We measured serum creatine kinase (CK), lactate dehydrogenase (LD), aspartate aminotransferase (AST), and serum alanine aminotransferase (ALT) in 26 heat stroke (HS) victims and 10 control (non-heat-exhausted) subjects during annual Hajj in Makkah, Saudi Arabia. On admission to the HS treatment unit, serum CK, AST, ALT, and LD were higher in HS victims than controls (P <0.05), and at 6, 12, and 24 h were higher than baseline concentration. The patient group was divided into three groups, (a) those who had a quick recovery, (b) those who were critically ill until the end of the Hajj period (7 days), and (c) those who died. Serum enzymes at the time of admission were significantly higher (P <0.05) in the nonsurviving group (n = 6) and the severely ill (n = 9) than in those who had a quick recovery (n = 11). ROC curves were plotted for each enzyme. The most useful indicator was LD, as it could distinguish significantly between the groups who died and those who had a quick recovery (area under the curve = 0.991 ± 0.0286). It was followed by CK and AST as useful prognostic factors. When compared with ROC curves for body temperature, anion gap, and serum potassium, the enzyme results were superior prognostic indicators.
- indexing terms: creatine phosphokinase
- lactate dehydrogenase
- aspartate aminotransferase
- alanine aminotransferase
- receiver-operating characteristic plots
Heat stroke (HS) is a medical emergency characterized by hyperpyrexia, impairment of the level of consciousness, and occasionally multiorgan damage and dysfunction (1)(2).1 It can be exercise induced (exertional) or non-exercise induced (classic), and occurs in humans when heat gain exceeds heat loss from the skin by radiation, convection, or evaporation (3). Several biochemical changes are associated with HS. These include increase in concentrations of serum transaminases, lactate dehydrogenase (LD), and creatine phosphokinase (CK) within 24 h of admission of patients with HS (4)(5). Rhabdomyolysis is a common presentation in the HS patient and is usually accompanied by hypocalcemia and hypomagnesemia (5). Most of these enzymes are found in skeletal and cardiac muscle, and their increase reliably reflects the extent of tissue damage due to thermal injury (6). Serum aspartate aminotransferase (AST) is the most sensitive of these enzymes (7). Although disturbances in serum enzymes and blood chemistry have been reported (4)(5), only few studies have discussed their prognostic value in the HS patient (7)(8).
This study was designed to observe the biochemical changes in HS victims and to determine whether the severity of these changes correlates with survival. ROC curves (9)(10)(11) were used as an index of accuracy to demonstrate the test’s ability to discriminate between survival and death and critical illness vs quick recovery.
Materials and Methods
This prospective study was carried out at the “Heat Stroke Centre” of King Faisal Hospital and Al-Noor Hospital, Makkah, Saudi Arabia, during the 1994 pilgrimage season. The study was conducted on 26 consecutive patients with HS. On admission to the Centre all patients fulfilled the criteria of HS, i.e., rectal temperature ≥40.6 °C associated with hot dry skin and deterioration of level of consciousness after exposure to hot conditions (3). Essential details of the patient were recorded immediately after admission, and a clinical assessment including full neurological examination, which was assisted by Glasgow coma scale, was carried out. Cooling was started immediately by evaporation method as described by Weiner and Khogali (12) and the cooling time was estimated. The cooling time is defined as the time required to reduce the rectal temperature to 38.9 °C. In terms of prognosis, the patients were divided into three groups: group A, those who survived and were discharged: 11 of 26 (42.3%); group B, those who survived but remained unconscious until the end of the pilgrimage period (7 days): 9 of 26 (34.6%); and group C, those who died: 6 of 26 (23.07%). The study group also included 10 controls attending the centre for other minor illnesses but free of hyperpyrexia due to HS or other causes. An initial blood sample was taken, before cooling, by venipuncture into EDTA tubes. After initiation of the cooling process, blood samples were extracted at 6, 12, and 24 h. Hematological parameters were estimated with a Coulter Counter (Coulter Electronics, Harpenden, UK). All blood samples were immediately centrifuged to separate the plasma from the red cells. The plasma was harvested and stored at −70 °C until required for analysis. The plasma was used to estimate the renal function profiles (plasma electrolytes, urea, and creatinine) and plasma enzymes, i.e., CK, LD, alanine aminotransferase (ALT), AST, and electrolytes with an autoanalyzer “American Monitor Parallel” at the Biochemistry Laboratory, College of Medicine, King Saud University, Riyadh. The isoenzymes of LD and CK were estimated after electrophoresis with kits from Helena (cat. nos. 3043 and 3042, respectively; Beaumont, TX).
The performance of LD, CK, ALT, AST, and LD isoenzymes were evaluated by ROC curve analysis (9)(10)(11). A ROC plot was produced for each indicator between groups, i.e., survival (group A) vs death (group C) and critically ill (group B) vs quick recovery (group A), using the 0 time values by plotting the true-positive fraction (sensitivity) against the false-positive fraction (1 − specificity) for multiple test value decision threshold with the original discrete test value data (9)(10)(11). The area under the ROC curves (AUC) were calculated by using “Sigma plot for windows” software. The relative diagnostic accuracies of the different tests were determined by comparing the calculated AUCs ± SE.
The data obtained for the 0-, 6-, 12-, and 24-h samples were analyzed on computers at King Saud University, Riyadh, with the univariate procedure of statistical analysis system (SAS, Cary, NC). To test the significance of the difference in the result of any two groups, the Student’s t-test was used. A P value <0.05 was considered statistically significant.
The essential clinical, physical, and laboratory data of the patients are presented in Table 1⇓ . There were 8 male and 18 female patients with an average age of 55.5 years. Of the 26 total patients, 24 had a rectal temperature of >41 °C, the highest being 43.2 °C. Also, 16 of 26 (61.5%) of the patients had hypotension (systolic blood pressure <90 mmHg), which was corrected in most of the cases after admission. Pretreatment electrocardiogram (ECG) was obtained in all the cases. One patient sustained acute inferior myocardial infarction and died despite inotropic support. The rest of the patients showed evidence of ECG changes as previously described in HS (13). The prognosis was poor in 15 of the patients, 6 of whom expired (5 due to circulatory failure), and the other 9 remained unconscious until the end of the pilgrimage period (they were classified as the critically ill group). One patient was lost to follow-up, and 11 recovered and were discharged (they were classified as the quick-recovery group). This gave us interesting circumstances to differentiate the concentrations of biochemical indicators in patients with different prognoses. The pretreatment values of CK, LD, AST, and ALT on admission were significantly higher than those of the controls (P <0.05) (Table 2⇓ ). The mean ± SD serum concentration of each enzyme for each group at different time intervals are shown in Fig. 1⇓ . The concentrations continued to increase and were significantly higher in the nonsurviving group at all time intervals after admission (P <0.05). The value of the LD isoenzyme proportion at different time intervals remained unchanged, as shown in Table 3⇓ .
ROC curves were plotted for each of the enzymes and the isoenzymes of LD. Fig. 2⇓ presents the ROC curves for LD, CPK, AST, and ALT between the survivor group (group A) and the group that died (group C). The highest prognostic accuracy for patients who survived compared with those who died was shown by total LD, followed by CK, AST, and ALT, which was the least predictive of all the enzymes tested. LD isoenzymes were also useful, where the best results were obtained with LD-1 and -2, followed by LD-3, whereas LD-4 and -5 were not of much significance (results not plotted). The AUCs were obtained for each indicator, including the body temperature and the anion gap, and the results are presented in Table 4⇓ . In the quick recovery vs dead group, total LD ROC had an AUC of 0.991 ± 0.0286 compared with 0.910 ± 0.0875 for CK, and 0.901 ± 0.0887 for AST and 0.550 ± 0.151 for ALT. For body temperature and anion gap the AUCs were 0.908 ± 0.0886 and 0.708 ± 0.1402 respectively.
Pilgrimage to Makkah is a ritual undertaken at least once during his or her lifetime by every Muslim and ∼2 million pilgrims perform it at the same time every year. HS is a frequently encountered problem among the pilgrims, as Makkah lies at a latitude of 20 degrees and the temperature exceeds 45 °C (113 °F) during the summer months. Lack of acclimatization and physical exertion play a significant role in the development of HS. Hypotension was a major problem in the majority (62%) of our patients. Sprung demonstrated that hypotension predicts poor outcome (14), although Hart et al. (8) did not demonstrate similar findings. Previous studies (14)(15)(16) showed that HS patients may present with hyperdynamic or hypodynamic state, the latter group presenting with circulatory failure and having a poorer outcome. In our study, 23% of our patients died of circulatory failure and all required inotropic support.
Data from human studies (17) suggest that temperature between 41.6 °C and 42 °C is the critical thermal maximum and results in the increase in the concentration of AST, ALT, LD, and serum alkaline phosphatase. Similarly, artificial heating of dogs to a rectal temperature of 41.4 °C resulted in an increase in the serum concentration of AST and ALT (18). It has been demonstrated that such enzymes increase as a result of physical exertion alone (19)(20), with the suggestion that physical training reduces the exercise-induced increments of most serum enzymes (21)(22). In experimental animals (23)(24), tissue damage and mortality with exertion-induced hyperthermia increases more than in simple hyperthermia due to heat exposure. Kew et al. (25) suggested that prognosis is poor in those with AST >1000 U in the first 24 h. In our patients the AST concentration was increased in the first 24 h and was significantly higher in the group that died. However, only one patient had AST >1000 U. The mean pretreatment concentrations were greater and remained so throughout the 24-h period in the group that died. Although patients who were critically ill showed a progressive increase in liver enzymes, the concentrations were lower than in those who died. In patients who survived, the concentrations of the liver enzymes tended to decline after 24 h. When the AUC of ROC plots were measured, LD, CK, and AST were much more useful in differentiating between survival and death (AUC >0.9), and this supports the previous reports (25). Furthermore, AST and LD were the best indicators to differentiate between the severely ill vs the quick recovery groups (AUC >0.8).
The total LD ROC curve appears to have the highest accuracy in predicting death. Similarly, LD-2 was the most predictive of the isoenzymes in predicting quick recovery or death and also in differentiating between quick recovery and critical illness. Even the anion gap (AUC = 0.908) was less sensitive in predicting death compared with LD. Van der Linde et al. (26) examined the role of LD isoenzymes in heated exercised rats, and concluded that the plasma isoenzymes could be useful diagnostically and prognostically in HS. Wolf et al. (27) also observed increased LD but did not demonstrate changes in the proportion of the five isoenzymes of LD after acute physical stress in exercising athletic men. Kew et al. (25) demonstrated increased LD-5 in HS patients. As HS damages many tissues with different LD isoenzyme, it is not surprising that the isoenzyme proportions were not changed although their absolute values were high.
Various pathological changes have been reported in the liver in HS (28). Rubel and Ishak (29) described fatty and nonfatty vascularization in addition to congestion in their patients who died within 6 h after therapeutic hyperthermia. The mechanism of hyperthermia-induced liver injury is uncertain; most likely it is related to direct thermal injury on the liver cells (30)(31). Rowell et al. (32) had suggested in experiments on volunteers subjected to a combination of exercise and hyperpyrexia that the temperature of hepatic venous blood is often as much as 1.5 °C warmer than the measured core body temperature, and this increase in the body temperature may predispose to hepatic cellular injury. Furthermore (32), glucose outpouring and increase of hepatic lactate concentration are the result of hepatic splanchnic hypoxia. Also, plasma renin activity increases during heating, which in turn increases vascular resistance of splanchnic blood vessels and induces further vasoconstriction and hypoxic damage (33). Also during heat stress, vasodilation of cutaneous blood vessels and redistribution of blood away from the internal circulation to the skin to maximize the heat dissipation may contribute to hepatic hypoxia and cell injury.
In our patients CK was significantly increased and differentiated between those who recovered and those who died (AUC = 0.910). Changes in the serum CK have been well documented [5, 8, 34] in patients with HS, and the enzyme is believed to be derived mainly from skeletal muscles and possibly from the liver (35). It may be accompanied by myoglobinuria, and significant rhabdomyolysis has been observed commonly in patients with the exertional type of HS (36). Fowler et al. (37) noted high total serum CK concentrations after muscular exercise in poorly conditioned men as compared with conditioned subjects. The pilgrimage rites necessitate considerable physical exertion in subjects who are generally poorly conditioned and unacclimatized to the high ambient temperature. Another explanation for the raised plasma CK is associated hypokalemia (29). Interestingly, hypokalemia was found in 42.3% of our patients, and in the rest of the patients the potassium concentrations were in the lower end of the normal range. This was found despite the presence of significant metabolic acidosis (which causes hyperkalemia), especially in those patients who died. The mean anion gap metabolic acidosis in the total group was 15.6, whereas in patients who did not survive it was 20.2. The possible explanation for hyperkalemia is excessive sweating (38), diarrhea, and vomiting, which are common features of HS. Furthermore, high levels of renin activity occur because of fluid depletion (38) and possibly excessive production of aldosterone, which leads to further conservation of sodium and depletion of potassium. Hypokalemia is well known to increase cellular permeability and loss of muscle cell membrane integrity, which will subsequently lead to increased production of CK. Knochel and Schlein had suggested, after experiments on dogs, that hypokalemia might lead to ischemia and muscle injury (39).
Another coexisting factor that might contribute to rhabdomyolysis in the presence of hypokalemia is the impairment of glycogen storage and synthesis in skeletal muscle, which may induce anaerobic metabolism and consequently the skeletal muscle becomes less efficient, thus making the subject vulnerable to HS illnesses (38). In addition, in vitro studies have demonstrated that membrane potential sharply decreases to abnormally low concentrations when potassium deficiency advances to 30% (38).
It has been questioned whether alterations observed in plasma concentrations of various indicators in the different groups may be a consequence of dehydration. The value of hematological indicators, in particular, hematocrit value, did not show any significant differences between the three groups (Table 1⇑ ) and pointed to the fact that the patients in our different groups were not dehydrated. In addition, earlier reports by Seraj et al. (40), who utilized central venous pressure to determine fluid depletion in HS patients, showed that 64.7% of the patients had normal or above-normal central venous pressure. Dahmash et al. (41), on the other hand, who measured pulmonary capillary wedge pressure (PCWP) after right-heart catheterization on 10 HS patients, showed that only one of these patients had low PCWP. Thus it was ruled that hypovolemia is not of major consequence in HS.
In conclusion, we have described the changes of serum enzymes and isoenzymes in HS and their role in the pathophysiology of HS, and also the role of potassium in rhabdomyolysis. This study is unique in that the patients were studied prospectively and compared with controls who were under the same environmental conditions, and shows that plasma enzymes can be a useful indicator of the prognosis of HS.
We thank the secretarial assistance of Ofelia S. Gurrea-Villamil, Hassan Al-Zahrani’s technical support, Ahmed Ali, and Edward De Vol for statistical assistance.
↵1 Nonstandard abbreviations: HS, heat stroke; LD, lactate dehydrogenase; CK, creatine kinase; AST, aspartate aminotransferase; ALT, alanine aminotransferase; AUC, area under the curve; ECG, electrocardiogram; and PCWP, pulmonary capillary wedge pressure.
GCS, Glasgow coma scale; MAP, mean arterial pressure; APACHE II, acute physiology and chronic health evaluation.
1 = time of admission; 2 = 6 h after cooling; 3 = 12 h after cooling; 4 = 24 h after cooling.
- © 1997 The American Association for Clinical Chemistry