crush syndrome sustained in the 1995 kobe, japan, earthquake; treatment and outcome

6
DISASTER MEDICINE/ORIGINAL CONTRIBUTION Crush Syndrome Sustained in the 1995 Kobe, Japan, Earthquake; Treatment and Outcome From the Department oj Anesthesiology and Intensive Care Medicine,* the Division of Critical Care Medicine,~ and the Intensive Care Unit, ~ Osaka City University Medical School, Osaka,Japan. Receivedjor publication September 28, 1995. Revisions received May 6 and December 3,/996; and April 4, 1997. Accepted for publication April 22, 1997. Copyright © by the Amer*can College of Emergency Physicians. Yutaka Oda, MD* Mitsuo Shindoh, MD* Hidekazu Yukioka, MD* Shinichi Nishi, MD § Mitsugu Fujimori, MD* Akira Asada, MD* Study objective: To assess the treatment and outcome of patients with crush syndrome sustained in an earthquake disaster. Methods: We conducteda retrospectiveanalysisof eight patients with crush syndrome and subsequent acute kidney failure who were treated in the ICU of a university hospital. All eight patients had been extricatedfrom buildings that collapsed in the 1995 Kobe, Japan, earthquake. Crush injury involved the upper extremities in one patient and the lower extremities in seven. Each patient re- ceived intravenousfluid infusion and diuretic drugs and underwent hemodialysis. Emergencyfasciotomy was performed in some patients, 17 to 100 hours after extrication. Results: All patients were consciousand lucid on admission, and blood pressure and heart rate were normal. All the patients dem- onstrated kidney failure with increased concentrations of serum creatinine (1.9 to 9.6 mg/dL [169 to 852 ~tmol/L]). Six patients were oliguric. Hyperkalemia (5.6 to 8.8 mEq/L) was present in six patients. We found close correlations between the serum potas- sium and creatine kinaseconcentrations,between the serum myo- globin and potassium concentrations, and between the serum myoglobin and creatine kinase concentrations.All the patients were weaned from hemodialysis.The serum creatinine concentra- tion decreasedto a normal level within 20 to 52 days of admis- sion in all patients. No patients underwent amputation. Muscle weakness and sensory deficits persisted in all patients 6 months after the earthquake. Conclusion: Our findings support current therapeutic strategies for crush syndrome,despite the long delay to initiation of intensive therapy. All the patients recovered kidney function and were weaned from hemodialysis; none required amputation. [0da Y, Shindoh M, Yukioka H, Nishi S, Fujimori M, Asada A: Crush syndrome sustained in the 1995 Kobe, Japan, earthquake: Treatment and outcome. Ann Emerg Med October 1997;30:507- 512.] OCTOBER 1997 30:4 ANNALS OF EMERGENCY MEDICINE 5 0 7

Upload: yutaka-oda

Post on 02-Jul-2016

221 views

Category:

Documents


4 download

TRANSCRIPT

Page 1: Crush Syndrome Sustained in the 1995 Kobe, Japan, Earthquake; Treatment and Outcome

DISASTER MEDICINE/ORIGINAL CONTRIBUTION

Crush Syndrome Sustained in the 1995 Kobe,

Japan, Earthquake; Treatment and Outcome

From the Department oj Anesthesiology and Intensive Care Medicine,* the Division of Critical Care Medicine, ~ and the Intensive Care Unit, ~ Osaka City University Medical School, Osaka, Japan.

Received jor publication September 28, 1995. Revisions received May 6 and December 3,/996; and April 4, 1997. Accepted for publication April 22, 1997.

Copyright © by the Amer*can College of Emergency Physicians.

Yutaka Oda, MD* Mitsuo Shindoh, MD* Hidekazu Yukioka, MD* Shinichi Nishi, MD § Mitsugu Fujimori, MD* Akira Asada, MD*

Study objective: To assess the treatment and outcome of patients with crush syndrome sustained in an earthquake disaster.

Methods: We conducted a retrospective analysis of eight patients with crush syndrome and subsequent acute kidney failure who were treated in the ICU of a university hospital. All eight patients had been extricated from buildings that collapsed in the 1995 Kobe, Japan, earthquake. Crush injury involved the upper extremities in one patient and the lower extremities in seven. Each patient re- ceived intravenous fluid infusion and diuretic drugs and underwent hemodialysis. Emergency fasciotomy was performed in some patients, 17 to 100 hours after extrication.

Results: All patients were conscious and lucid on admission, and blood pressure and heart rate were normal. All the patients dem- onstrated kidney failure with increased concentrations of serum creatinine (1.9 to 9.6 mg/dL [169 to 852 ~tmol/L]). Six patients were oliguric. Hyperkalemia (5.6 to 8.8 mEq/L) was present in six patients. We found close correlations between the serum potas- sium and creatine kinase concentrations, between the serum myo- globin and potassium concentrations, and between the serum myoglobin and creatine kinase concentrations. All the patients were weaned from hemodialysis. The serum creatinine concentra- tion decreased to a normal level within 20 to 52 days of admis- sion in all patients. No patients underwent amputation. Muscle weakness and sensory deficits persisted in all patients 6 months after the earthquake.

Conclusion: Our findings support current therapeutic strategies for crush syndrome, despite the long delay to initiation of intensive therapy. All the patients recovered kidney function and were weaned from hemodialysis; none required amputation.

[0da Y, Shindoh M, Yukioka H, Nishi S, Fujimori M, Asada A: Crush syndrome sustained in the 1995 Kobe, Japan, earthquake: Treatment and outcome. Ann Emerg Med October 1997;30:507- 512.]

OCTOBER 1997 30:4 ANNALS OF EMERGENCY MEDICINE 5 0 7

Page 2: Crush Syndrome Sustained in the 1995 Kobe, Japan, Earthquake; Treatment and Outcome

CRUSH S Y N D R O M E Oda et al

INTRODUCTION

Crush syndrome is the systemic manifestation of muscle injury caused by prolonged limb compression sustained in crush injury. It is characterized by hypovolemic shock, hyperkalemia, acute kidney failure, and muscle necrosis.*,2 In patients with crush syndrome, kidney failure develops shortly after muscle compression. 3 Mortality for crush syn- drome associated with acute kidney failure remains high, despite major advances in critical care. 3-5 Crush syndrome is typically encountered in war zones, in mining disasters, after earthquakes, and in industrial and traffic accidents. 6

Difficulties with communication and transportation in the wake of disasters often prevent early extrication and thera- peutic interventions. Since the first report of this syndrome by Bywaters, 1 many authors have described the management of crush syndrome. 6,r Early extrication and administration of intravenous fluid if possible, before or immediately after the release of compressed limbs--are crucially impor- tant in preventing kidney failure. 6,r Programs have been suggested to improve the survival rate of victims. 8

An earthquake on January 17, 1995, in Kobe, Japan, killed more than 5,000 people. Survivors extricated from the rubble who had sustained crush syndrome and kidney failure were ultimately transported to hospitals in Osaka. In this report we describe the profiles, treatment, and out- come of eight patients transported to our hospital after the Kobe earthquake.

MATERIALS AND METHODS

Eight patients who had been buried under wooden houses that collapsed in the earthquake were extricated between 1 and 18 hours after the earthquake and admitted to hospi-

tals m Kobe. The patients' limbs had been under continuous compression. Crush injury was diagnosed on the basis of the presence of swollen limbs and history of limb compres- sion. All eight patients were given intravenous fluid after arriving at the local hospitals, although exact fluid volumes were not noted. All eight patients, with suspected kidney failure, were transferred to our ICU by ambulance or boat between 17 and 100 hours after extrication for the treatment of suspected kidney failure because hemodialysis was not available at the local hospitals. Acute kidney failure was defined as a serum creatinine concentration of more than .5 mg/dL over baseline (>1.8 mg/dL).

On each patient's arrival, arterial blood-gas analysis, blood tests, skeletal radiography, and neurologic examina- tion were performed. A urinary catheter was inserted in each of the three patients who had not been fitted with a catheter in Kobe, and hourly urine output was measured for each of the eight patients. Aggressive intravenous vol- ume-loading with normal saline solution and albumin and continuous infusion of dopamine at 3 gg/kg/minute were begun, along with monitoring of arterial blood pressure, central venous pressure, heart rate, and arterial oxygen sat- uration of hemoglobin. The rate of dopamine infusion was increased up to 5 btg/kg/minute to increase urine output. Furosemide (1 rag/kg) was repeatedly administered intra- venously to all patients. Glucose and insulin were admin- istered and emergency hemodialysis performed in patients with serum potassium concentrations greater than 6 mEq/L. All patients underwent hemodialysis three or four times a week until urine output exceeded 400 mlJday and serum potassium and creatinine normalized without further hemodialysis.

Table 1. Characteristics of patients with crush syndrome.

Patient No. Age (Years) Sex Affected Limb

Time Between Duration of Burial Extrication and

(Hours) Treatment (Hours) Fasciotomy Complications

1 49 F One leg 2 69 M One leg 3 24 M Upper arm 4 41 M One leg 5 26 M Both legs 6 21 F Both legs 7 71 F Beth legs 8 72 F Both legs

*NR, not recorded. This patient was extricated within 18

8 17 Y - - I 62 N - - 5 73 Y - - 6 75 N - -

15 67 N Lumbar vertebral fracture 4 78 Y - - NR* NR N Pelvic bone fracture 6 100 Y Pelvic bone fracture,

rib fracture, hemothorax

hours, although the exact amount of time to extrication is unknown.

50 8 ANNALS OF EMERGENCY MED)CINE 30:4 OCTOBER 1997

Page 3: Crush Syndrome Sustained in the 1995 Kobe, Japan, Earthquake; Treatment and Outcome

A

The relationships between serum potassium and crea- tine kinase or creatinine concentration, between serum myo- globin and potassium or creatine kinase concentration, between the duration of burial or the time elapsed before initiation of treatment at our hospital and serum potassium, creatine kinase, or creatinine concentration were examined by analysis of linear regression with the use of the least- squares method and Spearman's rank correlation coefficient with the use of StatView J-4.5 software. P values less than .05 were considered significant.

The study was approved by the institutional review board of the Osaka City University Medical School.

RESULTS

The patients' mean age was 47 years (range, 21 to 72 years) (Table 1).

On admission, all the patients were alert, with Glasgow Coma Scale scores of 14 or 15 and normal vital signs. Each patient complained of the absence of sensory and motor activity m the affected extremity (Table 1). No fractures were detected in the affected limbs, but fractures were detected in the pelvis or lumbar vertebrae in three patients. Arterial blood pH ranged from 7.34 to 7.54. Pao; and Paco 2 were normal, except in a 72-year-old woman who presented with hemothorax with rib fracture. She was orotracheally intubated, underwent pleural drainage, and was mechani- cally ventilated for 3 days. Myoglobin was detected in the urine of all patients with the use of an orthotoluidine dip- stick test.

All patients had acute kidney failure at the time of admis- sion. In two patients, nonoliguric kidney failure developed, with urine output greater than 400 mI_/day and serum crea- tinine concentration greater than 4 mg/dL. Hyperkalemia with serum potassium concentrations ranging from 6.4 to 8.8 m E @ (normal range, 3.5 to 4.7 mEq/L) and elevated T waves on ECG were present in four patients, each of whom underwent emergency hemodialysis (Table 2). De- creased serum sodium concentrations ranging from 119 to 133 m E @ (normal range, 135 to 150 mEq/L) were present in six patients.

Urine output was increased in the two nonoliguric patients by the intravenous administrations of fluid, diuretic drugs, and dopamine, but serum creatinine concentrations remained increased. Urine output did not increase in other six oliguric patients, despite these measures.

All eight patients required hemodialysis three or four times a week. Hemodialysis was performed 3 to 20 times in the eight patients (mean, 9 times). Serum sodium and potassium were corrected to normal concentrations within

Figure 1. A, Correlation of serum potassium and creatine kinase concentrations. The linear regression correlation coefficient was .87 (P=.0047). B, Correlation of serum potassium and creatinine concentrations. The linear regression correlation coefficient was .57 (P=.I ~.). Potassium, creatine kinase, and creatinine concentrations were measured on admission.

I

Serum Creatine Kinase (105 U/L)

r=.87

4 5 6 7 8 9

Serum Potassium (mEq/L)

B Serum Creatinine {mg/dL)

r=.57

18

CRUSH S Y N D R O M E Oda et al

/ ,

Serum Potassium (mEq/L)

OCTOBER 1997 30:4 ANNALS OF EMERGENCY MEDICINE 5 O 9

Page 4: Crush Syndrome Sustained in the 1995 Kobe, Japan, Earthquake; Treatment and Outcome

C R U S H S Y N D R O M E

Oda et al

Figure 2. A, Correlation of serum myoglobin and potassium concentrations. The correlation coefficient (r ) obtained with the use of Spearman's rank correlation test was .81 (P=.030). B, Correlation of serum myoglobin and creatine kinase concentrations. The correlation coefficient (r) was .93 (P=.O14). Myoglobin was measured on the first, second, or third hospital day. Serum potassium and creatine kinase concentrations were measured on admission.

A ;erum Potassium (mEq/L)

1

y, ,

I I [ I

B

1 2 3

Serum Myoglobin (104gg/L)

Serum Creatine Kinase (10 ~ U/L)

2 -

1 -

I t I I 1 2 3 4

Serum Myoglobin (104p.g/L)

3 days of admission. The serum creatinine concentration was gradually reduced to less than 1.7 mg/dL within 20 to 52 days of admission.

We noted a strong correlation between serum potas- sium and creatine kinase concentrations (r=.87, P=.0047) (Figure 1A) but no correlation between serum potassium and creatinine concentrations (r=.57, P=. 14) (Figure 1B). No correlation was found between the duration of burial and time elapsed between extrication or initiation of treat- ment and serum potassium, creatine kinase, or creatinine concentration. A single serum myoglobin level for each patient measured on the first, second or third hospital day ranged from 400 to 33,000 btg/L (normal range 6 to 85) (Table 2). Significant correlations were found between this concentration and the admission potassium concentration (rs=.81, P=.030) (Figure 2A) and admission creatine kinase concentration (r =.93, P=.014) (Figure 2B).

Peripheral arterial pulse, measured with laser Doppler flowmetry (915-AL; Parks Medical Electronics, Incorpor- ated), was absent in four patients. Although intramuscular pressure was not measured in the patients, three patients underwent four compartment fasciotomies on their legs. One patient underwent fasciotomies on his upper arm while under general anesthesia (Table 1). We did not perform de- bridement because muscle necrosis was not observed by visual inspection during operation. Peripheral arterial pulse was regained after fasciotomy. We detected no significant changes in hemodynamic parameters, arterial blood gases or electrolyte concentrations, and serum creatine kinase levels decreased after fasciotomy.

No patients had infection or required amputation. Their fasciotomies were surgically closed 6 to 8 weeks after the initial procedures. Muscle weakness and sensory deficits persisted in all patients 6 months after injury. Of the patients who sustained lower-extremity injuries, at this writing three require braces to walk, three walk with canes, and the re- maining patient cannot walk. In the patient who sustained upper-extremity injury, Volkmann's contracture developed.

DISCUSSION

In this study of patients who sustained crush syndrome with kidney failure, reduced serum sodium and increased potassium concentrations were noted. Because seven of our patients did not demonstrate acidemia, their hyperkalemia was probably the result of an efflux of potassium from dam- aged muscle cells, not release of potassium from cells in ex- change for hydrogen ion. Hyperkalemia has been noted in almost every published report on crush syndrome. Four of

5 1 0 ANNALS OF EMERGENCY MEDICINE 30:4 OCTOBER 1997

Page 5: Crush Syndrome Sustained in the 1995 Kobe, Japan, Earthquake; Treatment and Outcome

CRUSH SYNDROME Oda et al

our eight patients required emergency hemodialysis for treatment of hyperkalemia.

Few authors have described decrease in serum sodium in patients with crush syndrome. 9,~° In those reports, re- duced serum sodium has been combined with increased serum potassium, just as we noted in our patients. These changes may have resulted from prior volume loading in Kobe with hypotonic fluids, diminished efflux of sodium from the sarcoplasm into serum, or both; Na+/K+-ATPase activity is inhibited by ischemia. 11

We noted a close correlation between the concentrations of serum potassium and creatine kinase, both of which are released from injured muscle ceils. In contrast, a very weak correlation was found between serum potassium and crea- tinine concentrations. This correlation may have resulted from the wide interindividual variation in serum creatinine level before injury and variation in creatinine clearance.

A close correlation was found between the concentrations of myoglobin and serum potassium and creatine kinase. Previously published reports on crush syndrome do not in- clude myoglobin data, and the relationship between myo- globin concentration and other laboratory findings remains unclear. On the other hand, no correlation was found be- tween duration of muscle compression or time elapsed be- tween extrication and treatment and parameters indicating the degree of muscle injury such as the serum concentrations of potassium, creatine kinase, and myoglobin. The reason for this lack of correlation is unclear, but it does suggest that muscle may be extensively damaged even if the victim is extricated soon after being buried in rubble. This finding emphasizes the importance of emergency blood testing, in- cluding measurement of serum electrolytes. Because early hemodialysis often cannot be performed after a disaster, prevention of acute kidney failure has been a major focus of investigation for many years. 12 Intravenous fluid infu- sion, particularly rapid infusion of isotonic saline solution,

has been recommended as a prophylactic treatment against the development of acute kidney failure. 3,5,r,13 Ron et al re- ported that kidney failure was successfully prevented with the initiation of aggressive fluid infusion within 10 hours of release of muscle compression, r Kidney failure with in- creased serum creatinine concentration had developed in all our patients by the time of admission to our hospital (range, 17 to 100 hours after extrication). We began aggres- sive intravenous fluid infusion with monitoring of arterial blood pressure, central venous pressure, heart rate, and blood gases. In no patient did heart Nilure or lung edema develop. Despite therapeutic measures, urine output in- creased only in the two nonoliguric patients, and all patients required hemodialysis in our study.

We performed fasciotomy in patients without peripheral arterial pulse as measured with laser Doppler flowmetry. Some authors suggest the use of fasciotomy to prevent the muscle necrosis caused by increased intramuscular pres- sure, 5 whereas others disagree because fasciotomy facilitates wound infection, thereby endangering the patient's life. 13 Muscle volume increases after fluid infusion, * and swollen muscle increases intramuscular pressure and decreases blood flow, resulting in further necrosis of affected muscle. Fasci- otomy was performed successfully, without the development of infection, in our patients.

Objective criteria predicting amputation after lower- extremity trauma are suggested byJohansen et al. 14 Accord- ing to those criteria, crush injury and limb ischemia are primary contributors to the need for limb amputation. How- ever, in this study limb amputation was not performed; fasciotomy may have prevented further circulatory distur- bances, and serum potassium levels were controlled by the course of our therapy.

Reperfusion injury after ischemia is considered an impor- tant mechanism of crush syndrome. 15 Microvascular damage caused by ischemia and oxygen free radicals, produced not

Table 2. Biochemical findings on admission.

Patient No. Sodium (mEq/L) Potassium (mEq/L) Creatinine (mg/dL) Creatine Kinase (U /L ) Myoglobin (#g/L}

1 133 5.6 1.9 213,000 5,200 2 140 5.7 6.8 12,139 400 3 119 7.1 9.6 199,200 7,100 4 120 7.4 9.0 205,900 7,500 5 123 8.8 8.0 334,000 33,000 6 126 6.4 4.8 168,400 4,100 7 128 4.2 5.8 t4,498 750 8 135 4.8 5.5 26,311 850

*Myoglobin data obtained on the first, second, or third hospital day.

OCTOBER 1997 30:4 ANNALS OF EMERGENCY MEDICJNE 5 1 1

Page 6: Crush Syndrome Sustained in the 1995 Kobe, Japan, Earthquake; Treatment and Outcome

C R U S H S Y N D R O M E Oda et aI

only by ischemic kidneys but by crushed muscle, contrib- utes to kidney failure. 15,~6 Reperfusion following ischemia is common after tourniquet release m daily clinical practice, and significant hemodynamic and metabolic changes have been reported. 17,18 In our study, blood pressure, heart rate, and blood gases were normal when muscle compression was released, and reperfusion occurred after fasdotomy

The mortality from crush syndrome sustained in earth- quakes ranges from 13% to 25% when kidney failure devel- ops. ~9,2o Although all our patients had kidney failure, with serum creatinine concentrations greater than 1.8 mg/dL, on arrival at our hospital, they survived and regained kidney function. Although other authors have reported that lddney failure resolves within 10 to 20 days in patients with crush syndrome, 2,9 in almost all of our patients, decreases in serum creatinine concentrations to the normal range required at least 20 days. No skin lacerations, fractures, or muscle necro- sis (which are associated with crush syndrome) were de- tected in the affected limbs of our patients, despite the severe muscle damage and remarkable increase of serum myo- globin and creatine kinase concentrations. This may have been due to the fact that our patients had been buried under demolished wooden houses instead of metal and glass build- ings, with less severe injuries than those sustained in a typ- ical case of crush syndrome. Rhabdomyolysis in patients buried under single-story structures, compared with that in patients buried under muldstory stone buildings, is also reported by Noji. 21 None of our patients required amputa- tion, but neurologic disorders, including muscle weakness and reduced sensitivity, persist in all patients, and protracted rehabilitation is required.

There is a wide disparity in the definitions of acute kidney failure; one commonly used definition is based on the in- crease of serum creatinine concentration over the baseline value. 22 In this study, acute kidney failure was defined as an increase in serum creatinine concentration above the upper limit of the normal level (1.8 mg/dL). 5 Because our patients' kidney status and serum creatinine concentrations before the earthquake were not known and patients were dehydrated, defining acute kidney failure as a creatinine concentration greater than 1.8 mg/dL might be misleading. Other limitations include inexact knowledge of the dura- tion of administration and volume of fluid administered before our patients' arrival at our institution. The use of a single myoglobin value not timed to a potassium value is certainly problematic.

In summary, our findings support the use of current thera- peutic strategies in the treatment of patients with crush syn- drome and acute kidney failure, despite the long delay from the occurrence of crush injury to treatment. All patients re-

gained kidney function and were weaned from hemodialy- sis, and none required amputation.

R E F E R E N C E S 1. Bywaters EGL Beall D: Crush injuries with impairment of renal function. BMJ1941;1:427-432.

2. Kikta M J, Meyer JP, Bishara RA, et ah Crush syndrome due to limb compression. Arch Surg 1987;122:1078-1081.

3. Reis ND, Michaelson M: Crush injury to the lower limbs. JBoneJointSurfl[Am]1986;68A: 414-418.

4. Gabew PA, Kaehny WD, Kelleher SP: The spectrum of rhabdomyolysis. Medicine 1982;61:141-152.

5. Shaw AD, SjNin SU, McQueen MM: Crush syndrome following unconsciousness: Need for urgent orthopaedic referral. BMJ 1994;309:857-869.

6. Bywaters EGL: 50 years on: The crush syndrome. BMJ 1990;301:1412q 416.

7. Ran D, Taitelman U, Michaelson M, et ah Prevention of acute renal failure in traumatic rheb- domyelysis. Arch Intern Mad 1984;144:277-280.

8. Koenig KL, Schultz CH, DiLorenzo RA: The crush injury cadaver lab: A new method of training physicians to perform fasciotomies and amputations on survivors of a catastrophic earthquake. Ann Emerg Med 1992;21:196.

9. Honda N, Kurokawa K: Acute renal failure and rhabdomynlysis. Kidneylnt1983;23:888-898.

10. Hawkins B J, Bays PN: Catastrophic complication of simple cast treatment: Case report. J Trauma 1993;34:760-762.

11. Bersehn MM, Philipson KD, Fukushima JY: Sodium-calcium exchange and sarcolemmal en- zymes in ischemic rabbit hearts. Am J Physio11982;242:C288-0295.

12. Solez K, 8ihari D, Collins A J, et el: International dialysis aid in earthquakes and other disas- ters. Kidney Int 1993;44:479-483.

13. Better OS, Stein JH: Early management of shock and prophylaxis of acute renal failure in traumatic rhabdomyolysis. N Enflf J Mad 1990;322:825-829.

14. Johansen K, Daines M, Howey T, et al: Objective criteria accurately predict amputation fol- lowing lower extremity trauma. J Trauma 1990;30:668-573.

15. Sexton WL, Korthuis R J, Laughlin MH: Ischemia-reperfusion injury in isolated rat hindquar- ters. J Appl Physiot 1990;68:387-392.

16. Odeh M: The role of reperfusion-induced injury in the pathogenesis of the crush syndrome. N Endl J Mad 1991 ;324:1417-1422.

17. Lynn AM, Fischer T, Brandford HG, et ah Systemic responses to tourniquet release in chil- dren. Anesth Analg 1986;65:865-872.

18. Klenerman L, Biswas M, Hulands GH, et al: Systemic end local effects of the application of a tourniquet. J Bone Joint Surg [Br] 1980;62B:385-388.

19. Richards NT, Tettersall J, Mcgann M. et al: Dialysis for acute renal failure due to crush injuries after the Armenian earthquake. BMJ 1989;298:443-445.

20. Santangelo ML, Usberti M, Di Salve E, et ah A study ef the pathology of the crush syndrome. Surg Gynecol Obstet 1982;154:372-374.

21. Noji EK: Prophylaxis of acute renal failure in traumatic rhabdomyolysis. N Engl J Med 1990;323:550-551.

22. Thadhani R, Pascual M, Bonventre JV: Acute renal failure. NEnglJMed 1996;334:1448-1460.

We are grateful to Dr M Shinozaki, Wakayama Medical College, for providing critical commentary on the manuscript.

Reprint no. 47/1/84404 Address for reprints:

Yutaka Oda, MD

Department of Anesthesiology and Intensive Care Medicine

Osaka City University Medical School

1-5-7, Asahimachi, Abeno-ku

Osaka 545, Japan

5 12. ANNALS OF EMERGENCY MEDICINE 30:4 OCTOBER 1997