Utilizing psoas muscle cross-sectional area to predict functional outcome

Abstract

Background

This study evaluated psoas muscle area (PMA) as a predictor of frailty and functional outcome in trauma patients.

Methods

The cohort included 211 trauma patients admitted to an urban level I trauma center from March 2012 to May 2014 who consented to participate in a longitudinal study and underwent abdominal-pelvic computed tomography scans during their initial evaluation. Physical component scores (PCS) of the Veterans RAND 12-Item Health Survey were administered to assess physical functionality at baseline and at 3, 6, and 12 months after injury. PMA in mm² and Hounsfield units was calculated using the Centricity PACS system. Statistical models were stratified by injury severity score (ISS), <15 or ≥15, and adjusted for age, sex, and baseline PCS. Follow-up PCS were analyzed using general linear regression models.

Results

For participants with an ISS <15, increased PMA was significantly associated with higher PCS at 3 (P = 0.008), 6 (P = 0.02), and 12 months (P = 0.002), although this relationship was not statistically significant for ISS ≥15 (P = 0.85, 0.66, 0.61).

Conclusion

For mild to moderately injured (but not seriously injured) patients, those with larger psoas muscles experience better functional outcomes after injury.

The term frailty attempts to define a patient’s vulnerability to clinical and functional decline over time or in response to injury.^1,2 Although clinicians are quick to ascribe frailty at the bedside in individual cases using a Stewart-esque “I know it when I see it” threshold, frailty has proven difficult to quantify, and few convenient metrics exist to compare frailty between patients.^3,4 Frailty is usually characterized as an age-related phenomenon. However, many factors beyond age contribute to a patient’s degree of frailty, including weight loss, negative energy balance, physical limitation, and nutritional deficiency.⁵ Sarcopenia, an age-related decrease in muscle size and/or mass leading to decreased strength, of the psoas muscle has been found in previous studies to correlate with frailty and to be predictive of poor outcomes in many patient populations.^6–8

While it is possible to assess frailty according to a variety of metrics preoperatively in certain patient populations, those who experience unexpected injury are unable to be evaluated by traditional frailty scoring systems. However, abdominal and pelvic computed tomography (CT) scans, which many trauma patients receive, can provide a mechanism to assess a surrogate of frailty through radiographic measurement of psoas muscle cross-sectional area (PMA) to evaluate for sarcopenia. These measurements have been shown to be strong surrogate metrics of frailty in studies of multiple other surgical populations, including cardiothoracic, orthopedic, transplant, and spine surgery.^9–19 The aim of this study was to evaluate psoas muscle sarcopenia, a surrogate measure of frailty, as a predictor of long-term functional outcomes in trauma patients.

METHODS

Approval for this study was obtained from the hospital’s institutional review board. Participants were selected from a convenience sample of a larger prospectively enrolling longitudinal outcome study, which included patients admitted to a large, urban American College of Surgeons Committee on Trauma–verified level I trauma center in the Southwestern United States from March 2012 to May 2014 for a traumatic injury. Inclusion criteria for this study included (1) admission to the trauma service and entry into the hospital’s trauma registry with total hospital admission of at least 24 hours, (2) age ≥18 years, and (3) completion of study follow-up measures at 3, 6, and 12 months. Exclusion criteria included patients with (1) effects of a traumatic brain injury and/or premorbid cognitive deficits (e.g., dementia, Alzheimer’s disease) that would interfere with the ability to provide informed consent and did not resolve during hospitalization and (2) inability to understand written and spoken English or Spanish.

Of the 506 participants in this prospective longitudinal study, 211 participants were included in the present cohort analysis. These included participants who underwent abdominal and pelvic CT scans, as a part of standard of care, during their initial evaluation. Participants were excluded from the analysis cohort if their admission CT scans included psoas hematoma, paravertebral hematoma, spine instrumentation, L4 fracture, previous spinal cord injury, severe scoliosis, motion degradation or artifact on CT, or obvious asymmetry in psoas diameter.

Functional outcome data were gathered prospectively during acute inpatient hospitalization (baseline) and at 3, 6, and 12 months using the Veterans RAND 12-Item Health Survey (VR-12), a self-report measure derived from the Veterans RAND 36-Item Health Survey that has been used as a measure of health-related quality of life.²⁰ The VR-12 includes aspects of physical functionality, role limitations due to physical problems, bodily pain, general perception of health, vitality, social function, role limitation due to emotional problems, and mental health. The VR-12 baseline measurement assesses preinjury functionality by asking participants to report physical functionality generally “prior to your injury” and specifically “during the 4 weeks prior to your injury,” and the follow-up VR-12 asks participants to report functionality over “the last 4 weeks.” The measure comprises two component scores, a physical component score (PCS) and a mental component score (MCS). The PCS emphasizes questions regarding general health, physical functioning, and bodily pain, and the MCS emphasizes questions regarding emotional vitality, mental health, and social functioning.²¹ PCS and MCS scores are standardized to a 1990 US population norm. Using this technique, the general US population would have a standardized score of 50 ± 10.

Radiographic data from the admission abdominal and pelvic CT scans were viewed on Centricity. A built-in measurement capability in the Centricity program allows the user to trace around a structure and automatically calculate the area in mm² and Hounsfield units (HU) of the selected region. A minimum of 20 points were used to surround and measure the PMA on one side, at the slice at which both transverse processes of L4 could be seen most in their entirety. The L4 vertebral body area and HU were also measured at the same slice. Intraclass correlations were used to measure interrater reliability for the radiographic measurements. The measurements of agreement for the raters are as follows: PMA, 0.93; psoas HU, 0.83; L4 area, 0.74; and L4 HU, 0.93.

Demographic and injury-related variables, including injury severity score (ISS), discharge disposition, and etiology of injury, were gathered from the trauma registry. Participant heights and weights at the time of admission were obtained from the hospital charts. All participant characteristics were summarized using counts and percentages or means and standard deviations (SD), as appropriate. General linear regression models were used to determine the association between PMA and the participants’ physical functionality (from their PCS) and hospital length of stay (LOS). Statistical models were also stratified by ISS <15 or ≥15 and adjusted for age, sex, and baseline PCS.

RESULTS

The demographic and injury-related variables and average PCS scores for the 211 participants analyzed in this study are presented in Table 1. The participants ranged in age from 18 to 86 years (mean = 42), and most were male (65%), Caucasian (70%), and experienced blunt traumatic injuries (85%), which is consistent with the typical trauma population at this trauma center. The most common etiology of injury was motor vehicle collision (39%), followed by motorcycle collision (16%). The mean ± SD PMA was 1339 ± 430 mm². The mean ± SD of the L4 vertebral body area was 1371 ± 254 mm². The majority of participants included in this analysis sustained mild to moderate injuries (ISS <15, 60%).

Table 1.

Patient demographic and injury-related variables

Variable	Mean ± SD; N (%)
Age, mean ± SD	42.0 ± 15.8
Male sex	137 (65%)
Race
White	147 (70%)
Black or African American	52 (25%)
Other/unknown	12 (5%)
Weight (kg), mean ± SD	88.7 ± 21.9
Height (cm), mean ± SD	172.8 ± 9.9
BMI (kg/m2), mean ± SD	29.4 ± 7.1
Blunt injury	179 (85%)
Cause of injury
Motor vehicle collision	79 (39%)
Motorcycle collision	32 (16%)
Fall	27 (13%)
Other	67 (32%)
ISS, mean ± SD	14.1 ± 8.9
<15	124 (60%)
≥15	81 (40%)
Total LOS, mean ± SD	8.9 ± 8.8
Disposition
Home	153 (75%)
Rehabilitation	32 (16%)
Nursing home	11 (5%)
Other	9 (4%)
PMA (mm2), mean ± SD	1339 ± 431
Psoas HU, mean ± SD	55.4 ± 10.5
L4 area (mm2), mean ± SD	1371 ± 254
Vertebral HU, mean ± SD	266.7 ± 61.9
Ratio PMA/L4 area, mean ± SD	1.0 ± 0.3
Ratio PMA/height, mean ± SD	7.6 ± 2.3
PCS, mean ± SD
Baseline	46.1 ± 10.4
3 months	30.0 ± 11.8
6 months	33.7 ± 12.9
12 months	34.9 ± 12.5

BMI indicates body mass index; HU, Hounsfield units; ISS, injury severity score; LOS, length of stay; PCS, physical component scores; PMA, psoas muscle area; SD, standard deviation.

Table 2 shows regression results for follow-up PCS scores with radiographic measurement variables, with all analyses stratified by ISS <15 or ≥15 and adjusted for age, sex, and baseline PCS. For ISS <15, PMA alone was significantly associated with higher PCS at baseline. Additionally, larger PMA corresponded with a higher PCS at all follow-up time points of 3 months (P = 0.008), 6 months (P = 0.02), and 12 months (P = 0.002) when controlling for age, sex, and baseline PCS. However, PMA was not significantly associated with PCS at any follow-up time point for more seriously injured patients with an ISS ≥15 (3 months, P = 0.85; 6 months, P = 0.66; 12 months, P = 0.60).

Table 2.

Associations of physical component score at 3, 6, and 12 months and radiographic measurements, stratified by injury severity score^a

ISS	Variable	PCS 3 months		PCS 6 months		PCS 12 months
		Beta (SE)	P value	Beta (SE)	P value	Beta (SE)	P value
<15	PMA	0.06 (0.02)	0.008	0.07 (0.03)	0.02	0.08 (0.02)	0.002
	Ratio PMA/L4 area	5.97 (9.46)	0.53	7.83 (11.43)	0.50	−8.14 (9.87)	0.41
	Ratio PMA/height	−9.37 (3.97)	0.02	−10.72 (5.50)	0.06	−10.38 (4.73)	0.03
	Psoas HU	−0.30 (0.14)	0.03	−0.02 (0.16)	0.89	0.24 (0.14)	0.10
	Vertebral HU	0.01 (0.03)	0.69	−0.01 (0.04)	0.84	0.02 (0.03)	0.55
≥15	PMA	−0.01 (0.03)	0.85	−0.01 (0.03)	0.66	−0.02 (0.04)	0.60
	Ratio PMA/L4 area	5.45 (14.28)	0.70	−8.69 (14.49)	0.55	−9.10 (19.63)	0.65
	Ratio PMA/height	−0.90 (5.85)	0.88	4.04 (6.45)	0.54	6.99 (8.87)	0.44
	Psoas HU	−0.13 (0.23)	0.57	−0.61 (0.24)	0.02	−0.11 (0.30)	0.72
	Vertebral HU	−0.02 (0.04)	0.69	0.01 (0.05)	0.79	0.03 (0.07)	0.61

HU indicates Hounsfield units; ISS, injury severity score; PCS, physical component score; PMA, psoas muscle area; SE, standard error.

^aAll models were stratified by ISS, <15 or ≥15, and adjusted for age, sex, and baseline physical component score.

In this analysis, for participants with an ISS <15, a 1 mm² increase in PMA correlated with a 0.06 increase in PCS at 3 months. Thus, after controlling for age, sex, and baseline PCS, an increase of 100 mm² in PMA predicted a 6-point increase in PCS 3 months after injury. The predicted differences in functionality were greater at later time points, such that a 100 mm² increase in PMA at the time of injury correlated with a 7-point increase in PCS at 6 months postinjury and an 8-point increase in PCS at 12 months postinjury.

Due to the variability of psoas muscle size among individuals of different heights, a similar analysis was undertaken utilizing the ratio of PMA to patient height. The association between ratios of PMA to patient height and PCS was significant for participants of ISS <15 at 3 months (P = 0.02) and 12 months (P = 0.03), and this association held a trend toward significance at 6 months (P = 0.06). The ratios of PMA to height were not statistically associated with PCS at any follow-up time point for participants with an ISS ≥15 (3 months, P = 0.88; 6 months, P = 0.54; 12 months, P = 0.44). The ratio of PMA to L4 area was also examined as a method to attempt to control for height differences. However, this ratio was not found to be significantly associated with PCS for participants of ISS <15 at 3 months (P = 0.53), 6 months (P = 0.50), or 12 months (P = 0.41) or ISS ≥15 at 3 months (P = 0.70), 6 months (P = 0.55), or 12 months (P = 0.65).

Table 3 provides the results of the cross-analysis of radiographic measurement variables and LOS. All models were again stratified by ISS, <15 or ≥15, and adjusted for age, sex, and baseline PCS. For ISS <15, the ratio of PMA to height was a significant predictor of decreased LOS (P = 0.04), where a one-unit increase in the ratio of PMA to height correlated with an expected decrease in LOS by 3.68 days. However, this association was not statistically significant for participants with an ISS ≥15 (P = 0.70). For ISS ≥15, vertebral HU was significantly associated with LOS (P = 0.04). However, this increase in LOS was negligible, 0.07 days, and likely holds little clinical significance. Additionally, this association was not found to be significant for participants of ISS <15 (P = 0.81). No other radiographic measurements, including PMA, were found to be significantly associated with LOS for ISS <15 or ≥15.

Table 3.

Associations of length of stay and radiographic measurements, stratified by injury severity score^a

ISS	Measure	Beta (SE)	P value
<15	PMA	0.02 (0.01)	0.08
	Ratio PMA/L4 area	3.66 (4.64)	0.43
	Ratio PMA/height	−3.68 (1.78)	0.04
	Psoas HU	−0.02 (0.07)	0.80
	Vertebral HU	0.00 (0.01)	0.81
≥15	PMA	−0.01 (0.02)	0.73
	Ratio PMA/L4 area	−9.45 (10.43)	0.37
	Ratio PMA/height	1.63 (4.17)	0.70
	Psoas HU	0.15 (0.15)	0.30
	Vertebral HU	0.07 (0.03)	0.04

HU indicates Hounsfield units; PMA, psoas muscle area; SE, standard error.

^aAll models were stratified by injury severity score, <15 or ≥15, and adjusted for age, sex, and baseline physical component score.

DISCUSSION

PMA was shown to be significantly correlated with improved functional outcomes at all follow-up time points for ISS <15, when controlling for age, sex, and baseline PCS. Based on these data, it may be reasonable to use the PMA at admission as a metric for baseline frailty in mild to moderately injured trauma patients. However, these results also show that PMA is not significantly related to such outcomes in more severely injured participants (ISS ≥15). An explanation for these findings may be that the benefit of a physiologic reserve is blunted once traumatic injuries become more severe, although further investigation is needed to confirm these findings.

The common use of CT scans to evaluate injured patients provides a single, convenient opportunity to measure PMA and may allow for early identification of sarcopenia and frailty without requiring formal functional measurements. This strategy provides an additional tool to supplement clinical decision-making at the time of admission and could contribute to the growing number of predictive metrics used to identify and stratify patients at risk for poorer outcomes. These patients may benefit from earlier involvement of physical therapy, more aggressive nutritional optimization, and earlier discharge planning to transitional or definitive care facilities. This information and early identification could also guide providers in establishing reasonable expectations for patients, family, and caregivers during care discussions.

The current study has limitations. These analyses were drawn from a convenience study sample from a large level I trauma center with trauma patients, so the results may not be generalizable to other trauma centers or patient populations. Additionally, this study was observational in nature and did not offer any randomization or interventions for participants. Further, the participants included only those who met inclusion criteria, consented to participate in the larger prospective, longitudinal study, and received abdominal-pelvic CT scans during initial presentation. These criteria inherently exclude patients who did not receive abdominal-pelvic CT scans and those who were unable to consent to participate in the study, including those with more severe injuries.

Additionally, as a measure of physical functionality, the VR-12 is limited to the information the participants self-reported and does not directly assess a patient’s physical functionality. However, it would not be feasible to take direct measurements of preinjury functionality for participants who have suffered a traumatic injury, so despite their limitations, validated self-report measures like the VR-12 are likely the best way to capture this information.²⁰ Further, the authors recognize that in stratifying the study population by ISS, age, and sex, the resulting participant cohorts contained small sample sizes.

As the burden of chronic disease in the trauma population increases, it is becoming increasingly important for clinicians to elucidate factors that may contribute to poorer outcomes for these patients and provide interventions to reduce the concomitant burden of injury. Frailty, and sarcopenia as a surrogate marker, has significant predictive implications for injured patients. Although there has been increased investigation of psoas sarcopenia in trauma patients in recent years, these studies have utilized a wide variety of inclusion criteria and methodologies, leaving the relationship between sarcopenia and long-term outcomes for patients still somewhat unclear.^22–24 While it is intuitive that the outcome for similar injuries is different for the frail versus the robust, capturing this reality in a readily available metric remains challenging. Larger studies are warranted to further characterize and hone the relationship between PMA and patient-centered outcomes.

Funding Statement

The authors gratefully acknowledge the support of the Stanley Seeger Surgical Fund of the Baylor Health Care System Foundation.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author(s).