values were based on two-tailed calculations, and all statistical analyses were performed using SYSTAT (version 7.0.1) software.
Results
Key subject characteristics expressed as means, geometric means, percentages, and complete ranges are depicted in Table 1. There was a trend for unadjusted geometric mean fasting tHcy levels to be greater in the men (n = 17) than in the women (n = 11) (12.1 versus 10.3 μM; P = 0.076). Analysis of covariance, adjusting for age, vitamin status, creatinine, and cystatin C, confirmed the independence of this gender difference (men = 12.4 μM, women = 9.8 μM, P = 0.017). Unadjusted Pearson correlations between the continuous variables examined are illustrated in Table 2. Creatinine was modestly correlated with tHcy (r = 0.214), as well as most of the usual determinants of tHcy levels, including pyridoxal 5′-phosphate, folate, age, and albumin (r ≥ 0.220). Although also correlated with tHcy (r = 0.306), cystatin C, in contrast to creatinine, was not correlated with either age, albumin, or B12 (r ≤ 0.044), and only weakly correlated with folate (r = 0.177) and pyridoxal 5′-phosphate (r = 0.150). General linear modeling with analysis of covariance (i.e., stepwise and “forced” forward or backward), revealed that only cystatin C (partial R = 0.494; P = 0.023) and gender (partial R = 0.486; P = 0.026) were independent predictors of fasting tHcy levels, after simultaneous adjustment for age, creatinine, albumin, folate, pyridoxal 5′-phosphate, and B12. Creatinine levels were not independently predictive of fasting tHcy levels in any of these models (partial R = 0.359; P = 0.110), or identical models that differed only by the exclusion of cystatin C levels (partial R = 0.332; P = 0.131).
Subject characteristicsa
Pearson correlation matrix of natural log-transformed data
Discussion
Previous reports (9,10,11,12), including a recent study of renal transplant recipients (12), have revealed that cystatin C is a more sensitive marker of mildly impaired GFR compared with creatinine, and that GFR itself, even into the normative range, is a powerful, independent predictor of fasting tHcy levels (13). Expanding on these combined observations, we report the first analysis indicatint that serum cystatin C levels may independently predict fasting tHcy levels among stable renal transplant recipients with normal serum creatinine levels. Clearly, larger studies will be required to confirm the external validity of the present findings.
At present, it is unknown whether the persistent mild hyperhomocysteinemia characteristic of chronic renal insufficiency and end-stage renal disease (16) results from the loss of intrarenal homocysteine metabolism (17) or uremia-induced extrarenal defects in homocysteine metabolism (18). The findings reported here from nonuremic subjects might suggest that intrarenal homocysteine metabolism is a major determinant of homocysteine levels. However, these data cannot rule out the possibility that subtle extrarenal defects in homocysteine metabolism that may accompany even such mild reductions in renal function could account for the resulting increases in tHcy levels.
Studies using serum cystatin C determinations to assess subclinical deterioration in renal function and homocysteinemia within other patient populations, e.g., patients with arteriosclerotic cardiovascular disease, may be warranted. An initial analysis (19) using cystatin C as a surrogate for GFR in healthy subjects has already indicated that the well characterized increase in plasma tHcy levels with advancing age may be due in part to an agerelated decline in renal function. Finally, given that clinical renal disease renders individuals refractory to the effect of low-dose folic acid supplementation on fasting tHcy levels (16), cystatin C data might also be useful in evaluating the influence of subclinical renal impairment on tHcy-lowering treatment responsiveness.
In summary, serum cystatin C levels may reflect subtle decreases in renal function that independently predict fasting tHcy levels among stable renal transplant recipients with normal serum creatinine levels.
Acknowledgments
Acknowledgments
This work was supported in part by a grant from Roche Vitamin Division. We thank Susan Ritter, Evelyn Tolbert, Marie Nadeau, and Bonnie Soupa, for their excellent technical assistance.
Footnotes
American Society of Nephrology
- © 1999 American Society of Nephrology