Enhanced recovery after surgery (ERAS) improves return of physiological function in frail patients undergoing 1-2 level TLIFs: an observational retrospective cohort study

Abstract

BACKGROUND CONTEXT:

The enhanced recovery after surgery (ERAS) protocol is a multimodal approach which has been shown to facilitate recovery of physiological function, and reduce early post-operative pain, complications, and length of stay (LOS) in open 1–2 level TLIF. The benefit of ERAS in specifically frail patients undergoing TLIF has not been demonstrated. Frailty is clinically defined as a syndrome of physiological decline that can predispose patients undergoing surgery to poor outcomes.

PURPOSE:

This study primarily evaluated the benefit of an ERAS protocol in frail patients undergoing 1- or 2-level open TLIF compared to frail patients without ERAS. Secondarily, we assessed whether outcomes in frail patients with ERAS approximated those seen in nonfrail patients with ERAS.

STUDY DESIGN:

Retrospective consecutive patient cohort with controls propensity-matched for age, body mass index, sex, and smoking status.

PATIENT SAMPLE:

Consecutive patients that underwent 1- or 2-level open TLIF for degenerative disease from August, 2015 to July, 2021 by a single surgeon. ERAS was implemented in December 2018.

OUTCOME MEASURES:

Primary outcome measure was return of postoperative physiological function defined as the summation of first day to ambulate, first day to bowel movement, and first day to void. Additional outcome measures included LOS, daily average pain scores, opioid use, discharge disposition, 30-day readmission rate, and reoperation.

METHODS:

A retrospective analysis of frail patients > 65 years of age undergoing 1–2 level open TLIF post-ERAS were compared to propensity matched frail pre-ERAS patients. Frailty was assessed using the Fried phenotype classification (score > 1). Patient demographics, LOS, first-day-to-ambulate (A1), first-day-to-bowel movement (B1), first-day-to-void (V1) were collected. Return of physiological function was defined as A1+B1+V1. Primary analysis was a comparison of frail patients pre-ERAS versus post-ERAS to determine effect of ERAS on return of physiologic function with frailty. Secondary analysis was a comparison of post-ERAS frail versus post-ERAS nonfrail patients to determine if return of physiologic function in frail patients with ERAS approximates that of nonfrail patients.

RESULTS:

In the primary analysis, 32 frail patients were included with mean age ± standard deviation of 72.8±4.4 years, mean BMI 28.8±5.5, 65.6% were male, 15 pre-ERAS and 17 post-ERAS. Patient characteristics were similar between groups. After ERAS implementation, return of physiological function improved by a mean 3.2 days overall (post-ERAS 3.4 vs. pre-ERAS 6.7 days) (p<.0001), indicating a positive effect of ERAS in frail patients. Additionally, length of stay improved by 1 day (4.8±1.6 vs. 3.8±1.9 days, p<.0001). Total daily intravenous morphine milligram equivalent (MME) as well as average daily pain scores were similar between groups. Secondarily, 26 nonfrail patients post ERAS were used as a comparison group with the 17 post-ERAS frail cohort. Mean age of this cohort was 73.4±4.6 years, mean BMI 27.4±4.9, and 61.9% were male. Return of physiologic function was similar between cohorts (post-ERAS nonfrail 3.5 vs. post-ERAS frail 3.4 days) (p=.938), indicating the benefit with ERAS in frail patients approximates that of nonfrail patients.

CONCLUSIONS:

ERAS significantly improves return of physiologic function and length of stay in patients with frailty after 1-2 level TLIF, and approximates improved outcomes seen in non-frail patients.

Introduction

The enhanced recovery after surgery (ERAS) protocol is an evidence-based approach to promote recovery and reduce length of hospitalization, postoperative complications, and stress response in patients undergoing surgical intervention. The ERAS protocol has been used successfully in various surgical specialties since its inception in 1997; many components of the protocol were first utilized in spine surgery in the mid-2000s [1-8]. Some of the key components of the ERAS protocol include its cost-effectiveness and safety [3]. Important features of the ERAS protocol include preoperative education, preoperative medical optimization, nausea control, early removal of drains, multimodal postoperative pain control, early postoperative mobilization, and nutrition management [4]. The protocol has since been widely adapted to various spinal surgical procedures, including microdiscectomy, cervical fusion, deformity correction, and elective spinal fusion [1,9-15]. Previous studies have shown that the ERAS protocol for lumbar spine fusion can reduce postoperative complication rates, limit postoperative pain, and reduce hospitalization length of stay, while promoting physiological function recovery [7-9,16-18].

There is increasing focus on improving perioperative outcomes in particularly vulnerable patient populations. Many indicators for high-risk patients including postoperative risk calculators have been developed [19-21]. Frailty is one high risk patient factor characterized by decreased physiological function, and has been shown to predispose patients to adverse outcomes [22-24]. Specifically, after spine surgery, frail patients have increased risk of worse outcome, higher complications, and longer hospital stay [25-30]. The Fried phenotype is a commonly used standardized phenotype of frailty that includes five measurements: grip strength, walking speed, weight loss, low physical activity, and exhaustion [31], and has been widely studied in various patient populations undergoing surgical intervention [10,32,33].

Transforaminal lumbar interbody fusion (TLIF) is a widely-used approach for surgical treatment of lumbar degenerative disease [34-37]. Advantages of the TLIF approach include circumferential neural decompression, restoration of disc space height, and augmenting posterior construct rigidity [38,39]. Previous studies have shown that an ERAS protocol is effective in spine patients, but there are no studies to date that have evaluated ERAS in frail patients [1,10,40-43]. The aims of this study are: 1. To compare physiological outcomes, length of stay, pain score, and opioid consumption in frail patients undergoing open 1- or 2-level TLIF before vs. after implementing a standardized ERAS protocol; 2. To determine if outcomes in frail patients with ERAS approximates that of non-frail patients with ERAS.

Methods

ERAS program

The current ERAS protocol was fully implemented by the neurosurgery department in December 2018, and has been previously published [18].

Population

This study was approved by the Institutional Review Board (IRB# 202000839). Informed consent was waived per the study design. A consecutive series of patients who underwent 1- or 2-level open TLIF by a single surgeon (DJH) before and after implementing an ERAS protocol were included for analysis. Inclusion criteria was surgery limited to 1- or 2-level TLIF for spondylolisthesis, spinal stenosis, nerve root compression, recurrent disc herniation, pseudoarthrosis, or adjacent segment disease in patients age > 65 years in which a Fried phenotype score was collected in the preoperative visit. Frailty was defined as having two or more frailty index components [31,44]. Exclusion criteria was surgery for or significant past history of neoplasm, trauma, infection, or concomitant cervical spine surgery or absence of a recorded Fried phenotype. ERAS-related order sets were confirmed to have been ordered in participants of the ERAS cohort. For Aim 1 of this study, frail post-ERAS patients were propensity matched (age, body mass index, sex, smoking status) to frail pre-ERAS patients. For Aim 2, frail post-ERAS patients were propensity matched (age, body mass index, sex, smoking status) to nonfrail post-ERAS patients.

Surgical technique

All patients underwent a standard open TLIF technique (as previously described) involving laminectomy with complete bilateral facetectomy, bilateral pedicle screw-rod construct (fluoroscopy assisted), and placement of a 10 or 12 mm nonlordotic cage (Concorde, Expedium, Depuy-Synthes, Raynham, MA, USA) [45]. There were no material differences or changes to the surgical technique, instrumentation, or intraoperative adjuncts during the study period.

Clinical evaluation variables

Recovery of physiological function was the primary outcome and was defined as the summation of time to first-day-to-ambulate (A1), first-day-to-bowel movement (B1), first-day-to-void (V1) (ie, A1+B1+V1). Length of hospital stay was evaluated as a secondary outcome measure. Additionally, pain scores and opioid consumption (in morphine milligram equivalents) were included as secondary outcome measures. Postoperative pain was recorded using the Defense & Veterans Pain Rating Scale (DVPRS), and overall pain scores were calculated as the average DVPRS during the hospital stay. Similarly, opioid consumption was recorded using oral morphine milligram equivalents (MME) and overall consumption was calculated as the average MME during the first three post-operative days. Estimated blood loss, operative time (interval from time of incision to time of closure), discharge location, 30-day readmission, and 30-day reoperation were also recorded. All information was gathered via chart review from physical therapy, occupational therapy, and physician progress notes, nursing notes, and electronic medical record charting.

Statistical methods

The study is primarily powered to detect a difference in physiological function. Based on the number of frail patients in the pre-ERAS cohort and the standard deviation of physiological function pulled from our previous study, a minimum of eight patients in the post-ERAS cohort would be required to detect a 1-day difference approaching a medium effect size (Cohen d = 0.4) with a statistical power of 0.80 and a two-tailed significance level of 0.05. For secondary outcomes, the study would require 18 patients for LOS, 645 for pain scores, 45 for composite complications, and 146 for estimated blood loss. Due to the similarity in opioid consumption from our prior study, a sample size determination was unobtainable. Homogeneity of variances was tested using Bartlett's test, and continuous data were compared across the pre-ERAS and ERAS cohorts using independent two-sample student’s t-test. When appropriate, logarithmic transformation was used to best approximate normal distributions due to skewness. Mann-Whitney U was used when normal distributions were unable to be approximated. Categorical data were compared using a chi-square test. All statistical analyses for this study were performed using SPSS ver. 23 software (SPSS Inc, Chicago, IL, USA). A probability (p) value of <0.05 was considered to indicate statistical significance.

Results

Descriptive data

There were 74 patients for analysis: 43 patients in the post-ERAS cohort from December 2018 to February 2021, and 31 in the pre-ERAS cohort (from August 2015 to December 2018). Sixteen patients were excluded as pre-ERAS non-frail patients. For Aim 1, 17 post-ERAS frail patients were compared to 15 pre-ERAS frail patients. For Aim 2, 17 post-ERAS frail patients were compared to 26 post-ERAS non-frail patients.

Aim 1:. Pre-ERAS versus post-ERAS in patients with frailty

Patient demographics

Demographic patient data and preoperative characteristics are shown in Table 1. Age, BMI, and sex were similar for both cohorts. Additionally, there were no significant differences in antiplatelet/anticoagulant use, smoking status, payer status, preoperative opioid use, anxiety/depression, weekday of admission, or surgical diagnosis. Fried phenotype scores were similar between cohorts (pre-ERAS: 2.4±0.6, post-ERAS: 2.5±0.7, p=.793) with no significant difference in subcategories.

Table 1.

Baseline demographics

Categorical variables: N (%) Continuous variables: mean ± SD	Pre-ERAS Frail	Post-ERAS Frail	Post-ERAS Non-frail	Pre-ERAS Frail vs. post-ERAS Frail p-value*	Post-ERAS Frail vs. Post ERAS Non-Frail p-value†
Patients, n	15	17	26
Age (y)	73.2 ± 4.7	72.5 ± 4.2	73.4 ± 4.6	.603	.465
Body mass index (kg/m2)	27.9 ± 4.5	29.5 ± 5.6	29.5 ± 4.8	.398	.210
Charlson comorbidity index	1.6 ± 1.3	1.2 ± 2.0	0.9 ± 1.1	.274	.589
Male	12 (80)	9 (52.9)	10 (38.5)	.147	.753
Fried phenotype	2.4 ± 0.6	2.5 ± 0.7	0.3 ± 0.5	.793	<.0001
No frailty
0	n/a	n/a	19 (73.1)	n/a	n/a
Pre-frail
1	n/a	n/a	7 (26.9)
2	10 (66.7)	11 (64.7	n/a	.793‡	n/a
Frail
3	4 (26.7)	4 (23.5)	n/a
4	1 (3.2)	2 (4.7)	n/a
5	0	0	n/a
Categories:
Unintended weight loss	3 (20.0)	1 (5.9)	0 (0)	.319	.395
Feels exhausted	7 (46.7)	5 (29.4)	1 (3.8)	.467	.028
Weak grip strength	3 (20.0)	9 (52.9)	2 (7.7)	.076	.003
Slow walking speed	11 (73.3)	14 (82.4)	3 (11.5)	.678	<.0001
Low physical activity	12 (80)	13 (76.5)	1 (3.8)	1.000	<.0001
Modified frailty index 11	1.9 ± 0.3	1.9 ± 1.5	1.3 ± 1.0	0.913	0.138
Anxiety	7 (46.7)	3 (17.6)	10 (38.5)	.128	.187
Depression	5 (33.3)	5 (29.4)	9 (34.6)	1.000	.753
L1 Hounsfield units	129 ± 27 N=4	130 ± 42 N=12	131 ± 41 N=7	0.862	0.837
Smoker status
Never	8 (53.3)	12 (70.6)	14 (53.8)	.467	.347
Former	7 (46.7)	5 (29.4)	12 (46.2)
Payer status
Medicare	13 (86.7)	16 (94.1)	21 (80.8)	.726‡	.778‡
Private	1 (6.7)	1 (5.9)	4 (15.4)
Managed care	1 (6.7)	0 (0)	0 (0)
Worker’s comp	0 (0)	0 (0)	1 (3.8)
Preoperative MME	10.9 ± 27.8	10.5 ± 22.9	7.4 ± 14.1	.908	.680
Opioid naïve (0 MME)	11 (73.3)	10 (58.8)	19 (73.1)	.472	.507
Low consumption (0–60 MME)	3 (20.0)	6 (35.3)	7 (26.9)	.444	.736
High consumption (60–90 MME)	0 (0)	1 (5.9)	0 (0)	.531	.395
Very high consumption (>90 MME)	1 (6.7)	0 (0)	0 (0)	.469	n/a
Antiplatelet/Anticoagulation use	2 (13.3)	5 (29.4)	7 (26.9)	.402	1.000
Admission day
Tuesday	4 (40.0)	7 (41.2)	10 (38.5)	.742‡	1.000‡
Wednesday	2 (20.0)	6 (35.3)	10 (38.5)
Thursday	4 (40.0)	4 (23.5)	6 (23.1)
Surgical diagnosis
Spondylolisthesis	7 (46.7)	10 (58.8)	18 (69.2)	.179‡	.437‡
Recurrent disc herniation	3 (20.0)	0 (0)	2 (7.7)
Adjacent segment disease	0 (0)	2 (11.8)	1 (3.8)
Pseudoarthrosis	1 (6.7)	2 (11.8)	1 (3.8)
Foraminal stenosis	2 (13.3)	1 (5.9)	3 (11.5)
Scoliosis	2 (13.3)	0 (0)	0 (0)
Other disc or facet disease	0 (0)	2 (11.8)	1 (3.8)

Bolded values represent p<.05.

^*p-value comparing pre-ERAS to ERAS in frail patients.

^†p-value comparing frail to nonfrail in post-ERAS patients.

^‡p-value represents the difference in distribution of variable among categories.

Surgical data

Surgical characteristics are shown in Table 2. Number of levels were similar between groups and the vertebral level distribution was not significantly different (p=.094). Operative time was not significantly different (pre-ERAS: 160.1±47.0 minute., post-ERAS: 147.9±51.6 min., p=.493). Intraoperative estimated blood loss was not significantly different (pre-ERAS: 527±273 mL, post-ERAS: 579±528 mL, p=.780). Incidence of durotomy trended toward reduction post-ERAS, but was not found to be significant (20% vs. 0%, p=.092). Total number of postoperative complications were not significantly different after implementation of ERAS (pre-ERAS: 1.5±0.9, post-ERAS: 1.1±0.9, p=.762). The only individual complication to reach significance was uncontrolled blood glucose with 4 occurrences in the pre-ERAS group (26.7%) and none post-ERAS (p=.038).

Table 2.

Surgical characteristics and post-operative primary and secondary outcomes

Categorical variables: N (%) Continuous variables: mean ± SD	Pre-ERAS Frail	Post-ERAS Frail	Post-ERAS Non-frail	Pre-ERAS Frail vs. post-ERAS Frail p-value*	Post-ERAS Frail vs. Post ERAS Non-Frail p-value†
TLIF level distribution
One-Level	9 (60.0)	11 (64.7)	22 (84.6)	1.000	.158
L3–L4 (% of one-level)	0 (0)	4 (36.4)	4 (18.2)	.094‡	.456‡
L4–L5 (% of one-level)	8 (88.9)	7 (63.6)	16 (82.7)
L5–S1 (% of one-level)	1 (11.1)	0 (0)	2 (9.1)
Two-Level	6 (40.0)	6 (35.3)	4 (15.4)	1.000	.158
L2–L4 (% of two-level)	0 (0)	1 (16.7)	0 (0)	1.000‡	1.000‡
L3–L5 (% of two-level)	1 (16.7)	1 (16.7)	1 (25.0)
L4–S1 (% of two-level)	5 (83.3)	4 (66.7)	3 (75.0)
Drain placement	15 (100)	10 (58.8)	9 (34.6)	.008	.209
Indwelling catheter placement	9 (60.0)	3 (17.6)	5 (19.2)	.027	1.000
Estimated blood loss (mL)	527 ± 273	579 ± 528	340 ± 141	.780	.126
Operative duration (minutes)	160.1 ± 47.0	147.9 ± 51.6	138.9 ± 31.6	.493
Complications	1.5 ± 0.9	1.1 ± 0.9	0.6 ± 0.8	.762	.020
Intraoperative durotomy	3 (20.0)	0 (0)	1 (3.8)	.092	1.000
Postoperative nausea/vomiting	10 (66.7)	8 (47.1)	9 (34.6)	.308	.528
Urinary retention	9 (60.0)	7 (41.2)	7 (26.9)	.479	.507
Hemodynamic instability	2 (13.3)	5 (29.4)	2 (7.7)	.402	.093
Atelectasis	3 (20.0)	1 (5.9)	1 (3.8)	.319	1.000
Uncontrolled blood glucose	4 (26.7)	0 (0)	0 (0)	.038	n/a
Acute kidney injury	1 (6.7)	2 (11.8)	0 (0)	1.000	.151
UTI	1 (6.7)	1 (5.9)	1 (3.8)	1.000	1.000
Altered mental status	0 (0)	1 (5.9)	1 (3.8)	1.000	1.000
Ileus	2 (13.3)	0 (0)	0 (0)	.212	n/a
Surgical site infection	0 (0)	1 (5.9)	0 (0)	1.000	.395
Delirium	0 (0)	0 (0)	1 (3.8)	n/a	1.000
Deep venous thrombosis	0 (0)	0 (0)	0 (0)	n/a	n/a
Pneumonia	0 (0)	0 (0)	0 (0)	n/a	n/a
Myocardial infarction	0 (0)	0 (0)	0 (0)	n/a	n/a
Hypoxia	0 (0)	0 (0)	0 (0)	n/a	n/a
Discharge disposition
Home	9 (60.0)	11 (64.7)	15 (57.7)	1.000	.755
Home with home health	1 (6.7)	2 (11.8)	9 (34.6)	1.000	.154
Skilled nursing facility	2 (13.3)	3 (17.6)	1 (3.8)	1.000	.284
Inpatient rehab	3 (20.0)	1 (5.9)	1 (3.8)	.319	1.000
Drain removal day	3.6 ± 0.9	3.5 ± 1.7	3.6 ± 0.9	.851	.932
Return of Physiological Function
1st ambulation POD	1.6 ± 1.0	0.7 ± 0.5	0.7 ± 0.5	.002	.764
1st bowel movement POD	3.0 ± 0.9	2.3 ± 0.9	2.3 ± 1.1	.058	.852
No bowel movement prior to discharge	2 (13.3)	6 (35.3)	12 (46.2)	.229	.541
1st void POD	2.1 ± 2.0	0.3 ± 0.6	0.4 ± 0.8	<.0001	.983
Length of stay	4.8 ± 1.6	3.8 ± 1.9	3.3 ± 1.3	.050	.375
30-d readmission	0 (0)	1 (5.9)	0 (0)	1.000	.395
30-d reoperation	0 (0)	1 (5.9)	0 (0)	1.000	.395

Bolded values represent p<.05.

^*p-value comparing pre-ERAS to post-ERAS in frail patients.

^†p-value comparing frail to nonfrail in post-ERAS patients.

^‡p-value represents the difference in distribution of variable among categories.

ERAS components

Minimizing intraoperative drain placement, foley catheter use, and postoperative patient-controlled analgesia (PCA) are primary elements of the ERAS protocol. The percentage of drains placed decreased from 100% pre-ERAS to 59% post-ERAS (p=.008). The average day of drain removal (for those patients with drains placed) was not significantly different between groups (3.6±0.9 vs. 3.5±1.7, p=.851). The proportion of subjects with indwelling urinary catheters placed decreased from 60% pre-ERAS to 18% post-ERAS (p=.027). PCA use decreased from 80% pre-ERAS to 0% post-ERAS (p<.0001).

Physiological function

The first day of assisted-walking occurred on average 0.9 days earlier with ERAS (1.6 days vs. 0.7 days, p=.002). First bowel movement occurred on average 0.7 days earlier (3.0 days vs. 2.3 days, p=.058). First day of bladder voiding occurred on average 1.8 days earlier (2.1 days vs. 0.3 days, p<.0001). Of note, 35% of patients were discharged without having a bowel movement post-ERAS compared to 13% pre-ERAS. Overall, the sum total days for recovery of physiological function was found to be significantly improved after implementation of ERAS (pre-ERAS: 6.7 days, post-ERAS: 3.4 days, p<.0001) (Fig. 1). In post-hoc analyses of the subgroups of the Fried phenotype (not shown in the tables), physiological function was strongly associated with those patients that exhibited unintended weight loss (p=.004) or weak grip strength (p=.033).

Fig. 1. Physiologic function.

Stacked bar graph showing physiologic function in days for pre-ERAS frail, post-ERAS frail, and post-ERAS nonfrail cohorts.

Pain and opioid consumption

Overall and daily pain scores and opioid consumption are described in Table 3 with no significant difference in average daily pain scores (Fig. 2). Similarly, 72-hour post-operative opioid consumption (MME) was not found to be significantly different (35.6 ± 22.5 vs. 40.2 ± 25.1, p=.526).

Table 3.

Postoperative pain and opioid use

mean ± SD	Pre-ERAS Frail	Post-ERAS Frail	Post-ERAS Non-frail	Pre-ERAS Frail vs. post-ERAS Frail p-value*	Post-ERAS Frail vs. Post ERAS Non-Frail p-value†
Patient controlled analgesia, N (%)	12 (80.0)	0 (0)	0 (0)	<.0001	<.0001
Postoperative mean pain scores (DVPRS)
POD 0	5.1 ± 1.8	4.9 ± 1.7	5.2 ± 1.7	.660	.517
POD 1	3.6 ± 2.1	5.0 ± 1.6	4.8 ± 1.2	.048	.674
POD 2	3.2 ± 2.1	3.9 ± 2.2	4.3 ± 1.6	.422	.466
POD 3	3.1 ± 2.3	3.9 ± 1.9	3.8 ± 1.7	.329	.837
Hospital Stay	3.6 ± 1.9	4.3 ± 1.6	4.6 ± 1.3	.238	.574
Total postoperative opioid use (MME)
POD 0	48.4 ± 28.3	57.1 ± 33.0	61.5 ± 39.9	.394	.876
POD 1	36.2 ± 37.4	45.7 ± 26.0	53.1 ± 33.7	.112	.651
POD 2	28.5 ± 25.9	29.1 ± 31.1	37.7 ± 35.5	.902	.473
POD 3	25.1 ± 17.5	20.3 ± 17.8	33.3 ± 31.8	.563	.364
Daily Average	35.6 ± 22.5	40.2 ± 25.1	50.8 ± 31.0	.526	.431
Total postoperative oral opioid use (MME)
POD 0	19.1 ± 16.1	40.6 ± 22.5	46.1 ± 28.9	.006	.703
POD 1	18.3 ± 31.3	43.4 ± 25.1	50.2 ± 30.6	.002	.416
POD 2	28.1 ± 22.8	27.8 ± 29.9	36.9 ± 35.7	.889	.466
POD 3	26.9 ± 16.7	20.3 ± 17.8	32.6 ± 30.7	.350	.375
Daily average	22.9 ± 17.1	34.0 ± 21.1	45.5 ± 25.1	.119	.149
Total Post-operative intravenous opioid use (MME)
POD 0	29.3 ± 26.4	16.5 ± 17.4	17.2 ± 20.3	.072	.983
POD 1	17.9 ± 28.1	2.3 ± 4.9	4.7 ± 8.1	.009	.349
POD 2	2.2 ± 6.4	1.2 ± 3.4	0.8 ± 2.8	.712	1.000
POD 3	0 ±0	0 ± 0	0.7 ± 2.9	n/a	.391
Daily average	12.7 ± 13.2	6.2 ± 8.4	7.1 ± 6.6	.062	.484

Bolded values represent p<.05.

^*p-value comparing pre-ERAS to post-ERAS in frail patients.

^†p-value comparing frail to non-frail in post-ERAS patients.

Fig. 2. Pain Scores and Opioid Use.

(A) Average daily Defense and Veterans Pain Rating Scale (DVPRS) scores over post-operative day (POD) 0-3. (B) Total daily average morphine milligram equivalents (MME) over postoperative day (POD) 0-3. There was no significant change with ERAS. Dashed line=pre-ERAS frail cohort; solid line = post-ERAS frail cohort; gray line = post-ERAS non-frail cohort. * = p<0.05

Stratifying opioid usage by method of administration revealed a significant difference in IV and oral opioid usage similarly observed in previously published data [18]. Daily IV opioid consumption decreased from POD 0 to POD 3, down-trending monotonically in both groups, however, there was a significant increase in oral consumption on POD 3 for the Pre-ERAS cohort (generally coinciding with PCA discontinuation).

Length of stay and discharge

Length of stay was significantly reduced from 4.8 ± 1.6 days (median 4 days) pre-ERAS to 3.8±1.9 (median 3 days) post-ERAS (p<.0001). Non-home discharge rates trended toward reduction post-ERAS but was not found to be significant (33.3% vs. 23.5%, p=.699). Discharge disposition matched physical therapy and occupational therapy recommendations for all patients. One patient underwent reoperation within 30 days requiring readmission in the post-ERAS cohort (5.9%), for instrumentation failure/wound dehiscence. There were no 30-day reoperations or readmissions in the pre-ERAS cohort.

Aim 2:. Frailty versus Non-Frailty in post-ERAS patients

Patient demographics

Demographic patient data and preoperative characteristics are shown in Table 1. Age, BMI, and sex were similar for both cohorts. Additionally, there was no significant difference in antiplatelet/anticoagulant use, smoking status, payer status, preoperative opioid use, anxiety/depression, weekday of admission, or surgical diagnosis. Fried phenotype score was expectedly different (Frail: 2.5±0.7, Non-Frail: 0.3±0.5, p<.0001) in all subcategories except unintended weight loss.

Surgical data

Surgical characteristics are shown in Table 2. Number of levels were similar between groups and the vertebral level distribution was not significantly different (p=.456). Operative time was not significantly different (Frail: 147.9±51.6 minutes, Non-Frail: 138.9±31.6, p=.479). Intraoperative estimated blood loss was not significantly different (Frail: 579±528 mL, Non-Frail: 340±141 mL, p=.126). Incidence of durotomy were similarly low with only one occurrence in the non-frail group. Total number of postoperative complications were significantly different (Frail: 1.1±0.9, Non-Frail: 0.6±0.8, p=.020).

ERAS components

All components of ERAS were similar between groups. The percentages of drains placed were 59% in frail patients versus 35% in nonfrail patients (p=.209). The average day of drain removal (for those patients with drains placed) was not significantly different between groups (3.5±1.7 vs. 3.6±0.9, p=.932). The proportions of subjects with indwelling urinary catheters placed were 18% in frail patients and 19% in non-frail patients (p=1.000). No incidences of PCA use occurred in either group.

Physiological function

The first day of assisted-walking occurred similarly between frail and non-frail patients (0.7 days vs. 0.7 days, p=.764). First bowel movement occurred similarly (2.3 days vs. 2.3 days, p=.852). First day of bladder voiding occurred similarly (0.3 days vs. 0.4 days, p=.983). Overall, the sum total days for recovery of physiological function were also similar (Frail: 3.4 days, Non-Frail: 3.5 days, p=.976) (Fig. 1).

Pain and opioid consumption

No significant difference was noted between frail and nonfrail patients for post-operative pain scores or opioid consumption (Table 3).

Length of Stay and Discharge

Length of stay was similar between groups (Frail: 3.8±1.9, Non-Frail: 3.3±1.3, p=.375; median 3 days for both). Non-home discharge rates trended toward increase in frail patients but was not significant (Frail: 23.5%, Non-Frail: 7.7%, p=.193). Discharge disposition matched physical therapy and occupational therapy recommendations for all patients. One patient underwent reoperation within 30 days requiring readmission in the frail cohort (5.9%), for instrumentation failure/wound dehiscence. There were no 30-day reoperations or readmissions in the nonfrail cohort.

Discussion

ERAS is an evidence-based multimodal perioperative care approach to reduce length of hospitalization, surgery-related complications, and improve early rehabilitation [9]. There have been no studies to our knowledge that have investigated the benefit of ERAS in specifically frail patients undergoing spine surgery. The results of this study indicate that a standardized ERAS protocol enhances return of physiological function and length of stay in frail patients undergoing open TLIF; and the improved outcomes in frail patients with ERAS approximate those in non-frail patients.

The existing literature for ERAS in frail patients has primarily been in colorectal surgery. In colorectal surgery patients who underwent an ERAS protocol, frail patients were found to have a higher incidence of ED visits, readmissions, and higher LOS compared to their non-frail counterparts [46]. Additionally, a dedicated work area for focused care of frail patients, with similar concepts to that of ERAS, was shown to reduce complications, length of stay, and readmissions [47]. Conversely in a randomized controlled trial of frail patients, postoperative care tailored for frailty in the form of physical therapy, nutritional intervention, prevention of delirium, or optimization of polypharmacy was shown to not affect the frequency of severe complications [48].

In a previously published retrospective study of patients undergoing open 1- or 2-level TLIF, our ERAS protocol demonstrated a reduction in LOS, improvement in return of post-operative physiological function, reduction in immediate postoperative pain scores, and reduction in overall inpatient IV opioid consumption and 72-hour total opioid consumption. In the current study specifically assessing ERAS with frailty, the primary benefit of ERAS for frail patients was limited to enhanced return of physiological function and length of stay. Frail patients were found to have earlier ambulation (1 day sooner), bowel movement (1 day sooner), and spontaneous bladder voiding (2 days sooner) after implementation of ERAS. Additionally, LOS was reduced by 1 day. There was no improvement in post-operative pain scores or opioid MME consumption with ERAS for frailty, which may indicate particularly resistant challenges to post-operative pain management with frailty. Alternatively, this study may have been underpowered to detect a difference for this specific outcome. Further investigation is needed to identify better pain management strategies in frail patients, as this population is potentially more susceptible to opioid associated adverse events.

As previously described in the literature, frailty itself was found to be associated with composite post-operative complication rates. This study is underpowered to detect a difference in other factors such as reoperation rate, readmission rate, or mortality [22-24].

It should be noted that in frail patients’ overall pain scores trended towards increasing with ERAS with no significant change noted in total daily opioid consumption and similar pain scores observed at the day of discharge. While the difference in overall pain scores was not found to be significant; this may be due to under-powering or sampling bias. When parsing the data into post-operative days, mean pain scores on POD 1 were found to be significantly higher post-ERAS (3.6 vs. 5.0, p=.048) with maximum pain scores on POD 1 trending towards significance (6.3 vs. 7.8, p=.092). Coinciding with this change, post-ERAS total PO opioid consumption on POD 1 was found to be increased by 25 MME whereas IV opioid use on POD 1 was found to be decreased by 16 MME. This is expected given the reduction in PCA usage; however, this may be revealing that frail patients respond better to IV medications in the immediate postop period. Further investigation is warranted on this topic as IV usage in frail patients conversely may increase risk for other adverse events (eg, delirium, urinary retention, impaired mobilization). It is possible that worse POD 1 pain scores in frail patients may be an unintended consequence of this ERAS protocol that ultimately benefits recovery of physiologic function and length of stay. Regardless, overall pain scores were found to be similar at discharge.

Limitations

This study has several limitations including its retrospective, single surgeon study design, which may impact generalizability. The relatively small frailty cohort sizes may have limited our secondary outcome analyses, however, likely reflects a patient-surgeon shared decision model that often defers elective spine surgery in patients with severe frailty. The cutoff for frailty included those with two or more components whereas the original cutoff by Fried et al was three or more [31]. Since frailty is known to be on a continuum, the cutoff for the sakes of this work is supported [44]. Future work will be aimed at expanding the study population to include multiple surgeons to potentially increase the statistical power. Additionally, the two cohorts were selected from two consecutive time periods, which may introduce confounding factors such as surgeon experience and other unaccounted hospital quality and patient safety improvement protocols. The equivalent operative technique, instrumentation, intraoperative adjuncts, and similar operative time between cohorts suggest against surgeon experience having a significant impact. Last, long term clinical and radiographic patient outcomes after TLIF were not included, as this is a study of primarily ERAS exposure with frailty as the variable (and patients undergoing TLIF representing the study population). Further, high quality prospective, multicenter studies may better elucidate the role of ERAS in a broader population of spine surgical patients, as this study was limited to only patients undergoing open TLIF.

Conclusion

An ERAS protocol improves return of physiological function and decreases length of post-operative hospitalization without increasing complications, readmission, or reoperation in frail patients undergoing open TLIF. Secondly, improved outcomes in frail patients with ERAS approximated those seen in non-frail counterparts, suggesting that ERAS has the potential to close the gap for this vulnerable population.

Abbreviations:

BMI

Body Mass Index

DVPRS

Defense & Veterans Pain Rating Scale

ERAS

Enhancing Recovery after Surgery

IV

Intravenous

LOS

Length of Stay

MME

Morphine Milligram Equivalent

PCA

Patient-controlled Analgesia

POD

Post-operative Day

PONV

Post-operative Nausea and Vomiting

SD

Standard Deviation

TLIF

Transforaminal Lumbar Interbody Fusion