Muscle loss phenotype predicts poor postoperative outcomes of colorectal cancer in US inpatients: a population-based analysis

Ko-Chao Lee; Sin-Ei Juang; Kuen-Lin Wu; Kung-Chuan Cheng; Ling-Chiao Song; Chien-En Tang; Hong-Hwa Chen; Kuan-Chih Chung

doi:10.3393/ac.2025.00129.0018

Articles

Page Path: HOME > Ann Coloproctol > Volume 41(5); 2025 > Article

Original Article Complications Muscle loss phenotype predicts poor postoperative outcomes of colorectal cancer in US inpatients: a population-based analysis: Ko-Chao Lee^1,2,3, Sin-Ei Juang⁴, Kuen-Lin Wu¹, Kung-Chuan Cheng¹, Ling-Chiao Song⁵, Chien-En Tang¹, Hong-Hwa Chen¹, Kuan-Chih Chung⁴; Annals of Coloproctology 2025;41(5):443-452.
DOI: https://doi.org/10.3393/ac.2025.00129.0018
Published online: October 24, 2025

¹Division of Colorectal Surgery, Department of Surgery, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung, Taiwan

²Division of Colorectal Surgery, Department of Surgery, Kaohsiung Municipal Fong Shan Hospital (under the management of Chang Gung Medical Foundation), Taiwan

³Division of Colorectal Surgery, Department of Surgery, Kaohsiung Municipal Ta-Tung Hospital, Taiwan

⁴Department of Anesthesiology, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung, Taiwan

⁵Division of Colon & Rectal Surgery, Department of Surgery, E-DA Hospital, I-Shou University, Kaohsiung, Taiwan

Correspondence to: Kuan-Chih Chung, MD Department of Anesthesiology, Kaohsiung Chang Gung Memorial Hospital, Chang Gung University College of Medicine, Kaohsiung 83301, Taiwan Email: s21096@ms24.hinet.net

• Received: February 6, 2025 • Revised: May 7, 2025 • Accepted: May 18, 2025

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

prev next

584 Views
15 Download

Full Article

Download PDF

Abstract
INTRODUCTION
METHODS
RESULTS
DISCUSSION
ARTICLE INFORMATION
SUPPLEMENTARY MATERIALS
REFERENCES

Abstract

Purpose
Muscle loss may lead to reduced therapy tolerance and survival. We aimed to assess whether colorectal cancer (CRC) patients with a muscle loss phenotype experience worse outcomes.
Methods
Data were extracted from the US Nationwide Inpatient Sample for hospitalized patients aged ≥20 years who underwent surgical resection for colorectal cancer (CRC) between 2005 and 2018. CRC and muscle loss phenotypes were identified using validated International Classification of Diseases (ICD) diagnosis and procedure codes. Propensity score matching was performed to balance characteristics. Regression analyses determined associations between muscle loss and in-hospital outcomes.
Results
A total of 209,171 patients were included, with a mean age of 67.9 years; 7.1% exhibited muscle loss phenotype. After matching, 60,295 patients remained in the sample. After adjustment, patients with muscle loss had significantly increased risks of postoperative complications (adjusted odds ratio [aOR], 2.99; 95% confidence interval [CI], 2.85–3.15), unfavorable discharge (aOR, 2.42; 95% CI, 2.30–2.53), prolonged length of stay (aOR, 4.34; 95% CI, 4.13–4.55), and higher total hospital costs (adjusted β, 70.86; 95% CI, 67.11–74.61) compared to patients without muscle loss. When stratified by age (≥65 years), results remained consistent. Among complications, muscle loss phenotype was most strongly associated with shock, sepsis, and respiratory failure.
Conclusion
Muscle loss phenotype among patients with CRC is strongly associated with poor postoperative outcomes, including higher complication rates, longer stays, and increased costs. These findings highlight the importance of preoperative muscle loss assessments and the necessity for targeted interventions.
Keywords: Muscular atrophy; Postoperative period; Colorectal neoplasms; Database

INTRODUCTION

Muscle loss is defined as a reduction in muscle mass resulting from various factors such as aging, obesity, chronic illnesses, and other medical conditions [1]. It is increasingly recognized as having significant clinical consequences, including physical impairment, reduced quality of life, decreased therapy tolerance, and shortened survival [2]. Additionally, muscle atrophy, coupled with systemic inflammation, has been associated with increased mortality in cancer patients [3–5]. Cancer itself can induce structural and functional alterations in skeletal muscle through increased collagen deposition, fat accumulation, and disruption of contractile elements, ultimately leading to diminished muscle quantity and strength [6, 7]. Furthermore, cancer treatments, such as chemotherapy and radiation therapy, can exacerbate muscle impairment [8–10].

Colorectal cancer (CRC) is the third most common and second most lethal cancer worldwide, responsible for 9.4% of cancer-related deaths in 2020 [10, 11]. Standard treatments for CRC typically include surgical resection, chemotherapy, and radiation therapy [12]. Recent studies suggest that patients with sarcopenia may have increased postoperative complications, prolonged hospitalizations, and higher healthcare costs compared to those without sarcopenia [13, 14]. However, current evidence remains inconclusive, and additional research is needed to fully elucidate the relationship between muscle loss and CRC outcomes.

To address this gap, the present study aimed to investigate the association between the muscle loss phenotype and surgical outcomes among CRC patients using a large US inpatient database.

METHODS

Ethics statement

This study sourced data from the Healthcare Cost and Utilization Project (HCUP) Central Distributor of the US National Institutes of Health in compliance with the US National (Nationwide) Inpatient Sample (NIS) data-use agreement. Given the anonymized nature of the data within the NIS database, the study was exempted from institutional review board approval by Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine.

Study design and data source

This retrospective study utilized data from the NIS, the largest continuous all-payer inpatient database in the United States. The NIS is managed by the HCUP and contains patient data, including primary and secondary diagnoses, procedures, demographics, hospital characteristics, and more. It includes about 8 million hospital stays per year, derived from approximately 1,050 hospitals across 44 states [15]. All admitted patients are considered for inclusion, representing a 20% stratified sample of US community hospitals.

Study population

The study population comprised adult inpatients aged ≥20 years, diagnosed with CRC who underwent CRC-directed surgical procedures between 2005 and 2018. Diagnoses and procedures were identified and confirmed using the International Classification of Diseases, 9th and 10th Revisions (ICD-9 and ICD-10) codes. Patients with metastatic disease, missing information on in-hospital outcomes, unknown sex, or missing sample weight values were excluded.

In-hospital outcomes

Outcomes measured included prolonged length of stay (LOS; defined as LOS ≥75th percentile for the study population), unfavorable discharge (defined as discharge to long-term care facilities), postoperative complications, and total hospital costs. Postoperative complications assessed were death during admission, acute myocardial infarction, cerebrovascular accident, venous thromboembolism, pneumonia, sepsis, infection, respiratory failure, mechanical ventilation, acute kidney injury, shock, bleeding, wound complications, device-related complication, and nervous and digestive system complications. ICD codes used to identify these conditions are detailed in Supplementary Table 1.

Definition of the muscle loss phenotype

We defined the muscle loss phenotype using a set of ICD-9 and ICD-10 diagnostic codes recorded in patients' medical records and diagnosed by physicians. This methodology for defining the muscle loss phenotype was proposed and validated by Attaway et al. [16]. Specifically, the muscle loss phenotype encompassed diagnostic codes indicative of muscle loss, including "other severe protein-calorie malnutrition," "malnutrition to a moderate degree," "other protein-calorie malnutrition," "unspecified protein-calorie malnutrition," "nutrition deficiency," "nutrition deficiency not otherwise specified," "other symptoms concerning nutrition metabolism and development," and "cachexia." Patients with any of these conditions were categorized as having the muscle loss phenotype.

Covariates

Patient baseline characteristics included age, sex, race/ethnicity, household income, and insurance status. Smoking status, tumor location, type of surgery (categorized as open and subtotal colectomy, laparoscopic, or robotic), and major comorbidities were identified using ICD codes. Major comorbidities analyzed included ischemic heart disease, congestive heart failure, atrial fibrillation, diabetes, anemia, hypertension, dyslipidemia, chronic obstructive pulmonary disease, cerebrovascular disease, obesity, severe liver disease, rheumatic disease, chronic kidney disease, and coagulopathy. Hospital-related characteristics, such as bed size, location/teaching status, and hospital region, were also extracted from the database.

Statistical analysis

All data were presented as numbers and weighted percentages (%) or means with standard error. Missing values were excluded from the calculation of percentages. Patients were divided into 2 groups: those with muscle loss (case group) and those without muscle loss (control group). Propensity score matching (PSM) was employed at a case-control ratio of 1:4 to ensure comparability between groups. The nearest neighbor matching method was used, prioritizing "best" matches and proceeding sequentially until no further matches were possible. Based on previous research identifying potential confounders affecting muscle mass, the matching process controlled for age, sex, tumor location, surgery type, comorbidities, and detailed hospital data [1, 17–19]. Multiple regression analyses were performed using PROC SURVEYLOGISTIC statements to evaluate associations between study variables and outcomes, excluding missing values. Results were reported as odds ratios (ORs) with 95% confidence intervals (CIs). Variables significantly different between groups (muscle loss vs. no muscle loss) after PSM (P<0.05) were adjusted for in the multivariable regression analysis. As the NIS database represents 20% of annual hospital admissions in the United States, nationwide estimates were derived using weighted samples (TRENDWT before 2011 and DISCWT from 2012 onward), along with stratum (NIS_STRATUM) and cluster (HOSPID) adjustments. Statistical significance was defined at P<0.05, and all P-values were 2-sided. Statistical analyses were performed using SAS ver. 9.4 (SAS Institute Inc).

RESULTS

Patient selection and matching

This study identified data from 292,117 patients aged ≥20 years who had CRC and underwent CRC-directed surgery. Of which, 82,946 patients were excluded due to metastatic disease, missing data on sex, mortality status, LOS, or missing sample weight values. Ultimately, 209,171 patients were included, representing an extrapolated total of 1,022,011 patients across the US population. According to our definition, 7.1% of these patients had the muscle loss phenotype. After performing PSM at a 1:4 ratio, the final analytic sample comprised 60,295 patients (Fig. 1).

Patient characteristics

Patient characteristics and in-hospital outcomes are presented in Tables 1, 2 and Supplementary Table 2. Before PSM, the mean age was 67.9 years; 106,689 (51.0%) were male, and 140,900 (76.3%) were White. Patients with muscle loss were older (72.4±0.12 years vs. 67.6±0.05 years), had a higher proportion of Black individuals (11.8% vs. 9.8%), lower household incomes (quartile 1: 28.6% vs. 25.1%; quartile 2: 27.1% vs. 26.5%), more frequent Medicare/Medicaid insurance coverage (78.4% vs. 62.7%), higher smoking rates (27.2% vs. 25.7%), and a greater likelihood of undergoing open and subtotal colectomy procedures (79.2% vs. 68.5%). Regarding comorbidities, patients with muscle loss had higher prevalence rates of most conditions, except diabetes, hypertension, and dyslipidemia. Additionally, patients with muscle loss had higher rates of emergent admissions (58.9% vs. 25.4%), hospitalizations in hospitals of small (15.0% vs. 13.9%) or medium bed sizes (27.9% vs. 25.9%), and admissions in the Midwest (26.0% vs. 23.1%) or West regions (20.7% vs. 19.4%; all P<0.001) (Supplementary Table 2).

After PSM, there were no significant differences in age, sex, tumor location, type of surgery, most comorbidities, or hospital-related characteristics between the 2 groups. However, differences existed in congestive heart failure (P<0.001), atrial fibrillation (P=0.018), and coagulopathy (P=0.002) (Table 1).

Regarding in-hospital outcomes before PSM, patients with the muscle loss phenotype had significantly higher rates of postoperative complications (78.1% vs. 54.7%), unfavorable discharge (42.7% vs. 13.0%), prolonged LOS (70.6% vs. 20.9%), and higher total hospital costs (US $162,500 vs. US $70,600; all P<0.001). After PSM, similar significant results persisted for these outcomes, except nervous system complications, which showed no significant difference (P=0.252) (Table 2).

Associations between the muscle loss phenotype and in-hospital outcomes

Associations between the muscle loss phenotype and in-hospital outcomes after PSM are summarized in Table 3. Compared with patients without the muscle loss phenotype, patients with the muscle loss phenotype had significantly increased risks of postoperative complications (adjusted OR [aOR], 2.99; 95% CI, 2.85–3.15), unfavorable discharge (aOR, 2.42; 95% CI, 2.30–2.53), prolonged LOS (aOR, 4.34; 95% CI, 4.13–4.55), and higher total hospital costs (adjusted β, 70.86; 95% CI, 67.11–74.61). After stratification by age (≥65 years), the muscle loss phenotype remained significantly associated with all outcomes, with these associations being more pronounced among younger patients (<65 years).

Associations between the muscle loss phenotype and specific complications

Associations between muscle loss and specific postoperative complications are detailed in Table 4. Patients with the muscle loss phenotype had significantly increased risks for all listed complications compared to patients without muscle loss (all P<0.05), except nervous system complications (P=0.280). Age-stratified analyses demonstrated findings similar to those in the overall population. Notably, the impact of the muscle loss phenotype appeared greater among younger patients. Specifically, younger patients with muscle loss had markedly higher risks of shock (aOR, 4.59; 95% CI, 3.91–5.38), sepsis (aOR, 4.28; 95% CI, 3.86–4.75), respiratory failure (aOR, 3.93; 95% CI, 3.46–4.47), and nervous system complications (aOR, 2.80; 95% CI, 1.07–7.32). For patients aged ≥65 years, significant associations were observed with shock (aOR, 3.19; 95% CI, 2.92–3.48), sepsis (aOR, 3.01; 95% CI, 2.83–3.20), and respiratory failure (aOR, 2.69; 95% CI, 2.53–2.86). However, there were no significant associations with cerebrovascular accident or nervous system complications in this older subgroup.

DISCUSSION

This study demonstrates significant and independent associations between muscle loss and adverse in-hospital outcomes in CRC patients undergoing surgery, including unfavorable discharge, prolonged LOS, increased total hospital costs, and postoperative complications. Patients with the muscle loss phenotype also had higher risks of acute myocardial infarction, cerebrovascular accident, venous thromboembolism, pneumonia, sepsis, infection, respiratory failure, mechanical ventilation, acute kidney injury, shock, bleeding, wound complications, device complications, nervous system and digestive system complications.

Our data revealed that approximately 1 in 14 patients with nonmetastatic CRC had a documented muscle loss phenotype. Sarcopenia, characterized by decreased muscle strength, function, quantity, and quality due to aging, is common among patients with advanced cancer, including CRC. Specifically, the prevalence of sarcopenia is significantly higher in CRC patients compared to age-matched healthy individuals [20]. In our analysis, age influenced postoperative complications among patients with the muscle loss phenotype. Notably, the strongest associations observed between muscle loss and postoperative complications were with shock, sepsis, and respiratory failure. Interestingly, when stratified by age at 65 years, the impact of muscle loss on postoperative complication risk was more pronounced in younger patients compared to older patients.

Despite advances in therapy, surgical resection remains the primary curative approach for CRC. However, surgical procedures impose additional metabolic demands on the body, including skeletal muscle, potentially exacerbating symptoms associated with cancer. Previous studies have indicated that both colon cancer and major surgeries are associated with skeletal muscle mass loss, increasing perioperative morbidity and mortality risk [21]. The present study confirms the link between muscle mass loss and poor postoperative outcomes in CRC patients, highlighting the importance of monitoring and addressing muscle loss to mitigate perioperative risks.

This study excluded metastatic CRC patients. Previous research has found low muscle mass in approximately 40% of metastatic CRC patients, a condition associated with a poor prognosis [22]. Body composition assessment through computed tomography imaging should become routine in clinical evaluations, as it is a superior predictor of surgical complications than body mass index [23]. Additionally, muscle loss is strongly associated with chemotherapy toxicity and could inform dosing based on lean body mass, particularly in the palliative setting. Studies have demonstrated that skeletal muscle loss occurring during palliative chemotherapy in metastatic CRC patients correlates with reduced survival [24]. Chemotherapy-related adverse effects and reduced physical activity during treatment may contribute to sarcopenia in CRC patients. Emerging evidence also suggests that overexpressed proapoptotic microRNAs from metastatic tissues may contribute to muscle wasting in metastatic CRC [25].

Low muscle mass and reduced aerobic function are frequently observed in colon cancer patients. Previous research has shown low preoperative pyruvate dehydrogenase activity in CRC patients normalizes post-surgery, suggesting muscle mass reduction results primarily from surgical intervention rather than from cancer itself [21]. Such reductions in mitochondrial enzyme activity may increase perioperative risks. Although mitochondrial enzyme activity was not measured in our study, we found CRC patients with muscle loss had increased postoperative complication risks compared to those without muscle loss, aligning with prior findings.

Chemotherapy can cause severe side effects such as weight loss, nausea, and vomiting in cancer patients [26]. Changes in body composition, particularly the loss of lean body mass, influence drug toxicity and prognosis. While chemotherapy dosing is typically based on body surface area, variability in lean and adipose tissue mass is not accounted for [27]. Research suggests that muscle wasting is linked to abnormal mitochondrial metabolism and reduced protein anabolism [28]. Anabolic interventions aimed at preserving muscle mass have shown promise in reducing chemotherapy-related toxicity. Accurate body composition assessment and preservation of muscle mass could enhance overall survival [29]. Muscle wasting is frequently exacerbated by anticancer treatments, including surgery, as demonstrated in previous and current studies.

Early identification and management of metabolic and nutritional changes in cancer patients are crucial, especially given the increased risk of complications and mortality associated with muscle loss in CRC patients. Targeted nutritional support, physical exercise, and pharmacological interventions might effectively manage these issues. Nonetheless, surgical complications and patient-related factors may negatively affect long-term survival. Muscle wasting in cancer patients leads to increased morbidity, treatment toxicity, diminished quality of life, and poor survival [30]. Recent research exploring the molecular mechanisms behind muscle wasting has identified potential therapeutic targets and promising pharmacological agents, although monotherapy is unlikely to succeed due to the multifactorial nature of muscle wasting [31]. Therefore, structured multimodal interventions are essential to prevent and treat cancer-related muscle wasting, necessitating further research in this area.

Our study found that muscle loss was associated with postoperative complications in both younger and older CRC patients. In the younger patient group, muscle loss correlated with higher risks of shock, sepsis, respiratory failure, and nervous system complications. In older patients, muscle loss similarly increased risks for shock, sepsis, and respiratory failure. However, no significant associations were observed between muscle loss and cerebrovascular accident or nervous system complications in older patients. These differences may reflect established compensatory mechanisms in older adults, which could partially mitigate the adverse impacts of muscle loss [32].

Strengths and limitations

This study has several notable strengths. First, it utilized the NIS database, providing a robust and comprehensive dataset that encompassed over a decade of patient information. The large sample size ensured high statistical power and generalizability of the findings. Second, the study applied PSM to balance baseline characteristics between patients with and without the muscle loss phenotype, reducing potential confounding and enhancing result reliability. Additionally, the detailed analysis of various postoperative outcomes offers a thorough understanding of the clinical impact of the muscle loss phenotype among CRC patients. Nevertheless, the study has several limitations. Its retrospective nature inherently provides a lower level of evidence compared to prospective studies. Moreover, reliance on ICD coding systems may introduce biases due to potential coding errors. The lack of detailed clinical data, such as tumor stage, grade, and histological subtype, may have introduced bias into the analysis. The study did not account for metabolic factors associated with CRC and muscle loss and lacked extended postdischarge follow-up, analyzing only data obtained during hospitalization. Consequently, we could not assess long-term postoperative outcomes, such as 5-year survival, which warrants further investigation in future research. Additionally, younger patients with the muscle loss phenotype might exhibit unfavorable lifestyle behaviors (e.g., poor dietary habits and physical inactivity), potentially negatively impacting cancer treatment outcomes and introducing selection bias into the study. Furthermore, inherent study design limitations prevent determining definitively whether muscle loss precedes and contributes to poor CRC outcomes or results from them. This ambiguity necessitates cautious interpretation of our findings and highlights the need for further research to clarify causal relationships.

Conclusions

In US patients undergoing surgery for CRC, the presence of the muscle loss phenotype is independently associated with poorer in-hospital outcomes. This finding emphasizes the importance of preoperative assessments for muscle loss and suggests the necessity of targeted interventions in clinical practice to improve patient outcomes.

ARTICLE INFORMATION

Conflict of interest

No potential conflict of interest relevant to this article was reported.

Funding

None.

Author contributions

Conceptualization: KCL, SEJ, KLW, KC Cheng, LCS, CET; Data curation: KCL, SEJ, KLW, KC Cheng, LCS, CET, HHC; Formal analysis: KCL, SEJ, KLW, KC Cheng, LCS, CET, HHC, KC Chung; Investigation: KCL, SEJ, KLW, KC Cheng, LCS, CET; Methodology: KCL, SEJ, KLW, KC Cheng, LCS, CET; Supervision: KCL, SEJ; Writing–original draft: KCL, SEJ, KC Chung; Writing–review & editing: all authors. All authors read and approved the final manuscript.

SUPPLEMENTARY MATERIALS

Supplementary Table 1.

ICD codes used in the study

ac-2025-00129-0018-Supplementary-Table-1.pdf

Supplementary Table 2.

Characteristics of the study population before propensity score matching

ac-2025-00129-0018-Supplementary-Table-2.pdf

Supplementary materials are available from https://doi.org/10.3393/ac.2025.00129.0018.

Fig. 1.

Flowchart of the patient selection. HCUP, Healthcare Cost and Utilization Project; NIS, US National (Nationwide) Inpatient Sample; CRC, colorectal cancer; PSM, propensity score matching. Representing an extrapolated total of ^a1,022,011 and ^b294,566 patients across the US population.

Table 1.

Characteristics of the study population after propensity score matching

Characteristic	Total (n=60,295)	Muscle loss phenotype		P-value	SMD
Characteristic	Total (n=60,295)	Yes (n=12,059)	No (n=48,236)	P-value	SMD
Age (yr)	72.0±0.1	71.9±0.1	72.0±0.1	0.933	0.002
20–39	908 (1.5)	178 (1.5)	730 (1.5)	0.859	0.006
40–64	15,202 (25.2)	3,061 (25.4)	12,141 (25.2)
≥65	44,185 (73.3)	8,820 (73.1)	35,365 (73.3)
Sex				0.820	0.002
Male	30,304 (50.3)	6,072 (50.4)	24,232 (50.2)
Female	29,991 (49.7)	5,987 (49.6)	24,004 (49.8)
Race/ethnicity				<0.001^*	0.084
White	40,532 (76.5)	8,176 (75.5)	32,356 (76.8)
Black	5,464 (10.3)	1,255 (11.6)	4,209 (10.0)
Hispanic	3,894 (7.4)	773 (7.1)	3,121 (7.4)
Asian or Pacific Islander	1,526 (2.9)	308 (2.8)	1,218 (2.9)
Other	1,558 (2.9)	323 (3.0)	1,235 (2.9)
Missing	7,321	1,224	6,097
Household income				<0.001^*	0.063
Quartile 1	15,862 (26.8)	3,371 (28.5)	12,491 (26.4)
Quartile 2	15,926 (26.9)	3,238 (27.4)	12,688 (26.8)
Quartile 3	14,822 (25.0)	2,927 (24.7)	11,895 (25.1)
Quartile 4	12,565 (21.2)	2,300 (19.4)	10,265 (21.7)
Missing	1,120	223	897
Insurance status				<0.001^*	0.083
Medicare/Medicaid	44,903 (74.6)	9,314 (77.4)	35,589 (73.9)
Private including HMO	12,828 (21.3)	2,250 (18.7)	10,578 (22.0)
Self-pay, no charge, other	2,493 (4.1)	477 (4.0)	2,016 (4.2)
Missing	71	18	53
Smoking status				0.292	0.011
No	44,617 (74.0)	8,878 (73.6)	35,739 (74.1)
Yes	15,678 (26.0)	3,181 (26.4)	12,497 (25.9)
Tumor location				0.793	0.007
Colon	47,098 (78.1)	9,447 (78.3)	37,651 (78.1)
Rectum	12,708 (21.1)	2,516 (20.9)	10,192 (21.1)
Colon and rectum	489 (0.8)	96 (0.8)	393 (0.8)
Type of surgery				0.068	0.021
Open and subtotal colectomy	47,610 (79.0)	9,434 (78.2)	38,176 (79.1)
Laparoscopic	10,341 (17.2)	2,153 (17.9)	8,188 (17.0)
Robotic	2,344 (3.9)	472 (3.9)	1,872 (4.0)
Comorbidity
Ischemic heart disease	11,563 (19.2)	2,374 (19.7)	9,189 (19.1)	0.112	0.018
Congestive heart failure	9,183 (15.2)	1,968 (16.3)	7,215 (15.0)	<0.001^*	0.036
Atrial fibrillation	10,516 (17.4)	2,191 (18.2)	8,325 (17.3)	0.018^*	0.022
Diabetes	13,194 (21.9)	2,690 (22.3)	10,504 (21.8)	0.207	0.013
Anemia	19,055 (31.6)	3,841 (31.9)	15,214 (31.5)	0.511	0.009
Hypertension	30,195 (50.1)	6,035 (50.0)	24,160 (50.1)	0.935	0.002
Dyslipidemia	16,018 (26.6)	3,263 (27.1)	12,755 (26.4)	0.171	0.016
Chronic obstructive pulmonary disease	11,394 (18.9)	2,331 (19.3)	9,063 (18.8)	0.175	0.014
Cerebrovascular disease	2,641 (4.4)	547 (4.5)	2,094 (4.3)	0.350	0.010
Obesity	9,816 (16.3)	1,998 (16.6)	7,818 (16.2)	0.337	0.010
Severe liver disease	534 (0.9)	119 (1.0)	415 (0.9)	0.185	0.015
Rheumatic disease	1,146 (1.9)	230 (1.9)	916 (1.9)	0.952	0.001
Chronic kidney disease	7,498 (12.4)	1,563 (13.0)	5,935 (12.3)	0.050	0.018
Coagulopathy	3,435 (5.7)	758 (6.3)	2,677 (5.5)	0.002^*	0.031
Hospital status
Emergent admission				0.074	0.019
No	26,417 (43.9)	5,374 (44.7)	21,043 (43.7)
Yes	33,762 (56.1)	6,660 (55.3)	27,102 (56.3)
Missing	116	25	91
Hospital bed size				0.900	0.013
Small	8,671 (14.4)	1,758 (14.6)	6,913 (14.4)
Medium	16,765 (27.9)	3,335 (27.7)	13,430 (27.9)
Large	34,726 (57.7)	6,938 (57.7)	27,788 (57.7)
Missing	133	28	105
Hospital location/teaching status				0.988	0.010
Rural	6,610 (11.0)	1,321 (11.0)	5,289 (11.0)
Urban nonteaching	20,483 (34.0)	4,109 (34.2)	16,374 (34.0)
Urban teaching	33,069 (55.0)	6,601 (54.9)	26,468 (55.0)
Missing	133	28	105
Hospital region				0.892	0.010
Northeast	10,641 (17.8)	2,107 (17.5)	8,534 (17.9)
Midwest	15,395 (25.6)	3,112 (25.9)	12,283 (25.6)
South	22,028 (36.4)	4,404 (36.5)	17,624 (36.4)
West	12,231 (20.1)	2,436 (20.1)	9,795 (20.1)

Values are presented as mean±standard error, number (weighted %), or number only. Weighted percentages were calculated after excluding missing values. Percentages may not total 100 because of rounding.

SMD, standardized mean difference; HMO, health maintenance organization.

Table 2.

Outcomes of the study population after propensity score matching

Outcome	Total (n=60,295)	Muscle loss phenotype		P-value	SMD
Outcome	Total (n=60,295)	Yes (n=12,059)	No (n=48,236)	P-value	SMD
Postoperative complication
Any	35,865 (59.5)	9,439 (78.3)	26,426 (54.8)	<0.001^*	0.512
Death	2,470 (4.1)	814 (6.8)	1,656 (3.4)	<0.001^*	0.152
Acute myocardial infarction	1,239 (2.1)	366 (3.0)	873 (1.8)	<0.001^*	0.081
Cerebrovascular accident	1,333 (2.2)	313 (2.6)	1,020 (2.1)	0.001^*	0.033
Venous thromboembolism	2,693 (4.5)	963 (8.0)	1,730 (3.6)	<0.001^*	0.184
Pneumonia	2,914 (4.8)	1,012 (8.4)	1,902 (3.9)	<0.001^*	0.183
Sepsis	7,654 (12.7)	3,105 (25.7)	4,549 (9.4)	<0.001^*	0.437
Infection	7,617 (12.6)	2,712 (22.5)	4,905 (10.2)	<0.001^*	0.336
Respiratory failure	6,903 (11.4)	2,620 (21.7)	4,283 (8.9)	<0.001^*	0.361
Mechanical ventilation	4,609 (7.6)	1,764 (14.6)	2,845 (5.9)	<0.001^*	0.290
Acute kidney injury	8,659 (14.4)	2,986 (24.8)	5,673 (11.8)	<0.001^*	0.339
Shock	3,024 (5.0)	1,343 (11.1)	1,681 (3.5)	<0.001^*	0.298
Bleeding	21,532 (35.7)	5,783 (48.0)	15,749 (32.6)	<0.001^*	0.316
Wound complication	3,839 (6.4)	1,443 (12.0)	2,396 (5.0)	<0.001^*	0.253
Device-related complication	1,610 (2.7)	490 (4.1)	1,120 (2.3)	<0.001^*	0.101
Nervous system complication	129 (0.2)	31 (0.3)	98 (0.2)	0.252	0.011
Digestive system complication	11,512 (19.1)	3,716 (30.8)	7,796 (16.2)	<0.001^*	0.350
Unfavorable discharge^a	15,613 (27.0)	4,648 (41.3)	10,965 (23.6)	<0.001^*	0.346
Prolonged LOS^a,b	24,869 (42.9)	7,903 (70.1)	16,966 (36.3)	<0.001^*	0.719
Total hospital cost (per US $1,000 dollars)	102.5±0.7	160.6±1.9	87.9±0.6	<0.001^*	0.236

Values are presented as number (%) or mean±standard error.

SMD, standardized mean difference; LOS, length of stay.

^aPatients with in-hospital mortality were excluded. ^bHospital LOS >75th percentile (10 days).

*P<0.05.

Table 3.

Associations between the muscle loss phenotype (yes vs. no) and in-hospital outcomes after propensity score matching stratified by age

Age group	Any postoperative complication		Unfavorable discharge^a		Prolonged LOS^a,b		Total hospital cost (per US $1,000)
Age group	aOR (95% CI)	P-value	aOR (95% CI)	P-value	aOR (95% CI)	P-value	Adjusted β (95% CI)	P-value
All	2.99 (2.85–3.15)	<0.001^*	2.42 (2.30–2.53)	<0.001^*	4.34 (4.13–4.55)	<0.001^*	70.86 (67.11–74.61)	<0.001^*
<65 yr	3.55 (3.25–3.88)	<0.001^*	3.65 (3.22–4.14)	<0.001^*	5.35 (4.90–5.85)	<0.001^*	89.08 (86.52–91.64)	<0.001^*
≥65 yr	2.80 (2.64–2.97)	<0.001^*	2.39 (2.27–2.52)	<0.001^*	3.99 (3.77–4.22)	<0.001^*	62.96 (59.48–66.43)	<0.001^*

In the multivariable analysis, models were adjusted for race, primary payer, household income, and comorbidities (congestive heart failure, atrial fibrillation, and coagulopathy).

LOS, length of stay; aOR, adjusted odds ratio; CI, confidence interval.

^aExcluded patients with in-hospital mortality. ^bHospital LOS >75th percentile (10 days).

*P<0.05.

Table 4.

Associations between the muscle loss phenotype (yes vs. no) and specific postoperative complications

Postoperative complication	Age group
	All		<65 yr		≥65 yr
	aOR (95% CI)	P-value	aOR (95% CI)	P-value	aOR (95% CI)	P-value
Death	2.00 (1.83–2.19)	<0.001^*	2.17 (1.68–2.81)	<0.001^*	2.00 (1.82–2.20)	<0.001^*
Acute myocardial infarction	1.67 (1.47–1.90)	<0.001^*	1.69 (1.17–2.43)	0.005^*	1.68 (1.47–1.92)	<0.001^*
Cerebrovascular accident	1.19 (1.05–1.36)	0.008^*	1.69 (1.20–2.40)	0.003^*	1.14 (0.99–1.31)	0.067
Venous thromboembolism	2.30 (2.11–2.50)	<0.001^*	2.69 (2.27–3.20)	<0.001^*	2.19 (1.99–2.41)	<0.001^*
Pneumonia	2.19 (2.02–2.38)	<0.001^*	2.75 (2.28–3.32)	<0.001^*	2.11 (1.93–2.30)	<0.001^*
Sepsis	3.31 (3.14–3.49)	<0.001^*	4.28 (3.86–4.75)	<0.001^*	3.01 (2.83–3.20)	<0.001^*
Infection	2.53 (2.40–2.67)	<0.001^*	2.73 (2.45–3.04)	<0.001^*	2.46 (2.31–2.62)	<0.001^*
Respiratory failure	2.89 (2.73–3.05)	<0.001^*	3.93 (3.46–4.47)	<0.001^*	2.69 (2.53–2.86)	<0.001^*
Mechanical ventilation	2.73 (2.56–2.91)	<0.001^*	3.75 (3.25–4.34)	<0.001^*	2.50 (2.33–2.69)	<0.001^*
Acute kidney injury	2.45 (2.33–2.59)	<0.001^*	3.35 (2.99–3.76)	<0.001^*	2.28 (2.15–2.42)	<0.001^*
Shock	3.49 (3.23–3.77)	<0.001^*	4.59 (3.91–5.38)	<0.001^*	3.19 (2.92–3.48)	<0.001^*
Bleeding	1.88 (1.80–1.97)	<0.001^*	2.24 (2.06–2.43)	<0.001^*	1.79 (1.70–1.88)	<0.001^*
Wound complication	2.58 (2.40–2.76)	<0.001^*	3.00 (2.64–3.42)	<0.001^*	2.40 (2.21–2.61)	<0.001^*
Device-related complication	1.78 (1.60–1.99)	<0.001^*	1.76 (1.42–2.18)	<0.001^*	1.79 (1.58–2.03)	<0.001^*
Nervous system complication	1.25 (0.83–1.88)	0.280	2.80 (1.07–7.32)	0.036^*	1.10 (0.70–1.72)	0.695
Digestive system complication	2.30 (2.19–2.41)	<0.001^*	2.86 (2.62–3.13)	<0.001^*	2.12 (2.00–2.24)	<0.001^*

In multivariable analysis, models were adjusted for race, primary payer, household income, and comorbidities (congestive heart failure, atrial fibrillation, and coagulopathy).

aOR, adjusted odds ratio; CI, confidence interval.

*P<0.05.

REFERENCES

1. Kalyani RR, Corriere M, Ferrucci L. Age-related and disease-related muscle loss: the effect of diabetes, obesity, and other diseases. Lancet Diabetes Endocrinol 2014;2:819–29. Article PubMed PMC
2. Wan Q, Yuan Q, Zhao R, Shen X, Chen Y, Li T, et al. Prognostic value of cachexia index in patients with colorectal cancer: a retrospective study. Front Oncol 2022;12:984459.Article PubMed PMC
3. Fearon KC, Glass DJ, Guttridge DC. Cancer cachexia: mediators, signaling, and metabolic pathways. Cell Metab 2012;16:153–66. Article PubMed
4. Melstrom LG, Melstrom KA Jr, Ding XZ, Adrian TE. Mechanisms of skeletal muscle degradation and its therapy in cancer cachexia. Histol Histopathol 2007;22:805–14. Article PubMed
5. van der Werf A, van Bokhorst QN, de van der Schueren MA, Verheul HM, Langius JA. Cancer cachexia: identification by clinical assessment versus international consensus criteria in patients with metastatic colorectal cancer. Nutr Cancer 2018;70:1322–9. Article PubMed
6. Huot JR, Pin F, Bonetto A. Muscle weakness caused by cancer and chemotherapy is associated with loss of motor unit connectivity. Am J Cancer Res 2021;11:2990–3001. PubMed PMC
7. Martin A, Freyssenet D. Phenotypic features of cancer cachexia-related loss of skeletal muscle mass and function: lessons from human and animal studies. J Cachexia Sarcopenia Muscle 2021;12:252–73. Article PubMed PMC PDF
8. Courneya KS. Exercise in cancer survivors: an overview of research. Med Sci Sports Exerc 2003;35:1846–52. Article PubMed
9. Dimeo F. Exercise for cancer patients: a new challenge in sports medicine. West J Med 2000;173:272–3. Article PubMed PMC
10. Kader YA, El-Nahas T, Sakr A. Adjuvant chemotherapy for luminal A breast cancer: a prospective study comparing two popular chemotherapy regimens. Onco Targets Ther 2013;6:1073–7. Article PubMed PMC
11. Hossain MS, Karuniawati H, Jairoun AA, Urbi Z, Ooi J, John A, et al. Colorectal cancer: a review of carcinogenesis, global epidemiology, current challenges, risk factors, preventive and treatment strategies. Cancers (Basel) 2022;14:1732.Article PubMed PMC
12. Krasteva N, Georgieva M. Promising therapeutic strategies for colorectal cancer treatment based on nanomaterials. Pharmaceutics 2022;14:1213.Article PubMed PMC
13. Miyamoto Y, Baba Y, Sakamoto Y, Ohuchi M, Tokunaga R, Kurashige J, et al. Negative impact of skeletal muscle loss after systemic chemotherapy in patients with unresectable colorectal cancer. PLoS One 2015;10:e0129742. Article PubMed PMC
14. Zhang Y, Zhu Y. Development and validation of risk prediction model for sarcopenia in patients with colorectal cancer. Front Oncol 2023;13:1172096.Article PubMed PMC
15. International Diabetes Federation (IDF). Facts & figures [Internet]. IDF; [cited 2025 Apr 14]. Available from: https://idf.org/aboutdiabetes/diabetes-facts-figures/
16. Attaway AH, Welch N, Hatipoğlu U, Zein JG, Dasarathy S. Muscle loss contributes to higher morbidity and mortality in COPD: an analysis of national trends. Respirology 2021;26:62–71. Article PubMed PDF
17. Silva AM, Shen W, Heo M, Gallagher D, Wang Z, Sardinha LB, et al. Ethnicity-related skeletal muscle differences across the lifespan. Am J Hum Biol 2010;22:76–82. Article PubMed PMC
18. Volpi E, Nazemi R, Fujita S. Muscle tissue changes with aging. Curr Opin Clin Nutr Metab Care 2004;7:405–10. Article PubMed PMC
19. Wan H, Hu YH, Li WP, Wang Q, Su H, Chenshu JY, et al. Quality of life, household income, and dietary habits are associated with the risk of sarcopenia among the Chinese elderly. Aging Clin Exp Res 2024;36:29.Article PubMed PMC PDF
20. Vergara-Fernandez O, Trejo-Avila M, Salgado-Nesme N. Sarcopenia in patients with colorectal cancer: a comprehensive review. World J Clin Cases 2020;8:1188–202. Article PubMed PMC
21. Phillips BE, Smith K, Liptrot S, Atherton PJ, Varadhan K, Rennie MJ, et al. Effect of colon cancer and surgical resection on skeletal muscle mitochondrial enzyme activity in colon cancer patients: a pilot study. J Cachexia Sarcopenia Muscle 2013;4:71–7. Article PubMed PMC
22. Deng CY, Lin YC, Wu JS, Cheung YC, Fan CW, Yeh KY, et al. Progressive sarcopenia in patients with colorectal cancer predicts survival. AJR Am J Roentgenol 2018;210:526–32. Article PubMed
23. Cespedes Feliciano E, Chen WY. Clinical implications of low skeletal muscle mass in early-stage breast and colorectal cancer. Proc Nutr Soc 2018;77:382–7. Article PubMed PMC
24. Park SE, Choi JH, Park JY, Kim BJ, Kim JG, Kim JW, et al. Loss of skeletal muscle mass during palliative chemotherapy is a poor prognostic factor in patients with advanced gastric cancer. Sci Rep 2020;10:17683.Article PubMed PMC PDF
25. Okugawa Y, Toiyama Y, Hur K, Yamamoto A, Yin C, Ide S, et al. Circulating miR-203 derived from metastatic tissues promotes myopenia in colorectal cancer patients. J Cachexia Sarcopenia Muscle 2019;10:536–48. Article PubMed PMC PDF
26. Eliasen A, Kornholt J, Mathiasen R, Wadt K, Stoltze U, Brok J, et al. Background sensitivity to chemotherapy-induced nausea and vomiting and response to antiemetics in paediatric patients: a genetic association study. Pharmacogenet Genomics 2022;32:72–8. Article PubMed
27. Herrstedt J, Lindberg S, Petersen PC. Prevention of chemotherapy-induced nausea and vomiting in the older patient: optimizing outcomes. Drugs Aging 2022;39:1–21. Article PubMed PMC PDF
28. Bonaldo P, Sandri M. Cellular and molecular mechanisms of muscle atrophy. Dis Model Mech 2013;6:25–39. Article PubMed PMC PDF
29. Onishi S, Tajika M, Tanaka T, Yamada K, Kamiya T, Abe T, et al. Effect of body composition change during neoadjuvant chemotherapy for esophageal squamous cell carcinoma. J Clin Med 2022;11:508.Article PubMed PMC
30. Fan Y, Yao Q, Liu Y, Jia T, Zhang J, Jiang E. Underlying causes and co-existence of malnutrition and infections: an exceedingly common death risk in cancer. Front Nutr 2022;9:814095.Article PubMed PMC
31. Clamon G, Byrne MM, Talbert EE. Inflammation as a therapeutic target in cancer cachexia. Cancers (Basel) 2022;14:5262.Article PubMed PMC
32. Jha SK. Compensatory cognition in neurological diseases and aging: a review of animal and human studies. Aging Brain 2022;3:100061.Article PubMed PMC

Figure & Data

References

Citations

Citations to this article as recorded by

Cite this Article

Cite this Article: export Copy Download Format; Close

Download Citation

Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.

Format:

RIS — For EndNote, ProCite, RefWorks, and most other reference management software
BibTeX — For JabRef, BibDesk, and other BibTeX-specific software

Include:

Citation for the content below
Citation and abstract for the content below

Muscle loss phenotype predicts poor postoperative outcomes of colorectal cancer in US inpatients: a population-based analysis

Ann Coloproctol. 2025;41(5):443-452. Published online October 24, 2025

DOI: https://doi.org/10.3393/ac.2025.00129.0018

XML Download

Figure

Muscle loss phenotype predicts poor postoperative outcomes of colorectal cancer in US inpatients: a population-based analysis

Fig. 1. Flowchart of the patient selection. HCUP, Healthcare Cost and Utilization Project; NIS, US National (Nationwide) Inpatient Sample; CRC, colorectal cancer; PSM, propensity score matching. Representing an extrapolated total of a1,022,011 and b294,566 patients across the US population.

Fig. 1.

Muscle loss phenotype predicts poor postoperative outcomes of colorectal cancer in US inpatients: a population-based analysis

Characteristic	Total (n=60,295)	Muscle loss phenotype		P-value	SMD
Characteristic	Total (n=60,295)	Yes (n=12,059)	No (n=48,236)	P-value	SMD
Age (yr)	72.0±0.1	71.9±0.1	72.0±0.1	0.933	0.002
20–39	908 (1.5)	178 (1.5)	730 (1.5)	0.859	0.006
40–64	15,202 (25.2)	3,061 (25.4)	12,141 (25.2)
≥65	44,185 (73.3)	8,820 (73.1)	35,365 (73.3)
Sex				0.820	0.002
Male	30,304 (50.3)	6,072 (50.4)	24,232 (50.2)
Female	29,991 (49.7)	5,987 (49.6)	24,004 (49.8)
Race/ethnicity				<0.001^*	0.084
White	40,532 (76.5)	8,176 (75.5)	32,356 (76.8)
Black	5,464 (10.3)	1,255 (11.6)	4,209 (10.0)
Hispanic	3,894 (7.4)	773 (7.1)	3,121 (7.4)
Asian or Pacific Islander	1,526 (2.9)	308 (2.8)	1,218 (2.9)
Other	1,558 (2.9)	323 (3.0)	1,235 (2.9)
Missing	7,321	1,224	6,097
Household income				<0.001^*	0.063
Quartile 1	15,862 (26.8)	3,371 (28.5)	12,491 (26.4)
Quartile 2	15,926 (26.9)	3,238 (27.4)	12,688 (26.8)
Quartile 3	14,822 (25.0)	2,927 (24.7)	11,895 (25.1)
Quartile 4	12,565 (21.2)	2,300 (19.4)	10,265 (21.7)
Missing	1,120	223	897
Insurance status				<0.001^*	0.083
Medicare/Medicaid	44,903 (74.6)	9,314 (77.4)	35,589 (73.9)
Private including HMO	12,828 (21.3)	2,250 (18.7)	10,578 (22.0)
Self-pay, no charge, other	2,493 (4.1)	477 (4.0)	2,016 (4.2)
Missing	71	18	53
Smoking status				0.292	0.011
No	44,617 (74.0)	8,878 (73.6)	35,739 (74.1)
Yes	15,678 (26.0)	3,181 (26.4)	12,497 (25.9)
Tumor location				0.793	0.007
Colon	47,098 (78.1)	9,447 (78.3)	37,651 (78.1)
Rectum	12,708 (21.1)	2,516 (20.9)	10,192 (21.1)
Colon and rectum	489 (0.8)	96 (0.8)	393 (0.8)
Type of surgery				0.068	0.021
Open and subtotal colectomy	47,610 (79.0)	9,434 (78.2)	38,176 (79.1)
Laparoscopic	10,341 (17.2)	2,153 (17.9)	8,188 (17.0)
Robotic	2,344 (3.9)	472 (3.9)	1,872 (4.0)
Comorbidity
Ischemic heart disease	11,563 (19.2)	2,374 (19.7)	9,189 (19.1)	0.112	0.018
Congestive heart failure	9,183 (15.2)	1,968 (16.3)	7,215 (15.0)	<0.001^*	0.036
Atrial fibrillation	10,516 (17.4)	2,191 (18.2)	8,325 (17.3)	0.018^*	0.022
Diabetes	13,194 (21.9)	2,690 (22.3)	10,504 (21.8)	0.207	0.013
Anemia	19,055 (31.6)	3,841 (31.9)	15,214 (31.5)	0.511	0.009
Hypertension	30,195 (50.1)	6,035 (50.0)	24,160 (50.1)	0.935	0.002
Dyslipidemia	16,018 (26.6)	3,263 (27.1)	12,755 (26.4)	0.171	0.016
Chronic obstructive pulmonary disease	11,394 (18.9)	2,331 (19.3)	9,063 (18.8)	0.175	0.014
Cerebrovascular disease	2,641 (4.4)	547 (4.5)	2,094 (4.3)	0.350	0.010
Obesity	9,816 (16.3)	1,998 (16.6)	7,818 (16.2)	0.337	0.010
Severe liver disease	534 (0.9)	119 (1.0)	415 (0.9)	0.185	0.015
Rheumatic disease	1,146 (1.9)	230 (1.9)	916 (1.9)	0.952	0.001
Chronic kidney disease	7,498 (12.4)	1,563 (13.0)	5,935 (12.3)	0.050	0.018
Coagulopathy	3,435 (5.7)	758 (6.3)	2,677 (5.5)	0.002^*	0.031
Hospital status
Emergent admission				0.074	0.019
No	26,417 (43.9)	5,374 (44.7)	21,043 (43.7)
Yes	33,762 (56.1)	6,660 (55.3)	27,102 (56.3)
Missing	116	25	91
Hospital bed size				0.900	0.013
Small	8,671 (14.4)	1,758 (14.6)	6,913 (14.4)
Medium	16,765 (27.9)	3,335 (27.7)	13,430 (27.9)
Large	34,726 (57.7)	6,938 (57.7)	27,788 (57.7)
Missing	133	28	105
Hospital location/teaching status				0.988	0.010
Rural	6,610 (11.0)	1,321 (11.0)	5,289 (11.0)
Urban nonteaching	20,483 (34.0)	4,109 (34.2)	16,374 (34.0)
Urban teaching	33,069 (55.0)	6,601 (54.9)	26,468 (55.0)
Missing	133	28	105
Hospital region				0.892	0.010
Northeast	10,641 (17.8)	2,107 (17.5)	8,534 (17.9)
Midwest	15,395 (25.6)	3,112 (25.9)	12,283 (25.6)
South	22,028 (36.4)	4,404 (36.5)	17,624 (36.4)
West	12,231 (20.1)	2,436 (20.1)	9,795 (20.1)

Outcome	Total (n=60,295)	Muscle loss phenotype		P-value	SMD
Outcome	Total (n=60,295)	Yes (n=12,059)	No (n=48,236)	P-value	SMD
Postoperative complication
Any	35,865 (59.5)	9,439 (78.3)	26,426 (54.8)	<0.001^*	0.512
Death	2,470 (4.1)	814 (6.8)	1,656 (3.4)	<0.001^*	0.152
Acute myocardial infarction	1,239 (2.1)	366 (3.0)	873 (1.8)	<0.001^*	0.081
Cerebrovascular accident	1,333 (2.2)	313 (2.6)	1,020 (2.1)	0.001^*	0.033
Venous thromboembolism	2,693 (4.5)	963 (8.0)	1,730 (3.6)	<0.001^*	0.184
Pneumonia	2,914 (4.8)	1,012 (8.4)	1,902 (3.9)	<0.001^*	0.183
Sepsis	7,654 (12.7)	3,105 (25.7)	4,549 (9.4)	<0.001^*	0.437
Infection	7,617 (12.6)	2,712 (22.5)	4,905 (10.2)	<0.001^*	0.336
Respiratory failure	6,903 (11.4)	2,620 (21.7)	4,283 (8.9)	<0.001^*	0.361
Mechanical ventilation	4,609 (7.6)	1,764 (14.6)	2,845 (5.9)	<0.001^*	0.290
Acute kidney injury	8,659 (14.4)	2,986 (24.8)	5,673 (11.8)	<0.001^*	0.339
Shock	3,024 (5.0)	1,343 (11.1)	1,681 (3.5)	<0.001^*	0.298
Bleeding	21,532 (35.7)	5,783 (48.0)	15,749 (32.6)	<0.001^*	0.316
Wound complication	3,839 (6.4)	1,443 (12.0)	2,396 (5.0)	<0.001^*	0.253
Device-related complication	1,610 (2.7)	490 (4.1)	1,120 (2.3)	<0.001^*	0.101
Nervous system complication	129 (0.2)	31 (0.3)	98 (0.2)	0.252	0.011
Digestive system complication	11,512 (19.1)	3,716 (30.8)	7,796 (16.2)	<0.001^*	0.350
Unfavorable discharge^a	15,613 (27.0)	4,648 (41.3)	10,965 (23.6)	<0.001^*	0.346
Prolonged LOS^a,b	24,869 (42.9)	7,903 (70.1)	16,966 (36.3)	<0.001^*	0.719
Total hospital cost (per US $1,000 dollars)	102.5±0.7	160.6±1.9	87.9±0.6	<0.001^*	0.236

Age group	Any postoperative complication		Unfavorable discharge^a		Prolonged LOS^a,b		Total hospital cost (per US $1,000)
Age group	aOR (95% CI)	P-value	aOR (95% CI)	P-value	aOR (95% CI)	P-value	Adjusted β (95% CI)	P-value
All	2.99 (2.85–3.15)	<0.001^*	2.42 (2.30–2.53)	<0.001^*	4.34 (4.13–4.55)	<0.001^*	70.86 (67.11–74.61)	<0.001^*
<65 yr	3.55 (3.25–3.88)	<0.001^*	3.65 (3.22–4.14)	<0.001^*	5.35 (4.90–5.85)	<0.001^*	89.08 (86.52–91.64)	<0.001^*
≥65 yr	2.80 (2.64–2.97)	<0.001^*	2.39 (2.27–2.52)	<0.001^*	3.99 (3.77–4.22)	<0.001^*	62.96 (59.48–66.43)	<0.001^*

Postoperative complication	Age group
	All		<65 yr		≥65 yr
	aOR (95% CI)	P-value	aOR (95% CI)	P-value	aOR (95% CI)	P-value
Death	2.00 (1.83–2.19)	<0.001^*	2.17 (1.68–2.81)	<0.001^*	2.00 (1.82–2.20)	<0.001^*
Acute myocardial infarction	1.67 (1.47–1.90)	<0.001^*	1.69 (1.17–2.43)	0.005^*	1.68 (1.47–1.92)	<0.001^*
Cerebrovascular accident	1.19 (1.05–1.36)	0.008^*	1.69 (1.20–2.40)	0.003^*	1.14 (0.99–1.31)	0.067
Venous thromboembolism	2.30 (2.11–2.50)	<0.001^*	2.69 (2.27–3.20)	<0.001^*	2.19 (1.99–2.41)	<0.001^*
Pneumonia	2.19 (2.02–2.38)	<0.001^*	2.75 (2.28–3.32)	<0.001^*	2.11 (1.93–2.30)	<0.001^*
Sepsis	3.31 (3.14–3.49)	<0.001^*	4.28 (3.86–4.75)	<0.001^*	3.01 (2.83–3.20)	<0.001^*
Infection	2.53 (2.40–2.67)	<0.001^*	2.73 (2.45–3.04)	<0.001^*	2.46 (2.31–2.62)	<0.001^*
Respiratory failure	2.89 (2.73–3.05)	<0.001^*	3.93 (3.46–4.47)	<0.001^*	2.69 (2.53–2.86)	<0.001^*
Mechanical ventilation	2.73 (2.56–2.91)	<0.001^*	3.75 (3.25–4.34)	<0.001^*	2.50 (2.33–2.69)	<0.001^*
Acute kidney injury	2.45 (2.33–2.59)	<0.001^*	3.35 (2.99–3.76)	<0.001^*	2.28 (2.15–2.42)	<0.001^*
Shock	3.49 (3.23–3.77)	<0.001^*	4.59 (3.91–5.38)	<0.001^*	3.19 (2.92–3.48)	<0.001^*
Bleeding	1.88 (1.80–1.97)	<0.001^*	2.24 (2.06–2.43)	<0.001^*	1.79 (1.70–1.88)	<0.001^*
Wound complication	2.58 (2.40–2.76)	<0.001^*	3.00 (2.64–3.42)	<0.001^*	2.40 (2.21–2.61)	<0.001^*
Device-related complication	1.78 (1.60–1.99)	<0.001^*	1.76 (1.42–2.18)	<0.001^*	1.79 (1.58–2.03)	<0.001^*
Nervous system complication	1.25 (0.83–1.88)	0.280	2.80 (1.07–7.32)	0.036^*	1.10 (0.70–1.72)	0.695
Digestive system complication	2.30 (2.19–2.41)	<0.001^*	2.86 (2.62–3.13)	<0.001^*	2.12 (2.00–2.24)	<0.001^*

Table 1. Characteristics of the study population after propensity score matching

SMD, standardized mean difference; HMO, health maintenance organization.

Table 2. Outcomes of the study population after propensity score matching

Values are presented as number (%) or mean±standard error.

SMD, standardized mean difference; LOS, length of stay.

^aPatients with in-hospital mortality were excluded. ^bHospital LOS >75th percentile (10 days).

*P<0.05.

Table 3. Associations between the muscle loss phenotype (yes vs. no) and in-hospital outcomes after propensity score matching stratified by age

In the multivariable analysis, models were adjusted for race, primary payer, household income, and comorbidities (congestive heart failure, atrial fibrillation, and coagulopathy).

LOS, length of stay; aOR, adjusted odds ratio; CI, confidence interval.

^aExcluded patients with in-hospital mortality. ^bHospital LOS >75th percentile (10 days).

*P<0.05.

Table 4. Associations between the muscle loss phenotype (yes vs. no) and specific postoperative complications

In multivariable analysis, models were adjusted for race, primary payer, household income, and comorbidities (congestive heart failure, atrial fibrillation, and coagulopathy).