- Email: [email protected]
Sarcopenia is a risk factor for complications and an independent predictor of hospital length of stay in trauma patients
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 1 8 ( 2 2 1 ) 1 6 1 e1 6 6
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Sarcopenia is a risk factor for complications and an independent predictor of hospital length of stay in trauma patients James DeAndrade, MD, Mark Pedersen, MD, MPH, Luis Garcia, MD, and Peter Nau, MD, MS* The University of Iowa Hospitals and Clinics, Iowa City, Iowa
article info
abstract
Article history:
Background: Sarcopenia is an independent risk factor for adverse outcomes in critically
Received 10 November 2016
ill patients. The impact of sarcopenia on morbidity and length of stay in a trauma popu-
Received in revised form
lation has not been completely defined. This project evaluated the influence of sarcopenia
28 June 2017
on patients admitted to the trauma service.
Accepted 10 August 2017
Materials and methods: A retrospective review of 778 patients presenting as a trauma alert at a
Available online xxx
single institution from 2012-2014 was completed. Records were abstracted for comorbidities and hospital complications. The Hounsfield Unit Area Calculation was collected from
Keywords:
admission computed tomography scans. Criteria for sarcopenia were based on the lowest 25th
Sarcopenia
percentile of muscle density measurements. Relationships to patient outcomes were evalu-
Length of stay
ated by univariate and multivariable regression or analyses of variance, when applicable.
Hospital outcomes
Results: A total of 432 (55.6%) patients suffered a complication. Sarcopenia was associated
Trauma
with overall complications (P < 0.0001, relative risk 2.54, confidence interval 1.78-3.61) and was an independent risk factor for catheter-associated urinary tract infections (P ¼ 0.011), wound infections (P ¼ 0.011), need for reintubation (P ¼ 0.0062), and length of hospitalization (P ¼ 0.0007). Incorporating sarcopenia into a novel length of stay calculator showed increased prognostic ability for prolonged length of stay over Abbreviated Injury Scale alone (P ¼ 0.0002). Conclusions: Sarcopenia is an independent risk factor for adverse outcomes and increased length of stay in trauma patients. Prognostic algorithms incorporating sarcopenia better predict hospital length of stay. Identification of patients at risk may allow for targeted interventions early in the patient’s hospital course. ª 2017 Elsevier Inc. All rights reserved.
Introduction Sarcopenia is the quantifiable loss of muscle mass that accompanies a deconditioned state of health.1 Often characterized as a progressive syndrome that accompanies aging and catabolism, the relevance of sarcopenia is not limited to
medical outcomes in cachectic elderly cohorts. In fact, sarcopenia is often an insidious process with few outward indicators of the disease.2-4 Sarcopenia correlates with adverse outcomes in several populations including patients with cancer, the critically ill, and those undergoing a diversity of surgical procedures.5-7
* Corresponding author. The University of Iowa Hospitals and Clinics, 4626 JCP, 200 Hawkins Drive, Iowa City, IA 52242-1086. Tel.: þ319 356-7675; fax: þ319 356-4609. E-mail address: [email protected] (P. Nau). 0022-4804/$ e see front matter ª 2017 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jss.2017.08.018
162
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 1 8 ( 2 2 1 ) 1 6 1 e1 6 6
After elective operations, sarcopenic patients are more likely to incur higher payer costs even after adjusting for patient and procedural factors.6 While less well defined, sarcopenia appears to play an important role in the outcomes of patients who suffer traumatic injuries or require emergent surgery. In elderly trauma patients, sarcopenia is a strong risk factor for loss of independence upon hospital discharge. One study noted that each additional 1 cm2 of psoas area predicted a 20% increase in functional independence.8 It is hypothesized that there is a pathologic sarcopenia distinct from that which is seen with aging alone and that this state is predictive of posttrauma morbidity. Furthermore, due to the significance of this finding on a patient’s overall wellbeing, it is postulated that sarcopenia is predictive of length of stay in this population.
Methods Patients and sarcopenia quantification A retrospective study of hospital charts of 778 patients presenting as a trauma alert at a single institution from 2012-2014 who underwent abdominal computed tomography (CT) scans was completed. CT scan measurements were recorded, and sarcopenia quantification was done using the validated method described by Joglekar et al.5 In short, sarcopenia was defined using a mean Hounsfield Unit Area Calculation within the lowest gender quartile as measured at the third lumbar vertebrae. Patient charts were also reviewed for clinical assessment data and complications. Appropriate approval from the University of Iowa Hospitals and Clinics (UIHC) institutional review board was obtained. A waiver of informed consent was obtained due to the retrospective nature of this research.
Statistical analysis Univariate logistic regression was used for categorical outcome variables and linear regression was used for continuous outcome variables to determine which predictor variables were significantly associated with complications. Patient variables abstracted as potential predictors for poor outcomes included the following factors: age, body mass index, comorbid conditions, smoking status, history of venous thrombosis, abdominal wall fat, hip girdle fat, visceral fat, and admission laboratory values including complete blood count, blood urea nitrogen, liver function tests, albumin, pre-albumin, and international normalized ratio. Trauma-related characteristics collected included Glasgow Coma Scale and the injury severity score. In addition, head, chest, abdomen, and extremity abbreviated injury scale (AIS) scores were recorded. The maximum AIS score was noted as its own variable. Additional variables included receipt of transfusion of any kind, total products transfused, total red blood cell transfusion, receipt of platelets, total platelet transfusion, receipt of fresh frozen plasma, and total fresh frozen plasma transfusion. Complications abstracted from records included the following factors: time in intensive care unit (ICU), length of stay, readmission, organ space infection, central lineeassociated blood stream infections (CLABSIs), catheterassociated urinary tract infections (CAUTIs), hospital-acquired pneumonia, sepsis, venous thrombosis requiring treatment,
discharge disposition, death while hospitalized, death within 30 d, death within 90 d, and total mortality. Multivariable logistic and linear regression models were constructed using all predictor variables found to be significantly associated with complications.
Length-of-stay prediction The finding of sarcopenia as a predictor of morbidity in a trauma cohort is not unexpected, given the available data in surgical populations. To date, its roll in a nongeriatric trauma population has not been completely defined. Having established the significance of sarcopenia in this group, it was thought to be critical to develop a scoring system that would evaluate the predictive capacity of sarcopenia on length of stay. To do this, the variables significantly associated with prolonged length of stay from the multivariate analysis were aggregated with the presence of sarcopenia. These variables included admission white blood cell count, maximum AIS value, and a diagnosis of congestive heart failure and hypertension. When linear regression was performed on these variables, an aspect of the data produced was the parameter estimate of the coefficients for these variables. These parameter estimates of coefficients served as the relative weights of each variable when creating the final calculator predicting the length of stay. In this way, each variable would have the appropriately weighted influence on the length of stay estimate. For continuous variables, the result that gave the greatest sensitivity for prolonged length of stay (>10 d) on the receiver operating characteristic (ROC) analysis was used as a cutoff point. Patient data for said variable, which was above and below this set point was assigned a number. In this way, a continuous variable such as white blood cell (WBC) count was translated into a categorical variable. A score was derived from this value multiplied by the respective linear regression coefficient parameter estimates to achieve the variable’s contribution to the composite score. For categorical variables, no translation was necessary. In this case, the calculation was derived from the weight defined by the linear regression analysis multiplied by the categorical value. Each of these results were then combined into a composite score predicting the length of stay. This final score did not include the injury severity score, as it was noted to be redundant with a patient’s maximum AIS score and did not add any significant prognostic power. These scores and the overall composite scoring system were developed using a random sampling of half of the patient population and then tested for prognostic power using the entire population. This assessment of prognostic power was done through ROC analysis of the scoring system for discrete lengths of stay.
Results A total of 978 patients presented as a trauma alert and also underwent CT imaging from 2012-2014 at the UIHC. Two hundred patients were excluded from the analysis due to missing data points for a total of 778 enrolled. Seventy-one percent of patients were male. The average age of those with sarcopenia was older than those without (63 versus 42 years). Sarcopenic patients were more likely to suffer from numerous medical comorbidities including obesity (Table 1).
163
deandrade et al sarcopenia predicts outcomes in trauma population
Table 1 e Comorbidities of patients included in study based on the presence or absence of sarcopenia.
Table 2 e Univariate relationship between sarcopenia and in-hospital morbidity.
Comorbidities
Complications
Hypertension
N (%) 238 (31%)
Sarcopenic? N (%)
P Value
Yes: 88 (45%)
<0.0001
No: 150 (26%) Dyslipidemia
73 (9.4%)
Yes: 36 (18.6%)
Diabetes
64 (8.2%)
Yes: 36 (18.5%)
<0.0001
No: 37 (6.4%) <0.0001
No: 28 (4.8%) Congestive heart failure
26 (3.4%)
Yes: 15 (7.7%)
0.0003
OR (CI)
P Value
Acute renal failure
1.11 (1.057-1.175)
<0.0001
Reintubation
1.09 (1.023-1.157)
0.0075
CLABSI
1.09 (1.001-1.193)
0.0485
HAP
1.05 (1.009-1.094)
0.0157
Wound infection
1.09 (1.019-1.167)
0.0128
CAUTI
1.10 (1.050-1.145)
<0.0001
OR ¼ odds ratio; CI ¼ confidence interval; HAP ¼ hospital-acquired pneumonia.
No: 11 (1.9%) Coronary artery disease
39 (5.0%)
Yes: 24 (12%)
Chronic obstructive pulmonary disease
24 (3.1%)
Yes: 11 (5.6%)
<0.0001
No: 15 (2.6%) 0.0136
Discussion
No: 13 (2.2%) Body mass index
Yes: 31.1 kg/m2
model of the three (area under the curve of 0.73 for UIHC calculator).
<0.0001
No: 28.2 kg/m2
Univariate analysis revealed that sarcopenia was associated with multiple complications. This included new onset acute renal failure, necessity of reintubation, CLABSI, hospital-acquired pneumonia, wound infection, and CAUTI (Table 2). Sarcopenia also correlated with the time spent in the ICU (P value ¼ 0.0026) as well as overall increased length of stay (P value ¼ 0.0007). Discharge disposition was related to the presence of sarcopenia. The loss of lean mass was also associated with discharge to an acute rehab facility and inversely related with freedom of assistance at home after hospital discharge (P value ¼ 0.0018 and < 0.0001, respectively). Finally, on univariate analysis, the presence of sarcopenia was predictive of 90 d mortality (P value ¼ 0.009; Table 3). On multivariable analysis several in-hospital complications remained independently associated with sarcopenia including CAUTI, the need for reintubation, CLABSI, and wound infections (Table 4). Additional independent variables such as age and comorbidities were also associated with these outcomes, but sarcopenia remained as an independent risk factor for these outcomes. When evaluating for the relationship between sarcopenia and length of stay, it was noted that the presence of hypertension and congestive heart failure as well as the maximum AIS score and WBC count on admission were independently associated with an increased time spent in the hospital (Table 5). The length of stay calculator was developed based on the aforementioned significant variables in the multivariable linear regression model for length of stay. The calculator was tested on the entire cohort of 778 patients at specified time points during the patient’s hospitalization. This score was compared to max AIS, as well as max AIS with the addition of sarcopenia using ROC analysis. It was found that the UIHC model was more predictive that max AIS or max AIS plus sarcopenia at 5, 10 and 20 d (Fig. 1). By 30 d, all three scoring systems became less accurate at predicting the length of stay. Notably, the UIHC calculator persisted as the most accurate
Cross-sectional imaging is a readily available and important element of the diagnostic algorithm for patients who suffer traumatic injuries. CT scans allow for prompt, highly accurate identification of pathology for nonelife-threatening and lifethreatening injuries.9,10 It has been noted that, in patients suffering blunt trauma without overt findings, CT imaging may identify occult management-altering injuries in 18.9% of cases.9 These images can also provide objective markers of sarcopenia. Synthesizing this data with specific patient characteristics provides the practitioner with the information that can aid in risk stratification for experiencing negative outcomes.11-13 Sarcopenia is defined as the loss of skeletal muscle mass, which occurs with aging.14 This relative deconditioning is pervasive and distinct from that which is seen in a persistently ill population. Sarcopenia associated with chronic medical conditions is an insidious finding that may be missed on subjective evaluation. In addition, it is often noted in obese cohorts and has been given the title of osteosarcopenic obesity.3 The loss of muscle quality and quantity has been associated with an increased risk of disability.15 It has also been shown to be a predictor of postoperative morbidity in numerous surgical populations.5,16,17 With the increasing emphasis on improving hospital outcomes and the well-defined correlation between sarcopenia and morbidity, defining the presence of sarcopenia will only become increasingly important in the future.
Table 3 e Univariate analysis identifying the association between discharge disposition and sarcopenia. Discharge disposition Time spent in ICU Length of stay Home without assistance
OR (CI) 0.94 (0.913-0.962)
P Value 0.0026 0.0007 <0.0001
Discharge to a facility
1.04 (1.015-1.068)
0.0018
90-d mortality
1.11 (1.026-1.206)
0.0094
OR ¼ odds ratio; CI ¼ confidence interval.
164
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 1 8 ( 2 2 1 ) 1 6 1 e1 6 6
Table 4 e Multivariable relationship between sarcopenia and the incidence of in-hospital complications. Complications
OR (CI)
P Value
CAUTI
1.06 (1.021-1.170)
0.0201
Reintubation
1.10 (1.028-1.180)
0.0062
CLABSI
1.09 (1.001-1.193)
0.0485
Wound infection
1.12 (1.026-1.221)
0.0111
OR ¼ Odds ratio; CI ¼ confidence interval.
It is clear that sarcopenia is independently associated with poor outcomes in surgical cohorts. It has also been identified as a risk factor for adverse discharge disposition in ICU patients.18 To date, there are little data pertaining to the relevance of this finding in a trauma population. The Canadian Study of Health and Aging Frailty Index and a modified frailty index have been used to identify patients at risk for inhospital complications.19,20 Frailty, as defined by the presence of sarcopenia and osteopenia, has been shown to be associated with traumatic thoracic and spine injuries.21 Muscle atrophy of the vastus lateralis was shown to be more prominent in patients who had suffered fall-related hip fractures versus healthy young or age-matched healthy controls.22 Owing to the relative void in the literature, this project was designed to investigate the effects of sarcopenia in a trauma population as it relates to clinical outcomes. Furthermore, given the prominent association between sarcopenia and morbidity in surgical cohorts, this study sought to create a predictive length-of-stay model in this unique population incorporating patient characteristics in addition to sarcopenia for a lengthof-stay calculator. In this cohort, patients with sarcopenia were noted to be older and suffer from more medical comorbidities (Table 1). The loss of lean muscle mass correlated with increased overall length of stay as well as time spent in the ICU. The presence of sarcopenia was associated with adverse discharge disposition and, perhaps most importantly, was predictive of mortality (Table 3). In line with the current literature, this population was more likely to be obese than nonsarcopenic individuals. These finding are significant in that sarcopenia is an occult process that is challenging to recognize on physical examination. This disease is not restricted to a frail, cachectic patient but rather can be identified in patients who appear vigorous and healthy. In the absence of an objective measurement for this disease, the identification of an at-risk population and a potential opportunity to intervene would be missed. Using readily available data in routinely ordered CT
Table 5 e Admission factors associated with length of stay on multivariable analysis. Variable
Parameter estimate
P Value
Sarcopenia
0.152
0.0038
Maximum AIS score
þ3.085
<0.0001
Admission WBC count
þ0.249
<0.0001
Congestive heart failure
þ5.971
0.0002
Hypertension
þ1.936
0.0027
Fig. 1 e Ability of UIHC model to predict 10 d length of stay (blue line e 0.79) versus max AIS D sarcopenia (red line e 0.76) and max AIS (brown line e 0.75).
imaging, sarcopenia can be identified, and this group flagged early in the hospital course. Multivariable analysis of the correlation between sarcopenia and the incidence of complications identified CAUTI, the need for reintubation, CLABSI, and wound infections as all being independently related to decreased lean mass. Previous studies have found that sarcopenia is more important than chronological age when considering perioperative morbidity. Englesbe et al.23 developed the theory of morphometric age, showing that it was able to more accurately predict the length of stay and postoperative mortality than chronological age. While novel in that there is a paucity of data on sarcopenia in a trauma setting, this was not altogether unexpected, given the available information in operative and ICU populations. Given this, the relationship between sarcopenia and length of stay was evaluated. On linear regression of the length of stay, it was noted that there were five statistically significant, independently predictive variables including sarcopenia, maximum AIS score, admission WBC count, and the presence or absence of congestive heart failure and hypertension. Using this data, a novel length-of-stay calculator was created using a random sampling of half of the cohort. This calculator was then validated for prognostic power using the entire population by means of an ROC analysis. The UIHC calculator was found to be more predictive than max AIS and max AIS with the addition of sarcopenia at 5-, 10- and 20-d hospital stays (Figs. 1-3). As the length of hospitalization increased, the prognostic capabilities decreased, while still remaining more accurate than max AIS. This is significant in that trauma patients with prolonged lengths of stay can be fiscally challenging for hospital systems. A retrospective review from a single academic institution found that its trauma center posted financial losses in caring for over half of patients with
deandrade et al sarcopenia predicts outcomes in trauma population
165
outcomes. The significance of these two findings in the ability to risk stratify for complications and also with discharge planning cannot be overstated. There are limitations to this research. This is a single institution, retrospective review of a trauma service with predominantly nonpenetrating admissions. Institutions with a preponderance of penetrating trauma with the attendant change in demographic make-up may find this calculator to be less accurate. Furthermore, in-hospital mortalities with a concomitant max AIS score of 6 were excluded, as those patients all sustained nonsurvivable injuries. This was done because the mortality of this population was likely secondary to severity of injury and less influenced by patient comorbidities including sarcopenia. Future directions of study would include a prospective assessment of the role of sarcopenia and the accuracy of this calculator across multiple centers.
Conclusion Fig. 2 e Ability of UIHC model to predict 5 d length of stay (blue line e 0.79) versus Max AIS D Sarcopenia (red e 0.76) and Max AIS (brown e 0.75).
hospitalizations of greater than 50 d.24 Targeting surgical patients with sarcopenia, early in admission with clinical resources, has been shown to not only reduce costs but also to shorten hospital lengths of stay.25 The UIHC calculator provides the traumatologist with an instrument that more accurately predicts the length of stay than any currently available tool. The identification of sarcopenia in a trauma patient also categorizes that person for an increased possibility of adverse
Sarcopenia is an insidious disease, which is difficult to identify on subjective examination alone. The loss of lean muscle is associated with multiple medical comorbidities and impaired perioperative outcomes. In a trauma population, it is an independent risk factor for adverse outcomes and increased length of stay. Incorporating an objective evaluation for lean mass, the UIHC calculator can accurately predict the length of stay. This allows for the identification of patients for early, targeted therapy and appropriate discharge planning.
Acknowledgment Author contributions De Andrade was responsible for data collection and manuscript drafting; Pedersen performed the data evaluation and study design; Garcia was responsible for study design and manuscript editing; Nau was responsible for data collection, study design, manuscript drafting and editing.
Disclosure The authors have no conflicts of interest to disclose.
references
Fig. 3 e Ability of UIHC model to predict 20 d length of stay (blue line e 0.78) versus Max AIS D Sarcopenia (red e 0.76) and Max AIS (brown e 0.72).
1. Cruz-Jentoft AJ, Baeyens JP, Bauer JM, et al. Sarcopenia: European consensus on definition and diagnosis: report of the European Working Group on sarcopenia in older people. Age Ageing. 2010;39:412e423. 2. Chung JY, Kang HT, Lee DC, Lee HR, Lee YJ. Body composition and its association with cardiometabolic risk factors in the elderly: a focus on sarcopenic obesity. Arch Gerontol Geriatr. 2013;56:270e278. 3. Ormsbee MJ, Prado CM, Ilich JZ, et al. Osteosarcopenic obesity: the role of bone, muscle, and fat on health. J Cachexia Sarcopenia Muscle. 2014;5:183e192.
166
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 1 8 ( 2 2 1 ) 1 6 1 e1 6 6
4. Prado CM, Wells JC, Smith SR, Stephan BC, Siervo M. Sarcopenic obesity: a critical appraisal of the current evidence. Clin Nutr. 2012;31:583e601. 5. Joglekar S, Nau PN, Mezhir JJ. The impact of sarcopenia on survival and complications in surgical oncology: a review of the current literature-author response. J Surg Oncol. 2015;112:910. 6. Sheetz KH, Waits SA, Terjimanian MN, et al. Cost of major surgery in the sarcopenic patient. J Am Coll Surg. 2013;217:813e818. 7. Moisey LL, Mourtzakis M, Cotton BA, et al. Skeletal muscle predicts ventilator-free days, ICU-free days, and mortality in elderly ICU patients. Crit Care. 2013;17:R206. 8. Fairchild B, Webb TP, Xiang Q, et al. Sarcopenia and frailty in elderly trauma patients. World J Surg. 2015;39:373e379. 9. Salim A, Sangthong B, Martin M, Brown C, Plurad D, Demetriades D. Whole body imaging in blunt multisystem trauma patients without obvious signs of injury: results of a prospective study. Arch Surg. 2006;141:468e473. discussion 473-475. 10. Rieger M, Czermak B, El Attal R, Sumann G, Jaschke W, Freund M. Initial clinical experience with a 64-MDCT wholebody scanner in an emergency department: better time management and diagnostic quality? J Trauma. 2009;66:648e657. 11. Peng P, Hyder O, Firoozmand A, et al. Impact of sarcopenia on outcomes following resection of pancreatic adenocarcinoma. J Gastrointest Surg. 2012;16:1478e1486. 12. House MG, Fong Y, Arnaoutakis DJ, et al. Preoperative predictors for complications after pancreaticoduodenectomy: impact of BMI and body fat distribution. J Gastrointest Surg. 2008;12:270e278. 13. Boutin RD, Yao L, Canter RJ, Lenchik L. Sarcopenia: current concepts and imaging implications. AJR Am J Roentgenol. 2015;205:W255eW266. 14. Englesbe MJ, Lee JS, He K, et al. Analytic morphomics, core muscle size, and surgical outcomes. Ann Surg. 2012;256:255e261.
15. Janssen I, Baumgartner RN, Ross R, Rosenberg IH, Roubenoff R. Skeletal muscle cutpoints associated with elevated physical disability risk in older men and women. Am J Epidemiol. 2004;159:413e421. 16. Joglekar S, Asghar A, Mott SL, et al. Sarcopenia is an independent predictor of complications following pancreatectomy for adenocarcinoma. J Surg Oncol. 2015;111:771e775. 17. Englesbe MJ, Patel SP, He K, et al. Sarcopenia and mortality after liver transplantation. J Am Coll Surg. 2010;211:271e278. 18. Mueller N, Murthy S, Tainter CR, et al. Can sarcopenia quantified by ultrasound of the rectus femoris muscle predict adverse outcome of surgical intensive care unit patients as well as frailty? A prospective, observational cohort study. Ann Surg. 2016;264:1116e1124. 19. Ali R, Schwalb JM, Nerenz DR, Antoine HJ, Rubinfeld I. Use of the modified frailty index to predict 30-day morbidity and mortality from spine surgery. J Neurosurg Spine. 2016;25:537e541. 20. Rockwood K, Song X, MacKnight C, et al. A global clinical measure of fitness and frailty in elderly people. CMAJ. 2005;173:489e495. 21. Oskutis MQ, Lauerman MH, Kufera JA, et al. Are frailty markers associated with serious thoracic and spinal injuries among motor vehicle crash occupants? J Trauma Acute Care Surg. 2016;81:156e161. 22. Kramer IF, Snijders T, Smeets JS, et al. Extensive type II muscle fiber atrophy in elderly female hip fracture patients. J Gerontol A Biol Sci Med Sci. 2017. http://dx.doi.org/10.1093/ gerona/glw253 [Epub ahead of print]. 23. Englesbe MJ, Terjimanian MN, Lee JS, et al. Morphometric age and surgical risk. J Am Coll Surg. 2013;216:976e985. 24. Fakhry SM, Couillard D, Liddy CT, Adams D, Norcross ED. Trauma center finances and length of stay: identifying a profitability inflection point. J Am Coll Surg. 2010;210:817e821, 821-823. 25. Friedman J, Lussiez A, Sullivan J, Wang S, Englesbe M. Implications of sarcopenia in major surgery. Nutr Clin Pract. 2015;30:175e179.
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.JournalofSurgicalResearch.com
Sarcopenia is a risk factor for complications and an independent predictor of hospital length of stay in trauma patients James DeAndrade, MD, Mark Pedersen, MD, MPH, Luis Garcia, MD, and Peter Nau, MD, MS* The University of Iowa Hospitals and Clinics, Iowa City, Iowa
article info
abstract
Article history:
Background: Sarcopenia is an independent risk factor for adverse outcomes in critically
Received 10 November 2016
ill patients. The impact of sarcopenia on morbidity and length of stay in a trauma popu-
Received in revised form
lation has not been completely defined. This project evaluated the influence of sarcopenia
28 June 2017
on patients admitted to the trauma service.
Accepted 10 August 2017
Materials and methods: A retrospective review of 778 patients presenting as a trauma alert at a
Available online xxx
single institution from 2012-2014 was completed. Records were abstracted for comorbidities and hospital complications. The Hounsfield Unit Area Calculation was collected from
Keywords:
admission computed tomography scans. Criteria for sarcopenia were based on the lowest 25th
Sarcopenia
percentile of muscle density measurements. Relationships to patient outcomes were evalu-
Length of stay
ated by univariate and multivariable regression or analyses of variance, when applicable.
Hospital outcomes
Results: A total of 432 (55.6%) patients suffered a complication. Sarcopenia was associated
Trauma
with overall complications (P < 0.0001, relative risk 2.54, confidence interval 1.78-3.61) and was an independent risk factor for catheter-associated urinary tract infections (P ¼ 0.011), wound infections (P ¼ 0.011), need for reintubation (P ¼ 0.0062), and length of hospitalization (P ¼ 0.0007). Incorporating sarcopenia into a novel length of stay calculator showed increased prognostic ability for prolonged length of stay over Abbreviated Injury Scale alone (P ¼ 0.0002). Conclusions: Sarcopenia is an independent risk factor for adverse outcomes and increased length of stay in trauma patients. Prognostic algorithms incorporating sarcopenia better predict hospital length of stay. Identification of patients at risk may allow for targeted interventions early in the patient’s hospital course. ª 2017 Elsevier Inc. All rights reserved.
Introduction Sarcopenia is the quantifiable loss of muscle mass that accompanies a deconditioned state of health.1 Often characterized as a progressive syndrome that accompanies aging and catabolism, the relevance of sarcopenia is not limited to
medical outcomes in cachectic elderly cohorts. In fact, sarcopenia is often an insidious process with few outward indicators of the disease.2-4 Sarcopenia correlates with adverse outcomes in several populations including patients with cancer, the critically ill, and those undergoing a diversity of surgical procedures.5-7
* Corresponding author. The University of Iowa Hospitals and Clinics, 4626 JCP, 200 Hawkins Drive, Iowa City, IA 52242-1086. Tel.: þ319 356-7675; fax: þ319 356-4609. E-mail address: [email protected] (P. Nau). 0022-4804/$ e see front matter ª 2017 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jss.2017.08.018
162
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 1 8 ( 2 2 1 ) 1 6 1 e1 6 6
After elective operations, sarcopenic patients are more likely to incur higher payer costs even after adjusting for patient and procedural factors.6 While less well defined, sarcopenia appears to play an important role in the outcomes of patients who suffer traumatic injuries or require emergent surgery. In elderly trauma patients, sarcopenia is a strong risk factor for loss of independence upon hospital discharge. One study noted that each additional 1 cm2 of psoas area predicted a 20% increase in functional independence.8 It is hypothesized that there is a pathologic sarcopenia distinct from that which is seen with aging alone and that this state is predictive of posttrauma morbidity. Furthermore, due to the significance of this finding on a patient’s overall wellbeing, it is postulated that sarcopenia is predictive of length of stay in this population.
Methods Patients and sarcopenia quantification A retrospective study of hospital charts of 778 patients presenting as a trauma alert at a single institution from 2012-2014 who underwent abdominal computed tomography (CT) scans was completed. CT scan measurements were recorded, and sarcopenia quantification was done using the validated method described by Joglekar et al.5 In short, sarcopenia was defined using a mean Hounsfield Unit Area Calculation within the lowest gender quartile as measured at the third lumbar vertebrae. Patient charts were also reviewed for clinical assessment data and complications. Appropriate approval from the University of Iowa Hospitals and Clinics (UIHC) institutional review board was obtained. A waiver of informed consent was obtained due to the retrospective nature of this research.
Statistical analysis Univariate logistic regression was used for categorical outcome variables and linear regression was used for continuous outcome variables to determine which predictor variables were significantly associated with complications. Patient variables abstracted as potential predictors for poor outcomes included the following factors: age, body mass index, comorbid conditions, smoking status, history of venous thrombosis, abdominal wall fat, hip girdle fat, visceral fat, and admission laboratory values including complete blood count, blood urea nitrogen, liver function tests, albumin, pre-albumin, and international normalized ratio. Trauma-related characteristics collected included Glasgow Coma Scale and the injury severity score. In addition, head, chest, abdomen, and extremity abbreviated injury scale (AIS) scores were recorded. The maximum AIS score was noted as its own variable. Additional variables included receipt of transfusion of any kind, total products transfused, total red blood cell transfusion, receipt of platelets, total platelet transfusion, receipt of fresh frozen plasma, and total fresh frozen plasma transfusion. Complications abstracted from records included the following factors: time in intensive care unit (ICU), length of stay, readmission, organ space infection, central lineeassociated blood stream infections (CLABSIs), catheterassociated urinary tract infections (CAUTIs), hospital-acquired pneumonia, sepsis, venous thrombosis requiring treatment,
discharge disposition, death while hospitalized, death within 30 d, death within 90 d, and total mortality. Multivariable logistic and linear regression models were constructed using all predictor variables found to be significantly associated with complications.
Length-of-stay prediction The finding of sarcopenia as a predictor of morbidity in a trauma cohort is not unexpected, given the available data in surgical populations. To date, its roll in a nongeriatric trauma population has not been completely defined. Having established the significance of sarcopenia in this group, it was thought to be critical to develop a scoring system that would evaluate the predictive capacity of sarcopenia on length of stay. To do this, the variables significantly associated with prolonged length of stay from the multivariate analysis were aggregated with the presence of sarcopenia. These variables included admission white blood cell count, maximum AIS value, and a diagnosis of congestive heart failure and hypertension. When linear regression was performed on these variables, an aspect of the data produced was the parameter estimate of the coefficients for these variables. These parameter estimates of coefficients served as the relative weights of each variable when creating the final calculator predicting the length of stay. In this way, each variable would have the appropriately weighted influence on the length of stay estimate. For continuous variables, the result that gave the greatest sensitivity for prolonged length of stay (>10 d) on the receiver operating characteristic (ROC) analysis was used as a cutoff point. Patient data for said variable, which was above and below this set point was assigned a number. In this way, a continuous variable such as white blood cell (WBC) count was translated into a categorical variable. A score was derived from this value multiplied by the respective linear regression coefficient parameter estimates to achieve the variable’s contribution to the composite score. For categorical variables, no translation was necessary. In this case, the calculation was derived from the weight defined by the linear regression analysis multiplied by the categorical value. Each of these results were then combined into a composite score predicting the length of stay. This final score did not include the injury severity score, as it was noted to be redundant with a patient’s maximum AIS score and did not add any significant prognostic power. These scores and the overall composite scoring system were developed using a random sampling of half of the patient population and then tested for prognostic power using the entire population. This assessment of prognostic power was done through ROC analysis of the scoring system for discrete lengths of stay.
Results A total of 978 patients presented as a trauma alert and also underwent CT imaging from 2012-2014 at the UIHC. Two hundred patients were excluded from the analysis due to missing data points for a total of 778 enrolled. Seventy-one percent of patients were male. The average age of those with sarcopenia was older than those without (63 versus 42 years). Sarcopenic patients were more likely to suffer from numerous medical comorbidities including obesity (Table 1).
163
deandrade et al sarcopenia predicts outcomes in trauma population
Table 1 e Comorbidities of patients included in study based on the presence or absence of sarcopenia.
Table 2 e Univariate relationship between sarcopenia and in-hospital morbidity.
Comorbidities
Complications
Hypertension
N (%) 238 (31%)
Sarcopenic? N (%)
P Value
Yes: 88 (45%)
<0.0001
No: 150 (26%) Dyslipidemia
73 (9.4%)
Yes: 36 (18.6%)
Diabetes
64 (8.2%)
Yes: 36 (18.5%)
<0.0001
No: 37 (6.4%) <0.0001
No: 28 (4.8%) Congestive heart failure
26 (3.4%)
Yes: 15 (7.7%)
0.0003
OR (CI)
P Value
Acute renal failure
1.11 (1.057-1.175)
<0.0001
Reintubation
1.09 (1.023-1.157)
0.0075
CLABSI
1.09 (1.001-1.193)
0.0485
HAP
1.05 (1.009-1.094)
0.0157
Wound infection
1.09 (1.019-1.167)
0.0128
CAUTI
1.10 (1.050-1.145)
<0.0001
OR ¼ odds ratio; CI ¼ confidence interval; HAP ¼ hospital-acquired pneumonia.
No: 11 (1.9%) Coronary artery disease
39 (5.0%)
Yes: 24 (12%)
Chronic obstructive pulmonary disease
24 (3.1%)
Yes: 11 (5.6%)
<0.0001
No: 15 (2.6%) 0.0136
Discussion
No: 13 (2.2%) Body mass index
Yes: 31.1 kg/m2
model of the three (area under the curve of 0.73 for UIHC calculator).
<0.0001
No: 28.2 kg/m2
Univariate analysis revealed that sarcopenia was associated with multiple complications. This included new onset acute renal failure, necessity of reintubation, CLABSI, hospital-acquired pneumonia, wound infection, and CAUTI (Table 2). Sarcopenia also correlated with the time spent in the ICU (P value ¼ 0.0026) as well as overall increased length of stay (P value ¼ 0.0007). Discharge disposition was related to the presence of sarcopenia. The loss of lean mass was also associated with discharge to an acute rehab facility and inversely related with freedom of assistance at home after hospital discharge (P value ¼ 0.0018 and < 0.0001, respectively). Finally, on univariate analysis, the presence of sarcopenia was predictive of 90 d mortality (P value ¼ 0.009; Table 3). On multivariable analysis several in-hospital complications remained independently associated with sarcopenia including CAUTI, the need for reintubation, CLABSI, and wound infections (Table 4). Additional independent variables such as age and comorbidities were also associated with these outcomes, but sarcopenia remained as an independent risk factor for these outcomes. When evaluating for the relationship between sarcopenia and length of stay, it was noted that the presence of hypertension and congestive heart failure as well as the maximum AIS score and WBC count on admission were independently associated with an increased time spent in the hospital (Table 5). The length of stay calculator was developed based on the aforementioned significant variables in the multivariable linear regression model for length of stay. The calculator was tested on the entire cohort of 778 patients at specified time points during the patient’s hospitalization. This score was compared to max AIS, as well as max AIS with the addition of sarcopenia using ROC analysis. It was found that the UIHC model was more predictive that max AIS or max AIS plus sarcopenia at 5, 10 and 20 d (Fig. 1). By 30 d, all three scoring systems became less accurate at predicting the length of stay. Notably, the UIHC calculator persisted as the most accurate
Cross-sectional imaging is a readily available and important element of the diagnostic algorithm for patients who suffer traumatic injuries. CT scans allow for prompt, highly accurate identification of pathology for nonelife-threatening and lifethreatening injuries.9,10 It has been noted that, in patients suffering blunt trauma without overt findings, CT imaging may identify occult management-altering injuries in 18.9% of cases.9 These images can also provide objective markers of sarcopenia. Synthesizing this data with specific patient characteristics provides the practitioner with the information that can aid in risk stratification for experiencing negative outcomes.11-13 Sarcopenia is defined as the loss of skeletal muscle mass, which occurs with aging.14 This relative deconditioning is pervasive and distinct from that which is seen in a persistently ill population. Sarcopenia associated with chronic medical conditions is an insidious finding that may be missed on subjective evaluation. In addition, it is often noted in obese cohorts and has been given the title of osteosarcopenic obesity.3 The loss of muscle quality and quantity has been associated with an increased risk of disability.15 It has also been shown to be a predictor of postoperative morbidity in numerous surgical populations.5,16,17 With the increasing emphasis on improving hospital outcomes and the well-defined correlation between sarcopenia and morbidity, defining the presence of sarcopenia will only become increasingly important in the future.
Table 3 e Univariate analysis identifying the association between discharge disposition and sarcopenia. Discharge disposition Time spent in ICU Length of stay Home without assistance
OR (CI) 0.94 (0.913-0.962)
P Value 0.0026 0.0007 <0.0001
Discharge to a facility
1.04 (1.015-1.068)
0.0018
90-d mortality
1.11 (1.026-1.206)
0.0094
OR ¼ odds ratio; CI ¼ confidence interval.
164
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 1 8 ( 2 2 1 ) 1 6 1 e1 6 6
Table 4 e Multivariable relationship between sarcopenia and the incidence of in-hospital complications. Complications
OR (CI)
P Value
CAUTI
1.06 (1.021-1.170)
0.0201
Reintubation
1.10 (1.028-1.180)
0.0062
CLABSI
1.09 (1.001-1.193)
0.0485
Wound infection
1.12 (1.026-1.221)
0.0111
OR ¼ Odds ratio; CI ¼ confidence interval.
It is clear that sarcopenia is independently associated with poor outcomes in surgical cohorts. It has also been identified as a risk factor for adverse discharge disposition in ICU patients.18 To date, there are little data pertaining to the relevance of this finding in a trauma population. The Canadian Study of Health and Aging Frailty Index and a modified frailty index have been used to identify patients at risk for inhospital complications.19,20 Frailty, as defined by the presence of sarcopenia and osteopenia, has been shown to be associated with traumatic thoracic and spine injuries.21 Muscle atrophy of the vastus lateralis was shown to be more prominent in patients who had suffered fall-related hip fractures versus healthy young or age-matched healthy controls.22 Owing to the relative void in the literature, this project was designed to investigate the effects of sarcopenia in a trauma population as it relates to clinical outcomes. Furthermore, given the prominent association between sarcopenia and morbidity in surgical cohorts, this study sought to create a predictive length-of-stay model in this unique population incorporating patient characteristics in addition to sarcopenia for a lengthof-stay calculator. In this cohort, patients with sarcopenia were noted to be older and suffer from more medical comorbidities (Table 1). The loss of lean muscle mass correlated with increased overall length of stay as well as time spent in the ICU. The presence of sarcopenia was associated with adverse discharge disposition and, perhaps most importantly, was predictive of mortality (Table 3). In line with the current literature, this population was more likely to be obese than nonsarcopenic individuals. These finding are significant in that sarcopenia is an occult process that is challenging to recognize on physical examination. This disease is not restricted to a frail, cachectic patient but rather can be identified in patients who appear vigorous and healthy. In the absence of an objective measurement for this disease, the identification of an at-risk population and a potential opportunity to intervene would be missed. Using readily available data in routinely ordered CT
Table 5 e Admission factors associated with length of stay on multivariable analysis. Variable
Parameter estimate
P Value
Sarcopenia
0.152
0.0038
Maximum AIS score
þ3.085
<0.0001
Admission WBC count
þ0.249
<0.0001
Congestive heart failure
þ5.971
0.0002
Hypertension
þ1.936
0.0027
Fig. 1 e Ability of UIHC model to predict 10 d length of stay (blue line e 0.79) versus max AIS D sarcopenia (red line e 0.76) and max AIS (brown line e 0.75).
imaging, sarcopenia can be identified, and this group flagged early in the hospital course. Multivariable analysis of the correlation between sarcopenia and the incidence of complications identified CAUTI, the need for reintubation, CLABSI, and wound infections as all being independently related to decreased lean mass. Previous studies have found that sarcopenia is more important than chronological age when considering perioperative morbidity. Englesbe et al.23 developed the theory of morphometric age, showing that it was able to more accurately predict the length of stay and postoperative mortality than chronological age. While novel in that there is a paucity of data on sarcopenia in a trauma setting, this was not altogether unexpected, given the available information in operative and ICU populations. Given this, the relationship between sarcopenia and length of stay was evaluated. On linear regression of the length of stay, it was noted that there were five statistically significant, independently predictive variables including sarcopenia, maximum AIS score, admission WBC count, and the presence or absence of congestive heart failure and hypertension. Using this data, a novel length-of-stay calculator was created using a random sampling of half of the cohort. This calculator was then validated for prognostic power using the entire population by means of an ROC analysis. The UIHC calculator was found to be more predictive than max AIS and max AIS with the addition of sarcopenia at 5-, 10- and 20-d hospital stays (Figs. 1-3). As the length of hospitalization increased, the prognostic capabilities decreased, while still remaining more accurate than max AIS. This is significant in that trauma patients with prolonged lengths of stay can be fiscally challenging for hospital systems. A retrospective review from a single academic institution found that its trauma center posted financial losses in caring for over half of patients with
deandrade et al sarcopenia predicts outcomes in trauma population
165
outcomes. The significance of these two findings in the ability to risk stratify for complications and also with discharge planning cannot be overstated. There are limitations to this research. This is a single institution, retrospective review of a trauma service with predominantly nonpenetrating admissions. Institutions with a preponderance of penetrating trauma with the attendant change in demographic make-up may find this calculator to be less accurate. Furthermore, in-hospital mortalities with a concomitant max AIS score of 6 were excluded, as those patients all sustained nonsurvivable injuries. This was done because the mortality of this population was likely secondary to severity of injury and less influenced by patient comorbidities including sarcopenia. Future directions of study would include a prospective assessment of the role of sarcopenia and the accuracy of this calculator across multiple centers.
Conclusion Fig. 2 e Ability of UIHC model to predict 5 d length of stay (blue line e 0.79) versus Max AIS D Sarcopenia (red e 0.76) and Max AIS (brown e 0.75).
hospitalizations of greater than 50 d.24 Targeting surgical patients with sarcopenia, early in admission with clinical resources, has been shown to not only reduce costs but also to shorten hospital lengths of stay.25 The UIHC calculator provides the traumatologist with an instrument that more accurately predicts the length of stay than any currently available tool. The identification of sarcopenia in a trauma patient also categorizes that person for an increased possibility of adverse
Sarcopenia is an insidious disease, which is difficult to identify on subjective examination alone. The loss of lean muscle is associated with multiple medical comorbidities and impaired perioperative outcomes. In a trauma population, it is an independent risk factor for adverse outcomes and increased length of stay. Incorporating an objective evaluation for lean mass, the UIHC calculator can accurately predict the length of stay. This allows for the identification of patients for early, targeted therapy and appropriate discharge planning.
Acknowledgment Author contributions De Andrade was responsible for data collection and manuscript drafting; Pedersen performed the data evaluation and study design; Garcia was responsible for study design and manuscript editing; Nau was responsible for data collection, study design, manuscript drafting and editing.
Disclosure The authors have no conflicts of interest to disclose.
references
Fig. 3 e Ability of UIHC model to predict 20 d length of stay (blue line e 0.78) versus Max AIS D Sarcopenia (red e 0.76) and Max AIS (brown e 0.72).
1. Cruz-Jentoft AJ, Baeyens JP, Bauer JM, et al. Sarcopenia: European consensus on definition and diagnosis: report of the European Working Group on sarcopenia in older people. Age Ageing. 2010;39:412e423. 2. Chung JY, Kang HT, Lee DC, Lee HR, Lee YJ. Body composition and its association with cardiometabolic risk factors in the elderly: a focus on sarcopenic obesity. Arch Gerontol Geriatr. 2013;56:270e278. 3. Ormsbee MJ, Prado CM, Ilich JZ, et al. Osteosarcopenic obesity: the role of bone, muscle, and fat on health. J Cachexia Sarcopenia Muscle. 2014;5:183e192.
166
j o u r n a l o f s u r g i c a l r e s e a r c h j a n u a r y 2 0 1 8 ( 2 2 1 ) 1 6 1 e1 6 6
4. Prado CM, Wells JC, Smith SR, Stephan BC, Siervo M. Sarcopenic obesity: a critical appraisal of the current evidence. Clin Nutr. 2012;31:583e601. 5. Joglekar S, Nau PN, Mezhir JJ. The impact of sarcopenia on survival and complications in surgical oncology: a review of the current literature-author response. J Surg Oncol. 2015;112:910. 6. Sheetz KH, Waits SA, Terjimanian MN, et al. Cost of major surgery in the sarcopenic patient. J Am Coll Surg. 2013;217:813e818. 7. Moisey LL, Mourtzakis M, Cotton BA, et al. Skeletal muscle predicts ventilator-free days, ICU-free days, and mortality in elderly ICU patients. Crit Care. 2013;17:R206. 8. Fairchild B, Webb TP, Xiang Q, et al. Sarcopenia and frailty in elderly trauma patients. World J Surg. 2015;39:373e379. 9. Salim A, Sangthong B, Martin M, Brown C, Plurad D, Demetriades D. Whole body imaging in blunt multisystem trauma patients without obvious signs of injury: results of a prospective study. Arch Surg. 2006;141:468e473. discussion 473-475. 10. Rieger M, Czermak B, El Attal R, Sumann G, Jaschke W, Freund M. Initial clinical experience with a 64-MDCT wholebody scanner in an emergency department: better time management and diagnostic quality? J Trauma. 2009;66:648e657. 11. Peng P, Hyder O, Firoozmand A, et al. Impact of sarcopenia on outcomes following resection of pancreatic adenocarcinoma. J Gastrointest Surg. 2012;16:1478e1486. 12. House MG, Fong Y, Arnaoutakis DJ, et al. Preoperative predictors for complications after pancreaticoduodenectomy: impact of BMI and body fat distribution. J Gastrointest Surg. 2008;12:270e278. 13. Boutin RD, Yao L, Canter RJ, Lenchik L. Sarcopenia: current concepts and imaging implications. AJR Am J Roentgenol. 2015;205:W255eW266. 14. Englesbe MJ, Lee JS, He K, et al. Analytic morphomics, core muscle size, and surgical outcomes. Ann Surg. 2012;256:255e261.
15. Janssen I, Baumgartner RN, Ross R, Rosenberg IH, Roubenoff R. Skeletal muscle cutpoints associated with elevated physical disability risk in older men and women. Am J Epidemiol. 2004;159:413e421. 16. Joglekar S, Asghar A, Mott SL, et al. Sarcopenia is an independent predictor of complications following pancreatectomy for adenocarcinoma. J Surg Oncol. 2015;111:771e775. 17. Englesbe MJ, Patel SP, He K, et al. Sarcopenia and mortality after liver transplantation. J Am Coll Surg. 2010;211:271e278. 18. Mueller N, Murthy S, Tainter CR, et al. Can sarcopenia quantified by ultrasound of the rectus femoris muscle predict adverse outcome of surgical intensive care unit patients as well as frailty? A prospective, observational cohort study. Ann Surg. 2016;264:1116e1124. 19. Ali R, Schwalb JM, Nerenz DR, Antoine HJ, Rubinfeld I. Use of the modified frailty index to predict 30-day morbidity and mortality from spine surgery. J Neurosurg Spine. 2016;25:537e541. 20. Rockwood K, Song X, MacKnight C, et al. A global clinical measure of fitness and frailty in elderly people. CMAJ. 2005;173:489e495. 21. Oskutis MQ, Lauerman MH, Kufera JA, et al. Are frailty markers associated with serious thoracic and spinal injuries among motor vehicle crash occupants? J Trauma Acute Care Surg. 2016;81:156e161. 22. Kramer IF, Snijders T, Smeets JS, et al. Extensive type II muscle fiber atrophy in elderly female hip fracture patients. J Gerontol A Biol Sci Med Sci. 2017. http://dx.doi.org/10.1093/ gerona/glw253 [Epub ahead of print]. 23. Englesbe MJ, Terjimanian MN, Lee JS, et al. Morphometric age and surgical risk. J Am Coll Surg. 2013;216:976e985. 24. Fakhry SM, Couillard D, Liddy CT, Adams D, Norcross ED. Trauma center finances and length of stay: identifying a profitability inflection point. J Am Coll Surg. 2010;210:817e821, 821-823. 25. Friedman J, Lussiez A, Sullivan J, Wang S, Englesbe M. Implications of sarcopenia in major surgery. Nutr Clin Pract. 2015;30:175e179.