End-stage renal disease has been associated with elevations in plasma Lp(a) levels. Analysis of patients treated with long-term hemodialysis revealed that Lp(a) levels were approximately three times higher than those of matched healthy controls [8, 9]. However, studies of Lp(a) concentration in patients with end-stage renal disease that control for apo(a) isoform have not been reported. Patients with proteinuria have also been shown to have increased Lp(a) levels [10, 11]; however, the proteinuria was relatively low grade and the patients were not classically nephrotic.

Therefore, we undertook a comprehensive evaluation of Lp(a) concentrations in a series of 62 patients with classic nephrotic syndrome, controlling for apo(a) isoform, and compared the results with data obtained in a healthy control group.

Methods

Top Methods Results Discussion Author & Article Info References

A total of 62 patients (26 women, 36 men; mean age, 47.2 ± 2.2; age range, 18 to 76 years) with the nephrotic syndrome were studied. The patients were consecutively entered at one clinic (University Clinic, Freiburg, Germany) from 1990 to 1992. All patients had severe proteinuria, low serum albumin, hyperlipoproteinemia, and edema. Patients were divided into two subgroups: those with primary kidney disease (n = 47) and those with diabetic nephropathy (n = 15). The histologic findings of patients with primary kidney disease as determined by renal biopsy were membranous glomerulonephritis (n = 22), focal and segmental glomerulosclerosis (n = 13), minimal change disease (n = 7), mesangiocapillary glomerulonephritis (n = 1), and immunotactoid glomerulonephritis (n = 1). In three patients with primary glomerulonephritis tissue, a diagnosis could not be obtained. Patients with diabetic glomerulopathy were diagnosed as having diabetic nephropathy according to the duration of underlying type I (n = 4) or type II (n = 11) diabetes mellitus and the presence of retinopathy. Patients with diabetic nephropathy were all being treated with pharmacologic therapy with an angiotensin-converting inhibitor at the time they were sampled. Ninety-one healthy members of the staff of the National Institutes of Health acted as controls and were studied in 1991.

Thirteen patients with primary kidney disease were resampled 6.7 ± 0.7 months after remission of the nephrotic syndrome was induced by immunosuppressive therapy. Four patients had normal Lp(a) levels, and nine exhibited high levels (> 30 mg/dL) before remission. A second follow-up was performed in six of the nine patients with high Lp(a) 13.1 ± 0.5 months after they achieved remission. Drugs used for induction of remission were cyclosporine/prednisone (n = 4), chlorambucil/prednisone (n = 6), prednisone monotherapy (n = 1), and acetylsalicylic acid/prednisone (n = 1). One patient had a spontaneous remission. Blood sampling was done when patients were free of steroids and chlorambucil, but four patients still were being treated with cyclosporine monotherapy. Diabetic patients were followed for a total period of 19 ± 2.3 months. Ten patients developed end-stage renal disease. Because two patients died of myocardial infarction and one had no detectable Lp(a) levels, serum was analyzed in seven patients 9.8 ± 0.9 months after initiation of hemodialysis (n = 5) or peritoneal dialysis (n = 2). The remaining five patients are still nephrotic.

Blood samples were collected after an overnight fast of at least 12 hours. All chemistry, lipid, and apolipoprotein measurements were performed on fresh samples. At least two 24-hour urine specimens were collected and subjected to gradient sodium dodecyl sulfate-polyacrylamide gel electrophoresis (8% to 25%) (Phast System, Pharmacia; Uppsala, Sweden) to determine selectivity of protein excretion.

Lipid and Apolipoprotein Analysis

Very-low-density lipoprotein was isolated by ultracentrifugation [12]. Cholesterol and triglycerides were determined enzymatically (Boehringer-Mannheim, Mannheim, Germany). Apolipoprotein B (apoB) concentrations were determined by rate nephelometry on the "Array" nephelometer (Beckman; Palo Alto, California). Lyophilized control serum was used as a calibrator as previously described [13].

Lipoprotein(a) concentrations in the nephrotic patients were determined by two independent methods. A commercial two-site immunoradiometric assay (Pharmacia) was performed on fresh samples from the nephrotic patients. Lipoprotein(a) was remeasured in the same samples stored at –20°C for less than 1 year by a differential enzyme-linked immunosorbent assay (ELISA) based on the method of Fless and colleagues [14]. A monoclonal antibody against apo(a) (2-D1, Cappel; Durham, North Carolina) was used to coat microtiter plates at a concentration of 10 µg/mL. This antibody recognizes all sizes of apo(a) (on immunoblotting) and does not cross-react with plasminogen. After blocking with 2% bovine serum albumin, plasma samples at a 1:5000 dilution were added to wells and incubated for 60 minutes at 37 °C. A sheep, polyclonal anti-apoB (Biodesign; Kennebunkport, Maine) labeled with horseradish peroxidase was added to the wells at a 1:2000 dilution and incubated for 60 minutes. Substrate was then added and absorbance read at 450 nm. The standard was a secondary plasma standard calibrated against two commercial standards (Terumo; Elkton, Maryland; Immuno; Vienna, Austria). Two controls were run with each assay. Intra-assay and interassay coefficents of variation were less than 3% and less than 10%, respectively. Between the two assays (ELISA and radioimmunoassay), a good correlation (r = 0.884) was found. Lipoprotein(a) concentrations in control subjects were measured using only the differential ELISA. The values reported here are those from the ELISA.

Apoprotein(a) Isotyping

Apoprotein(a) isoform determination was done on plasma using a modification of an immunoblotting technique, as previously described [15]. Briefly, plasma samples were delipidated twice in chloroform-methanol 8:5 (vol/vol) and washed twice with phosphate-buffered saline. The samples were reduced with 100 mmol/L dithiothreitol in 8 mol/L urea and incubated at 37 °C for 30 minutes. The samples were solubilized in a solution containing 75% glycerol, 0.5% bromophenol blue, and 10% sodium dodecyl sulfate and applied to 7.5% polyacrylamide gel electrophoresis with 0.1% crosslinker. Gels were run for approximately 2.5 hours at 60 mA. Subsequent immunoblotting was carried out as described previously [16]. After transfer of the proteins to Immobilon polyvinylidene difluoride transfer membrane (Millipore; Bedford, Massachusetts), incubation of the membranes with antibodies was performed using an anti-apo(a) monoclonal antibody (2-D1, 1:2000; Cappel) as first antibody and the Vectastain ABC anti-mouse test kit (Vector Laboratories; Burlingame, California) for detection. Several plasma samples of a defined apo(a) isoform were used as standards on each gel. One nephrotic patient had no detectable apo(a) band, 30 patients had only one apo(a) band, and 29 patients had two apo(a) bands. In both patients and controls with two apo(a) bands, one band was usually of substantially greater intensity. For purposes of statistical analysis, these patients and controls were assigned a single apo(a) isotype based on the apo(a) band with the greater immunostaining. This dominant band accounts for the majority of Lp(a) in plasma.

Statistical Analysis

Data are given as mean ±SE. Data analysis was performed using the Statistical Analysis System software package (SAS Institute; Cary, North Carolina) using the Wilcoxon rank-sum tests or the Student t-test for group comparisons of continuous variables with non-normal or normal population distributions, respectively. Point estimates and confidence interval estimates were calculated by nonparametric procedures [17, 18]. P values less than 0.05 were considered to be statistically significant.

Results

Top Methods Results Discussion Author & Article Info References

The clinical and laboratory characteristics of the nephrotic patients are shown in Table 1. Patients with primary kidney disease were younger, had lower body mass indices, and had lower serum creatinine and serum albumin values than did patients with diabetic nephropathy (P < 0.05 for all values). The lipid and apolipoprotein data for the nephrotic patients are shown in Table 2 and compared with those of the control sample (n = 91). Total cholesterol, low-density lipoprotein cholesterol, triglyceride, and apoB concentrations were elevated in both groups of patients compared with healthy controls (P < 0.05). Lipoprotein(a) levels were also elevated in the nephrotic patients compared with the control group (P < 0.01). Sixty percent of the nephrotic patients had Lp(a) levels greater than 30 mg/dL, whereas only 18% of the controls had levels in this range. The percentage of elevated Lp(a) levels were identical: >30 mg/dL among patients with primary kidney disease (60%; n = 28) and among patients with diabetic nephropathy (60%; n = 9). No significant correlation between renal function (r = –0.12,P > 0.2), serum albumin, or 24-hour protein excretion and Lp(a) levels could be detected (Table 3). The histologic findings of renal disease also did not correlate with Lp(a) levels. However, apoB, total cholesterol, and low-density lipoprotein cholesterol were positively correlated with Lp(a) levels (see Table 3).

Because apo(a) phenotype has a substantial effect on Lp(a) levels, the nephrotic patients were separated into groups based on apo(a) phenotype. Lipoprotein(a) levels within each apo(a) isoform class were compared with those of the control population (Table 4). Within S1, S2, S3, S4, and S5 apo(a) isoform classes, the Lp(a) levels in the nephrotic patients were significantly greater than those of the controls.

Figure 1 is a flow chart of the 62 patients with the nephrotic syndrome and their follow-up data. Thirteen patients with primary kidney disease were resampled 6.7 ± 0.7 months after remission of the nephrotic syndrome was induced by immunosuppressive therapy. Four patients had normal Lp(a) levels and nine exhibited high levels (> 30 mg/dL) before remission. Induction of remission was accompanied by a substantial decrease of the Lp(a) level in all nine patients, with a pretreatment mean Lp(a) of 90 ± 15 mg/dL decreasing to a remission mean Lp(a) of 31 ± 8 mg/dL (P < 0.001; (Figure 2). Remission also decreased serum cholesterol [527 ± 68 versus 224 ± 15 mg/dL], low-density lipoprotein cholesterol (376 ± 43 versus 147 ± 11 mg/dL), apoB (240 ± 25 versus 104 ± 6 mg/dL), and daily urinary protein excretion (13.0 ± 1.6 versus 0.67 ± 0.2 g), whereas serum albumin increased (21.8 ± 1.9 to 40.4 ± 1.0 g/L). A good correlation was found between Lp(a) levels and daily protein excretion (r = 0.78, P < 0.01) after remission was achieved. However, apoB, serum cholesterol, low-density lipoprotein cholesterol, and serum albumin did not correlate (see Table 3). A second follow-up was performed in six of the nine patients with high Lp[a] 13.1 ± 0.5 months after they achieved remission. Lipoprotein(a) levels remained decreased at the second follow-up (32 ± 7 mg/dL). In the four patients with normal Lp(a) levels before immunosuppressive treatment, induction of remission had no effect on Lp(a) levels (11.3 ± 5.3 versus 12.1 ± 5.7 mg/dL, mean ± SE). Five of nine patients were not immunosuppressed, and protein excretion did not exceed the normal range. The other four patients were receiving cyclosporine monotherapy, three of them with a proteinuria level ranging from 0.8 to 2.2 g/day. Thirty-four patients with primary kidney disease did not achieve complete remission. Two patients died of cerebral hemorrhage. The remaining 32 patients were treated with immunosuppressive protocols (n =14) or conservatively treated with angiotensin-converting inhibitors and diuretics (n = 12). Of these patients, 10 still have the nephrotic syndrome, 5 exhibit nephrotic-range proteinuria (>3 g/d), 8 have moderate proteinuria, 4 have developed end-stage renal disease, and 5 have been lost to follow-up. The 10 patients with unremitted nephrotic syndrome were resampled 20 ± 3.3 months after the first examination and evaluation of lipoprotein status. At this follow-up, Lp(a) values were found to be similar (72.0 ± 19.2 versus 70.7 ± 14.0 mg/dL; (Figure 3).

View larger version (19K): [in this window] [in a new window]   Figure 1. Flow chart of the 62 patients with the nephrotic syndrome. *Lipoprotein(a) determinations made. High Lp(a) at baseline was defined as a value greater than 30 mg/dL. Low Lp(a) at baseline includes the patient with an 0-allele.

View larger version (17K): [in this window] [in a new window]   Figure 2. Serum lipoprotein(a) values in patients with primary kidney disease (n = 9) before and after induction of remission of the nephrotic syndrome. Five patients with primary kidney disease were receiving no immunosuppressive therapy at time of resampling, and four were receiving monotherapy with cyclosporine but no steroids. Mean values ± SE are given by the vertical bars. **P = 0.001.

View larger version (16K): [in this window] [in a new window]   Figure 3. Serum lipoprotein(a) values in a group of patients (n = 10) with unremitted nephrotic syndrome. At the time of follow-up (20 ± 3.3 months after the initial measurement), patients were without immunosuppressive therapy. Mean values ± SE are shown by the vertical bars.

Diabetic patients were followed for 19.1 ± 2.3 months. Ten patients developed end-stage renal disease. Two patients died of myocardial infarction and one had no detectable Lp(a) levels. Serum was analyzed in seven patients 9.8 ± 0.9 months after initiation of hemodialysis (n = 5) or peritoneal dialysis (n = 2). At the time of resampling, the residual urinary excretion of the patients was 1.1 ± 0.3 g/d (range, 0.6 to 1.8 g). Lipoprotein(a) levels decreased in all seven diabetic patients, with a pretreatment mean Lp(a) of 56 ± 11 mg/dL (range, 24 to 87 mg/dL) decreasing to a mean Lp(a) of 34 ± 10 mg/dL (range, 3 to 73 mg/dL), (P< 0.05, Figure 4. In these patients serum cholesterol [312 ± 32 versus 236 ± 15 mg/dL], low-density lipoprotein cholesterol (224 ± 24 versus 111 ± 19 mg/dL), apoB (160 ± 13 versus 101 ± 15 mg/dL), and daily urinary protein excretion (8.1 ± 1.9 versus 0.64 ± 0.1 g) decreased, whereas serum albumin increased (32 ± 1.9 to 40.5 ± 0.8 g/L). No correlation could be found between Lp(a) levels and the other variables (see Table 3). The remaining five diabetic patients are still nephrotic.

View larger version (15K): [in this window] [in a new window]   Figure 4. Serum lipoprotein(a) values in patients with diabetic nephropathy (n = 7) and after development of end-stage renal disease treated by dialysis. Mean values ± SE are shown by the vertical bars. *P < 0.05.

Discussion

Top Methods Results Discussion Author & Article Info References

The factors that modulate plasma Lp(a) concentrations are poorly understood. Variation in apo(a) phenotype accounts for approximately 40% of the variation in Lp(a) levels [19] by an unknown mechanism. Other genetic factors also strongly influence Lp(a) levels [1, 20]. The rate of Lp(a) production appears to be a major determinant of plasma Lp(a) levels in the healthy population [21, 22]. Recent kinetic studies in persons with the same apo(a) phenotype but widely different Lp(a) levels indicated that the Lp(a) production rate was the major determinant of differences in Lp(a) levels [7]. Other conditions that have been proposed to influence Lp(a) concentrations include familial hypercholesterolemia [23] and end-stage renal disease [8, 9].

Our study conclusively demonstrates that the nephrotic syndrome should be added to the list of conditions that modulate Lp(a) levels. Patients with classic nephrotic syndrome had substantially elevated plasma Lp(a) concentrations when compared with a control group controlled for apo(a) isoform. The observation that induction of remission in a subgroup of these patients resulted in a dramatic decrease in Lp(a) concentrations, further strengthens the conclusion that the nephrotic syndrome directly affects Lp(a). Furthermore, patients with diabetic nephropathy who were being treated by dialysis in the follow-up study also exhibited a decrease in Lp(a) concentrations. It can be argued that immunosuppressive drugs may acutely reduce Lp(a) levels independent of nephrosis remission. Recent observations in 20 patients, studied prospectively for 6 months after successful renal transplantation, demonstrated a marked decrease in apo(a) levels [24]. No correlation, however, was found between immunosuppressant dose and apo(a) concentration. Another report observed higher Lp(a) levels in a group of patients who had renal transplants and were taking cyclosporine when compared to a group of patients taking azathioprine [25].

Hyperlipidemia is a consistent feature of the nephrotic syndrome. Investigators have attempted to establish a causal relationship between proteinuria and altered lipoprotein metabolism [26, 27]. It has been proposed that hypoalbuminemia causes reduced serum oncotic pressure, which in turn stimulates hepatic synthesis of albumin and other liver-derived proteins, including apolipoproteins [28]. Most or all of plasma apo(a) is derived from the liver in humans [29]. If the nephrotic syndrome causes increased hepatic protein synthesis, then the elevated Lp(a) levels in the nephrotic syndrome may result from overproduction by the liver. Alternatively, the kidney may play a role in the catabolism of Lp(a), and therefore the nephrotic syndrome could affect plasma Lp(a) levels by affecting the rate of catabolism. This could be the mechanism by which end-stage renal disease results in increased levels of Lp(a) [8]. In-vivo studies will be required to investigate the mechanism by which the nephrotic syndrome and end-stage renal disease influence Lp(a) metabolism.

Several of our patients had the nephrotic syndrome due to diabetes mellitus. It has been suggested that diabetic patients have higher Lp(a) levels than do healthy controls [30], although one report found that only the subgroup of diabetics with microalbuminuria had elevated Lp(a) levels [31], and a recent study reported that improved glycemic control did not lower Lp(a) levels in patients with type II diabetes [32]. None of the studies in diabetic patients has controlled for apo(a) isoform. The association between diabetes and Lp(a) therefore remains uncertain. In this study, the Lp(a) levels did not differ significantly between the diabetic group and the nondiabetic group; in fact, the mean Lp(a) level in the diabetic group was slightly lower than that in the nondiabetic group. Therefore, the inclusion of patients with diabetic nephropathy in this report has not biased our conclusion that the nephrotic syndrome causes increased levels of Lp(a).

Recent studies have shown that lipoproteins may play a role in the progression of renal disease [33]. In focal and segmental glomerulosclerosis, microthrombosis with occlusion of glomerular capillaries, as well as renal foam cell formation, is common. Because of the striking homology of apo(a) with plasminogen [34], it has been suggested that Lp(a) may interfere with the fibrinolytic process [35]. It has also been suggested that Lp(a) may promote foam cell formation [3]. Therefore, Lp(a) may contribute to the microthrombosis and lipid deposition in the kidney seen in some forms of renal disease [36].

Relatively little has been published on the risk for atherosclerotic vascular disease in the nephrotic syndrome, and what exists is poorly controlled and conflicting in conclusions [37-40]. If the risk for premature cardiovascular disease is increased, this may be due in part to the elevated levels of Lp(a) in the nephrotic syndrome. The increase in blood coagulability and thrombotic events in the nephrotic syndrome is well documented [41], and although this has been attributed to relative antithrombin III deficiency, it is possible that the elevated levels of Lp(a) may also play a role [42].

In summary, the nephrotic syndrome results in substantial elevation of plasma Lp(a) concentrations by an as yet unknown physiologic mechanism. However, because not all nephrotic patients have high Lp(a) levels, genetic factors may play an important role in influencing Lp(a) levels even in the nephrotic population. It may be of clinical value to quantitate Lp(a) in the nephrotic syndrome because the subgroup of patients with high levels may be at increased risk for accelerated progression of renal disease, atherosclerotic vascular disease, and thrombotic events.