Opinions and Management of Hypothermic Sepsis: Results from an Online Survey

Introduction

An altered core body temperature is a hallmark of sepsis. Although fever is most commonly associated with sepsis, hypothermia occurs spontaneously in 10–35% of septic patients at the time of admission (Wiewel et al., 2016) and in an additional 30% of the patients during the intensive care unit (ICU) stay (Fonseca et al., 2016). Development of hypothermia is associated with a twofold increase in mortality compared to patients presenting with fever (Wiewel et al., 2016). This association has led to the hypothesis that hypothermia is a dysregulated, detrimental phenomenon in sepsis.

More recently, however, it has been increasingly recognized that the prognostic value of hypothermia does not imply that it is the hypothermia itself that worsens outcome, the main reason being that those septic patients who develop hypothermia are usually sicker than those who do not (Jiang et al., 2015). The notion that septic hypothermia reflects dysfunction has also been challenged by the finding that hypothermia in septic patients not subjected to active rewarming is a transient, self-limiting response that becomes rare when organ dysfunction is at a maximum in the hours preceding death (Fonseca et al., 2016). The same finding also raises the question as to whether it is necessary to actively rewarm those septic patients who develop hypothermia.

It is currently unknown how these scientific lines of evidence relate to the perceptions of health care professionals regarding septic hypothermia, as is unknown how such perceptions drive clinical conduct concerning the use of active rewarming in these cases. We conducted an online survey on septic hypothermia to address these questions. Using this survey, we aimed to characterize differences in definition and rewarming practices of spontaneous hypothermia during sepsis and determine to what extent these practices are protocolized.

Second, we aimed to assess current opinion on the effect of spontaneous hypothermia and rewarming patients with spontaneous hypothermia during sepsis. We hypothesized that rewarming practices and opinions regarding spontaneous hypothermia would be highly variable.

Methods

Results

In total, there were 440 respondents to the survey. Of these, 288 (66%) of respondents were working in Europe, 67 in Asia (15%), and 39 (9%) in Latin America. 341 (78%) of respondents were intensivists. The main specialty area of respondents was intensive care (288 [59%]) followed by anesthesiology (114 [26%]) and internal medicine (40 [9%]). Respondents mainly worked in an academic hospital (290 [66%]) versus nonacademic hospital (150 [34%]). Reported years of ICU experience were evenly distributed; the largest portion of respondents (28%) reported 0–5 years of ICU experience. Only three (1%) nurses filled in the survey. Table 1 shows the characteristics of respondents in the survey.

Rewarming: definitions and practice

Among the 440 respondents, 96% indicated that there was no written protocol for the management of hypothermic septic patients at their institution. Sixty-two percent of respondents reported that they actively rewarmed patients with spontaneous hypothermia during sepsis, compared to 38% who did not rewarm patients. Respondents were asked to specify per 0.1°C what their definition of hypothermia was. There was a wide range of definitions of hypothermia ranging from 34.0°C to 37.5°C. The most frequent answer was 36.0°C (44%) followed by 35.0°C (15%) (Fig. 1A shows answers grouped per 0.5°C).

OPEN IN VIEWER

FIG. 1.

(A) The temperatures below which respondents consider a patient with sepsis hypothermic. (B) The trigger temperature at which respondents consider rewarming patients with hypothermic sepsis. (C) The target temperature to which respondents rewarm patients. (D) The reported rewarming rates of respondents. (Answers are grouped per 0.5°C).

Respondents who actively rewarmed patients were asked about their rewarming practices. On average, most respondents considered rewarming patients from a temperature of 36.0°C (31%) (Fig. 1B). Forty-eight percent of respondents delayed rewarming of patients starting at a lower temperature than their hypothermia definition.

The target temperature to which patients are rewarmed also varied; most patients were rewarmed to a target temperature of 37°C (28%) (Fig. 1C). There was a large variation in rewarming practice, which ranged from 0.5°C per hour to as fast as possible (Fig. 1D). Among respondents, 58% said that the decision to rewarm patients was predominantly physician driven compared with 42% who said that it was nurse driven.

Respondents were asked which factors influenced their decision to rewarm patients on a scale of 0 (not influential at all) to 5 (extremely influential). Shivering was the most influential (3 [2–4]), followed by the presence of shock (3 [2–3]). The predominant first-line method of rewarming was forced warm air (90%), followed by warmed IV fluids (36%).

Discussion

In this study, we describe the results of a survey on the management of patients with spontaneous hypothermia. The main findings of this study are as follows: (1) both the definition of spontaneous hypothermia and the practice of active rewarming in these patients are extremely variable; (2) a considerable proportion of respondents do not rewarm patients with spontaneous hypothermia during sepsis; and (3) there is no consensus on the etiology of spontaneous hypothermia.

This study is the first to provide detailed insight into temperature management of patients with spontaneous hypothermia in sepsis. The practice of rewarming remains controversial. Considering that hypothermia is associated with an increased mortality, rewarming is understandable, likely in an attempt to negate any adverse effects of hypothermia. The practice of rewarming may be further motivated by physicians' drive to reach physiological values. A majority of respondents in our survey (62%) indicated that they rewarmed patients with hypothermic sepsis. In a recent survey among UK physicians, a majority of physicians (84%) also rewarmed patients, whereas 16% indicated that they would not rewarm septic patients below a temperature of 34°C (Beverly et al., 2016). Only 60% of respondents in this study were concerned with a body temperature of 35–36°C in this patient population (Beverly et al., 2016).

Taken together, the differences in temperature management practice is indicative of a lack of clarity on this issue. The results from this survey make it clear that a robust definition of hypothermia is warranted. This would enable comparisons of incidences and outcome of rewarming practices.

Clinical observational studies overwhelmingly point to hypothermia adversely effecting outcome (Wiewel et al., 2016). In light of these studies, it is noteworthy that none of the respondents reported to have a protocol in place for the management of hypothermic sepsis, most probably due to lack of strong clinical evidence. The aforementioned clinical studies are observational in nature and therefore cannot infer causation between hypothermia and outcome. Moreover, there are no randomized trials studying the impact of rewarming in hypothermic septic patients. If rewarming benefits patients, it is unclear at which temperature they should be rewarmed as well as to which target temperature and how fast patients should be rewarmed. It would be interesting to see the effect that rewarming to target temperatures is in respect to outcome and specifically immune response. In vitro experiments have shown that rewarming can improve immune responses (Billeter et al., 2015).

On the contrary, experimental evidence points to hypothermia being an adaptive response (Steiner et al., 2017). In hypothermic rodents with endotoxemia, thermogenic capacity is intact, arguing against hypothermia being a defect in thermogenic capacity (Romanovsky et al., 1996). Also, hypothermic rodents actively seek colder environments (Almeida et al., 2006) and spontaneous hypothermia also improved survival in mice infected with E. coli (Liu et al., 2012). Moreover, rewarming can potentially have adverse effects such as ischemia reperfusion injury and increased metabolism (Bisschops et al., 2012). A randomized clinical trial would clarify the discrepancy between experimental studies and clinical observations.

There are several limitations that apply to this study. An inherent limitation of this online survey is that we do not know the percentage of respondents or duplicates. To provide an indication of the rate of response to this survey, we have added the number of members of each contacted ESICM section, which provided most of the responses. We used IP address to address potential duplicate responses to the survey. In total, there were 15 duplicate addresses. The majority of these duplicates had different respondent characteristics. In three cases, there were similar respondent characteristics. We did not remove these. Also, the respondents in this survey were predominantly from European countries, limiting the translatability to other countries. Finally, a large proportion of respondents said that rewarming decisions were nurse driven and only three nurses filled in the survey. We do not know how well the physicians' answers reflect rewarming practices by nurses.

Conclusions

In conclusion, policies, procedures, and beliefs about spontaneous hypothermia and active rewarming in patients with sepsis are variable.

The variation in practice must be taken into consideration in designing trials on temperature management in hypothermic sepsis patients. As a starting point, we propose a working group to define hypothermic sepsis to improve comparability of research on this subject.