The Efficacy of Renal Replacement Therapy for Rewarming of Patients in Severe Accidental Hypothermia—Systematic Review of the Literature

Background: Renal replacement therapy (RRT) can be used to rewarm patients in deep hypothermia. However, there is still no clear evidence for the effectiveness of RRT in this group of patients. This systematic review aims to summarize the rewarming rates during RRT in patients in severe hypothermia, below or equal to 32 °C. Methods: This systematic review was registered in the PROSPERO International Prospective Register of Systematic Reviews (identifier CRD42021232821). We searched Embase, Medline, and Cochrane databases using the keywords hypothermia, renal replacement therapy, hemodialysis, hemofiltration, hemodiafiltration, and their abbreviations. The search included only articles in English with no time limit, up until 30 June 2021. Results: From the 795 revised articles, 18 studies including 21 patients, were selected for the final assessment and data extraction. The mean rate of rewarming calculated for all studies combined was 1.9 °C/h (95% CI 1.5–2.3) and did not differ between continuous (2.0 °C/h; 95% CI 0.9–3.0) and intermittent (1.9 °C/h; 95% CI 1.5–2.3) methods (p > 0.9). Conclusions: Based on the reviewed literature, it is currently not possible to provide high-quality recommendations for RRT use in specific groups of patients in accidental hypothermia. While RRT appears to be a viable rewarming strategy, the choice of rewarming method should always be determined by the specific clinical circumstances, the available resources, and the current resuscitation guidelines.


Introduction
The management of the patients suffering from accidental hypothermia, defined as an unintentional drop of the body core temperature below 35 • C, is challenging. The rewarming approach depends mainly on the patient's core temperature, clinical condition, and environmental setting [1]. In the pre-hospital setting, the highest priority should be given to the early recognition of hypothermia, thermal insulation to prevent further heat loss, and prompt transfer to the hospital [2]. Among the active rewarming methods available in the hospital setting, non-invasive and invasive can be distinguished. The management often combines techniques from different categories. Active external methods include warm blankets or forced warm air, with a rewarming rate of 0.5-4 • C/h. Active internal rewarming includes body cavities lavage with warm fluids (bladder, gastric, thoracic, or peritoneal lavage) with the rewarming rate estimated at 0.5-2 • C/h, application of renal replacement therapy (RRT) or extracorporeal life support (ECLS), mainly with extracorporeal membrane oxygenation (ECMO) [3]. The rewarming rate is the highest for ECMO with values up to 6 • C/h [4]. Due to limited availability and risk of severe complications, the use of ECMO is reserved for patients undergoing cardiopulmonary resuscitation or in deep hypothermia with signs of circulatory instability in select institutions [5]. On the contrary, RRT techniques are widely available and associated with a lower complication rate than ECMO therapy. However, no conclusive evidence on the efficacy of RRT in deep accidental hypothermia patients is currently available. This systematic review aims to assess the rewarming capacity of RRT in patients in severe hypothermia, below or equal to 32 • C.

Materials and Methods
This systematic review was registered in the PROSPERO International Prospective Register of Systematic Reviews (identifier CRD42021232821). A professional library assistant searched Embase, Medline, and Cochrane databases using the following keywords: "hypothermia" AND ("renal replacement therapy" OR "hemodialysis" OR "hemofiltration" OR "hemodiafiltration" OR "HD" OR "CVVH" OR "CVVHD" OR "CVVHDF"). The search included only articles in English with no time limit, up until 30 June 2021. The obtained results were independently assessed for their relevance by two authors (TD and PPod) according to the predefined criteria: only articles including human subjects with accidental moderate-to-severe hypothermia (defined as a body temperature below or equal to 32 • C) treated with RRT were included in the review. Reviews, animal studies, publications on therapeutic hypothermia or iatrogenic hypothermia were excluded. Conference abstracts were not analyzed. In addition, the references of the included studies were searched manually to find any additional studies. Any disagreements were resolved by team discussion. Two review authors (KM and SK) independently extracted data for each included study. The main objective was to summarize the RRT rewarming rates reported in the literature. Details about the type of RRT, other types of rewarming used, rewarming time, initial and end temperatures, the incidence of cardiac arrest, and patient outcome were also recorded. The quality of the publications included in the systematic review was assessed using the tool introduced by Murad et al. and presented in Table 1 [6].

Results
From the 795 revised articles, 18 studies including 21 patients were selected for the final assessment and data extraction ( Figure 1). The mean rate of rewarming calculated for all studies combined was 1.9 • C/h (95% CI 1.5-2.3) and did not differ between continuous (2.0 • C/h; 95% CI 0.9-3.0) and intermittent (1.9 • C/h; 95% CI 1.5-2.3) methods (p > 0.9). The lowest reported rewarming rate was 0.6 • C/h and the highest 4 • C/h, both were achieved with the continuous RRT [17,18]. For continuous renal replacement therapy (CRRT), the median rewarming time was 4.5 h (IQR 3.3-6.1), the initial median temperature 26. The mean age of the patients was 49.5 years ranging from 17 to 77 years. Most patients were males (67%; 14/21) who were found outdoors (68%; 13/19), including one trauma patient. In two cases, the circumstances of hypothermia were not reported. Data on comorbidities were provided for less than half of the patients (43%; 9/21 patients) with alcohol abuse disorders predominating. The most frequently used RRT technique was hemodialysis (57%; 12/21), with dialysate temperature 37 • C (IQR 35-39 • C), median blood flow 250 mL/min (IQR 150-300 mL/min), and dialysate flow 500 mL/min (only in one publication dialysate flow was 800 mL/min [21]). Only two publications provided technical information on flows during CRRT, with a median blood flow of 175 mL/min (IQR 125-250 mL/min). Detailed information is presented in Table 2.

Discussion
To our best knowledge, this systematic review is the first to provide structured evidence on the rewarming rates using RRT in hypothermic patients. The commonly cited rewarming rate during RRT is 2-3 • C/h [3]. Based on our review, the mean rewarming rate calculated for all studies combined was 1.9 • C/h and did not differ between continuous and intermittent methods. Interestingly, both the lowest rewarming rate and the highest rewarming rate were reported under continuous RRT techniques. Our study shows that renal replacement therapy may be an appealing alternative for rewarming of hypothermic patients.
The first element in the chain of survival in patients suffering from accidental hypothermia is pre-hospital protection from cold, wet, and wind. The lack of adequate thermal insulation during transport poses the risk of further heat loss, thus may increase the risk of cardiac arrest (CA) [3]. The subsequent management of the hypothermic patient should be multidirectional and determined by factors such as the patient's hemodynamic stability, the degree of hypothermia, and existing comorbidities [5]. There are no definite recommendations based on randomized trials regarding the patients' eligibility for specific rewarming methods in the current literature. The current ERC/AHA resuscitation guidelines leave the decision for choosing the rewarming method to the physician, except for the cardiac arrest situation or hemodynamically unstable patients, in whom ECLS rewarming should be the method of choice [5,25]. Other therapies remain an option, not a recommendation, and each can be implemented not as the only main, but as an additional rewarming method.
RRT is a promising form of extracorporeal rewarming, but there is little scientific evidence to support its use in this setting. The rewarming rate for active internal rewarming techniques varies, ranging from 1 • C/h for bladder lavage to over 6 • C/h for ECMO. These are mostly invasive techniques, but gaining access to the central compartment allows rapid rewarming and thus can minimize the risk of CA or, if it occurs, hasten the return of spontaneous circulation [3]. Every technique used should have the potential for rapid rewarming and minimal risk of inducing hemodynamic instability. For example, during a resuscitation attempt, pleural lavage may compromise the quality of chest compressions, which may not be desirable. It is also a common belief that extracorporeal devices which do not support the circulation (such as hemodialysis) are relatively contraindicated as they can compromise the hemodynamics and are ineffective in the absence of spontaneous circulation. However, as shown in this review, reports exist of RRT being used to rewarm a patient in CA with good neurological outcomes [16,18,19,22,23].
The ECMO therapy is the most effective in rewarming the hypothermic patient, both with preserved circulation and in CA. The device's capabilities are greater than RRT. However, the optimal rewarming rates are still unknown. In patients in accidental hypothermia, mortality increases for every 0.5 • C/h drop in rewarming rate below 2 • C/h, but higher values cause a significant increase of neuronal damage markers and thus indicate the possibility of central nervous system injury [26,27]. Moreover, the use of ECMO poses the risk of severe and potentially fatal complications, both associated with severe bleeding and thromboembolism, whose incidence reaches almost 30% [28]. Therefore, ECMO therapy is reserved for highly specialized centers, often with cardiac surgery departments.
Renal replacement therapy is a method of active internal rewarming, particularly interesting due to its widespread availability and relatively high rewarming potential. In addition to a much faster initiation time than ECMO [29,30], it is available not only in ICU's but also in many other hospital units. Although the use of hemodialysis (HD) as a rewarming method was first described in 1965, only single cases have been reported in the literature since then [31]. As we have demonstrated, this surprisingly small number of publications provide limited details on the indications, duration, and specifics of the RRT protocols used in hypothermic patients.
The advantages of the RRT may be the relative simplicity and short onset time of the therapy, the ability of electrolyte correction, and precise control of fluid balance [32].
Also, using a smaller vascular cannula and avoiding systemic anticoagulation is beneficial. The use of systemic anticoagulation may hinder invasive procedures, such as pleural and peritoneal lavage, and may be contraindicated in trauma patients. Also, heparin anticoagulation should be avoided in hypothermic patients who have coagulopathy. For RRT, local anticoagulation with citrate is now the most commonly used method instead of systemic anticoagulation. It is the method of choice in critically ill patients, offering a relatively low (although not zero) risk of thromboembolic and hemorrhagic complications [33].
Currently, there is no uniform protocol for rewarming patients in hypothermia with renal replacement therapy. A recent study published by our team evaluated the heating capabilities of a CRRT device. We rewarmed a 5 L tank of 25 • C fluid (central compartment) using the CVVHDF technique. Initially, we assumed that the filter in which the dialysis fluid flows in the opposite direction to the blood might act as a countercurrent heat exchanger and be the potential element where the flowing blood can be rewarmed. However, according to mathematical calculations and acquired experimental data, we discovered that during CVVHDF, the filter is an inefficient heat exchanger and acts rather as a buffer-the blood flow (ml/min) was 60 times higher than the dialysate flow (ml/h). The main heat flux takes place on the uninsulated drains. We concluded that using two heaters instead of the single manufacturer's heater (mounted on the inflow and outflow blood drains) allows the most efficient rewarming of the central compartment.
We also found that substitution fluid at room temperature, which is used to replace ultrafiltrate volume, significantly reduces the blood temperature. Our study showed that the measured rewarming rate increases with fast blood flow and minimal replacement fluid flow. The median rewarming rate was highest during blood flow 150 mL/min and substitution 0 ml/h, reaching the maximum value of 6.0 • C/h (IQR 4.8-6.0) [34].
Worthy of note, the use of renal replacement therapy may carry a risk of hemodynamic instability, water-electrolyte, and acid-base disturbances. Also, even small size vascular access itself poses a risk of venous damage, hematoma, thrombosis, and hemorrhage. Moreover, in normothermic patients, the most frequently described adverse event of CRRT is hypothermia, occurring in up to 40% of the therapies [35]. Nowadays, RRT devices use various heating systems, including heating coils, heating sleeves, and heating systems incorporated directly into the bloodstream. This results in improved thermal efficiency of the devices and minimizes patient heat loss.
There are some limitations of our study. Due to the inclusion in the review of only case reports and case series, the risk of bias was rated high. Meta-analysis cannot be performed, and the result is very likely to be subject to publication bias. In addition, no attempt was made to contact the authors in the case of missing data. It should be emphasized that in the reviewed literature, RRT was never the only treatment provided, and the technical aspects and the RRT setting differed between the papers. Moreover, it is challenging to compare continuous methods such as CVVHDF with hemodialysis therapy. Flow rates of warm dialysate in HD ranged from 500 to 800 mL per minute, whereas flow rates in constant RRT are typically set in units of ml per hour. Therefore, they cannot be extrapolated to each other.

Conclusions
The cases reported in the literature did not have sufficient detail to allow other researchers to replicate the study or to allow practitioners to draw conclusions related to their own practice. Based on the reviewed literature, it is currently not possible to provide high-quality recommendations for RRT use in specific groups of patients in accidental hypothermia. The calculated rewarming rate of 1.9 • C/h is the mean value for various RRT techniques, both continuous and intermittent. However, in analyzed studies, RRT was never the only rewarming technique for patients. While RRT appears to be a viable rewarming strategy, the choice of rewarming method should always be determined by the specific clinical circumstances, the available resources, and the current resuscitation guidelines.