Letter: A Pathway to Safe Spine Surgery in Underweight Frail Patients: The Revised Risk Analysis Index Displays Remarkable Discrimination for 30-Day Postoperative Mortality and Nonhome Discharge.

To the Editor: The United States has the highest incidence of spine surgery with approximately 900 000 adult patients undergoing various spine procedures annually.1 Therefore, identifying risk factors for unfavorable postoperative outcomes is crucial for improving patient-centered decision-making, coordinating perioperative care, and optimizing quality improvement efforts.2,3 Underweight frail patients may encounter challenges with spine surgery and postoperative recovery.2,4 Their increased risk of unfavorable postoperative outcomes may be attributed to reduced system-wide resilience, potentially impeding wound healing and recovery.2,4 These patients have increased risk of developing major and life-threatening complications, 30-day postoperative mortality, and nonhome discharge (NHD).2,4 Frailty is defined as a multimodal, age-related decline in physiological reserve, resulting in a decreased ability to cope with imposed stressors.5,6 A variety of clinical and biochemical variables can be used to assess frailty by examining various frailty domains and offering a comprehensive insight into a patient's overall health status.5,6 Factors such as the presence of multiple comorbidities, limitations in daily activities/increased functional dependence, and diminished cognitive function all play a part in measuring patients' frailty.4-6 Preoperative frailty risk assessment is vital in shared decision-making because it can significantly enhance preoperative patient education, surgical planning, precise risk-benefit discussions, and provide multidisciplinary approach to care delivery.3-6 Ultimately, frailty assessment may lead to improved patient outcomes because it allows for focusing extra resources and time for more vulnerable patients.3 The paucity of data between frailty and low body mass index (BMI) provides an opportunity for investigation. We evaluated the impact of frailty in underweight (BMI <18.0 kg/m2) patients within a subset of spine surgery cases involving a mix of single and concurrent procedures. Most of these procedures were fusion procedures, consisting of anterior, posterior, or combined arthrodesis, but there were also single-level and multilevel laminotomy with decompression. Our evaluation focused on the discrimination of the Risk Analysis Index-revised (RAI-rev), 5-Factor Modified Frailty Index-5 (mFI-5), greater patient age, and BMI for 30-day postoperative mortality and NHD. We retrospectively analyzed data from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database from 2012 to 2020. We included 2718 underweight patients who underwent spine procedures. We examined the distribution of patient characteristics, including demographics, clinical parameters, and postoperative outcomes (Table). Our Institutional Review Board approved this study, and we adhered to the ACS-NSQIP data use agreement. We used the RAI-rev,6 a simple-to-administer 11-item questionnaire that assesses 5 domains of frailty, including social, physical, cognitive, nutritional, and functional aspects, and the mFI-5,4 which is based on a 5-factor index: diabetes, chronic obstructive pulmonary disease, functional dependency, congestive heart failure, and hypertension (Table), to classify patients into 1 of 4 categories: robust, typical, frail, and very frail. We present descriptive statistics and receiver operator characteristic comparisons for RAI-rev, mFI-5, greater patient age, and BMI. TABLE. - Demographics and Clinical Characteristics of Underweight Spine Surgery Patients Variables TotalN = 2718 Age, median (IQR), y 60.0 (50.0, 70.0) Sex, n (%) Male 914 (33.6) Female 1804 (66.4) Race, n (%) White 2076 (76.4) Black 247 (9.1) Asian 99 (3.6) Other a 296 (10.9) BMI, median (IQR), kg/m2 17.7 (16.8, 18.1) Underweight category, n (%) Mild (17.00-18.49 kg/m2) 1950 (71.7) Moderate (16.00-16.99 kg/m2) 445 (16.4) Severe (<16.00 kg/m2) 323 (11.9) HTN, n (%) 938 (34.5) Diabetes, n (%) 173 (6.4) COPD, n (%) 296 (10.9) CHF, n (%) 11 (0.4) Functional status, n (%) Dependent 237 (8.7) Chronic kidney disease, n (%) 21 (0.78) Cancer diagnosis, n (%) 183 (6.7) RAI-rev, n (%) Robust 328 (12.1) Typical 1124 (41.4) Frail 886 (32.6) Very frail 380 (14.0) 5-Factor Modified Frailty Index, n (%) Robust 1438 (52.9) Typical 957 (35.2) Frail 273 (10.0) Very frail 50 (1.8) Operative approach Emergency 615 (22.6) Elective 2103 (77.4) Operative time, median (IQR), min 120.0 (75.0, 184.0) Major complications, n (%) b Clavien-Dindo (III-IV) 146 (5.4) LOS, median (IQR), d 3.0 (1.0, 6.0) NHD, n (%) 495 (18.2) 30-day postoperative mortality, n (%) 61 (2.2) BMI, body mass index; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; HTN, hypertension, NHD, nonhome discharge; LOS, length of hospital stay; mFI-5, Modified Frailty Index-5; RAI-rev, Risk Analysis Index-revised.aOther: American Indian/Alaska Native, Native Hawaiian/Pacific Islander, and Unknown.bMajor complications consists of all postoperative complications categorized by Clavein-Dindo classification grades III-IV (life-threatening complications).RAI-rev: Total = 81: ≤10—“Robust,” 11–20—“Typical,” 21–30—“Frail,” and ≥31—“Very Frail.”Score assignment: age (cancer positive diagnosis) <19 (+28), 20–29 (+29), 30–39 (+30), 40–49 (+31), 50–59 (+32), 60–64 (+33), 65–74 (+34), 75–84 (+35), 85–89 (+36), and (cancer negative diagnosis) <19 (+0), 20–24 (+1), 25–29 (+4), 30–34 (+6), 35–39 (+8), 40–44 (+10), 45–49 (+12), 50–54 (+14), 55–59 (+16), 60–64 (+18), 65–69 (+20), 70–74 (+22), 75–79 (+24), 80–84 (+26), 85–89 (+28). Male sex (+3); chronic kidney disease (end-stage and/or dialysis) (+8); congestive heart failure (+5); dyspnea (+3); poor appetite (+4); unintentional weight loss >10 lbs within 6 months (+4); residence at high-level of care facility (decrease reserve) (+1); assistance required for completion of activities of daily living (without cognitive evaluation): partially dependent (+7) and totally dependent (+14).mFI-5: Total = 5: 0—“Robust,” 1—“Typical,” 2—“Frail,” and ≥3—“Very Frail.”Chronic obstructive pulmonary disease (+1); congestive heart failure (+1); diabetes mellitus (requiring oral medication or insulin) (+1); hypertension (requiring oral medication) (+1); partial and total functional dependence (+1). The median age of these underweight patients was 60 years (IQR: 50-70), and a larger proportion of patients were White 2076/2718 (76.4%) and female 1804/2718 (66.4%). The median BMI was 17.7 kg/m2 (IQR: 16.8-18.1), and a larger proportion of patients were mildly underweight 1950/2718 (71.7%) (Table). By RAI-rev, 886/2718 (32.6%) were frail and 380/2718 (14.0%) were very frail. While by mFI-5, 273/2718 (10.0%) were frail and 50/2718 were (1.8%) very frail. The NHD rate was 18.2% (495/2718), and the 30-day postoperative mortality rate was 2.2% (61/2718) (Table). RAI-rev demonstrated better discrimination for 30-day postoperative mortality and NHD (Figure A and B). For 30-day postoperative mortality, the C-statistics were RAI-rev 0.900 (95% CI: 0.866-0.935) vs mFI-5 0.692 (95% CI: 0.632-0.752) vs greater patient age 0.677 (95% CI: 0.618-0.737) vs BMI 0.430 (95% CI: 0.357-0.503), Delong P < .001 (Figure A). While for NHD, the C-statistics were RAI-rev 0.709 (95% CI: 0.684-0.734) vs mFI-5 0.624 (95% CI: 0.599-0.649) vs greater patient age 0.659 (95% CI: 0.633-0.685) vs BMI 0.440 (95% CI: 0.412-0.469), Delong P < .001 (Figure B).FIGURE 1.: Outcome-based predictive thresholds of frailty screening tools, age, and BMI for underweight patients undergoing spine surgery outlined by A, 30-day postoperative mortality and B, NHD. A, Overall 30-day postoperative mortality C-statistics (95% CI): RAI-revised 0.900 (0.866-0.935) vs mFI-5 0.692 (0.632-0.752) vs age 0.677 (0.618-0.737) vs BMI 0.430 (0.357-0.503), Delong P < .001. Pairwise Delong P values: all RAI-revised pairwise comparisons P < .001; mFI-5 vs Age P = .81; mFI-5 vs BMI P < .001. Emergency procedures: 30-day postoperative mortality C-statistics (95% CI): RAI-revised 0.817 (0.763-0.871) vs mFI-5 0.633 (0.559-0.706) vs age 0.669 (0.597-0.741) vs BMI 0.508 (0.421-0.594), Delong P < .001. Pairwise Delong P values: all RAI-revised pairwise comparisons P < .001; mFI-5 vs Age P = .39; mFI-5 vs BMI P = .02. Elective procedures: 30-day postoperative mortality C-statistics (95% CI): RAI-revised 0.890 (0.804-0.976) vs mFI-5 0.744 (0.633-0.855) vs age 0.643 (0.513-0.772) vs BMI 0.331 (0.190-0.471), Delong P < .001. Pairwise Delong P values: RAI-revised vs mFI-5 P = .03; RAI-revised vs age P = .002; RAI-revised vs BMI P < .001; mFI-5 vs age P = .0.5; mFI-5 vs BMI P < .001. B, Overall NHD C-statistics (95% CI): RAI-revised 0.709 (0.684-0.734) vs mFI-5 0.624 (0.599-0.649) vs age 0.659 (0.633-0.685) vs BMI 0.440 (0.412-0.469), Delong P < .001. Pairwise Delong P values: all RAI-revised pairwise comparisons P < .001; mFI-5 vs age P = .03; mFI-5 vs BMI P < .001. Emergency procedures: NHD C-statistics (95% CI): RAI-revised 0.817 (0.763-0.871) vs mFI-5 0.633 (0.559-0.706) vs Age 0.669 (0.597-0.741) vs BMI 0.508 (0.421-0.594), Delong P < .001. Pairwise Delong P values: RAI-revised vs mFI-5 P = .26; RAI-revised vs Age P = .94; RAI-revised vs BMI P < .001; mFI-5 vs age P = .29; mFI-5 vs BMI P = .01. Elective procedures: NHD C-statistics (95% CI): RAI-revised 0.724 (0.693-0.756) vs mFI-5 0.644 (0.611-0.677) vs age 0.696 (0.663-0.729) vs BMI 0.453 (0.416-0.490), Delong P < .001. Pairwise Delong P values: RAI-revised vs mFI-5 P < .001; RAI-revised vs age P = .02; RAI-revised vs BMI P < .001; mFI-5 vs age P = .01; mFI-5 vs BMI P < .001. BMI, body mass index; mFI-5, Modified Frailty Index-5; NHD, nonhome discharge; RAI, Risk Analysis Index.These findings emphasize that frailty is a substantial predictor of 30-day postoperative mortality and NHD after spinal surgery in underweight patients. These results demonstrate that the RAI-rev demonstrates superior discrimination compared with the mFI-5, greater patient age, and BMI. This information advocates for precise risk assessment in high-risk patients, allowing for targeted preoperative optimization strategies.2,5,6 Using preoperative optimization strategies, such as prehabilitation, may improve spine surgery outcomes for frail or underweight frail patients, thereby minimizing the risk of adverse postoperative outcomes.2,3,5,6 The data used in this study are publicly available and can be accessed by researchers on request. Detailed information about the data set, including access requirements and procedures, can be found on the ACS website. Data sharing adheres to the guidelines and policies set forth by the data provider. A limitation of this study was the reliance on BMI scores from the ACS-NSQIP, which hindered our ability to assess sarcopenia. Nevertheless, our findings emphasize the importance of incorporating additional frailty risk assessments, such as the RAI-rev, into preoperative evaluations for spine surgery patients, particularly for underweight patients. This tool can identify older male or female patients with a cancer diagnosis and weight loss as very frail (RAI-rev score >31), enabling health care providers to prioritize prehabilitation and multidisciplinary surgery planning, as well as tailor postoperative management for successful recovery.3 In conclusion, it is crucial to raise awareness about the risks of major spine surgery in frail underweight patients and emphasize the importance of comprehensive preoperative risk assessments. Both patients and health care providers must be mindful of these factors to make informed decisions about the proposed surgery. By encouraging a collaborative approach, we can work together to ensure the best possible outcomes for these high-risk patients.