Posttransplant Acidosis and Associated Disorders of Mineral Metabolism in Patients With a Renal Graft

Abstract

Background. 

Persisting disturbances in acid/base homeostasis may have an impact on several metabolic aspects of individuals with a kidney graft, specifically with regard to mineral metabolism and bone.

Methods. 

We undertook a cross-sectional analysis among 823 unselected patients being transplanted with a functioning renal allograft who had at least one measurement of venous serum bicarbonate available within a 4-year period before May 1, 2005. As a determinate of metabolic acidosis bicarbonate was measured along with serum calcium, phosphate, parathyroid hormone, and other routine serological and epidemiological parameters. Data were assessed according to quartiles of serum bicarbonate and by univariate analysis. A multivariate regression model examined the effects of potential predictors of acidosis.

Results. 

Mean serum bicarbonate was 22.5±4 mmol/L, with 58.1% of the examined renal transplant patients having metabolic acidosis as defined by a venous bicarbonate of <24 mmol/L. Bicarbonatemia was highly associated with serum parathyroid hormone, phosphate, and calcium but also with renal graft function (determined as calculated glomerular filtration rate). Multiple stepwise regression analysis revealed age, glomerular filtration rate, parathyroid hormone, and albumin to be the strongest predictors of serum bicarbonate concentration. Therapy with any calcineurin inhibitor was not associated with an increased likelihood of acidosis (odds ratio 1.04), but a significant difference was found between cyclosporine A and tacrolimus, which had an attributed odds ratio for acidosis of 0.6 and 1.8, respectively.

Conclusions. 

Metabolic acidosis is highly prevalent among an unselected cohort of renal transplant patients. A clear association exists between the severity of acidosis and disturbances of mineral metabolism. Thus, persisting acid/base disorders may accentuate bone disease in a setting with other factors predisposing for posttransplant osteopathy.

Successful renal transplantation can restore glomerular filtration rate (GFR) to some degree. However, other functions of the transplanted kidney, such as renal acid/base handling, have not been given much consideration so far and may not be normal with an otherwise-functioning graft. Specifically in the early course after transplantation subclinical metabolic acidosis (low serum bicarbonate with or without acidic serum pH) has been documented (¹). Renal tubular acidosis has been reported since the beginning of kidney transplantation (^2–6) up to present days (⁷). However, no systematic assessment with regard to the prevalence of acidosis has been performed and published in a large unselected cohort of contemporary renal transplant patients. Prolonged persistence of acidosis may negatively impact on many metabolic aspects, mainly the regulation of mineral metabolism and bone (⁸). Therefore, we conducted a cross-sectional analysis in 823 patients transplanted for various periods of time, in order to determine the prevalence, degree and associated factors of metabolic acidosis.

METHODS

To assess the prevalence, severity, and associated disorders of posttransplant metabolic acidosis, a cross-sectional analysis of all 983 patients with a functioning renal graft from our institution was performed in June 2005. Their medical records were screened, and 823 patients (84%) were identified who had at least one measurement of venous serum bicarbonate available within a 4-year period before May 1 2005. These patients were further analyzed regarding clinical and additional laboratory parameters, with special emphasis on renal graft function and determinants of mineral metabolism, such as serum calcium, phosphate, and parathyroid hormone (PTH). GFR was calculated according to the Cockcroft-Gault formula and using lean body mass. Total serum calcium and phosphate measurements were available in 818 (99%) and 819 (99%) patients, respectively. Calcium was normalized for serum albumin concentration. Intact PTH was determined by radioimmunoassay (RIA; Nichols Institute Diagnostics, San Juan Capistrano, CA) and was available in only 133 patients (16%) coincident with serum bicarbonate measurements, and in an additional 291 patients (35%) within 9.3 months around the date of bicarbonate determination. Because both PTH and bicarbonate were not statistically different between these two groups with regard to their respective average (PTH: 141±178 vs. 142±173 ng/L, P=0.944; bicarbonate: 22.5 vs. 22.4, P=0.68) and their correlation coefficient (PTH/bicarbonate: 6.5 vs. 7.1, P=0.581), the results were combined.

Statistical Analysis

Data are presented as mean±standard deviation. Differences between groups were calculated using analysis of variance (ANOVA) for parametric data. Univariate and multivariate analyses were performed to assess correlation between variables. A multivariate regression model was created to control for the effects of other predictors of acidosis. Variables used in the model were those with P<0.05 identified in the univariate analysis. Using a backward elimination procedure, the least significant variables were dropped from the model stepwise until all variables remaining in the model had a P<0.05. Statistical significance was defined as P<0.05. Statistical calculations were performed using SPSS for Windows software version 12.0.1 (SPSS Inc., Chicago, IL).

RESULTS

Table 1 shows the results for serum bicarbonate, parameters of renal function and mineral metabolism along with epidemiological data for the entire cohort as well as stratified in quartiles of serum bicarbonate concentration. Almost 60% of the patients were acidotic based on a venous bicarbonate of less than 24 mmol/L. The degree of acidosis was more pronounced in individuals with lower GFR, but was unrelated to age, time since transplantation, or body mass index. Lower bicarbonate was associated with higher serum chloride and potassium concentration. Conversely, normal bicarbonate was related to greater serum albumin levels. Finally, more severe acidosis was accompanied with lower total serum calcium but greater phosphorous and PTH concentrations in our transplant cohort. Univariate analysis revealed a significant correlation of serum bicarbonate with GFR and serum albumin and an inverse correlation with serum phosphate and PTH (Table 2).

TABLE 1: Epidemiological and biological characteristics of the study population and subgroups according to quartiles of serum bicarbonate concentration

TABLE 2: Univariate correlation analysis

Disturbances in mineral metabolism often persist despite successful kidney transplantation and are considered to be the consequence of reduced nephron mass resulting in lower than normal GFR. However, the natural course of posttransplant hyperparathyroidism and its associated disorders is not as well documented and understood compared with chronic insufficiency of native kidneys and during dialysis therapy. Particularly, the relationship between posttransplant acidosis and disorders of mineral metabolism has not been given much consideration. Only 27% of our renal transplant population had a PTH within normal range (i.e., 15–65 ng/L), whereas 70% presented with hyperparathyroidism (PTH>65 ng/L), and 2.8% were hypoparathyroid (PTH<15 ng/L). Even if applying the criteria for patients with native chronic kidney disease, the mean PTH values determined in our transplant cohort by far exceed the target ranges defined by the National Kidney Foundation Kidney Disease Outcome Quality Initiative (NKF-K/DOQI) guidelines (⁹). PTH was clearly increased in patients with reduced GFR (Table 2) but independent of time since transplantation (data not shown). A high level of PTH was inversely correlated with serum calcium but even stronger with serum bicarbonate (Table 2). This interrelation between PTH and bicarbonate is particularly obvious among individuals with impaired renal function (GFR<52 mL/min; Fig. 1A). Similarly, a stronger correlation between PTH and GFR is apparent in patients with acidosis (bicarbonate <24 mmol/L; Figure 1B). Serum phosphate was within normal range in 74% of the patients (0.85–1.45 mmol/L) and increased in only 3.6%.

FIGURE 1.: Univariate correlation analyses of serum bicarbonate and renal graft function (GFR) with serum PTH, phosphate and calcium. (A) The concentration of serum PTH (expressed as logarithm) is shown as dependent variable of venous serum bicarbonate concentration both for individuals with a GFR less than (•) and greater than 52 ml/min (□). (B) The concentration of serum PTH (expressed as logarithm) is shown as a dependent variable of GFR both for individuals with a venous serum bicarbonate concentration less than (•) and greater than 24 mmol/L (□). (C) Serum phosphate concentration is shown as a dependent variable of GFR both for individuals with a venous serum bicarbonate concentration less than (•) and greater than 24 mmol/L (□). (D) Serum calcium concentration is shown as a dependent variable of glomerular filtration rate (GFR) both for individuals with a venous serum bicarbonate concentration less than (•) and greater than 24 mmol/L (□).

Interestingly, 21.6% of the individuals of our renal transplant cohort were hypophosphatemic (<0.8 mmol/L). Hypophosphatemia was significantly associated with better graft function (GFR 57±17 vs. 52±16 ml/min in hypo- and normophosphatemic patients, respectively; P=0.001) but not with PTH (142±15 vs. 128±10 ng/L; P=0.475). As for PTH, serum phosphate was negatively and significantly correlated with both bicarbonate concentration and GFR. Both hyper- and hypophosphatemia were accentuated by acidosis (bicarbonate<24mmol/L) independent of transplant function (Fig. 1C). Finally, serum calcium was within normal range in most subjects (85.9%), with only 2.8 and 11.3% being hypo- and hypercalcemic, respectively. Hypocalcemia (<2.10 mmol/L) was significantly associated with lower GFR (46±16 vs. 53±16 mL/min in hypo- and normo-/hyperphosphatemic patients, respectively) and lower bicarbonate (20.4±4 vs. 22.6±4 mmol/L; P=0.017) but not with PTH and phosphate. Low serum calcium was more pronounced in acidotic patients, regardless of renal function (Fig. 1D). Accordingly, hypercalcemia (>2.60 mmol/L) was significantly and positively associated with bicarbonate (23.4±4 vs. 22.4±4 mmol/L in hyper- vs. normo-/hypophosphatemic patients, respectively, P=0.035), but not with GFR, PTH, and phosphate.

Multivariate logistic regression was performed using the significant continuous variables identified by univariate regression in Table 2. Age at time of examination, time posttransplant, GFR, albumin, and PTH were identified as the variables which correlated best with serum bicarbonate (not shown). Using these variables for stepwise inclusion into multivariate analysis models revealed GFR, age at time of examination, albumin and PTH to contribute individually and significantly to the prediction of serum bicarbonate concentration in renal transplant patients (Table 3).

TABLE 3: Multiple stepwise linear regression analysis of variables associated with serum bicarbonate

Immunosuppressive Medication

Several immunosuppressive drugs have been implicated into the pathogenesis of posttransplant acidosis, mainly calcineurin inhibitors (CNIs) such as cyclosporine A (CyA) and tacrolimus. Most of the 823 analyzed patients were on a CNI (CyA, n=646 [78%]; tacrolimus, n=135 [16%]). Interestingly, as shown in Table 4, CNI therapy was not associated with an increased likelihood of acidosis (odds ratio [OR] 1.04). However, a substantial difference was found between CyA and tacrolimus, which had an attributed OR for acidosis of 0.6 and 1.8, respectively. Similarly, the odds to be acidotic were clearly lower when treated with azathioprine compared with mycophenolate mofetil (MMF). However, the higher likelihood of acidosis from tacrolimus versus CyA cannot be explained by the somewhat greater proportion of patients on MMF in the tacrolimus versus the CyA group. Similarly, the higher bicarbonate from CyA±MMF versus tacrolimus±MMF is independent of GFR. In contrast, the finding of a slightly and significantly lower bicarbonate on a steroid versus a steroid free regimen (22.1±4 vs. 22.9±3 mmol/L, P=0.002) may be confounded by the clearly lower GFR in the former group (49±17 vs. 56±15 ml/min, P=0.000).

TABLE 4: Odds ratio for acidosis (serum bicarbonate <24 mM/L) depending on immunosuppressive medication

DISCUSSION

To the best of our knowledge, this study is the first to determine the prevalence of metabolic acidosis in a large renal transplant cohort. Three major findings result from our analysis. First, more than half of the patients after kidney transplantation presented with a venous serum bicarbonate <24 mmol/L, which is consistent with metabolic acidosis. Second, impaired renal function and age are the main predictors of metabolic acidosis. Third, low bicarbonate concentration is independently associated with disturbances in posttransplant mineral metabolism, such as hyperparathyroidism and hyperphosphatemia.

Several mechanisms may account for metabolic acidosis after kidney transplantation. Reduced nephron mass results in general impairment of renal tubular processes, contributing to secretion of acid as well as the reclamation and regeneration of bicarbonate and base equivalents. Alternatively, each process may be disturbed selectively and individually, and in relation to distinguishable factors. In this regard, immunosuppressive drugs may play an important role, and have been described to result in renal tubular dysfunction (¹⁰), i.e., renal tubular acidosis (RTA). In a contemporary analysis, the whole spectrum of RTA has been described in renal transplant patients, with classical distal RTA (type Ia) being the predominant form (⁷). As transplantation specific factors, allo-immune mechanisms also could be operational. Finally, metabolic acidosis might ensue from a defect in proximal tubule ammonia synthesis as the result of insulin resistance. Insulin induces proximal tubular ammonia synthesis, but may not be effective because immunosuppressive drugs, such as corticosteroids (¹¹) and calcineurin-inhibitors (¹²), may confer insulin resistance in renal transplant patients. For a comprehensive review on the pathogenesis of posttransplant acidosis refer to (¹³).

In accordance with the postulated mechanisms, we found serum bicarbonate to be strongly correlated with GFR. In the multivariate model, including age, albumin, and PTH, each milliliter decrease in GFR results in an approximately 0.1 mmol reduction in serum bicarbonate concentration. The majority of the analyzed patients was on a calcineurin inhibitor (CyA, 646 [78%]; tacrolimus, 135 [16%]). However, only tacrolimus was found to be associated with a greater likelihood of acidosis, a finding that can not be explained by renal function or co-medication with MMF. Patients on MMF had a clearly lower bicarbonate and were more likely to be acidotic compared to those treated with azathioprine. Interestingly, albumin was found to be a relevant predictor for posttransplant acidosis in our analysis. This may be explained by the well established inhibitory effect of metabolic acidosis on albumin synthesis (^{14, 15}).

Patients with end-stage renal disease have a broad spectrum of bone disorders (¹⁶) and osteopenia, which may worsen after renal transplantation secondary to immunosuppression with corticosteroids (¹⁷) or calcineurin inhibitors, such as cyclosporine A (^{18, 19}). Metabolic acidosis could be an additional contributor in the pathogenesis of posttransplant bone disease. A decrease in environmental pH can directly increase the resorptive activity of osteoclasts in vitro (^20–22). Acidotic rats have increased bone resorption versus control animals (²³). Whether metabolic acidosis in humans results in bone resolution due to proton buffering is being discussed controversially for decades and remains unresolved (^{8, 24, 25}). Classical studies in humans demonstrate the powerful effect of chronic metabolic acidosis, either induced experimentally or resulting from chronic kidney disease, on the loss of bone minerals (^{24, 26–28}). On a hormonal level, chronic metabolic acidosis reduces renal synthesis of 1,25(OH)₂D₃ (²⁹) and alters the homeostatic relationships between blood ionized calcium, PTH, and 1,25(OH)₂D₃ such that bone dissolution is exaggerated (^{8, 30}). In accordance, recent studies using dynamic histomorphometry have demonstrated a reduction in bone mineral density and in bone formation rates (^{31, 32}). Not surprisingly, bone fractures are a relatively common manifestation of chronic metabolic acidosis and have been shown to develop without concomitant renal insufficiency (³³). Comparable studies on the interrelationship between acidosis and bone disorders in patients after kidney transplantation are lacking.

In the present study, we found a clear association between serum bicarbonate concentration and parameters of mineral metabolism. In our transplant cohort low bicarbonate was related with higher serum phosphate and PTH concentrations, and lower serum calcium (Table 1 and 2). These alterations do not necessarily cause bone disorders, however, they clearly represent disturbances in the regulation of mineral metabolism that are very likely to promote and/or to reflect osteopathy. Several mechanisms may be operative to explain these findings. Acidosis induces calcium excretion (³⁴) and reduces vitamin D synthesis (²⁹) which, in turn stimulate PTH release. Accordingly, correction of metabolic acidosis has been shown to reduce PTH levels (³⁵), probably through increased sensitivity of the parathyroid glands to calcium (³⁶). On the other hand, evidence suggests, that PTH promotes metabolic acidosis through inhibition of bicarbonate reabsorption from the glomerular filtrate (³⁷) and alterations in renal phosphate handling (³⁸). Obviously, several factors besides acidosis may contribute to the observed results, most importantly GFR, which correlated well with bicarbonate, phosphate, and PTH in our analysis. However, multiple linear regression revealed both GFR and PTH to be independently related to acidosis. Nevertheless, the causal relationship between posttransplant acidosis and disturbances in mineral metabolism observed in this study can not be proven by our analysis. To better understand the impact of acidosis to the regulation of calcium/phosphate homeostasis and its potential implications on bone in renal transplant patients, a prospective controlled interventional trial is warranted assessing the effect of base supplementation in this setting. Until the advent of such studies clinicians caring for renal transplant patients should implement the NKF-K/DOQI guidelines for maintenance dialysis patients. These guidelines recommend maintenance of serum bicarbonate levels at 22 mM/L or greater (⁹).