Comparison of outcomes by modality for critically ill patients requiring renal replacement therapy: a single-centre cohort study adjusting for time-varying illness severity and modality exposure.
Prolonged intermittent renal replacement therapy (PIRRT) is a more recently defined acute modality that is also known as extended (daily) dialysis or sustained low-efficiency (daily) dialysis/diafiltration (9-16). PIRRT involves the application of conventional iHD machinery with reduced blood and dialysate flow rates for prolonged periods. PIRRT combines the superior detoxification and haemodynamic stability of CRRT with the operational convenience and low cost of iHD. However, there are fewer outcomes studies with PIRRT than there are with are CRRT or iHD. The cohort studies that do exist are often confounded by selection bias, time-varying modality exposure and time-varying illness severity, factors that can all lead to potentially incorrect estimates of mortality risk (12,17,18). A suitably powered randomised clinical trial would provide definitive outcome data, but is unlikely given the other more pressing priorities for research within critical care nephrology.
A number of advanced statistical techniques have recently been described which may improve the validity of cohort studies in this area. Using one such technique, we studied critically ill patients treated with renal replacement therapy in our centre and compared mortality risk by modality. We used the Marginal Structural Modelling (MSM) approach of Robins to adjust for multiple confounders including time-varying modality exposure, and also time-varying illness severity that is both affected by previous modality exposure while also affecting subsequent modality choice (19). We present partially and fully adjusted estimates of the mortality risks associated with different acute modalities in our centre, including PIRRT.
We performed an inception cohort study, using an as-treated framework ("did the modality that the patient actually received affect outcomes?"), as opposed to an intention-to-treat framework ("did the initial modality that the patient receive affect outcomes, irrespective of subsequent changes that occurred along the way?") (20). Data were sourced from patients' hospital and primary care clinical records. We performed analyses of repeated measures within patients over time.
Setting and participants
We created an inception cohort of all adult renal replacement therapy patients (>16 years) admitted to the intensive care unit (ICU) of Middlemore Hospital, Counties Manukau District Health Board, Auckland, New Zealand from January 2002 to December 2008. Middlemore Hospital is a 900-bed tertiary referral centre for plastic surgery, burns, orthopaedics and a range of medical sub-specialities. Any patient requiring neurosurgical or cardio-thoracic surgical intervention is referred on; all other patients remain at Middlemore Hospital. Currently, the ICU in the hospital is a 12 funded-bed level 3 facility (21). Patients were followed until death, hospital discharge or 90 days post renal replacement therapy inception, whichever occurred first. Censoring was at the time of loss to follow-up. We included patients with AKI or acute-on-chronic kidney disease as defined by the 'Failure' stage of the RIFLE (Risk, Injury, Failure, Loss, End-stage kidney disease) criteria (22), and excluded those with end-stage kidney failure on maintenance dialysis and those with normal renal function requiring renal replacement therapy for treatment of poisoning. Of note, we only analysed a patient's first ICU admission if they were re-admitted to the ICU during the same hospitalisation. This is a standard approach and avoids counting more than one outcome for the same patient.
Primary exposure and outcome variables
The primary exposure was time-varying acute modality: PIRRT, CRRT or iHD (9,23). All treatments in our centre irrespective of modality were delivered via standard veno-venous access, with unfractionated heparin for anticoagulation, using synthetic haemofilters and haemodialysers (polysulfone, polyamide, AN-69). PIRRT was performed using Fresenius 4008S ARrT-Plus or 5008 machines (Fresenius Medical Care (New Zealand) Pty Ltd, Auckland, New Zealand), using blood flow rate (QB) of 200 to 300 ml/minute, with counter-current bicarbonate-based dialysate with flow rate (QD) of 200 ml/minute, and haemofiltration between 25 to 100 ml/minute in prefilter mode. Treatment duration was prescribed at between 8 to 10 hours on a daily or alternate day basis as clinically indicated for solute or fluid control. On-line replacement fluid was produced using portable reverse osmosis water and at two in-line dialysate filters, a technique that has been validated previously. Ultrapure water quality was ensured by regular endotoxin and microbiological testing according to standard guidelines (ISO 11663 and ISO 13959--www.iso.org, AAMI RD52:2004/A3--www.aami.org). The dialy-sate used constituent concentrations (mmol/l) as clinically appropriate: [[Na.sup.+]] 140-145, [[K.sup.+]] 3.0-4.3, [[Ca.sup.2+]] 1.25 and [HC[O.sub.3.sup.-]] 26-32.
CRRT was performed using Baxter Aquarius (Baxter Healthcare New Zealand, Auckland, New Zealand) or Prisma machines (Gambro New Zealand Pty Ltd, Botany, New Zealand) by haemo-filtration or haemodiafiltraion using bicarbonate-based replacement fluid (Hemosol B0) in postdilution mode. (QB ranged from 150 to 200 ml/ minute, with pump-controlled ultrafiltration or dialysis at 1500 to 2500 ml/hour).
iHD was performed using Gambro AK200S machines with QB 250 to 300 ml/minute, and counter-current bicarbonate-based dialysate with QD 500 to 700 ml/minute. Treatment duration was prescribed at between four and six hours on a daily or alternate day basis as clinically indicated for solute or fluid control. Dialysate was produced using portable reverse osmosis water and one in-line dialysate filter. Ultrapure water quality was ensured by regular endotoxin and microbiological testing according to standard guidelines. The dialysate used constituent concentrations (mmol/l) as clinically appropriate: [[Na.sup.+]] 140-145, [[K.sup.+]] 2-3, [[Ca.sup.2+]] 1.25 and [HC[O.sub.3.sup.-]] 35-40.
The primary outcome was patient death determined at hospital discharge. The secondary outcome was patient death determined at 90 days post renal replacement therapy inception.
Data measurement and quantitative variables
We adjusted for known risk factors as recorded in patients' clinical records: baseline illness severity and co-morbidity as measured by the Acute Physiological and Chronic Health Evaluation (APACHE) IV score and predicted risk of death (24), time-varying illness severity as measured by daily Sepsis-Related Organ Failure Assessment (SOFA) scores, time from ICU admission to renal replacement therapy inception (25), and year of admission to account for any secular variation.
Data for analyses consisted of T daily waves for each of n patients for a total of n*T patient-days. Each wave contained the primary exposure and outcome and also potential confounders for that day. Variables were modelled as baseline or time-varying as appropriate by carrying forward either patients' first or most recent observations respectively.
We used the MSM technique of Robins to adjust for time-varying modality exposure and illness severity (19,26-29). MSMs relate potential outcomes to exposure history, using weights that redistribute the population so that the primary exposure is unconfounded. In our study the main source of confounding is illness severity. CRRT is used rather than PIRRT and IHD for patients with higher degrees of illness severity and adjustment is necessary to avoid selection bias. However, illness severity is also an intermediary variable on the causal pathway to death; modality exposure may affect the development of new haemodynamic instability, which is an independent prognostic factor for death. Controlling for illness severity alone would result in biased estimates of the effect of modality, as any effect of modality exposure acting via illness severity would be lost (19,30,31). In our analyses, weights remove associations between dialysis modality and prior illness severity, and also between prior modality exposure and the development of worsening illness severity.
Using this technique, we calculated the hazard ratio for mortality for each of the three possible acute modalities. The estimated hazard functions can be interpreted as individuals' mortality outcomes had they (possibly contrary to fact) been continuously treated with the given modality. Results from such 'counterfactual' analyses more closely approximate those of clinical trials than standard regression models (19,32-35).
We estimated MSM parameters by pooled multinomial logistic regression using programming that has previously been published (36). We computed inverse-probability-of-treatment weights as the patients' probability of having the acute modality that they actually had on day t, as a function of predictive baseline and time-varying covariates (measured at baseline, t and t - 24 hours), carrying forward observations in the first day. Stable versions of these weights were computed to avoid bias/imprecision as a result of over-inflated weights from individuals with estimated probabilities very close to 0. We specified MSMs in a manner to meet the assumption of experimental treatment assignment (or positivity). We assessed for model misspecification (leading to incorrect inverseprobability-of-treatment weights) from the distribution of weights (37,38). We used a 24-hour lag for attribution of outcomes to modality in the main-effects model.
We computed inverse-probability-of-treatment weights using all potentially confounding variables to avoid the risk of eliminating any that might be true confounders. For the main-effects MSMs, we considered all potentially confounding variables initially and estimated reduced models after removing those with P <0.2 jointly adjusted for other covariates. Statistical significance was assessed at a level of P <0.05.
We used Stata Intercooled MP/11.2 for all analyses (StataCorp, College Station, TX, USA).
Patient and outcome data
Approval for this study was granted by the National (New Zealand) Health and Disability Ethics Committee under the provisions made for clinical audit. We identified an inception cohort of 151 patients with a total of 655 treatment-days. One hundred and forty-six with a total of 633 treatment-days had sufficient data for modelling; five patients and 20 patient-days were omitted because of missing data or because of a second ICU admission in the same patient (n=1). No patient was lost to follow-up. We recorded 49 ICU deaths, 59 hospital deaths and 69 deaths at day 90 post renal replacement therapy inception. We recorded dialysis dependence in 52 of the 97 ICU survivors, 15 of the 87 hospital survivors, and in eight of the 77 survivors at day 90 post renal replacement therapy inception.
Patient characteristics by modality exposure are provided in Table 1. In general, those treated with PIRRT and CRRT had higher degrees of illness severity than those treated with iHD.
Renal replacement therapy characteristics are shown in Table 2. The total number of cumulative days in the cohort between each patient's first and last renal replacement therapy treatments was 845 days. Allowing for 144 days of CRRT, the frequency of intermittent renal replacement therapy with PIRRT or iHD was one treatment per 1.45 days.
Main results are shown in Table 3. Unadjusted models show a trend to lower associated hospital mortality risk for iHD relative to both PIRRT and CRRT. After adjustment for baseline and then time varying factors (year of treatment; tertial of lead-time category, baseline SOFA score category, and baseline APACHE IV risk of death category), there was a progressive increase in the mortality risk associated with iHD. In the final fully adjusted model, there were no apparent trends to differences in outcomes by modality.
Supplementary analysis of the secondary endpoint is shown in Table 4. Unadjusted models show a similar associated 90-day mortality risk for all modalities. However, adjustment for baseline and then time varying factors (as above) resulted in a marked increase in the mortality risk associated with iHD, without any marked change in those associated with PIRRT and CRRT. In the final fully adjusted model, there was a trend to higher 90-day mortality in those receiving iHD relative to both PIRRT and CRRT.
In this study, there were no observed differences in hospital mortality between those treated with the different modalities of renal replacement therapy, including PIRRT. This conclusion is consistent with randomised controlled trials comparing iHD and CRRT, and might be considered predictable given the results from these other studies (6,8,39). However, there are specific clinical concerns with PIRRT that need to be allayed by studies such as ours. There are concerns that PIRRT may not provide optimal outcomes for very haemodynamically unstable or shocked patients for whom it is frequently used. Moreover, exposure of critically ill patients for long periods to on-line dialysate (rather than sterile batched dialysate) might carry a risk of exposure to microbial contaminants thereby exacerbating inflammatory milieu. Finally, there are concerns that the relatively greater changes in solutes during PIRRT compared to other modalities might lead to threatening electrolyte abnormalities such as hypokalaemia and hypophosphataemia. Our findings suggest that these issues are not likely to be factors in the routine clinical use of PIRRT by experienced practitioners.
In comparing our findings to the literature, there are only a relatively small number of studies describing mortality risk with PIRRT. Several single-centre observational studies with PIRRT have reported hospital mortality to be at least comparable to that predicted by the APACHE II illness severity scoring system (11,13,15,17,23,40,41). Several single-centre clinical trials have compared hospital mortality between PIRRT and continuous therapies, and reported similar hospital mortality between those treated with PIRRT and those treated with CRRT (Table 5) (42-44). Multi-centre studies are rarer. One observational study used a time series approach in three centres from three different countries, and showed no change in hospital mortality over time with a change from CRRT to PIRRT as the predominant therapy for sicker patients9. Potentially the most definitive data was to come from the Stuivenberg Hospital Acute Renal Failure investigators, who provided interim analyses of a large, multi-centre, prospective, randomised clinical trial of 996 patients reporting a hospital mortality of 68% for those treated with PIRRT and 64% for those treated with CRRT (45,46). Unfortunately, the most recent publication from the Stuivenberg Hospital Acute Renal Failure investigators described the premature termination of their study, and it is likely that the results of this study will not be seen in final form (47).
As with most other studies, the group of patients who received CRRT in our study had the highest illness severity as shown by their SOFA scores and APACHE IV score and predicted risk of death. However, our entire patient population had a generally high degree of illness severity, including those who received PIRRT who had severe haemodynamic instability as indicated by their SOFA scores. Our experience reflects findings of a previous research including two small randomised clinical trials: in the hands of experienced practitioners, the effects of CRRT and PIRRT on blood pressure and other haemodynamic parameters are clinically in distinguishable (42,43,48-52).
In our study, there is a difference between the association between iHD and mortality risk at hospital discharge versus that at 90 days post renal replacement therapy inception. The endpoints of mortality at 90 days post randomisation (53,54) or at 60 days post randomisation (55,56) have been used in recent large clinical trials involving renal replacement therapy. The former is recommended by the United Kingdom Medical Research Council International Working Party for Clinical Trials in Patients with Sepsis and Septic Shock (57), although there is evidence that mortality for patients with AKI may reach a plateau at between 60 and 90 days (58). In cohort studies that use 'as treated' framework such as ours, the choice of hospital mortality as the primary end-point is important. In our study, the lower associated mortality risk for iHD at hospital discharge versus that at 90 days post renal replacement therapy inception reflects the prognosis of patients who are alive but on iHD at hospital discharge. These patients have delayed or non-recovery of AKI, and are thus at high mortality risk after hospital discharge. It could be argued that modality might affect recovery of renal function to some degree, and by acting through this mechanism, also affect late mortality. However, randomised clinical trials in this area have not shown a difference in renal recovery from AKI by modality (6-8). Our study suggests therefore, that mortality at hospital discharge is the most appropriate end-point for cohort studies that use an 'as treated' framework.
The potential for selection bias in our study is evident from the differences in baseline and time-varying illness severity scores between groups. In our study, selection bias is minimised in two ways. First, the routine utilisation of PIRRT is somewhat higher in our ICU than it is elsewhere, diminishing selection bias at a patient level. Second, we have made extensive statistical adjustments for risk factors as discussed. The major strength of our study is that it compares all three modalities of renal replacement therapy. Furthermore, the as-treated framework compares deaths in purely one modality with deaths in purely another modality. The intention-to-treat framework in other comparable studies gives biased results if patients change modalities, since the deaths attributed to the allocated modality are actually a mixture of deaths in all modalities (59,60). Our results are likely to be a closer approximation of the true 'causal' effect of modality on mortality, as might be determined from of a clinical trial randomising patients to different modalities.
As with all observational studies, associations do not prove causality. There are a number of limitations of this study. The most important shortcoming is the small number of patients, raising the possibility of type II error (i.e. failure to reject a false null hypothesis). The wide confidence intervals of statistical estimates allow for a large number of possible associations, including at worst a 1.5-fold higher associated mortality with PRRT relative to CRRT. Another limitation of this study is likely residual confounding from limited collection of co-morbidity outside of that already recorded in the APACHE IV score (e.g. no collection of functional status, or left ventricular function) and lack of data about blood pressure/ fluid volume status. Notwithstanding, clinical trials comparing outcomes by modality are notoriously difficult and results are often difficult to generalise. Cohort studies are therefore likely to remain a useful source of evidence to guide clinical practice, especially ones such as ours that use contemporary cohorts and state-of-the-art methods for statistical adjustment.
Our observations should be regarded as exploratory and hypothesis-generating, rather than confirmatory and definitive. Further study is needed in patient populations where there is less equipoise about PIRRT over CRRT, such as those with cardiogenic shock, fulminant hepatic failure (61) and brain injury. Acknowledging the significant limitations of our study and others, PIRRT appears to provide non-inferior clinical outcomes compared to CRRT in the general ICU setting.
(1.) Uchino S, Kellum JA, Bellomo R, Doig GS, Morimatsu H, Morgera S et al. Acute renal failure in critically ill patients: a multinational, multicenter study. JAMA 2005; 294:813-818.
(2.) Waikar SS, Curhan GC, Wald R, McCarthy EP, Chertow GM. Declining mortality in patients with acute renal failure, 1988 to 2002. J Am Soc Nephrol 2006; 17:1143-1150.
(3.) Ympa YP, Sakr Y, Reinhart K, Vincent J-L. Has mortality from acute renal failure decreased? A systematic review of the literature. Am J Med 2005; 118:827-832.
(4.) Basso F, Ricci Z, Cruz D, Ronco C. International survey on the management of acute kidney injury in critically ill patients: year 2007. Blood Purif 2010; 30:214-220.
(5.) Overberger P, Pesacreta M, Palevsky PM. Management of renal replacement therapy in acute kidney injury: a survey of practitioner prescribing practices. Clin J Am Soc Nephrol 2007; 2:623-630.
(6.) Rabindranath K, Adams J, Macleod AM, Muirhead N. Intermittent versus continuous renal replacement therapy for acute renal failure in adults. Cochrane Database Syst Rev 2007;CD003773.
(7.) Tonelli M, Manns B, Feller-Kopman D. Acute renal failure in the intensive care unit: a systematic review of the impact of dialytic modality on mortality and renal recovery. Am J Kidney Dis 2002; 40:875-885.
(8.) Pannu N, Klarenbach S, Wiebe N, Manns B, Tonelli M. Renal replacement therapy in patients with acute renal failure: a systematic review. JAMA 2008; 299:793-805.
(9.) Marshall MR, Creamer JM, Foster M, Ma TM, Mann SL, Fiaccadori E et al. Mortality rate comparison after switching from continuous to prolonged intermittent renal replacement for acute kidney injury in three intensive care units from different countries. Nephrol Dial Transplant 2011; 26:2169-2175.
(10.) Lonnemann G, Floege J, Kliem V, Brunkhorst R, Koch KM. Extended daily veno-venous high-flux haemodialysis in patients with acute renal failure and multiple organ dysfunction syndrome using a single path batch dialysis system. Nephrol Dial Transplant 2000; 15:1189-1193.
(11.) Marshall MR, Golper TA, Shaver MJ, Alam MG, Chatoth DK. Sustained low-efficiency dialysis for critically ill patients requiring renal replacement therapy. Kidney Int 2001; 60:777-785.
(12.) Kumar VA, Craig M, Depner TA, Yeun JY. Extended daily dialysis: A new approach to renal replacement for acute renal failure in the intensive care unit. Am J Kidney Dis 2000; 36:294-300.
(13.) Naka T, Baldwin I, Bellomo R, Fealy N, Wan L. Prolonged daily intermittent renal replacement therapy in ICU patients by ICU nurses and ICU physicians. Int J Artif Organs 2004; 27:380-387.
(14.) Berbece AN, Richardson RMA. Sustained low-efficiency dialysis in the ICU: cost, anticoagulation, and solute removal. Kidney Int 2006; 70:963-968.
(15.) Fiaccadori E, Maggiore U, Parenti E, Giacosa R, Picetti E, Rotelli C et al. Sustained low-efficiency dialysis (SLED) with prostacyclin in critically ill patients with acute renal failure. Nephrol Dial Transplant 2007; 22:529-537.
(16.) Bellomo R, Baldwin I, Fealy N. Prolonged intermittent renal replacement therapy in the intensive care unit. Crit Care Resusc 2002; 4:281-290.
(17.) Holt BG, White JJ, Kuthiala A, Fall P, Szerlip HM. Sustained low-efficiency daily dialysis with hemofiltration for acute kidney injury in the presence of sepsis. Clin Nephrol 2008; 69:40-46.
(18.) Wu VC, Wang C-H, Wang W-J, Lin Y-F, Hu F-C, Chen Y-W et al. Sustained low-efficiency dialysis versus continuous venovenous hemofiltration for postsurgical acute renal failure. Am J Surg 2010; 199:466-476.
(19.) Robins JM, Hernan MA, Brumback B. Marginal structural models and causal inference in epidemiology. Epidemiology 2000; 11:550-560.
(20.) Vonesh EF, Schaubel DE, Hao W, Collins AJ. Statistical methods for comparing mortality among ESRD pateints: examples of regional/international variations. Kidney Int 2000; 57 (Suppl 74):S19-27.
(21.) College of Intensive Care Medicine of Australia and New Zealand. Minimum standards for intensive care units. From http://cicm.org.au/policydocs.php. Accessed April 2010.
(22.) Bellomo R, Ronco C, Kellum JA, Mehta RL, Palevsky P. Acute renal failure--definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group. Crit Care 2004; 8:R204-212.
(23.) Marshall MR, Ma TM, Galler D, Rankin APN, Williams AB. Sustained low-efficiency daily diafiltration (SLEDD-f) for critically ill patients requiring renal replacement therapy: towards an adequate therapy. Nephrol Dial Transplant 2004; 19:877884.
(24.) Zimmerman JE, Kramer AA, McNair DS, Malila FM. Acute Physiology and Chronic Health Evaluation (APACHE) IV: hospital mortality assessment for today's critically ill patients. Crit Care Med 2006; 34:1297-1310.
(25.) Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A, Bruining H et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996; 22:707-710.
(26.) Cole SR, Hernan MA. Constructing inverse probability weights for marginal structural models. Am J Epidemiol 2008; 168:656-664.
(27.) Hernan MA, Brumback B, Robins JM. Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men. Epidemiology 2000; 11:561-570.
(28.) Hernan MA, Brumback BA, Robins JM. Estimating the causal effect of zidovudine on CD4 count with a marginal structural model for repeated measures. Stat Med 2002; 21:1689-1709.
(29.) Marshall MR, Hawley CM, Kerr PG, Polkinghorne KR, Marshall RJ, Agar JWM et al. Home hemodialysis and mortality risk in Australian and New Zealand populations. Am J Kidney Dis 2011; 58:782-793.
(30.) Hernan MA, Hernandez-Diaz S, Robins JM. A structural approach to selection bias. Epidemiology 2004; 15:615-625.
(31.) Robins JM. Estimation of the time-dependent accelerated failure time model in the presence of confounding factors. Biometrika 1992; 79:321-334.
(32.) Bodnar LM, Davidian M, Siega-Riz AM, Tsiatis AA. Marginal structural models for analyzing causal effects of time-dependent treatments: an application in perinatal epidemiology. Am J Epidemiol 2004; 159:926-934.
(33.) Cole SR, Hernan MA, Margolick JB, Cohen MH, Robins JM. Marginal structural models for estimating the effect of highly active antiretroviral therapy initiation on CD4 cell count. Am J Epidemiol 2005; 162:471-478.
(34.) Cole SR, Hernan MA, Robins JM, Anastos K, Chmiel J, Detels R et al. Effect of highly active antiretroviral therapy on time to acquired immunodeficiency syndrome or death using marginal structural models. Am J Epidemiol 2003; 158:687-694.
(35.) Cook NR, Cole SR, Hennekens CH. Use of a marginal structural model to determine the effect of aspirin on cardiovascular mortality in the Physicians' Health Study. Am J Epidemiol 2002; 155:1045-1053.
(36.) Fewell Z, Hernan MA, Wolfe F, Tilling K, Choi H, Sterne J. Controlling for time-dependent confounding using marginal structural models. The Stata Journal 2004; 4:402-420.
(37.) Comment: demystifying double robustness: a comparison of alternative strategies for estimating a population mean from incomplete data.
(38.) Schafer JL, Kang J. Average causal effects from nonrandomized studies: a practical guide and simulated example. Psychol Methods 2008; 13:279-313.
(39.) Kellum JA, Angus DC, Johnson JP, Leblanc M, Griffin M, Ramakrishnan N et al. Continuous versus intermittent renal replacement therapy: a meta-analysis. Intensive Care Med 2002; 28:29-37.
(40.) Faulhaber-Walter R, Hafer C, Jahr N, Vahlbruch J, Hoy L, Haller H et al. The Hannover Dialysis Outcome study: comparison of standard versus intensified extended dialysis for treatment of patients with acute kidney injury in the intensive care unit. Nephrol Dial Transplant 2009; 24:2179-2186.
(41.) Gashti CN, Salcedo S, Robinson V, Rodby RA. Accelerated venovenous hemofiltration: early technical and clinical experience. Am J Kidney Dis 2008; 51:804-810.
(42.) Kumar VA, Yeun JY, Depner TA, Don BR. Extended daily dialysis vs. continuous hemodialysis for ICU patients with acute renal failure: a two-year single center report. Int J Artif Organs 2004; 27:371-379.
(43.) Kielstein JT, Kretschmer U, Ernst T, Hafer C, Bahr MJ, Haller H et al. Efficacy and cardiovascular tolerability of extended dialysis in critically ill patients: a randomized controlled study. Am J Kidney Dis 2004; 43:342-349.
(44.) Abe M, Okada K, Suzuki M, Nagura C, Ishihara Y, Fujii Y et al. Comparison of sustained hemodiafiltration with continuous venovenous hemodiafiltration for the treatment of critically ill patients with acute kidney injury. Artif Organs 2010; 34:331338.
(45.) Lins R, Elseviers M, Van der Niepen P. Outcome of different treatment modalities in ARF: randomized population. J Am Soc Nephrol 2004; 14:195A.
(46.) Malbrain M, Elseviers M, Van Der Niepen P, Damas P, Hoste E, Devriendt J et al. Interim results of the SHARF4 study: outcome of acute renal failure with different treatment modalities. Critical Care 2004; 8 (Suppl 1):P153.
(47.) Lins RL, Elseviers MM, Van der Niepen P, Hoste E, Malbrain ML, Damas P et al. Intermittent versus continuous renal replacement therapy for acute kidney injury patients admitted to the intensive care unit: results of a randomized clinical trial. Nephrol Dial Transplant 2009; 24:512-518.
(48.) Gedmintas A, Crilly J, Richards B, Comadira GP, Creamer J, Lind J et al. Haemodynamic stability is maintained during extended daily diafiltration in critically ill septic patients. Crit Care Resusc 2010; 12:203-208.
(49.) Baldwin I, Bellomo R, Naka T, Koch B, Fealy N. A pilot randomized controlled comparison of extended daily dialysis with filtration and continuous veno-venous hemofiltration: fluid removal and hemodynamics. Int J Artif Organs 2007; 30:10831089.
(50.) Fieghen HE, Friedrich JO, Burns KE, Nisenbaum R, Adhikari NK, Hladunewich MA et al. The hemodynamic tolerability and feasibility of sustained low efficiency dialysis in the management of critically ill patients with acute kidney injury. BMC Nephrol 2010; 11:32.
(51.) Kron J, Kron S, Wenkel R, Schuhmacher HU, Thieme U, Leimbach T et al. Extended daily on-line high-volume haemodiafiltration in septic multiple organ failure: a well-tolerated and feasible procedure. Nephrol Dial Transplant 2011; [Epub ahead of print].
(52.) Marcelino P, Marum S, Fernandes AP, Ribeiro JP. [Hybrid or continuous renal replacement techniques for unstable haemodynamic patients in the intensive care unit]. Acta Med Port 2006; 19:275-280.
(53.) Bellomo R, Cass A, Cole L, Finfer S, Gallagher M, Goldsmith D et al. Design and challenges of the Randomized Evaluation of Normal versus Augmented Level Replacement Therapy (RENAL) Trial: high-dose versus standard-dose hemofiltration in acute renal failure. Blood Purif 2008; 26:407-416.
(54.) Bellomo R, Cass A, Cole L, Finfer S, Gallagher M, Lo S et al. Intensity of continuous renal-replacement therapy in critically ill patients. N Engl J Med 2009; 361:1627-1638.
(55.) Palevsky PM, O'Connor T, Zhang JH, Star RA, Smith MW. Design of the VA/NIH Acute Renal Failure Trial Network (ATN) Study: intensive versus conventional renal support in acute renal failure. Clin Trials 2005; 2:423-435.
(56.) Palevsky PM, Zhang JH, O'Connor TZ, Chertow GM, Crowley ST, Choudhury D et al. Intensity of renal support in critically ill patients with acute kidney injury. N Engl J Med 2008; 359:7-20.
(57.) Cohen J, Guyatt G, Bernard GR, Calandra T, Cook D, Elbourne D et al. New strategies for clinical trials in patients with sepsis and septic shock. Crit Care Med 2001; 29:880-886.
(58.) Bagshaw SM, Laupland KB, Doig CJ, Mortis G, Fick GH, Mucenski M et al. Prognosis for long-term survival and renal recovery in critically ill patients with severe acute renal failure: a population-based study. Crit Care 2005; 9:R700-709.
(59.) Petitti DB, Freedman DA. Invited commentary: how far can epidemiologists get with statistical adjustment? Am J Epidemiol 2005; 162:415-418; discussion 419-420.
(60.) Prentice RL, Langer R, Stefanick ML, Howard BV, Pettinger M, Anderson G et al. Combined postmenopausal hormone therapy and cardiovascular disease: toward resolving the discrepancy between observational studies and the Women's Health Initiative clinical trial. Am J Epidemiol 2005; 162:404-414.
(61.) Bandyopadhyay S, Jakobson D, Chhabra KD, Baker A. Improvement of cerebral blood flow patterns in hepatorenal syndrome using sustained low-efficiency dialysis. Br J Anaesth 2010; 105:547-548.
N. KHANAL *, M.R. MARSHALL ([dagger]), T.M. MA ([double dagger]), P.J. PRIDMORE ([section]), A.B. WILLIAMS **, A.P.N. RANKIN **
Department of Intensive Care Medicine, Counties Manukau District Health Board, Auckland, New Zealand
* FCPS, Nephrologist, Department of Nephrology, Capital and Coast District Health Board.
([dagger]) MB, ChB, MPH, FRACP, Nephrologist, Department of Renal Medicine, Counties Manukau District Health Board and Senior Lecturer, Faculty of Medical and Health Sciences, University of Auckland.
([double dagger]) MD, PhD Student, Knowledge Engineering and Discovery Research Institute, Auckland University of Technology.
([section]) FACEM, FCICM, Intensivist, Intensive Care Unit, Fremantle Hospital and Health Services, Fremantle, Western Australia, Australia.
** FCICM, Intensivist.
Address for correspondence: Dr M. R. Marshall, Department of Renal Medicine, Middlemore Hospital, Counties Manukau District Health Board, Private Bag 93311, Auckland 1640, New Zealand. Email: mark.marshall@ middlemore.co.nz
Accepted for publication on December 31, 2011.
Table 1 Clinical characteristics of the study cohort by modality exposure. Patients may be classified in multiple modality categories due to multiple exposures over the period of observation. All data are presented as n (%) or mean (SD) Variable PIRRT CRRT Patients 118 48 Age, y 57.6 (14.9) 58.0 (16.1) Gender Male 70 (59.3) 33 (68.8) Female 48 (40.7) 15 (31.3) Ethnicity Asian 8 (6.8) 3 (6.3) Caucasian 64 (54.2) 20 (41.7) New Zealand Maori 16 (13.6) 8 (16.7) Pacific people 30 (25.4) 17 (35.4) Body mass index, kg/[m.sup.2] 31.6 (9.7) 28.6 (6.8) (pre-morbid weight) Pre ICU eGFR, ml/kg/1.73 [m.sup.2] 59.5 (29.4) 71.5 (30.3) APACHE IV, Acute Physiology Score 108.9 (38.5) 114.9 (36.7) APACHE IV, risk of death % 67.7 (28.2) 77.3 (25.4) SOFA at ICU admission 12.7 (4.7) 14.2 (4.2) SOFA on RRT days 13.2 (4.1) 13.9 (3.9) Lead time to RRT, days 2.6 (2.4) 2.5 (2.3) ICU length of stay, days 10.1 (9.5) 12.4 (10.0) Serum creatinine at RRT inception, 359 (214) 301 (127) mmol/l Serum urea at RRT inception, mmol/l 21.6 (11.6) 18.9 (9.2) Serum K at RRT inception, mmol/l 5.2 (1.1) Arterial pH at RRT inception 7.17 (0.15) 7.18 (0.14) Sepsis 82 (69.5) 34 (70.8) Posoperative status 52 (44.1) 25 (52.1) Vasopressors 103 (88) 47 (97.9) Exposure to PIRRT during ICU admission - 25 (52.1) Exposure to CRRT during ICU admission 25 (21.2) - Exposure to iHD during ICU admission 13 (11) 7 (14.6) Variable iHD Total Patients 20 146 Age, y 58.6 (17.56) 58.5 (15.2) Gender Male 11 (55) 88 (60.3) Female 9 (45) 58 (39.7) Ethnicity Asian 1 (5) 9 (6.2) Caucasian 8 (40) 74 (50.7) New Zealand Maori 3 (15) 20 (13.7) Pacific people 8 (40) 43 (29.5) Body mass index, kg/[m.sup.2] 30.4 (5.9) 30.8 (9.1) (pre-morbid weight) Pre ICU eGFR, ml/kg/1.73 [m.sup.2] 59.8 (40) 60.4 (30.4) APACHE IV, Acute Physiology Score 100.7 (35.5) 110.2 (37.9) APACHE IV, risk of death % 68.3 (26.2) 70.7 (27.2) SOFA at ICU admission 10.6 (4.4) 12.9 (4.5) SOFA on RRT days 10.5 (4.2) 13.1 (4.1) Lead time to RRT, days 1.9 (1.1) 2.5 (2.4) ICU length of stay, days 11.2 (11.2) 9.3 (8.9) Serum creatinine at RRT inception, 479 (171) 356 (210) mmol/l Serum urea at RRT inception, mmol/l 29.9 (16.3) 22.2 (12.7) Serum K at RRT inception, mmol/l 5.1 (1.1) 5.1 (1.1) Arterial pH at RRT inception 7.25 (0.13) 7.18 (0.15) Sepsis 14 (70) 103 (70.6) Posoperative status 6 (30) 63 (43.2) Vasopressors 15 (75) 128 (88.3) Exposure to PIRRT during ICU admission 13 (65) - Exposure to CRRT during ICU admission 7 (35) - Exposure to iHD during ICU admission - - PIRRT=prolonged intermittent renal replacement therapy, CRRT=continuous renal replacement therapy, iHD=intermittent haemodialysis, ICU=intensive care unit, eGFR=epidermal growth factor receptor, APACHE=Acute Physiology and Chronic Health Evaluation, SOFA=Sepsis-Related Organ Failure Assessment, RRT=renal replacement therapy. Table 2 Renal replacement therapy characteristics. All results are mean (SD) or range Variable PIRRT CRRT iHD N 413 166 55 Duration, h 7.16 (7.8) 24-48 3.43 (1.8) QB, ml/min 263 (75) 236 (39) 271 (101) QD, ml/min 215 (206) 38 (22) 510 (252) QF, ml/min 92 (80) 43 (20) - Single-pool Kt/V 1.53 (0.34) - 1.19 (0.54) PIRRT=prolonged intermittent renal replacement therapy, CRRT=continuous renal replacement therapy, iHD=intermittent haemodialysis, QB=blood flow rate, QD=counter-current bicarbonate-based dialysate with flow rate, QF=filtration rate. Table 3 Hazard ratio (95% CIs) for patient death as determined at hospital discharge, by modality Adjustments in statistical Modality Heart rate 95% CI model None PIRRT 1.0 (Ref) CRRT 1.11 0.57-2.18 iHD 0.70 0.21-2.29 Baseline variables PIRRT 1.0 (Ref) CRRT 1.37 0.65-2.84 iHD 1.10 0.32-3.83 Baseline and time PIRRT 1.0 (Ref) Varying variables CRRT 1.31 0.60-2.90 iHD 1.22 0.33-4.43 CI=confidence interval, PIRRT=prolonged intermittent renal replacement therapy, CRRT=continuous renal replacement therapy, iHD=intermittent haemodialysis. Table 4 Hazard ratio (95% CIs) for patient death as determined at 90 days post renal replacement inception, by modality. Adjustments in Modality HR 95% CI statistical model None PIRRT 1.0 (Ref) CRRT 0.89 0.42-1.83 iHD 0.88 0.17-4.51 Baseline variables PIRRT 1.0 (Ref) CRRT 0.99 0.39-2.53 iHD 2.16 0.45-10.33 Baseline and time PIRRT 1.0 (Ref) Varying variables CRRT 0.96 0.39-2.36 iHD 2.22 0.49-10.11 CI=confidence interval, HR=heart rate, PIRRT=prolonged intermittent renal replacement therapy, CRRT=continuous renal replacement therapy, iHD=intermittent haemodialysis. Table 5 Published clinical trials comparing outcomes between PIRRT and CRRT Author (year) Outcome n PIRRT CRRT Kumar (2004) Hospital mortality 54 54% 71% Kielstein (2004) 10-day mortality 39 60% 58% Holt (2010) 30-day mortality 21 0% 62% Abe (2010) ICU/30-day mortality 60 13% 30% Hospital mortality 17% 37% Lins (2004), Malbrain (2004) Hospital mortality 996 68% 64% PIRRT=prolonged intermittent renal replacement therapy, CRRT=continuous renal replacement therapy, ICU=intensive care unit.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Original Papers|
|Author:||Khanal, N.; Marshall, M.R.; Ma, T.M.; Pridmore, P.J.; Williams, A.B.; Rankin, A.P.N.|
|Publication:||Anaesthesia and Intensive Care|
|Date:||Mar 1, 2012|
|Previous Article:||Prevention of gastrointestinal bleeding due to stress ulceration: a review of current literature.|
|Next Article:||Pharmacokinetics of oxycodone after subcutaneous administration in a critically ill population compared with a healthy cohort.|