Relationship of body mass index, serum creatine kinase, and acute kidney injury after severe trauma

Abstract

Objectives:

Body mass index (BMI) is associated with acute kidney injury (AKI) after trauma, but underlying mechanisms are unclear. BMI correlates with both excess adiposity and increased muscle mass. Since the latter could predispose to severe rhabdomyolysis after trauma, we hypothesized that the BMI-AKI association may be partially explained by a direct relationship of BMI with serum creatine kinase (CK).

Methods:

Prospective cohort study of 463 critically ill patients admitted to a Level I trauma center from 2005-2015 with injury severity score (ISS)>15 and serum CK measured in the first 7 days. We defined AKI by AKI Network creatinine criteria. We used simple linear regression to determine the association of BMI with peak CK, and multivariable logistic regression to adjust the BMI-AKI association for peak CK and confounders.

Results:

Median age was 43 years, 350 (76%) were male, 366 (79%) had blunt mechanism, and median ISS was 24. BMI was associated with peak CK (R² 0.05, p<0.001). AKI developed in 148 patients (32%), and median time to peak CK was 29 (IQR 15-56) hours after presentation. BMI was significantly associated with AKI in multivariable models adjusted for age, race, sex, diabetes, injury mechanism and severity, and red blood cell transfusions (OR 1.31 per 5 kg/m², 95% CI 1.09-1.58, p=0.004). Adding peak CK to the model partially attenuated the association of BMI with AKI (OR 1.26 per 5 kg/m², 95% CI 1.04-1.52, p=0.018), and peak CK was also associated with AKI (OR 1.19 per natural log, 95% CI 1.00-1.41, p=0.049). Peak CK remained associated with AKI when restricted to patients with values<5000 U/l (OR 1.31 per natural log, 95% CI 1.01-1.69, p=0.043).

Conclusion:

Serum CK correlated with BMI and partially attenuated the association of BMI with AKI after major trauma, suggesting that excess muscle injury may contribute to the BMI-AKI association.

Level of Evidence:

Epidemiologic study, Level III

BACKGROUND

Acute kidney injury (AKI), is a common complication of trauma that contributes to the >190,000 trauma-related deaths per year in the United States^1,2. Obesity, defined as body mass index (BMI) ≥30 kg/m², is associated with AKI after trauma as well as in post-surgical, acute respiratory distress syndrome, and general critical illness populations^3-6. While obesity is known to increase the risk of chronic kidney disease (CKD) via its contributions to insulin resistance, hypertension, and glomerular hyperfiltration, we and others previously showed that the association of BMI with post-trauma AKI was independent of diabetes, hypertension, and CKD as well as injury characteristics^7-10. The mechanisms underlying the BMI-AKI association, and whether they are therapeutically modifiable in the acute setting, remain unclear. While visceral adiposity is characterized by increased baseline circulating levels of nephrotoxic inflammatory mediators, quantitative estimates of both visceral and subcutaneous adiposity in trauma patients failed to demonstrate stronger associations with AKI than BMI¹¹. BMI itself is not a specific measure of adiposity as it is also dependent on muscle mass, particularly in young patients typical of the trauma population. Increased muscle mass, therefore, may plausibly play a role in the development of AKI after severe injury and contribute to the association of BMI with AKI.

Rhabdomyolysis, the destruction and breakdown of skeletal muscle, is one mechanism by which greater muscle mass could increase AKI risk. Rhabdomyoloysis is common after trauma and may result from direct muscle damage or ischemia related to vascular injury or systemic hypoperfusion^12-14. Intracellular myoglobin is released into the extra-cellular fluid and circulation, and may cause renal injury via renal vasoconstriction, oxidative endothelial damage, and formation of pigmented tubular casts^15-18. Since circulating myoglobin rapidly rises and falls within several hours of injury, rhabdomyolysis is usually clinically detected by elevation of serum creatine kinase (CK), which is also released by skeletal muscle and starts to rise 2-12 hours after injury, typically peaking between 24 and 72 hours^19,20. Though descriptions of the nephrotoxic effects of rhabdomyolysis are decades old, no study has examined whether rhabdomyolysis could account for the known association of BMI with AKI after trauma. Further, nearly all studies of rhabdomyolysis and AKI have been limited only to patients with elevation of peak serum creatine kinase to an arbitrary cut-off of >5,000 U/l, and most commonly focus on AKI requiring renal replacement therapy (RRT)^21,22. Though Byerly and colleagues recently showed that a broader range of peak CK levels may connote increased AKI risk, AKI was defined as a serum Cr ≥2 mg/dl during hospitalization rather than by the smaller, relative creatinine changes over specified time intervals recommended by AKI consensus definition statements^23,24. Thus, important questions remain regarding the contribution of rhabdomyolysis to the BMI-AKI association, the association of modest CK elevations with consensus-defined AKI, and whether early, pre-peak CK measurements can aid in quantifying AKI risk after trauma. Bridging this knowledge gap may help to determine whether a larger group of trauma patients may be targets for both established and novel AKI-preventive therapies after muscle injury, compared to current practice^25-27.

We therefore aimed to analyze the relationship between BMI, post-injury elevation in serum CK, and AKI in severely injured trauma patients. Secondarily, we sought to determine the associations of peak CK levels <5000 U/l and early CK levels measured within 24 hours of presentation with AKI. We hypothesized that BMI would correlate with CK levels, that a broad range of CK levels would be associated with AKI, and that CK elevation would contribute to the BMI-AKI association.

METHODS

Every patient who sustained a traumatic injury and was admitted to the surgical ICU of the University of Pennsylvania Level I Trauma Center from 2005 to 2015 was screened for enrollment in the prospective Penn Trauma Acute Organ Dysfunction Study (PETROS)^28,29. Patients were included if they were aged 14 years old or older and had an Injury Severity Score (ISS)≥ 16. Key exclusion criteria were severe isolated head injury, and death or discharge from the ICU within 24 hours of arrival to the emergency department. For the current study, we included PETROS participants who had at least one serum creatine kinase (CK) value measured with seven days of presentation, a time frame chosen to reflect rhabdomyolysis related to the initial injury while minimizing patient exclusions (Table S2). We also excluded those with end-stage kidney disease. This study was approved by the Institutional Review Board at the University of Pennsylvania utilizing a waiver of informed consent.

Demographic, medical history, injury characteristics and information regarding the patient’s treatment and hospital course were obtained from the medical record. BMI was calculated and categorized using standard World Health Organization definitions as follows: underweight (BMI <18.5 kg/m²), normal weight (BMI 18.5-24.9 kg/m²), overweight (BMI 25.0-29.9 kg/m²), and obese (BMI ≥30 kg/m²)³⁰. Height and weight were measured on admission. We used serum CK levels available in the medical record that were ordered by the managing team and tested in the clinical laboratory. CK was measured by using standard enzymatic rate methodology and spectrophotometry, with normal serum CK levels are 45-260 U/L based upon University of Pennsylvania Health System laboratory standards. We defined AKI during the first 6 days after presentation using the Acute Kidney Injury Network (AKIN) creatinine and dialysis criteria (Table S1)²³. Urine output criteria were not utilized due to lack of precision with urine output monitoring during the hospitalization. In order to minimize ascertainment bias, we did not use pre-hospital baseline serum creatinine values to define AKI in the few patients in whom they were available.

Differences in baseline characteristics by AKI status were analyzed using the Wilcoxon rank-sum and t-tests for continuous variables, and chi-squared and Fisher’s exact tests for categorical variables, as appropriate. We tested the unadjusted association of BMI with peak CK using Spearman’s rho and utilized a linear regression model to visually represent these data. We constructed multivariable logistic regression models to test the associations of both peak CK and BMI with AKI. We adjusted for known AKI risk factors, as well as covariates with an unadjusted association with AKI at p<0.20 that changed the odds ratio of the association of peak CK or BMI with AKI by ≥10%³¹. We used natural log-transformed CK values for these models given significant non-normality of raw CK levels. Models of the association of BMI with AKI were inspected both with and without adjustment for peak CK to determine potential contribution of rhabdomyolysis to the BMI-AKI association. We excluded Acute Physiology and Chronic Health Evaluation (APACHE) II scores from our multivariable models because these scores include both creatinine and acute renal failure. In secondary analyses, we determined the association of early CK, measured in the first 24 hours after presentation, with AKI. We also performed an analysis restricted to patients with peak CK<5,000 U/L to determine the association of modest CK elevations with AKI. For these secondary analyses, we constructed multivariable logistic regression models using an approach similar to that used in the primary analysis. All statistical analyses were performed using Stata/IC 15.1 (StataCorp LP, College Station, TX), with p<0.05 considered statistically significant.

RESULTS

From 2005-2015, 1104 injured patients without end-stage kidney disease were enrolled in PETROS. Of these, 553 had a serum CK measurement during the first 7 days and were included in the primary analysis. The median age was 42, 76% were male, and 44% were black. Trauma mechanism was blunt in 78%, and median ISS was 24. Compared with PETROS patients excluded for lack of CK measurement, those in the primary analytic sample were older, had higher median BMI, and had higher rates of hypertension (Table S2). In addition, included patients had higher rates of blunt mechanism, higher Injury Severity Score (ISS), higher APACHE II scores, as well as higher rates of AKI and in-hospital mortality.

Of 553 patients in the primary analysis population, 138 (25%) were obese (BMI ≥30 kg/m²), 155 (28%) were overweight (BMI 25.0-29.9 kg/m²), 159 (29%) had a normal weight (BMI 18.5-24.9 kg/m²) and 11 (2%) were underweight (BMI <18.5 kg/m²). The median peak CK was 1,876 U/L (IQR 512-5,405 U/L) with a median time to peak CK of 34 hours (IQR 15-80 hours). First CK was measured at a median of 14 hours after presentation (IQR 5-50 hours). Overall, 148 (32%) patients developed AKI. Patients who developed AKI were more likely to be African-American, have higher BMI, and have diabetes mellitus (Table 1). In addition, patients with AKI had higher injury and illness severity.

Table 1:

Demographic and clinical characteristics of patients in the primary analysis, stratified by presence of acute kidney injury

Results*	No AKI N=315	AKI N=148	P-value**
Demographics
Age (y)	43 (26-61)	42 (28-60)	0.988
Male Sex	230 (73.0)	120 (81.1)	0.060
Race			0.003
African American	116 (36.8)	81 (54.7)
Caucasian	185 (58.7)	59 (39.9)
Other	14 (4.4)	8 (5.5)
BMI (kg/m2)	26.5 (22.6 - 30.2)	28.0 (24.3-32.3)	0.001
Baseline Comorbid Disease
Hypertension	88 (27.9)	39 (26.4)	0.415
Diabetes mellitus	15 (4.8)	16 (10.8)	0.006
Chronic kidney disease	2 (1.4)	1 (1.6)	0.299
First serum creatinine	1.1 (0.9-1.3)	1.2 (1.0-1.4)	0.001
Congestive heart failure	2 (1.4)	2 (3.3)	0.199
Injury Characteristics
Blunt mechanism	261 (82.9)	105 (70.9)	0.003
ISS	24 (20-29)	25 (19-29)	0.899
APACHE II score	17 (12-22)	23 (18-30)	<0.001
AIS score
Abdomen	0 (0-2)	2 (0-3)	<0.001
External	1 (0-1)	1 (0-1)	0.517
Extremity	2 (0-3)	2 (1-3)	0.131
Face	0 (0-2)	0 (0-0)	0.150
Head	3 (0-4)	0 (0-3)	<0.001
Thorax	3 (0-4)	3 (0-4)	0.569
Peak serum CK (U/L)	1,850 (583-5,139)	3,650 (1,328-9,322)	<0.001
Peak serum CK category			<0.001
≤260 U/L (upper limit of normal)	43 (14%)	4 (3%)
261 − 4,999 U/L	192 (61%)	84 (57%)
≥5,000 U/L	80 (25%)	60 (41%)
Initial Treatment
Operation prior to ICU admission	128 (40.6)	84 (56.8)	0.001
Total crystalloid prior to ICU (L)	2.4 (1.0-4.7)	3.3 (1.3-6.0)	0.011
PRBC transfused within first 24 Hours	1 (0-4)	5 (2-13)	<0.001

^*Variable Format: categorical, n (%); continuous, median (IQR).

^**Wilcoxon rank-sum and t-tests for continuous variables, and X² and Fisher’s exact tests for categorical variables

Abbreviations: AKI, acute kidney injury; BMI, body mass index (kg/m²); ISS, injury severity score; AIS, abbreviated injury scale; APACHE II, acute physiology and chronic health evaluation; ICU, intensive care unit; PRBC, packed red blood cells; IQR, interquartile range

BMI, serum CK, and AKI relationship

In the unadjusted linear regression model, BMI was significantly associated with natural log-transformed peak CK levels (R² 0.048, p<0.001, Figure 1). Although the scatterplot shows a fair degree of variability in this relationship, predicted serum CK levels were nearly 1 natural log higher for an obese patient with a BMI of 35 kg/m² (near the mean BMI of the obese patients in our study) compared with a normal weight patient with a BMI of 20 kg/m². Analyzing patients by WHO BMI categories (Table 2) showed a >1,700 U/l increase in median peak CK levels among obese patients compared with normal weight patients, with underweight patients having substantially lower peak CK.

Figure 1: Scatterplot of body mass index (BMI) vs. natural log-transformed creatine kinase (CK).

Superimposed on the scatterplot is the estimated association of BMI with serum peak CK (solid line), 95% confidence limits (dashed lines), and p-value derived from a linear regression model. Note: Exclusion of the outlier patient with BMI 58.8 did not significantly change the slope of the regression line or statistical significance of the association.

Table 2:

Serum creatine kinase levels and acute kidney injury incidence by body mass index class

	Overall (n=553)	Underweight (n=11)	Normal weight (n=159)	Overweight (n=155)	Obese (n=138)	p*
Peak CK
Peak CK (U/L)	2,225 (704-6,223)	1,045 (452-2,828)	1,813 (538-4,877)	2,220 (760-6,101)	3,529 (1,171-8,637)	<0.001
Time to Peak CK (h)	29 (15-56)	16 (1-36)	25 (11-55)	27 (13-55)	35 (17-59)	0.012
Peak CK ≥ 5,000	140 (25%)	1 (9%)	39 (25%)	46 (30%)	54 (39%)	0.043
First CK
First CK (U/L)	1,247 (452-3,341)	691 (208-1,944)	1,034 (339-3,339)	1,247 (460-3,156)	1,487 (626-3,913)	0.024
Time to First CK (h)	11 (4-26)	2 (1-34)	12 (4-32)	12 (5-25)	11 (4-24)	0.924
First CK ≥ 5,000	79 (14%)	0 (0%)	26 (16%)	25 (16%)	28 (20%)	0.065
AKI
Any AKI, n (%)	148 (32)	2 (18)	42 (26)	46 (30)	58 (42)	0.003
Stage 1	99	1	28	36	34
Stage 2	22	0	11	3	8
Stage 3	27	1	3	7	16

CK and time to CK data shown as median (IQR). CK≥5,000 shown as n (%).

^*Nonparametric test for trend across ordered groups

BMI classification: Underweight (BMI<18.5 kg/m²), Normal (BMI 18.5-24.9 kg/m²), Overweight (BMI 25.0-29.9 kg/m²), Obese (BMI >30 kg/m²)

Abbreviations: CK, creatine kinase; IQR, interquartile range; AKI, acute kidney injury; BMI, body mass index (kg/m²)

Both BMI and peak CK were significantly higher among those who developed AKI (Table 1). AKI incidence increased consistently across WHO BMI groups, with a particularly notable increase in stages 2 and 3 AKI cases among obese patients (Table 2). In multivariable analyses, BMI remained significantly associated with AKI when adjusting for previously described risk factors, including age, diabetes, transfused PRBCs and other potential confounders (Table 3, Model 1)⁹. Peak CK was also associated with AKI adjusting for the same risk factors (Figure 2; Table 3, Model 2). The associations of both BMI and peak CK with AKI were partially attenuated though still statistically significant when both were included in the multivariable model (Table 3, Model 3).

Table 3.

Logistic regression models of risk factors for acute kidney injury.

Covariate	Unadjusted associations*		Model 1** (without peak CK)		Model 2** (without BMI)		Model 3** (all variables)
	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p
Body mass index (per 5 kg/m2)	1.32 (1.13-1.55)	0.001	1.31 (1.09-1.57)	0.004	-	-	1.26 (1.04-1.52)	0.018
Peak CK (ln U/L)	1.41 (1.25-1.61)	<0.001	-	-	1.24 (1.05-1.47)	0.010	1.19 (1.00-1.41)	0.049
Age (per 5 years)	0.99 (0.95-1.04)	0.694	1.06 (1.00-1.13)	0.063	1.09 (1.02-1.16)	0.013	1.08 (1.01-1.15)	0.020
Male sex	1.58 (0.98-2.56)	0.061	1.40 (0.80-2.43)	0.238	1.34 (0.76-2.35)	0.314	1.31 (0.75-2.30)	0.348
African American race	2.07 (1.40-3.08)	<0.001	1.55 (0.93-2.57)	0.093	1.47 (0.88-2.47)	0.139	1.43 (0.85-2.40)	0.176
Diabetes mellitus	2.47 (1.20-5.09)	0.014	3.22 (1.44-7.17)	0.004	3.68 (1.66-8.18)	0.001	3.40 (1.51-7.64)	0.003
Blunt trauma mechanism	0.51 (0.32-0.80)	0.004	1.09 (0.57-2.08)	0.797	0.968 (0.50-1.86)	0.921	0.97 (0.50-1.87)	0.929
AIS score abdomen (per point)	1.31 (1.16-1.50)	<0.001	1.06 (0.91-1.24)	0.469	1.08 (0.93-1.26)	0.323	1.06 (0.90-1.24)	0.501
PRBC transfusion (per unit)	1.14 (1.10-1.18)	<0.001	1.14 (1.09-1.19)	<0.001	1.12 (1.07-1.17)	<0.001	1.13 (1.08-1.18)	<0.001

^*Odds ratios in this column denote the effect estimate of each variable on AKI risk without adjustment for other variables.

^**Models 1-3 include all variables for which odds ratios are shown. Abbreviations: OR, odds ratio; CI, confidence interval; AIS, abbreviated injury scale; PRBC, packed red blood cells (within first 24 hours from admission); CK, creatine kinase (natural log-transformed).

Figure 2: Predicted risk of AKI by peak CK elevation.

Adjusted probability of acute kidney injury (AKI) across the range of peak serum creatine kinase (CK) values. Estimated probabilities (line) with 95% confidence intervals (gray shading) determined using post-estimation marginal analysis after multivariable logistic regression modeling. Estimated probabilities adjusted for age, race, sex, diabetes status, trauma mechanism, abbreviated injury scale (AIS) abdomen score, and packed red blood cell transfusion within 24 hours. Natural log-transformed CK levels were used in the multivariable model and the CK axis is on the natural log scale, though for clarity the corresponding non-transformed CK values are labeled across the x-axis.

Modest peak CK and early CK levels and AKI

We performed several secondary analyses to shed further light on the CK-AKI relationship. First, we assessed the risk of AKI associated with modest peak CK elevations, below the threshold of 5,000 U/l utilized to define rhabdomyolysis in most prior studies of AKI^21,22. Among patients with peak CK<5,000 U/l (n=323), CK was significantly associated with AKI in unadjusted (OR 1.42 per natural log increase; 95% CI 1.14-1.80, p=0.004) and adjusted (aOR 1.31 per natural log increase; 95% CI 1.01-1.69, p=0.043) analyses (Table s3). We also tested whether the first CK values measured within 24 hours of presentation, which might further aid in early AKI risk stratification, were associated with AKI. Early CK levels (available in 338 patients) increased across higher BMI category (Table 2) and showed an unadjusted association with AKI (Table s3). This association was attenuated by adjustment for confounders (Table s3).

DISCUSSION

Our primary objective in this study was to test the following hypotheses in a prospective cohort of severely injured trauma patients: that BMI would correlate with CK levels, that a broad range of CK levels would be associated with AKI, and that CK elevation would contribute to the BMI-AKI association. We found an association between higher BMI and increased post-injury CK elevation. We also showed that the association of BMI with AKI is partially attenuated after adjustment for CK. Finally, we demonstrated that both modest (<5,000 U/l) peak CK values and early (within 24 hours after injury) CK elevation confer significant AKI risk.

Grigorian et al recently showed an association of BMI category with ICD-10-defined rhabdomyolysis using the National Trauma Data Bank¹⁰. Our findings regarding the association of BMI with peak CK, as well as the impact of CK on the association of BMI with AKI, are novel in the trauma population. Data from NHANES III demonstrate strong correlations (ρ>0.70) of BMI with lean body mass in both males and females ³². We hypothesize that trauma patients with elevated BMI may also have greater muscle mass, which provides a larger reservoir available to contribute to the release of both myoglobin and CK after injury. However, it is also possible that obesity itself predisposes to increased CK release after traumatic injury, irrespective of baseline muscle mass. In a study of exercise-induced CK release, Heled et al. demonstrated that increased body fat mass, but not increased muscle mass, was an independent predictor of higher post-exercise CK release⁴⁶. In addition, Haan et al. demonstrated that resting CK levels were independently associated with BMI in a diverse, multi-ethnic patient population⁴⁷. However, despite this association, baseline CK levels still remained well below the upper limit of normal across a broad range of BMI⁴⁷. Our study adds to this limited evidence base by demonstrating a significant association of BMI with a broad range of CK values, 75% of which were below the standard rhabdomyolysis threshold of 5000 U/l. Further, the impact of adjusting the BMI-AKI association for CK suggests that excess muscle injury may indeed be partially responsible for higher AKI rates in obese trauma patients. Since myoglobin, rather than CK, is the principal established nephrotoxic mediator in rhabdomyolysis, subsequent studies testing circulating myoglobin in the very early timeframe after trauma may be warranted to more clearly establish the contribution of rhabdomyolysis to the BMI-AKI relationship.

Treatment-related factors may contribute to our observations as well. Prior studies showed that obese trauma and burn patients, though receiving a greater overall volume of fluid in the first 24-48 hours after injury, received proportionally less fluid when corrected for body weight, and obese trauma patients had more persistent base deficit and lactic acidosis^33,34. Under-resuscitation of obese patients could potentiate rhabdomyolysis-mediated AKI for several reasons. Trauma patients sequester fluid within injured muscle, reducing intra-vascular volume. Decreased effective circulating volume may worsen the renal arteriolar vasoconstriction that results from myoglobin-induced lipid peroxidation and isoprostane production¹². Inadequate glomerular filtration may worsen tubular precipitation of myoglobin with resultant cast formation. It remains unclear whether targeting more liberal fluid resuscitation, the standard treatment for rhabdomyolysis, to obese patients would reduce rhabdomyolysis-related AKI²⁶. Our analyses and prior studies, however, raise the possibility that testing such a strategy, as well as potential novel rhabdomyolysis therapeutics, may ultimately be warranted in this population.

Myoglobin-specific renal injury may not be the only mechanism explaining the relationship of BMI, CK, and AKI. Acutely damaged muscle cells release a variety of effectors such as mitochondrial and nuclear DNA, microRNAs, and uric acid, which create a positive feedback loop for generation of pro-inflammatory cytokines. This contributes to activation of the multiprotein inflammasome complex and leads to renal tubular injury^16,35. This mechanism of renal toxicity may be further exacerbated by increased adiposity, which is known to be associated with a chronic increase in pro-inflammatory cytokines, such as IL-1β, IL-6, IL-8 and tumor necrosis factor alpha⁷.

In addition, iron overload could link increased adiposity with rhabdomyolysis-induced AKI. Non-hospitalized obese patients have increased serum ferritin levels and dysmetabolic iron overload³⁶. Guerrero-Hue et al recently showed that ferroptosis, an iron-dependent form of programmed cell death, was a key mechanism of kidney injury in an animal model of rhabdomyolysis³⁷. Curcumin, an antioxidant and anti-inflammatory herbal compound, was renal-protective in this model. This builds upon an existing body of literature on the importance of iron metabolism in AKI and may represent an additional locus of susceptibility for obese patients³⁸. Limited animal and human studies of the iron chelators deferoxamine and deferiprone have shown potential efficacy in reducing AKI severity³⁹. Further study of how iron metabolism contributes to the associations of BMI, CK, and AKI may help to identify a subgroup of trauma patients who could benefit from iron-targeted therapies.

While most prior studies of rhabdomyolysis in trauma patients have focused exclusively on very high peak CK values, defining rhabdomyolysis as a value ≥5,000 U/l, we found that both early and modest CK elevations were associated with AKI risk^40,41. These findings may be clinically relevant: early CK levels could be used in AKI risk stratification and identifying patients for specific kidney-protective therapies, and even apparently limited muscle injury may be an important pathophysiologic contributor to AKI. Prior reports referencing peak CK values <5,000 U/l have been limited. In a study of unselected trauma patients, Byerly et al found that older patients may develop AKI at lower peak CK levels than younger patients²⁴. However, AKI was defined simply as a peak serum creatinine ≥2 mg/dl during hospitalization, with an incidence of ~5%, and patients who met this creatinine threshold, whether young or old, had a median peak CK >5,000 U/l. In pediatric trauma patients, Talving et al identified an unadjusted association of lower CK values (1,000-5,000 U/l) with AKI risk, but did not perform analyses adjusting for injury severity or other confounders⁴². Published predictive models for AKI in rhabdomyolysis are based on critical illness populations with diverse underlying insults and utilize very high peak CK value cut-offs, in some cases >40,000 U/l^22,43. Despite the described lag between muscle injury and CK peak, our findings suggest that early CK testing could have utility in determining AKI risk across the severity spectrum. Prospective studies with serial early CK measurements, potentially also including early myoglobin testing, may help to clarify this utility.

Our study has several important limitations. First, though we used a cohort with prospectively enrolled patients, the potential for selection bias was introduced by the necessary limitation of the analysis to those with serum CK measured by the clinical team. Included patients were significantly different in several ways, including older age, higher average BMI, and higher injury severity, as well as higher rates of AKI compared to the cohort of patients without CK measurements. Our cohort includes relatively fewer obese patients compared to the adult U.S. population (25% vs. 39%) ⁴⁴. However, obesity rates in our study were similar to those reported in a retrospective study of ~11,000 abdominal trauma patients within the U.S. National Trauma Data Bank⁴⁵. These factors and the use of a single center limit both the sample size of patients with high BMI and the generalizability of our findings to the population in general. Further, timing of CK measurement was not standardized across patients, which likely diminishes the accuracy of peak CK determination. Multicenter studies with uniform timing and application of CK measurement are warranted. Additionally, we considered peak CK measurements as late as 7 days after presentation. While in some cases these values may have reflected secondary muscle insults that occurred after AKI, over 80% of peak values were within 72 hours of presentation and, based on described CK kinetics, likely reflected muscle injury on admission¹⁹. In addition, since free serum myoglobin, rather than CK, is hypothesized as the causative agent of rhabdomyolysis-induced AKI, we did not conduct a formal causal mediation analysis between BMI, CK, and AKI¹⁵. Since circulating myoglobin levels peak and fall rapidly after muscle injury, future studies of serum myoglobin tested shortly after presentation may facilitate such mediation analyses to better define the contribution of rhabdomyolysis to the BMI-AKI association in trauma patients.

In conclusion, we demonstrated that higher BMI was associated with higher peak CK levels across a broad range of CK values, that early and modest CK elevations were associated with AKI, and that adjustment for peak CK partially attenuated the association of BMI with AKI. If confirmed in subsequent studies, this may reflect a causal, and potentially targetable, pathway by which obesity contributes to AKI risk after trauma.