BRIEF COMMUNICATIONS J Am Soc Nephrol 10: 382–386, 1999

Acute Renal Impairment after Immersion and Near-Drowning

S. TIMOTHY SPICER,* DAVID QUINN,* NYEIN N. NYI NYI,* BRIAN J. NANKIVELL,† JAMES M. HAYES,* and ELLIOTT SAVDIE* Departments of Nephrology, *St. Vincent’s Hospital and †Westmead Hospital, Sydney, Australia.

Abstract. Acute renal impairment (ARI) secondary to immer- more common in patients with ARI (P ϭ 0.016), and patients sion and near-drowning is rarely described and poorly under- with2to3ϩ of admission dipstick had a higher stood. A retrospective case-control study was performed: (1)to peak serum than patients with less proteinuria (P Ͻ determine the incidence of ARI associated with near-drowning 0.05). Admission predictors of ARI by univariate logistic re- or immersion and (2) to define the clinical syndrome and to gression analysis included reduced serum bicarbonate (P ϭ assess clinical predictors of ARI. Of 30 patients presenting 0.002), pH (P ϭ 0.001), and (P Ͻ 0.001). The best after immersion or near-drowning, 50% were identified with predictor of ARI on multivariate analysis was a negative base ARI, with a mean admission serum creatinine of 0.24 Ϯ 0.33 excess (P ϭ 0.01). In summary, acute renal impairment com- mmol/L (2.7 Ϯ 3.7 mg/dl). These patients were a heteroge- monly occurs after immersion and near-drowning and is a neous group: Eight had mild reversible ARI, three had ARI heterogeneous condition. Although mild reversible renal im- related to shock and multisystem failure, two had rhabdomy- pairment (serum creatinine Ͻ 0.30 mmol/L) (3.4 mg/dl) is olysis-related ARI, and two had severe isolated ARI. Two usual, severe acute renal failure requiring dialysis can occur. It patients required supportive hemodialysis and two died. Pa- is recommended that any patient who presents after near- tients with ARI experienced more marked acidosis than control drowning or immersion should be assessed for potential ARI patients, as measured by serum bicarbonate (P Ͻ 0.001), pH by serial estimations of serum creatinine, particularly when (P Ͻ 0.001), and base excess (P Ͻ 0.001). There was also a there is an increase in the initial serum creatinine, marked higher admission lymphocyte count in the ARI group (P ϭ metabolic acidosis, an abnormal urinalysis, or a significant 0.056). Dipstick on admission was significantly lymphocytosis.

Acute renal impairment (ARI) due to immersion and near- drowning and immersion has been described previously drowning in either salt or fresh water is a recognized but poorly (1–4,11), but a clear pattern of the entity has not emerged from defined disorder, with only a few sporadic cases described these anecdotal case reports. As our institution commonly (1–4). Although the diuretic effect of immersion has long been receives immersion and near-drowning victims from nearby recognized, the underlying physiologic changes following Sydney beaches, a retrospective study was undertaken (1)to whole-body immersion have only come to light within the last determine the incidence of ARI associated with near-drowning century (5–9). Much of the literature of near-drowning has or immersion and (2) to define the clinical syndrome and to concentrated on the respiratory effects of aspiration of salt and assess clinical predictors of ARI. fresh water, and on the management of both early and late respiratory complications such as aspiration pneumonia and Materials and Methods adult respiratory distress syndrome (10,11). However, near- Review of the records of 30 near-drowning victims admitted to our drowning and immersion can have profound effects on other institution was undertaken, and subjects with ARI were defined by a bodily systems. Cerebral (hypoxic brain injury/cerebral ab- serum creatinine Ͼ0.12 mmol/L (Ͼ1.36 mg/dl) during the admission, scess) (11,12), cardiac (atrial fibrillation) (13,14), and hema- which later normalized. A control group of near-drowning victims Յ tologic complications (disseminated intravascular coagulation with normal serum creatinine levels ( 0.12 mmol/L) was also iden- and hemolysis) (15) can result from immersion and near- tified. Patients and control subjects were compared with respect to a variety of demographic, clinical, and laboratory parameters, including drowning. Multisystem failure resulting from near-drowning is the circumstances of the immersion, circulatory indices, acid-base also well described (16). studies, and hematologic markers. Urinalysis was performed using The association of acute renal failure (ARF) with near- Ames Multistix (Bayer Diagnostics, Victoria, Australia). To correct for variations in concentration, the measured dipstick proteinuria and hematuria was divided by the urine-specific gravity. Received April 15, 1998. Accepted August 2, 1998. Correspondence to Dr. Timothy Spicer, Level 4, Health Services Building, Liverpool Hospital, Liverpool, Sydney, 2170 Australia. Phone: 61-2-9828- Statistical Analyses 3000; Fax: 61-2-9828-4970; E-mail: [email protected] Data were compared using unpaired t test or the Wilcoxon rank 1046-6673/1002-0382$03.00/0 sum test as appropriate. Categorical data were compared using the 2 Journal of the American Society of Nephrology conditional binomial test or Upton ␹ test. Risk factors were analyzed Copyright © 1999 by the American Society of Nephrology by univariate and multivariate logistic regression analysis, after back- J Am Soc Nephrol 10: 382–386, 1999 Acute Renal Impairment after Near-Drowning 383 ward elimination (Statistical Package for Interactive Data Analysis, parameters were consistent with a metabolic acidosis, but version 5.0, Sydney, Australia). All data are expressed as mean Ϯ SD. serum lactate levels were not available. On univariate logistic Ͻ A level of statistical significance was determined as P 0.05. regression analysis, serum bicarbonate (P ϭ 0.002), pH (P ϭ 0.001), and base excess (P Ͻ 0.001) were predictors of ARI. Results Base excess was the best predictor of acute renal impairment There were 15 subjects with ARI and 15 without renal (P Ͻ 0.01) by multivariate logistic regression analysis. Abso- impairment (Table 1). Overall, 80% of patients were male, lute lymphocytosis (Ͼ4.0 ϫ 109/L) was seen more commonly Ϯ with an average age of 35 15 yr. Immersion events occurred in the ARI group, a trend that approached statistical signifi- in salt water in 97% of cases, during the summer months in cance (P ϭ 0.056). Seven of the patients with ARI had an 37%, and between the hours of 1 p.m. and 7 p.m. in 77%. absolute lymphocytosis, with lymphocyte counts ranging from Overall, 26.4% of all patients required cardiopulmonary resus- 6.8 to 10.8 ϫ 109/L (mean 8.2 ϫ 109/L). citation (CPR), the mean admission duration was 2.9 Ϯ 3.4 d The admission serum creatine phosphokinase (CPK) (nor- (range, Ͻ24 h to 14 d), and the total mortality was 10%. The mal Ͻ130 U/L) was 8,764 Ϯ 23,416 U/L in the ARI group two groups were comparable with regard to immersion time, versus 496 Ϯ 743 U/L for the near-drowning control subjects CPR requirement, duration of admission, and admission clin- (P ϭ 0.13). This apparent marked difference was primarily due ical parameters (temperature, BP, heart rate, and urine output over the first 12 h). Atrial fibrillation was more common in the to two patients with and a clinical diagnosis of renal impairment group (33%) compared with non-ARI control rhabdomyolysis, who had admission serum CPK of 8,520 and subjects (13%), but this did not reach statistical significance 75,000 U/L, respectively. When these patients were excluded (P ϭ 0.10). In patients with ARI, the admission and peak from analysis, the CPK was comparable between ARI and Ϯ Ϯ serum creatinine values were 0.24 Ϯ 0.33 mmol/L and 0.34 Ϯ non-ARI groups (515 683 U/L versus 496 743 U/L, ϭ 0.41 mmol/L, respectively. Renal impairment was mild (peak respectively, P NS). Both patients with myoglobinuria pre- serum creatinine Ͻ0.30 mmol/L) in 73% of patients, but 27% sented with muscle pain, had been surfing in tightly fitting wet had a peak serum creatinine greater than 0.30 mmol/L. Serum suits, and had a longer immersion time than other patients with creatinine peaked on day 1 in 33%, and in the remainder it ARI (75 Ϯ 23 versus 14.3 Ϯ 16 min, respectively, P Ͻ continued to rise after admission. The ARI patients could be 0.0001). Neither of the myoglobinuric patients had seizures. divided into four clinical categories, which included mild re- Overall, only two patients had seizures, both of which were versible ARI (53.3%), ARI related to shock and multisystem described as “hypoxic” and which occurred soon after circu- failure (20%), rhabdomyolysis/myoglobinuric ARI (13.3%), latory arrest. Regarding the use of wet suits, only two other and severe isolated ARF (13.3%). Two patients required he- patients had been wearing wet suits. Both were divers who modialysis and two died. suffered cardiac arrest under water after brief immersions. One All patients with ARI were acidotic on arrival, as were 67% had ARI and neither developed rhabdomyolysis. of non-ARI control subjects. The degree of acidosis was sig- Admission dipstick urinalysis corrected for specific gravity nificantly greater in the ARI group, as measured by serum pH, showed that ARI patients had significantly more hematuria bicarbonate, and base excess (Table 2). Arterial blood gas than control patients (P ϭ 0.016). Those patients with 2 to 3ϩ

Table 1. Comparative demographic and clinical features at presentation according to the presence of renal impairmenta

No Renal Impairment Renal Impairment Parameter (n ϭ 15) (n ϭ 15) P Value

Age (yr) 36.9 Ϯ 18.9 33 Ϯ 11.2 NS Gender: male 67% 93% NS Admission duration (days) 2.1 Ϯ 2.7 3.8 Ϯ 4.1 NS Mortality 6.7% 13.3% NS Salt water immersion 93% 93% NS Immersion time (min) 9.8 Ϯ 7.4 25.2 Ϯ 28.5 NS CPR required 27% 33% NS CPR duration (min) 11.6 Ϯ 16.1 25 Ϯ 13.5 NS Admission temperatureb 36 Ϯ 1.0 36.1 Ϯ 1.0 NS Systolic BP (mmHg) 117 Ϯ 49 117 Ϯ 39.3 NS Diastolic BP (mmHg) 68 Ϯ 22 72 Ϯ 24 NS Heart rate (bpm) 98 Ϯ 33 103 Ϯ 41 NS Atrial fibrillation 13% 40% 0.10 Urine (ml/12 h) 1415 Ϯ 745 1087 Ϯ 662 NS

a CPR, cardiopulmonary resuscitation. b Results given in degrees Celsius. 384 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 382–386, 1999

Table 2. Laboratory data of 30 near-drowning admissions according to presence of renal impairmenta

No Renal Impairment Renal Impairment Parameter (n ϭ 15) (n ϭ 15) P Value Ϯ Ϯ pO2 (mmHg) 201 155.7 208 175 NS pH 7.28 Ϯ 0.13 7.01 Ϯ 0.22 Ͻ0.001 Serum bicarbonate (mmol/L) 20.6 Ϯ 5.0 12.7 Ϯ 6.6 Ͻ0.001 Base excess Ϫ5.4 Ϯ 5.9 Ϫ21 Ϯ 10 Ͻ0.001 Serum urea (mmol/L) 5.2 Ϯ 1.6 8.9 Ϯ 8.4 NS Admission serum creatinine (mmol/L) 0.09 Ϯ 0.01 0.24 Ϯ 0.33 Ͻ0.001 Peak serum creatinine (mmol/L) 0.09 Ϯ 0.01 0.34 Ϯ 0.2 Ͻ0.001 Hemoglobin (g/dl) 14.4 Ϯ 1.1 14.9 Ϯ 1.6 NS Leukocyte count (ϫ109/L) 13.2 Ϯ 5.5 15.4 Ϯ 5.8 NS Lymphocytesb (ϫ109/L) 2.3 Ϯ 1.9 4.4 Ϯ 2.9 0.056

a pO2, partial pressure of oxygen. To convert creatinine from mmol/L to mg/dl, multiply by 11.3. b Normal range Ͻ4.0 ϫ 109/L. of dipstick proteinuria developed higher peak serum creatinine Lymphocytic infiltrate was not prominent (the patient’s admis- levels than those with less abnormal dipstick results (P Ͻ 0.05) sion lymphocyte count was 9.0 ϫ 109/L). (Figure 1). The presence of 3ϩ was also associated with a higher peak serum creatinine, but this failed Discussion to reach statistical significance (P ϭ 0.08). Alcohol and drug This study demonstrates that near-drowning-associated ARI screens were only performed on clinical suspicion, but showed is more common than previously recognized, occurring in 50% a definite yield. Drug screens were performed in two patients, of a population of near-drowning emergency room admissions, and one non-ARI patient tested positive for benzodiazepines. and is a heterogeneous clinical entity. Although the resultant Of six blood alcohol screens, two patients registered a blood ARI was usually mild and reversible, we occasionally observed alcohol level greater than 0.05, one each from the ARI and ARI associated with shock and multisystem failure, severe non-ARI groups. isolated ARF, and ARI due to rhabdomyolysis and myoglo- Regarding the course of the renal impairment, 30% of pa- binuria. Severe ARI requiring dialysis occurred in 7% of tients were oliguric for at least the first 48 h, the rest having immersion victims. Although severe acidosis on admission was nonoliguric ARI. Only one renal biopsy was performed (in a the best predictor of subsequent ARI, lymphocytosis or prom- patient who required dialysis), and revealed vacuolization of inent dipstick proteinuria may also presage renal impairment. the tubular epithelium, loss of tubular cell nuclei, and intersti- Immersion and near-drowning injury has been widely re- tial edema, consistent with acute tubular injury (Figure 2). viewed (10–13), and widespread tissue hypoxia and subse- quent reperfusion injury are thought to be the predominant underlying , although hypovolemia and hypo- thermia may also contribute to tissue damage. The best pre- dictor of acute renal impairment in our study was the severity of presenting acidemia, as measured by a low base excess and reduced serum bicarbonate and pH. Metabolic acidosis is com- mon after near-drowning, and is due to lactic acidosis induced by widespread tissue hypoxia (11,16,17). That lactate was the probable cause of acidosis in our series was supported by the increased and reversible nature of the acidosis. It is likely that the degree of initial acidosis is a surrogate marker of the severity of the hypoxic insult. Acidosis is a better marker of

injury in this setting than the initial pO2, because acidosis persists longer, after reoxygenation and reperfusion have oc- curred. Admission dipstick hematuria was associated with the development of ARI, and prominent urinary protein (2 to 3ϩ) on admission dipstick urinalysis was associated with a higher peak serum creatinine. Admission dipstick urinalysis may therefore be a simple and useful screening test for potential Figure 1. Admission urinalysis versus mean peak serum creatinine ARI in immersion victims. Prominent microhematuria (3ϩ) (mean Ϯ SEM). Solid bars represent hematuria and hatched bars should also prompt a check for myoglobinuria and rhabdomy- proteinuria. *P Ͻ 0.05. olysis. J Am Soc Nephrol 10: 382–386, 1999 Acute Renal Impairment after Near-Drowning 385

Figure 2. after near-drowning. Renal biopsy specimen from a patient with serum creatinine of 0.86 mmol/L, 2 d after near-drowning. A proximal tubule with necrotic and denuded epithelium is shown.

Although leukocytosis is common after near-drowning hypoxemia and acidosis were quickly corrected in these pa- (12,14), an absolute lymphocytosis has not previously been tients, but the initial mild renal impairment progressed to reported. Interestingly, lymphocytosis has been observed in established ARF and required supportive dialysis. Previous response to the cold pressor stress test with immersion of hands case reports of near-drowning-associated ARI have described in ice-cold water (18). However, as there were no significant “isolated ARF,” and although its pathophysiology is contro- differences between the groups with regard to admission tem- versial, postulated contributors include hypovolemia, hypo- perature or hemodynamic status, the association of peripheral thermia, and rhabdomyolysis. (1,2,4). Our data, however, sug- lymphocytosis with renal impairment in our study is likely to gest that this entity represents an ischemia/reperfusion injury reflect the effects of prior hypoxic injury. Whether the lym- that follows the multifactorial hypoxemia resulting from near- phocytosis contributes to renal injury is unresolved, but it is drowning. Why renal injury predominates in the absence of interesting to note recent studies suggesting a link between other postimmersion injury sequelae is not clear from our data. hypoxia-reperfusion injury and cytotoxic immune effector A renin-angiotensin surge when returning to dry land after mechanisms (19,20). whole-body immersion may play a role (7,13). Acute tubular The ARI associated with cardiac arrest and shock occurred necrosis, seen in the one renal biopsy performed in our series, in patients who suffered a profound insult requiring CPR. has been reported previously (1,4) and represents this more These patients usually developed multisystem failure and had severe end of the spectrum. a substantial mortality and morbidity. ARI associated with In summary, ARI after near-drowning is a relatively com- clinical rhabdomyolysis and myoglobinuria was characterized mon event, is usually mild and self-limited, and comprises a by longer immersion times and the wearing of a wet suit. Wet heterogeneous group of conditions. Myoglobinuric renal fail- suit-related ARF has not been previously described, and we ure due to prolonged immersion and the wearing of a tight wet hypothesize that an excessively tight wet suit may function as suit is part of this spectrum, as is renal failure associated with a tourniquet in creating muscle ischemia, especially when shock and multisystem failure, and isolated severe acute renal associated with muscle swelling after prolonged and vigorous failure requiring dialysis. It is recommended that any patient muscular effort. presenting after near-drowning or immersion should have se- Of the remaining patients with ARI, most experienced only rial serum creatinine measurements, particularly in the pres- a mild transient increase in serum creatinine; however, there ence of an elevated initial creatinine, marked metabolic acido- was a subgroup that developed a disproportionately severe sis, abnormal urinalysis, or a peripheral lymphocytosis. acute renal failure, which we have labeled “isolated ARF.” These patients did not usually require CPR after their near- drowning episode, and are similar to the two patients originally References described by Grausz et al. in 1971 (1), who presented with ARI 1. Grausz H, Amend WJ, Farley LE: Acute renal failure compli- several days after the immersion. In our study, the admission cating submersion in sea water. JAMA 217: 207–209, 1971 386 Journal of the American Society of Nephrology J Am Soc Nephrol 10: 382–386, 1999

2. Agar J: Rhabdomyolysis and acute renal failure after near drown- 11. Fuller RH: Drowning and the post immersion syndrome: A ing in cold salt water. Med J Aust 161: 686–687, 1994 clinicopathologic study. Mil Med 128: 22–40, 1963 3. Neale TJ, Dewar JM, Parr R, Kimber J, Hatfield PJ, Dixon P: 12. Oakes D, Sherk J: Prognosis and management of victims of near Acute renal failure following near drowning in salt water. NZ drowning. J Trauma 22: 544–549, 1982 Med J 97: 319–322, 1984 13. Pearn J: Pathophysiology of drowning. Med J Aust 142: 586– 4. Oren A, Etziou Z, Broitman D, Yagil R: Renal involvement 588, 1985 following near-drowning in the sea. Comp Biochem Physiol A 14. Modell J, Graves S: Clinical course of 91 consecutive near 73: 175–179, 1982 drowning victims. Chest 70: 231–238, 1976 5. Hartshorne H: Water vs Hydrotherapy—Or an Essay on Water 15. Ports TA, Deuel J: Intravascular coagulation in freshwater sub- and Its True Relation to Medicine, Philadelphia, Lloyd Smith mersion: Report of 3 cases. Ann Intern Med 87: 60–61, 1977 Press, 1847, p 28 16. Hoff BH: Multisystem failure: A review with special reference to 6. Epstein M: Renal effects of head out of water immersion in man. drowning. Crit Care Med 7: 310–320, 1979 Physiol Rev 58: 529–581, 1978 17. Modell J, Davis J: Blood gas and electrolyte changes in human 7. Gooden BA: Mechanism of the human diving response. Med J near drowning victims. JAMA 203: 99–105, 1968 Aust 157: 629–632, 1992 18. Chi DS, Neumann JK, Mota-Marquez M, Dubberley FA: Effects 8. Ogihara T, Shima J, Tabuchi Y, Hashizune K, Kumahara Y: Significant increases in plasma immunoreactive atrial polypep- of acute stress on lymphocyte beta 2-adrenoreceptors in white tide concentration during head out of water immersion. Life Sci males. J Psychosom Res 37: 763–770, 1993 38: 2413–2418, 1986 19. Takada M, Chandraker A, Nadeau KC, Sayegh MH, Tilney NL: 9. Larah JH, Atlas SA: Atrial natriuretic hormone: A regulator of The role of the B7 costimulatory pathway in experimental cold blood pressure and volume homeostasis. Int 34[Suppl ischemia/reperfusion injury. J Clin Invest 100: 1199–1203, 1997 25]: 564–571, 1988 20. Azuma H, Nadeau K, Takada M, Tilney NL: Cellular and mo- 10. Levin DL, Morriss FC, Toro LO, Brink LW, Turner GR: Drown- lecular predictors of chronic renal dysfunction after initial isch- ing and near-drowning. Pediatr Clin North Am 40: 321–336, emia/reperfusion injury of a single kidney. Transplantation 64: 1993 190–197, 1997