High Modified 5 Factor Frailty Index is Associated With Worse PROMs and Decreased Return to Activities After 1 or 2 Level MI-TLIF

Abstract

Study Design

Retrospective cohort study.

Objective

Frailty is defined as a state of minimal “physiologic reserve.” The modified 5 factor frailty index (mFI-5) is a recently proposed metric for assessing frailty and has been previously studied as a predictor of morbidity and mortality. However, its utility in predicting postoperative patient reported outcomes (PROMs) and return to activities remains understudied.

Methods

This is a retrospective cohort study of patients undergoing 1 or 2 level minimally invasive transforaminal lumbar interbody fusion queried from a prospectively maintained multi-surgeon registry. Comorbidities including hypertension, congestive heart failure, diabetes, chronic obstructive pulmonary disease, and functional status were utilized to calculate the mFI-5. Patients were separated into 3 cohorts, “non-frail,” “moderately frail” or “severely frail” based on their mFI-5. Patient outcomes were compared between the cohorts.

Results

392 patients were included of which 227 patients were non-frail, 134 were moderately frail, and 31 were severely frail. Frail patients had longer operative time (P = 0.002), greater estimated blood loss (P = 0.038). Despite similar preoperative PROMs, at a longer term follow up, frail patients also had worse PROM scores for ODI (P = 0.009), VAS-back (P = 0.028), and VAS-leg (P = 0.004). Frail patients had worse preoperative (P = 0.017) and postoperative (P < 0.001) SF-12 PCS. At 1-year, frail patients had lower rates of fusion (P = 0.002). Frail patients also demonstrated lower rates of returning to work (P = 0.018), returning to driving (P = 0.027), and discontinuation of narcotics (P = 0.004).

Conclusion

Frail patients as measured by the mFI-5 index demonstrated worse postoperative outcomes following 1 or 2 level MI-TLIF. Careful patient selection and preoperative optimization may be especially important in frail patients.

Keywords

frailty MI-TLIF PROMs outcomes return to activities mFI-5

Introduction

As the demand for surgical treatment of degenerative lumbar spinal disease continues to grow,¹ it becomes increasingly important to understand which patients will benefit the most and which patients may experience suboptimal outcomes.

Frailty is one such patient specific factor that has the potential to influence how patients improve after spine surgery. Frailty represents a state of minimal “physiologic reserve” or the decreased ability of a patient to rebound after significant stressors.² Specifically, since first described in 2018, the 5-item Modified Frailty Index (mFI-5) has been extensively studied within the orthopedic and spine surgery literature as a prognostic indicator for morbidity and mortality following surgery.^3-6 These studies were largely conducted utilizing large national databases such as the American College of Surgeons National Surgical Quality Improvement Program database. While these databases can be highly valuable for discerning overarching trends in morbidity and mortality, especially when these complication rates are low, they do not capture long-term outcomes.

Very few studies have analyzed the impact of frailty on patient reported outcomes measures (PROMs) following minimally invasive transforaminal lumbar interbody fusion (MI-TLIF). Furthermore, prior studies have all utilized the mFI-11 index, while the mFI-5 may be more applicable clinically and for research purposes.^2,3 In this context, the relationship between mFI-5 and long-term PROMs and recovery kinetics deserves investigation. The purpose of the present study is to provide more granular data on longer term patient outcomes in frail patients defined with the mFI-5 index following one- or two-level MI-TLIF for the treatment of degenerative lumbar disorders. Specifically, we aim to 1) study the relationship between PROMs and mFI-5 and 2) evaluate whether mFI-5 influences recovery kinetics.

Methods

Study Design and Patient Cohort

This was a retrospective cohort study conducted following approval by the Institutional Review Board. Patients who underwent primary one or two-level MI-TLIF for the treatment of degenerative conditions of the lumbar spine between April 2017 and August 2023 with a minimum follow up of 6 months were included. The data was derived from a single-center multi-surgeon prospectively maintained database. MI-TLIF was performed as previously described in the literature.^7,8 Patients less than 18 years old, revision surgeries, or those with a surgical indication of trauma were excluded.

Data Collection

Data was collected and managed using REDCap (Research Electronic Data Capture)^9,10 hosted at Weill Cornell Medicine Clinical and Translational Science Center supported by the National Center For Advancing Translational Science of the National Institute of Health under award number: UL1 TR002384. Following data was collected and analyzed:

1. Demographics: age, gender, race, insurance type, body mass index (BMI), age adjusted Charlson Comorbidity Index (CCI), American Society of Anesthesiologists (ASA) class were all recorded. Past medical history collected additionally included hypertension (HTN) requiring medication, diabetes mellitus (DM), chronic obstructive pulmonary disease (COPD), congestive heart failure (CHF), renal disease, tumor (both benign or malignant), autoimmune disease, and functional status (independent, partially dependent, or fully dependent.

2. PROMs: Oswestry Disability Index (ODI) Visual Analog Scale, (VAS) back and leg, 12-Item Short Form Survey Physical Component Score (SF-12 PCS), and global rating change (GRC) were collected at 2 weeks, 6 weeks, 12 weeks, 6 months, 1 year, and 2 years after the surgery. GRC is a single question “Compared to preoperative, you feel 1) much better, 2) somewhat better, 3) about the same, 4) somewhat worse, or 5) much worse?).” Responses of “much better” and “somewhat better” were defined as “feeling better” while “much worse” and “somewhat worse” were defined as “feeling worse.” Two postoperative timepoints - early (<6 months) and late (>6 months) - were defined for analyses.^11,12

3. Surgical details and outcomes: Estimated blood loss (EBL), operative time, number of levels operated, and specific operated levels were recorded. Postoperative length of stay (LOS) was also collected. Evidence of fusion was assessed on the 1-year CT scan as part of our institution’s patient standard of care. Intraoperative and postoperative complications as well as all reoperations through 2 years postoperatively when available were collected.

4. Return to activities: patients completed a return to activities survey as part of their standard of care. Patients who were driving and working preoperatively were asked if they had begun driving and working again at every follow-up visit until they returned to that activity. Patients who took opioids postoperatively were asked the date of cessation if applicable.

5-item Modified Frailty Index (mFI-5)

The 5-item modified frailty index is a validated measure of frailty and has been shown to be a reliable predictor of complications and mortality in the spine literature and across various surgical subspecialties. The index accounts for 5 comorbidities in patient medical history - hypertension requiring medication, history of chronic obstructive pulmonary disease, history of congestive heart failure, diabetes mellitus, and functional health status. For the calculation of frailty index, patients received 1 point on the mFI-5 index for each comorbidity present preoperatively (HTN, DM, COPD, CHF) and 1 point if their functional status was either partially or fully dependent. For the purpose of this analysis, patients were then separated into 3 cohorts based on their mFI-5 index: 1) 0 comorbidities (non-frail), 2) 1 comorbidity (moderate frailty), or 3) > 2 comorbidities (severe frailty).

PROMs and Statistical Analysis

VAS-back, VAS-leg, ODI, and SF-12 PCS were assessed for improvement. Changes in PROMs from their preoperative baseline were assessed at both the early and late timepoints. Additionally, the percentage of patients achieving patient acceptable symptom state (PASS) at the late time point and minimal clinically important difference (MCID) at both the early and late time points were calculated. PASS achievement was defined as an absolute postoperative ODI <25.2 as previously defined by Shahi et al.¹³ MCID achievement was defined as a decrease in ODI by 12.8, VAS-back by 1.2, VAS-leg by 1.6, and SF-12 PCS by 4.1 as previously defined by Copay et al.¹⁴ Since preoperative baseline affects the utility of PASS and MCID in predicting patient improvement,^15-17 patients with preoperative ODI<25 were appropriately excluded from PASS achievement calculations. Patients with ODI<15, VAS-back/leg<2.5, and SF-12 PCS>92.1 were similarly excluded from MCID achievement calculations.

Statistical analyses were conducted to understand differences between the 3 defined groups. Normally distributed continuous variables between groups were compared with a one-way Analysis of Variance (ANOVA) or 2-tailed independent sample t test when appropriate. Non-normally distributed variables were compared between groups with a Kruskal-Wallis test or a Mann-Whitney U test as appropriate. Continuous variables were reported as mean ± standard deviation (SD) or as median [interquartile range (IQR)] when non-normally distributed. Categorical variables were compared between groups with a χ2 test or a fisher exact test when the number of subjects was less than 10. Categorical variables are reported as sample size (%). Statistical significance was taken at P value r <0.05. For comparisons between 3 groups that were significant, a post hoc analysis between each group was conducted to determine specific intergroup differences. Analyses were performed using the IBM Statistical Package for the Social Sciences (SPSS) version 25 (IBM Corp., Armonk, NY).

Results

A total of 392 patients met the inclusion criteria for the present study of which 227 (57.9%) patients were non-frail, 134 (34.2%) were moderately frail, and 31 (7.9%) were severely frail. On average, patients were operated on 3.02 years prior to data analysis. Frail patients (mFI-5 ≥ 1) were older (55.9 years vs 64.5 years vs 65.7 years, P < 0.001), had elevated BMI (27.3 kg/m² vs 29.9 kg/m²vs 31.5 kg/m², P < 0.001), had elevated CCI (1.8 vs 2.7 vs 3.9, P < 0.001), and had a greater percentage of patients with an ASA class ≥ 3 (5.3% vs 14.2% vs 38.7%) when compared with non-frail patients. There additionally were significant differences in the gender (P = 0.013), race (P < 0.001), and insurance type (P < 0.001) distributions (Table 1).

Table 1.

Patient Demographics. Superscripts a, b Indicates Post-hoc Significance Between Groups.

	Modified Frailty Index			p-value	Post-hoc p-value
	0	1	2+	p-value	Post-hoc p-value
n	227	134	31
Age (average ± SD)	55.86 ± 13.99a,b	64.52 ± 9.72a	65.68 ± 8.13b	<0.001	a<0.001, b<0.001
Male Gender (n (%))	105 (46.26%)a	81 (60.45%)a,b	12 (38.71%)b	0.013	a <0.001, b 0.028
Race (n (%))				<0.001	a 0.031, b <0.001, c 0.024
White	188 (82.82%)a,b	108 (80.6%)a,c	17 (54.84%)b,c
Black	8 (3.52%)	13 (9.7%)	7 (22.58%)
Asian	2 (0.88%)	3 (2.24%)	1 (3.23%)
Other/Unknown	29 (12.78%)	10 (7.46%)	6 (19.35%)
Insurance Type (n (%))				<0.001	a 0.005, b 0.001
Commercial/Private	167 (73.57%)a,b	77 (57.46%)a	12 (38.71%)b
Medicare	43 (18.94%)	50 (37.31%)	16 (51.61%)
Medicaid	3 (1.32%)	0 (0%)	1 (3.23%)
Workers' compensation	6 (2.64%)	4 (2.99%)	0 (0%)
Other/unknown	8 (3.52%)	3 (2.24%)	2 (6.45%)
BMI (average ± SD)	27.27 ± 5.32a,b	29.92 ± 7.57a	31.47 ± 4.09b	<0.001	a<0.001, b<0.001
CCI (average ± SD)	1.76 ± 1.49a,b	2.71 ± 1.62a,c	3.9 ± 1.66b,c	<0.001	a<0.001,b<0.001, c <0.001
ASA (n (%))				<0.001	a<0.001,b<0.001, c 0.008
I	30 (13.22%)a,b	2 (1.49%)a,c	0 (0%)b,c
II	185 (81.5%)	112 (83.58%)	19 (61.29%)
III	12 (5.29%)	19 (14.18%)	12 (38.71%)
IV	0 (0%)	1 (0.75%)	0 (0%)
Individual comorbidities
Renal disease	4 (1.8%)a,b	9 (6.7%)a	4 (12.9%)b	0.004	^a 0.019, ^b 0.009
Cancer (benign or malignant)	24 (10.6%)	22 (16.4%)	6 (19.4%)	0.166
Autoimmune disorders	9 (4.0%)	3 (2.2%)	1 (3.2%)	0.676

^a,bIndicates Post‐hoc significance between groups.

Bold values significant at p < 0.05.

The number of levels and specific levels operated on were similar between the 3 groups. (Table 2). Patients with severe frailty had significantly greater EBL (81.8 mL vs 85.7 mL vs 164.5 mL, P = 0.002) and longer operative time (145 minutes vs 149 minutes vs 195 minutes, P < 0.001) (Table 2).

Table 2.

Surgical Details. Superscripts a, b Indicates Post-hoc Significance Between Groups.

	Modified Frailty Index			P-value	Post-hoc P-value
	0	1	2+	P-value	Post-hoc P-value
n	227	134	31
Number of levels (n (%))				0.411
1	187 (82.38%)	113 (84.33%)	23 (74.19%)
2	40 (17.62%)	21 (15.67%)	8 (25.81%)
Surgical level (n (%))				0.081
L1-L2	1 (0.44%)	0 (0%)	0 (0%)
L2-L3	1 (0.44%)	0 (0%)	1 (3.23%)
L3-L4	18 (7.93%)	10 (7.46%)	2 (6.45%)
L4-L5	138 (60.79%)	103 (76.87%)	23 (74.19%)
L5-S1	109 (48.02%)	42 (31.34%)	13 (41.94%)
EBL (average mL ± SD)	81.79 ± 116.68a	85.71 ± 131.51b	164.5 ± 232.25a,b	0.038	a0.002,b0.013
Operative time (average minutes ± SD)	145.39 ± 69.28a	149.14 ± 82.44b	195.29 ± 83.06a,b	0.002	a<0.001, b0.005
Length of stay (average hours ± SD)	48.18 ± 26.35	50 ± 33.68	59.22 ± 29.38	0.137

^a,bIndicates Post-hoc significance between groups.

Bold values significant at p < 0.05.

Frail patients demonstrated worse postoperative outcomes. Frail patients had similar preoperative symptom state for all PROMs collected except SF-12 PCS which was lower among the severe frailty cohort (32.5 vs 32.2 vs 26.9, P = 0.017). Postoperatively, patients with severe frailty had delayed ODI improvement (P = 0.075 at the early time point and P = 0.001 at the late time point). Comparison of PROMs between groups at early and late time points showed significantly worse ODI (at both early and late time points), VAS-Back (at late time point), VAS-leg (at both early and late time points), and SF12-PCS (at late time point) (Table 3). Magnitude of improvement was largely the same between cohorts and only showed a significant difference for patients with moderate frailty in SF12-PCS (Table 4).

Table 3.

PROMs Comparison. Superscripts a, b, c Indicates Post-hoc Significance Between Groups.

		Modified Frailty Index			P-value	Post-hoc P-value
		0	1	2+	P-value	Post-hoc P-value
ODI	Preoperative	36.92 ± 19.21	39.85 ± 21.67	40.91 ± 21.36	0.24
	<6 mo	23.54 ± 17.43a	26.43 ± 19.11	34.17 ± 18.25a	0.021	a 0.006
		<0.001	<0.001	0.075
	≥6 mo	15.98 ± 16.57a,b	20.58 ± 17.51b	26.97 ± 23.35a	0.009	a0.008,b0.035
		<0.001	<0.001	0.001
VAS-back	Preoperative	5.65 ± 3.18	6.2 ± 3.23	6.52 ± 3.35	0.114
	<6 mo	2.39 ± 2.21	2.74 ± 2.38	3.46 ± 2.34	0.067
		<0.001	<0.001	<0.001
	≥6 mo	2.28 ± 2.51a,b	3 ± 2.68a	3.45 ± 3.05b	0.028	a0.029, ^b0.044
		<0.001	<0.001	<0.001
VAS-leg	Preoperative	5.16 ± 3.28	5.82 ± 3.53	6.2 ± 3.51	0.081
	<6 mo	1.83 ± 2.29a	2.01 ± 2.48b	3.35 ± 2.87a,b	0.017	a0.004, ^b0.023
		<0.001	<0.001	0.004
	≥6 mo	1.68 ± 2.44a,b	2.5 ± 2.97a	3.36 ± 3.14b	0.004	a0.012, ^b0.004
		<0.001	<0.001	0.003
SF12-PCS	Preoperative	32.52 ± 13.9a	32.16 ± 15.52b	26.93 ± 13.32a,b	0.017	a0.002, ^b0.019
	<6 mo	38.55 ± 10.21	35.87 ± 10.28	35.15 ± 10.38	0.055
		<0.001	0.002	0.004
	≥6 mo	44.99 ± 10.93a,b	38.94 ± 11.16a	37.44 ± 12.05b	<0.001	a<0.001, ^b0.007
		<0.001	<0.001	0.002

^a,b,cIndicates Post-hoc significance between groups.

Bold values significant at p < 0.05.

Table 4.

Comparison of Magnitude of Improvement. Superscript a Indicates Post-hoc Significance Between Groups.

		Modified Frailty Index			P-value	Post-hoc P-value
		0	1	2+	P-value	Post-hoc P-value
ΔODI	<6 mo	−13.39 ± 17.74	−14.25 ± 20.03	−7.47 ± 19.15	0.284
ΔODI	≥6 mo	−19.2 ± 16.12	−17.23 ± 20	−15.19 ± 18.19	0.526
ΔVAS-back	<6 mo	−3.28 ± 3.05	−3.52 ± 3.02	−3.2 ± 2.55	0.773
ΔVAS-back	≥6 mo	−3.18 ± 3.27	−3.16 ± 3.16	−3.21 ± 3.18	0.958
ΔVAS-leg	<6 mo	−3.36 ± 3.49	−3.36 ± 3.49	−2.75 ± 3.49	0.396
ΔVAS-leg	≥6 mo	−3.38 ± 3.61	−3.2 ± 3.45	−2.48 ± 3.3	0.584
ΔSF12-PCS	<6 mo	6.16 ± 10.08	3.62 ± 11.59	8.48 ± 10.79	0.074
ΔSF12-PCS	≥6 mo	11.54 ± 10.47a	6.06 ± 13.4a	10.74 ± 11.8	0.004	a <0.001

^aIndicates Post-hoc significance between groups.

Bold values significant at p < 0.05.

Severely frail patients had lower VAS-leg MCID achievement at the early time point (86.7% vs 81.4% vs 61.1%, P = 0.022). Frail patients also had lower MCID achievement for SF-12 PCS at the later time point (72.6% vs 48.1% vs 64.7%, P < 0.001). Other PROMs did not show significant differences in MCID achievement rates but similarly trended towards worse improvement in frail patients (Table 5).

Table 5.

MCID, PASS, GRC Achievement. Superscript a Indicates Post-hoc Significance Between Groups.

		Modified Frailty Index			P-value	Post-hoc P-value
		0	1	2+	P-value	Post-hoc P-value
MCID <6m	ODI	56.6%	57.1%	40.9%	0.352
	VAS-back	85.5%	82.1%	78.3%	0.585
	VAS-leg	86.7%a	81.4%	61.1%a	0.022	a 0.012
	SF12-PCS	53.1%	44.6%	61.1%	0.258
MCID ≥6m	ODI	71.2%	63.4%	63.2%	0.419
	VAS-back	79.3%	77.5%	55.0%	0.056
	VAS-leg	78.0%	74.6%	68.8%	0.665
	SF12-PCS	72.6%a	48.1%a	64.7%	<0.001	a <0.001
PASS ≥6m	ODI	68.9%	59.7%	50.0%	0.190
GRC	Worse	9.4%	9.2%	22%	0.564
	Same	5.6%	7.1%	0%
	Better	85.0%	83.7%	78%

^aIndicates Post-hoc significance between groups.

Bold values significant at p < 0.05.

Frail patients returned to driving (98% vs 93.3% vs 84.6%, P = 0.027), working (90.6% vs 75% vs 85.7%, P = 0.018), and discontinued narcotics (98.2% vs 89.4% vs 86.7%, P = 0.004) at significantly lesser rates. The time to return to activities was not significantly different between the groups (Table 6).

Table 6.

Return to Activities. Superscripts a, b Indicates Post-hoc Significance Between Groups.

		Modified Frailty Index			P-value	Post-hoc P-value
		0	1	2+	P-value	Post-hoc P-value
Driving (n driving preoperatively = 153 (0), 89 (1), 13 (2+)	Return to driving (%)	98.0%a	93.3%	84.6%a	0.027	a 0.049
	Days to return to driving	25 [14-35]	23 [15-33.5]	25 [22-33]	0.683
Working (n working preoperatively = 127 (0), 60 (1), 7 (2+)	Return to working (%)	90.6%a	75.0%a	85.7%	0.018	a 0.007
Working (n working preoperatively = 127 (0), 60 (1), 7 (2+)	Days to return to working	17 [7-35]	16 [8.25-30.25]	21 [19.5-22.5]	--
Narcotics (n took narcotics postoperatively = 166 (0), 94 (1), 15 (2)	Stopped narcotic use (%)	98.2%a,b	89.4%a	86.7%b	0.004	a 0.003, b 0.009
	Days to Stopping to narcotic use	10 [4-21.5]	11 [5.5-22]	10.5 [0.75-37]	0.890

^a,bIndicates Post-hoc significance between groups.

Bold values significant at p < 0.05.

Of the 392 total patients, 226 (57.7%) patients had a 1-year CT scan to assess for fusion. Frail patients had significantly lower rates of fusion (96.4% vs 80.6% vs 82.4%, P = 0.002) than non-frail patients (Table 7).

Table 7.

Surgical Outcomes and Complications. Superscripts a, b Indicates Post-hoc Significance Between Groups.

	Modified Frailty Index			P-value	Post-hoc P-value
	0	1	2+	P-value	Post-hoc P-value
Fusion (n analyzed = 137 (0), 72 (1), 17 (2+)	96.4%a,b	80.6%a	82.4%b	0.002	a <0.001, b 0.045
Complications
Intraoperative	2 (0.88%) - 1 Durotomy, 1 Screw malposition requiring removal	0 (0.0%)	1 (3.2%) - Durotomy	--
In hospital	27 (11.9%) - 6 POUR, 3 anemia requiring transfusion, 1 DVT, 1 cardiac event, 2 ileus, 11 PONV, 2 hypotension, 1 hyponatremia	8 (6.0%) - 3 POUR, 1 Cardiac event, 2 ileus, 2 PONV, 1 hypotension, 1 fever	4 (12.9%) - 2 POUR, 1 AKI, 1 Anemia requiring transfusion	0.163
Postoperative	12 (5.3%) - 3 wound drainage, 2 seroma, 2 wound dehiscence, 1 bowel/bladder dysfunction, 1 cage subsidence not requiring reoperation, 1 new onset sensory symptoms not requiring reoperation, 2 ASD not requiring reoperation	8 (6.0%) - 1 wound drainage, 1 bowel/bladder dysfunction, 1 wound dehiscence, 1 PE requiring readmission, 3 new onset sensory symptoms not requiring reoperation, 1 cage subsidence not requiring reoperation	3 (9.7%) - 1 PE requiring readmission, 1 cage subsidence not requiring reoperation, 1 ASD not requiring reoperation	0.620
Reoperation	8 (3.5%) - 4 revision fusion for painful hardware, 4 revision fusion for new onset sensory symptoms	7 (5.2%) - 2 revision fusion for cage displacement, 3 revision fusion for recurrent stenosis, 2 revision fusion for ASD	2 (6.5%) - 2 revision fusion for painful hardware	0.622

^a,bIndicates Post-hoc significance between groups.

Bold values significant at p < 0.05.

While the severe frailty cohort experienced higher rates of intraoperative, in-hospital medical, or postoperative complications, these differences were not statistically significant. Reoperation rate between the 3 cohorts was similar as well (Table 7).

Discussion

As the aging population continues to expand, a corresponding rise in lumbar spondylosis and frailty is anticipated. Recent evidence suggests that the assessment of perioperative risks and outcomes following spine surgery may be better informed by biological age rather than chronological age.¹⁸ Biological age encompasses the accumulation of physiological challenges over time, alongside an individual’s physiological reserve, which reflects their capacity to withstand stressors such as surgery.¹⁹ Frailty indices have been established to enhance the predictive accuracy in identifying patients at risk of experiencing unfavorable perioperative and postoperative outcomes. Specifically, the mFI-5 is a newer index, simplified from the previous mFI-11 index focusing on 5 key comorbidities—functional dependence, diabetes, hypertension, COPD, and congestive heart failure—chosen for their strong predictive association with frailty. We found that frail patients as defined by the mFI-5 had longer operative time, greater EBL, worse PROMs postoperatively, lower rates of return to activities of daily living, lower rate of narcotic discontinuation, and lower fusion rates compared to non-frail patients. Taken together, frailty status likely impacts the patient’s “overall health state” and may make it more difficult to recover from spine surgery.

Frail patients had significantly greater operative time and EBL with similar LOS. While it is not what we found in this cohort, longer LOS has been a well described outcome of frail patients following lumbar fusion.^20,21 Moses et al. previously found a 3 day longer LOS among severely frail patients compared with non-frail patients after single level MI-TLIF.²⁰ Comparing these may be difficult however as the baseline patients with no frailty had a LOS of 3.1 days, a day longer than non-frail patients in our cohort. It is possible differences in institutional postoperative protocols clouds the LOS finding. Nonetheless, frail patients likely do have longer LOS and it is important to consider this within the context of postoperative recovery. Contrastingly, the literature regarding operative time and EBL in frail patients is less clear. Previous studies that have reported operative time and EBL in frail patients following TLIF have not found significant differences.^20,21 In our cohort, severely frail patients experienced almost a 100% and 35% increase in operative time and EBL respectively compared to non and moderately frail patients. It is possible patients with significant frailty had more degeneration and therefore complex anatomy, predisposing to challenging surgeries and difficult recovery. Importantly however, frail patients did not experience greater rates of intraoperative complications. While these patients may undergo surgery safely, surgeons should be wary of the potential risks associated with prolonged anesthesia time and increased blood loss in frail patients.

Although the magnitude of improvement was similar between the cohorts, our findings demonstrated that frail patients had significantly worse PROMs at both the early and late time points despite similar preoperative symptom states. To our knowledge, 2 previous studies have reported PROMs in frail patients after TLIF, both utilizing the mFI-11 index.^20,21 Postoperatively, in an analysis of 488 one or 2 level TLIF, not necessarily conducted using minimally invasive techniques, Tran et al. found that frail patients are associated with worse one-year postoperative PROMs including ODI, SF-12 PCS, VAS-back, and VAS-leg.²² Frail patients demonstrated worse MCID achievement for SF-12 PCS and VAS-back. In their cohort the magnitude of recovery was significantly worse in frail patients as well for ODI, SF-12 PCS, and VAS-back. The second study, Moses et al. included a uniform cohort of 198 only single level MI-TLIF procedures and demonstrated frail patients had significantly worse one-year ODI, SF12-PCS, and SRS function score with no significant differences in delta change or MCID achievement.²⁰ We found similarly worse PROMs for ODI, VAS-back, VAS-leg, and SF12-PCS in frail patients and significantly worse MCID achievement in VAS-leg and SF12-PCS. For predicting postoperative PROMs, both mFI-5 and mFI-11 may serve as beneficial indices for providing adequate preoperative counseling concerning expectations for symptom improvement among frail patients.

Our study found that frail patients had lower rates of return to activities including driving and working compared to non-frail patients. While to the best of our knowledge, no prior analysis has reported recovery kinetics in frail patients, these findings are in line with the current literature on predictors of delayed or non-return to activities after lumbar spine surgery. Shinn et al. and Singh et al. have both identified that patients with worse physical function and higher comorbidity burden were associated with slower return to activities and non-returning to work.^23-25 Similarly, prior studies have identified frailty as an independent risk factor for predicting driving cessation^26,27 and have been shown to predict early retirement from both manual and non-manual labor^28,29 among elderly individuals. To our knowledge, no prior studies have investigated the rate of return to driving and working rates after spine surgery specifically among frail patients. These findings further confirm that frailty represents a state of minimal “physiologic reserve” or the decreased ability for a patient to rebound after significant stressors.² Hence, particularly among frail patients who prioritize returning to work and/or driving postoperatively, it may be important to offer thorough counseling regarding the higher risk of not achieving these goals. The rate of narcotic discontinuation was also lower among frail patients within our cohort. The current data regarding frailty and postoperative narcotic use is limited and mixed. Among patients undergoing elective non-spine-related surgeries, Auckley et al. found that frailty is associated with increased opioid use whereas no associations were found by Admiraal et al.^30,31 Shinn et al. demonstrated that elevated pain scores were independent predictor non-cessation of narcotics after elective lumbar surgery.²³ While frailty may not be a direct predictor for increased postoperative narcotic use, frail patients had higher early and late postoperative pain scores, which likely explains the correlation with the higher rate of narcotic use among these patients.

Successful fusion rates were also lower among frail individuals within our study cohort. Frail patients had evidence of fusion on 1 year CT at ∼80% compared with >95% in non-frail patients. This is consistent with a prior multicenter study of adult spinal deformity (ASD) patients by Miller et al. who utilized the adult spinal deformity frailty index (ASD-FI) and found a 13-fold higher likelihood of pseudarthrosis among frail patients undergoing ASD surgery.³² The mFI-5 evaluates the presence of one or more baseline comorbidities including HTN, DM, COPD, and CHF, as well as the presence of partially or fully dependent functional status. Narain et al. conducted an analysis on predictors of surgical complications after single level MI-TLIF. In their cohort DM was found as an independent predictor of surgical complications of which pseudarthrosis represented 64% of.³³ In a cohort of MI-TLIF with rhBMP-2 or mesenchymal stem cells, Overley et al. found hypertension was predictive of radiographic nonunion.³⁴ Although no prior studies have identified a robust association between each of these individual variables with pseudarthrosis after MI-TLIF, our studies demonstrate that the evaluation of frailty that comprehensively accounts for the aggregate of all 5 variables is potentially a reliable and clinically useful predictor.

Postoperative complications and reoperation demonstrated a non-significant “dose response” trend towards greater rates in frail patients. It is possible we were not powered to demonstrate significant differences. Much of the literature does support the notion that frail patients have increased rates of medical and surgical complications after spine fusion.^20,21,35 A systematic review of the literature by Chan et al. including 32 studies with a total of 127,813 patients concluded that regardless of how frailty was defined it was associated with increased serious adverse events, mortality, increased LOS, readmission, and non-home discharge.³⁶ Interestingly Burke et al. reported that MIS TLIF is protective against revision surgery and non-home discharge compared to open surgery in frail patients.³⁷ The exact relationship between MIS and open surgery for these patients remains unclear and the present study is not designed to answer that question.

Several limitations constrain the scope of this study. First, given the retrospective design of the study, the accuracy of the data hinges on the precision of charting and could also be prone to selection bias. Additionally, this study examines data solely from one institution, which may not accurately represent outcomes from other centers where surgical techniques or management approaches could differ. Moreover, the study encompasses a relatively small patient cohort of individuals with mFI-5 of 2+, possibly resulting in insufficient statistical power to identify certain associations such as differences magnitude of PROMs improvement, in LOS, as well as peri- and postoperative complications. Similarly, we are limited by our follow up. It is possible frail patients take longer to improve but will eventually plateau at the same state as non-frail patients. This question remains unanswered by our analysis and may be a subject of future research. Further studies investigating the external validity of this study may be important prior to the application in practice. We additionally did not analyze potential cofounder variables such as spinopelvic parameters, degree of degeneration, muscle health, and mental fitness that can certainly impact a patient’s ability to recover postoperatively, some of which are included in other frailty indices such as the Edmonton Frailty score. Despite these limitations, to our knowledge, this is the largest comprehensive study to date that investigates the association of frailty defined by mFI-5 with outcomes after one- or two-level MI-TLIF.

Conclusion

Increased mFI-5 is associated with inferior postoperative outcomes after one- or two-level MI-TLIF surgery including worse postoperative PROMs, lower rates of return to activities, lower rate of narcotic discontinuation, and higher pseudarthrosis rates. Whether these differences are modulated by overall poor health state or a decreased ability to recover requires further investigation. Additionally, more research is necessary to fully understand the correct patient selection and how best to optimize highly frail patients before surgical intervention.

Footnotes

Declaration of Conflicting Interests

The author(s) declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: HS2, LLC Innovasis Globus Medical, Inc. Intrinsic Therapeutics Inc. Bioventus HSS AND ASC Development Network.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Pratyush Shahi

Maximilian Korsun

Joshua Zhang

Eric Kim

Austin C. Kaidi

Kasra Araghi

Avani Vaishnav

Francis C. Lovecchio

Sheeraz A. Qureshi

Sravisht Iyer

References

Araghi

Subramanian

Haque

, et al. Provider referral patterns and surgical utilization among new patients seen in spine clinic. Spine. 2023;48(12):885-891. doi:10.1097/BRS.0000000000004656

Kweh

BTS

Lee

Tan

Liew

Hunn

Wee Tee

. Posterior instrumented spinal surgery outcomes in the elderly: a comparison of the 5-item and 11-item modified frailty indices. Global Spine J. 2024;14(2):593-602. doi:10.1177/21925682221117139

Subramaniam

Aalberg

Soriano

Divino

. New 5-factor modified frailty index using American College of surgeons NSQIP data. J Am Coll Surg. 2018;226(2):173-181.e8. doi:10.1016/j.jamcollsurg.2017.11.005

Weaver

Malik

Jain

Kim

Khan

. The modified 5-item frailty index: a concise and useful tool for assessing the impact of frailty on postoperative morbidity following elective posterior lumbar fusions. World Neurosurg. 2019;124:e626-e632. doi:10.1016/j.wneu.2018.12.168

Tracy

Wilson

Smith

Schenker

Gelbard

. The 5-item modified frailty index predicts adverse outcomes in trauma. J Surg Res. 2020;253:167-172. doi:10.1016/j.jss.2020.03.052

Pierce

Naessig

Kummer

, et al. The five-item modified frailty index is predictive of 30-day postoperative complications in patients undergoing spine surgery. Spine. 2021;46(14):939-943. doi:10.1097/BRS.0000000000003936

Sarmiento

Shahi

Melissaridou

Fourman

Araghi

Qureshi

. Step-by-step guide to robotic-guided minimally invasive transforaminal lumbar interbody fusion (MI-TLIF). Ann Transl Med. 2023;11(5):221. doi:10.21037/atm-22-3273

Shahi

Subramanian

Maayan

, et al. Surgeon experience influences robotics learning curve for minimally invasive lumbar fusion: a cumulative sum analysis. Spine. doi:10.1097/BRS.0000000000004745

Harris

Taylor

Thielke

Payne

Gonzalez

Conde

. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377-381. doi:10.1016/j.jbi.2008.08.010

10.

Harris

Taylor

Minor

, et al. The REDCap consortium: building an international community of software platform partners. J Biomed Inform. 2019;95:103208. doi:10.1016/j.jbi.2019.103208

11.

Subramanian

Merrill

Shahi

, et al. Predictors of subsidence and its clinical impact following expandable cage insertion in minimally invasive transforaminal interbody fusion. Spine. doi:10.1097/BRS.0000000000004619

12.

Shahi

Subramanian

Araghi

, et al. Comparison of robotics and navigation for clinical outcomes after minimally invasive lumbar fusion. Spine. 2023;48(19):1342-1347. doi:10.1097/BRS.0000000000004721

13.

Shahi

Shinn

Singh

, et al. ODI <25 denotes patient acceptable symptom state after minimally invasive lumbar spine surgery. Spine. 2023;48(3):196-202. doi:10.1097/BRS.0000000000004479

14.

Copay

Glassman

Subach

Berven

Schuler

Carreon

. Minimum clinically important difference in lumbar spine surgery patients: a choice of methods using the Oswestry Disability Index, Medical Outcomes Study questionnaire Short Form 36, and pain scales. Spine J. 2008;8(6):968-974. doi:10.1016/j.spinee.2007.11.006

15.

Shahi

Maayan

Subramanian

, et al. Preoperative disability influences effectiveness of minimal clinically important difference and patient acceptable symptom state in predicting patient improvement following cervical spine surgery. Global Spine J. 2023;20:21925682231215765. doi:10.1177/21925682231215765. Published online November.

16.

Shahi

Subramanian

Maayan

, et al. Preoperative disability influences effectiveness of MCID and PASS in predicting patient improvement following lumbar spine surgery. Clin Spine Surg. 2023;36(10):E506-E511. doi:10.1097/BSD.0000000000001517

17.

Singh

Shahi

Asada

, et al. Poor muscle health and low preoperative ODI are independent predictors for slower achievement of MCID after minimally invasive decompression. Spine J. 2023;23(8):1152-1160. doi:10.1016/j.spinee.2023.04.004

18.

Safaee

Dwaraka

Lee

, et al. Epigenetic age biomarkers and risk assessment in adult spinal deformity: a novel association of biological age with frailty and disability. J Neurosurg Spine. 2024;40(3):312-323. doi:10.3171/2023.10.SPINE23435

19.

Phan

Kim

Lee

, et al. Frailty is associated with morbidity in adults undergoing elective anterior lumbar interbody fusion (ALIF) surgery. Spine J. 2017;17(4):538-544. doi:10.1016/j.spinee.2016.10.023

20.

Moses

Fontes

RBV

Deutsch

O’Toole

Fessler

. The modified frailty index and patient outcomes following transforaminal lumbar interbody fusion surgery for single-level degenerative spine disease. J Neurosurg Spine. 2021;4:1-7. doi:10.3171/2020.11.SPINE201263. Published online June.

21.

Patel

Coban

Changoor

Sinha

Hwang

Emami

. The 5-factor modified frailty index is associated with increased risk of reoperations and adjacent level disease following single-level transforaminal lumbar interbody fusion. Global Spine J. Published online August. 2023;19:21925682231196828. doi:10.1177/21925682231196828

22.

Tran

Lambrechts

Issa

, et al. Modified frailty index does not provide additional value in predicting outcomes in patients undergoing elective transforaminal lumbar interbody fusion. World Neurosurg. 2023;170:e283-e291. doi:10.1016/j.wneu.2022.11.011

23.

Shinn

Mok

Vaishnav

, et al. Recovery kinetics after commonly performed minimally invasive spine surgery procedures. Spine. 2022;47(21):1489-1496. doi:10.1097/BRS.0000000000004399

24.

Subramanian

Shinn

Korsun

, et al. Recovery kinetics after cervical spine surgery. Spine. 2023;48(24):1709-1716. doi:10.1097/BRS.0000000000004830

25.

Singh

Ailon

McIntosh

, et al. Time to return to work after elective lumbar spine surgery. J Neurosurg Spine. 2021;1:1-9. doi:10.3171/2021.2.SPINE202051. Published online.

26.

Crowe

Kannoth

Andrews

, et al. Associations of frailty status with low-mileage driving and driving cessation in a cohort of older drivers. Geriatrics. 2020;5(1):19. doi:10.3390/geriatrics5010019

27.

Ishii

Doi

Tsutsumimoto

Nakakubo

Kurita

Shimada

. Driving cessation and physical frailty in community-dwelling older adults: a longitudinal study. Geriatr Gerontol Int. 2021;21(11):1047-1052. doi:10.1111/ggi.14272

28.

Haapanen

von Bonsdorff

Perttilä

, et al. Retirement age and type as predictors of frailty: a retrospective cohort study of older businessmen. BMJ Open. 2020;10(12):e037722. doi:10.1136/bmjopen-2020-037722

29.

Norheim

Bøggild

Andersen

Omland

Bønløkke

Madeleine

. Retirement status and frailty: a cross-sectional study of the phenotype of manual workers aged 50-70 years. Eur J Public Health. 2021;31(1):116-121. doi:10.1093/eurpub/ckaa188

30.

Auckley

Bentov

Zelber-Sagi

Jeong

Reed

Bentov

. Frailty status as a potential factor in increased postoperative opioid use in older adults. BMC Geriatr. 2021;21(1):189. doi:10.1186/s12877-021-02101-4

31.

Admiraal

van Zuylen

Hermanns

, et al. The effect of preoperative disability, cognitive impairment, frailty and opioid use on acute postoperative pain in older patients undergoing surgery A prospective cohort study. J Pain. 2023;24(10):1886-1895. doi:10.1016/j.jpain.2023.05.013

32.

Miller

Neuman

Jain

, et al. An assessment of frailty as a tool for risk stratification in adult spinal deformity surgery. Neurosurg Focus. 2017;43(6):E3. doi:10.3171/2017.10.FOCUS17472

33.

Narain

Parrish

Jenkins

, et al. Risk factors for medical and surgical complications after single-level minimally invasive transforaminal lumbar interbody fusion. Int J Spine Surg. 2020;14(2):125-132. doi:10.14444/7018

34.

Overley

McAnany

Anwar

, et al. Predictive factors and rates of fusion in minimally invasive transforaminal lumbar interbody fusion utilizing rhBMP-2 or mesenchymal stem cells. Int J Spine Surg. 2019;13(1):46-52. doi:10.14444/6007

35.

Leven

Lee

Kim

, et al. Frailty is predictive of adverse postoperative events in patients undergoing lumbar fusion. Global Spine J. 2017;7(6):529-535. doi:10.1177/2192568217700099

36.

Chan

Wilson

JRF

Ravinsky

, et al. Frailty adversely affects outcomes of patients undergoing spine surgery: a systematic review. Spine J. 2021;21(6):988-1000. doi:10.1016/j.spinee.2021.01.028

37.

Burke

Garcia

Safaee

, et al. Patient frailty influences outcome after open, but not minimally invasive, transforaminal interbody fusion: a case series of 115 patients with lumbar degenerative disease. Oper Neurosurg (Hagerstown). 2023;24(6):565-571. doi:10.1227/ons.0000000000000665