Serum potassium changes during hypothermia and rewarming: a case series and hypothesis on the mechanism

ABSTRACT

Introduction

Hypokalemia is known to occur in association with therapeutically induced hypothermia and is usually managed by the administration of potassium (K⁺).

Methods

We reviewed data from 74 patients who underwent a therapeutic hypothermia protocol at our medical institution.

Results

In four patients in whom data on serum K⁺ and temperature were available, a strong positive correlation between serum K⁺ and body temperature was found. Based on the close positive relationship between serum K⁺ and total body temperature, we hypothesize that serum K⁺ decreases during hypothermia owing to decreased activity of temperature-dependent K⁺ exit channels that under normal conditions are sufficiently active to match cellular K⁺ intake via sodium/K⁺/adenosine triphosphatase. Upon rewarming, reactivation of these channels results in a rapid increase in serum K⁺ as a result of K⁺ exit down its concentration gradient.

Conclusion

Administration of K⁺ during hypothermia should be done cautiously and avoided during rewarming to avoid potentially life-threatening hyperkalemia. K⁺ exit via temperature-dependent K⁺ channels provides a logical explanation for the rebound hyperkalemia. K⁺ exit channels may play a bigger role than previously appreciated in the regulation of serum K⁺ during normal and pathophysiological conditions.

Graphical Abstract

Graphical Abstract.

INTRODUCTION

Induced therapeutic hypothermia (ITH) has been a commonly recommended intervention to improve survival as well as neurological outcomes in comatose patients who have survived prolonged cardiac resuscitation [1–8]. Target temperatures of 32°–34°C (or 91°–92°F) were initially used in hypothermia protocols [1–3]. More recently, a less aggressive approach has been recommended with a target temperature of 36°C (96.8°F) [7, 8]. Nevertheless, hypokalemia is well known to be associated with hypothermia [9–17], but the mechanism whereby hypothermia changes extracellular potassiumm (K⁺) concentration has never been elucidated. Mechanisms that have been suggested to explain the hypokalemia include activation of the sympathetic nervous system [18], altered insulin secretion [19–23], a membrane-stabilizing effect of hypothermia [24] and enhanced renal K⁺ excretion during cold diuresis [25]. While hypothermia can be associated with an increase in circulating catecholamines, animal studies have demonstrated that β-adrenergic blockade had no effect on hypothermia-induced hypokalemia [12]. Moreover, hypothermia is associated with a decrease, not an increase, in insulin secretion [19–23]. Enhanced kaliuresis is also highly unlikely since serum K⁺ still decreased with hypothermia in bilaterally nephrectomized dogs [9] and during bilateral ligation of rat ureters [12]. Therefore none of these proposed explanations satisfactorily explains the hypokalemia associated with hypothermia.

Internal redistribution of K⁺ is the most likely mechanism involved in the development of hypokalemia during hypothermia. Much of the clinical knowledge of changes in serum K⁺ due to internal shifts is attributed to K⁺ entry into the cell via the sodium (Na⁺)/K⁺/adenosine triphosphatase (ATPase) pump. Consequently, alterations in this pump have been considered as a cause of hypokalemia during hypothermia [26–28]. however, hypothermia, has an inhibitory effect on metabolic processes, including decreasing ATP, which should reduce the activity of the Na⁺/K⁺/ATPase pump [26–28]. In fact, these suppressive effects are considered part of the cellular-protective effect of hypothermia [29]. Hypothermia-induced suppression of the Na⁺/K⁺/ATPase pump activity would decrease cell K⁺ entry and thereby result in an increase in extracellular K⁺, not a decrease. Recovery of pump activity upon rewarming would promote cell K⁺ entry and decrease extracellular K⁺. Such effects on serum K⁺ are just the opposite of what is actually observed as documented in this report. Accordingly, mechanisms other than primary changes in Na⁺/K⁺/ATPase activity need to be considered to explain changes in extracellular K⁺ during hypothermia and rewarming. The present study investigated the changes in serum K⁺ in relation to body temperature during hypothermia and rewarming. This was prompted by a patient with hypothermia-induced hypokalemia who developed cardiac arrest and expired from severe fatal hyperkalemia during rewarming. In this patient a striking relationship between serum K⁺ and body temperature could be documented during rewarming. This prompted a search for more cases of changes in serum K⁺ during hypothermia and rewarming in our medical center. The rapid changes in serum K⁺ and the striking correlation of serum K⁺ with temperature that we were able to document in four cases prompted us to hypothesize that the cellular exit pathway, likely via temperature-dependent K⁺ channels, is the mechanism responsible for the development of hypokalemia during hypothermia and hyperkalemia during rewarming.

MATERIALS AND METHODS

The index case that prompted this study of K⁺ changes during ITH was a 60-year-old male who died during rewarming with rapid hyperkalemia (see below). We subsequently reviewed the records of 74 patients who underwent the ITH protocol after cardiac arrest at Northwestern Memorial Hospital. This review was approved by the institutional review board of Northwestern University.

All patients had to meet the following inclusion criteria: ≥18 years of age, hypothermia started within 12 hours of return of spontaneous circulation following cardiac arrest, serum K⁺ of ≤3.5 mEq/L, systolic blood pressure ≥90 mmHg (either spontaneously or with any combination of fluids, vasopressors or blood pressure augmentation devices), no encephalopathy as evidenced by not being able to follow commands after return of spontaneous circulation and a prearrest functional status with acceptable quality of life according to family or the primary team attending physician if family was not available. Patients were cooled using an external blanket to achieve a target core body temperature of 91.4°F (33°C). The temperature was kept as close to 91.4°F (33°C) as possible for 22–24 hours. After 22–24 hours, the temperature was gradually increased to 98.6°F (37°C) over 16 hours at a rate of 0.45°F/h (rewarming phase). Sedation and analgesia were administered to prevent shivering and provide patient comfort.

Exclusion criteria included other causes of coma unlikely to be reversible or that may be exacerbated by hypothermia treatment (e.g. intracerebral hemorrhage), an imminently pre-morbid state prior to cardiac arrest, known pre-existing severe coagulopathy that posed a high risk of life-threatening pathological bleeding, do not resuscitate status and severe sepsis or uncontrolled infection in the opinion of the primary medical team. In an attempt to eliminate confounding factors that could affect K⁺ homeostasis, we also excluded those patients who had been hospitalized for >48 hours prior to the induction of hypothermia, who had a serum creatinine >1.3 mg/dL  and whose body temperature did not achieve the hypothermia target of 91.4°F (33°C). In addition, patients who were given any K⁺ supplements during the rewarming phase were excluded so the relationship between serum K⁺ and body temperature could be examined without this confounding variable. Many patients had to be excluded because there were no data on serum K⁺ levels to match the temperature within 30 minutes of each other. With these criteria only three patients in addition to the index case were available with complete data during the hypothermia and rewarming phases. The relationship between serum K⁺ and body temperature was examined by simple linear regression analysis. The data were converted to Celsius in all graphs.

RESULTS

Index case (Case 1)

The patient was a 60-year-old Caucasian male admitted with ventricular fibrillation and cardiac arrest. He underwent cardiopulmonary resuscitation and was intubated. Subsequently, ITH was implemented. Serum K⁺ declined during ITH from 4.1 mEq/L before hypothermia to a lowest level of 2.7 mEq/L. Because of the precipitous and continuous decrease in serum K⁺, the management team administered increasing doses of K⁺ for a total of 420 mEq of K⁺. Despite this, serum K⁺ remained <3 mEq during the cooling phase (Fig. 1) and ventricular ectopy occurred. The patient had normal kidney function based on serum creatinine (1.1–1.3 mg/dL) and urine output was consistently between 100 and 200 cc/h.

Figure 1:

Serum K⁺ and temperature over time during hypothermia (blue) and during rewarming (red) in Case 1. Temperature as a function of time and serum K⁺ as a function of time are shown in the upper and the middle panels, respectively. The measurements of K⁺ and temperature are shown in the lower panel and were taken within no more than 30-minute intervals of each other. A very strong positive correlation was found (r = 0.915, P < .001) between K⁺ and temperature (lower panel).

During rewarming, serum K⁺ rose to a peak of 6.8 mEq/L, the electrocardiogram showed sine waves resulting in cardiac arrest and he expired. Of importance, no K⁺ was administered during rewarming and serum K⁺ did not increase until rewarming, which was started 24 hours later (Fig. 1). A direct positive correlation was found between body temperature and serum K⁺ during the rewarming phase (r = 0.915, P < .001) (Fig. 1). Blood pH was in the range of 7.01–7.33 and bicarbonate was 13–24 mEq/L throughout the entire episode.

Additional patients

In order to examine the effect of temperature on serum K⁺, we retrospectively examined the patients who had undergone ITH at our hospital (see the Methods section). Of a total of 74 patients, 44 (59.5%) developed hypokalemia (range 1.9–3.5 mEq/dL) on at least one measurement. K⁺ was administered during the cooling phase as deemed clinically necessary.

As noted in the Methods section, we found only three patients in addition to the index case in whom complete data on serum K⁺ data and temperature were available within 30 minutes of each other to examine the relationship between serum K⁺ and temperature during hypothermia and rewarming (Fig. 2). In Case 2, serum K⁺ prior to hypothermia was 5 mEq/L and decreased to 3.3 mEq/L despite a total of 160 mEq of potassium chloride (KCl) intravenously administered during hypothermia. During rewarming, serum K⁺ increased to 5.1 even though no K was administered during this phase. A direct positive correlation was found between body temperature and serum K⁺ (r = 0.878, P = .002) (Fig. 2, left panel).

Figure 2:

Serum K⁺ and temperature over time during hypothermia (blue) and during rewarming (red) in Cases 2, 3, and 4. Temperature and serum K⁺ as a function of time are shown in the upper and middle panels, respectively. The measurements of K⁺ and temperature are shown in the lower panel and were taken within no more than 30-minute intervals of each other. A strong positive correlation between serum K⁺ and temperature was found in all three cases (Case 2: r = 0.878, P = .002; Case 3: r = 0.945, P < .001; Case 4: r = 0.878, P < .001) (lower panel).

In Case 3, serum K⁺ prior to hypothermia was 4 mEq/L and decreased to 3.3 mEq/L despite a total of 180 mEq of KCl intravenously administered only during hypothermia. During the rewarming phase, serum K⁺ increased to 4.5 mEq/L even though no K was administered. A direct positive correlation was found between body temperature and serum K⁺ (r = 0.945, P < .001) (Fig. 2, middle panel).

In Case 4, serum K⁺ prior to hypothermia was 4 mEq/L and decreased to 3.5 mEq/L despite a total of 140 mEq/L of KCl intravenously administered during hypothermia. During rewarming, serum K⁺ increased to 5.8 mEq/L even though no K was administered. A direct positive correlation was found between body temperature and serum K⁺ (r = 0.878, P < .001) (Fig. 2, right panel).

DISCUSSION

ITH is known to be associated with hypokalemia but, as noted, the mechanism is not well understood [10–12, 15, 17, 27, 30]. Out of the 74 patients we were able to survey who had undergone ITH at our institution, 44 (59.5%) developed hypokalemia despite the administration of K⁺ supplements during the cooling phase, which is consistent with other studies [9–11]. The development of hyperkalemia during rewarming, however, has been described less commonly [31], although a recent multicenter study by Kirkegaard et al. [32] reported mild disturbances in serum K⁺ and other electrolytes, both during hypothermia and rewarming. The present report shows that hypokalemia during hypothermia can convert during rewarming to hyperkalemia that could be life threatening, as in the index case that prompted this study. In this patient, serum K⁺ during rewarming increased to 6.8 mEq/L and resulted in cardiac standstill and death. We examined the relationship between serum K⁺ and temperature during the rewarming phase and found a positive correlation in this patient and in three additional patients in whom sufficient data for analysis of such a relationship were available (Figs. 1 and 2). Based on the direct correlation between the increase in temperature and in serum K⁺, we propose the hypothesis that the changes in serum K⁺ are primarily attributable to exit K⁺ channels that are temperature dependent (Fig. 3 and as discussed below).

Figure 3:

Schematic diagram of a skeletal muscle cell with cell K⁺ entry via Na⁺/K⁺/ATPase and exit via the K⁺ exit channels, which are temperature dependent (green) and independent (orange). During hypothermia (middle panel), hypokalemia develops because of closing/inactivation of the K⁺ temperature-dependent K⁺ exit channels. During rewarming, serum K⁺ increases because of the opening/reactivation of these temperature–dependent K⁺ channels. In the scheme and for simplicity, it is assumed that the temperature-independent K⁺ channels are not involved and that Na⁺/K⁺/ATPase either is not affected by temperature or is decreased during hypothermia and increased during rewarming but the dominant effect on K⁺ transport is that of the temperature-dependent K⁺ channels.

About 98% of the total body content of K⁺ resides within the cell and serum K⁺ is determined by the interplay between internal K⁺ distribution and external K⁺ balance [33, 34]. Acute shifts of intracellular K⁺ into or out of the extracellular space can cause severe, even lethal, derangements of extracellular K⁺ concentration [33–36]. The movement of K⁺ between the intracellular and the extracellular compartments likely is the result of the balance between the activity of the Na⁺/K⁺/ATPase pump (entry pathway) and the K⁺ channels (exit pathway, predominantly in skeletal muscle) (Fig. 3). Current thinking in this area places the extensively studied Na⁺/K⁺/ATPase pump at the center of these transcellular shifts [37, 38]. This pump clearly offers the pathway of K⁺ entry into cells, but its activity alone cannot account for the transcellular shifts that are observed clinically and experimentally as this study shows using hypothermia and rewarming as the paradigm. Evidence that a decrease in temperature suppresses Na⁺/K⁺/ATPase activity is available [28]. However, suppression of this pump would cause an increase in extracellular K⁺ in hypothermia and not a decrease. This increase in extracellular K⁺ would then normalize or decrease during rewarming, which is exactly the opposite of what was observed in the four patients studied in detail. Current clinical reviews of internal K⁺ transfers provide limited information regarding the nature and role of K⁺ exit channels [16]. The role of K⁺ exit channels in extrarenal K⁺ handling, however, can be inferred from the effect of barium toxicity causing profound hypokalemia by blocking the K⁺ inward rectifier channels [39]. The changes we observed in serum K⁺ during hypothermia and rewarming point to a critical role of K⁺ exit channels that are temperature dependent and allow the transfer of K⁺ from the intracellular to the extracellular compartment down its concentration gradient. We therefore hypothesize that hypokalemia develops during hypothermia owing to decreased exit of K⁺ via such channels and that during rewarming the extracellular K⁺ concentration rises (Fig. 3) when such channels are reactivated.

Temperature dependence is inherent in practically all types of ion currents as corresponding current-conducting channels change their biophysical properties with temperature [40]. Hodgkin and Huxley showed, >60 years ago, that temperature could profoundly influence membrane excitability [41, 42]. Heat contributes to activation energy for channel conformational changes and the shape of action potentials [43, 44], although within the physiological temperature range, thermal energy alone is generally insufficient to initiate channel activation [42]. K⁺ channels that are expressed in skeletal muscle and are temperature dependent are likely those involved in the changes in serum K⁺ observed during hypothermia and rewarming. What could these channels be? There are three major families of K⁺ channels: voltage-gated (Kv), inward rectifier (Kir) and two-pore domain (K2P) channels [45–47]. Of relevance to the proposed mechanism of temperature-dependent changes in serum K⁺ during hypothermia and rewarming are Kv channels, which conduct K⁺ ions extracellularly along the concentration gradient, leading to repolarization [48], and K2P channels, which are described as leaky K⁺ channels [47, 49].

Kv channels are found in several cell types, including skeletal muscle [48]. A study done on human T-cell lymphocyte K⁺ conductance revealed profound effects of temperature on the properties of Kv channels [50]. In this study, Kv channels demonstrated an increase in preexisting rates of inactivation of K⁺ currents at 22°C (71.6°F). The whole cell conductance increased when temperature was increased from 22°C (71.6°F) to 37°C (98.6°F), with the effect being reversed when temperature was returned to 22°C (71.6°F). Another study reported that native Kv2.1 in pancreatic β-islet cells of rats showed greatly enhanced inactivation at near physiological temperatures of 32–35°C (89.6–95°F) [51], while Yang and Zheng [42] revealed the gating of Kv2.1 and Kv4.3 channels to be highly temperature sensitive.

Temperature sensitivity was also described in the K2P channels, specifically in the TREK-1, TREK-2 and TRAAK channels [40, 49, 52]. Of note, TREK-1 and TRAAK have been cloned from human skeletal muscle [47]. At room temperature, TREK-1 exhibits only background K⁺ leak, which increases with temperature, reaching a maximum at 42°C (107.6°F) [49]. TREK-1 has its half-maximal temperature activation point at 37°C, implying that the midpoint of the channel's dynamic range is centered on the homeostatic thermal set point for most mammals. If this property is maintained in native cells, then TREK-1 activity is set to be maximally sensitive to minute variations in physiological temperature [49].

The temperature-dependent activation of the K⁺ channels mentioned above is consistent with the observed K⁺ shifts during hypothermia and recovery that must take place in skeletal muscles, the main reservoir of intracellular K⁺. If we extrapolate that the K⁺ shifts across the cell parallel the K⁺ conductance, one can easily see that during recovery from hypothermia (rewarming), K⁺ exit from these channels down the chemical gradient must result in an increase in extracellular K⁺ (Fig. 3). From the preceding description of the characteristics of temperature-dependent K⁺ channels, it is reasonable to postulate as potential candidates that the channels likely involved in K⁺ exit from skeletal muscle cells are Kv2.1, which are voltage gated, and TREK-1 and TRAAK, two-pore channels, which are not voltage gated [47, 48]. We speculate that with therapeutic hypothermia, skeletal muscle activity of these K⁺ exit channels (obviously we cannot determine specific subtypes involved based on our clinical observations) decreases markedly during hypothermia, leading to a decrease in serum K⁺, and recovers during rewarming, leading to an increase in serum K⁺ (Fig. 3).

From our observations, one can also infer that temperature-dependent K⁺ exit channels are not totally quiescent at physiological temperatures but are sufficiently active to keep up with K⁺ entry into the cell via Na⁺/K⁺/ATPase. When K⁺ entry slows down in hypothermia due to inactivation of Na⁺/K⁺/ATPase, the net decrease in serum K⁺ must be from of a cessation of K⁺ exit via these otherwise active channels. During rewarming, serum K⁺ increases in the absence of K⁺ supplementation in our selected cases, which would reflect the restoration of K⁺ exit via reactivation of these temperature-dependent K⁺ channels (Fig. 3). Accordingly, the activity of skeletal muscle K⁺ exit channels, both temperature dependent and possibly others that are not temperature dependent, probably plays a previously unrecognized role in the day-to-day regulation of K⁺ transfers in and out of the cell.

While the regulation of K⁺ entry by insulin and β-adrenergic receptors is well studied [12, 53–56], very little is known about the regulation of K⁺ exit channels. Moreover, in pathophysiologic situations, changes in serum K⁺ may involve K⁺ exit channels. The most obvious situation is the hyperkalemia of vigorous exercise, a phenomenon that cannot be attributed to changes in Na⁺/K⁺/ATPase activity, since it is a pathway of cell K⁺ entry, not exit. Yet there are no studies, to our knowledge, investigating specifically the channels involved in cellular K⁺ exit during exercise. Other conditions such as metabolic acidosis with acidemia may cause a net shift of K⁺ from the intracellular to the extracellular space [57]. The effect of acidemia and/or low bicarbonate on K⁺ exit channels is not known, to our knowledge, again reflecting the limited information about putative K⁺ exit channels in skeletal muscle. The role of aldosterone on transcellular shifts of K⁺ has been debated, with most studies showing no effect, as recently reviewed [58]. The presence of the mineralocorticoid receptor in nonepithelial tissues like the kidney, however, is well described. Of note, brown adipose tissue, which is involved in thermogenesis, has a mineralocorticoid receptor [59].

In conclusion, administration of K⁺ should be done cautiously during hypothermia and avoided during the rewarming phase to prevent severe hyperkalemia that can be lethal, as in our index case. We hypothesize that this increase in serum K⁺ during rewarming is most reasonably explained by reactivation of temperature-dependent K⁺ exit channels. These channels are likely inactivated during hypothermia, leading to impaired cell K⁺ exit and thus hypokalemia. Moreover, our findings illustrate the need for further research on K⁺ exit channels that act in concert with the activity of Na⁺/K⁺/ATPase, which is the best studied pathway for K⁺ entry into the cell.

CONFLICT OF INTEREST STATEMENT

The results presented in this article have not been published previously in part or in whole, except in abstract format. D.B. is coinventor of patents entitled ‘Active low molecular weight variants of angiotensin converting enzyme 2 (ACE2)’, ‘Active low molecular weight variants of angiotensin converting enzyme 2 (ACE2) for the treatment of diseases and conditions of the eye’ and ‘Soluble ACE2 variants and uses therefor’; is founder of Angiotensin Therapeutics Inc.; has received consulting fees from AstraZeneca, Relypsa and Tricida, all unrelated to this work and received unrelated support from the National Institute of Diabetes and Digestive and Kidney Diseases (grant R01DK104785) as well as from a grant from AstraZeneca. All remaining authors have nothing to disclose related to this publication.