Vinsonneau C, Camus C, Combes A, et al. The Hemodiafe Study Group. Continuous venovenous haemodiafiltration versus intermittent haemodialysis for acute renal failure in patients with multiple-organ dysfunction syndrome: a multicentre randomised trial. Lancet. 2006 Jul 29;368(9533):379-85
Until the early 1980s, intermittent hemodialysis (IHD) was the standard modality of renal replacement therapy (RRT) in critically ill patients with acute renal failure. Hemodialysis effectively cleared potassium, urea, creatinine, and other molecules that accumulate in renal failure; however, hemodynamic instability was common. Besides acute shifts in the intravascular volume, the non-biocompatible, cellulose-based membranes of that era often triggered a profound inflammatory response, leading to severe hypotension. Perhaps due to fear of potential complications, clinicians were reluctant to perform dialysis in acutely ill patients until it seemed inevitable. Hemodialysis was delayed until the blood urea nitrogen level rose to well above 100 mg/dl, or the ominous appearance of “uremic frost” – deposition of urea crystals on the skin – during the late phase of renal failure. Outcomes were predictably poor with this approach, and mortality remained unacceptably high. It was abundantly clear that critically ill patients tolerated conventional hemodialysis poorly.
The concept of continuous renal replacement therapy (CRRT) arose at the 1982 meeting of the American Society for Artificial Internal Organs (ASAIO). Peter Kramer, a German physician, presented his observations when he interposed a microporous filter between the femoral artery and vein. Driven by the femoral arterial pressure, large volumes of a yellow-tinged fluid – the ultrafiltrate – poured out through the filter. Kramer replaced the ultrafiltrate with an intravenous electrolyte solution devoid of potassium, creatinine, urea, and other molecules that are excreted by the kidney. He went on to explain that this technique could be performed continuously, thereby facilitating solute clearance and fluid removal at a constant rate (1). His words were prophetic, triggering a paradigm shift in the technique of renal replacement therapy in the critically ill.
Decades later, Robert Bartlett, widely regarded as the pioneer of extracorporeal membrane oxygenation, described his early experience with the technique of continuous arterio-venous hemofiltration that Kramer had described, among patients who remained anuric despite industrial doses of diuretics. “I wept with joy”, he reminisced; most of the problems of hemodialysis were eliminated, and they could feed patients without the fear of fluid overload (2).
Background to the Hemodiafe trial
Until the early 1990s, IHD remained the mainstay of RRT in most parts of the world. However, ICUs were generally ill-equipped to perform IHD, with special requirements including reverse osmosis-enabled water purification. ICU nurses were not trained to perform hemodialysis; the nephrologist had to prescribe and supervise treatment – all leading to a tedious process. The concept of convection-based continuous therapy led to the introduction of automated machines that were less cumbersome to operate and relatively easy to perform by the bedside. The learning curve was short, the process simpler, and nephrologist supervision was no longer required. It was apparent that continuous therapies carried several advantages, including hemodynamic stability and superior control over fluid and solute removal. IHD-related rapid solute removal often led to a drastic drop in serum osmolality, with fluid shift from the intravascular to the interstitial compartment, leading to life-threatening cerebral edema in patients with neurological injury. CRRT, with slow, controlled solute removal, held the potential to avert this devastating complication. Recovery of renal function was also considered to occur more often with continuous compared with intermittent therapies. Unsurprisingly, ICUs around the world embraced the new technique with unbridled enthusiasm.
Despite the widespread use of the technique, the question of whether CRRT resulted in improved clinical outcomes compared with IHD in critically ill patients remained unanswered. Early, non-randomized studies had suggested improved outcomes with CRRT, although patients were more severely ill at baseline (3,4). A randomized controlled trial (RCT) reported higher mortality with CRRT compared with IHD, although patients in the CRRT group had adverse predictors of mortality, including more profound organ dysfunction (5). Two other RCTs (6,7) and a meta-analysis (8) did not demonstrate any significant difference in clinical outcomes between the two modalities of therapy. These studies did not employ standardized methods of therapy in either arm, making it difficult to draw definitive conclusions. The Hemodiafe trial aimed to compare CRRT with IHD using standard polyacrylonitrile membranes and evaluate the impact on mortality in patients with acute renal failure as part of multiorgan failure (9). The study was pertinent to many French ICUs, including that of the lead investigator, Christophe Vinsonneau, that employed both modalities of treatment.
Population and design
The Hemodiafe trial was conducted between October 1999 and March 2003 across 21 French ICUs among patients with acute renal failure as part of multiorgan dysfunction syndrome. Acute renal failure was defined as a serum urea level of 36 mmol/l or a serum creatinine level of 310 μmol/l (3.5 mg/dl). A logistic dysfunction score ≥ represented multiorgan dysfunction syndrome. Patients were randomly allocated to receive CRRT or IHD. Randomization was stratified by center.
Patients with chronic renal failure and acute renal failure due to obstructive or a vascular etiology were excluded. Patients on angiotensin-converting-enzyme inhibitor, those with coagulopathies, uncontrolled bleeding, a SAPS II score ≤37, a moribund state, and those not expected to survive beyond eight days were also excluded.
The CRRT group
CRRT was commenced in the continuous veno-venous hemodiafiltration mode (CVVHDF) at a blood flow rate of ≥120 ml/min, a dialysate flow of 500 ml/hour, and an ultrafiltration rate of ≥1000 ml/hour. Prisma machines were used in the predilution mode with a bicarbonate-based replacement fluid. The filter was routinely changed every 48 hours. The settings were adjusted depending on the clinical condition, aiming to maintain the serum urea level at <30 mmol/l. Once multiorgan dysfunction had resolved and hemodynamic stability was attained, patients could be switched over to hemodialysis.
The IHD group
In the IHD group, the initial settings included a blood flow rate of 250 ml/min, and dialysate flow of 500 ml/min. The dialysate was cooled to 35°C and the sodium level in the dialysate fluid was set at 150 mmol/l for improved hemodynamic stability. Dialysis was performed for at least 4 hours once in 48 hours in the presence of continued anuria or oliguria. The frequency was customized to maintain a serum urea level of <40 mmol/l. Therapy was individualized, aiming for a urea reduction ratio of >65% during each session. IHD was carried out using machines available at the study center. All treatment was performed using the same membrane polymer with a large surface area (2 m2), with a bicarbonate-based dialysate.
Other therapies, including fluid resuscitation, pharmacological support of the circulation, mechanical ventilation, and antibiotic therapy, were left to clinician judgment. Anticoagulation was carried out using unfractionated or low molecular-weight heparin based on a standard protocol.
According to the study protocol, a switch of treatment modality was not allowed until after discussion with the coordinating center. The possible reasons for a change of treatment modality were predefined, including persisting hypotension in the absence of hypovolemia, unfavorable fluid balance or poor metabolic control, technique-related adverse events, and technical problems that prevented continued treatment with the allocated modality.
The 60-day mortality, the primary endpoint, was assumed to be 45% in the IHD group. Assuming a 15% lower mortality in the CRRT group, the authors calculated a sample size of 240 patients in each group allowing for a 10% dropout, providing 90% power at an alpha level of 0.05.
A total of 360 patients were randomized; 184 patients from the IHD group and 175 from the CRRT group were included in the final analysis. Patients were well-matched at baseline. Most patients (70%) were medical; 96% were on mechanical ventilation, and 87% were on catecholamine support at baseline.
Intensity of therapy
The mean dose of CVVHDF was 29 (± 11) ml/kg/hour. IHD was performed for a mean of 3.6 (± 2) sessions during the first week of treatment; seven patients underwent daily IHD during this period. IHD was carried out for a median duration of 5.2 (5.1–5.3) hours per session. The median net fluid daily fluid removal was similar in the IHD and CRRT groups [2213 ml (2141–2285) vs. 2107 ml (2011–2203)]. The efficacy of treatment, based on the mean daily urea level, did not differ between the two groups.
Primary and secondary outcomes
There was no difference in the 60-day survival, the primary outcome, between the IHD and the CRRT groups (31.5% vs. 32.6%, p = 0.98). Among the secondary outcomes, the 28-day (41.8% vs. 38.9%) and the 90-day survival (27.2% vs. 28.5%) were also similar between the two groups. The median duration of renal support was 11 days in both groups. The duration of ICU stay (20 vs. 19 days) and hospital stay (30 vs. 32 days) were also not significantly different.
Recovery of renal function was similar in both groups, with 6/61 (10%) in the IHD group and 4/61 (7%) patients in the CRRT group remaining dialysis-dependent following discharge from the ICU. There was a solitary patient in the CRRT group who remained on dialysis at the time of hospital discharge.
Adverse events and switch of therapy
Contrary to expectations, the incidence of hypotensive episodes, defined as a systolic BP ≤ 80 mm Hg or a fall by >50 mm Hg from baseline, was similar between the IHD and CRRT groups (39% vs. 35%, p = 0.47). The incidence of arrhythmias was also similar.
Hypothermia was less commonly observed with IHD compared with CRRT. Six patients in the IHD group were switched to CRRT – three for hemodynamic instability and two due to technical problems. Seventeen patients in the CRRT group were switched to IHD due to contraindications to anticoagulant administration, a high risk of bleeding, and recurrent clotting of the filter despite adequate anticoagulation.
The authors concluded that IHD might be feasible in nearly all critically ill patients with acute renal failure as part of multiorgan dysfunction.
Strengths of the study
The Hemodiafe trial was conducted during an era in which CRRT was in widespread use in critically ill patients. However, the outcome benefits of CRRT compared with IHD remained uncertain. This landmark RCT compared IHD with CRRT in a sick cohort of ICU patients; most were on mechanical ventilation and required circulatory support with catecholamines. The severity of illness in both groups were well-matched at baseline. The treatment protocol was standardized in both arms, and crossover was not allowed unless it was approved by the coordinating center. Among the 360 patients who were randomized, all except one were followed up and included in the final analysis.
Although one of the largest RCTs that compared continuous vs. intermittent modalities of renal replacement therapy, the study was underpowered. It fell short of the calculated sample size was 480, enrolling only 360 patients. Criticisms were directed towards likely underdosing with CRRT, contrasting with a more balanced approach with IHD. The duration and frequency of IHD were greater compared to contemporaneous practice; this may have led to improved outcomes with IHD. The study may have been influenced by the findings of the Schiffl et al. trial of 2002, which showed improved survival with daily compared with alternate day IHD (10). In fact, the frequency of IHD was noted to increase during the course of the study. The increase in the frequency of IHD sessions may have resulted in the progressively lower mortality observed during the course of the study in the IHD group. In contrast, the mortality in the CRRT group remained stable during the entire course of the study. The recommended dose of CRRT was 35 ml/kg/hour during the period that the study was conducted (11); however, the delivered dose was lower at 29 ml/kg/hour.
An unclear “moribund” state was one of the exclusion criteria. Did the investigators exclude patients who were on high-dose vasopressors, who may have benefited the most with CRRT?
Furthermore, the investigators modified IHD to enable therapy in hemodynamically unstable patients. The dialysate was cooled to 35 °C; cooling the dialysate has been shown to improve hemodynamic stability during high-efficiency hemodialysis (12). The sodium level of the dialysate fluid was also set at 150 mmol/l to improve hemodynamic stability.
RRT in the ICU has undergone considerable refinement since the publication of the Hemodiafe trial. This landmark trial provided the impetus for modifications of conventional hemodialysis to adapt to the specialized requirements of critically ill patients who were at risk of hemodynamic instability. Currently, prolonged, intermittent renal replacement therapies (PIRRT) carried out over an extended duration are being increasingly employed in patients who are at high risk of hemodynamic instability (13). Hybrid therapies, variably described as sustained low-efficiency dialysis (SLED), extended daily dialysis (EDD), or slow, continuous dialysis (SCD) have evolved, and are the preferred modality of RRT in many ICUs (14).
The Hemodiafe trial addressed the long-debated question of clinical outcomes following CRRT compared with IHD in critically ill patients. The trial was conducted against a background of inconclusive findings from preceding studies. IHD was already being carried out in many ICUs in France as the standard modality of RRT. Modifications were made to the technique of IHD to reduce hemodynamic instability by cooling the dialysate to 35°C and maintaining the sodium level at 150 mmol/l. As the study progressed, the frequency of IHD also seemed to have increased, probably related to newly available evidence. The study was criticized for comparing a modified, optimized adaptation of IHD with a suboptimal dose of CRRT. However, this landmark trial spurred the development of further refinements to IHD that culminated in several modes of PIRRT employed in many ICUs today. Hence, their conclusion that “virtually all patients can be treated with intermittent hemodialysis” remains valid.
1. Kramer P, Böhler J, Kehr A, Gröne HJ, Schrader J, Matthaei D, et al. Intensive care potential of continuous arteriovenous hemofiltration. Trans – Am Soc Artif Intern Organs. 1982;28:28–32.
2. Bartlett RH. The Origins of Continuous Renal Replacement Therapy. ASAIO J Am Soc Artif Intern Organs 1992. 2018;64(3):427–30.
3. Guérin C, Girard R, Selli JM, Ayzac L. Intermittent versus continuous renal replacement therapy for acute renal failure in intensive care units: results from a multicenter prospective epidemiological survey. Intensive Care Med. 2002 Oct;28(10):1411–8.
4. Bellomo R, Farmer M, Parkin G, Wright C, Boyce N. Severe acute renal failure: a comparison of acute continuous hemodiafiltration and conventional dialytic therapy. Nephron. 1995;71(1):59–64.
5. Mehta RL, McDonald B, Gabbai FB, Pahl M, Pascual MT, Farkas A, et al. A randomized clinical trial of continuous versus intermittent dialysis for acute renal failure. Kidney Int. 2001 Sep;60(3):1154–63.
6. Uehlinger DE, Jakob SM, Ferrari P, Eichelberger M, Huynh-Do U, Marti HP, et al. Comparison of continuous and intermittent renal replacement therapy for acute renal failure. Nephrol Dial Transplant Off Publ Eur Dial Transpl Assoc – Eur Ren Assoc. 2005 Aug;20(8):1630–7.
7. Augustine JJ, Sandy D, Seifert TH, Paganini EP. A randomized controlled trial comparing intermittent with continuous dialysis in patients with ARF. Am J Kidney Dis Off J Natl Kidney Found. 2004 Dec;44(6):1000–7.
8. 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 Jan;28(1):29–37.
9. Vinsonneau C, Camus C, Combes A, Costa de Beauregard MA, Klouche K, Boulain T, et al. Continuous venovenous haemodiafiltration versus intermittent haemodialysis for acute renal failure in patients with multiple-organ dysfunction syndrome: a multicentre randomised trial. Lancet Lond Engl. 2006 Jul 29;368(9533):379–85.
10. Schiffl H, Lang SM, Fischer R. Daily hemodialysis and the outcome of acute renal failure. N Engl J Med. 2002 Jan 31;346(5):305–10.
11. Ronco C, Bellomo R, Homel P, Brendolan A, Dan M, Piccinni P, et al. Effects of different doses in continuous veno-venous haemofiltration on outcomes of acute renal failure: a prospective randomised trial. Lancet Lond Engl. 2000 Jul 1;356(9223):26–30.
12. Kaufman AM, Morris AT, Lavarias VA, Wang Y, Leung JF, Glabman MB, et al. Effects of controlled blood cooling on hemodynamic stability and urea kinetics during high-efficiency hemodialysis. J Am Soc Nephrol JASN. 1998 May;9(5):877–83.
13. Schwenger V, Weigand MA, Hoffmann O, Dikow R, Kihm LP, Seckinger J, et al. Sustained low efficiency dialysis using a single-pass batch system in acute kidney injury – a randomized interventional trial: the REnal Replacement Therapy Study in Intensive Care Unit PatiEnts. Crit Care Lond Engl. 2012 Jul 27;16(4):R140.
14. Berbece AN, Richardson RMA. Sustained low-efficiency dialysis in the ICU: cost, anticoagulation, and solute removal. Kidney Int. 2006 Sep;70(5):963–8.
Best 35 critical care blogs to follow in 2023: https://blog.feedspot.com/critical_care_blogs/