OBJECTIVES. The goals were to (1) compare the causes, clinical presentation, and prevalence of acute renal failure in pediatric rhabdomyolysis with the published data for adults; (2) determine predictors of acute renal failure in pediatric patients with rhabdomyolysis; and (3) explore the relationship of acute renal failure with treatment modalities such as fluid and bicarbonate administration.
METHODS. We performed a retrospective chart review to identify patients with creatinine kinase levels of >1000 IU/L who were treated in the emergency department of a tertiary pediatric hospital between 1993 and 2003, and we constructed regression models.
RESULTS. Two hundred ten patients were studied. One hundred ninety-one patients met study eligibility (128 male and 63 female), with a median age of 11 years. The most common documented symptoms were muscle pain (45%), fever (40%), and symptoms of viral infection (39%). The most common causes of pediatric rhabdomyolysis were viral myositis (38%), trauma (26%), and connective tissue disease (5%). Six of 37 patients with creatinine kinase levels of ≥6000 IU/L had previously undiagnosed dermatomyositis or hereditary metabolic disease, compared with 10 of 154 patients with creatinine kinase levels of 1000 to 5999 IU/L. Nine of 191 patients developed acute renal failure. None of 99 patients with initial urinary heme dipstick results of <2+ developed acute renal failure, compared with 9 of 44 patients with urinary heme dipstick results of ≥2+. Higher initial creatinine kinase levels and higher fluid administration rates were associated with higher maximal creatinine levels.
CONCLUSIONS. The cause of acute pediatric rhabdomyolysis is different from that of adult rhabdomyolysis. The risk of acute renal failure in children is much less than the risk reported for adults.
Rhabdomyolysis was first described in German medical literature by Fleischer in 1881. However, it was not until the work of Bywaters and colleagues,1,2 who reported 4 cases of renal failure in crush victims during the bombing of London in the Battle of Britain, that the relationship between rhabdomyolysis and acute renal failure (ARF) was established.
Since the early work by Bywaters and colleagues,1,2 numerous descriptive studies of rhabdomyolysis in adults have been published.3–6 The pathogenesis of the condition has been well characterized in adults, and the associated rate of ARF has been estimated to be between 17% and 35%.3,6
Data on pediatric rhabdomyolysis are limited. Only 2 small case series, with <20 patients each, have attempted to characterize the pathogenesis and prevalence of ARF.7,8 These case series reported ARF rates of 42%8 and 50%.7 The pathogenesis of rhabdomyolysis in the pediatric ambulatory population and the risk of developing ARF are poorly characterized.9–11
Our clinical impression has been that rhabdomyolysis in children is a more benign condition, with only a small risk of ARF. Our study objectives were to establish more clearly the pathogenesis of rhabdomyolysis and the associated rates of ARF in pediatric patients, to establish predictors of ARF in these patients, and to explore the relationship of maximal creatinine levels with fluid and bicarbonate administration.
By using the hospital laboratory system, we identified all patients with creatinine kinase (CK) levels of >1000 IU/L within 72 hours after admission to the emergency department (ED) of a tertiary urban pediatric hospital during a 10-year period (June 1993 to June 2003), and we reviewed their medical records. Patients were excluded if they had (1) a documented history of muscular dystrophy or other metabolic muscle disorders, (2) a history of myocardial damage with a documented CK-MB fraction of >5%, or (3) age of >21 years.
The medical records were reviewed by using a standardized form. All charts underwent duplicate review by any 2 of the 4 authors. Discrepancies were resolved through consensus before data analysis.
We abstracted historical, examination, and laboratory data from the ED and inpatient records. Historical information included age, weight, gender, fever, viral infection symptoms (cough, sore throat, rhinitis, diarrhea, or vomiting), exercise, muscle or abdominal pain, dark urine, intramuscular injection, surgery, trauma, and medication use. Physical examination data included maximal temperature, heart rate (mean of heart rates measured in the ED), abdominal or muscle tenderness, strength, and reflexes. Laboratory data included CK, electrolyte, serum urea nitrogen, creatinine, phosphorus, lactic dehydrogenase, alanine and aspartate aminotransferase, and myoglobin levels, urinary heme orthotolidine level, urinalysis for red blood cells, urinary myoglobin level, and urine toxicology screens (for amphetamine, opiates, barbiturates, phencyclidine, cannabis, and benzodiazepines). Other clinical data included the amount of fluid administered within the first 24 hours, bicarbonate therapy, development of ARF (defined below), death, and length of stay. Follow-up data included the date of the last visit, last creatinine value, documentation of diagnosis of a metabolic or myopathic cause for the rhabdomyolysis, any recurrence of rhabdomyolysis, and any chronic renal sequelae.
We defined rhabdomyolysis as a serum CK level of >1000 IU/L in the absence of myocardial infarction. CK decline was calculated for admitted patients by using the peak CK level subtracted from the next CK level, divided by the time interval, as recorded by the laboratory.
ARF was defined as a creatinine level of >97.5th percentile for age and gender.12 The final cause of the rhabdomyolysis and ARF was determined on the basis of the patient's discharge diagnosis or diagnosis through subspecialty follow-up assessments. If the final diagnosis was rhabdomyolysis, then we reviewed the medical records to assign a cause (Table 1). Discrepancies were resolved through consensus.
Our data analysis focused on the following aims and hypotheses. (1) What is the cause, clinical presentation, and prevalence of ARF among children with rhabdomyolysis? (2) What are the predictors of ARF for patients with rhabdomyolysis? (3) What is the relationship between fluid and bicarbonate administration and the development of ARF? To address aim 1, we assigned causes of rhabdomyolysis by using the criteria listed in Table 1, tabulated the number of children with ARF, and calculated an unadjusted binomial 95% confidence interval (CI).
In the next stage of our analysis, we sought predictors of maximal creatinine levels in children with rhabdomyolysis. A priori, we chose gender, age, initial CK level, initial serum bicarbonate level, urinary heme dipstick result, and rate of fluid administration as the most likely predictors, on the basis of biological considerations. This approach was chosen to avoid well-known problems with stepwise logistic regression techniques, in which the final model may vary depending on the order in which variables are entered. Because our biological assumptions of association might not be complete, we also explored whether all covariates measured were associated with maximal creatinine levels in bivariate analyses. Student's t test or analysis of variance was used for bivariate analysis. The correlation and bivariate association between potential predictors and serum creatinine levels were measured by using Pearson's correlation and standard linear regression analyses. Any association between clinical/demographic parameters and maximal serum creatinine levels was explored with the use of bivariate scatter plots for continuous variables and cross-tabulations for discrete variables. The clinical variables that were selected a priori as predictors, as described above, or were found to be predictors of elevated serum creatinine levels on the basis of these bivariate associations (ie, P < .1 in the bivariate analysis) were then entered into a multivariate linear regression model with initial CK levels, with maximal creatinine levels as the dependent variable.
To test for colinearity, we generated new models excluding 1 covariate (age, gender, initial CK level, intravenous fluid rate, urinary heme dipstick result, serum bicarbonate level at presentation, or whether the patient received bicarbonate) and compared the R2 of the full model with that of the model missing a covariate. We also used the variance inflation factor (VIF) function in Stata 8.0 (Stata Corp, College Station, TX) to determine whether any of the covariates had a VIF of >10 (a VIF of >10 is comparable to a tolerance of <0.1). Stata 8.0 for Windows was used for all statistical calculations.
Subjects and Initial Findings
A total of 211 subjects were identified through the laboratory database search. One chart was missing. Two hundred ten charts were reviewed and 19 subjects were excluded, because of a previous diagnosis of McArdle's disease, carnitine palmitoyl transferase II deficiency, or muscular dystrophy (6 patients), cardiac abnormalities noted on echocardiograms and CK-MB fractions of >5% (3 patients), age of >21 years (5 patients), or lack of evaluation in the ED (5 patients; these patients were admitted directly to the ICU or to the inpatient unit). For the remaining 191 subjects (128 male subjects and 63 female subjects), the median age was 11 years (interquartile range [IQR]: 6–16 years).
Viral myositis and trauma were the most common causes, together accounting for 64% of patients (Table 2). Of the patients diagnosed as having viral myositis (n = 73), 4 were found to be positive for influenza B and 2 were positive for influenza A.
Ten (6%) of 154 of patients with CK levels of 1000 to 5999 IU/L had previously undiagnosed dermatomyositis or hereditary metabolic disease, compared with 6 (16%) of 37 patients with CK levels of >6000 IU/L (P = .02). These patients presented with acute-onset rhabdomyolysis and received diagnoses subsequently through follow-up evaluations at specialty clinics.
The leading diagnosis in the 0- to 9-year age group was presumed viral infection, and the leading diagnosis in the 9- to 18-year age group was trauma (Fig 1). The most common documented presenting symptoms were muscle pain (45%), fever (40%), symptoms of viral infection (39%), and muscle weakness (38%). Only 3.6% of patients reported a history of dark urine. Only 1 patient presented with the “classic triad” of myalgia, weakness, and dark urine. Muscle tenderness was documented for 39% of patients, abdominal tenderness for 12% of patients, decreased strength for 9.4% of patients, and decreased reflexes for 5.2% of patients.
Rate of ARF
Nine (5%; 95% CI: 2%–8%) of 191 patients developed ARF during the course of their hospitalization. The overall rate of ARF solely attributable to rhabdomyolysis was 3 (1.6%; 95% CI: 0%–3.3%) of 191 patients. We found no seasonal association with rhabdomyolysis or the development of ARF.
Of the 143 patients with initial urinary heme dipstick results recorded, 0 of 99 patients with initial urinary heme dipstick results of <2+ developed ARF (95% CI upper limit: 3%), compared with 9 (20%; 95% CI: 10%–35%) of 44 patients with urinary heme dipstick results of ≥2+. In bivariate analysis, no statistically significant difference was found in the initial sodium, potassium, bicarbonate, or calcium levels of the ARF group versus the non-ARF group. The ARF group did receive bicarbonate therapy more frequently (36% in the ARF group and 12% in the non-ARF group; P < .05).
Six of 9 patients developed ARF in the setting of multiorgan system failure; 4 of those patients died (Table 3). The remaining 3 patients who developed ARF were thought to have ARF solely attributable to rhabdomyolysis; 2 of those patients developed symptoms after an episode of extreme exercise and 1 developed symptoms after consuming a large quantity of ethyl alcohol. One of the 3 patients who developed rhabdomyolysis-induced ARF required dialysis. All 3 patients survived and had normal creatinine levels documented at a follow-up visit. One of the 3 patients was diagnosed as having McArdle's disease. The other 2 patients underwent extensive metabolic and genetic evaluations, but no cause could be determined.
Fluid and Sodium Bicarbonate Therapy
One hundred twenty-three patients (64%) were admitted. There was no difference in median CK levels between patients who were admitted to the hospital (2315 IU/L; IQR: 1410–5890 IU/L) and those who were discharged from the ED (2210 IU/L; IQR: 1440–3790 IU/L). Forty-eight of 123 admitted patients had follow-up CK levels measured after the peak levels.
One hundred twenty-three of 191 patients had their fluid intake and output documented during their hospitalization. The results of the linear regression model for peak serum creatinine levels are shown in Table 4. There was no evidence of colinearity, and all VIF values were <2. Our results indicated that higher levels of fluid therapy and higher initial CK levels were associated with higher maximal creatinine levels. There was no association between bicarbonate administration and maximal creatinine levels.
Morbidity and Mortality Rates
The overall mortality rate was 13 of 191 patients. Nine of the 13 patients arrived in the ED in cardiopulmonary arrest and could not be resuscitated. The cause was trauma (6 of 9 patients), drowning (2 of 9 patients), or acute drug overdose (1 of 9 patients). Four of 13 patients died later, 2 as a result of sepsis, 1 as a result of multiorgan failure, and 1 as a result of malignant hyperthermia. The patient with malignant hyperthermia was diagnosed after death as having aldolase A deficiency. No patient died as a result of rhabdomyolysis-related complications.
Of the 178 patients who survived, 112 (63%) had documented follow-up care through either a specialist outpatient clinic or a return visit to the ED. Fifty-three of the 112 patients had follow-up creatinine levels measured after discharge, and all levels were normal. There was no report of any patient developing chronic renal sequelae as a result of the acute rhabdomyolysis. All 5 patients with ARF who survived had normal creatinine levels at follow-up assessments (5–52 months).
Rhabdomyolysis, which literally means “dissolution of striped [skeletal] muscle,” is the final common pathway of many different processes, all of which end in skeletal muscle injury and the subsequent release of muscle cell contents, including CK, myoglobin, potassium, and phosphorus, into plasma. The classic triad of symptoms of rhabdomyolysis includes myalgia, weakness, and dark urine, although these findings may be inconsistent.13 The definitive diagnosis of rhabdomyolysis requires an elevation of CK levels to >5 times normal in the absence of significant elevations of brain or cardiac CK fractions. The most dangerous sequela of rhabdomyolysis is ARF, the exact mechanisms of which is unclear but may be attributable to vasoconstriction/hypoperfusion, renal tubular dysfunction/cast formation, and/or myoglobin-induced tubular cytotoxicity.7,14,15 The mainstay of treatment for rhabdomyolysis, directed at preventing ARF, is fluid therapy. Many clinicians advocate alkalinization of urine with sodium bicarbonate (sometimes with concomitant forced diuresis with mannitol). There are no data that suggest that this strategy prevents ARF in children with rhabdomyolysis.
Most of the literature on rhabdomyolysis is drawn from adult studies. In one of the largest case series, Gabow et al3 found that the most common risk factors for rhabdomyolysis in adults included alcohol abuse (67%), recent soft-tissue compression (39%), and seizure activity (24%). In contrast to most pediatric patients, the majority of adult patients studied by Gabow et al3 had multiple risk factors. The 1988 review by Ward6 of 157 patients found trauma (38%), ischemia (14%), and polymyositis (8%) to be the most frequent causes of rhabdomyolysis. Viral illness accounted for only 1% of the patients examined by Ward.6 These findings are very different from our clinical experience and led us to hypothesize that the cause and perhaps treatment and outcomes for children with rhabdomyolysis are far different from the adult reports.
The data for children are very limited. Compared with the larger studies by Ward6 and Gabow et al,3 2 of the largest pediatric case series to date included a total of 39 patients.7,8 In one of those studies, Watemberg et al8 reviewed 19 pediatric cases and found that trauma (n = 5), metabolic disorders (n = 4), and viral myositis (n = 2) were the most common causes. These findings seemed contrary to our clinical experience, in which viral myositis accounted for a large number of children with rhabdomyolysis. Our series confirmed our clinical expectation and showed that viral myositis accounted for more than one third of the cases. However, the cause of rhabdomyolysis in patients >9 years of age in our series resembled more closely the data of Ward6 and Gabow et al.3 Viral myositis caused the majority of rhabdomyolysis in the first decade of life, whereas trauma and drug-related causes had their peaks in the second decade of life. Our study also confirmed the findings by Watemberg et al8 that a relatively large proportion of patients (1 of 6 patients with CK levels of >6000 IU/L presenting to a tertiary care center) had an underlying metabolic or rheumatologic cause. Given the 7% rate of “unknown” causes, we think that the rate of metabolic causes actually might be higher.
Although the classic clinical presentation of a patient with rhabdomyolysis is that of myalgia, dark urine, and muscle weakness, our study found that only 45% of patients presented with myalgia and 38% had documented complaints of muscle weakness. Only 3.6% of patients noted dark urine. Only 1 patient had all 3 symptoms recorded.
As shown previously in adult studies, muscular signs on initial examination do not seem to be discriminatory or indicative of rhabdomyolysis. On examination, 39% of patients had documented muscle tenderness but only 9.4% had decreased muscle strength. We attribute these findings to the difficulty of examining children and the potentially insidious onset of rhabdomyolysis.
Previous studies of rhabdomyolysis suggested that the rates of ARF secondary to rhabdomyolysis range from 17% to 35% in adults and from 42% to 50% in children.3,6–8 Our study shows that the rate of ARF with pediatric rhabdomyolysis is in fact much lower than reported previously (5%; 95% CI: 2%–8%). This low rate may be elevated by the presence of other comorbid factors; the majority of patients who developed ARF in our study had documented multiorgan system disease at presentation. In an attempt to address this limitation, additional examination of the ARF group was undertaken, and the rate of ARF solely attributable to rhabdomyolysis was found to be 1.6% (95% CI: 0%–3.3%). Of the patients who developed ARF solely attributable to rhabdomyolysis, all recovered and none developed chronic renal failure.
Many researchers compared the chemical findings for patients who developed rhabdomyolysis-associated ARF eventually with those for patients who experienced rhabdomyolysis without ARF, in an attempt to characterize the risk factors for the development of ARF.3,6–8,16,17 There is a lack of such data in the pediatric literature. In our study, the rate of ARF was lower than in adult studies and we were unable to detect many significant differences in initial serum sodium, potassium, bicarbonate, or calcium levels for patients who developed ARF, compared with those who did not.
Urinalysis was more clinically revealing, because no patient with an initial urinary heme dipstick result of <2+ went on to develop ARF. Although our data demonstrated that patients with negative heme dipstick results are at much lower risk of developing ARF, the diagnosis of rhabdomyolysis should not be excluded solely on the basis of negative heme dipstick results. Fifty-six percent of patients with CK levels of >1000 IU/L had negative heme orthotolidine dipstick results. Our data suggest that this test may be useful as a screen for determining the need to monitor for renal toxicity patients who present with only modest elevations in serum CK levels.
Fluid therapy and bicarbonate therapy are standard prophylactic treatments for the prevention of ARF secondary to rhabdomyolysis. Our study found that higher fluid rates predicted higher maximal creatinine levels. This may be an effect of clinicians' responses to clinical severity, rather than a true causal effect, because patients who were more ill might have had higher initial CK and creatinine levels and therefore treating clinicians might have administered more fluid. No association was found between bicarbonate administration and maximal creatinine levels. Additional prospective studies are needed to determine the effects of fluid rates and bicarbonate administration on renal toxicity in patients with rhabdomyolysis.
Our study represents the largest case series of pediatric rhabdomyolysis. Our findings demonstrate that the pathogenesis of pediatric rhabdomyolysis is quite different from that reported in the adult literature. We found a much lower rate of ARF, compared with previous studies in adults and children. Although no single symptom, sign, or laboratory result is highly predictive of ARF, urinary heme dipstick results of <2+ seem to indicate a much-reduced risk of developing ARF. Although the administration of fluids was associated with increased maximal creatinine levels, this effect might be attributable to more clinically ill patients receiving higher fluid rates, rather than a causal link.
- Accepted July 18, 2006.
- Address correspondence to Rebekah Mannix, MD, Division of Emergency Medicine, Children's Hospital, 300 Longwood Ave, Boston, MA 02115. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Bywaters EG, Beall D. Crush injuries with impairment of renal function. Br Med J.1941;1 :427– 432
- ↵Bywaters EG, Stead J. The production of renal failure following injection of solution containing myohaemoglobin. Q J Exp Physiol.1944;33 :53– 70
- ↵Watemberg N, Leshner RL, Armstrong BA, Lerman-Sagie T. Acute pediatric rhabdomyolysis. J Child Neurol.2000;15 :222– 227
- ↵Goebel J, Harter HR, Boineau FG, el-Dahr SS. Acute renal failure from rhabdomyolysis following influenza A in a child. Clin Pediatr (Phila).1997;36 :479– 481
- Friedman BI, Libby R. Epstein-Barr virus infection associated with rhabdomyolysis and acute renal failure. Clin Pediatr (Phila).1986;25 :228– 229
- ↵Taylor WR, Prosser DI. Acute renal failure, acute rhabdomyolysis and falciparum malaria. Trans R Soc Trop Med Hyg.1992;86 :361
- ↵Walsh PC, Retik AB, Vaughan ED, Wein AJ, eds. Campbell's Urology. 8th ed. Philadelphia, PA: WB Saunders; 2002
- ↵Braun SR, Weiss FR, Keller AI, Ciccone JR, Preuss HG. Evaluation of the renal toxicity of heme proteins and their derivatives: a role in the genesis of acute tubule necrosis. J Exp Med.1970;131 :443– 460
- Copyright © 2006 by the American Academy of Pediatrics