Original Article
Health plan administrative databases can efficiently identify serious myopathy and rhabdomyolysis

https://doi.org/10.1016/j.jclinepi.2004.10.004Get rights and content

Abstract

Objective

We evaluated the positive predictive values (PPVs) of specific criteria based upon International Classification of Diseases, 9th revision (ICD-9-CM) codes documented in health plan administrative databases for identification of cases of serious myopathy and rhabdomyolysis.

Study design and setting

We conducted a retrospective study among patients enrolled in 11 geographically dispersed managed care organizations. Cohorts of new users of specific statins and fibrates were identified by selecting patients with an initial dispensing of the drug during the period 1 January 1998 to 30 June 2001. Potential cases of serious myopathy or rhabdomyolysis were identified using specific criteria based upon ICD-9-CM codes suggesting a muscle disorder or acute renal failure.

Results

A total of 194 hospitalizations meeting the criteria for chart review selection were identified among 206,732 new users of statins and 15,485 new users of fibrates. Overall, 31 cases of serious, clinically important myopathy or rhabdomyolysis (18%) were confirmed through chart review. Of these, 26 (84%) had a claim including codes for myoglobinuria (ICD-9-CM 791.3) or other disorders of muscle, ligament, and fascia (ICD-9-CM 728.89). A PPV of 74% (26 of 35 patients meeting criteria) was found for a composite definition that included (1) a primary or secondary discharge code for myoglobinuria, (2) a primary code for “other disorders of muscle,” or (3) a secondary code for “other disorders of muscle” accompanied by a claim for a CK test within 7 days of hospitalization or a discharge code for acute renal failure.

Conclusion

For rare adverse events such as serious myopathy or rhabdomyolysis, large population-based databases that include diagnosis and laboratory test claims data can facilitate epidemiologic research.

Introduction

Serious myopathy and rhabdomyolysis are muscle-related adverse events that have been reported for a number of drugs, most notably lipid-lowering drugs, including statins and fibrates [1], [2], [3], [4]. Clinically important myopathy and rhabdomyolysis are generally defined by muscle symptoms (e.g., pain, fatigue, or weakness) with creatine kinase (CK) elevations greater than 10 times the upper limit of normal (ULN) [1], [2], [3]. Rhabdomyolysis is often associated with myoglobinuria, myoglobinemia, and end organ damage (e.g., acute renal failure) and may be fatal.

Data on the incidence of rhabdomyolysis among users of lipid-lowering drugs are limited and are based largely on data from spontaneous reporting systems and clinical trials [1], [4], [5]. Administrative databases may be useful to identify potential adverse events, such as serious myopathy or rhabdomyolysis, for assessment of the incidence and risk factors for the disease. However, the International Classification of Diseases, 9th revision (ICD-9-CM) code for rhabdomyolysis is nonspecific, and a number of different muscle-related diagnoses, in addition to diagnoses related to end organ damage, may be suggestive of serious myopathy and rhabdomyolysis. Thus, assessment of the validity of diagnostic criteria for identification of rhabdomyolysis is warranted. We evaluated the positive predictive values (PPVs) of specific criteria based upon ICD-9-CM codes to assess the utility of using automated administrative databases to identify cases of serious myopathy and rhabdomyolysis.

Section snippets

Methods

We conducted a retrospective study among patients enrolled in 11 geographically dispersed managed care organizations (five in the Midwest, three in the Northeast, two in the Southeast, and one in the West) that included independent practice associations and staff and group practice models. Each of the organizations maintains computerized databases of pharmacy dispensings, inpatient and outpatient diagnoses and procedures, and enrollment and demographic data.

Institutional Review Board (IRB)

Results

Among 206,732 new statin users and 15,485 new fibrate users, 194 hospitalizations met the criteria for chart review. Of these hospitalizations, 174 charts (90%) were reviewed and abstracted. The mean age of patients at hospitalization was 64 years (range 32–86), and 56% were female. The gender and age distributions of the 20 patients whose records could not be reviewed were similar to those of patients whose records were reviewed.

Thirty-one cases of clinically important myopathy or

Discussion

Our results suggest that administrative databases may be useful for signal detection of drug exposures potentially associated with rhabdomyolysis. Given that the large majority of cases were identified with criteria that included a discharge diagnosis of myoglobinuria (criterion 1) or other disorder of the muscle (using criteria 2, 3, and 4) and that the overall PPV using only these criteria was 74%, investigators with limited resources might choose to restrict case identification using these

Acknowledgments

This work was supported by grants FD-U-001641, FD-U-001643, and FD-U-002067 from the US Food and Drug Administration. We are grateful to Rachel Kasper, Nicole Boudreau, Claire Canning, Kristi Paulsen, and Margaret Burgess for their technical support. The views expressed herein are those of the authors and not necessarily those of the Food and Drug Administration.

References (7)

  • P.D. Thompson et al.

    Statin-associated myopathy

    JAMA

    (2003)
  • R.L. Wormann

    Lipid-lowering agents and myopathy

    Curr Opin Rheumatol

    (2002)
  • C.M. Ballantyen et al.

    Risk for myopathy with statin therapy in high-risk patients

    Arch Intern Med

    (2003)
There are more references available in the full text version of this article.

Cited by (42)

  • Risk factors and future directions for preventing and diagnosing exertional rhabdomyolysis

    2021, Neuromuscular Disorders
    Citation Excerpt :

    These cases were the first to associate RM with acute kidney injury (AKI). Despite being a fairly well-described clinical phenomenon, RM only received an international disease classification code (ICD-9) in 2004 [5]. There are different causes for RM, one of which is due to physical exertion and known as exertional rhabdomyolysis (ERM).

  • Statin-associated side effects

    2016, Journal of the American College of Cardiology
    Citation Excerpt :

    Subjects were observed from 0.5 to 6.1 years, giving an extremely low yearly incidence of rhabdomyolysis (15). The incidence of rhabdomyolysis in clinical practice has been examined using national health records (20) and health insurance databases (21,22). An examination of health claims from 473,343 patients treated with lipid-lowering agents, of whom 86% received statin monotherapy, found 144 claims coded for rhabdomyolysis, of which 44 were confirmed by physician review (22).

  • Comparative Effectiveness of Statin Therapy in Chronic Kidney Disease and Acute Myocardial Infarction: A Retrospective Cohort Study

    2015, American Journal of Medicine
    Citation Excerpt :

    Diabetes was identified using an ICD-9-CM code-based approach modified from Nichols et al16 (1 inpatient stay or at least 2 outpatient visits coded with any of 250.x, 357.2, 366.41, 362.01-362.07). Myopathy was identified using ICD-9-CM codes that have a positive predictive value of 74%17 (inpatient stay coded with 791.3, or inpatient stay with primary discharge code 728.89, or secondary discharge code of 728.89 along with a creatine kinase test performed within 7 days of the hospitalization, or code 584. xx along with secondary discharge code of 728.89).

  • Simvastatin dose and risk of rhabdomyolysis: Nested case-control study based on national health and drug dispensing data

    2014, International Journal of Cardiology
    Citation Excerpt :

    These were the patients for whom NHIs were provided. As the specificity of this broad search strategy was inevitably low, we then applied further diagnostic criteria (Appendix 2) which were adapted from an ICD-9 based algorithm developed by others [21]. For patients who met these criteria, we then sought copies of hospital discharge letters, necropsy reports, and any other death records in order to validate the diagnosis of rhabdomyolysis.

View all citing articles on Scopus
View full text