Acid-base profile in patients on PD

Transcription

1 Kidney International, Vol. 6, Supplement 88 (23), pp. S26 S36 Acid-base profile in patients on PD SALIM MUJAIS Renal Division, Baxter Healthcare Corporation, McGaw Park, Illinois Acid-base profile in patients on PD. Secular trends in dialysis dose in peritoneal dialysis (PD) and modes of dialysis delivery [automated PD () versus continuous ambulatory PD (C)] require a reexamination of acid-base status in patients treated with these renal replacement modalities. We explored steady-state acid-base profile and its determinants in 17 patients on C and 77 patients on. The majority (62% to 7%) of patients had serum bicarbonate levels in the normal range, and a minority (17% to 27%) had values just above 28 meq/l. Only a small percentage (1% to 12%) of patients in either the C or the groups had a serum HCO 3 less than 22 meq/l, an indication of the successful correction of acidosis in most patients. The anion gap was elevated (>16 meq/l) in the majority of patients on C and and bore an inverse relationship to serum HCO 3 and a direct relationship to serum albumin and serum phosphate. In C patients, but not patients, a significant inverse relationship was observed between the anion gap and peritoneal permeability as assessed by four-hour D/P creatinine. The correction of acidosis in PD appears to be predominantly achieved by the continuous supplementation of alkali via dialysis, with residual renal function not differentiating the degree of correction. Steady-state serum bicarbonate in patients on C appeared to be responsive to the underlying peritoneal membrane permeability characteristics of the patient that govern alkali loss and gain, but the higher dialysate volumes in appear to override this effect. Higher albumin, blood urea nitrogen (BUN), and phosphate in patients with lower HCO 3 suggest a discrepancy between daily acid load and dialysis dose. Acidosis is one of the cardinal manifestations of renal failure and its correction is one of the elementary and obvious goals of renal replacement therapy. A variety of studies have examined the correction of acidosis and the relevance of such corrections on the health of patients on dialysis [1 2]. While these contributions have greatly advanced our understanding of mechanisms of generation and correction of acidosis [13, 1], the ongoing changes in dialysis technology as to delivery modes and changing buffer use in both hemodialysis (HD) and peritoneal dialysis (PD) require that examination of contemporary patient groups be undertaken to delineate current Key words: peritoneal dialysis, acidosis, residual renal function, creatinine clearance. C 23 by the International Society of Nephrology successes. Changes in buffer sources have been undertaken in both HD (from acetate to bicarbonate) and in PD (from acetate to lactate, and to bicarbonate alone or supplemented with lactate) [1, 3, 6 1, 12]. Furthermore, changes in dialysis dose (increasing in both HD and PD) [21 28], dialysis frequency (shorter times in in-center HD and the emergence of short daily and nocturnal HD), and modifications in dialysis delivery modes (an increasing use of ), may have an impact on the correction of uremic acidosis. It is the aim of the present paper to examine the status of acidosis correction in patients on PD and compare the effects of C,, and patient factors that may modulate response to alkali therapy delivered by dialysis. METHODS The present analysis is based on cohorts of patients originally recruited to participate in a randomized controlled trial for the evaluation of a new dialysis solution [29, 3]. Only measurements obtained at baseline while patients were on standard dialysis regimen and solution are considered herein. The criteria for patient selection, and hence, inclusion in the current analysis, were as follows: (1) Patients who had given written informed consent; (2) patients who were at least 18 years old; (3) patients who were treated with PD for at least 9 days before the screening visit; and () patients whose standard prescription was stable for at least 3 days prior to the screening visit. The following patients were excluded from the study and hence the present analysis: (1) Patients who had an acute or chronic exit site or tunnel infection; (2) patients who required antibiotics for the treatment of an episode of peritonitis in the 3 days prior to the screening visit; (3) patients who were known to be HIV positive; () patients who had other serious disease such as active malignancy, or, if previously treated, residual malignancy or systemic infection; () patients who had active liver disease such as cirrhosis of the liver, active hepatitis, or other active liver disease; (6) patients who had an illness or injury requiring hospitalization during the 3 days preceding the screening visit; and (7) patients who were pregnant or lactating. It is recognized that any and each S-26

2 Mujais: Acid-base profile in patients on PD S-27 Table 1. Overall demographic profile Parameter C P value N Age.29 ± ± 1.66 NS Height 16.2 ± ± Weight 76. ± ± M:F 6:11 9:28 Duration of dialysis years 2.1 ± ±..1 Data are mean ± SE. Table 2. Overall biochemical profile Parameter C N BUN mg/dl 1.8 ± ± 1.92 Creatinine mg/dl 9.67 ± ±.9 Na meq/l ± ±. K meq/l 3.91 ±. 3.8 ±.7 Cl meq/l 9.7 ± ±.6 HCO 3 meq/l 2.3 ± ±.36 AGAP meq/l ± ±.1 Ca mg/dl 9. ± ±.1 PO mg/dl.7 ± ±.19 Albumin g/dl 3.21 ± ±.21 Cholesterol mg/dl ± ±.8 Data are mean ± SE. Abbreviations are: BUN, blood urea nitrogen; Na, sodium; K, potassium; Cl, chloride; HCO 3, bicarbonate; AGAP, anion gap; Ca, calcium; PO, phosphate. of the above exclusion conditions can have an effect on patient acid-base balance. As part of their baseline evaluation, all patients described had standard blood chemistry and hematology measures, as well as a standard peritoneal equilibration test (PET). All patients, who had urine output >1 ml/2 hours, collected their 2-hour urine output during the 2 hours preceding baseline. Urea nitrogen and creatinine were used to determine the residual renal function (RRF). RRF was calculated as the average of renal creatinine clearance (CrCl) and renal urea nitrogen clearance (Ucl) by standard formulas. RESULTS General characteristics of the population The overall profile of the study population is illustrated in Tables 1 3. The demographic characteristics of this population are similar to the dialysis population in North America in general and can thus be considered as representative of the state of stable patients on PD. Patients on were slightly larger than patients on C, had lower residual renal function (RRF), and used a higher volume of dialysis solution daily. All of these differences reflect common practice patterns in North America. A small percentage (1% to 12%) of patients in either the C or the groups had a serum HCO 3 less than 22 meq/l, an indication of the successful correction of acidosis in most patients. The majority (62% to Table 3. Overall renal and dialytic profile Parameter C N Urine volume ml/2 hr 718 ± 6 1 ± 61 CrCl ml/min.23 ± ±.8 Urea Cl ml/min 2.3 ±.2.96 ±.19 RRF ml/min 3.88 ± ±.31 Dialysis volume ml/2 hr 936 ± ± 31 -hour D/P creatinine.67 ±.1.69 ±.1 D /D glucose.38 ±.1.39 ±.2 -hour D/P urea.89 ±.1.9 ±.1 PET-UF ml/ hr 36 ± ± 26 LD-UF ml 29 ± ± Data are mean ± SE. Abbreviations are: CrCl, creatinine clearance; urea Cl, urea clearance; RRF, average of creatinine and urea clearances; -hour D/P creatinine, ratio of dialysate creatinine at hours of dwell to serum creatinine; D /D glucose, ratio of -hour dialysate glucose to hour dialysate glucose; -hour D/P urea, ratio of dialysate urea at hours of dwell to BUN; PET-UF, net ultrafiltration during a -hour PET with 2.% dextrose solution; LD-UF, net ultrafiltration during an overnight dwell with 2.% dextrose solution. 7%) of patients had serum bicarbonate levels in the normal range, and a small group (17% to 27%) had values just above 28 meq/l (Fig. 1). The anion gap was elevated (>16 meq/l) in the majority of patients on C and (Fig. 2) and bore an inverse relationship to serum HCO 3 (Fig. 3) and a direct relationship to serum albumin (Fig. ) and serum phosphate (Fig. ). In C patients, but not patients, a significant inverse relationship was observed between the anion gap and peritoneal permeability as assessed by four-hour D/P creatinine (Fig. 6). The correlation between the anion gap and albumin in C patients (r =.3, P <.1) was reduced when it was controlled for phosphate levels (r =.363, P <.1), but remained significant and was further reduced when controlled for both phosphate and four-hour D/P creatinine (r =.276, P <.1). In patients, however, the correlation between albumin and the anion gap (r =.3, P <.1) was minimally affected by controlling for phosphate (r =.28, P <.1) or both phosphate and four-hour D/P creatinine (r =.16, P <.1). In C patients, the correlation between the anion gap and four-hour D/P creatinine (r =.319, P <.1) was reduced but remained significant when it was controlled for phosphate (r =.269, P <.1). Controlling for albumin, however, abolished the correlation between the anion gap and the four-hour D/P creatinine (r =.13, P = ns). In neither C nor did the dialysis solution volume correlate with any of the biochemical markers of interest, either absolute or factored for body weight or height. Profile by residual renal function Because renal function plays a crucial role in acid excretion, the impact of residual renal function (RRF) on the acid-base status of patients on dialysis was examined to explore whether the effects of continuous dialysis

3 S-28 Mujais: Acid-base profile in patients on PD Patient number C HCO 3, meq/l HCO 3, meq/l Fig. 1. Distribution of serum HCO 3 values in patients on continuous ambulatory peritoneal dialysis (C) and automated peritoneal dialysis (). Numerals within each bar reflect number of patients falling within the indicated interval. C 16 Patient number Fig. 2. Distribution of serum anion gap values in patients on continuous ambulatory peritoneal dialysis (C) and automated peritoneal dialysis (). Numerals within each bar reflect number of patients falling within the indicated interval C 3 HCO 3, meq/l R =.667, P <.1 R =.76, P < Fig. 3. Inverse significant correlations were observed between serum HCO 3 and anion gap in continuous ambulatory peritoneal dialysis (C) and automated peritoneal dialysis () patients.

4 Mujais: Acid-base profile in patients on PD S-29.. C.. Albumin, g/dl R =.3, P <.1 1. R =.3, P < Fig.. Direct significant correlations were observed between serum albumin and anion gap in continuous ambulatory peritoneal dialysis (C) and automated peritoneal dialysis () patients. Phosphate, mg/dl C R =.7, P <.1 1 R =.39, P < Fig.. Direct significant correlations were observed between serum phosphate and anion gap in continuous ambulatory peritoneal dialysis (C) and automated peritoneal dialysis () patients. hour D/P creatinine 1. C R =.319, P < R =.27, P = NS Fig. 6. Inverse significant correlations were observed between four-hour D/P creatinine and anion gap in continuous ambulatory peritoneal dialysis (C) and automated peritoneal dialysis () patients.

5 S-3 Mujais: Acid-base profile in patients on PD were dominant, or whether the role of RRF was still apparent in the face of continuous alkali supplementation. Among patients on C, 86 had measurable urine output (mean ± SE, 718 ± 6 ml/2 h) and consequently measurable renal function (CrCl,.1 ±. ml/min; urea Cl, 2. ±.2 ml/min; RRF, 3.8 ±.3 ml/min), and 87 patients were considered anuric. The two groups were comparable for demographic characteristics (except for duration on dialysis, which was longer in the anuric patients: 2.8 ±.3 vs. 1. ±.1 years, P <.1). On examination of biochemical profile, albumin, blood urea nitrogen, cation levels (calcium, phosphate), electrolytes (Na, K, Cl, HCO 3 ) and anemia control were similar between the two groups. Serum bicarbonate levels were equivalent (2.3 ±.3 vs. 2.3 ±.3 meq/l), but the anion gap was slightly higher in the anuric group (22.1 ±. vs. 2.3 ±. meq/l, P <.1). The other notable difference between the two groups was serum creatinine, which was higher in the anuric patients compared with those with RRF (11. ±. vs. 8.3 ±.3 mg/dl, respectively, P <.1). Among patients on, 39 had measurable urine output (mean ± SE, 1 ± 61 ml/2 h) and consequently measurable renal function (CrCl,. ±.8 ml/min; urea Cl, 1.9 ±.3 ml/min; RRF, 3.1 ±. ml/min), and 38 patients were considered anuric. The two groups were comparable for demographic characteristics (except for duration on dialysis, which was longer in the anuric patients). On examination of biochemical profile, albumin, blood urea nitrogen, cation levels (calcium, phosphate), electrolytes (Na, K, Cl, HCO 3 ) and anemia control were similar between the two groups. Serum bicarbonate levels were equivalent (2.6 ±.6 vs. 2.9 ±. meq/l) and so were the anion gaps (2.7 ±.7 vs ±.7 meq/l). The only notable difference between the two groups was serum creatinine, which was higher in the anuric patients compared with those with residual renal function (13. ±.7 vs. 8.6 ±. mg/dl, respectively, P <.1). Profile by HCO 3 level To gain better insight into the factors underlying the broad range of values of bicarbonate, the C and groups were further subdivided into subsets based on serum bicarbonate levels (Tables 7). In C patients, those with HCO 3 lower than 22 had higher blood urea nitrogen (BUN), anion gap, phosphate, and cholesterol than the other two groups (P <.1 to P <.1) (Figs. 7 and 8). Furthermore, albumin levels were higher than the other two groups but were statistically different only compared with the high HCO 3 group (P <.) (Fig. 9). These differences in biochemical profile were paralleled by differences in peritoneal membrane characteristics, with the low bicarbonate group being characterized by lower membrane permeability and higher Table. C biochemical profile by HCO 3 status Parameter HCO 3 <22 22 HCO 3 <28 HCO 3 28 N% 22 (12.6%) 122 (7.1%) 3 (17.3%) BUN mg/dl 6.1 ± ± 1.. ± 2.6 Creatinine mg/dl 1.8 ±.9 9. ± ±.7 Na meq/l ± ± ±. K meq/l 3.96 ± ± ±.11 Cl meq/l 9.2 ± ±. 9.7 ±.7 HCO 3 meq/l 2.3 ±.2 a 2.9 ± ±. AGAP meq/l 2. ±.9 a 21.7 ± ±.7 Ca mg/dl 9. ± ±.7 9. ±.18 PO mg/dl.8 ±.3 b. ±.1.6 ±.2 Albumin g/dl 3.37 ± ±. 3. ±.11 Cholesterol mg/dl 23 ± 12 2 ± 19 ± 8 Data are mean ± SE. Abbreviations are: BUN, blood urea nitrogen; Na, sodium; K, potassium; Cl, chloride; HCO 3, bicarbonate; AGAP, anion gap; Ca, calcium; PO, phosphate. a P <.1 vs. each of the other two groups; b P <.1 vs. HCO 3 >28 group Table. C renal and dialytic profile by HCO 3 status Parameter <22 22 x <28 28 N% 22 (12.6%) 122 (7.1%) 3 (17.3%) Urine volume ml/2 hr 3 ± ± ± 96 CrCl ml/min 2.2 ± ± ±.96 Urea Cl ml/min 2.2 ± ± ±. RRF ml/min 3.76 ± ± ±.96 Dialysis volume ml/2 hr 929 ± ± ± 26 -hour D/P creatinine.61 ±.1.67 ±.1.71 ±.2 D /D glucose.2 ±.1.38 ±.7.3 ±.1 -hour D/P urea.86 ±.1.89 ±.7.9 ±.1 PET-UF ml/ hr ± 6 3 ± 19 3 ± 31 LD-UF ml 79 ± ± 32 2 ± 62 Data are mean ± SE. Abbreviations are: CrCl, creatinine clearance; urea Cl, urea clearance; RRF, average of creatinine and urea clearances; -hour D/P creatinine, ratio of dialysate creatinine at hours of dwell to serum creatinine; D /D glucose, ratio of -hour dialysate glucose to hour dialysate glucose; hour D/P urea, ratio of dialysate urea at hours of dwell to BUN; PET-UF, net ultrafiltration during a -hour PET with 2.% dextrose solution; LD-UF, net ultrafiltration during an overnight dwell with 2.% dextrose solution. ultrafiltration response than the other two groups (P <.1 for all comparisons). Patients in the intermediate HCO 3 group differed from the high HCO 3 group only in terms of a higher anion gap (P <.1), slightly elevated phosphate and BUN (P <.1), and marginally significant lower membrane permeability and higher UF response. In patients, those with HCO 3 lower than 22 had higher BUN, anion gap, and phosphate than the other two groups, but only the anion gap was significantly different from the intermediate HCO 3 group (P <.1) (Figs. 7 and 8). BUN, anion gap, and phosphate were significantly different from the higher HCO 3 group (P <.1). Furthermore, albumin levels were higher than the other two groups but were statistically different only compared with the high HCO 3 group (P <.2) (Fig. 9). Unlike patients on C, the three groups were not distinct in peritoneal membrane characteristics and ultrafiltration response. Patients in the intermediate HCO 3 group differed from the high HCO 3 group only in terms of a higher anion gap (P <.1) and higher BUN (P <.1).

6 Mujais: Acid-base profile in patients on PD S-31 Table 6. biochemical profile by HCO 3 status Parameter HCO 3 <22 22 HCO 3 <28 HCO 3 28 N% 8 (1.%) 8 (62.3%) 21 (27.3%) BUN mg/dl 6. ± ± ± 2.7 Creatinine mg/dl 12.6 ± ± ±.8 Na meq/l 13.7 ± ± ±.9 K meq/l.19 ± ± ±.1 Cl meq/l 93.2 ± ±. 9.6 ±.9 HCO 3 meq/l 21. ± ±.2 3. ±. AGAP meq/l 2.7 ± ± ±.7 Ca mg/dl 8.8 ± ± ±.17 PO mg/dl 6.3 ±.7.2 ±.2. ±.3 Albumin g/dl 3. ± ± ±.1 Cholesterol mg/dl 21 ± ± 6 29 ± 1 Data are mean ± SE. Abbreviations are: BUN, blood urea nitrogen; Na, sodium; K, potassium; Cl, chloride; HCO 3, bicarbonate; AGAP, anion gap; Ca, calcium; PO, phosphate. No correlation was found between HCO 3 and peritoneal transport as reflected in the four-hour D/P creatinine.no correlation was found between HCO 3 and serum creatinine or serum albumin or any measure of renal clearance (CrCl, urea Cl, or RRF). In patients, a significant correlation was found between the anion gap and each of the following parameters: BUN (r =.382, P <.1), albumin (r =.3, P <.1), serum creatinine (r =.36, P <.1), phosphate (r =.39, P <.1), and urea Cl (r =.29, P <.1). Unlike C patients, no correlation was found in patients between the anion gap and peritoneal transport profile, and the correlations with measures of renal clearance were of lower value and significant only for urea clearance. Table 7. renal and dialytic profile by HCO 3 status Parameter HCO 3 <22 22 HCO 3 <28 HCO 3 28 N% 8 (1.%) 8 (62.3%) 21 (27.3%) Urine volume ml/2 hr 1 ± 17 1 ± 7 67 ± 181 CrCl ml/min.92 ± ± ±.7 Urea Cl ml/min.8 ± ± ±. RRF ml/min.7 ± ± ±.9 Dialysis volume ml/2 hr ± ± ± 736 -hour D/P creatinine.66 ±.3.7 ±.1.68 ±.2 D /D glucose. ±.2.1 ±.3.36 ±.2 -hour D/P urea.88 ±.3.9 ±.1.89 ±.2 PET-UF ml/ hr 13 ± ± ± 9 LD-UF ml 3 ± ± ± 7 Data are mean ± SE. Abbreviations are: CrCl, creatinine clearance; urea Cl, urea clearance; RRF, average of creatinine and urea clearances; -hour D/P creatinine, ratio of dialysate creatinine at hours of dwell to serum creatinine; D /D glucose, ratio of -hour dialysate glucose to hour dialysate glucose; hour D/P urea, ratio of dialysate urea at hours of dwell to BUN; PET-UF, net ultrafiltration during a -hour PET with 2.% dextrose solution; LD-UF, net ultrafiltration during an overnight dwell with 2.% dextrose solution. Because categorization by HCO 3 status is empiric and arbitrary, we further explored whether HCO 3 as a continuous variable was correlated with other parameters of biochemical control. In patients on C, there was an inverse correlation between HCO 3 and BUN (r =.3, P <.1), albumin (r =.19, P <.1), and phosphate levels (r =.279, P <.1). A direct correlation was found between HCO 3 and peritoneal transport as reflected in the four-hour D/P creatinine (r =.2, P <.1). No correlation was found between HCO 3 and serum creatinine or any measure of renal clearance (CrCl, urea Cl, or RRF). In C patients, a significant correlation was found between the anion gap and each of the following parameters: BUN (r =.29, P <.1), albumin (r =.3, P <.1), serum creatinine (r =.362, P <.1), four-hour D/P creatinine (r =.319, P <.1), phosphate (r =.7, P <.1), and measures of renal clearance [CrCl (r =.371, P <.1), urea Cl (r =.386, P <.1), and RRF (r =.393, P <.1)]. In patients on, there was an inverse correlation found only between HCO 3 and BUN (r =.371, P <.1) and phosphate levels (r =.336, P <.). Profile by transport status To gain further insight into the impact of peritoneal transport status on biochemical parameters, the C and groups were divided into four subsets corresponding to the standard definitions of transport subtypes by PET. In C patients, no notable differences were observed between the four subgroups with respect to BUN, creatinine, electrolytes, calcium, and cholesterol (Tables 8 and 9). The only significant difference in HCO 3 levels was between high transporters and low-average transporters (P <.7), but considering the proximity of the values for the four subgroups it is not likely to represent a clinically meaningful finding. A tendency for phosphate levels to increase with declining peritoneal permeability parameters was observed with the difference between high, high-average, and low-average of borderline statistical significance (P <.8), but attaining statistical significance compared to low transporters (P <.2). A similar tendency for the anion gap to increase with declining peritoneal permeability was also observed and was statistically significant between high and high-average (P <.17), high and low-average (P <.2), and marginally significant between high-average and lowaverage (P =.6). The trend for increasing serum albumin with diminishing peritoneal permeability was statistically significant for all comparisons (P <.1). Among C patients, peritoneal transport profile did not have an effect on measures of renal function, but, as expected, was manifest in the parameter of peritoneal transport and ultrafiltration. Ultrafiltration during the PET and during the overnight dwell with 2.% dextrose showed a tendency to increase with declining peritoneal permeability measures. In patients on, statistically significant differences in biochemical profile were not found in intergroup comparisons for any of the biochemical markers measured (Tables 9 and 1). This may have been due to the small size of the two groups (high and low transporters) represented in the sample.

7 S-32 Mujais: Acid-base profile in patients on PD 12 1 C 12 1 BUN, mg/dl <22 22<xx<28 >28 HCO 3, meq/l <22 22<xx<28 >28 HCO 3, meq/l Fig. 7. Blood urea nitrogen (BUN) values were higher in patients with lower HCO 3 values in both continuous ambulatory peritoneal dialysis (C) and automated peritoneal dialysis () patients. 3 C High High Av Low Av PET category Low High High Av Low Av PET category Low Fig. 8. Distribution of anion gap values by peritoneal equilibration test (PET) category. C Albumin, g/dl High High Av Low Av Low High High Av Low Av Low PET category PET category Fig. 9. Albumin values were lower in patients with higher peritoneal equilibration test (PET) categories in both continuous ambulatory peritoneal dialysis (C) and automated peritoneal dialysis () patients. The differences, however, were more marked in C patients.

8 Mujais: Acid-base profile in patients on PD S-33 Table 8. C biochemical profile by transport status Parameter High High Avg. Low Avg. Low N% BUN mg/dl 6. ± ± ± ± 7.3 Creatinine mg/dl 9.2 ± ± ±. 1.9 ± 1. Na meq/l ± ± ± ±.9 K meq/l 3.83 ± ± ±.8.2 ±.29 Cl meq/l 9.7 ± ±. 9.6 ± ± 1.7 HCO 3 meq/l 26.6 ±.6 2. ±. 2.6 ±. 2.6 ± 1.1 AGAP meq/l 19.1 ± ± ± ± 1.1 Ca mg/dl 9.2 ±.2 9. ± ± ±.27 PO mg/dl. ±.2.1 ±.1.2 ±.2.7 ±. Albumin g/dl 2.62 ± ±. 3. ± ±.9 Cholesterol mg/dl 198 ± 1 23 ± 216 ± ± 13 Data are mean ± SE. Abbreviations are: BUN, blood urea nitrogen; Na, sodium; K, potassium; Cl, chloride; HCO 3, bicarbonate; AGAP, anion gap; Ca, calcium; PO, phosphate. Because categorization by PET category is empiric and arbitrary, we further explored whether measures of peritoneal permeability as continuous variables were correlated with other parameters of biochemical control. In C patients, four-hour D/P creatinine was significantly inversely correlated with BUN (r =.179, P <.), albumin (r =.91, P <.1), anion gap (r =.319, P <.1), and phosphate (r =.19, P <.12), and directly correlated with HCO 3 (r =.2, P <.1). A similar pattern was observed for four-hour D/D glucose, which was significantly correlated with albumin (r =.6, P <.1), anion gap (r =.32, P <.1), and phosphate (r =.183, P <.2), and inversely correlated with HCO 3 (r =.28, P <.1). The correlations of four-hour D/P urea with biochemical measures was similar to that of four-hour D/P creatinine, but of lower magnitude. In patients on, no correlations were found between any of the peritoneal permeability parameters and measures of biochemical control. DISCUSSION The findings of this study can be summarized as follows: (1) Patients on PD have, as a group, adequate correction of acidosis as reflected by normal mean bicarbonate levels; (2) the correction of acidosis in PD appears to be predominantly achieved by the continuous supplementation of alkali via dialysis with RRF not differentiating the degree of correction; (3) steady-state serum bicarbonate in patients on C appears to be responsive to the underlying peritoneal membrane permeability characteristics of the patient which govern alkali loss and gain, but the higher dialysate volumes in appear to override this effect; and () higher albumin, BUN, and phosphate in patients with lower HCO 3 suggest a discrepancy between daily acid load and dialysis dose as the latter is usually adjusted to the uninformative Kt/V measure rather than indices of metabolic correction. The observation that the majority of patients on either C or had adequate correction of their metabolic acidosis does not corroborate statements made in the literature about the status of acidosis correction in these populations. Indeed, the statements of lack of full correction of metabolic acidosis [1] in reference to the use of mmol/l lactate are dated as they reference studies in the early 198s [31] and do not represent the contemporary status, particularly in view of the higher dose of dialysis used currently [21]. The results of the present study also call into doubt the representativeness of recent studies that show values for bicarbonate in patients on lactate-based solutions significantly lower than those observed in current North American practice [3]. These aberrantly lower values have been used as reference for comparative efficiency of dialysis solutions with different alkali preparations, which is clearly problematic. The present report represents the largest database of the status of acidosis correction in patients on C and. Our findings are consistent with several publications in the literature of smaller series. Kung et al [32] reported mean HCO 3 values of 2.9 meq/l in 3 patients, with nine patients having values lower than 22, a proportion slightly higher than ours. Dumler and Gallan [33] described values of HCO 3 of 2 ± meq/l, stable over a one-year period of observation with 13 of 8 patients having values lower than 22 meq/l. Interestingly, Kung et al [32] found that the lowest values of HCO 3 were seen in well nourished patients (as assessed by subjective global assessment [SGA]), suggesting that a higher food intake may be responsible for an acid load higher than the rate of alkali replenishment by dialysis. Higher values for BUN and albumin were also observed in the Korean patients with the lowest HCO 3. Kang et al [3] found a negative correlation between npna and HCO 3, suggesting that increased dietary acid load is responsible for the lower HCO 3. The elevation of the anion gap would not be surprising in patients with untreated uremic metabolic acidosis as it is a hallmark of the condition. The persistent elevation in the face of normal bicarbonate levels invites further scrutiny [3]. While the correlation between the anion gap and serum bicarbonate is expected for the condition, the curve is shifted upward because of the normal levels of bicarbonate. The increase in the anion gap cannot in this case be explained by decreased cations as the concentrations of K and Ca were normal, and Mg, while not measured in the present study, is not likely to be low in patients with advanced renal failure. Additionally, unidentified anions of a hyperosmolar state can also be ruled out. Elevations in either inorganic anions (phosphate and sulfate) or organic anions (lactate, ketones, and uremic organic anions of various identities) must be invoked as residual causality factors. Phosphate levels were elevated in these patients and correlated with the anion gap. While

9 S-3 Mujais: Acid-base profile in patients on PD Table 9. C renal and dialytic profile by transport status Parameter High High Avg. Low Avg. Low N% Urine volume ml/2 hr 6 ± ± 13 6 ± 78 3 ± 121 CrCl ml/min 3.6 ± ± ± ± 1.17 Urea Cl ml/min 1.7 ± ± ± ±.1 RRF ml/min 2.6 ±.77.1 ±.1.26 ± ±.83 Dialysis volume ml/2 hr 978 ± ± ± ± 327 -hour D/P creatinine.88 ±.1.71 ±.1.8 ±.1. ±.3 D /D glucose.26 ±.1.37 ±.1. ±.1. ±.3 -hour D/P urea.98 ±.1.92 ±..8 ±.1.72 ±.6 PET-UF ml/ hr 2 ± 33 9 ± 2 72 ± 2 9 ± 1 LD-UF ml 133 ± ± 16 ± ± 81 Data are mean ± SE. Abbreviations are: CrCl, creatinine clearance; urea Cl, urea clearance; RRF, average of creatinine and urea clearances; -hour D/P creatinine, ratio of dialysate creatinine at hours of dwell to serum creatinine; D /D glucose, ratio of -hour dialysate glucose to hour dialysate glucose; -hour D/P urea, ratio of dialysate urea at hours of dwell to BUN; PET-UF, net ultrafiltration during a -hour PET with 2.% dextrose solution; LD-UF, net ultrafiltration during an overnight dwell with 2.% dextrose solution. Table 1. biochemical profile by transport status Parameter High High Avg. Low Avg. Low N% BUN mg/dl.3 ± ± ± ± 1.6 Creatinine mg/dl 1.76 ± ± ± ± 3.9 Na meq/l ± ± ± ± 1.19 K meq/l. ± ± ± ±.1 Cl meq/l ± ± ± ±.8 HCO 3 meq/l 2.77 ± ± ± ± 1.73 AGAP meq/l 19.9 ± ± ± ±.8 Ca mg/dl 9. ± ± ± ±.72 PO mg/dl.82 ± ±.2.3 ± ± 1.6 Albumin g/dl 3.8 ± ± ±.8 3. ±.22 Cholesterol mg/dl 27 ± ± ± 8 2 ± 23 Data are mean ± SE. Abbreviations are: BUN, blood urea nitrogen; Na, sodium; K, potassium; Cl, chloride; HCO 3, bicarbonate; AGAP, anion gap; Ca, calcium; PO, phosphate. Table 11. renal and dialytic profile by transport status Parameter High High Avg. Low Avg. Low N% Urine volume ml/2 hr 16 ± ± 9 2 ± 9 23 ± CrCl ml/min 2.1 ± ±.8 1. ±..9 ±.9 Urea Cl ml/min 1. ±. 1.2 ±.3.6 ± ±.26 RRF ml/min 1. ± ±. 1.1 ±..9 ±.9 Dialysis volume ml/2 hr 16 ± ± ± ± 172 -hour D/P creatinine.87 ±.2.72 ±.1.61 ±.1. ± D /D glucose.2 ±.2.38 ±.2.7 ±..36 ±.1 -hour D/P urea.99 ±.1.92 ±.1.8 ±.2.6 ± PET-UF ml/ hr 168 ± ± 29 ± 7 1 ± 11 LD-UF ml 99 ± ± ± ± 37 Data are mean ± SE. Abbreviations are: CrCl, creatinine clearance; urea Cl, urea clearance; RRF, average of creatinine and urea clearances; -hour D/P creatinine, ratio of dialysate creatinine at hours of dwell to serum creatinine; D /D glucose, ratio of -hour dialysate glucose to hour dialysate glucose; hour D/P urea, ratio of dialysate urea at hours of dwell to BUN; PET-UF, net ultrafiltration during a -hour PET with 2.% dextrose solution; LD-UF, net ultrafiltration during an overnight dwell with 2.% dextrose solution. these patients were on lactate-containing dialysis fluid, the likelihood of an elevation in serum lactate in stable patients on PD is extremely remote. The load of lactate absorbed in PD is greatly inferior to the rate of production of lactic acid in normal humans (1 to 3 meq/kg/day), and elevation in serum lactate has not been reported in patients on PD [36]. The inclusion criteria for the study also rule out ketones as a source of organic anions. We are therefore left with the inorganic anions [phosphate (documented) and sulfate (assumed)] and the elevation in organic uremic anions. The low albumin concentration observed in these patients, while typical of patients on dialysis in North America, is expected to lead to a narrowing of the anion gap. Indeed, the correlation between serum albumin and serum anion gap confirms this expectation (Fig. ). The lowest anion gap values are observed in the patients with the lowest albumin. A correction of the anion gap for the decline in albumin would have

10 Mujais: Acid-base profile in patients on PD S-3 resulted in higher values for the anion gap in the majority of subjects in this study. Finally, some of the increase in the anion gap may be due to the level of alkalosis observed in some patients, as alkalosis affects the charge on serum albumin. It is interesting to note, however, that the lowest levels of anion gap were observed in the patients with the highest values of serum bicarbonate (Fig. 3). The combination of a low anion gap and high bicarbonate may imply that the acid burden from diet and metabolism is lower than alkali gain from dialysis and may identify patients with lower dietary protein intake. Correlations between the anion gap and other biochemical measures in patients on PD were previously explored by Dumler et al [37]. They found a direct correlation (P <.1) between anion gap and serum albumin (R =.2), BUN (R =.88), and serum creatinine (R =.73) concentrations [37]. These authors used a definition of acidosis based exclusively on the elevation in anion gap [37]. However, as our analysis illustrates, the relationship between bicarbonate and anion gap is shifted in such a way that an elevated anion gap will continue to be observed in these patients despite a normalization of the metabolic acidosis. The impact of the PD modality used (C vs. ) on acid base balance has been invoked as unexplored in recent reviews [13]. The present study serves to address this issue in that it shows equal correction of metabolic acidosis in both modalities of PD. The use of short cycles in nighttime period may be expected to alter the kinetics of buffer exchange (discussed in a separate paper in this supplement), but the net overall effect of acidosis correction appears to be identical. In neither group did the level of residual renal function influence acidbase status, suggesting that the provision of alkali via PD overrides quantitatively the small contribution of RRF. There were differences, however, observed in the relationships between certain measures of acid-base status and transport characteristics that were different between C and. These were notably the independence of HCO 3 levels in from the patient transport status and their dependence on peritoneal permeability in patients on C. This may relate to the higher dialysate volume and, hence, alkali exposure in patients on. patients are recurrently exposed to high lactate concentrations by virtue of the solution replenishment during frequent short exchanges during the nocturnal phase of the therapy. In patients on C, peritoneal transport status appeared to be correlated with HCO 3 levels, BUN, albumin, and the anion gap. In both C and, lower levels of HCO 3 were observed in association with higher BUN and phosphate, suggesting a slight discrepancy between the acid burden (reflected in the high BUN and phosphate from protein catabolism) and the dialytic provision of alkali. It is plausible that, had dialysis dose been directed at general metabolic correction rather than a fixed Kt/V target, such a discrepancy may have been bridged. Kt/V is notoriously uninformative when used as the sole guidance for dialysis dose. The source of higher acid load was also associated with higher albumin level reflecting likely a better nutritional status. CONCLUSION Patients on PD have as a group adequate correction of acidosis as reflected by normal mean bicarbonate levels. The correction of acidosis in PD appears to be predominantly achieved by the continuous supplementation of alkali via dialysis with RRF not differentiating the degree of correction. Steady-state serum bicarbonate in patients on C appears to be responsive to the underlying peritoneal membrane permeability characteristics of the patient that govern alkali loss and gain, but the higher dialysate volumes in appear to override this effect. Higher albumin, BUN, and phosphate in patients with lower HCO 3 suggest a discrepancy between daily acid load and dialysis dose, as the latter is usually adjusted to the uninformative Kt/V measure rather than indices of metabolic correction. Adjustment of dialysis dose to achieve full metabolic correction needs to be considered. REFERENCES 1. FERIANI M: Use of different buffers in peritoneal dialysis. Semin Dial 13:26 26, 2 2. FERIANI M, DELL AQUILA R: Buffer content in automated peritoneal dialysis solutions. Contrib Nephrol 129:187 19, FERIANI M, KIRCHGESSNER J, LA GRECA G, PASSLICK-DEETJEN J: Randomized long-term evaluation of bicarbonate-buffered C solution. Kidney Int : , FERIANI M: Behaviour of acid-base control with different dialysis schedules. Nephrol Dial Transplant 13(Suppl 6):62 6, FERIANI M, DELL AQUILA R: Acid-base balance and replacement solutions in continuous renal replacement therapies. Kidney Int 66(Suppl):S16 19, FERIANI M: Bicarbonate-buffered C solutions: From clinical trials to clinical practice. Perit Dial Int 17(Suppl 2):S1, FERIANI M, CAROBI C, LA GRECA G, et al: Clinical experience with a 39 mmol/l bicarbonate-buffered peritoneal dialysis solution. Perit Dial Int 17:17 21, FERIANI M: Buffers: Bicarbonate, lactate and pyruvate. Kidney Int 6(Suppl):S7 8, FERIANI M, RONCO C, LA GRECA G: Acid-base balance with different C solutions. Perit Dial Int 16(Suppl 1):S , FERIANI M, PASSLICK-DEETJEN J, LA GRECA G: Factors affecting bicarbonate transfer with bicarbonate-containing C solution. Perit Dial Int 1:336 31, FERIANI M: Adequacy of acid base correction in continuous ambulatory peritoneal dialysis patients. Perit Dial Int 1(Suppl 3):S , FERIANI M, DISSEGNA D, LA GRECA G, PASSLICK-DEETJEN J: Shortterm clinical study with bicarbonate-containing peritoneal dialysis solution. Perit Dial Int 13:296 31, GENNARI FJ: Acid-base balance in dialysis patients. Semin Dial 13:23 239, 2 1. GENNARI FJ: Acid-base balance in dialysis patients. Kidney Int 28: , TRANAEUS A: A long-term study of a bicarbonate/lactate-based peritoneal dialysis solution Clinical benefits. The Bicarbonate/Lactate Study Group. Perit Dial Int 2:16 23, 2

11 S-36 Mujais: Acid-base profile in patients on PD 16. BERGSTROM J: Metabolic acidosis and nutrition in dialysis patients. Blood Purif 13: , CARRASCO AM, RUBIO MA, SANCHEZ TOMMERO JA, et al: Acidosis correction with a new 2 mmol/l bicarbonate/1 mmol/l lactate peritoneal dialysis solution. Perit Dial Int 21:6 3, COLES GA, GOKAL R, OGG C, et al: A randomized controlled trial of a bicarbonate- and a bicarbonate/lactate-containing dialysis solution in C. Perit Dial Int 17:8 1, GRAHAM KA, REAICH D, CHANNON SM,et al: Correction of acidosis in C decreases whole body protein degradation. Kidney Int 9:1396 1, STEIN A, MOORHOUSE J, ILES-SMITH H, et al: Role of an improvement in acid-base status and nutrition in C patients. Kidney Int 2:189 19, FRANKENFIELD DL, PROWANT BF, FLANIGAN MJ,et al: Trends in clinical indicators of care for adult peritoneal dialysis patients in the United States from 199 to ESRD Core Indicators Workgroup. Kidney Int : , FLANIGAN MJ, ROCCO MV, PROWANT B, et al: Clinical performance measures: The changing status of peritoneal dialysis. Kidney Int 6: , FLANIGAN MJ, ROCCO MV, FRANKENFIELD D, et al: 1997 Peritoneal Dialysis-Core Indicators Study: Dialysis Adequacy and Nutritional Indicators Report. Am J Kidney Dis 33:e3, ROCCO MV, FLANIGAN MJ, PROWANT B, et al: Cycler adequacy and prescription data in a national cohort sample: The 1997 core indicators report. Health Care Financing Administration Peritoneal Dialysis Core Indicators Study Group. Kidney Int :23 239, FLANIGAN MJ, ROCCO MV, FRANKENFIELD DL, et al: 1996 peritoneal dialysis Core indicators report. Am J Kidney Dis 32:E3, FREDERICK PR, FRANKENFIELD DL, BIDDLE MG, SIMS TW: Changes in dialysis units quality improvement practices from 199 to Anna J 2:69 78, HELGERSON SD, MCCLELLAN WM, FREDERICK PR, et al: Improvement in adequacy of delivered dialysis for adult in-center hemodialysis patients in the United States, 1993 to 199. Am J Kidney Dis 29:81 861, ROCCO MV, BEDINGER MR, MILAM R, et al: Duration of dialysis and its relationship to dialysis adequacy, anemia management, and serum albumin level. Am J Kidney Dis 38: , WOLFSON M, HAGEN T, OGRINC F, MARTIS L, for the Icodextrin Study Group: Effects of icodextrin on ultrafiltration (UF) and small solute clearance in continuous ambulatory peritoneal dialysis patients (C). J Am Soc Nephrol 12:317A, WOLFSON M, HAGEN T, OGRINC F, MARTIS L, for the Icodextrin Study Group: One year results Icodextrin vs dextrose for the long dwell in peritoneal dialysis. J Am Soc Nephrol 12:317A, NOLPH K, PROWANT B, SERKES KD, et al: Multicenter evaluation of a new peritoneal dialysis solution with a high lactate and a low magnesium concentration. Perit Dial Bull 3:63 6, KUNG SC, MORSE SA, BLOOM E, RAJA RM: Acid-base balance and nutrition in peritoneal dialysis. Adv Perit Dial 17:23 237, DUMLER F, GALAN M: Impact of acidosis on nutritional status in chronic peritoneal dialysis patients. Adv Perit Dial 12:37 31, KANG DH, YOON KI, LEE HY, Han DS: Impact of peritoneal membrane transport characteristics on acid-base status in C patients. Perit Dial Int 18:29 32, SALEM MM, MUJAIS SK: Gaps in the anion gap. Arch Intern Med 12: , SENNESAEL JJ, DE SMEDT GC, VAN DER NIEPEN P, VERBEELEN DL: The impact of peritonitis on peritoneal and systemic acid-base status of patients on continuous ambulatory peritoneal dialysis. Perit Dial Int 1:61 6, DUMLER F, FALLA P, BUTLER R, et al: Impact of dialysis modality and acidosis on nutritional status. Asaio J :13 17, 1999