Metabolic acidosis in hemodialysis: a neglected problem in Brazil

ABSTRACT Introduction: Metabolic acidosis is associated with the high mortality seen in hemodialysis patients. The panorama of metabolic acidosis in hemodialysis in Brazil is unclear since 1996 when the analysis of bicarbonate levels was no longer a compulsory exam. We aimed to establish the prevalence of metabolic acidosis in a hemodialysis population and analyze the factors associated with low bicarbonate levels. Methods: A cross-sectional study was carried out to assess the prevalence of metabolic acidosis in adults undergoing regular hemodialysis from January to April 2017, in four dialysis centers from Niteroi, Rio de Janeiro, Brazil, and surroundings. For blood gas analysis, samples of 2 mL were collected in heparinized syringes before a midweek dialysis session. Results: 384 patients with a mean age of 58.1 ± 15.8 years (54.5% men and 63.0%, non-white) were included. Approximately 30% had diabetes and 48%, hypertension. Nearly 88% used primary arteriovenous fistula as vascular access. The pre-dialysis mean serum tCO2 in the midweek session was 22.7 ± 3.0 mEq/L. The prevalence rate of serum bicarbonate below DOQI recommendation (22 mEq/L or higher) was 40.3%, and 6.5% had serum bicarbonate < 18 mEq/L. The dialyzer use count and the use of low-flux dialyzers were negatively associated whereas age and the standard Kt/V values were positively associated with the serum bicarbonate levels. Conclusion: The findings were in agreement with global data reported in previous studies. However, because the sample was relatively small and non-representative of the Brazilian population, a more comprehensive study, addressing national data is necessary to substantiate our findings.

Introduction: Metabolic acidosis is associated with the high mortality seen in hemodialysis patients. The panorama of metabolic acidosis in hemodialysis in Brazil is unclear since 1996 when the analysis of bicarbonate levels was no longer a compulsory exam. We aimed to establish the prevalence of metabolic acidosis in a hemodialysis population and analyze the factors associated with low bicarbonate levels. Methods: A cross-sectional study was carried out to assess the prevalence of metabolic acidosis in adults undergoing regular hemodialysis from January to April 2017, in four dialysis centers from Niteroi, Rio de Janeiro, Brazil, and surroundings. For blood gas analysis, samples of 2 mL were collected in heparinized syringes before a midweek dialysis session. Results: 384 patients with a mean age of 58.1 ± 15.8 years (54.5% men and 63.0%, non-white) were included. Approximately 30% had diabetes and 48%, hypertension. Nearly 88% used primary arteriovenous fistula as vascular access. The pre-dialysis mean serum tCO 2 in the midweek session was 22.7 ± 3.0 mEq/L. The prevalence rate of serum bicarbonate below DOQI recommendation (22 mEq/L or higher) was 40.3%, and 6.5% had serum bicarbonate < 18 mEq/L. The dialyzer use count and the use of low-flux dialyzers were negatively associated whereas age and the standard Kt/V values were positively associated with the serum bicarbonate levels. Conclusion: The findings were in agreement with global data reported in previous studies. However, because the sample was relatively small and nonrepresentative of the Brazilian population, a more comprehensive study, addressing national data is necessary to substantiate our findings.

resumo IntroductIon
The prevalence of end-stage renal disease (ESRD) has been growing worldwide. 1 Hemodialysis patients have a high mortality rate, with cardiovascular disease accounting for about 50% of the fatalities. 2 Several studies indicate that the presence of metabolic acidosis is associated with high mortality. 3,4,5,6 Conventional hemodialysis treatment may not be sufficient for the adequate control of acidosis in ESRD patients 7,8 and complementary interventions, such as oral bicarbonate supplementation, may be required.
Brazil has the third largest number of patients on dialysis of the world with 126,593 people in 2017. 9 The panorama of metabolic acidosis in hemodialysis in Brazil is unclear, especially because, since 1996, the determination of bicarbonate levels is no longer required as a compulsory laboratory exam in hemodialysis patients, following a Brazilian Ministry of Health ordinance. 10 The present study aimed to establish the prevalence of metabolic acidosis in a hemodialysis population of four centers in a metropolitan area of Rio de Janeiro, Brazil, and to analyze the factors associated with low bicarbonate levels including demographics, comorbidities, and characteristics of the dialysis treatment.

methods
This was a cross-sectional study of the prevalence of metabolic acidosis in adult patients undergoing regular hemodialysis in Niteroi city and surroundings in the State of Rio de Janeiro. Data were collected from January to April 2017. The convenience sample consisted of patients from four centers who agreed to participate. Patients less than 18 years old, with less than 3 months in hemodialysis, or positive serology for hepatitis and HIV were excluded. The Ethics Committee of the Medical School of Universidade Federal Fluminense approved the study under the number CAAE 57822316900005243 and all patients provided signed informed consent. The study conformed to the principles of the Declaration of Helsinki.

PrOcedures
Dialysis centers were numbered from 1 to 4. In centers 1, 2, and 3 dialysis sessions were performed using 4008S hemodialysis machines (Fresenius Medical Care, Schweinfurt, Germany). In center 4, hemodialysis machines consisted of Diamax or Surdial (both made by Nipro Co., Osaka, Japan). Centers 1 and 2 used high-flux polysulfone dialyzers (HF-80S, Fresenius Medical Care, Bad Homburg, Germany); center 3, high--flux polyamix dialyzers (Polyflux 2.1 ® , Gambro, Bad Homburg, Germany); and center 4, low-flux dialyzers (Hemoflow F10 HPS, Fresenius Medical Care, Bad Homburg, Germany). In every center, dialyzers were reprocessed with peracetic acid as the sterilant solution by an automated system for a maximal of 20 times following Brazilian renal replacement therapy regulations. Fresenius Medical Care (Jaguariúna -SP, Brazil) manufactured all the dialysis concentrates used. Characteristics of the dialysis treatment of each center are listed in Table  1.
For blood gas analysis, samples of 2 mL were collected in heparinized syringes (S-Monovette Blood Collection System, SARSTEDT AG & Co. KG, Nümbrecht, Germany) before a midweek dialysis session. Species were collected from the indwelling needle used to puncture the arterial limb of the arteriovenous fistula or directly from the central intravenous catheter after discarding the volume corresponding to the priming capacity of the device. After collection, the syringe was stored on ice and collectively transported to the laboratory within one hour.

Parameters and estimates
We collected all clinical data and routine laboratory tests from patients' charts. Blood pressure values were taken as the mean of pre-dialysis measurements of the last three dialysis sessions before enrollment. Routine laboratory tests were extracted from the patient's chart and calculated   mean dialysis vintage was 58 ± 55 months; twenty--eight percent were smokers. The laboratory data of the participants are shown in Table 3. The mean total CO 2 (tCO 2 ) was 22.7 ± 3.0 mEq/L, and 40.3% of patients had levels < 22 mEq/L. The prevalence rates of the assigned ranges of tCO 2 in the whole sample and each dialysis center are shown in Table 4. A logistic regression model in which the dependent variable was tCO 2 values < 22 mEq/L is presented in Table 5. Both higher age and higher standard Kt/V were found to be associated with a reduced risk for tCO 2 values < 22 mEq/L; the dialyzer utilization number in the day of the blood gas analysis and the use of low-flux membrane dialyzers were associated with an increased risk of low tCO 2 .

dIscussIon
Patients on hemodialysis have a high mortality rate 2 with several studies indicating a relationship between metabolic acidosis and mortality. 3,4,5,6,12 We aimed to determine the prevalence of metabolic acidosis in hemodialysis patients and to analyze the factors associated with low bicarbonate levels in a sample derived from four centers of the metropolitan area of Rio de Janeiro, Brazil.
Fifty percent of the participants of the present study were receiving treatment form the Public Health System, which is markedly lower than the national population under publicly funded treatment (82%). 9 This aspect may have affected the dialysis prescription: for instance, the fraction of patients on more than 4 sessions of dialysis per week in the present study (10.2%) was strikingly higher than the national average (2.4%). 9 Perhaps as a reflex of the substantial heterogeneity regarding the source of funding, centers 1 and 2 (exclusively funded by private health insurance companies) had a higher fraction of patients on more than 4 sessions of dialysis per week and higher mean values for standard Kt/V. The most common dialysate calcium concentration was 3.0 mEq/L. Regarding dialysate bicarbonate, only center number tCO 2 ranges in mEq/L, n (%) All Dialysis centers    9 Judging from the mean values of Kt/V and body mass index, and from the mean serum levels of albumin, calcium, phosphorus, iPTH, and hemoglobin, participants received a dialysis treatment well within the standards of adequacy on average. 13 The mean pre-dialysis serum tCO 2 in a midweek session of our sample (22.7 mEq/L) was very similar to global data reported in 2 studies, 21.9 mEq/L 3 and 22.9 mEq/L. 4 As a whole, 40.3% of patients had serum bicarbonate levels below DOQI recommendation (22 mEq/L). 14 Values >27 mEq/L and < 18mEq/L, which had already been associated with higher mortality and higher hospitalization rate, 3,8 were found in 14.1% and 6.5% respectively. Once more, results varied between centers, with the highest mean value for tCO 2 in center 1, and the lowest one in center 4. Discrepancies in the dialysis prescription may again have accounted for these differences: for instance, the center with the lower mean tCO 2 values was the only one in which all patients underwent three dialysis sessions per week. A logistic regression model was designed to test associations with tCO 2 below 22 mEq/L, which included the following variables: age in decades, gender, diabetes, body mass index, hemoglobin, blood flow, dialyzer use count, use of low-flux membrane dialyzers, hours of dialysis per week, and standard Kt/V. Four variables in the model showed a statistically significant association with a blood tCO 2 below 22 mEq/L: older age and higher standard Kt/V were protective factors whereas higher dialyzer use count and the use of low-flux membrane dialyzers were associated with increased risk.
The negative impact of dialyzer reuse and the use of low flux membrane dialyzer on serum bicarbonate, even after extensive adjustment, did come as a surprise and should await further confirmation. The dialyzer reuse does not seem to affect the effectiveness of dialysis and is a widely used practice in countries with restricted resources. 15,16 In addition, although some studies show that the all-cause mortality in patients undergoing high flux dialysis may be lower compared to low flux hemodialysis, 15,17 this subject is still a matter of controversy. 18 The association of age and higher levels of bicarbonate had already been reported 3,18 and can be interpreted as a consequence of the lower acid generation in the dialysis interval due to the sarcopenia and reduced physical activity that can affect older hemodialysis patients. Although higher values for Kt/V 3,19 and duration of dialysis session 18 have been individually associated with higher levels of serum bicarbonate, we could not find studies addressing the role of the standard Kt/V in this regard. However, higher values of standard Kt/V can conceivably be associated with higher serum bicarbonate values taking in consideration that its calculation requires the last two mentioned variables (the Kt/V itself and the weekly frequency of dialysis).
The study presents some limitations, the main one being a relatively small sample not representative of the Brazilian population. These limitations restrict the generalizability of the findings; however, the information generated regarding this important and unexplored subject in Brazil, the metabolic acidosis in maintenance hemodialysis, is of great value.

conclusIons
The mean pre-dialysis serum tCO 2 in a midweek session of the studied sample was 22.7 mEq/L. The prevalence rate of serum bicarbonate below DOQI recommendation (22mEq/L or higher) was 40.3%, and 6.5% had serum bicarbonate below 18 mEq/L. The dialyzer use count and the use of low-flux membrane dialyzers were negatively associated with serum bicarbonate levels. Older age and higher standard Kt/V values were positively associated with serum bicarbonate levels.
Our findings conform to global data reported in previous studies but a more comprehensive study addressing national data is necessary to substantiate our findings.

Author's contrIbutIon
Jorge Paulo Strogoff de Matos and Jocemir Ronaldo Lugon equally contributed to the design of the study and final reviewing of the manuscript. Ana Paula Ramos Silva was involved with data collection and manuscript writing.