Metabolic acidosis adversely affects nutritional status and the musculoskeletal system as well as the cardiovascular system leading to worsening morbidity and mortality in patients in chronic renal failure1). Also, the experimental and clinical data showed the adverse effects of metabolic acidosis in maintenance dialysis patients regarding poor nutritional status due to its catabolic effects; which include increased protein breakdown, subnormal essential branched-chain amino acids, and abnormal bone metabolism2, 3). However, most of the recent epidemiologic studies have revealed an inverse correlation between metabolic acidosis and nutritional status, i.e. better nutritional status and lower relative risk for mortality and morbidity in hemodialysis patients with mild to moderate predialysis acidosis4, 5). Therefore, the optimal acidbase status in maintenance hemodialysis patients remains as an on-going debate6).

Though the method of hemodialysis in patients with end-stage renal disease (ESRD) is well known as a method of correction of metabolic acidosis by the supplement of bicarbonate, and NKF KDOQI (The National Kidney Foundation Kidney Disease Outcomes Quality Initiative) guideline recommends predialysis serum bicarbonate (HCO₃^-) concentration ≥22 mEq/L as a target to correct metabolic acidosis associated adverse effects7), the optimal target for the acid-base correction in hemodialysis by delivering base supply, i.e. bicarbonate buffer, is to maintain patients close to the physiological plasma bicarbonate range during the intradialytic as well as the interdialytic period. However, the variation in blood bicarbonate occurs in maintenance hemodialysis with the highest HCO₃^- just after dialysis and the lowest just before the next dialysis, the so-called "sawtooth pattern." In fact, mild to moderate predialysis metabolic acidosis with 22 mEq/L or less is very common in hemodialysis patients, which can be abolished by using high dialysate bicarbonate concentration (35-40 mEq/L), but only at the cost of postdialysis alkalosis8).

In general, the specific dialysate base concentration among the available commercial dialysate bicarbonate ranging from 25 to 40 mEq/L appears to be arbitrarily chosen without clear-cut guidelines for its selection so-far by physicians in charge of the dialysis unit. Therefore, there are two main issues to be resolved at present for maintaining the best acid-base status in maintenance hemodialysis patients. First, which one among varying dialysate bicarbonate solutions would be optimal as the standard concentration of bicarbonate in dialysate for maintaining or correcting acidbase status in maintenance hemodialysis patients. Second, which stage among predialysis, post dialysis, and interdialytic phase would be the optimal time for assessing the acid-base status in these patients?

To answer these two major unresolved questions, we assessed the acid-base status by arterial blood gas analysis at both predialysis and postdialysis stages in chronic maintenance hemodialysis patients allocated to one of 3 different conventional bicarbonate solutions, i.e. 25 mEq/L (low), 30 mEq/L (average), and 35 mEq/L (high).

This study evaluated the impact of varying dialysate bicarbonate concentrations of 25 mEq/L, 30 mEq/L, and 35 mEq/L on acid-base status just before and after dialysis in 53 chronic maintenance hemodialysis patients. In addition, we tried to find out the optimal dialysate base (bicarbonate) concentration among commercially available various dialysate solutions in maintenance hemodialysis patients to correct the acid-base imbalance, which had been on continuous debate9). The main results of this study are summarized as follows: 1) the prevalence of metabolic acidosis in maintenance hemodialysis patients is high, when defined by pre-dialysis pH <7.4 (77%) and pre-dialysis HCO₃^- <22 mEq/L (60%) (Table 2), 2) predialysis acid-base status showing mild to moderate degree of metabolic acidosis is similar regardless of the amount of base (bicarbonate) delivery by hemodialysis using three different dialysate bicarbonate concentrations, which suggests the more important role of acid loading and generation or alkali regeneration during the interdialytic period (Table 1), and 3) post-dialysis acid-base status changes from pre-dialysis metabolic acidosis toward metabolic alkalosis in all three different kinds of dialysate bicarbonate concentration with more prominent alkalemia in higher dialysate bicarbonate concentration (Table 2).

It has been well recognized that the correction of metabolic acidosis is one of the major goals in hemodialysis treatment of chronic maintenance hemodialysis patients since metabolic acidosis, if left untreated, may lead to systemic detrimental effects such as malnutrition due to negative nitrogen balance and increased protein catabolism, anorexia, fatigue, alterations in bone metabolism and formation, secondary hyperparathyroidism, and decreased cardiovascular functions10). Therefore, the treatment of metabolic acidosis can be assumed to reverse its many side effects in chronic hemodialysis patients. But the level of bicarbonate concentration at which the correction of harmful effects of metabolic acidosis is seen has not been clearly defined so far6).

Among several preferred recommendations in the world, NKF KDOQI guideline recommends to preserve the predialysis HCO₃^- concentration at or above 22 mEq/L7). However, most of the maintenance hemodialysis patients can be found to have low pre-dialysis HCO₃^- concentration. In fact, the prevalence of metabolic acidosis based on pre-dialysis HCO₃^- less than 22 mEq/L in our study was more than half of the total patients (60%), which is higher than that of the hemodialysis (HEMO) study of 1,000 patients (50%) by Uribarri et al.4).

The clinical importance of metabolic acidosis in dialysis patients has not been clearly clarified due to contradictory findings between the clinical or experimental data of hemodialysis patients revealing previously well-known detrimental effects and the vast majorities of recent epidemiologic studies of them showing the pattern of reverse epidemiology5, 6). One of the recent epidemiologic studies, DOPPS (Dialysis Outcomes and Practice Pattern Study) shows that moderate Predialysis metabolic acidosis seems to be associated with better nutritional status and lower relative risk for mortality or hospitalization than is observed in patients with normal or severe pre-dialysis metabolic acidosis5). Furthermore, whether pre-dialysis bicarbonate level alone in maintenance hemodialysis patients can reflect their acid-base status has only been addressed in few studies8).

The delivered quantity of bicarbonate in dialysate solution during dialysis could be dependent by the concentration of bicarbonate, flow speed of solution, filtration rate, type of the machine and dynamics of bicarbonate. A formula to predict HCO₃^- concentration after dialysis was created as 15.87122 - 0.03889 (dialysis hours) - 0.0337 (fluid speed) + 0.01621 (injected dialysis solution) + 0.44359 (bicarbonate concentration before dialysis) + 0.04048 (body weight) + coefficient of used dialysis membrane11). However, the authors suggested that this calculation predicting HCO₃^- concentration can be more accurate when the hydrogen ion generation in the body besides base supply by dialysis can be estimated. In our study, since the predialysis acid-base status showing mild to moderate degree of metabolic acidosis is similar regardless of the amount of base (bicarbonate) delivery by hemodialysis using three different kinds of dialysate bicarbonate concentration (Table 1), it suggests that pre-dialysis metabolic acidosis is independent from bicarbonate supply in the intradialytic phase, and the base supply by dialysis does not seem to represent the main mechanism for the correction of acid-base imbalance by hemodialysis12). Diet, intestine, bone, and other intermediate metabolism including the effects of medications would be involved for acid-base balance in addition to the base delivery by dialysis in maintenance hemodialysis patients9). In fact, besides the base supply by dialysis, Soudan et al. reported that HCO₃^- concentration decreases by 1 mEq/per 0.2 g/kg/d of increased protein intake. HCO₃^- concentration decreases by 0.37 mEq/L per 2 g/d of sevelamer intake. It also decreases by 0.76 mEq/L per 0.2 of increased Kt/V and by 0.90 mEq/L per 2 g/d of calcium carbonate intake13).

In a study using 30 mEq/L of dialysate bicarbonate solution, the patients reached normal acid-base balance following dialysis but returned to a mild degree of acidosis before the next dialysis. However, when using 35 mEq/L of dialysate bicarbonate solution, patients did not show acidosis before dialysis but showed alkalosis following dialysis8). Another study divided 46 patients using 35 mEq/L of dialysate bicarbonate solution into group A (n=21) and group B (n=25). A lowevel bicarbonate (30 mEq/L) and high-level bicarbonate (40 mEq/L) dialysate solution were used alternatively to each group with an interval of 6 months, respectively. As a result, serum total carbon dioxide concentration showed a significant difference between the low-level and high-level of dialysate bicarbonate solutions. But, increased pH during dialysis was not meaningful before and after dialysis14). This phenomenon was explained by bicarbonate-transfer, achieving a new balanced situation. Thus, it has been suggested that even though the high-level of bicarbonate dialysis solution is used, progressive alkalosis will not occur and patients will be safe14, 15).

In this research, along with universal pre-dialysis metabolic acidosis, post-dialysis metabolic alkalosis immediately after dialysis was noted in all three different dialysate bicarbonate concentrations with more prominent alkalemia in higher dialysate bicarbonate concentration (Table 2). Therefore, it could be considered the time-averaged bicarbonate concentration may be within the normal range, but at the cost of post-dialysis metabolic alkalosis16). The risks of post-dialysis metabolic alkalosis include cardiac arrhythmias, respiratory suppression, hypotension, and muscle cramps as acute episodes, and vascular or metastatic calcifications as a chronic detrimental effect, particularly in the presence of elevated calcium-phosphorus product17).

The results of our study have limitations due to the observational retrospective analysis of cross-sectional study and because analysis was done on only pre- and post-dialysis arterial pH and HCO₃^-. To evaluate clear-cut advantages and disadvantages of either pre-dialysis metabolic acidosis or post-dialysis metabolic alkalosis shown in our results would require the prospective cross-over study on varying dialysate bicarbonate concentrations in which various parameters associated with acid-base imbalance can be assessed simultaneously.

In conclusion, the optimal dialysate among various dialysate bicarbonate concentrations ranging from 25 to 40 mEq/L could be selected for the need of acidbase correction of maintenance hemodialysis patients currently. However, the base supply by hemodialysis alone does not seem to be the main factor to determine the predialysis acidosis in ESRD patients on chronic maintenance hemodialysis. Therefore, besides dialysate bicarbonate concentration, other factors involved in acid-base metabolism in both interdialytic and intradialytic phases should be considered to maintain the best physiologic acid-base status in maintenance hemodialysis patients. Also, this study warrants for the particular attention to the ratio of benefits and risks of post-dialysis metabolic alkalosis when choosing high dialysate bicarbonate concentration for maintenance hemodialysis. Meantime, it would be prudent to choose a more tailored regimen of dialysate bicarbonate concentration for hemodialysis patients, which could avoid severe post-dialysis alkalemia.