Indications for and Contraindications to Intermittent Hemodialysis in Critically Ill Patients

Operational Characteristics of Intermittent Hemodialysis and Hemofiltration

In IHD, molecule removal is driven by a concentration gradient between the vascular compartment and the dialysate side. This method favors removal of small molecules because their high diffusibility across the membrane provides a high efficiency (clearance around 200 mL/min). This high clearance is responsible for a rapid decrease in the concentration gradient, which in turn leads to a drop in the removal rate, thus limiting the amount of solute removed ( Fig. 153.1 ). These characteristics explain why IHD is used discontinuously, usually for 4 to 6 hours every day or every other day. Taking into account the high urea volume distribution and the high efficiency of the treatment, the refilling of urea from the interstitium to the vascular compartment is limited during the IHD session but occurs soon after the end of treatment. This explains the increase in serum urea after each session, called urea rebound. This phenomenon limits IHD efficiency.

Because of the rapid exchange of solute, high and fast osmolality variations may occur during treatment. These variations involve the vascular compartment and may induce or worsen cellular edema, leading to cerebral edema. In addition, along with the high ultrafiltration rate of IHD needed by the shortness of the session, these osmolality variations participate in hemodynamic impairment. However, the short duration of IHD sessions offers some advantages ( Table 153.1 ). The nurse's workload is diminished, the patient's mobility is preserved, and bleeding risk is decreased because of low exposure to anticoagulants. Moreover, treatment can be performed without anticoagulation, with good efficiency given the short duration and the high blood flow. In addition, from a practical point of view, one machine can treat several patients a day, whereas continuous therapies require one monitor for each patient-day. Yet IHD presents some technical limitations (see Table 153.1 ): it demands a specific water production, more complex training of care providers and, in many countries, the intervention of a nephrology team.

Hemofiltration refers to all extrarenal therapies that use convection as the mechanism of solute or water removal. Therefore solute and water removal is driven by a pressure gradient between the blood and ultrafiltrate sides of the membrane. The solute concentration in the ultrafiltrate side is then similar to the blood concentration, and small molecule clearance rate exactly correlates with the ultrafiltration rate (around 25 mL/min). This low clearance rate explains the necessity to use hemofiltration continuously. Two other RRT methods use continuous patterns, either based on diffusion (continuous venovenous hemodialysis [CVVHD]) or combining diffusion and convection (continuous venovenous hemodiafiltration). All of these continuous therapies are collectively called CRRT. The specific characteristics of hemofiltration account for many advantages: no abrupt variation of osmolality, the management of net ultrafiltration over 24 hours, and an increase in the amount of urea removed, considering the interstitium's potential to refill the plasma compartment. This explains the better hemodynamic tolerance and efficiency usually reported with the use of hemofiltration. In addition, the convection mechanism allows a higher efficiency of removal of middle-molecular-weight substances, with a potential effect on inflammatory mediators. In contrast, the continuous aspect of this method entails some limitations (see Table 153.1 ): high dose of anticoagulation, lack of patient mobility, higher nurse's workload, and frequent unplanned interruptions of treatment.

Intermittent Hemodialysis and Continuous Renal Replacement Therapy: Is One Better Than the Other?

The debate between the proponents of IHD and CRRT is ongoing, with valuable arguments on both sides. Several studies have compared the two methods, but most of them were nonrandomized, retrospective trials. Many methodologic biases preclude conclusions to be drawn from the results of these studies; the membranes were not standardized (biocompatible in CRRT, cuprophane in IHD), different therapies were pooled in CRRT (arteriovenous and venovenous methods) or in IHD (peritoneal dialysis and IHD), and some studies compared two groups enrolled at different times (historical IHD group). Probably the most important limitation is the lack of standardization for efficiency (i.e., dialysis dose) and hemodynamic tolerance in IHD. Indeed, we know that hemodynamic tolerance can be improved significantly with the use of specific settings in IHD for critically ill patients and that dialysis dose needs specific attention. Nevertheless, these studies reported conflicting results, either in favor of CRRT or not.

Eight prospective randomized studies have been published. The study by Mehta et al. found a significantly higher mortality in the CRRT group, whereas the seven other studies found no significant difference between the methods in terms of mortality. In the Mehta study, however, despite randomization, the IHD and CRRT patient groups were not comparable for several covariates (number of organ failures and severity score), but the multivariate analysis showed no relation between the mode of RRT and mortality. Most of those studies present major weaknesses, such as randomization failure, modifications of therapeutic protocol during the study period, combination of different types of CRRT, and small number of heterogeneous groups of patients enrolled. However, the multicenter Hemodiafe study, conducted by Vinsonneau et al., enrolled 360 patients and found no significant difference in survival between the two groups (60-day survival, 32% for IHD vs. 33% for CRRT). In that study, both techniques were standardized for membrane polymers and dialysis buffers, factors known to affect the ability of patients to tolerate renal replacement therapies. In addition, guidelines based on results of the study by Schortgen et al. were provided to improve hemodynamic tolerance of IHD.

Regarding renal recovery, the literature provides conflicting results. Retrospective or observational studies report a higher rate of dialysis dependency at ICU or hospital discharge when IHD was used as first line compared with CRRT. These studies have some important limitations. Allocation of treatment (IHD or CRRT) depended on patient's baseline characteristics. CRRT was applied in more severe patients, leading to a higher mortality and probably a lower number of patients at risk to become dialysis dependent. Thus the evaluation of dependency in survivors induced an evident bias in favor of CRRT. Two recent observational studies used propensity score or marginal structural Cox model to decrease the impact of allocation bias and differences in baseline or time-dependant confounding factors. In the study from Liang et al. of 4738 patients included with Kidney Disease: Improving Global Outcomes (KDIGO) stage 3 AKI, 28.2% received RRT. In multivariable analysis no difference was found in the odds ratio of recovery at day 90 (OR:1.19, 95% CI: 0.91–1.55, p = .20) or day 365 (OR:0.93, 95% CI: 0.72–1.2, p = .55) for patients initially treated with IHD or CRRT. The study from Truche et al. found similar results including 1360 patients receiving RRT with a composite primary end point (mortality and dialysis dependency at day 30). In this study, subgroup analysis showed that CRRT was favored in patients with higher weight gain at RRT initiation (HR for mortality and dialysis dependency at 30-day: 0.54, 95% CI: 0.29–0.99, p = .05) but deleterious in patients without shock (HR for mortality and dialysis dependency at 30-day: 2.24, 95% CI: 1.24–4.04, p =.01). These results are in agreement with those found in prospective randomized studies.

Therefore the two methods seem to provide similar outcomes in critically ill patients as long as they are performed by experienced teams with strict adherence to guidelines to improve hemodynamic tolerance. Therefore the operational characteristics of each method with its advantages and limitations (see Table 153.1 ) permit one to propose some good indications for IHD and some debatable ones. In fact, there is no a priori contraindication to IHD, given that prospective studies report similar survival rates for patients with ARF, even with multiple-organ dysfunction syndrome, who undergo the treatment.

Nevertheless, in some cases, hemofiltration may appear more suitable, as for example, in patients with severe hemodynamic instability, especially when high ultrafiltration rates are needed. Finally, the advantages of one method compensate for the limitations of the other—situations in which one probably should not be used are ideal for use of the other. This is in agreement with the recent Kidney Disease: Improving Global Outcomes (KDIGO) Clinical Practice Guidelines for Acute Kidney Injury that recommend to “use continuous and intermittent RRT as complementary therapies in AKI.” Therefore it is possible to propose more specific indications for IHD, even though either method can provide adequate treatment for ARF in the ICU. This is all the more true when new developments are implemented such as high efficiency hemofiltration to enhance delivered dose or sustained low-efficiency dialysis to enhance the tolerance of IHD.

The choice should be determined in light of the two main objectives of RRT, adequate delivered dialysis dose and good hemodynamic tolerance to avoid ischemic events. Therefore the better method is the one that permits these objectives to be achieved for each patient.

Situations in Which Intermittent Hemodialysis Should Be Preferred

IHD is indicated to treat the metabolic syndrome of acute ARF and to manage fluid balance. The best indications are acute metabolic or toxic situations in acutely ill patients without uncontrolled hemodynamic instability. The need to treat a patient without using anticoagulation and the preference to permit patient mobility are other good indications. Inefficient hemofiltration for repeated filter clotting despite adequate anticoagulation and insufficient metabolic control can be good indications as well. Given the low efficiency of diffusion in removing middle-molecular-weight substances, IHD cannot be considered for modulation of inflammatory processes, but to date, no evidence does support the use of RRT to modulate inflammation.