Discriminatory Value of the Risk Analysis Index Versus the 5-Factor Modified Frailty Index for Major Outcome Measures in Degenerative Cervical Myelopathy

Abstract

Study Design

Observational Retrospective Cohort Study.

Objectives

To compare the discriminatory abilities of the Revised Risk Analysis Index (RAI-Rev) and the 5-Factor Modified Frailty Index (mFI-5) to predict major postoperative outcomes in DCM patients overall, and by anterior (ADF) or posterior (PDF) approaches for decompression and fusion.

Methods

The American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database was queried for DCM patients undergoing ADF or PDF from 2015-2020. Logistic regression and Receiver Operating Characteristic analysis were used to compare the predictive and discriminatory value of mFI-5 and RAI-Rev for 30-day outcomes.

Results

18,138 DCM patients were included (median age: 61 years). Both RAI-Rev and mFI-5 predicted 30-day outcomes; however, increasing frailty as measured by RAI-Rev demonstrated greater odds ratios. RAI-rev demonstrated superior discrimination for non-fatal adverse outcomes, notably non-home discharge (NHD) [C-statistic 0.72 (95% CI 0.71-0.73, P < 0.001)], and for all non-fatal measures except major complications (P < 0.05). Sub-analysis by approach showed frailty had greater predictive accuracy for adverse outcomes in ADF compared to PDF.

Conclusions

The RAI-Rev demonstrated superior discrimination predicting non-fatal outcomes following DCM when compared to the mFI-5, with equivalent mortality prediction. Further, frailty plays a stronger role in predicting mortality and morbidity in ADF compared to frailty’s impact on predicting outcomes with PDF. This finding demonstrates the utility of the RAI-Rev in preoperative risk stratification with an increasingly frail patient population and provides initial evidence for selecting posterior approaches for DCM patients as frailty increases.

Keywords

degenerative cervical myelopathy frailty risk analysis index

Introduction

Degenerative cervical myelopathy (DCM) is a degeneration of the cervical spine with subsequent symptomatic compression of the spinal cord, which has increasingly affected aging populations.¹ As the elderly population is expected to almost double from 58 million in 2022 to 82 million in 2050, the incidence of DCM will continue to rise.² Surgical decompression of the spinal cord is the mainstay of treatment in moderate-to-severe disease, with data to suggest benefits in earlier disease as well.^3,4 Older patients frequently have more advanced disease, require higher risk surgical approaches, and have lower functional scores postoperatively; although, the functional improvement with surgery is comparable to younger patients.⁵

Frailty, or decreased baseline physiologic reserve, is a multifaceted syndrome that manifests across multiple organ systems and reflects a patient’s vulnerability to internal and external stressors, resulting in adverse postoperative outcomes.^6,7 Previously, the modified frailty index-11 (mFI-11) or its abbreviated form, the modified frailty index-5 (mFI-5), were used to measure frailty.^8-11 However, the mFI has been critiqued in recent years for more so resembling a comorbidity index rather than a measure of true frailty, especially when compared to the Risk Analysis Index (RAI).^12-14 The RAI was developed by Hall et al to more effectively capture the factors influencing frailty.^15,16 There is now a growing body of evidence, especially regarding spinal pathologies, to support superior outcome discrimination when using the Revised Risk Analysis Index (RAI-Rev).^17,18

As of this time, there is no current literature assessing discrimination of the RAI-Rev for predicting adverse outcomes for anterior (ADF) compared to posterior (PDF) approaches for decompression and fusion in degenerative cervical myelopathy. This study evaluates the discriminative thresholds of the RAI-Rev and mFI-5 on 30-day major postoperative outcomes in DCM patients using data from a large national cohort drawn from American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) and analyzes subgroups by anterior vs posterior approach.¹⁹

Methods

Data Source

The present study investigated the ACS-NSQIP database, in keeping with the NSQIP Data Use Agreement. Established within the Veterans Health Administration in 1994, the NSQIP was designed to moderate and improve healthcare quality. Following its effective implementation within the Veterans Health Administration, it became available to the private sector in 1999. We utilized the NSQIP data from 2015 to 2020, with approximately six million patient records, comprising over 200 variables including patient demographics, comorbidities, and postoperative outcomes. To maintain data integrity, onsite surgical clinical reviewers are employed alongside regular internal audits to ensure inter-rater reliability.

Patient Population and Baseline Characteristics

Patients eligible for this study were those diagnosed with DCM who underwent cervical decompression and fusion surgery. This procedure included anterior approaches and/or posterior approaches (both instrumented and non-instrumented). Supplemental Table 1 summarizes the specific ICD-9/10-CM and CPT codes utilized in this study.

The study examined various demographic and baseline variables such as age, gender, height, weight, Body Mass Index (BMI), length of hospital stay, transfer status, operation time, and history of medical conditions including diabetes mellitus, Chronic Obstructive Pulmonary Disease (COPD), Congestive Heart Failure (CHF), dyspnea, hypertension, cancer, open wounds, steroid use, significant unintentional weight loss (>10%), bleeding disorders, preoperative transfusion, sepsis, systemic inflammatory response syndrome (SIRS), functional dependence, and smoking status.

Frailty Indices

Frailty was assessed using the RAI-Rev scoring system, adapted from previous studies by Hall et al and Arya et al^12,14 RAI incorporates sex, age, cancer diagnosis (excluding melanoma), weight loss (unintentional weight loss of 4.5 kg over 3 months), renal failure, CHF, poor appetite, shortness of breath at rest, residence status (independence of living), cognitive decline, and activities of daily living (Supplemental Table 2). However, preoperative cognitive decline, another variable originally included in RAI, was not included in the present study due to NSQIP data limitations. Scores were categorized into robust, normal, frail, and very frail tiers based on specific score ranges. Score stratification of the RAI-Rev is as follows: robust ≤10, normal 11-20, frail 21-30, and very frail ≥31 RAI-Rev scores.

The modified frailty index was initially developed with 11 variables (mFI-11).²⁰ However, it was later modified to include only five variables as a result of changes in reporting requirements.²¹ These variables include history of diabetes, hypertension, functional dependence, COPD, and CHF, with one point attributed to each condition. The mFI scores range from 0 to 5, and the tiers are as follows: nonfrail 0, prefrail 1, frail 2, and severely frail ≥3.

Outcome Measures

The primary outcomes included 30-day mortality, non-home discharge (NHD), extended length of stay (eLOS), minor complications, major complications, Clavien-Dindo Physical Status (PS) classification grade IV complications (CDIV), 30-day unplanned readmission, and 30-day unplanned reoperation. Minor complications included intra-/postoperative blood transfusion, renal insufficiency, or urinary tract infection (UTI). Major complications encompassed prolonged intubation exceeding 48 hours, unplanned reintubation, sepsis, septic shock, pneumonia, deep vein thrombosis (DVT)/thrombophlebitis, pulmonary embolism (PE), acute cerebrovascular accident or stroke with neurological deficit(s), acute renal failure, myocardial infarction (MI), cardiac arrest requiring cardiopulmonary resuscitation, superficial surgical site infection (SSI), deep incisional SSI, organ space SSI, or wound disruption. CDIV were characterized by life-threatening organ dysfunction requiring intensive care, namely sepsis or septic shock, acute renal failure, pulmonary embolism (PE), myocardial infarction (MI), cardiac arrest requiring cardiopulmonary resuscitation, ventilation >48 hours, and unplanned reintubation. eLOS was defined as hospital stays exceeding the 75th percentile of the study population. Discharge designations were categorized into home and a facility that is home, or NHD, namely rehabilitation facility, skilled nursing facility, hospice care, and patients leaving against medical advice.

Statistical Analysis

Continuous variables with non-normal distribution are reported as median with Interquartile Range (IQR). The distribution of the data was determined by employing the Kolmogorov-Smirnov test. Comparisons of medians of non-normally distributed continuous variables were carried out using Kruskal-Wallis test while the categorical variables were analyzed using Chi-square tests. Univariable and multivariable logistic regression analyses were performed for mFI-5, and RAI-Rev, and the following outcomes: 30-day mortality, NHD, eLOS, minor complications, major complications, CDIV, 30-day unplanned readmission, and 30-day unplanned reoperation. The effect sizes for dichotomous outcomes were presented as odds ratio (OR) and 95% Confidence Interval (CI). Receiver Operating Characteristic (ROC) curve analysis including the Area Under the ROC Curve (AUC)/C-statistics calculations were used to evaluate the discriminatory accuracy of the mFI-5 and RAI-Rev for primary outcomes. The DeLong test was used to compute the significance of the C-statistic difference between the frailty indices. Statistical analyses were carried out using Excel (Microsoft, Redmond, WA, USA), IBM SPSS Statistics ver. 27.0 (IBM Co., Armonk, NY, USA), and STATA 12.0 (StataCorp, College Station, TX, USA). P < 0.05 was considered statistically significant.

Results

Study Population Characteristics

A total of 18,138 patients with DCM who met the inclusion criteria were included in this study (Figure 1). The median age of the cohort was 61 years (IQR: 53 - 69 years) and 56.4% were male. The median BMI was 29.45 kg/m² (IQR: 25.75 - 34.01 kg/m²) and the most common comorbidities within the cohort were hypertension (56.7%), smoking (25.3%) and diabetes mellitus (21.8%). Stratification of the cohort by frailty group as measured by the mFI-5 was as follows: robust 36.9%, normal: 39.9%, frail 20.1%, and very frail 3.1%. Distribution of frailty tiers as measured by RAI-Rev was as follows: robust: 5.6%, normal: 50.7%, frail: 40.7%, and very frail: 3.0%.

Figure 1.

Selection process for study cohort. Abbreviations: ACS-NSQIP; American College of Surgeons National surgical quality improvement program.

Of the total population, the median operative time was 138 minutes (IQR: 100 - 190 minutes) and the median postoperative length of stay was 2 days (IQR: 1 - 3 days). Major postoperative complications occurred in 4.8% of patients with the most common being pneumonia (1.4%) and unplanned reintubation (0.9%). Additionally, CDIV complications were reported in 2.5% of the cohort and 17.7% of patients had a NHD destination. A total of 5.6% of patients were readmitted, 2.8% underwent reoperation, and the total mortality rate of the cohort was 0.4%. All patient demographics, clinical characteristics, and comparisons by RAI-Rev score are further described in Table 1.

Table 1.

Overview of baseline, preoperative, intraoperative and postoperative characteristics of the study cohort, stratified by frailty tier as defined by the RAI-Rev.

Variable	Total	RAI-Rev≤ 10	RAI-Rev = 11-20	RAI-Rev = 21-30	RAI-Rev ≥31	P-value
Demographics
Age (years)	61 (53-69)	39 (33-42)	55 (50-59)	70 (64-74)	76 (71-83)	<0.001
Sex, male: female	10,228 (56.4):7.910 (43.6)	181 (17.8):836 (82.2)	5062 (55.1):4132 (44.9)	5520 (74.8):1857 (25.2)	395 (71.8):155 (28.2)	<0.001
Body mass index (kg/m²)	29.45 (25.75-34.01)	29.69 (25.11-35.18)	30.02 (25.94-34.79)	28.94 (25.69-32.93)	27.62 (24.26-31.68)	<0.001
Functional status						<0.001
Partially dependent	779 (4.3)	4 (0.4)	52 (0.5)	430 (5.8)	303 (55.1)
Totally dependent	84 (0.5)	0 (0)	2 (0)	10 (0.1)	72 (13.1)
Medical comorbidities
Diabetes mellitus	3962 (21.8)	72 (7.1)	1667 (18.1)	2016 (27.3)	207 (37.6)	<0.001
COPD	1033 (5.7)	9 (0.9)	422 (4.6)	539 (7.3)	63 (11.5)	<0.001
CHF	88 (0.5)	0 (0)	4 (0)	43 (0.6)	41 (7.5)	<0.001
Current smoker	4585 (25.3)	357 (35.1)	2843 (30.9)	1,3.7 (17.7)	78 (14.2)	<0.001
Dyspnea at rest	76 (0.4)	0 (0)	19 (0.2)	43 (0.6)	14 (2.5)	<0.001
Hypertension	10,293 (56.7)	166 (16.3)	4467 (48.6)	5227 (70.9)	433 (78.7)	<0.001
Disseminated cancer	27 (4.9)	0 (0.0)	0 (0.0)	0 (0.0)	27 (4.9)	<0.001
Open wound	130 (0.7)	1 (0.1)	30 (0.3)	67 (0.9)	32 (5.8)	<0.001
Steroid use	785 (4.3)	32 (3.1)	359 (3.9)	352 (4.8)	42 (7.6)	<0.001
Weight loss	87 (0.5)	1 (0.1)	3 (0.0)	40 (0.5)	43 (7.8)	<0.001
Bleeding disorders	330 (1.8)	14 (1.4)	104 (1.1)	179 (2.4)	33 (6.0)	<0.001
Preop transfusion	23 (0.1)	1 (0.1)	3 (0.0)	14 (0.2)	5 (0.9)	<0.001
Residence other than independent living	127 (0.7)	1 (0.1)	13 (0.1)	57 (0.8)	56 (10.2)	<0.001
Procedure
Outpatient procedure	3396 (18.7)	293 (28.8)	1960 (21.3)	1106 (15.0)	37 (6.7)	<0.001
Length of stay after operation (day)	2 (1-3)	1 (1-2)	2 (1-3)	2 (1-4)	4 (2-7)	<0.001
Operative time (min)	138 (100-190)	122 (90-162)	137 (100-191)	142 (102-193)	139 (100-205)	<0.001
Outcomes and postoperative complications
Mortality	77 (0.4)	0 (0.0)	20 (0.2)	38 (0.5)	19 (3.5)	<0.001
Nonroutine discharge destination (rehab, SNF, and hospice, AMA)	3206 (17.7)	44 (4.3)	971 (10.6)	1836 (24.9)	355 (64.5)	<0.001
Minor postoperative complications	870 (4.8)	27 (2.7)	309 (3.4)	448 (6.1)	86 (15.6)
Postoperative transfusion	283 (1.6)	3 (0.3)	101 (1.1)	147 (2.0)	32 (5.8)	<0.001
Renal insufficiency	15 (0.1)	0 (0.0)	5 (0.1)	7 (0.1)	3 (0.5)	0.001
UTI	229 (1.3)	3 (0.3)	61 (0.7)	133 (1.8)	32 (5.8)	<0.001
Major postoperative complications	870 (4.8)	27 (2.7)	309 (3.4)	448 (6.1)	86 (15.6)	<0.001
Prolonged intubation (≥48 hr)	130 (0.7)	1 (0.1)	49 (0.5)	67 (0.9)	13 (2.4)	<0.001
Unplanned reintubation	171 (0.9)	1 (0.1)	55 (0.6)	93 (1.3)	22 (4.0)	<0.001
Sepsis	97 (0.5)	0 (0.0)	39 (0.4)	44 (0.6)	14 (2.5)	<0.001
Septic shock	39 (0.2)	0 (0.0)	5 (0.1)	23 (0.3)	11 (2.0)	<0.001
Pneumonia	257 (1.4)	5 (0.5)	71 (0.8)	143 (1.9)	38 (6.9)	<0.001
DVT/thrombophlebitis	106 (0.6)	4 (0.4)	36 (0.4)	56 (0.8)	10 (1.8)	<0.001
Pulmonary embolism	90 (0.5)	4 (0.4)	26 (0.3)	53 (0.7)	7 (1.3)	<0.001
CVA/stroke with neurological deficit	29 (0.2)	0 (0.0)	14 (0.2)	10 (0.1)	5 (0.9)	<0.001
Acute renal failure	26 (0.1)	0 (0.0)	8 (0.1)	12 (0.2)	6 (1.1)	<0.001
Myocardial infarction	61 (0.3)	0 (0.0)	14 (0.2)	31 (0.4)	16 (2.9)	<0.001
Cardiac arrest requiring CPR	49 (0.3)	0 (0.0)	14 (0.2)	28 (0.4)	7 (1.3)	<0.001
Superficial SSI	118 (0.7)	10 (1.0)	59 (0.6)	45 (0.6)	4 (0.7)	0.574
Deep incisional SSI	54 (0.3)	2 (0.2)	23 (0.3)	27 (0.4)	2 (0.4)	0.514
Organ space SSI	69 (0.4)	3 (0.3)	30 (0.3)	33 (0.4)	3 (0.5)	0.537
Wound disruption	59 (0.3)	1 (0.1)	16 (0.2)	37 (0.5)	5 (0.9)	<0.001
Clavien-Dindo grade IV	457 (2.5)	5 (0.5)	157 (1.7)	236 (3.2)	59 (10.7)	<0.001
Readmission	1010 (5.6)	27 (2.7)	371 (4.0)	534 (7.2)	78 (14.2)	<0.001
Reoperation	514 (2.8)	12 (1.2)	192 (2.1)	277 (3.8)	33 (6.0)	<0.001

Abbreviations: RAI-Rev; recalibrated version of the Risk Analysis Index, COPD; chronic obstructive pulmonary disease, CHF; congestive heart failure, SNF; skilled nursing facility, AMA; against medical advice, UTI; urinary tract infection, DVT; deep vein thrombosis, CVA; cerebral vascular accident, CPR; cardiopulmonary resuscitation, SSI; surgical site infection.

Univariate Analysis of mFI-5 and RAI-Rev for 30-Day Postoperative Outcomes

Increasing frailty had greater odds of all adverse outcomes for RAI-Rev except NHD. The odds of poor postoperative outcomes increased across frailty tiers for all measures using the mFI-5 (Table 2). Increasing frailty, measured by RAI-Rev, increased a patient’s risk of 30-day mortality (frail: OR 2.64, 95% CI, 1.53-4.54, P < 0.001 and severely frail: OR 18.23, 95% CI, 9.67-34.37, P < 0.001), however frail patients had greater risk of NHD compared to severely frail (frail: OR 7.33, 95% CI, 5.39-9.96, P < 0.001 and severely frail: OR 4.26, 95% CI, 28.34-57.07, P < 0.001). Frailty as measured by the mFI-5 had similar odds of mortality (severely frail: OR 19.28, 95% CI, 8.31-44.74; P < 0.001) and NHD (severely frail: OR 8.55, 95% CI, 7.11-10.29; P < 0.001) within 30-days of decompression when compared to the RAI-Rev.

Table 2.

Univariate analysis of the mFI-5 and RAI-Rev for 30-day postoperative outcomes in DCM patients.

Variable	Mortality	NHD	eLOS	Minor Complication	Major Complication	Clavien-Dindo Grade IV	Unplanned Readmission	Reoperation
mFI-5
Prefrail (mFI-5 = 1)	2.58 (1.20-5.53)^*	1.89 (1.71-2.09)	1.52 (1.40-1.65)	1.98 (1.55-2.52)	1.46 (1.22-1.75)	1.64 (1.26-2.14)	1.50 (1.28-1.76)	1.15 (0.93-1.43)^**
Frail (mFI-5 = 2)	5.96 (2.82-12.60)	3.27 (2.94-3.64)	2.25 (2.05-2.47)	3.06 (2.37-3.96)	2.86 (2.38-3.45)	3.76 (2.90-4.88)	2.19 (1.84-2.61)	1.64 (1.29-2.07)
Severely frail (mFI-5≥ 3)	19.28 (8.31-44.74)	8.55 (7.11-10.29)	4.67 (3.91-5.59)	7.20 (5.09-10.19)	4.50 (3.36-6.01)	6.72 (4.64-9.74)	3.64 (2.75-4.83)	2.50 (1.69-3.72)
RAI-Rev
Prefrail (RAI-Rev = 21-30)	^a	2.61 (1.92-3.56)	1.77 (1.44-2.17)	3.06 (1.35-6.93)*	1.28 (0.86-1.90)**	3.52 (1.44-8.59)^*	1.54 (1.04-2.29)^*	1.79 (0.99-3.21)^**
Frail (RAI-rev = 31-40)	2.64 (1.53-4.54)	7.33 (5.39-9.96)	3.56 (2.91-4.36)	6.65 (2.95-14.96)	2.37 (1.60-3.52)	6.69 (2.75-16.26)	2.86 (1.93-4.24)	3.27 (1.83-5.85)
Severely frail (RAI-rev ≥41)	18.23 (9.67-34.37)	4.26 (28.34-57.07)	11.50 (8.87-14.91)	21.02 (9.03-48.96)	6.80 (4.35-10.62)	24.32 (9.70-60.98)	6.06 (3.86-9.51)	5.35 (2.74-10.44)

Outcomes included: mortality, discharge to nonhome destination, eLOS, minor complication, major complication, Clavien-Dindo grade IV complication, unplanned readmission, and reoperation. Values presented as odds ratio (95% confidence interval). Abbreviations: mFI-5; the 5-factor modified frailty index, RAI-Rev; recalibrated version of the risk analysis index, eLOS; extended length of stay. All categories had P < 0.001 unless indicated otherwise.

*for 0.001 < P < 0.05.

** indicated P ≥ 0.5.

^aThe prefrail and nonfrail groups were merged due to the absence of mortality among patients with RAI-Rev ≤10, resulting in a combined group (RAI-Rev ≤20) that was used as the reference for comparison with frail and severely frail groups.

ROC Curve Analysis and C Statistic

AUROC curve analysis was conducted for DCM patients to determine the discriminatory accuracy of the mFI-5 and RAI-Rev for predicting the primary postoperative outcomes (Figure 2 and Table 3). For NHD prediction, RAI-Rev had significantly greater performance (C-statistic = 0.72, 95% CI, 0.71 - 0.73) than the mFI-5 (C-statistic = 0.64, 95% CI, 0.63 - 0.65, P < 0.00. Similarly, for extended length of stay (eLOS), RAI-Rev significantly outperformed mFI-5 (RAI-Rev: C-statistic = 0.64, 95% CI, 0.63 - 0.65 vs mFI-5: C-statistic = 0.60, 95% CI, 0.59 - 0.61; P < 0.0001). RAI-Rev further outperformed the mFI-5 in predicting minor complications and readmission. It also demonstrated superior discrimination than mFI-5 only for reoperation (P = 0.0007). There was no significant difference in discriminative thresholds for predicting postoperative mortality.

Figure 2.

AUC curves [(a) mortality, (b) nonhome discharge, (c) eLOS, (d) minor complications, (e) major complications, (f) Clavien-Dindo IV complications, (g) readmission, (h) reoperation] and forest plot for prediction of 30-day outcomes after surgery for DCM by the mFI-5 and RAI-Rev.

Table 3.

Area under the curve analysis with concordance statistic to compare the predictive capabilities of the mFI-5 and RAI-Rev for 30-day outcomes.

	Frailty Index	AUC (95% CI)	P-value
Nonhome discharge	mFI-5	0.64 (0.63-0.65)	<0.0001
Nonhome discharge	RAI-Rev	0.72 (0.71-0.73)	Reference
Mortality	mFI-5	0.71 (0.66-0.77)	0.5610
Mortality	RAI-Rev	0.73 (0.67-0.79)	Reference
eLOS	mFI-5	0.60 (0.59-0.61)	<0.0001
eLOS	RAI-Rev	0.64 (0.63-0.65)	Reference
Minor complication	mFI-5	0.64 (0.61-0.66)	0.0325
Minor complication	RAI-Rev	0.67 (0.64-0.69)	Reference
Major complication	mFI-5	0.62 (0.60-0.64)	0.1601
Major complication	RAI-Rev	0.64 (0.62-0.65)	Reference
Clavien-Dindo IV complication	mFI-5	0.65 (0.63-0.68)	0.4534
Clavien-Dindo IV complication	RAI-Rev	0.66 (0.64-0.69)	Reference
Readmission	mFI-5	0.59 (0.57-0.61)	0.0011
Readmission	RAI-Rev	0.63 (0.61-0.64)	Reference
Reoperation	mFI-5	0.56 (0.53-0.58)	0.0007
Reoperation	RAI-Rev	0.61 (0.58-0.63)	Reference

Abbreviations: mFI-5; the 5-factor modified frailty index, RAI-Rev; recalibrated version of the Risk Analysis Index, eLOS; extended Length of Stay, AUC; area under the curve.

A subgroup AUROC analysis was performed as well to understand the discriminatory accuracy of frailty measures stratified by ADF or PDF (Table 4). For ADF, the RAI-Rev had superior discrimination compared to mFI-5 for NHD (RAI-Rev: C-statistic = 0.70, 95% CI, 0.68 - 0.71 vs mFI-5: C-statistic = 0.64, 95% CI, 0.63 - 0.66), eLOS (RAI-Rev: C-statistic = 0.62 , 95% CI, 0.61-0.64 vs mFI-5: C-statistic = 0.58, 95% CI, 0.56-0.60), and readmission (RAI-Rev: C-statistic = 0.64, 95% CI, 0.62-0.67 vs mFI-5: C-statistic = 0.60, 95% CI, (0.58-0.63, but not mortality. RAI-Rev had superior discrimination compared to mFI-5 for predicting minor complications, major complications, CDIV complications, and reoperation.

Table 4.

Area under the curve analysis with delong test to compare the discriminatory accuracy of the mFI-5 and RAI-Rev for 30-day outcomes for anterior versus posterior approaches.

		Anterior Approach		Posterior Approach
	Frailty Index	AUC (95% CI)	P-value	AUC (95% CI)	P-value
Nonhome discharge	mFI-5	0.64 (0.63-0.66)	<0.0001	0.61 (0.60, 0.63)	<0.0001
Nonhome discharge	RAI-Rev	0.70 (0.68-0.71)	Reference	0.66 (0.65-0.68)	Reference
Mortality	mFI-5	0.69 (0.60-0.77)	0.157	0.74(0.66- 0.82)	0.568
	RAI-Rev	0.75 (0.68-0.81)	Reference	0.71 (0.60-0.81)	Reference
eLOS	mFI-5	0.61 (0.59-0.62)	0.031	0.56 (0.54-0.57)	0.400
	RAI-Rev	0.62 (0.61-0.64)	Reference	0.56 (0.55-0.58)	Reference
Minor complication	mFI-5	0.64 (0.60-0.68)	0.398	0.60 (0.57-0.64)	0.858
	RAI-Rev	0.66 (0.62-0.70)	Reference	0.61 (0.57-0.64)	Reference
Major complication	mFI-5	0.63 (0.60-0.66)	0.118	0.59 (0.56-0.62)	0.030
	RAI-Rev	0.654 (0.63-0.68)	Reference	0.56 (0.53-0.59)	Reference
Clavien-Dindo IV complication	mFI-5	0.66 (0.62-0.69)	0.208	0.65 (0.61-0.69)	0.846
	RAI-Rev	0.68 (0.65-0.72)	Reference	0.61 (0.57-0.65)	Reference
Readmission	mFI-5	0.60 (0.58-0.63)	0.007	0.56 (0.52-0.59)	0.687
	RAI-Rev	0.64 (0.62-0.67)	Reference	0.56 (0.53-0.59)	Reference
Reoperation	mFI-5	0.56 (0.53-0.60)	0.005	0.53 (0.49-0.57)	0.721
	RAI-Rev	0.614 (0.58-0.65)	Reference	0.54 (0.50-0.58)	Reference

Abbreviations: mFI-5; the 5-factor modified frailty index, RAI-Rev; recalibrated version of the risk analysis index, eLOS; extended length of stay, AUC; area under the curve.

For PDF, the RAI-Rev demonstrated superior discrimination compared to mFI-5 for NHD (RAI-Rev: C-statistic = 0.70, 95% CI, 0.68-0.71 vs mFI-5: C-statistic = 0.64, 95% CI, 0.63 - 0.66). There were no significant differences between frailty indexes for mortality, CDIV complications, readmission or reoperation. Further, discriminatory accuracy was poor with all measures demonstrating a C statistic ≤0.60 for all outcomes except NHD, minor complications, CDIV complications and mortality with the posterior approach.

Discussion

In this surgical registry analysis of 18,138 cervical decompression and fusion DCM patients, increasing patient frailty was associated with nonhome discharge, extended LOS, readmission, and minor complications. In AUROC curve analysis, the RAI demonstrated superior discrimination compared to the mFI for predicting NHD (C-statistic: 0.72 vs 0.65), extended LOS (C-statistic: 0.64 vs 0.60), minor complications, and readmission. When stratified by surgical approach, the RAI-Rev was superior to the mFI-5 in predicting outcomes following ADF, while all frailty metrics demonstrated poor discrimination with PDF.

The incidence of age-related diseases, including DCM, is increasing in proportion with the increasingly aging and frail population. Surgical treatment for DCM is relatively safe with in-hospital mortality ranging from 0.03 - 1.33%; however, complication rates are as high as 14.1%.^22,23 To the authors’ knowledge, this is the first study which directly analyzes the discriminatory threshold of two commonly used frailty indices in DCM patients. These findings support the RAI as the leading frailty index to aid in preoperative risk stratification of DCM patients. Furthermore, ADF for DCM patients may play a more important role in predicting surgical outcomes in increasingly frail patients, PDF may be the most risk-averse option for frail DCM patients. Adverse outcomes are expected in patients with decreased baseline physiologic reserve; however, the accurate quantification of a patient’s frailty is critical. The present findings support the RAI as the most practical frailty index, as it is used at the point of care, with the best discrimination and suggest its utility in identifying the more high-risk DCM patients for improved risk-benefit discussion preoperatively.

Frailty in spine disorders is well described within the literature; however, there are no studies which examine RAI as a measure of frailty in the context of only DCM. Wilson et al demonstrated the impact of frailty on functional and quality of life metrics following surgery in DCM patients using the 5-factor modified frailty index (mFI-5).²⁴ Multiple studies demonstrated the utility of the 11-factor modified frailty index (mFI-11) and mFI-5 in predicting mortality and morbidity across a host of spinal pathologies.^{10,11,18,25-27} For DCM specifically, studies have shown that the mFI predicts adverse outcomes better than age alone and may have clinical utility in understanding improvement in quality of life measures following surgical treatment of DCM. However, the mFI have been increasingly scrutinized in recent years for resembling comorbidity indices rather than measures of true decreased baseline physiologic reserve.^14,28 The RAI has emerged as a more quantitatively robust frailty scale and has demonstrated superior outcome discrimination compared to the mFI across a range of cranial and spinal pathologies.^{13,17,18,29-31} The present results corroborate these findings as the RAI outperformed the mFI in predicting NHD, eLOS, minor complications and readmission following decompression and fusion in DCM patients.

In addition to consideration of a patient’s frailty status, a surgeon’s decision regarding anterior vs posterior surgery is a critical aspect of the preoperative decision-making process. ADF and PDF are performed at similar rates in DCM patients.³² ADF is associated with decreased blood loss, shorter LOS, lower odds of infection and C5 palsy, and superior outcomes in robust patients.³³ However, ADF is associated with higher odds of dysphagia.³² Interestingly, frailty has been investigated as a predictor of adverse outcomes in DCM, however much of the current literature focuses on ADF. Further, the few studies examining frailty in the context of PDF did not perform an analysis of discriminatory accuracy, necessary for risk stratification.^8,33 One explanation for a general lack of literature on this topic may be that frailty does not effectively predict outcomes in PDF, like the results presented in this study. The present results suggest that ADF in frail DCM patients is more influential in leading to adverse postoperative outcomes. The association between dysphagia and frailty is well-established, with some studies suggesting inclusion of dysphagia in baseline frailty assessment.³⁴ Posterior approaches in older patients are often selected by surgeons to avoid dysphagia.^35-37 Additionally, posterior approaches to decompression are associated with significantly superior neurological and clinical improvement than anterior in cervical spondylotic myelopathy.^38,39 Our findings support the understanding that patient frailty is superior to patient age in predicting surgical adverse outcomes, while independently supporting the association between a posterior approach to DCM results in a lower incidence of adverse outcomes in frail patients. These results offer quantitative evidence of frailty’s greater impact on ADF outcomes for DCM patients and support the commonly observed surgeon preference of selecting posterior approaches for DCM as patient frailty increases.

Limitations

The present study utilized a large, multicenter national database, which limits the granularity of the results and relies solely on the variables coded within the NSQIP. Specifically, the NSQIP does not include variables related to preoperative cognitive status, and therefore these variables were excluded from the RAI-Rev frailty calculations. Additionally, the NSQIP does not differentiate between “poor appetite” and “weight loss,” which are both reported in the database under “WTLOSS.” These limitations are similarly reported in previous studies which show that even with variable modifications to the RAI-Rev, there is still excellent discriminative ability of these frailty measures. Finally, unlike previous studies which used conventional RAI frailty tiers (robust: 0-20; normal, 21-30; frail, 31-40; very frail, 41+), our study utilized robust ≤10, normal 11-20, frail 21-30, and very frail ≥31, as the patient cohort when utilizing 41+ for very frail was minimal. This suggests that patients undergoing decompression for DCM are not as frail as other spine surgery populations, and a larger patient cohort may be necessary to capture outcomes in very frail patients.

Conclusion

The RAI-Rev demonstrated superior discrimination for predicting NHD, eLOS, readmission, and minor complications in DCM patients undergoing decompression and fusion compared to the mFI-5. Further, the role of frailty is less pronounced with posterior approaches compared to anterior approaches for DCM, providing evidence for utilizing PDF in increasingly frail DCM patients. The RAI-Rev may aid in surgical candidate selection and inform preoperative risk stratification of these patients based on frailty strata.

Supplemental Material

Supplemental Material - Discriminatory Value of the Risk Analysis Index Versus the 5-Factor Modified Frailty Index for Major Outcome Measures in Degenerative Cervical Myelopathy

Supplemental Material for Discriminatory Value of the Risk Analysis Index Versus the 5-Factor Modified Frailty Index for Major Outcome Measures in Degenerative Cervical Myelopathy by Nithin Gupta, Sina Zoghi, Michael M. Covell, Chase Smitterberg, Stefan T. Prvulovic, William T. DiCiurcio, Johnny Delashaw, Meic H. Schmidt, Marc D. Moisi, and Christian A. Bowers in Global Spine Journal.

Footnotes

Author Contributions

Conceptualization (NKG, SZ, CAB), Methodology (NKG, SZ, CAB), Formal Analysis (SZ), Writing - Original Draft (NKG, SZ, MMC, CS, STP), Writing - Review & Editing (NKG, MMC, WTD, JD, MHS, MDM, CAB), Supervision (WTD, JD, MHS, MDM, CAB), Project Administration (MHS, MDM, CAB).

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

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

Ethical Statement

ORCID iDs

Nithin K. Gupta

Sina Zoghi

Stefan T. Prvulovic

Christian A. Bowers

Data Availability Statement

Data used to produce this manuscript are available from the corresponding author upon reasonable request.*

Supplemental Material

Supplemental material for this article is available online.

Appendix

References

Saunders

Sandhu

McBride

Maniarasu

Taylor

Dhokia

. Degenerative cervical myelopathy: an overview. Cureus. 2023;15(12):e50387. doi:10.7759/cureus.50387

Jones

Dolsten

. Healthcare on the brink: navigating the challenges of an aging society in the United States. [published correction appears in NPJ Aging. 2024;10(1):25]. NPJ Aging. 2024;10(1):22. doi:10.1038/s41514-024-00148-2

Badhiwala

Ahuja

Akbar

, et al. Degenerative cervical myelopathy: update and future directions. Nat Rev Neurol. 2020;16(2):108-124. doi:10.1038/s41582-019-0303-0

Gulati

Vangen-Lønne

Nygaard

ØP

, et al. Surgery for degenerative cervical myelopathy: a nationwide registry-based observational study with patient-reported outcomes. [published correction appears in Neurosurgery. 2021;89(5):943]. Neurosurgery. 2021;89(4):704-711. doi:10.1093/neuros/nyab259

Vargas Castillo

Das

Diwan

. Degenerative cervical myelopathy: insights into its pathobiology and molecular mechanisms. J Clin Med. 2021;10:1214. doi:10.3390/jcm10061214

Xue

. The frailty syndrome: definition and natural history. Clin Geriatr Med. 2011;27(1):1-15. doi:10.1016/j.cger.2010.08.009

Covell

Warrier

Rumalla

Dehney

Bowers

. RAI-measured frailty predicts non-home discharge following metastatic brain tumor resection: national inpatient sample analysis of 20,185 patients. J Neuro Oncol. 2023;164(3):663-670. doi:10.1007/s11060-023-04461-w

Elsamadicy

Sayeed

Sherman

, et al. Impact of preoperative frailty on outcomes in patients with cervical spondylotic myelopathy undergoing anterior vs. posterior cervical surgery. J Clin Med. 2024;13(114):1-20. doi:10.3390/jcm13010114

Pedro

Alvi

Hejrati

Moghaddamjou

Fehlings

. Elderly patients show substantial improvement in health-related quality of life after surgery for degenerative cervical myelopathy despite medical frailty: an ambispective analysis of a multicenter international data set. Neurosurgery. 2024;94:1122-1131. doi:10.1227/neu.0000000000002818

10.

Wilson

JRF

Badhiwala

Moghaddamjou

Yee

Wilson

Fehlings

. Frailty is a better predictor than age of mortality and perioperative complications after surgery for degenerative cervical myelopathy: an analysis of 41,369 patients from the NSQIP database 2010–2018. J Clin Med. 2020;9(11):3491. doi:10.3390/jcm9113491

11.

Wilson

JRF

Badhiwala

Moghaddamjou

, et al. Adverse effects of frailty on the outcomes of surgery for degenerative cervical myelopathy: results from a prospective multicenter international data set of 757 patients. J Neurosurg Spine. 2023;39(6):815-821. doi:10.3171/2023.6.SPINE23461

12.

Arya

Varley

Youk

, et al. Recalibration and external validation of the risk analysis index: a surgical frailty assessment tool. Ann Surg. 2020;272(6):996-1005. doi:10.1097/SLA.0000000000003276

13.

Yocky

Owodunni

Courville

, et al. The risk analysis index has superior discrimination compared with the modified frailty index-5 in predicting worse postoperative outcomes for the octogenarian neurosurgical patient. Neurosurg Pract. 2023;4(3):e00044. doi:10.1227/neuprac.0000000000000044

14.

Owodunni

Uzoukwu

Courville

Schmidt

Bowers

. The fine line between simplicity and oversimplification: comparing the risk analysis index and 5-factor modified frailty index as frailty assessment tools. Neurospine. 2023;20(2):728-730. doi:10.14245/ns.2346496.248

15.

Hall

Arya

Schmid

, et al. Development and initial validation of the risk analysis index for measuring frailty in surgical populations. JAMA Surg. 2017;152(2):175-182.

16.

Hall

Arya

Schmid

, et al. Association of a frailty screening initiative with postoperative survival at 30, 180, and 365 days. JAMA Surg. 2017;152(3):233-240.

17.

Bowers

Varela

Conlon

, et al. Comparison of the risk analysis index and the modified 5-factor frailty index in predicting 30-day morbidity and mortality after spine surgery. J Neurosurg Spine. 2023;39(1):136-145. doi:10.3171/2023.2.SPINE221019

18.

Conlon

Thommen

Kazim

, et al. Risk analysis index and its recalibrated version predict postoperative outcomes better than 5-factor modified frailty index in traumatic spinal injury. Neurospine. 2022;19(4):1039-1048. doi:10.14245/ns.2244326.163

19.

Velanovich

Antoine

Swartz

Peters

Rubinfeld

. Accumulating deficits model of frailty and postoperative mortality and morbidity: its application to a national database. J Surg Res. 2013;183(1):104-110. doi:10.1016/j.jss.2013.01.021

20.

Chimukangara

Helm

Frelich

, et al. A 5-item frailty index based on NSQIP data correlates with outcomes following paraesophageal hernia repair. Surg Endosc. 2017;31(6):2509-2519. doi:10.1007/s00464-016-5253-7

21.

Fehlings

Tetreault

Dettori

, et al. Change in functional impairment, disability, and quality of life following operative treatment for degenerative cervical myelopathy: a systematic review and meta-analysis. GSJ. 2017;7:53-69.

22.

Wang

Chan

Maiman

Kreuter

Deyo

. Complications and mortality associated with cervical spine surgery for degenerative disease in the United States. Spine. 2007;32:342-347.

23.

Wilson

JRF

Badhiwala

Moghaddamjou

24.

Jang

Chin

Park

, et al. Influence of frailty on life expectancy in octogenarians after lumbar spine surgery. Neurospine. 2021;18:303-310.

25.

Shinall

Jr Arya

Youk

, et al. Association of preoperative patient frailty and operative stress with postoperative mortality. JAMA Surg. 2020;155(1):e194620. doi:10.1001/jamasurg.2019.4620

26.

Weaver

Malik

Jain

, et al. 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;S1878-8750(19):30038.

27.

Covell

Roy

Rumalla

, et al. The limited utility of the hospital frailty risk score as a frailty assessment tool in neurosurgery: a systematic review. Neurosurgery. 2024;94(2):251-262. doi:10.1227/neu.0000000000002668

28.

Moya

Owodunni

Harrison

, et al. Preoperative frailty risk in cranioplasty patients: risk analysis index predicts adverse outcomes. Plast Reconstr Surg Glob Open. 2023;11(6):e5059. doi:10.1097/GOX.0000000000005059

29.

Covell

Rumalla

Kassicieh

, et al. Frailty measured by risk analysis index and adverse discharge outcomes after adult spine deformity surgery: analysis of 3104 patients from a prospective surgical registry (2011-2020). Spine J. 2023;23(5):739-745. doi:10.1016/j.spinee.2022.12.014

30.

Dicpinigaitis

Kazim

Al-Mufti

, et al. Frailty in aneurysmal subarachnoid hemorrhage: the risk analysis index. J Neurol. 2023;270(10):4820-4826. doi:10.1007/s00415-023-11805-z

31.

Sattari

Ghanavatian

Feghali

, et al. Anterior cervical discectomy and fusion versus posterior decompression in patients with degenerative cervical myelopathy: a systematic review and meta-analysis. J Neurosurg Spine. 2023;38:631-643. doi: 10.3171/2023.1.SPINE221244. Published online February 24.

32.

Evaniew

Bailey

Rampersaud

, et al. Anterior vs posterior surgery for patients with degenerative cervical myelopathy: an observational study from the Canadian Spine outcomes and research network. Neurosurgery. 2024;95:53–62. doi: 10.1227/neu.0000000000002842. Published online February 2.

33.

Shin

Kothari

Phan

, et al. Frailty index as a predictor of adverse postoperative outcomes in patients undergoing cervical spinal fusion. Spine. 2017;42(5):304-310. doi:10.1097/BRS.0000000000001755

34.

Yang

Chen

Lee

Chen

. Association between dysphagia and frailty in older adults: a systematic review and meta-analysis. Nutrients. 2022;14(9):1812. doi:10.3390/nu14091812

35.

Johansen

Vangen-Lønne

Holmberg

, et al. Surgery for degenerative cervical myelopathy in the elderly: a nationwide registry-based observational study with patientreported outcomes. Acta Neurochir. 2022;164(9):2317-2326. doi:10.1007/s00701-022-05282-y

36.

Audat

Fawareh

Radydeh

, et al. Anterior versus posterior approach to treat cervical spondylotic myelopathy: clinical and radiological results with long period of follow-up. SAGE Open Med. 2018;6:2050312118766199. doi:10.1177/2050312118766199

37.

Epstein

. Spine surgery in geriatric patients: sometimes unnecessary, too much, or too little. Surg Neurol Int. 2011;2:188. doi:10.4103/2152-7806.91408

38.

AOSpine North America . Anterior versus posterior surgical approaches to treat cervical spondylotic myelopathy: outcomes of the prospective multicenter AOSpine North America CSM study in 264 patients. Spine. 2013;38:2247-2252. doi:10.1097/BRS.0000000000000047

39.

Wilson

JRF

Badhiwala

, et al. Surgery for degenerative cervical myelopathy: an update with clinical guidelines. J Clin Med. 2024;13(2):445-467. doi:10.3390/jcm13020445

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.01 MB