Trends in Pediatric Rhabdomyolysis and Associated Renal Failure: A 10-Year Population-Based Study Free

This was a retrospective analysis of the Healthcare Cost and Utilization Project Kids’ Inpatient Database (KID), a publicly available all-payer database that estimates pediatric hospitalizations (age <21 years) across the United States, estimating roughly 7 million inpatient stays (when weighted). In 2006, the KID database included over 3700 nonrehabilitation hospitals in 38 states; this increased to 4200 hospitals in 47 states in 2016. All hospitalized patients in 2006, 2009, 2012, and 2016 with a primary diagnosis code for rhabdomyolysis (International Classification of Diseases, Ninth and Tenth Revision, Clinical Modification [ICD], ICD-9 code 728.88 [2006–2012]; ICD-10 code M62.82 [2016]) were included. To exclude patients with other diagnoses that may independently cause renal injury, like septic shock or medication-induced, we chose not to include patients who only had a secondary diagnosis of rhabdomyolysis. US Census Bureau yearly population estimates by age group and geographic region were used to determine incidence and rates over time.¹⁵  This study, utilizing deidentified data, was not considered human subjects research by the official institutional review board.

Demographic characteristics included US census-designated age groups (infant to young child [less than 5 years], older child [5 to 9 years], early adolescence [10 to14 years], late adolescence [15 to 20 years]), gender (male or female), race (non-Hispanic White, non-Hispanic Black, Hispanic, Asian or Pacific Islander, Native American, or other), and payor (Medicare, Medicaid, private insurance, self-pay, or other). Race was included, as a social construct, to evaluate for diverse sociodemographic representation of groups that may impact generalizability of data or suggest greater burden of disease within various groups. The number of complex chronic conditions (CCCs) were counted (0, 1, 2+CCCs) using a prior described methodology, and patient disease severity was quantified using the All Patient Refined Diagnosis Related Groups classification system with relative weights from Hospitalization Resource Intensity Scores for Kids.^16,17  Hospital length of stay (days) represented resource utilization. Finally, we grouped hospitals by United States census-designated geographic region (Northeast, Midwest, South, West), and admission month.¹⁸ 

The primary outcome was incidence of hospitalizations for pediatric rhabdomyolysis and associated renal morbidity or death. Overall, incidence was calculated using the number of pediatric hospitalizations for rhabdomyolysis compared with the general population of children reported by the United States census from 2006 to 2016. To generate national estimates, we used the discharge-level weight available in the database to estimate patients treated at community hospitals (excluding rehabilitation and long-term acute care facilities) in the United States. These rates were analyzed for trends over time, by geographic region, and by month of the year. To estimate the disease burden among hospitalized children, the number of hospitalizations for rhabdomyolysis was also compared with the overall number of children hospitalized over the same period. Renal morbidity was defined as acute renal insufficiency and/or acute renal failure (including need for RRT). ICD-9 and ICD-10 diagnosis codes were used to identify patients with concurrent diagnoses of rhabdomyolysis and acute renal insufficiency, injury and/or failure; ICD-9 and ICD-10 procedure codes for dialysis or hemodialysis were used to identify patients who received RRT (Table 3). Patients with the ICD-10 diagnosis code for “disorder of kidney and ureter, unspecified” (N28.9) were not included, as this was less specific for acute renal failure. As a secondary outcome, patient characteristics were analyzed for independent predictors of renal morbidity diagnoses.

All analyses made use of the sampling frames and weights available in the Healthcare Cost and Utilization Project KID database. Bivariate analyses describing renal failure by baseline demographics was conducted using χ² and Wilcoxon Rank Sum tests as appropriate. Because of the non-normal distribution of continuous data, geometric means and standard deviation (SD) were used. Rates of rhabdomyolysis were analyzed over time with a generalized linear model with a Poisson distribution and log link using a fixed effect for year. The log transformation of the general population of children for each year (derived from census data) was used as an offset in the model. Comparisons across geographic regions were performed by adding an interaction term to this model between census region and year. Multivariable logistic regression model with backward selection was used to estimate adjusted odds ratio. Variables that were statistically significant in the bivariate analysis were considered for the multivariable logistic regression. All variables were evaluated in the full model and retained at 0.05 by backward elimination. Analysis was performed with SAS Enterprise Guide v7.1 (SAS Institute, Cary, NC) at a significant level of 0.05. Figures were produced in GraphPad Prism v.9 (GraphPad Software, San Diego, CA).

Among the 8599 included patients with rhabdomyolysis diagnosis, the majority were male (80.9%), in late adolescence or early adulthood (15 to 20 years; 75.1%), and predominantly non-Hispanic White and non-Hispanic Black (41.7% and 36.5%, respectively). Few patients had a complex chronic condition (17.8%), and the majority had private insurance (51.2%). The geometric mean length of stay was 3.46 days (SD 0.06). Most patients presented to hospitals in the Southern region (43.7%) (Table 1).

In 2006, the rate of pediatric hospitalizations for rhabdomyolysis was 0.8 per 100 000 children, increasing to 2.8 per 100 000 children by 2016, a 218% increase (Fig 1). In 2006 and 2009, there were no significant differences in the rates of rhabdomyolysis across the 4 geographic regions. In 2012, the South and Midwest had significantly higher rates than the West (29% difference, P = .009 and 33% difference, P = .021, respectively). In 2016, rates in the South were 47% higher than the West (P < .001) and 36% higher than the Midwest (P < .001), and rates in the Northeast were 27% higher than the Midwest (P = .033) and 47% higher than the West (P < .001). There was notable seasonal variation in the incidence of rhabdomyolysis with peaks in March and August (Fig 2). The peak in March was dominated by the Northeast and Midwest, whereas the peak in August was dominated by the South and Northeast regions.

Diagnosis of acute renal failure occurred at a rate of 85.2 per 1000 hospitalized patients with rhabdomyolysis (95% confidence interval [CI] 79.3–91.1). Over time, rates increased from 6.6% in 2006 to 10.3% in 2016. Males had significantly higher rates of renal failure compared with females (9.8% versus 3.5%, respectively); similarly, the older patients compared with the younger cohort (<15 years old) had higher rates of renal failure (10.4% versus 2.7%, respectively). With concurrent renal failure diagnosis, non-Hispanic Black patients surpassed non-Hispanic White patients to represent the dominant group (41.6% versus 36.5%, respectively). Similarly, a majority of patients diagnosed with renal failure (57.2% total) were hospitalized during the summer and early fall months (June–September). The largest peak of renal failure diagnoses was in August (21.1%), with almost twice as many cases than other months. Patients with 2 or more CCCs also had a higher proportion of concurrent renal failure diagnoses than patients without any CCCs (17% versus 7.6%, respectively).

In the overall cohort of hospitalized patients with rhabdomyolysis, only 35 required reported RRT (4.1 per 1000 rhabdomyolysis patients [95% CI 2.7–5.4]); all of these patients had an associated renal failure diagnosis. Among the 733 children with rhabdomyolysis and acute renal failure diagnoses, 4.8% required RRT. Most of these patients were male and older adolescents or young adults (77.1% and 85.7%, respectively), consistent with the rates in the overall rhabdomyolysis cohort and in those without a renal failure diagnosis. Those requiring RRT had significantly higher disease severity overall (geometric mean 5.37 [SD 0.32] versus 1.43 [SD 0.02], P < .001), longer hospital length of stay (geometric mean 18.5 days [SD 2.14] versus 3.46 days [SD 0.06]), and at least 1 CCC. Though diagnoses of renal failure have increased over time, associated RRT has decreased from 12 cases in 2006 to none reported in 2016. There was no significant pattern of seasonality or regional predominance in the group of patients with RRT, and none were reported to have died.

Overall, 3 patients died (0.4 deaths per 1000 rhabdomyolysis patients [95% CI 0.0, 0.7]; none of these patients reported RRT while hospitalized, though 1 had a concurrent renal failure diagnosis. Compared with the patients who required RRT and survived, the patients who died had higher disease severity (geometric mean 8.7 [SD 0.89] versus 5.4 [SD 0.32], P < .001). Of the 3 patients who died, associated diagnoses included seizures, disseminated intravascular coagulation, cardiac arrest, shock, acute respiratory distress or failure, and nonspecific viral infection.

In multivariable modeling, the characteristics most significantly associated with renal failure diagnosis included male sex (odds ratio [OR] 3.07; 95% CI 2.12–4.98) and older patients compared with those less than 15 years old (OR 3.55; 95% CI 2.53–4.98). Presence of CCCs was also associated with diagnosis of renal failure, with 2 or more CCCs nearly double the likelihood of only 1 CCC (OR 3.97 [95% CI 2.56–6.16] versus OR 1.84 [95% CI 1.41–2.39]). Finally, patients presenting to hospitals in the Northeast region were less likely to have a diagnosis of renal failure, compared with those presenting to hospitals in the Southern region (OR 0.53; 95% CI 0.38–0.73) (Table 3).

Yearly hospitalizations for rhabdomyolysis in children are relatively infrequent but have increased from 0.8 to 2.8 per 100 000 children over the past decade. The majority of rhabdomyolysis hospitalizations occur in the Southern region with higher peaks in March and August. Though renal morbidity was diagnosed in 8.5% of children requiring hospitalization for rhabdomyolysis, very few of these patients required RRT (0.41%), and death was rare (0.03%). Demographic characteristics associated with concurrent diagnosis of renal failure included male sex, adolescent, or young adult age (15 to 20 years old), non-Hispanic Black race, Southern region, and presence of 1 or more CCC.

Prior studies were mostly single or dual institution case series and unable to estimate incidence, rates over time, or regional variability. By using a large national database, we were able to establish the number of hospitalizations for rhabdomyolysis over time, noting geographic and temporal peaks; this seasonal variation remains consistent across all geographic regions. Overall, there are increased rates during March and August; prior studies were unable to observe any seasonal variability.¹  Additionally, the majority of patients with concurrent diagnosis of renal failure presented in the summer months (57.2% in June to September). We were unable to make conclusions about the rates of specific etiologies of rhabdomyolysis in this cohort of patients because of the limitations of the administrative database used; however, this seasonal pattern introduces the hypothesis of possible environmental contributors. Prior studies in children report exercise or exertion and viral myositis, particularly influenza, as the most common etiologies for pediatric rhabdomyolysis.^{1–3,7,8,11,12}  Our study’s reports of higher rates of rhabdomyolysis in males and adolescents and young adults is consistent with military literature on exertional rhabdomyolysis in active duty military.^19–21  Higher temperatures in the Southern region, paired with increased intensity of school athletic training outdoors in the late summer or early fall (August) and during spring break (March), may contribute to increased rates of heat-related illnesses, including exertional rhabdomyolysis.²²  However, heat-related illnesses, including dehydration, may independently cause elevations in serum creatinine. Additionally, there is persistence of influenza and other respiratory viruses through May, which may contribute to the increase in rhabdomyolysis seen in March, though it not fully consistent with peak winter respiratory viral season.^{1–3,10,11,13} 

Most pediatric patients hospitalized with rhabdomyolysis did not have CCCs, suggesting pediatric patients with rhabdomyolysis are generally healthy. Presence of CCCs did not increase risk for rhabdomyolysis hospitalization, but greater number of CCCs did confer greater likelihood of diagnosis of renal failure with rhabdomyolysis. Pediatric patients with rhabdomyolysis also have generally low disease severity, which suggests lower acuity of illness while hospitalized. Though rhabdomyolysis and associated renal failure diagnoses are increasing over time, need for renal replacement has decreased over time, suggesting either decreasing acuity of hospitalized patients versus early diagnosis and intervention preventing complications. Increased use of diagnostic testing may contribute to increased diagnoses without true increased disease incidence.

To our knowledge, this is the largest study to date to report incidence of renal morbidity in children hospitalized with rhabdomyolysis. We report a much lower rate of associated renal morbidity, at 8.5% out of 8599 hospitalized patients, compared with previously reported rates up to 50%.^{1–3,5,7–13,23}  Our reported rate is also lower than the 14% to 48% reported for rhabdomyolysis-associated renal failure in adults.^6,24–28  It is important to note that definitions of renal morbidity (including acute kidney injury, acute renal insufficiency, and acute renal failure) vary across these studies, thus direct comparison is challenging. Despite a clinically significant rate of renal failure, few patients require RRT (0.41%) and death was rare (0.03%), consistent with most prior studies, with the exception of Watanabe et al who reported rates of 33% for both renal failure and death (total 18 patients).^2,3,8,10,11  Further studies are needed to investigate whether this reported difference is because of the natural history of underlying etiology or because of regional differences in management.

Non-Hispanic Black patients, along with non-Hispanic White patients, made up the majority of pediatric patients hospitalized with rhabdomyolysis. A greater percentage of non-Hispanic Black children are hospitalized with rhabdomyolysis (32%) compared with the percentage of non-Hispanic Black children in the general population (13.5% in 2016).²⁹  Previous case reports have noted an association between Black race, exertional rhabdomyolysis, and sudden death with exertion.^30–33  Thus, presence of sickle cell trait was hypothesized as a risk factor, since there is a greater incidence of sickle cell trait in Black newborns.³⁴  A large scale military study of over 47 000 Black soldiers determined that sickle cell trait conferred greater adjusted risk of exertional rhabdomyolysis.³⁵  In a similar study using an administrative database, Black race was found to be an independent risk factor for rhabdomyolysis and subsequent acute kidney injury in trauma patients, despite only having <0.1% of Black patients reporting sickle cell trait. Sickle cell trait did not independently confer greater risk rhabdomyolysis or acute kidney injury, though it may not have been powered enough to show any difference. Our study shows the majority of those with renal failure identified as non-Hispanic Black, though they were not the predominant group in the overall cohort with rhabdomyolysis. Prior studies have shown worse outcomes for Black patients presenting with trauma or sepsis caused by systemic under-recognition and under-resuscitation in this group, rather than biologic differences. Similarly, when looking at these differences in outcomes for pediatric rhabdomyolysis, the possible contribution of systemic racism and inherent biases cannot be ignored; however, further investigation is needed, as we are unable to comment on etiology for rhabdomyolysis, severity at presentation, or progression of disease while hospitalized.

This study should be viewed with several limitations in mind. Prior studies, in both pediatric and adult populations, have used varying definitions of rhabdomyolysis and acute renal failure. By using administrative billing data, patients were identified as having rhabdomyolysis or renal failure based on discharge diagnosis codes assigned by individual clinicians, which may limit the consistency in the diagnostic definition across patients, institutions, and regions in this study. The KID itself is dependent on the completeness of data reported by each participating hospital; subsequently, there are missing values reported in both race and diagnosis coding. The decision to not include patients with a secondary diagnosis of rhabdomyolysis may have limited inclusion of higher acuity patients with greater rates of renal failure and longer hospital stays. From ICD-9 to ICD-10, the specific diagnosis code for acute renal insufficiency was replaced with a more generic code for “disorder of the kidney and ureter, unspecified,” which we chose not to include. Although this kept our scope narrow to higher acuity renal failure, it may have excluded lower acuity patients with mild renal insufficiency. Whereas there was a change from ICD-9 to ICD-10, the code definition and scope for rhabdomyolysis did not change, though provider billing practices might have changed. Additionally, we were unable to account for changes in diagnostic testing patterns or thresholds for diagnosis that may influence apparent trends in incidence, since the KID does not provide specific information about diagnostic testing or individual laboratory values.³⁶ 

Rhabdomyolysis is an uncommon disease process in children with infrequent rate of hospitalizations. However, among patients who are hospitalized, a significant portion of patients can have associated renal failure, though severe complications continue to be rare. Clinicians should have a high index of suspicion for rhabdomyolysis in children during peak months (March and August), especially in geographic areas with higher incidence (South and Northwest). Further studies are needed to correlate etiology with outcomes and investigate other risk factors for morbidity and mortality.

RRT

renal replacement therapy

HCUP

Healthcare Cost and Utilization Project

KID

Kids’ Inpatient Database

ICD

International Classification of Diseases, Clinical Modification

CCC

complex chronic conditions

APR-DRG

All Patient Refined Diagnosis Related Groups