FREE NEJM E-TOC    HOME   |   SUBSCRIBE   |   CURRENT ISSUE   |   PAST ISSUES   |   COLLECTIONS   |   Search Term  Advanced Search

Sign in | Get NEJM's E-Mail Table of Contents — Free | Subscribe

Previous Volume 330:19-24 January 6, 1994 Number 1 Next

Abstract Add to Personal Archive Add to Citation Manager Notify a Friend E-mail When Cited PubMed Citation

ABSTRACT

Background Arterial gas embolism due to pulmonary barotrauma and the resultant cerebral gas embolism are catastrophic complications of diving. Previous studies have only rarely noted evidence of gas embolism to noncranial sites.

Methods Among 142 persons with diving-related injuries evaluated between January 1982 and July 1991, we identified 29 who had arterial gas embolism and who underwent biochemical studies indicative of muscle injury. Of the 29 patients, 4 were excluded because cardiopulmonary resuscitation had been performed and 3 were excluded because the duration of their dives met or exceeded standard limits set for dives not requiring staged decompression. The outcome at the time of hospital discharge in the remaining 22 patients was correlated with clinical factors and the results of biochemical studies. We also studied 22 subjects after uncomplicated dives and 11 patients who had sustained blunt trauma.

Results All the patients with diving-associated gas embolism had elevated serum creatine kinase activity (normal, 175 U per liter); the values were markedly elevated (>900 U per liter) in 14. The MB isoenzyme of creatine kinase was detected in the serum of 13 of 20 patients in whom it was measured and was 4 percent of total creatine kinase activity in 6 patients. In three patients electrocardiography showed myocardial injury. Changes in serum creatine kinase activity of similar magnitude were not present in the subjects who had uncomplicated dives or in the patients with blunt trauma. Thirteen patients recovered fully, four had minor residual neurologic deficits, three were severely impaired, and two died. Logistic-regression analysis revealed a significant correlation between peak serum creatine kinase values and clinical outcome.

Conclusions Biochemical evidence of muscle injury is frequently found after diving-associated arterial gas embolism. The correlation between serum creatine kinase activity and outcome suggests that serum creatine kinase is a marker of the size and severity of arterial gas embolism.

We describe our experience with diving-associated arterial gas embolism in patients treated between 1982 and 1991 at a referral center for hyperbaric medicine and diving-related accidents. All had elevated serum creatine kinase activity, suggesting the presence of widespread skeletal-muscle embolization. These observations suggest that the gas emboli caused by a diving accident are distributed widely through the systemic vascular bed, possibly including the coronary circulation.

Methods

We reviewed the medical records of 142 patients treated for diving-related injuries from January 1982 to July 1991. Among them we identified 29 with barotraumatic arterial gas embolism whose records contained the results of assays for creatine kinase in serum. The diagnosis of arterial gas embolization was based on the presence of neurologic symptoms consistent with the obstruction of blood flow to any part of the cerebral circulation and on the onset of these symptoms during ascent or immediately on reaching the surface from a dive requiring the breathing of compressed air. Four patients were excluded because they underwent cardiopulmonary resuscitation, including chest compression, either before or during hospitalization. Three additional patients were excluded because the duration of their dives met or exceeded the limits recommended by the U.S. Navy for dives not requiring staged decompression. These patients were excluded in order to reduce the likelihood that patients in this series might have a combination of arterial gas embolism and decompression sickness or might have cerebral decompression sickness rather than arterial gas embolism.

During the hospitalization, diagnostic studies, including determinations of arterial-blood gases, serum biochemistry tests, and electrocardiography, were ordered at the discretion of the treating physician and performed with the use of routine techniques. Serum creatine kinase activity was measured on an automated analyzer with a spectrophotometric assay (Beckman ASTRA or Kodak EKTACHEM) (normal level, 175 U per liter). Assays to identify and quantitate the isoenzymes of creatine kinase were performed by cellulose acetate electrophoresis of appropriately diluted samples⁸. Normal values for the MM (from skeletal muscle), MB (from myocardial muscle), and BB (from brain tissue) isoenzymes were 96 to 100 percent, 0 to 4 percent, and 0 percent, respectively. Outcome was based on the patient's status at the time of discharge from the hospital as documented in the medical record. Patients who survived to leave the hospital were considered to have no impairment if there were no neurologic abnormalities at the time of discharge, to have minimal impairment if there were residual neurologic abnormalities that did not interfere with the activities of daily living, and to have severe impairment if the residual neurologic abnormalities interfered with the activities of daily living.

The results in these patients were compared with those of two control groups. One was a diving control group of 22 subjects who consented to venipuncture after undergoing compressed-air dives (typically training dives). None of these subjects had neurologic symptoms of arterial gas embolism or decompression sickness. We also studied a group of 11 patients who were evaluated for myocardial contusion after severe blunt trauma; patients with fractures of long bones were excluded.

Correlations between outcome and clinical data were examined by logistic-regression analysis^9,10. The correlations were considered to be significant if the P value (two-tailed) was less than 0.05. Unless otherwise indicated, the results are presented as means ±SE when values were normally distributed. Confidence intervals for proportions were estimated with standard techniques¹¹. To examine the relation between the length of time after the diving accident and subsequent biochemical alterations, patients with at least four separate determinations of serum creatine kinase were identified, and their serum creatine kinase values were expressed as a fraction of their peak value and plotted against the interval between gas embolism and sample collection. Polynomial regression was performed (orders 1 to 5), and 95 percent confidence intervals were calculated with nonlinear least-squares techniques on a microcomputer.

Results

Characteristics of the Subjects

The 22 patients with arterial gas embolism ranged in age from 16 to 54 years (mean, 28). Three were women, and 19 were men. No patient had any history of cardiac disease before the diving accident. Seventeen of these patients had less than one year of experience as a diver, and 11 had not yet received certification as a diver. In all patients, the durations of the dives were less than 90 percent of the U.S. Navy time limits recommended for dives not requiring staged decompression, and in 19 patients the durations of the dives were less than 50 percent of the recommended time limits (Figure 1). A history of an abnormally rapid ascent was obtained for 18 of the 20 patients who could give a history of the ascent or whose ascent was witnessed. The other two patients were unable to provide a history, and their ascents were not witnessed. Fifteen patients were described as being initially unresponsive at the surface, six were confused, and one had acute visual changes. After the onset of symptoms, a mean of 5 ±1 hours elapsed before recompression therapy was instituted. All but one patient received recompression therapy according to a standard U.S. Navy schedule. There was no history of muscle trauma in any patient, and no evidence of muscle injury was described on initial or subsequent examinations. No patient received intramuscular medication during or after resuscitation efforts. For each patient, 1 to 11 assays of serum creatine kinase activity were performed. The assays were performed more frequently in the patients with the highest initial creatine kinase values, but no consistent schedule was followed.

View larger version (4K): [in this window] [in a new window]   Figure 1. Relation of Patients' Depth-Time Profiles to U.S. Navy Limits for Dives Not Involving Decompression. Dives preceding the development of air embolism are shown (circles); the duration of the dive is plotted against the maximal depth of the dive. The U.S. Navy recommended limits for dives not involving decompression are denoted by the solid line, with the actual points denoted by the triangles. All patients' dives were well within the recommended time limits. To convert feet to meters, multiply by 0.3048.

 
The subjects in the diving control group ranged in age from 19 to 47 years (mean, 33). Four were women, and 18 were men. Their dive profiles were similar to those of the patients with arterial gas embolism. Blood samples were obtained from these subjects soon (2 ±1 hours, 12 samples) or late (11 ±1 hours, 12 samples) after the dive. The 11 patients with blunt trauma ranged in age from 16 to 71 years (mean, 38); 3 were women, and 8 were men. Blood samples for measurement of serum creatine kinase activity were obtained from the patients in the trauma control group every six hours from the time of admission until creatine kinase activity decreased or the results of isoenzyme studies were available; two to four measurements were made for each patient.

Outcome

Thirteen (59 percent) of the patients with arterial gas embolism recovered with no objective or subjective neurologic deficit at the time of discharge from the hospital. Of the remaining nine patients, four had residual deficits consisting of abnormalities of fine-motor function or impairments in memory, calculating skills, or word-finding skills; these patients were classified as minimally impaired. Three patients had residual hemiplegia of varying degrees or severe cognitive impairment and were classified as severely impaired; two patients died during their hospitalization. There was no correlation between outcome and the length of time before recompression therapy was instituted, the depth or duration of the dive, the initial mental status, or the mode of transport to the medical center.

Diagnostic Studies

All 22 patients with arterial gas embolism (95 percent confidence interval, 85 to 100 percent) had elevated serum creatine kinase values (>175 U per liter) at some time during hospitalization. The median peak serum creatine kinase value for the group as a whole was 1218 U per liter, but the values varied widely and were not distributed normally (range, 221 to 45,170; mean, 5956) (Figure 2). Fourteen of the 22 patients had markedly elevated peak values (>900 U per liter). The value was maximal 15 ±2 hours (median, 11) after the diving accident and fell rapidly thereafter (Figure 3). In all but 2 patients the values were highest within 24 hours after the diving accident, and in 18 the values were highest immediately before or immediately after recompression therapy. The mean serum creatine kinase value in the diving control group was 177 ±23 U per liter (Figure 2), and there was no significant difference between the samples obtained soon after a dive and those obtained 11 hours later. Similarly, the mean serum creatine kinase value in the trauma control group was 249 ±63 U per liter; none of these patients had a value exceeding 900 U per liter. The initial or peak serum creatine kinase values were above normal more often in the patients with arterial gas embolism than in the diving control group (P<0.001 by Fisher's exact test) or in the trauma control group (P<0.001).

View larger version (8K): [in this window] [in a new window]   Figure 2. Peak Serum Creatine Kinase Activity and Creatine Kinase MB Isoenzyme Activity in Patients with Arterial Gas Embolism after Diving (Circles) and Control Groups of Patients with Blunt Trauma (Triangles) and Subjects without Arterial Gas Embolism (Inverted Triangles). The patients with arterial gas embolism are grouped according to the degree of residual neurologic impairment present at the time of discharge from the hospital. The samples in the normal subjects were obtained an average of 2 hours (Early) and 11 hours (Late) after a dive. Serum samples in which the MB isoenzyme activity was 4 percent or more of total creatine kinase activity are denoted by the open circles.

 

View larger version (13K): [in this window] [in a new window]   Figure 3. Temporal Relation between Arterial Gas Embolism and Serum Creatine Kinase Activity. In Panel A, the mean serum creatine kinase values in the 12 patients in whom values were determined at least four times are expressed as fractions of the peak value and plotted against the time after the occurrence of the arterial gas embolism. The regression line (solid line) and 95 percent confidence intervals (broken lines) are shown. In Panel B, the interval between the arterial gas embolism and peak creatine kinase value was calculated for all 22 patients, and the cumulative number of patients whose peak value occurred in each six-hour interval is shown (solid line). The normalized cumulative fraction based on a gaussian distribution is indicated by the dashed line. In both panels, the mean (±SD) times at which recompression therapy was started and stopped are shown.

 
Creatine kinase isoenzymes were measured in 20 patients with arterial gas embolism (Figure 2). The MB isoenzyme was detectable in 13 of these patients in amounts ranging from 1 to 18 percent of serum total creatine kinase activity; it was 4 percent of total activity in 6 patients. The mean MB isoenzyme activity for the group as a whole was 268 U per liter, but as with serum total creatine activity, the range of values was broad (0 to 1743 U per liter). Only one patient had detectable BB isoenzyme activity (serum creatine kinase activity, 259 U per liter; BB activity, 2 percent of the total); this patient recovered completely.

There was a significant correlation between the peak serum creatine kinase activity and neurologic outcome in the 22 patients. This relation was present when the patients with no neurologic deficit were compared with all others (P<0.001) and when patients with minimal or no impairment were compared with those who had severe neurologic impairment or who died (P<0.001). There were also significant correlations between the serum MB isoenzyme value and outcome (P<0.001) and between the fraction of MB isoenzyme (relative to total creatine kinase activity) and outcome (P<0.001) (Figure 2). However, both the MB fraction and absolute MB values were significantly correlated with serum creatine kinase values and therefore did not appear to be independent predictors of neurologic deficit after arterial gas embolization. There was no apparent relation between serum creatine kinase values and the initial neurologic status, the length of time from the accident to evaluation or recompression therapy, or any other clinical variable.

Arterial-blood gas determinations were performed in all patients with arterial gas embolism at the time of the initial evaluation. Five patients had severe hypoxemia (partial pressure of arterial oxygen, <50 mm Hg); all five had changes on chest roentgenograms consistent with the aspiration of seawater. Twelve patients had abnormalities of gas exchange as defined by a ratio of the partial pressure of arterial oxygen (in mm Hg) to the fraction of inspired oxygen of 400, and in three of these gas exchange was markedly abnormal (ratio, <200). There was a significant correlation between the degree of impairment of gas exchange and the presence of bilateral infiltrates on the chest roentgenogram (P = 0.002). However, there was no correlation between abnormalities of gas exchange and the final neurologic outcome, between hypoxemia and outcome, or between hypoxemia and serum creatine kinase activity.

Twenty-one patients underwent electrocardiography at least once during their hospitalization; no tracings that predated the diving accident were available for any patient. In 15 patients, electrocardiograms obtained immediately after recompression therapy were described as abnormal by the treating physicians. A review of the tracings showed that 12 of these patients had nonspecific changes not necessarily indicative of myocardial injury. The remaining three patients, including the only two patients who also had electrocardiography before recompression, had abnormalities consistent with myocardial injury. In these three patients, the changes (abnormal Q waves in one and ST-segment elevations of 1 mm in the anterior chest leads in two) subsequently resolved during their hospitalization.

Attempts were made to confirm the presence of myocardial injury in two patients. A 16-year-old boy with a peak serum creatine kinase value of 2520 U per liter (MB isoenzyme, 7 percent of the total) and T-wave inversions of the inferior wall on electrocardiography was noted to have inferior-wall hypokinesis on M-mode and two-dimensional echocardiography (no subsequent echocardiograms were obtained). A technetium pyrophosphate scan obtained 48 hours after the diving accident showed no focal accumulation of the radiopharmaceutical agent. A 29-year-old man with a peak serum creatine kinase value of 12,065 U per liter (MB isoenzyme, 8 percent of the total) and ST-segment elevations of 1 mm in the anterior leads had no abnormalities on gated radionuclide scans or M-mode and two-dimensional echocardiograms.

No patient had any abnormalities of renal function. Myoglobin was not found in the urine of the one patient in whom it was measured; this patient's peak serum creatine value was 306 U per liter.

Discussion

All the patients with arterial gas embolism whom we studied had an elevated serum creatine kinase value during their hospitalization. Among these patients, 59 percent had complete neurologic recovery, and 23 percent had severe impairment or died. Since these percentages are similar to those in other series of patients with arterial gas embolism,^5,6 we believe the population we studied is representative. Although the clinical manifestations of arterial gas embolism have been well described, no consistent association between abnormalities of serum creatine kinase activity and arterial gas embolism has been noted. Two reports did describe single patients with elevated values in association with scuba-diving accidents^12,13.

Many tissues contain small amounts of creatine kinase, but large quantities are found only in skeletal muscle, cardiac muscle, and brain^14,15. The BB isoenzyme is the sole form of creatine kinase in the brain, and its detection in serum is a relatively specific indicator of central nervous system injury^16,17,18. The detection of BB isoenzyme in only one patient suggests that the brain is not a major source of the elevated serum creatine kinase activity after arterial gas embolism. The magnitude of the elevations in serum creatine kinase activity, the predominance of the MM isoenzyme, and the presence of the MB isoenzyme support the conclusion that the source was skeletal or myocardial muscle.

The finding of MB isoenzyme values that were at least 4 percent of total creatine kinase activity in 6 of the 22 patients with arterial gas embolism suggests a myocardial source for some of the creatine kinase. However, the MB isoenzyme is present in small amounts in nonmyocardial muscle as well,^14,19,20 and the specificity of elevations in the MB isoenzyme for myocardial injury may be questionable when there is also diffuse injury of skeletal muscle. Indeed, although the MB isoenzyme is present in high concentrations (15 to 22 percent) in diseased heart muscle, it may be present only in low concentrations (1 percent) in myocardium from subjects (such as our patients) without preexisting cardiac disease²¹. This suggests that the MB isoenzyme we detected may originate from a nonmyocardial source. Similarly, although electrocardiographic abnormalities suggesting myocardial injury were found in some patients, cerebral gas embolism is known to lead to reflex changes in the electrocardiogram that can mimic ischemia²². Hence, the extent of myocardial damage, if any, in our patients is uncertain. It is important to note there was no direct evidence of myocardial necrosis, and indeed, in the three patients with electrocardiographic abnormalities suggestive of myocardial injury, these changes completely resolved by the time the patients left the hospital.

The mechanisms leading to the release of creatine kinase after arterial gas embolism are unclear. It is possible that focal muscle injury during the process of resuscitation or removal from the water led to the elevation of creatine kinase activity. Evidence of substantial muscle injury was not found on physical examination, nor was there information in the patients' histories that would have suggested direct trauma as an explanation for the elevated values. Elevations in serum creatine kinase of similar magnitudes were not found in patients in the trauma control group, who were transported to our institution by the same emergency system. In addition, the absence of any relation between serum creatine kinase activity and initial neurologic impairment or the duration or mode of transport to the hospital further suggests that trauma during or after the accident did not cause the elevation in serum creatine kinase after arterial gas embolism. The serum creatine kinase values in the diving control group were also not elevated to the same degree as in the patients with arterial gas embolism. Arterial-blood gas determinations did not reveal hypoxemia or acidosis consistently, and when hypoxemia was observed, it did not correlate with the elevation in serum creatine kinase values. Thus, traumatic injury, resuscitative efforts, diving itself, hypoxemia, and the mode of transport do not appear to account for the elevations in serum creatine kinase. We are, therefore, led to the conclusion that gas emboli entering the systemic circulation after barotraumatic pulmonary injury result in the release of creatine kinase from skeletal muscles. Centrally released gas bubbles are known to disseminate throughout the systemic circulation in normotensive animals²³. A similar pattern of dissemination after diving-associated arterial gas embolism would cause bubbles to reach skeletal muscle. We can only speculate whether the release of creatine kinase is caused by direct injury to the muscle cells during transit or occlusion of the adjacent microcirculation by gas bubbles, by reperfusion injury of muscle cells, or by some secondary effect of the bubbles. Any process that allows leakage of creatine kinase out of muscle cells without causing cell necrosis could explain our results.

We conclude that arterial gas embolism associated with a diving accident is a systemic process and that the eventual neurologic outcome of patients who have arterial gas embolization correlates well with the magnitude of the increase in serum creatine kinase. Marked elevation of serum creatine kinase may be an indicator of arterial gas embolism.

We are indebted to Mr. Al Jones and his students for their kind assistance.

Source Information

From the Divisions of Pulmonary and Critical Care Medicine (R.M.S.) and Emergency Medicine (T.S.N.), Department of Medicine, University of California, San Diego School of Medicine, San Diego.

Address reprint requests to Dr. Smith at the Division of Pulmonary and Critical Care Medicine, 111-J, VA Medical Center, 3350 La Jolla Village Dr., San Diego, CA 92161.

References

1. Neuman TS. Diving medicine. Clin Sports Med 1987;6:647-661. [Medline]
2. Ghosh PK, Kaplan O, Barak J, Lubliner J, Vidne BA. Massive arterial air embolism during cardiopulmonary bypass. J Cardiovasc Surg (Torino) 1985;26:248-250. [Medline]
3. Horrow JC, Laucks SO. Coronary air embolism during venous cannulation. Anesthesiology 1982;56:212-214. [Medline]
4. King MW, Aitchison JM, Nel JP. Fatal air embolism following penetrating lung trauma: an autopsy study. J Trauma 1984;24:753-755. [Medline]
5. Dick AP, Massey EW. Neurologic presentation of decompression sickness and air embolism in sport divers. Neurology 1985;35:667-671. [Free Full Text]
6. Leitch DR, Green RD. Pulmonary barotrauma in divers and the treatment of cerebral arterial gas embolism. Aviat Space Environ Med 1986;57:931-938. [Medline]
7. Whitcraft DD III, Karas S. Air embolism and decompression sickness in scuba divers. JACEP 1976;5:355-361. [Medline]
8. Trainer TD, Gruenig D. A rapid method for the analysis of creatine phosphokinase isoenzymes. Clin Chim Acta 1968;21:151-154. [CrossRef][Medline]
9. Glantz SA. Primer of biostatistics. 2nd ed. New York: McGraw-Hill, 1987.
10. Hosmer DW Jr, Lemeshow S. Applied logistic regression. New York: John Wiley, 1989.
11. Berry CC. A tutorial on confidence intervals for proportions in diagnostic radiology. AJR Am J Roentgenol 1990;154:477-480. [Free Full Text]
12. Williams J, Collins DM, Heenan A. Creatine kinase increased after a scuba diving incident. Clin Chem 1988;34:1514-1515. [Free Full Text]
13. Feinstein EI, Akmal M, Telfer N, Massry SG. Delayed hypercalcemia with acute renal failure associated with nontraumatic rhabdomyolysis. Arch Intern Med 1981;141:753-755. [Free Full Text]
14. Dawson DM, Fine IH. Creatine kinase in human tissues. Arch Neurol 1967;16:175-180. [Free Full Text]
15. Trask RV, Billadello JJ. Tissue-specific distribution and developmental regulation of M and B creatine kinase mRNAs. Biochim Biophys Acta 1990;1049:182-188. [Medline]
16. Hamburg RJ, Friedman DL, Olson EN, et al. Muscle creatine kinase isoenzyme expression in adult human brain. J Biol Chem 1990;265:6403-6409. [Free Full Text]
17. Longstreth WT Jr, Clayson KJ, Sumi SM. Cerebrospinal fluid and serum creatine kinase BB activity after out-of-hospital cardiac arrest. Neurology 1981;31:455-458. [Free Full Text]
18. Kaste M, Somer H, Konttinen A. Brain-type creatine kinase isoenzyme: occurrence in serum in acute cerebral disorders. Arch Neurol 1977;34:142-144. [Free Full Text]
19. Bendz R, Strom S. Diagnostic significance of serum CK-MB elevations following surgical damage to skeletal muscles. Scand J Thorac Cardiovasc Surg 1981;15:199-204. [Medline]
20. Roberts R, Gowda KS, Ludbrook PA, Sobel BE. Specificity of elevated serum MB creatine phosphokinase activity in the diagnosis of acute myocardial infarction. Am J Cardiol 1975;36:433-437. [CrossRef][Medline]
21. Ingwall JS, Kramer MF, Fifer MA, et al. The creatine kinase system in normal and diseased human myocardium. N Engl J Med 1985;313:1050-1054. [Abstract]
22. Evans DE, Kobrine AI, Weathersby PK, Bradley ME. Cardiovascular effects of cerebral air embolism. Stroke 1981;12:338-344.
23. Butler BD, Laine GA, Leiman BC, et al. Effect of the Trendelenburg position on the distribution of arterial air emboli in dogs. Ann Thorac Surg 1988;45:198-202. [Abstract]

Abstract Add to Personal Archive Add to Citation Manager Notify a Friend E-mail When Cited PubMed Citation

This article has been cited by other articles:

• Salam, A. M., Albinali, H. A., Gehani, A. A., Suwaidi, J. A. (2003). Acute Myocardial Infarction in a Professional Diver After Jellyfish Sting. Mayo Clin Proc. 78: 1557-1560 [Abstract]  
• Neuman, T. S. (2002). Arterial Gas Embolism and Decompression Sickness. Physiology 17: 77-81 [Abstract] [Full Text]  
• Bove, A. A. (2002). Medical Disorders Related to Diving. J Intensive Care Med 17: 75-86 [Abstract]