Hemodialysis patients receiving a greater Kt dose than recommended have reduced mortality and hospitalization risk
2016; Elsevier BV; Volume: 90; Issue: 6 Linguagem: Inglês
10.1016/j.kint.2016.08.022
ISSN1523-1755
AutoresFrancisco Maduell, Rosa Ramos, Javier Varas, Alejandro Martín‐Malo, Manuel Molina, Rafael Pérez‐García, Daniele Marcelli, Francesc Moreso, Pedro Aljama, José Ignacio Merello,
Tópico(s)Acute Kidney Injury Research
ResumoAchieving an adequate dialysis dose is one of the key goals for dialysis treatments. Here we assessed whether patients receiving the current cleared plasma volume (Kt), individualized for body surface area per recommendations, had improved survival and reduced hospitalizations at 2 years of follow-up. Additionally, we assessed whether patients receiving a greater dose gained more benefit. This prospective, observational, multicenter study included 6129 patients in 65 Fresenius Medical Care Spanish facilities. Patients were classified monthly into 1 of 10 risk groups based on the difference between achieved and target Kt. Patient groups with a more negative relationship were significantly older with a higher percentage of diabetes mellitus and catheter access. Treatment dialysis time, effective blood flow, and percentage of on-line hemodiafiltration were significantly higher in groups with a higher dose. The mortality risk profile showed a progressive increase when achieved minus target Kt became more negative but was significantly lower in the group with 1 to 3 L clearance above target Kt and in groups with greater increases above target Kt. Additionally, hospitalization risk appeared significantly reduced in groups receiving 9 L or more above the minimum target. Thus, prescribing an additional 3 L or more above the minimum Kt dose could potentially reduce mortality risk, and 9 L or more reduce hospitalization risk. As such, future prospective studies are required to confirm these dose effect findings. Achieving an adequate dialysis dose is one of the key goals for dialysis treatments. Here we assessed whether patients receiving the current cleared plasma volume (Kt), individualized for body surface area per recommendations, had improved survival and reduced hospitalizations at 2 years of follow-up. Additionally, we assessed whether patients receiving a greater dose gained more benefit. This prospective, observational, multicenter study included 6129 patients in 65 Fresenius Medical Care Spanish facilities. Patients were classified monthly into 1 of 10 risk groups based on the difference between achieved and target Kt. Patient groups with a more negative relationship were significantly older with a higher percentage of diabetes mellitus and catheter access. Treatment dialysis time, effective blood flow, and percentage of on-line hemodiafiltration were significantly higher in groups with a higher dose. The mortality risk profile showed a progressive increase when achieved minus target Kt became more negative but was significantly lower in the group with 1 to 3 L clearance above target Kt and in groups with greater increases above target Kt. Additionally, hospitalization risk appeared significantly reduced in groups receiving 9 L or more above the minimum target. Thus, prescribing an additional 3 L or more above the minimum Kt dose could potentially reduce mortality risk, and 9 L or more reduce hospitalization risk. As such, future prospective studies are required to confirm these dose effect findings. Adequate dialysis dose is one of the most important goals in hemodialysis (HD) treatment and should be appropriately prescribed. Achieving a minimum dialysis dose is the responsibility of nephrologists and represents an area open to improvement. Because age, gender, and comorbidity cannot be changed, dialysis parameters should be adjusted to ensure that the patient receives the optimal treatment. Several clinical practice guidelines1NKF-DOQI Clinical Practice Guideline for hemodialysis adequacy: 2015 update.Am J Kidney Dis. 2015; 66: 884-930Abstract Full Text Full Text PDF PubMed Scopus (613) Google Scholar, 2European Best Practice Guidelines for Haemodialysis. Nephrol Dial Transplant 2002;17(Suppl 7):17–21.Google Scholar, 3The Canadian Society of Nephrology: Clinical practice guidelines the delivery of haemodialysis.J Am Soc Nephrol. 1999; 10: S306-S310PubMed Google Scholar, 4Greenwood R, Tomson C, Hoenich N. Haemodialysis – clinical standards and targets. In: The Renal Association, Royal College of Physicians of London, eds. Treatment of Adult and Children with Renal Failure. Standards and Audit Measure. Third edition, Chapter 3. London: The Lavenham Press Ltd; 2002:19–35.Google Scholar, 5Maduell F. García M. Alcázar R. Dosificación y adecuación del tratamiento dialítico. Guías SEN: Guías de Centros de hemodiálisis.Nefrología. 2006; 26: 15-21PubMed Google Scholar have recommended a minimum Kt/V or urea reduction ratio (URR) as methods for monitoring dialysis dose. Because the urea kinetic method requires pre- and post-dialysis urea determinations, monitoring is performed monthly, bimonthly, or quarterly, and the result of this 3% to 7% of total sessions is extrapolated to the totality of the treatments. Given the relevance of dialysis dose to survival and that multiple factors can influence dialytic efficacy in each session, it seems reasonable to incorporate biosensors to quantify the dose in each session and in real time. Most monitors have incorporated ionic dialysance (ID), which allows calculation of dialysis dose in all sessions, without involving any additional workload, analytical determinations, or cost.6Peticlerc T. Bene B. Jacobs C. et al.Non-invasive monitoring of effective dialysis dose delivered to the haemodialysis patient.Nephrol Dial Transplant. 1995; 10: 212-216PubMed Google Scholar Consequently, many dialysis units have already abandoned urea determinations. In 1999, Lowrie et al.7Lowrie E.G. Chertow G.M. Lew N.L. et al.The urea {clearance x dialysis time} product (Kt) as an outcome-based measure of hemodialysis dose.Kidney Int. 1999; 56: 729-737Abstract Full Text Full Text PDF PubMed Scopus (141) Google Scholar proposed Kt as a method of monitoring dialysis dose and mortality. These authors observed a J-shaped survival curve when they distributed the patients into quintiles from the smallest to the highest URR, while the curve descended with Kt for the same patients.8Chertow G.M. Owen W.F. Lazarus J.M. et al.Exploring the reverse J-shaped curve between urea reduction ratio and mortality.Kidney Int. 1999; 56: 1872-1878Abstract Full Text Full Text PDF PubMed Scopus (111) Google Scholar In 2005, the minimum Kt dose was individualized according to body surface area (BSA)9Lowrie E.G. Li Z. Ofsthun N.J. Lazarus J.M. The online measurement of hemodialysis dose (Kt): Clinical outcome as a function of body surface area.Kidney Int. 2005; 68: 1344-1354Abstract Full Text Full Text PDF PubMed Scopus (62) Google Scholar and validated in a further study.10Lowrie E.G. Li Z. Ofsthun N.J. Lazarus J.M. Evaluating a new method to judge dialysis treatment using online measurements of ionic clearance.Kidney Int. 2006; 70: 211-217Abstract Full Text Full Text PDF PubMed Scopus (36) Google Scholar Since 2006, the Guidelines of the Spanish Society of Nephrology5Maduell F. García M. Alcázar R. Dosificación y adecuación del tratamiento dialítico. Guías SEN: Guías de Centros de hemodiálisis.Nefrología. 2006; 26: 15-21PubMed Google Scholar have proposed that dialysis centers with dialysis machines that have ionic dialysance use Kt to monitor dialysis dose. The Optimizing Results in Dialysis research initiative began in 2010 with the aim of improving HD patient outcomes by elucidating patient characteristics and practice of care in Spain.11Aljama P. ORD Work and Initiative Group ("Optimising Results in Dialysis").Nefrología. 2012; 32: 701-703PubMed Google Scholar In a previous retrospective study published by this group,12Maduell F. Ramos R. Palomares I. et al.ORD GroupImpact of targeting Kt instead of Kt/V.Nephrol Dial Transpl. 2013; 22: 2595-2603Crossref Scopus (18) Google Scholar monitoring the dialysis dose with Kt instead Kt/V was evaluated. The authors concluded that the advantage of this method is that it identifies 25.8% of patients who did not reach the minimum Kt while achieving Kt/V. This difference was particularly evident in women, patients with low body weight, and those with a venous central catheter (VCC). To define and validate the minimum Kt recommendations in the current Spanish dialysis population, therefore, the aim of the Optimizing Results in Dialysis research initiative was to design this prospective, observational, multicenter study. The goal was to assess whether patients receiving the current recommendations for an adequate dialysis dose by Kt individualized for BSA had improved survival and reduced hospitalizations at 2 years compared to those who did not. In addition, we assessed whether patients receiving a greater dose experienced greater benefit. During the recruitment period, 8095 patients were assessed for eligibility in 65 Fresenius Medical Care (FMC) Spanish facilities. Following the inclusion criteria, 6129 patients were subsequently included in this prospective study (Supplementary Table S1). The mean age was 68.9 ± 14 years, 62.3% were male, 65.5% had cardiovascular risk factors, 36.4% had diabetes mellitus, and the mean Charlson index was 5.5 ± 1.9. Vascular access through an autologous arteriovenous fistula (AVF) was present in 67.9% of patients, through a prosthetic arteriovenous fistula in 3.8%, and through VCCs in 28.3%. Patients were categorized monthly into 1 of 10 risk groups based on the difference between achieved Kt and target Kt (achieved–targetKt). The influences of patient characteristics per each dialysis dose group are summarized in Table 1. The patient groups that had a more negative relationship with the achieved–targetKt were significantly older and had a higher percentage of diabetes mellitus, higher comorbidity index, and higher percentage of catheters as vascular access. By contrast, the groups were balanced for gender and for the percentage of patients with cardiovascular risk.Table 1Patient characteristics and baseline parameters in the 10 groups based on Kt individualized or body surface areaachieved–targetKt categories (L)NumberAge (yr)Gender (% female)CV risk (%)DM (%)CIWeight (kg)VA (%)Kt/V (median)Kt/V on target< –820372.2 (11.8)36.9566.5046.806.1 (1.9)76.4 (19.6)80.791.26 (1.11–1.43)27.8%–8 to –427473.0 (11.6)41.2464.9644.166.0 (1.7)73.5 (16.9)77.371.45 (1.32–1.62)62.0%–4 to –143571.8 (12.8)40.4664.3742.305.9 (1.8)72.8 (15.0)71.031.57 (1.46–1.71)84.4%–1 to +143771.6 (13.4)40.7360.8745.545.9 (1.8)71.9 (15.8)62.701.65 (1.51–1.80)90.3%1 to +357970.4 (13.4)38.6965.8041.285.7 (1.9)71.6 (15.4)54.921.71 (1.53–1.89)91.8%3 to +6100069.8 (13.5)35.5064.6041.005.6 (1.9)70.9 (14.2)40.801.78 (1.61–1.97)97.0%6 to +9109368.6 (14.1)36.4167.0636.145.4 (1.9)69.6 (14.1)30.101.85 (1.68–2.07)98.2%9 to +1291868.0 (14.8)38.8964.6032.245.3 (1.9)67.7 (14.2)22.661.96 (1.76–2.22)98.9%12 to +1564166.1 (15.7)34.4869.8928.085.2 (2.0)66.7 (13.9)17.162.06 (1.84–2.31)99.7%> +1554963.1 (16.9)38.6264.8520.584.7 (1.9)63.2 (14.4)11.292.22 (1.93–2.51)99.8%P value<0.0010.250.21<0.001<0.001<0.001<0.001<0.001<0.001Data are presented as mean (SD), percentages (%), or median and interquartile range (25th–75th percentile).CI, Charlson Index (age adjusted); CV, cardiovascular; DM, diabetes mellitus; VA, venous central catheter. Open table in a new tab Data are presented as mean (SD), percentages (%), or median and interquartile range (25th–75th percentile). CI, Charlson Index (age adjusted); CV, cardiovascular; DM, diabetes mellitus; VA, venous central catheter. The mean effective dialysis duration (Td) was 240.3 ± 14 min, blood flow rate (QB) was 413 ± 64 ml/min, dialysate flow rate was 511±10 ml/min, and dry body weight was 69.7 ± 15 kg. Post-dilution online hemodiafiltration (HDF) treatments were performed for a total of 45.3% of the sessions, and the mean substitution volume was 24.2 ± 3.6 L. High-flux nonreusable dialyzers used were FX-50 (0.02%), FX CorDiax 60 (50.9%), FX CorDiax 600 (38.9%), FX CorDiax 80 or 800 (8.8%), FX CorDiax 1000 (0.2%), (Fresenius Medical Care, Bad Homburg, Germany), and Sureflux (Nipro, Osaka, Japan) 19UX or 21L (1.1%). The mean target Kt was 48.6 ± 3.8 L, and the mean achieved Kt was 55.1 ± 9.3 L. Moreover, the median achieved–targetKt was 6.4 (1.7–10.8) L, ranging from –20.1 to +30.5 L. In addition, the minimum target Kt dose was achieved in 80.5% of patients. Nevertheless, significant differences were found for the percentage of women who achieved the target Kt dose compared with men (79.0% vs. 81.5%; P < 0.001). Significant differences also were found between the percentage of sessions with the target Kt achieved and performed by VCC or AVF (59.2% vs. 88.9%; P < 0.001). Length of effective treatment dialysis time and QB were significantly longer in groups with higher achieved–targetKt (P < 0.001, both). In contrast, the lower adjusted Kt groups showed significantly lower percentages with the online HDF treatment option (P < 0.001) (Figure 1). The patients included in the project had average albumin levels from 3.6 to 3.9 g/dl. Despite the short range, significant differences were found between groups (Figure 2a). The corresponding post hoc Scheffé test revealed significant differences comparing the first achieved–targetKt group (lower than –8 L) to the 5 highest groups analyzed (from +3 L and up). Patients had median C-reactive protein levels of 5.0 (1.8–12.1) mg/l. ANOVA for the logarithmic-transformed variable indicated significant differences for the groups (P < 0.001), and the Scheffé test showed that the patients included in the 2 first achieved–targetKt groups (the <–8 L group and the –4 to –1 L group) had higher levels of C-reactive protein compared to the last 4 groups (+6 to +9 L group and up) (Figure 2b). This result is consistent with the higher percentage of catheters used in these groups (Table 1). Finally, the average hemoglobin values for each group ranged from 11.1 to 11.7 g/dl. Again, despite the narrowness of this range, ANOVA revealed significant differences among groups (P < 0.001). Of interest, the post hoc test showed that groups with better outcomes for achieved–targetKt had significantly lower levels of hemoglobin compared to the 3 highest groups (Figure 2c). To identify different predictors that could affect the nonachievement of the target Kt, unadjusted and adjusted generalized linear mixed models were constructed. In every case, these models were built considering repeated measures nested within patients. Moreover, the different Spanish FMC facilities involved in the study were introduced into the different models as a random effect component. The corresponding ORs (odds ratios) and 95% confidence intervals (CIs) are shown in Table 2. After the analysis of a total of 1,076,252 HD sessions, we found that all covariates recorded could affect the outcome. Among nonmodifiable factors, the presence of diabetes mellitus was the covariate associated with a higher risk of not achieving the target Kt, whereas VCC was identified among the modifiable factors. On the other hand, patients treated by HDF had a significantly higher probability of making the target Kt.Table 2Generalized Linear Mixed Models to identify fixed effects predictors with a potential negative impact on achieving the Kt individualized for body surface area targetUnivariateMultivariatePOR95% CIPOR95% CILowerUpperLowerUpperNon–modifiable factorsVintage (yr)Ref: <2.00––––––––2.00 – 2.990.9791.0000.9851.0140.0021.0271.0121.0433.00 – 4.990.1181.0100.9971.024<0.0011.0531.0391.0685.00 – 8.00<0.0011.0261.0131.039 8.00<0.0010.9600.9470.9720.0341.0151.0011.030GenderRef: female<0.0011.0551.0471.064<0.0011.0261.0121.043CV riskRef: No0.4801.0030.9951.011––––DMRef: No<0.0011.1401.1311.150<0.0011.0421.0331.052Age (yr)Ref: < 51.00––––––––51.00 – 60.00<0.0011.0861.0681.1040.0860.9850.9671.00261.00 – 70.00<0.0011.1131.0961.1300.0110.9790.9640.99571.00 – 80.00 80.00<0.0011.2321.2151.2500.6600.9860.9701.002Modifiable factorsHDFRef: HD<0.0010.6700.6650.675<0.0010.7870.7800.794VCCRef: AVF<0.0011.9121.8961.929<0.0011.3321.3181.346Weight (kg)Ref: < 58.00––––––––58.00 – 64.000.2680.9930.9801.006<0.0011.1091.0941.12465.00 – 72.00<0.0011.0421.0291.055<0.0011.1951.1791.21173.00 – 81.00<0.0011.0781.0641.092 81.00<0.0011.1901.1751.250<0.0011.5241.5031.546QB(ml/min)<0.0010.9940.9940.995<0.0010.9950.9940.995Td(min)<0.0010.9830.9820.984<0.0010.9840.9830.984Albumin (g/dl)< 3.50<0.0011.1631.1461.1740.0111.0181.0041.0313.50 – 4.00 4.00––––––––Hb (g/dl)< 10.00<0.0010.9390.9260.952<0.0010.9180.9050.93210.00 – 11.00<0.0010.9650.9550.952<0.0010.9620.9520.973Ref: 11.00 – 12.00––––––––12.00 – 13.00<0.0011.0851.0731.098 13.00<0.0011.2951.2781.313<0.0011.2211.2031.238CRP (mg/l)Ref: < 1.20––––––––1.20 – 3.400.6891.0030.9901.0160.7701.0020.9891.0153.40 – 7.00<0.0011.0531.0401.0670.0021.0211.0081.0357.00 – 15.00<0.0011.0791.0651.093 15.00<0.0011.1031.0891.118<0.0011.0531.0391.068Generalized Linear Mixed Models considering repeated measures nested within patients; the different Spanish Fresenius Medical Care facilities were introduced in the model as a random effect component; link function: logarithmic.CRP, C–reactive protein; CV, cardiovascular; DM, diabetes mellitus; Hb, hemoglobin; HDF, hemodiafiltration; QB, blood flow; Td, effective treatment time; VCC, venous central catheter. Open table in a new tab Generalized Linear Mixed Models considering repeated measures nested within patients; the different Spanish Fresenius Medical Care facilities were introduced in the model as a random effect component; link function: logarithmic. CRP, C–reactive protein; CV, cardiovascular; DM, diabetes mellitus; Hb, hemoglobin; HDF, hemodiafiltration; QB, blood flow; Td, effective treatment time; VCC, venous central catheter. Patients were monitored for 2 years or until premature termination or death. During the observation period, 790 patients prematurely exited the study because of kidney transplantation (n = 685), change in dialysis unit (n = 62), or other reasons (n = 43). All of these patients were censored at the time of premature termination. There were 1004 deaths (16.4%) during the follow-up. The main causes of death were cardiovascular diseases (45.6%), infectious diseases (17.9%), sudden death (10.4%), gastrointestinal diseases (6.2%), tumor (5.2%), and other (14.7%). The independent predictors for all-cause mortality were identified exploring the nonlinear effects of the continuous variables. The corresponding univariate time-dependent Cox model is shown in Table 3. Moreover, the possible center effect on the outcome was identified as a nonsignificant covariate in a univariate analysis (Supplementary Table S2). The independent predictors for all-cause mortality were age, gender, cardiovascular risk, diabetes mellitus, VCC, weight, hemoglobin, albumin, C-reactive protein, and achieved–targetKt.Table 3Univariate and multivariate time dependent Cox regression analysis for all–cause mortality with variables grouped as nonmodifiable or modifiable factorsTime dependent Cox modelUnivariateMultivariatePHR95% CIPHR95% CILowerUpperLowerUpperNon–modifiable factorsGenderRef: female0.011.4071.0831.8280.0031.2401.0781.425CV RiskRef: No0.041.1201.1021.2780.8881.010.8791.161DMRef: No<0.0011.3221.1661.4980.0041.2141.0631.386Age (yr)Ref: < 51.0––––––––51.0 – 60.00.0251.6941.0682.6860.3451.2570.7822.01961.0 – 70.0<0.0013.2502.1754.858<0.0012.2671.5073.41271.0 – 80.0<0.0014.5313.0796.668 81.0<0.0016.3644.3449.323<0.0013.5202.3795.208Modifiable factorsachieved–targetKt(L)<0.0010.9500.9420.958<0.0010.9450.9330.957HDFRef: HD0.010.8450.7420.9610.1100.8830.7581.029VCCRef: AVF<0.0011.4991.3241.6970.0361.1731.0111.362Weight (kg)Ref: < 58.0––––––––58.0 – 64.0<0.0010.7210.6050.8600.0030.7570.6300.91065.0 – 72.0<0.0010.6400.5380.762<0.0010.6040.5000.72973.0 – 81.0<0.0010.5070.4160.618 81.0<0.0010.4160.3390.511<0.0010.4150.3280.527Td(min)<0.0010.9830.9790.9870.3051.0030.9981.008Albumin (g/dl)< 3.5<0.0014.9054.0915.881<0.0012.3831.9392.9303.5 – 4.0 4.0––––––––Hb (g/dl)< 10.0<0.0016.1274.9707.553<0.0015.1884.1516.48510.0 – 11.0<0.0011.6721.4371.945<0.0011.6421.4061.918Ref: 11.0 – 12.0––––––––12.0 – 13.0<0.0011.4241.1721.732 13.00.0321.3291.0241.7260.0841.2770.9671.686CRP (mg/l)Ref: < 1.2––––––––1.2 – 3.40.3811.1680.8251.6530.7281.0640.7501.5103.4 – 7.00.0341.4261.0271.9790.3781.1610.8331.6197.0 – 15.00.0021.6221.1862.2190.3501.1640.8461.603>15.0<0.0013.5962.6724.841<0.0012.0321.4932.766The plausible center effect was identified as a nonsignificant variable in a univariate analysis.CRP, C-reactive protein; CV, cardiovascular risk; DM, diabetes mellitus; Hb, hemoglobin; HDF, hemodiafiltration; Td, effective treatment time; VCC, venous central catheter. Open table in a new tab The plausible center effect was identified as a nonsignificant variable in a univariate analysis. CRP, C-reactive protein; CV, cardiovascular risk; DM, diabetes mellitus; Hb, hemoglobin; HDF, hemodiafiltration; Td, effective treatment time; VCC, venous central catheter. To assess whether patients receiving an adequate dialysis dose could have a reduced mortality risk, unadjusted and adjusted time-dependent Cox analyses were performed. All covariates that were previously identified as independent predictors for all-cause mortality were introduced into the Cox adjusted model. The reference group selected was the one that incorporated the null value (achieved–targetKt –1 to +1 L/treatment). Results with hazard ratios (HRs) for each group are shown in Figure 3. The mortality risk profile showed a significant trend (P < 0.001) toward a progressive increase in mortality risk with increasingly negative achieved–targetKt. The HR for the lower achieved–targetKt group was associated with significantly higher mortality risk (univariate HR, 1.69; 95% confidence interval [CI], 1.25 to 2.28; P < 0.001; multivariate HR, 1.95; 95% CI, 1.41 to 2.71; P < 0.001). Of interest, the HR fell as the achieved–targetKt parameter moved into the positive range and became significantly lower at the +1 to +3 L group and up for both univariate and adjusted models. From +9 L and up, the mortality risk profile remained flattened and significantly lower than the reference group (Figure 3). Further Cox models were used for the analysis of different causes of death that were identified as represented in >10% of cases. We found significant differences for these outcomes (Table 4). The risk appeared to be lower as the achieved–targetKt parameter moved into the positive range for cardiovascular, sudden death, and infection-related mortality, from +6 to +9 L and up.Table 4Primary outcome: mortalityDeath cause and number of eventsachieved–targetKtUnivariateMultivariateP valueHR95.0% CIP valueHR95.0% CILowerUpperLowerUpperDeath from any cause1004< –80.0011.6881.2512.277<0.0011.9541.4052.717–8 to –40.2831.1810.8721.5980.6581.0760.7791.486–4 to –10.4520.8980.6781.1890.6190.9300.6971.240–1 to +1Ref.–––Ref.–––+1 to +30.0080.6860.5190.9060.0080.6780.5080.904+3 to +60.0010.6490.5070.8300.0220.7450.5790.959+6 to +9<0.0010.5630.4390.721<0.0010.5830.4500.755+9 to +12<0.0010.4900.3770.637<0.0010.4850.3680.640+12 to +15<0.0010.3940.2910.533 +15<0.0010.4200.3070.574<0.0010.3640.2590.511Cardiovascular cause458 (45.62%)< –80.4191.2250.7492.0030.2191.3950.8212.372–8 to –40.4391.1940.7621.8710.4901.1850.7321.917–4 to –10.5700.8860.5831.3460.9520.9870.6411.519–1 to +1Ref.–––Ref.–––+1 to +30.0670.6770.4471.0270.1190.7110.4641.092+3 to +60.0560.7040.4911.0090.1720.7710.5311.119+6 to +90.0010.5370.3700.7790.0020.5330.3600.789+9 to +120.0010.5170.3520.761<0.0010.4570.2990.697+12 to +15<0.0010.4280.2760.663 +150.0010.4840.3100.755<0.0010.3460.2050.584Infection180 (17.93%)< –80.0302.0461.0723.9060.0312.0771.0704.032–8 to –40.9190.9620.4612.0090.3240.6610.2901.504–4 to –10.2331.4180.7992.5170.4171.2750.7092.294–1 to +1Ref.–––Ref.–––+1 to +30.5600.8350.4561.5300.5790.8400.4541.555+3 to +60.0210.4970.2740.8990.0950.6010.3311.093+6 to +90.0160.4900.2740.8780.0250.5090.2820.920+9 to +120.0030.3860.2040.7310.0050.3960.2080.753+12 to +15 +150.0010.2430.1030.5750.0010.2100.0830.530Sudden death104 (10.36%)< –80.2311.7460.7024.3400.3541.6070.5894.385–8 to –40.0742.0780.9314.6380.1511.8490.7984.282–4 to –10.2440.5540.2051.4980.1890.4910.1701.419–1 to +1Ref.–––Ref.–––+1 to +30.3890.6860.2911.6160.2180.5630.2261.405+3 to +60.3650.7090.3381.4910.3900.7170.3361.532+6 to +90.0320.4090.1810.9280.0200.3740.1630.857+9 to +120.1240.5380.2441.1860.0340.4100.1800.935+12 to +150.0210.3100.1150.8390.0030.2070.0730.593> +150.0300.3090.1070.8900.0020.1680.0540.528The variables included in the multivariate analysis were gender, cardiovascular risk, diabetes mellitus, age, treatment option, vascular access, weight, effective treatment time, albumin, hemoglobin, C-reactive protein and Spanish Fresenius Medical Care facilities. Variables not deemed significant were excluded from final analysis. Open table in a new tab The variables included in the multivariate analysis were gender, cardiovascular risk, diabetes mellitus, age, treatment option, vascular access, weight, effective treatment time, albumin, hemoglobin, C-reactive protein and Spanish Fresenius Medical Care facilities. Variables not deemed significant were excluded from final analysis. Reviewing the demographic features of the study groups (Table 1), it would be expected that the more elderly and frail groups and those with catheter access would fail to achieve target Kt (Table 2) and therefore have greater mortality rates. Thus, although the previous adjusted Cox models included all of the independent mortality predictors collected for this study, we cannot exclude a reverse-causality phenomenon. To address this problem, we decided to use a propensity score matching approach. We combined the 10 study categories into 2 population groups according to baseline HD dose: equal to or less than the reference dose group (achieved–targetKt –1 to +1 L) and greater than the group with reduced mortality according to our predictive models (+1 to +3 L). We tried to balance these 2 resulting populations for every covariate described as predictors with an impact on achieving the target Kt (Table 2). The resulting balanced population was properly assessed (Supplementary Table S3), and the adjusted cohort was used to estimate the corresponding mortality risk. The Cox model indicated that having an HD dose defined as achieved–targetKt above +1 L was associated with a 25% reduction in risk for all-cause mortality compared to the lower HD group (< +1 L) (HR, 0.753; 95% CI, 0.578–0.979; P = 0.034). There were 6939 hospital admissions for 3042 of the total included patients (49.6%) during the follow-up. The main causes of hospital admissions were vascular access (15.3%), cardiovascular diseases (16.5%), infectious diseases (19.2%), digestive diseases (10.6%), respiratory diseases (6.0%), and other (32.4%). Again, a negative trend (P < 0.001) in hospitalization risk with the progressive increases in dialysis dose was identified (Figure 4). In both models, the HR increased progressively as the achieved–targetKt groups became more negative. On the other hand, as occurred in the mortality risk models, the hospitalization risk appeared to decrease until the last achieved–targetKt group, becoming significant with the +6 to +9 L group in the univariate model and with the +9 to +12 L group in the multivariate model. In this study, the current recommendations for monitoring dialysis dose with Kt individualized for BSA were validated in the current Spanish dialysis population and found to be suitable. A dose higher or lower than the recommended minimum is predictably associated with lower or higher death and hospitalization risk. To our knowledge, this is the first prospective trial showing that prescribing more than 1 to 3 L of the current Kt individualized for BSA recommendations reduces mortality risk and more than 9 L reduces hospitalization risk. Thus, both adjusted and unadjusted risk profile results indicated that an adequate dialysis dose as measured by Kt individualized for BSA improved the morbidity–mortality rate in HD patients. Traditionally, dialysis dose recommendations are based on analytical pre- and post-dialysis urea determinations. US, European, Canadian, United Kingdom, and Spanish guidelines1NKF-DOQI Clinical Practice Guideline for hemodialysis adequacy: 2015 update.Am J Kidney Dis. 2015; 66: 884-930Abstract Full Text Full Text PDF PubMed Scopus (613) Google Scholar, 2European Best Practice Guidelines for Haemodialysis. Nephrol Dial Transplant 2002;17(Suppl 7):17–21.Google Scholar, 3The Canadian Society of Nephrology: Clinical practice guidelines the delivery of haemodialysis.J Am Soc Nephrol. 1999; 10: S306-S310PubMed Google Scholar, 4Greenwood R, Tomson C, Hoenich N. Haemodialysis – clinical standards and targets. In: The Renal Association, Royal College of Physicians of London, eds. Treatment of Adult and Children with Renal Failure. Standards and Audit Measure. Third edition, Chapter 3. London: The Lavenham Press Ltd; 2002:19–35.Google Scholar, 5Maduell F. García M. Alcázar R. Dosificación y adecuación del tratamiento dialítico. Guías SEN: Guías de Centros de hemodiálisis.Nefrología. 2006; 26: 15-21PubMed Google Scholar recommend a minimum Kt/V of 1.2 and/or a URR of 65% (or Kt/V of 1.3 and a URR of 70% to ensure that these minimum requirements are reached). If urea measurements are carried out only monthly, bimonthly, or quarterly to calculate the dialysis dose,
Referência(s)