[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.158.119.60. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Download PDF
Figure 1.
Rates of In-hospital Mortality or Acute Kidney Injury Requiring Renal Replacement Therapy, Stratified by Cause of Rhabdomyolysis
Rates of In-hospital Mortality or Acute Kidney Injury Requiring Renal Replacement Therapy, Stratified by Cause of Rhabdomyolysis

AKI indicates acute kidney injury; RRT, renal replacement therapy.

Figure 2.
Probability of In-hospital Mortality or Acute Kidney Injury Requiring Renal Replacement Therapy
Probability of In-hospital Mortality or Acute Kidney Injury Requiring Renal Replacement Therapy

Dark blue bars and light blue bars show the observed risk at Massachusetts General Hospital (MGH) and Brigham and Women’s Hospital (BWH), respectively. RRT indicates renal replacement therapy.

Table 1.  
Baseline Clinical Characteristics, Cause of Rhabdomyolysis, Laboratory Values, and Outcomes Associated With Rhabdomyolysis
Baseline Clinical Characteristics, Cause of Rhabdomyolysis, Laboratory Values, and Outcomes Associated With Rhabdomyolysis
Table 2.  
Outcomes by Category of Risk Score
Outcomes by Category of Risk Score
Table 3.  
Risk Score
Risk Score
1.
Bosch  X, Poch  E, Grau  JM.  Rhabdomyolysis and acute kidney injury. N Engl J Med. 2009;361(1):62-72.
PubMedArticle
2.
Melli  G, Chaudhry  V, Cornblath  DR.  Rhabdomyolysis: an evaluation of 475 hospitalized patients. Medicine (Baltimore). 2005;84(6):377-385.
PubMedArticle
3.
Delaney  KA, Givens  ML, Vohra  RB.  Use of RIFLE criteria to predict the severity and prognosis of acute kidney injury in emergency department patients with rhabdomyolysis. J Emerg Med. 2012;42(5):521-528.
PubMedArticle
4.
de Meijer  AR, Fikkers  BG, de Keijzer  MH, van Engelen  BG, Drenth  JP.  Serum creatine kinase as predictor of clinical course in rhabdomyolysis: a 5-year intensive care survey. Intensive Care Med. 2003;29(7):1121-1125.
PubMedArticle
5.
Waikar  SS, Wald  R, Chertow  GM,  et al.  Validity of International Classification of Diseases, Ninth Revision, Clinical Modification codes for acute renal failure. J Am Soc Nephrol. 2006;17(6):1688-1694.
PubMedArticle
6.
Rhee  CM, Bhan  I, Alexander  EK, Brunelli  SM.  Association between iodinated contrast media exposure and incident hyperthyroidism and hypothyroidism. Arch Intern Med. 2012;172(2):153-159.
PubMedArticle
7.
Kidney Disease:Improving Global Outcomes (KDIGO) Acute Kidney Injury Work Group.  KDIGO clinical practice guideline for acute kidney injury. Kidney Int. 2012;2(suppl):1-138.Article
8.
Wijeysundera  DN, Karkouti  K, Dupuis  JY,  et al.  Derivation and validation of a simplified predictive index for renal replacement therapy after cardiac surgery. JAMA. 2007;297(16):1801-1809.
PubMedArticle
9.
Mehran  R, Aymong  ED, Nikolsky  E,  et al.  A simple risk score for prediction of contrast-induced nephropathy after percutaneous coronary intervention: development and initial validation. J Am Coll Cardiol. 2004;44(7):1393-1399.
PubMed
10.
Minne  L, Abu-Hanna  A, de Jonge  E.  Evaluation of SOFA-based models for predicting mortality in the ICU: a systematic review. Crit Care. 2008;12(6):R161. doi:10.1186/cc7160.
PubMedArticle
11.
Ho  KM, Lee  KY, Williams  T, Finn  J, Knuiman  M, Webb  SA.  Comparison of Acute Physiology and Chronic Health Evaluation (APACHE) II score with organ failure scores to predict hospital mortality. Anaesthesia. 2007;62(5):466-473.
PubMedArticle
12.
Peres Bota  D, Melot  C, Lopes Ferreira  F, Nguyen Ba  V, Vincent  JL.  The Multiple Organ Dysfunction Score (MODS) versus the Sequential Organ Failure Assessment (SOFA) score in outcome prediction. Intensive Care Med. 2002;28(11):1619-1624.
PubMedArticle
13.
Ward  MM.  Factors predictive of acute renal failure in rhabdomyolysis. Arch Intern Med. 1988;148(7):1553-1557.
PubMedArticle
14.
Gabow  PA, Kaehny  WD, Kelleher  SP.  The spectrum of rhabdomyolysis. Medicine (Baltimore). 1982;61(3):141-152.
PubMed
15.
Hoffman  MD, Ingwerson  JL, Rogers  IR, Hew-Butler  T, Stuempfle  KJ.  Increasing creatine kinase concentrations at the 161-km Western States Endurance Run. Wilderness Environ Med. 2012;23(1):56-60.
PubMedArticle
16.
Clarkson  PM, Kearns  AK, Rouzier  P, Rubin  R, Thompson  PD.  Serum creatine kinase levels and renal function measures in exertional muscle damage. Med Sci Sports Exerc. 2006;38(4):623-627.
PubMedArticle
17.
Clarkson  PM, Hoffman  EP, Zambraski  E,  et al.  ACTN3 and MLCK genotype associations with exertional muscle damage. J Appl Physiol. 2005;99(2):564-569.
PubMedArticle
18.
Hubal  MJ, Devaney  JM, Hoffman  EP,  et al.  CCL2 and CCR2 polymorphisms are associated with markers of exercise-induced skeletal muscle damage. J Appl Physiol. 2010;108(6):1651-1658.
PubMedArticle
19.
Heled  Y, Bloom  MS, Wu  TJ, Stephens  Q, Deuster  PA.  CK-MM and ACE genotypes and physiological prediction of the creatine kinase response to exercise. J Appl Physiol. 2007;103(2):504-510.
PubMedArticle
20.
Woodrow  G, Brownjohn  AM, Turney  JH.  The clinical and biochemical features of acute renal failure due to rhabdomyolysis. Ren Fail. 1995;17(4):467-474.
PubMedArticle
21.
Akmal  M, Bishop  JE, Telfer  N, Norman  AW, Massry  SG.  Hypocalcemia and hypercalcemia in patients with rhabdomyolysis with and without acute renal failure. J Clin Endocrinol Metab. 1986;63(1):137-142.
PubMedArticle
Original Investigation
October 28, 2013

A Risk Prediction Score for Kidney Failure or Mortality in Rhabdomyolysis

Author Affiliations
  • 1Renal Division, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts
  • 2Framingham Heart Study, National Heart, Lung, and Blood Institute, and Center for Population Studies, Framingham, Massachusetts
  • 3Department of Nephrology, West China Hospital of Sichuan University, Chengdu, China
JAMA Intern Med. 2013;173(19):1821-1827. doi:10.1001/jamainternmed.2013.9774
Abstract

Importance  Rhabdomyolysis ranges in severity from asymptomatic elevations in creatine phosphokinase levels to a life-threatening disorder characterized by severe acute kidney injury requiring hemodialysis or continuous renal replacement therapy (RRT).

Objective  To develop a risk prediction tool to identify patients at greatest risk of RRT or in-hospital mortality.

Design, Setting, and Participants  Retrospective cohort study of 2371 patients admitted between January 1, 2000, and March 31, 2011, to 2 large teaching hospitals in Boston, Massachusetts, with creatine phosphokinase levels in excess of 5000 U/L within 3 days of admission. The derivation cohort consisted of 1397 patients from Massachusetts General Hospital, and the validation cohort comprised 974 patients from Brigham and Women’s Hospital.

Main Outcomes and Measures  The composite of RRT or in-hospital mortality.

Results  The causes and outcomes of rhabdomyolysis were similar between the derivation and validation cohorts. In total, the composite outcome occurred in 19.0% of patients (8.0% required RRT and 14.1% died during hospitalization). The highest rates of the composite outcome were from compartment syndrome (41.2%), sepsis (39.3%), and following cardiac arrest (58.5%). The lowest rates were from myositis (1.7%), exercise (3.2%), and seizures (6.0%). The independent predictors of the composite outcome were age, female sex, cause of rhabdomyolysis, and values of initial creatinine, creatine phosphokinase, phosphate, calcium, and bicarbonate. We developed a risk-prediction score from these variables in the derivation cohort and subsequently applied it in the validation cohort. The C statistic for the prediction model was 0.82 (95% CI, 0.80-0.85) in the derivation cohort and 0.83 (0.80-0.86) in the validation cohort. The Hosmer-Lemeshow P values were .14 and .28, respectively. In the validation cohort, among the patients with the lowest risk score (<5), 2.3% died or needed RRT. Among the patients with the highest risk score (>10), 61.2% died or needed RRT.

Conclusions and Relevance  Outcomes from rhabdomyolysis vary widely depending on the clinical context. The risk of RRT or in-hospital mortality in patients with rhabdomyolysis can be estimated using commonly available demographic, clinical, and laboratory variables on admission.

Rhabdomyolysis is characterized by muscle injury leading to the release of intracellular muscle contents into the systemic circulation. Because muscle injury from any cause can lead to rhabdomyolysis, the causes are numerous and include trauma or muscle compression as well as nontraumatic causes.1 The outcomes following rhabdomyolysis are similarly variable, ranging from asymptomatic elevations of creatine phosphokinase (CPK) concentration to life-threatening electrolyte abnormalities and acute kidney injury (AKI) requiring hemodialysis or continuous renal replacement therapy (RRT). Acute kidney injury is a feared and common complication of rhabdomyolysis, occurring in 13% to 50% of patients,2,3 with reported mortality rates as high as 59% in critically ill patients.4 Currently, clinicians lack tools to predict adverse outcomes in patients with rhabdomyolysis. The degree of CPK elevation is often used clinically as a marker of disease severity but has been reported to have a weak correlation with risk.1

The ability to identify patients early in the course of rhabdomyolysis who are likely to have adverse clinical outcomes would be useful for risk stratification in the emergency department, early institution of aggressive prophylactic measures, and communication with patients and families about prognosis. In this study, we used clinical and laboratory data from 2 large hospitals to derive and validate a risk prediction equation to estimate the chance of RRT or death in patients with rhabdomyolysis.

Methods
Data Collection

We obtained data from 2 teaching hospitals in the northeastern United States (Brigham and Women’s Hospital [BWH] and Massachusetts General Hospital [MGH]) on patients admitted between January 1, 2000, and March 31, 2011, through the Partners Healthcare System Research Patient Data Registry, a centralized clinical data warehouse designed for research and quality improvement purposes that has been accessed for clinical studies.5,6 We obtained information on patient demographics (age, sex, and race), length of stay, vital status at hospital discharge, billing codes (International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM]; Current Procedural Terminology [CPT]; and diagnosis-related group [DRG]), electronic discharge summaries, and inpatient laboratory values, including CPK, creatinine, phosphate, calcium, potassium, albumin, bicarbonate, white blood cell count, hemoglobin, and platelet count. Approval for this study was granted by the institutional review board at Partners, and the need for informed consent was waived.

Study Population

The inclusion criteria were age older than 18 years and CPK levels in excess of 5000 U/L within 72 hours of admission. We excluded patients with preexisting end-stage renal disease and those transferred from an outside facility who were receiving RRT for at least 24 hours. We also excluded 539 patients with CPK elevation considered due to acute myocardial infarction by reviewing ICD-9-CM codes (410.x), DRG codes, and medical record review of selected cases. For those patients with multiple admissions during the study period, we included the first admission only.

Ascertainment of Clinical Characteristics and Outcomes

We determined the most likely cause of rhabdomyolysis by having 1 of us (G.M.M.) review the DRG code assigned at discharge, ICD-9-CM codes, CPT codes, and electronic discharge summaries when administrative data were not informative; patients were then further classified as “medical” or “surgical.” We confirmed in-hospital mortality and dates of death obtained by administrative data by reviewing electronic death notes and/or discharge summaries from the electronic medical record. Acute kidney injury was defined and staged according to the definition by the Kidney Disease: Improving Global Outcomes group using creatinine values but not urine output because accurate urine output data were not available.7 For the baseline creatinine, we used the preadmission creatinine value or, if not available, the lowest creatinine measured during hospitalization. We confirmed every case of RRT during hospitalization by reviewing the electronic medical records of all patients with appropriate diagnostic and procedure codes (ICD-9-CM codes 39.95 and 54.98 and CPT codes 90935, 90937, 90945, and 909475).

Statistical Analysis

Categorical covariates were described by frequency distribution, while continuous covariates were expressed in terms of their mean (SD) or median and interquartile range as appropriate. Unadjusted associations between the covariates and the primary outcome were evaluated using χ2 tests for categorical data, while for continuous data, the t test was used for normally distributed variables and the Kruskal-Wallis test for nonparametric data. Unadjusted restricted cubic spline analysis was performed to explore the reported nonlinear relationship between initial CPK and risk of the composite outcome. Adjusted odds ratios were estimated by multivariable logistic regression.

We used the MGH cohort to derive a risk prediction score and then externally validated the score in the BWH cohort. We chose MGH as the derivation cohort because complete data were available in 98.9% of patients (16 were missing calcium and/or phosphate values) compared with 69.3% of patients in the BWH cohort, due primarily to missing phosphate values in the latter (299 [30.7%] with missing phosphate values, of whom 4 were also missing calcium values). Phosphate values were more likely to be missing in the BWH cohort than in the MGH cohort because of differences in ordering laboratory tests between the 2 hospitals. At BWH, in contrast to MGH, a standard chemistry panel includes calcium but not phosphate, whereas at MGH, calcium and phosphate are typically ordered together with the standard chemistry panel. We excluded those with missing values from the derivation cohort and assumed normal values for missing laboratory values in the validation cohort. We first selected clinically plausible variables that were potentially associated with the composite outcome of RRT or in-hospital mortality. Those variables associated with the composite outcome in bivariable analyses were then subjected to a backward stepwise logistic regression procedure based on a Wald χ2P value of more than .05. The final variables included in the model were age; sex; quintiles of initial phosphate, calcium, creatinine, and bicarbonate; CPK levels in excess of 40 000; and cause of rhabdomyolysis (seizure, syncope, statins, myositis, or exercise vs other causes). We selected the first available result for each laboratory variable for analysis. For each variable, quintiles with β coefficients that were not significantly different were grouped for the final logistic regression model. We then divided the β coefficients for each category by the smallest value of a β coefficient in the model to allocate an integer or half integer score for each variable. We generated a score for each individual based on the sum of the scores for every variable; this score was incorporated into a final model. We assessed the model’s discrimination using the C statistic and calibration with the Hosmer-Lemeshow (H-L) goodness-of-fit test using deciles. Confidence intervals (95%) were calculated using a nonparametric bootstrap method. We validated the final multivariable logistic regression model in the BWH cohort. Because of the large number of missing phosphate values in the validation cohort, we performed a sensitivity analysis to determine whether the model performed equally well in patients who did and did not have complete data.

Results

Between January 1, 2000, and March 31, 2011, we identified 3501 hospitalizations of adults with a CPK level in excess of 5000 U/L. After excluding 58 patients with end-stage renal disease or RRT on transfer from an outside institution, 74 multiple admissions, 459 with CPK elevation more than 3 days after admission, and 539 admitted with an acute myocardial infraction, the final study population included 2371 patients (1397 from MGH and 974 from BWH). Mean (SD) age was 50.7 (19.2) years, and 73.8% were male (Table 1). Review of administrative codes and electronic discharge summaries identified medical conditions as the clinical setting for rhabdomyolysis in 47.1% of patients compared with 52.9% for surgical settings. Quiz Ref IDThe clinical conditions most frequently associated with rhabdomyolysis were trauma (26.3%), immobilization (18.1%), sepsis (9.9%), and vascular and cardiac operations (8.1% and 5.9%, respectively). Other causes are shown in Table 1.

Outcomes Associated With Rhabdomyolysis

Quiz Ref IDAmong patients with rhabdomyolysis in the 2 cohorts, 47.7% developed AKI as described by the Kidney Disease: Improving Global Outcomes consensus definition.7 The composite outcome occurred in 19.0% of patients (8.0% required RRT and 14.1% died during hospitalization). In-hospital mortality was higher in patients with AKI (22.5% vs 7.1%, P < .001) and substantially higher in patients requiring RRT (40.0% vs 11.9%, P < .001). Outcomes were comparable between MGH and BWH (Table 1). Quiz Ref IDThe clinical conditions with the highest rates of the composite outcome of RRT or in-hospital mortality were compartment syndrome (41.2%), sepsis (39.3%), neuroleptic malignant syndrome (36.4%), abdominal or thoracic surgery (27.4%), and following cardiac arrest (58.5%). The lowest rates of the primary outcome were from myositis (1.7%), exercise (3.2%), and seizures (6.0%) (Figure 1). When stratified by cause of rhabdomyolysis, the rates of the primary outcome were similar in the MGH and BWH cohorts (Table 1).

Degree of CPK Elevation

In the 2 cohorts, we found no evidence of a linear association between CPK and risk of the composite outcome. In unadjusted logistic regression models with restricted cubic splines, we found increased risk of the composite outcome only with CPK levels in excess of 40 000 U/L. The C statistics for logistic regression models containing initial and peak CPK were only 0.52 and 0.61, respectively.

Derivation and Validation of Prediction Rule

Using MGH as the derivation cohort, we constructed a multivariable logistic regression model with RRT or in-hospital mortality as the outcome. Significant predictors of the primary outcome included age, female sex, cause of rhabdomyolysis, initial creatinine, CPK, phosphate, calcium, and bicarbonate (Table 2). The C statistic for the model was 0.82 (95% CI, 0.79-0.85). The model was well calibrated according to the H-L test (P = .07). A risk score was generated for each individual using the β coefficients from the logistic regression model. The parameters for the risk score are shown in Table 3. The mean (SD) risk score was 6.4 (3.2) (range, 0-17.5). A low risk score predicted a favorable outcome: a score of less than 5 identified patients with a 3% risk of the primary outcome, while a score of more than 10 was associated with a risk of 59.2%. Using 5 as the cutoff, the negative predictive value for the primary outcome was 97.0%, while the positive predictive value was 29.6%. In a logistic regression model using the score as a predictor, the C statistic was 0.82 (95% CI, 0.80-0.85) and the H-L P value was .14. Each 1-point increase in the risk score was associated with an adjusted odds ratio of 1.49 (95% CI, 1.42-1.57) for the primary outcome.

We performed external validation of the model in the BWH cohort. The C statistic for the model including the risk score was 0.83 (95% CI, 0.80-0.86), and the H-L P value was .28. Figure 2 shows the observed probabilities in the derivation and validation cohorts. A score of less than 5 was associated with a risk of the primary outcome of 2.3%, while a score of more than 10 was associated with a risk of 61.2%. The negative predictive value for a score of less than 5 in the validation cohort was 97.7%, while the positive predictive value was 27.2%. In the validation cohort, 30.1% of patients were missing phosphate values. The risk of the primary outcome was lower in patients with missing values (8.7% vs 20.5%, P < .001), supporting the imputation of normal values in those with missing phosphate values. In addition, in sensitivity analyses, we found similar calibration and discrimination when restricting the validation cohort to those with nonmissing phosphate values (area under the curve [AUC] = 0.81 [0.77-0.85] and H-L P = .73). In the validation cohort, the C statistics for the prediction of AKI (stage I and higher), AKI stage II and higher, AKI stage III, in-hospital mortality, and RRT were 0.68 (95% CI, 0.65-0.72), 0.75 (0.72-0.79), 0.80 (0.76-0.83), 0.80 (0.77-0.84), and 0.83 (0.78-0.88), respectively.

Discussion

In this study involving more than 2000 patients admitted with rhabdomyolysis during a 10-year period to 2 teaching hospitals, we have derived and externally validated a simple risk prediction score to estimate the risk of RRT or in-hospital mortality. The risk score has several important attributes for clinical utility. It is easily calculated with readily available clinical, demographic, and laboratory values; has good discrimination (ie, accuracy) and calibration (ie, accuracy across a range of predicted probabilities); and provides estimates of “hard” clinical end points of importance to clinicians caring for patients with rhabdomyolysis. Compared with other risk prediction scores in nephrology and critical care—for example, prediction of the risk of RRT after cardiac surgery (C statistic, 0.78-0.81),8 prediction of contrast nephropathy following cardiac catheterization (C statistic, 0.67),9 and prediction of mortality using the admission Sequential Organ Failure Assessment score in the intensive care unit (C statistic, 0.67-0.90)1012—the rhabdomyolysis risk score has comparable or better discrimination. We chose to develop a tool to predict the composite outcome of RRT or in-hospital mortality because of considerations of clinical relevance: a score that predicted just mortality would not provide information on RRT, an important, common, costly, and morbid complication of rhabdomyolysis, and a score that predicted just RRT would be limited since death may be a competing outcome. In post hoc analyses, the risk score had reasonable discrimination for the 2 individual components of the composite outcome (AUC = 0.83 and H-L P = .36 for RRT; AUC = 0.80 and H-L P < .001 for in-hospital mortality).

A key finding from our analyses relates to the nature of the relationship between CPK levels and risk. Previous smaller studies of rhabdomyolysis have shown that CPK has a weak relationship with the risk of RRT.2,13,14 However, other groups have documented large rises in CPK after vigorous exercise with no deleterious consequences,15,16 and there is evidence of a genetic component to the variability in CPK levels following injury.1719Quiz Ref IDWe found an association between elevated initial CPK levels in excess of 40 000 U/L and the risk of RRT or in-hospital mortality, but admission CPK levels alone were not sufficiently predictive to enable clinical decision making. The incorporation of cause of rhabdomyolysis provides important additional prognostic information, as suggested by previous smaller studies.2,20 The risk of the composite outcome in the overall cohort with low-risk causes (seizures, syncope, exercise, statins, or myositis) was only 5.1% compared with 21.1% in those with other causes. The components of the risk score also include age, sex, and clinical laboratory values. All laboratory variables are plausible as risk factors: high creatinine reflects AKI or CKD, which are both strong and independent risk factors for poor outcomes; high phosphate values may signify greater severity of muscle injury and impaired renal function; and low calcium results from deposition within damaged skeletal muscle and may be a marker of increased muscle injury.21

The risk score may be particularly useful in the emergency department to evaluate and triage patients with exertional rhabdomyolysis. For example, a 30-year-old female patient with myalgia following exercise with a CPK level of 20 000 U/L but normal laboratory values for calcium, phosphate, creatinine, and bicarbonate has a risk score of 1, corresponding to less than a 1% risk of RRT or in-hospital mortality. Knowledge of the predicted risk of adverse outcomes may lead clinicians to choose intravenous fluid administration in the emergency department followed by discharge with plans for repeated outpatient laboratories rather than inpatient hospitalization for observation. The safety of this approach and clinical implementation of the risk score in this manner would require additional study. The risk score has particular utility in identifying low-risk patients. Although the documentation of a high risk score may not directly influence treatment decisions, it may be useful in conveying prognosis and expectations to the care team, patient, and family members.

Clinical trials of therapeutic agents for rhabdomyolysis may also be aided by the use of a risk score as inclusion and exclusion criteria. Low-risk patients, in whom the event rate may be too low to warrant inclusion, and high-risk patients, for whom interventions may not carry clinical benefit due to severity of illness, may be excluded to optimize study design.

This study has a number of important strengths. To our knowledge, it is the largest study to date of rhabdomyolysis and, more important, has identified patients on the basis of CPK levels rather than administrative codes. Only patients with moderate to severe biochemical evidence of rhabdomyolysis were included, thus removing some of the ambiguity in previous studies in which a lower CPK cutoff for inclusion was used. The study drew data from 2 large university teaching hospitals, allowing external validation of the risk prediction model. However, we recognize some limitations. The study was retrospective in nature; the cause of rhabdomyolysis was ascertained by medical record review and administrative codes and may not have been accurate in all cases. Quiz Ref IDThe timing of the insult was not certain in many cases, and patients came to the hospitals at varying intervals following injury. There was no information on urine output or the treatments that were used or their relative effectiveness in preventing the primary outcome. Creatine phosphokinase was not routinely checked in all patients seeking treatment at the hospital, and therefore a number of patients may have developed rhabdomyolysis without being diagnosed. Finally, because this study includes patients from 2 large tertiary teaching hospitals in a single city in the northeastern United States, it may not be generalizable to other non–tertiary care hospitals in different geographical locations. Further validation studies may be informative.

In summary, we have described the incidence, causes, and outcomes of rhabdomyolysis in a large cohort of patients seeking treatment at 2 large university hospitals during a 10-year period. From these data, we derived and validated an easy-to-use risk score based on readily available parameters that can aid in estimating the probability of RRT or in-hospital mortality in patients with rhabdomyolysis.

Back to top
Article Information

Accepted for Publication: June 9, 2013.

Corresponding Author: Gearoid M. McMahon, MB, BCh, Renal Division, Brigham and Women’s Hospital, 75 Francis St, Boston, MA 02115 (gearoidmm@gmail.com).

Published Online: September 2, 2013. doi:10.1001/jamainternmed.2013.9774.

Author Contributions: Dr McMahon had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: McMahon, Waikar.

Acquisition of data: All authors.

Analysis and interpretation of data: All authors.

Drafting of the manuscript: McMahon.

Critical revision of the manuscript for important intellectual content: Zeng,Waikar.

Statistical analysis: McMahon, Zeng.

Study supervision: Waikar.

Conflict of Interest Disclosures: Dr Waikar reported serving as a consultant to CVS Caremark, BioTrends Research Group, Harvard Clinical Research Institute, and Takeda; providing expert testimony for GE Healthcare, Northstar Rx, and Salix; and receiving grants from the National Institute of Diabetes and Digestive Kidney Diseases, Otsuka, Merck, Genzyme, and Satellite Healthcare.

Funding/Support: Dr Zeng was supported by the China Scholarship Council. Dr Waikar is supported by grants DK093574 and DK085660.

Previous Presentation: The results of this study were presented in part as a poster at the American Society of Nephrology Annual Meeting; November 3, 2012; San Diego, CA.

Additional Contributions: Shawn Murphy, MD, PhD, Henry Chueh, MD, MS, and the Partners Healthcare System Research Patient Data Registry group facilitated the use of their database.

References
1.
Bosch  X, Poch  E, Grau  JM.  Rhabdomyolysis and acute kidney injury. N Engl J Med. 2009;361(1):62-72.
PubMedArticle
2.
Melli  G, Chaudhry  V, Cornblath  DR.  Rhabdomyolysis: an evaluation of 475 hospitalized patients. Medicine (Baltimore). 2005;84(6):377-385.
PubMedArticle
3.
Delaney  KA, Givens  ML, Vohra  RB.  Use of RIFLE criteria to predict the severity and prognosis of acute kidney injury in emergency department patients with rhabdomyolysis. J Emerg Med. 2012;42(5):521-528.
PubMedArticle
4.
de Meijer  AR, Fikkers  BG, de Keijzer  MH, van Engelen  BG, Drenth  JP.  Serum creatine kinase as predictor of clinical course in rhabdomyolysis: a 5-year intensive care survey. Intensive Care Med. 2003;29(7):1121-1125.
PubMedArticle
5.
Waikar  SS, Wald  R, Chertow  GM,  et al.  Validity of International Classification of Diseases, Ninth Revision, Clinical Modification codes for acute renal failure. J Am Soc Nephrol. 2006;17(6):1688-1694.
PubMedArticle
6.
Rhee  CM, Bhan  I, Alexander  EK, Brunelli  SM.  Association between iodinated contrast media exposure and incident hyperthyroidism and hypothyroidism. Arch Intern Med. 2012;172(2):153-159.
PubMedArticle
7.
Kidney Disease:Improving Global Outcomes (KDIGO) Acute Kidney Injury Work Group.  KDIGO clinical practice guideline for acute kidney injury. Kidney Int. 2012;2(suppl):1-138.Article
8.
Wijeysundera  DN, Karkouti  K, Dupuis  JY,  et al.  Derivation and validation of a simplified predictive index for renal replacement therapy after cardiac surgery. JAMA. 2007;297(16):1801-1809.
PubMedArticle
9.
Mehran  R, Aymong  ED, Nikolsky  E,  et al.  A simple risk score for prediction of contrast-induced nephropathy after percutaneous coronary intervention: development and initial validation. J Am Coll Cardiol. 2004;44(7):1393-1399.
PubMed
10.
Minne  L, Abu-Hanna  A, de Jonge  E.  Evaluation of SOFA-based models for predicting mortality in the ICU: a systematic review. Crit Care. 2008;12(6):R161. doi:10.1186/cc7160.
PubMedArticle
11.
Ho  KM, Lee  KY, Williams  T, Finn  J, Knuiman  M, Webb  SA.  Comparison of Acute Physiology and Chronic Health Evaluation (APACHE) II score with organ failure scores to predict hospital mortality. Anaesthesia. 2007;62(5):466-473.
PubMedArticle
12.
Peres Bota  D, Melot  C, Lopes Ferreira  F, Nguyen Ba  V, Vincent  JL.  The Multiple Organ Dysfunction Score (MODS) versus the Sequential Organ Failure Assessment (SOFA) score in outcome prediction. Intensive Care Med. 2002;28(11):1619-1624.
PubMedArticle
13.
Ward  MM.  Factors predictive of acute renal failure in rhabdomyolysis. Arch Intern Med. 1988;148(7):1553-1557.
PubMedArticle
14.
Gabow  PA, Kaehny  WD, Kelleher  SP.  The spectrum of rhabdomyolysis. Medicine (Baltimore). 1982;61(3):141-152.
PubMed
15.
Hoffman  MD, Ingwerson  JL, Rogers  IR, Hew-Butler  T, Stuempfle  KJ.  Increasing creatine kinase concentrations at the 161-km Western States Endurance Run. Wilderness Environ Med. 2012;23(1):56-60.
PubMedArticle
16.
Clarkson  PM, Kearns  AK, Rouzier  P, Rubin  R, Thompson  PD.  Serum creatine kinase levels and renal function measures in exertional muscle damage. Med Sci Sports Exerc. 2006;38(4):623-627.
PubMedArticle
17.
Clarkson  PM, Hoffman  EP, Zambraski  E,  et al.  ACTN3 and MLCK genotype associations with exertional muscle damage. J Appl Physiol. 2005;99(2):564-569.
PubMedArticle
18.
Hubal  MJ, Devaney  JM, Hoffman  EP,  et al.  CCL2 and CCR2 polymorphisms are associated with markers of exercise-induced skeletal muscle damage. J Appl Physiol. 2010;108(6):1651-1658.
PubMedArticle
19.
Heled  Y, Bloom  MS, Wu  TJ, Stephens  Q, Deuster  PA.  CK-MM and ACE genotypes and physiological prediction of the creatine kinase response to exercise. J Appl Physiol. 2007;103(2):504-510.
PubMedArticle
20.
Woodrow  G, Brownjohn  AM, Turney  JH.  The clinical and biochemical features of acute renal failure due to rhabdomyolysis. Ren Fail. 1995;17(4):467-474.
PubMedArticle
21.
Akmal  M, Bishop  JE, Telfer  N, Norman  AW, Massry  SG.  Hypocalcemia and hypercalcemia in patients with rhabdomyolysis with and without acute renal failure. J Clin Endocrinol Metab. 1986;63(1):137-142.
PubMedArticle
×