Morphometrics as a predictor of perioperative morbidity after lumbar spine surgery

Free access

OBJECT

Improved objective assessments of perioperative risk after spine surgery are necessary to decrease postoperative morbidity and mortality rates. Morphometric analysis has proven utility in predicting postoperative morbidity and mortality in surgical disciplines. The aim of the present study was to evaluate whether morphometrics can be applied to the cases of patients undergoing lumbar spine surgery.

METHODS

The authors performed a retrospective review of the perioperative course of 395 patients who underwent lumbar surgery at their institution from 2013 to 2014. Preoperative risk factors such as age, diabetes, smoking, coronary artery disease, and body mass index (BMI) were recorded. Preoperative MRI was used to measure the psoas muscle area at the L-4 vertebra and paraspinal muscle area at the T-12 vertebra. Primary outcomes included unplanned return to the operating room, 30- and 90-day readmissions, surgical site infection, wound dehiscence, new neurological deficit, deep vein thrombosis, pulmonary embolism, myocardial infarction, urinary tract infection, urinary retention, hospital-acquired pneumonia, stroke, and prolonged stay in the intensive care unit.

RESULTS

The overall rate of adverse events was 30%, the most common event being urinary retention (12%). Greater age (p = 0.015) and tobacco usage (p = 0.026) were both significantly associated with complications for all patients, while diabetes, coronary artery disease, and high BMI were not. No surgery-related characteristics were associated with postoperative morbidity, including whether surgery required instrumentation, whether it was a revision, or the number of vertebral levels treated. Using multivariate regression analysis, male and female patients with the lowest psoas tertile had an OR of 1.70 (95% CI 1.04–2.79, p = 0.035) for having postoperative complications. Male patients in the lowest psoas tertile had an OR of 2.42 (95% CI 1.17–5.01, p = 0.016) for having a postoperative complication. The paraspinal muscle groups did not provide any significant data for postoperative morbidity, even after multivariate analysis.

CONCLUSIONS

The morphometric measurement of psoas muscle size may be a sensitive predictive tool compared with other risk factors for perioperative morbidity in male patients undergoing lumbar surgery.

ABBREVIATIONSBMI = body mass index; CAD = coronary artery disease; DVT = deep vein thrombosis; HFHS = Henry Ford Health System; ICC = intraclass correlation coefficient; LOS = length of stay; Ml = myocardial infarction; PE = pulmonary embolism; UTI = urinary tract infection.

OBJECT

Improved objective assessments of perioperative risk after spine surgery are necessary to decrease postoperative morbidity and mortality rates. Morphometric analysis has proven utility in predicting postoperative morbidity and mortality in surgical disciplines. The aim of the present study was to evaluate whether morphometrics can be applied to the cases of patients undergoing lumbar spine surgery.

METHODS

The authors performed a retrospective review of the perioperative course of 395 patients who underwent lumbar surgery at their institution from 2013 to 2014. Preoperative risk factors such as age, diabetes, smoking, coronary artery disease, and body mass index (BMI) were recorded. Preoperative MRI was used to measure the psoas muscle area at the L-4 vertebra and paraspinal muscle area at the T-12 vertebra. Primary outcomes included unplanned return to the operating room, 30- and 90-day readmissions, surgical site infection, wound dehiscence, new neurological deficit, deep vein thrombosis, pulmonary embolism, myocardial infarction, urinary tract infection, urinary retention, hospital-acquired pneumonia, stroke, and prolonged stay in the intensive care unit.

RESULTS

The overall rate of adverse events was 30%, the most common event being urinary retention (12%). Greater age (p = 0.015) and tobacco usage (p = 0.026) were both significantly associated with complications for all patients, while diabetes, coronary artery disease, and high BMI were not. No surgery-related characteristics were associated with postoperative morbidity, including whether surgery required instrumentation, whether it was a revision, or the number of vertebral levels treated. Using multivariate regression analysis, male and female patients with the lowest psoas tertile had an OR of 1.70 (95% CI 1.04–2.79, p = 0.035) for having postoperative complications. Male patients in the lowest psoas tertile had an OR of 2.42 (95% CI 1.17–5.01, p = 0.016) for having a postoperative complication. The paraspinal muscle groups did not provide any significant data for postoperative morbidity, even after multivariate analysis.

CONCLUSIONS

The morphometric measurement of psoas muscle size may be a sensitive predictive tool compared with other risk factors for perioperative morbidity in male patients undergoing lumbar surgery.

Low-back pain is a common disorder, contributing to a large portion of health care costs.15,20 Over the past decade, there has been a noticeable increase in the frequency of elective lumbar surgery (decompressions and arthrodesis),19,28,35 perhaps due its proven effectiveness.32–34 However, this is associated with an increase in the frequency of postoperative complications after spinal surgery.3,11,25 With the current focus on cost-efficient health care, it is important to identify patients who are at greater risk for postoperative morbidity and mortality, which may require greater resource utilization.3,11

A patient’s general health status is an important consideration when weighing which surgical approaches are appropriate for a given pathology, always with the aim of minimizing the potential perioperative morbidity. However, the factors that contribute to a surgeon’s decision regarding patient selection and surgical approach are generally empirical and without the benefit of a validated risk assessment tool. Factors such as age, smoking status, obesity, coronary artery disease (CAD), chronic obstructive pulmonary disease, diabetes, severity of symptoms, and complexity of the surgery are all important, but their relative contribution to overall perioperative risk is not well quantified. Validated risk-stratification tools provide objective data to surgeons to one aspect of a patient’s operative risk (cardiovascular, pulmonary, psychological, etc.),8,10,16,23,26,29,30 but their utility to spinal surgery is limited. Given that the vast majority of spinal procedures are elective, surgical candidates are generally without any active exacerbations of their medical problems. Despite diligent patient selection and preoperative screening and optimization, elective spinal procedures are still potentially associated with morbidity.27 Ultimately, risk calculations may be insufficient to assess the overall health of a patient, as two patients with the same preoperative risk factors (i.e., age, American Society of Anesthesiologists score, comorbidities, and disease stage) are often in clearly different stages of health. Surgeons will often “eyeball” patients to see if they are fit for surgery, and surgeons make operative decisions regardless of risk stratification.

Recently, the concept of frailty has been introduced as a measure of a patient’s health status.9,21 Frailty is defined as “a biologic syndrome of decreased reserve and resistance to stressors, resulting from cumulative declines across multiple physiologic systems.”9,21 While it is a common end point to human senescence, this process is pathological and has been shown to be associated with adverse outcomes after surgery.1,5,14 Unfortunately, measuring human frailty relies on subjective assessments and prolonged patient cooperation, which is impractical in application.1,14 Surrogate markers of frailty, such as sarcopenia, may be more useful to clinicians. Morphometrics is the measurement of patient attributes that are indicative of sarcopenia and thus frailty by proxy. Morphometric analysis has proven utility in predicting postoperative morbidity and mortality following major general, vascular, and transplant surgery.2,4,6,7,12,13,17,18,22,24,31 This study evaluates whether morphometrics can be used as a reliable predictor of perioperative morbidity in patients undergoing lumbar spine surgery.

Methods

Chart Review

This is a retrospective cohort study of patients undergoing lumbar surgery within the Henry Ford Health System (HFHS). After obtaining institutional review board approval, we compiled a list of all patients who had undergone lumbar surgery (T-11 through S-1, inclusive) at HFHS from 2013 to 2014. We used the Current Procedural Terminology (CPT) codes for lumbar laminectomy, lumbar posterior/posterolateral arthrodesis, and lumbar interbody arthrodesis and identified 466 patients. Using the electronic medical records, each patient’s preoperative risk factors, including age, sex, and status pertaining to diabetes, smoking, CAD, and body mass index (BMI), were entered into the database. Variables pertaining to operative details as well as immediate postoperative stay included the number of vertebral levels treated, whether the surgery was a revision, whether there was instrumentation, the length of stay (LOS), and disposition at discharge (home, skilled nursing facility, acute rehabilitation facility). Primary outcomes recorded included any 90-day postoperative complications including: death, unplanned return to surgery, 30- and 90-day hospital readmission, surgical site infection, wound dehiscence, new neurological deficit, deep vein thrombosis (DVT), pulmonary embolism (PE), myocardial infarction (MI), urinary tract infection (UTI), urinary retention, hospital-acquired pneumonia, stroke, and prolonged stay (> 3 days) in the intensive care unit.

Analytical Morphometrics

Our analytical morphometrics methodology has been adopted from that described by Englesbe et al.2,31 and applied to preoperative MRI. Using the freehand region of interest tool on Philips ePACS viewer, the circumferential area of each patient’s psoas muscles at L-4 and paraspinal muscles at T-12 were measured on axial T2-weighted MRI sequences (Fig. 1). Muscle area was calculated in cubic centimeters, and the areas for each muscle were entered in our database. For a reliability assessment of the freehand identification of the region of interest, a random sample of 40 patients (20 males and 20 females) was selected and measured by 3 observers (H.Z., B.G., and V.C.) independently.

FIG. 1.
FIG. 1.

Examples of morphometric measurements on T2-weighted MR images obtained in patient with (A–C) and without (D–F) sarcopenia. A: Sagittal image of a patient with sarcopenia. The white lines through the L-4 and T-12 vertebrae are example cross-sections from which morphometric measurements were taken. B: Axial image, of the same patient, taken at the L-4 vertebra showing a small psoas muscle measuring 5.33 cm2. C: Axial image, of the same patient, obtained at the T-12 vertebra showing a small paraspinal muscle measuring 9.32 cm2. D: Sagittal image of a patient without sarcopenia. The white lines through the L-4 and T-12 vertebrae again show example cross-sections from which morphometric measurements were taken. E: Axial image, of the patient, demonstrating a large psoas muscle measuring 21.48 cm2. F: Axial image, of the same patient, demonstrating a large paraspinal muscle measuring 23.32 cm2.

Statistical Analysis

To assess differences in demographic and surgical information and morphometric measurements associated with complications, chi-square tests were performed for categorical/binary variables and 2-sample t-tests for the continuous variables. A Wilcoxon 2-sample nonparametric test was used to compare patients with and without complications for number of vertebral levels treated. Due to sex differences in the morphometric measurements, cut-points for the tertiles were determined separately for males and females. Chi-square and Kruskal-Wallis tests were performed to assess the association between the tertile measurements and specific complications and total number of complications. Multivariate logistic regression was done to adjust for potential confounders when assessing the relationship between experiencing any complication and psoas tertiles. Kruskal-Wallis tests were used to assess the association of LOS with total psoas and total paraspinous tertiles. The associations of discharge disposition and tertiles for total psoas and paraspinous muscles were investigated using chi-square tests. Sex differences in the morphometric measurements were tested using 2-sample t-tests. Intraclass correlation coefficients (ICCs) were computed to assess the reliability among observers of the morphometric measurements of the right and left psoas and paraspinous muscles, as well as the total psoas and paraspinous muscles. All testing was done at the 0.05 level. SAS version 9.4 was used for data analysis.

Results

Participants

A total of 466 patients were identified and their data entered in the database. Of the 466 patients, 8 were excluded due to duplicate records, 19 were excluded due to having undergone nonlumbar surgery, and 30 were excluded due to inadequate imaging (i.e., MR images not available or incorrectly formatted for morphometric analysis). Of the remaining 409 patients, 14 were excluded due to incomplete medical records, leaving 395 patients for statistical analysis. Three patients had two elective surgeries recorded, but only the first one was used in the analysis.

Descriptive Data

A summary of descriptive information can be found in Table 1. The mean age of all patients was 63.3 years (SD ± 12.48, median 63, range 23–88). Of the 395 patients, 51% (n = 203) were female, 24% (n = 94) had diabetes, 27% (n = 105) were smokers, 16% (n = 62) had CAD, and mean BMI was 30.92 (SD ± 6.44, median 30, range 16.45 to 55.51). Of the 395 surgeries, 54% (n = 213) were instrumentations and 29% (n = 113) were revisions. Regarding complexity, 30% (n = 119) were 1-level procedures, of which 43% were instrumented and 25% were revisions; 31% (n = 122) were 2-level procedures, of which 58% were instrumented and 28% were revisions; the remaining 39% (n = 154) were multilevel operations, of which 60% were instrumented and 34% were revisions. Only 14 cases were minimally invasive surgeries (MIS), complementing 4% of our patient population.

TABLE 1.

Summary of descriptive features*

VariableAll Patients (n = 395)
Demographics
 Age (yrs)
 Mean ± SD63.30 ± 12.48
 Median (range)64 (23–80)
Sex
 Female203 (51%)
 Male192 (49%)
Risk factors
 Diabetes94 (24%)
 Smoking105 (27%)
 CAD62 (16%)
 BMI
  Mean ± SD
  Median (range)30.92 ± 6.44 30 (16.45 to 55.51)
Surgical information
 Instrumented213 (54%)
 Revision113 (29%)
 MIS14 (4%)
 No. of levels treated
  1119 (30%)
  2122 (31%)
  377 (19%)
  448 (12%)
  521 (5%)
  63 (1%)
  73 (1%)
  81 (0%)
  101 (0%)
 Mean ± SD
 Median (range)2.40 ± 1.37 2 (0 to 10)

MIS = minimally invasive surgery.

Values are presented as the mean ± SD, median (range), or number (%).

Outcome Data

Table 2 contains the types of the complications and their respective rates, as well as LOS and discharge disposition. Of the 395 patients who had surgery, 30% (n = 120) had at least one complication, with the most common complication being urinary retention (12%, n = 46), followed by surgical site infection (9%, n = 36) and unplanned readmission within 90 days (9%, n = 36) (Table 2). The mean LOS was 3.89 days (SD ± 2.97, median 3, range 0–27 days), with 71% (n = 278) of patients being discharged home immediately thereafter (Table 2).

TABLE 2.

Outcomes and complications

VariableAll Patients (n= 395)
Complications
 Mortality1 (0%)
 Unplanned return to OR20 (5%)
 Unplanned readmission 30 days30 (8%)
 Unplanned readmission 90 days36 (9%)
 Surgical site infection36 (9%)
 Wound dehiscence24 (6%)
 New neurological deficit7 (2%)
 DVT12 (3%)
 PE6 (2%)
 MI4 (1%)
 UTI18 (5%)
 Urinary retention 46 (12%)
 Pneumonia 9 (2%)
 Prolonged ICU stay >3 days7 (2%)
 CVA1 (0%)
 Any complication 120 (30%)
Other outcomes
 LOS, days
  Mean + SD3.89 ± 2.97
  Median (range)3 (0–27)
 Disposition
  Home278 (71%)
  Acute rehab38 (10%)
  Skilled nursing facility78 (20%)

CVA = cerebrovascular accident; ICU = intensive care unit; OR = operating room.

Postoperative Complication Analysis

Table 3 contains the comparisons for patients with and without complications. The differences between the groups were significant for age (patients who had complications were an average of 3 years older [65.6 vs 62.3 years, p = 0.015]) and smoking (patients who had complications were less likely to be smoking [19% vs 30%, p = 0.026]). Female sex approached significance in terms of having more complications than males (p = 0.068). No significant association was found for diabetes, CAD, BMI, number of vertebral levels treated, and whether surgery involved instrumentation or was a revision. These same comparisons of patients with and without complications were performed for females and males due to the observed sex differences. For females, the only significant difference observed was that patients with complications had a higher mean BMI than patients without complications (p = 0.021), while age (p = 0.098) and smoking (p = 0.064) trended toward significance. For males, trends were also seen for age (males with complications on average were older [p = 0.076]) and CAD (30% patients with complications had CAD, whereas 18% those without complications had CAD [p = 0.063]). No significant differences in complication rates were seen when comparing cases of minimally invasive surgery and open surgery (Table 3).

TABLE 3.

Comparison of patients with and without complications*

VariableNo Complications (n=275)Complications (n = 120)P Value
All patients
 Mean age (yrs)62.29 ± 12.3165.61 ± 12.610.015
 Sex0.068
  Female133 (48%)70 (58%)
  Male142 (52%)50 (42%)
 Diabetes63 (23%)31 (26%)0.530
 Smoking82 (30%)23 (19%)0.026
 CAD39 (14%)23 (19%)0.198
 Mean BMI30.65 ± 6.1631.52 ± 7.020.215
 Instrumented147 (53%)66 (55%)0.777
 Revision83 (30%)30 (25%)0.295
 MIS12 (4%)2 (2%)0.182
 Mean no. of levels treated2.42 ± 1.362.35 ± 1.410.562
Female patients
 Mean age (yrs)62.37 ± 12.6565.46 ± 12.420.098
 Diabetes25 (19%)18 (26%)0.252
 Smoking34 (26%)10 (14%)0.064
 CAD14 (11%)8 (12%)0.817
 Mean BMI30.26 ± 6.5332.62 ± 7.590.021
 Instrumented87 (65%)44 (63%)0.717
 Revision46 (35%)21 (30%)0.509
 MIS5 (4%)1 (1%)0.351
 Mean no. of levels treated2.41 ± 1.292.37 ± 1.520.568
Male patients
 Mean age (yrs)62.22 ± 12.0265.82 ± 13.010.076
 Diabetes38 (27%)13 (26%)0.917
 Smoking48 (34%)13 (26%)0.295
 CAD25 (18%)15 (30%)0.063
 Mean BMI31.02 ± 5.7829.99 ± 5.870.279
 Instrumented60 (42%)22 (44%)0.830
 Revision37 (26%)9 (18%)0.251
 MIS7 (5%)1 (2%)0.372
 Mean no. of levels treated2.43 ± 1.422.32 ± 1.240.792

Mean values are presented ± SD. Boldface indicates statistical significance.

Morphometric Results

The ICCs for the psoas muscles were extremely high or almost perfect with a range from 0.988 to 0.991, indicating that the 3 independent observers were able to each precisely measure psoas muscle sizes. For the paraspinous muscles, the ICCs were lower than the psoas, but were still very good with a range from 0.842 to 0.857. Table 4 contains descriptive information for the morphometric measurements for all patients. For each morphometric location, the average area and total muscle area (psoas or paraspinous) were used. Average here refers to the mean between the right and left muscle area within a particular individual, and the total refers to the sum of both muscles for each individual. For all measurements, the differences were significant, with males having higher muscle sizes than females when comparing average psoas areas (14.98 ± 4.24 vs 9.66 ± 2.57, p < 0.001), total psoas areas (29.96 ± 8.49 vs 19.32 ± 5.14, p < 0.001), average paraspinous areas (19.01 ± 6.28 vs 14.65 ± 4.56, p < 0.001), and total paraspinous areas (37.96 ± 12.61 vs 28.98 ± 8.41, p < 0.001). Paraspinous area measurements were not obtained in all patients due to the heterogeneity of the overall MRI protocols (i.e., many studies were performed outside HFHS and then uploaded to our ePACs system), several of which did not include axial images up to T-12.

TABLE 4.

Morphometric measurements*

MeasurementsAll Patients (n = 395)Female (n = 203)Male (n = 192)p Value
Average L-4 psoas area12.25 ± 4.389.66 ± 2.5714.98 ± 4.24<0.001
Total L-4 psoas area24.48 ± 8.7619.32 ± 5.1429.96 ± 8.49<0.001
Average T-12 paraspinous area16.70 ± 5.8514.65 ± 4.5619.01 ± 6.28<0.001
Total T-12 paraspinous area§33.18 ± 11.4928.98 ± 8.4137.96 ± 12.61<0.001

Values presented as the mean ± SD. “Average” refers to the mean between the right and left muscle area within a particular individual, and “total” refers to the sum of both muscles for each individual.

Overall 394 patients; 203 females and 191 males.

Overall 355 all patients; 188 females and 167 males.

Overall 346 all patients; 184 females and 162 males.

We then compared the morphometric psoas and paraspinal muscle measurements of patients who did not have complications with those who did have complications (Table 5). We divided patients into tertiles based on the size of their psoas and paraspinal muscles, with the highest tertile having the largest muscle sizes and the lowest tertiles having the smallest. Cases in the smallest tertile in terms of average and total psoas areas approached significance in predicting postoperative morbidity (p = 0.096 and p = 0.052, respectively). Paraspinal muscle tertiles did not show any significant associations with morbidity. Given the significant difference in psoas size between sexes, we analyzed males and females separately. For female patients, there was no observed trend for either psoas muscles or paraspinal muscles. Male patients with smaller average psoas areas and with smaller total psoas areas were significantly more likely to have postoperative morbidity (p = 0.014 for both). For the morphometric tertiles, differences were significant for average psoas area (p = 0.032) and total psoas area (p = 0.008), with patients in the lowest tertile having a greater incidence of complications. There was no observed trend for paraspinal muscle area and morbidity for male patients.

TABLE 5.

Morphometric comparisons of patients with and without complications

Variable & ResponseNo ComplicationsComplicationsp Value*
All patients275120
 Average psoas tertiles0.096
  Lowest83 (30%)48 (40%)
  Middle102 (37%)33 (28%)
  Highest90 (33%)39 (33%)
 Total psoas tertiles0.052
  Lowest83 (30%)50 (42%)
  Middle99 (36%)31 (26%)
  Highest92 (34%)39 (33%)
 Average paraspinous tertiles0.826
  Lowest78 (32%)40 (35%)
  Middle81 (33%)37 (33%)
  Highest83 (34%)36 (32%)
 Total paraspinous tertiles0.791
  Lowest76 (32%)39 (36%)
  Middle82 (35%)36 (33%)
  Highest79 (33%)34 (31%)
Female patients13370
 Average L-4 psoas areaf9.58 ± 2.409.82 ± 2.880.512
 Total L-4 psoas areaf19.15 ± 4.8019.65 ± 5.750.512
 Average T-12 paraspinous areaf14.47 ± 4.4014.99 ± 4.850.458
 Total T-12 paraspinous areaf28.48 ± 7.5529.91 ± 9.830.315
 Average psoas tertiles0.673
  Lowest43 (32%)24 (34%)
  Middle48 (36%)21 (30%)
  Highest42 (32%)25 (36%)
 Total psoas tertiles0.673
  Lowest43 (32%)24 (34%)
  Middle48 (36%)21 (30%)
  Highest42 (32%)25 (36%)
 Average paraspinous tertiles0.848
  Lowest40 (33%)23 (35%)
  Middle42 (34%)20 (30%)
  Highest40 (33%)23 (35%)
 Total paraspinous tertiles0.821
  Lowest39 (33%)22 (34%)
  Middle43 (36%)20 (31%)
  Highest38 (32%)22 (34%)
Male patients14250
 Average L-4 psoas areaf15.43 ± 4.3213.71 ± 3.730.014
 Total L-4 psoas areaf30.86 ± 8.6827.43 ± 7.450.014
 Average T-12 paraspinous areaf19.35 ± 6.8518.14 ± 4.470.183
 Total T-12 paraspinous areaf38.68 ± 13.7636.09 ± 8.840.159
 Average psoas tertiles0.032
  Lowest40 (28%)24 (48%)
  Middle54 (38%)12 (24%)
  Highest48 (34%)14 (28%)
 Total psoas tertiles0.008
  Lowest40 (28%)26 (52%)
  Middle51 (36%)10 (20%)
  Highest50 (35%)14 (28%)
 Average paraspinous tertiles0.602
  Lowest38 (32%)17 (36%)
  Middle39 (33%)17 (36%)
  Highest43 (36%)13 (28%)
 Total paraspinous tertiles0.573
  Lowest37 (32%)17 (38%)
  Middle39 (33%)16 (36%)
  Highest41 (35%)12 (27%)

Boldface indicates statistical significance,

Values are presented as the mean ± SD.

Odds Ratio and Multivariate Analysis of Morphometrics

Multivariate logistic regression was done to adjust for potential confounders when assessing the relationship between the lowest total psoas tertile versus the middle/highest tertile (Table 6). The association of complications and BMI remained significant for all patients and for female patients after adjusting for age, CAD, and smoking. The difference between the lowest total psoas tertile and the upper two tertiles remained significant; patients in the lowest tertile had an OR of 1.70 that they would experience complications compared with the other two tertiles (95% CI 1.04–2.79, p = 0.035). This association was especially prominent in male patients, with an OR of 2.42 (95% CI 1.17–5.01, p = 0.016) of experiencing a postoperative morbidity.

TABLE 6.

Multivariate results using lowest total psoas tertile vs other two tertiles*

VariableAll PatientsMales OnlyFemale Only
OR (95% CI)p ValueOR (95% CI)p ValueOR (95% CI)p Value
Female sex1.46 (0.93–2.28)0.103
Age (increase of 10)1.14 (0.94–1.40)0.1851.06 (0.78–1.45)0.7041.19 (0.91–1.56)0.195
CAD1.34 (0.73–2.48)0.3461.65 (0.75–3.64)0.2140.96 (0.36–2.59)0.940
Smoking0.65 (0.38–1.12)0.1180.75 (0.35–1.61)0.4650.55 (0.25–1.21)0.138
BMI (increase of 5)1.22 (1.02–1.46)0.0300.99 (0.72–1.36)0.9421.32 (1.05–1.66)0.016
Total psoas (low vs middle & high tertiles)1.70 (1.04–2.79)0.0352.42 (1.17–5.01)0.0161.22 (0.62–2.43)0.564

Boldface indicates statistical significance.

Discussion

Key Results

The overall rate of adverse events after lumbar spinal surgery is high, with 30% of patients experiencing some sort of postoperative morbidity. For all patients, the preoperative characteristics of age and tobacco usage were both significantly associated with complications (p = 0.015 and p = 0.026, respectively), while diabetes, CAD, and BMI were not. No surgical characteristics were associated with postoperative morbidity, including whether the surgery required instrumentation, whether it was a revision, or the number of levels treated. Important sex differences were observed. Female sex trended toward greater risk for morbidity (p = 0.068), and BMI was associated with increased morbidity in females (p = 0.021).

For all patients, small total psoas tertiles approached significance for being associated with postoperative morbidity (p = 0.052). These results were significant with multivariate regression analysis, with the lowest psoas tertile having an odds ratio of 1.70 (95% CI 1.04–2.79, p = 0.035) for a postoperative complication for all patients. For male patients, a small total psoas size and being in the lowest psoas tertile were both significant predictors of postoperative morbidity (p = 0.014 and p = 0.008, respectively). For males, the multivariate regression analysis calculated an OR of 2.42 (95% CI 1.17–5.01, p = 0.016). The paraspinal muscle groups did not provide any significant data for postoperative morbidity, even after multivariate analysis. It is important to note that for males, the only variable in the multivariate analysis that was associated with higher risk of complication was being in the lowest psoas tertile.

Interpretation

This is the first time that morphometric analysis has been validated in spinal surgery, illustrating that psoas size may be used as a sensitive preoperative risk factor for perioperative morbidity in patients undergoing lumbar spine surgery. Morphometric analysis of psoas size was more sensitive than analyses of other preoperative risk factors (age, diabetes, smoking, CAD, and BMI) in predicting postoperative morbidity. Psoas size is particularly accurate in predicting postoperative morbidity, with smaller psoas sizes having statistically significant increased odds of complications. This is in agreement with the work of authors in other surgical disciplines.3,7,12,13,17,18,22,24 The association with preoperative psoas size and postoperative morbidity is especially apparent in male patients. It is important to note that our morphometric analysis was not predictive or statistically significant in female patients. There are two reasons for this. First, women have statistically significant smaller psoas muscles than males (Table 2). If the proposed model of sarcopenia and frailty is correct, female sex could represent an indirect risk factor for complications. Indeed, we did observe that women trended toward having an increased rate of complications compared with males (Table 4, p = 0.068). Second, the psoas muscle may not be the best choice to measure sarcopenia and frailty in women. This is evidenced by the fact that the size of the paraspinal muscles, which have an antagonistic action to the psoas muscles, was not predictive of postoperative morbidity in any situation. This suggests that not all muscle groups may be accurate when used for morphometric analysis and that there may be a better, more sensitive muscle group yet to be identified.

It is not surprising that age is associated with increased risk for postoperative complications (p = 0.015). Increasing age is a known risk factor for senescence and sarcopenia, which is indirectly being measured with our morphometric analysis. However, while age is sensitive for predicting sarcopenia, it is not specific, as there are persons of advanced age who have not lost muscle mass.9,21 It is noteworthy that diabetes, CAD, and BMI were not predictive of postoperative morbidity in all patients. Given that the majority of surgery performed in this study was elective in nature, these disease processes may have been under adequate control such that they did not impact the rate of postoperative complications; however, this might not be the case in the setting of an acute neurological deficit following an urgent or semielective operation. Our findings also underscore the relative imprecision of traditional methods for risk stratification in predicting postoperative morbidity. Our multivariate analysis shows that being in the lowest psoas tertile was more accurate at predicting postoperative morbidity than were age and all other premorbid states except for BMI in females. This suggests that morphometrics as a quantitative measure of sarcopenia and frailty may be more indicative of a patient’s general health and their physiological reserve for tolerating surgery.

These findings are especially important in the current health care environment in the United States, as multiple parties (i.e., payers, the federal government, hospital administrators, and regulatory bodies) have begun to focus on the cost and value of care. As payments become bundled by procedure, and health care providers can potentially be penalized or withheld compensation for complications after routine procedures, it will be vital to identify predictors that put patients at a higher risk of morbidity. If a surgeon must perform a needed procedure in a potentially high-risk patient, a reliable model for risk stratification could provide some justification of perioperative morbidity. Regardless of the economic implications, a better understanding of the factors that will put a patient at increased risk of an adverse event after surgery is an important consideration for surgeons as they counsel patients and their families. Setting realistic expectations and providing accurate prognostic information is a key component in the delivery of care.

Limitations

The chief limitation of this study is its retrospective nature and relatively heterogeneous study population. Despite the robustness of our electronic medical record, we cannot account for potential bias and reporting error without standardization in data collection with each individual case. In addition, we considered all lumbar operations as a whole and did not stratify by indication (degenerative vs infection, tumor, trauma, etc.). Our findings highlight the need for future prospective multicenter trials.

Conclusions

Morphometric analysis of psoas size can be used as a sensitive tool to predict perioperative morbidity after lumbar spine surgery. Our results can be generalized to other institutions, as they were based on the experience of 12 surgeons across multiple hospitals, each with different practice patterns and techniques (i.e., minimally invasive surgery versus traditional open techniques).

Author Contributions

Conception and design: Zakaria, Chang, Griffith. Acquisition of data: Zakaria, Chang, Mossa-Basha, Griffith. Analysis and interpretation of data: Chang, Zakaria, Schultz, Griffith. Drafting the article: Zakaria, Chang, Schultz. Critically revising the article: Chang, Zakaria, Schultz, Griffith. Reviewed submitted version of manuscript: Chang, Zakaria, Schultz, Griffith. Approved the final version of the manuscript on behalf of all authors: Chang. Statistical analysis: Schultz. Study supervision: Chang.

References

  • 1

    Amrock LGDeiner S: The implication of frailty on preoperative risk assessment. Curr Opin Anaesthesiol 27:3303352014

  • 2

    Canvasser LDMazurek AACron DCTerjimanian MNChang ETLee CS: Paraspinous muscle as a predictor of surgical outcome. J Surg Res 192:76812014

    • Search Google Scholar
    • Export Citation
  • 3

    Deyo RAMirza SKMartin BIKreuter WGoodman DCJarvik JG: Trends, major medical complications, and charges associated with surgery for lumbar spinal stenosis in older adults. JAMA 303:125912652010

    • Search Google Scholar
    • Export Citation
  • 4

    Dodson RMFiroozmand AHyder OTacher VCosgrove DPBhagat N: Impact of sarcopenia on outcomes following intra-arterial therapy of hepatic malignancies. J Gastrointest Surg 17:212321322013

    • Search Google Scholar
    • Export Citation
  • 5

    Dunne MJAbah UScarci M: Frailty assessment in thoracic surgery. Interact Cardiovasc Thorac Surg 18:6676702014

  • 6

    Englesbe MJLee JSHe KFan LSchaubel DESheetz KH: Analytic morphomics, core muscle size, and surgical outcomes. Ann Surg 256:2552612012

    • Search Google Scholar
    • Export Citation
  • 7

    Englesbe MJPatel SPHe KLynch RJSchaubel DEHarbaugh C: Sarcopenia and mortality after liver transplantation. J Am Coll Surg 211:2712782010

    • Search Google Scholar
    • Export Citation
  • 8

    Fleisher LAEagle KA: Clinical practice. Lowering cardiac risk in noncardiac surgery. N Engl J Med 345:167716822001

  • 9

    Fried LPTangen CMWalston JNewman ABHirsch CGottdiener J: Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci 56:M146M1562001

    • Search Google Scholar
    • Export Citation
  • 10

    García-Miguel FJSerrano-Aguilar PGLópez-Bastida J: Preoperative assessment. Lancet 362:174917572003

  • 11

    Goz VWeinreb JHMcCarthy ISchwab FLafage VErrico TJ: Perioperative complications and mortality after spinal fusions: analysis of trends and risk factors. Spine (Phila Pa 1976) 38:197019762013

    • Search Google Scholar
    • Export Citation
  • 12

    Hasselager RGögenur I: Core muscle size assessed by perioperative abdominal CT scan is related to mortality, postoperative complications, and hospitalization after major abdominal surgery: a systematic review. Langenbecks Arch Surg 399:2872952014

    • Search Google Scholar
    • Export Citation
  • 13

    Lee JSHe KHarbaugh CMSchaubel DESonnenday CJWang SC: Frailty, core muscle size, and mortality in patients undergoing open abdominal aortic aneurysm repair. J Vasc Surg 53:9129172011

    • Search Google Scholar
    • Export Citation
  • 14

    Makary MASegev DLPronovost PJSyin DBandeen-Roche KPatel P: Frailty as a predictor of surgical outcomes in older patients. J Am Coll Surg 210:9019082010

    • Search Google Scholar
    • Export Citation
  • 15

    Manchikanti LSingh VDatta SCohen SPHirsch JA: Comprehensive review of epidemiology, scope, and impact of spinal pain. Pain Physician 12:E35E702009

    • Search Google Scholar
    • Export Citation
  • 16

    Oresanya LBLyons WLFinlayson E: Preoperative assessment of the older patient: a narrative review. JAMA 311:211021202014

  • 17

    Peng PHyder OFiroozmand AKneuertz PSchulick RDHuang D: Impact of sarcopenia on outcomes following resection of pancreatic adenocarcinoma. J Gastrointest Surg 16:147814862012

    • Search Google Scholar
    • Export Citation
  • 18

    Peng PDvan Vledder MGTsai Sde Jong MCMakary MNg J: Sarcopenia negatively impacts short-term outcomes in patients undergoing hepatic resection for colorectal liver metastasis. HPB (Oxford) 13:4394462011

    • Search Google Scholar
    • Export Citation
  • 19

    Rajaee SSKanim LEBae HW: National trends in revision spinal fusion in the USA: patient characteristics and complications. Bone Joint J 96-B:8078162014

    • Search Google Scholar
    • Export Citation
  • 20

    Rubin DI: Epidemiology and risk factors for spine pain. Neurol Clin 25:3533712007

  • 21

    Ruiz MCefalu CReske T: Frailty syndrome in geriatric medicine. Am J Med Sci 344:3953982012

  • 22

    Sabel MSTerjimanian MConlon ASGriffith KAMorris AMMulholland MW: Analytic morphometric assessment of patients undergoing colectomy for colon cancer. J Surg Oncol 108:1691752013

    • Search Google Scholar
    • Export Citation
  • 23

    Scandrett KGZuckerbraun BSPeitzman AB: Operative risk stratification in the older adult. Surg Clin North Am 95:1491722015

  • 24

    Sheetz KHZhao LHolcombe SAWang SCReddy RMLin J: Decreased core muscle size is associated with worse patient survival following esophagectomy for cancer. Dis Esophagus 26:7167222013

    • Search Google Scholar
    • Export Citation
  • 25

    Shen YSilverstein JCRoth S: In-hospital complications and mortality after elective spinal fusion surgery in the united states: a study of the nationwide inpatient sample from 2001 to 2005. J Neurosurg Anesthesiol 21:21302009

    • Search Google Scholar
    • Export Citation
  • 26

    Smetana GWLawrence VACornell JE: Preoperative pulmonary risk stratification for noncardiothoracic surgery: systematic review for the American College of Physicians. Ann Intern Med 144:5815952006

    • Search Google Scholar
    • Export Citation
  • 27

    Street JTLenehan BJDiPaola CPBoyd MDKwon BKPaquette SJ: Morbidity and mortality of major adult spinal surgery. A prospective cohort analysis of 942 consecutive patients. Spine J 12:22342012

    • Search Google Scholar
    • Export Citation
  • 28

    Taylor VMDeyo RACherkin DCKreuter W: Low back pain hospitalization. Recent United States trends and regional variations. Spine (Phila Pa 1976) 19:120712131994

    • Search Google Scholar
    • Export Citation
  • 29

    Trayner E JrCelli BR: Postoperative pulmonary complications. Med Clin North Am 85:112911392001

  • 30

    van Meenen LCvan Meenen DMde Rooij SEter Riet G: Risk prediction models for postoperative delirium: a systematic review and meta-analysis. J Am Geriatr Soc 62:238323902014

    • Search Google Scholar
    • Export Citation
  • 31

    Waits SAKim EKTerjimanian MNTishberg LMHarbaugh CMSheetz KH: Morphometric age and mortality after liver transplant. JAMA Surg 149:3353402014

    • Search Google Scholar
    • Export Citation
  • 32

    Weinstein JNLurie JDTosteson TDHanscom BTosteson ANBlood EA: Surgical versus nonsurgical treatment for lumbar degenerative spondylolisthesis. N Engl J Med 356:225722702007

    • Search Google Scholar
    • Export Citation
  • 33

    Weinstein JNLurie JDTosteson TDSkinner JSHanscom BTosteson AN: Surgical vs nonoperative treatment for lumbar disk herniation: the Spine Patient Outcomes Research Trial (SPORT) observational cohort. JAMA 296:245124592006

    • Search Google Scholar
    • Export Citation
  • 34

    Weinstein JNTosteson TDLurie JDTosteson ANBlood EHanscom B: Surgical versus nonsurgical therapy for lumbar spinal stenosis. N Engl J Med 358:7948102008

    • Search Google Scholar
    • Export Citation
  • 35

    Yoshihara HYoneoka D: National trends in the surgical treatment for lumbar degenerative disc disease: United States, 2000 to 2009. Spine J 15:2652712015

    • Search Google Scholar
    • Export Citation

If the inline PDF is not rendering correctly, you can download the PDF file here.

Article Information

Correspondence Victor Chang, Henry Ford West Bloomfield Hospital, Department of Neurosurgery, 6777 W. Maple Blvd., West Bloomfield Township, MI 48322. email: vicchang@gmail.com.

INCLUDE WHEN CITING DOI: 10.3171/2015.7.FOCUS15257.

Disclosure Dr. Chang reports being a consultant for Globus Medical.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Examples of morphometric measurements on T2-weighted MR images obtained in patient with (A–C) and without (D–F) sarcopenia. A: Sagittal image of a patient with sarcopenia. The white lines through the L-4 and T-12 vertebrae are example cross-sections from which morphometric measurements were taken. B: Axial image, of the same patient, taken at the L-4 vertebra showing a small psoas muscle measuring 5.33 cm2. C: Axial image, of the same patient, obtained at the T-12 vertebra showing a small paraspinal muscle measuring 9.32 cm2. D: Sagittal image of a patient without sarcopenia. The white lines through the L-4 and T-12 vertebrae again show example cross-sections from which morphometric measurements were taken. E: Axial image, of the patient, demonstrating a large psoas muscle measuring 21.48 cm2. F: Axial image, of the same patient, demonstrating a large paraspinal muscle measuring 23.32 cm2.

References

  • 1

    Amrock LGDeiner S: The implication of frailty on preoperative risk assessment. Curr Opin Anaesthesiol 27:3303352014

  • 2

    Canvasser LDMazurek AACron DCTerjimanian MNChang ETLee CS: Paraspinous muscle as a predictor of surgical outcome. J Surg Res 192:76812014

    • Search Google Scholar
    • Export Citation
  • 3

    Deyo RAMirza SKMartin BIKreuter WGoodman DCJarvik JG: Trends, major medical complications, and charges associated with surgery for lumbar spinal stenosis in older adults. JAMA 303:125912652010

    • Search Google Scholar
    • Export Citation
  • 4

    Dodson RMFiroozmand AHyder OTacher VCosgrove DPBhagat N: Impact of sarcopenia on outcomes following intra-arterial therapy of hepatic malignancies. J Gastrointest Surg 17:212321322013

    • Search Google Scholar
    • Export Citation
  • 5

    Dunne MJAbah UScarci M: Frailty assessment in thoracic surgery. Interact Cardiovasc Thorac Surg 18:6676702014

  • 6

    Englesbe MJLee JSHe KFan LSchaubel DESheetz KH: Analytic morphomics, core muscle size, and surgical outcomes. Ann Surg 256:2552612012

    • Search Google Scholar
    • Export Citation
  • 7

    Englesbe MJPatel SPHe KLynch RJSchaubel DEHarbaugh C: Sarcopenia and mortality after liver transplantation. J Am Coll Surg 211:2712782010

    • Search Google Scholar
    • Export Citation
  • 8

    Fleisher LAEagle KA: Clinical practice. Lowering cardiac risk in noncardiac surgery. N Engl J Med 345:167716822001

  • 9

    Fried LPTangen CMWalston JNewman ABHirsch CGottdiener J: Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci 56:M146M1562001

    • Search Google Scholar
    • Export Citation
  • 10

    García-Miguel FJSerrano-Aguilar PGLópez-Bastida J: Preoperative assessment. Lancet 362:174917572003

  • 11

    Goz VWeinreb JHMcCarthy ISchwab FLafage VErrico TJ: Perioperative complications and mortality after spinal fusions: analysis of trends and risk factors. Spine (Phila Pa 1976) 38:197019762013

    • Search Google Scholar
    • Export Citation
  • 12

    Hasselager RGögenur I: Core muscle size assessed by perioperative abdominal CT scan is related to mortality, postoperative complications, and hospitalization after major abdominal surgery: a systematic review. Langenbecks Arch Surg 399:2872952014

    • Search Google Scholar
    • Export Citation
  • 13

    Lee JSHe KHarbaugh CMSchaubel DESonnenday CJWang SC: Frailty, core muscle size, and mortality in patients undergoing open abdominal aortic aneurysm repair. J Vasc Surg 53:9129172011

    • Search Google Scholar
    • Export Citation
  • 14

    Makary MASegev DLPronovost PJSyin DBandeen-Roche KPatel P: Frailty as a predictor of surgical outcomes in older patients. J Am Coll Surg 210:9019082010

    • Search Google Scholar
    • Export Citation
  • 15

    Manchikanti LSingh VDatta SCohen SPHirsch JA: Comprehensive review of epidemiology, scope, and impact of spinal pain. Pain Physician 12:E35E702009

    • Search Google Scholar
    • Export Citation
  • 16

    Oresanya LBLyons WLFinlayson E: Preoperative assessment of the older patient: a narrative review. JAMA 311:211021202014

  • 17

    Peng PHyder OFiroozmand AKneuertz PSchulick RDHuang D: Impact of sarcopenia on outcomes following resection of pancreatic adenocarcinoma. J Gastrointest Surg 16:147814862012

    • Search Google Scholar
    • Export Citation
  • 18

    Peng PDvan Vledder MGTsai Sde Jong MCMakary MNg J: Sarcopenia negatively impacts short-term outcomes in patients undergoing hepatic resection for colorectal liver metastasis. HPB (Oxford) 13:4394462011

    • Search Google Scholar
    • Export Citation
  • 19

    Rajaee SSKanim LEBae HW: National trends in revision spinal fusion in the USA: patient characteristics and complications. Bone Joint J 96-B:8078162014

    • Search Google Scholar
    • Export Citation
  • 20

    Rubin DI: Epidemiology and risk factors for spine pain. Neurol Clin 25:3533712007

  • 21

    Ruiz MCefalu CReske T: Frailty syndrome in geriatric medicine. Am J Med Sci 344:3953982012

  • 22

    Sabel MSTerjimanian MConlon ASGriffith KAMorris AMMulholland MW: Analytic morphometric assessment of patients undergoing colectomy for colon cancer. J Surg Oncol 108:1691752013

    • Search Google Scholar
    • Export Citation
  • 23

    Scandrett KGZuckerbraun BSPeitzman AB: Operative risk stratification in the older adult. Surg Clin North Am 95:1491722015

  • 24

    Sheetz KHZhao LHolcombe SAWang SCReddy RMLin J: Decreased core muscle size is associated with worse patient survival following esophagectomy for cancer. Dis Esophagus 26:7167222013

    • Search Google Scholar
    • Export Citation
  • 25

    Shen YSilverstein JCRoth S: In-hospital complications and mortality after elective spinal fusion surgery in the united states: a study of the nationwide inpatient sample from 2001 to 2005. J Neurosurg Anesthesiol 21:21302009

    • Search Google Scholar
    • Export Citation
  • 26

    Smetana GWLawrence VACornell JE: Preoperative pulmonary risk stratification for noncardiothoracic surgery: systematic review for the American College of Physicians. Ann Intern Med 144:5815952006

    • Search Google Scholar
    • Export Citation
  • 27

    Street JTLenehan BJDiPaola CPBoyd MDKwon BKPaquette SJ: Morbidity and mortality of major adult spinal surgery. A prospective cohort analysis of 942 consecutive patients. Spine J 12:22342012

    • Search Google Scholar
    • Export Citation
  • 28

    Taylor VMDeyo RACherkin DCKreuter W: Low back pain hospitalization. Recent United States trends and regional variations. Spine (Phila Pa 1976) 19:120712131994

    • Search Google Scholar
    • Export Citation
  • 29

    Trayner E JrCelli BR: Postoperative pulmonary complications. Med Clin North Am 85:112911392001

  • 30

    van Meenen LCvan Meenen DMde Rooij SEter Riet G: Risk prediction models for postoperative delirium: a systematic review and meta-analysis. J Am Geriatr Soc 62:238323902014

    • Search Google Scholar
    • Export Citation
  • 31

    Waits SAKim EKTerjimanian MNTishberg LMHarbaugh CMSheetz KH: Morphometric age and mortality after liver transplant. JAMA Surg 149:3353402014

    • Search Google Scholar
    • Export Citation
  • 32

    Weinstein JNLurie JDTosteson TDHanscom BTosteson ANBlood EA: Surgical versus nonsurgical treatment for lumbar degenerative spondylolisthesis. N Engl J Med 356:225722702007

    • Search Google Scholar
    • Export Citation
  • 33

    Weinstein JNLurie JDTosteson TDSkinner JSHanscom BTosteson AN: Surgical vs nonoperative treatment for lumbar disk herniation: the Spine Patient Outcomes Research Trial (SPORT) observational cohort. JAMA 296:245124592006

    • Search Google Scholar
    • Export Citation
  • 34

    Weinstein JNTosteson TDLurie JDTosteson ANBlood EHanscom B: Surgical versus nonsurgical therapy for lumbar spinal stenosis. N Engl J Med 358:7948102008

    • Search Google Scholar
    • Export Citation
  • 35

    Yoshihara HYoneoka D: National trends in the surgical treatment for lumbar degenerative disc disease: United States, 2000 to 2009. Spine J 15:2652712015

    • Search Google Scholar
    • Export Citation

TrendMD

Cited By

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 288 284 54
PDF Downloads 203 181 9
EPUB Downloads 0 0 0

PubMed

Google Scholar