Thirty-Day Complications Following Anterior Lumbar Interbody Fusion Versus Lumbar Disc Arthroplasty: A Propensity Score Matched Analysis

Abstract

The anterior lumbar interbody fusion (ALIF) and lumbar disc arthroplasty (LDA) procedures are both commonly performed to improve the quality of life and pain in people with lower back pain. However, few recent studies have compared 30-day complications on a large scale. The objectives of this study were to compare the 30-day complications seen after ALIF and LDA and identify risk factors for these complications. The National Surgical Quality Improvement Program (NSQIP) database was queried between the years 2012–2021 (10 years in total) for records of patients who underwent either ALIF or LDA as a primary procedure. Patients in each group underwent a 1:1 propensity match for age, gender, BMI, ASA status, diabetes mellitus (DM), hypertension requiring medication, congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), inpatient/outpatient status, smoking status, and bleeding disorders. Bivariate and multivariate analyses were performed to determine significant differences in complications and risk factors between these cohorts. A total of 1218 propensity-matched subjects, with 609 receiving ALIF and 609 receiving LDA, were included in the analyses of this study. The incidence of extended length of stay (LOS) (>4 days) was higher in the ALIF cohort compared to the LDA cohort (14.6% vs. 4.76%, p < 0.001). Multivariate analysis demonstrated that subjects who underwent LDA had lower odds (Odds Ratio [OR]: 0.457; 95% Confidence Interval [CI]: 0.283–0.738, p = 0.001) of experiencing extended LOS compared to the ALIF cohort. Longer operative times increased the odds of prolonged LOS in both cohorts. The results of this study suggest that ALIF is associated with longer LOS than LDA when baseline demographic data are controlled. Further, longer operative times increase the odds that subjects receiving either ALIF or LDA will experience a prolonged LOS. Besides extended LOS, ALIF and LDA produce a relatively similar 30-day complication profile.

1. Introduction

Interbody lumbar fusion has historically been the mainstay of surgical treatment for degenerative disc disease (DDD) of the spine [1,2]. The anterior lumbar interbody fusion (ALIF) procedure utilizes an anterior approach that avoids disruption of the paraspinal muscles and mitigates the risk of neural injury which is more commonly seen in posterior approach techniques [3]. Additionally, the ALIF may provide superior radiological outcomes postoperatively and mitigate blood loss when compared to a posterior lumbar interbody fusion [3]. For these reasons, the frequency of ALIF procedures is increasing annually [4]. A common criticism of the ALIF is that it limits the lumbar range of motion (ROM) at the fused vertebrae which can lead to accelerated degeneration at adjacent levels [5,6,7,8,9]. In contrast, the lumbar disc arthroplasty (LDA) aims to address DDD without compromising lumbar mobility [10]. Additionally, LDA may reduce adjacent disc disease [11,12,13].

The National Surgery Quality Improvement Program (NSQIP) database, via the American College of Surgeons (ACS), tracks 30-day complications after numerous inpatient and outpatient surgeries at 700 participating institutions. Two prior studies have compared 30-day outcomes after ALIF and LDA using the NSQIP database [14,15]. Schultz et al. analyzed 30-day outcomes after ALIF and LDA from the years 2011–2015. They found that subjects undergoing ALIF had a significantly higher risk of requiring postoperative blood transfusion and a significantly longer length of stay (LOS) (0.28 days longer) compared to a propensity-matched LDA cohort [14]. Song et al. also compared propensity-matched ALIF and LDA cohorts from the years 2011–2019 and found that ALIF was associated with significantly higher relative value units (RVUs) per minute than LDA, indicating potentially higher reimbursement potential for ALIF procedures. Further, they found the 30-day morbidity rate to be significantly higher in the ALIF cohort compared to the LDA cohort [15]. To the authors’ knowledge, no prior study has compiled the past 10 years of NSQIP-reported outcomes after ALIF versus LDA procedures.

The primary purpose of this retrospective cohort study was to compare 30-day complication incidence between propensity-matched ALIF and LDA cohorts from the years 2012–2021. The secondary objective of this study was to identify independent risk factors for perioperative complications after both ALIF and LDA.

2. Methods

This study is exempt from the informed consent requirement and institutional review board review due to its utilization of de-identified patient information.

This is a retrospective cohort study of prospectively collected data as part of the ACS NSQIP database which contains demographic information and baseline comorbidities with corresponding readmission and complication rates within 30 days of the indexed procedure. Current Procedural Terminology (CPT) and International Classification of Diseases (ICD), Ninth and Tenth Revision codes are used to identify patients [16]. The NSQIP outcome reporting system has been shown to be reliable in prior research [17]. The ACS NSQIP database is subjected to rigorous review from assigned surgical and clinical reviewers at each institution. To ensure validity, NSQIP employs quality assurance measures, including biweekly random internal audits, which have reported <1.8% inter-rater disagreement [18,19].

Patients who underwent ALIF (CPT 22558) or LDA (CPT 22857) as the primary operation from the years 2012 to 2021 were retrieved from the NSQIP database. Inclusion criteria were as follows: patients who underwent either isolated, primary, non-emergent, single-level ALIF or LDA between the years of 2012–2021. Exclusion criteria were as follows: patients with concomitant procedures such as laminectomies, laminotomies, cervical or thoracic procedures, or posterior instrumentation; procedures including two levels or greater, revision cases, emergent/urgent cases, or patients with missing demographic outcome data. Patient demographics, including age, smoking status, body mass index (BMI), gender, and American Society of Anesthesiologists (ASA) physical status classification score were collected along with complication data for each. Using 1:1 propensity score matching, both cohorts were made to be as similar as possible according to baseline characteristics such as gender, age, ASA status, BMI, hypertension requiring medication, congestive heart failure (CHF), diabetes mellitus (DM), chronic obstructive pulmonary disease (COPD), smoking status, inpatient/outpatient status, and bleeding disorders.

For each patient, LOS, readmission rate, death, and 30-day complications were collected. Queried complications included the following: both superficial and deep surgical site infections (SSI), wound dehiscence, pneumonia, unplanned intubation, pulmonary embolism, postoperative renal insufficiency or failure, urinary tract infection, cerebrovascular accident, cardiac arrest, myocardial infarction (MI), transfusions, deep vein thrombosis (DVT), sepsis, and return to the operating room. The variable “any adverse event” (AAE) was made to include any of the above-mentioned complications. LOS was defined as the number of days from procedure to postoperative discharge. Extended LOS was defined as greater than the 75% quartile (>4 days) LOS for the entire sample.

Propensity score matching, univariate, bivariate, and multivariable logistic regression analyses were performed using R-Studio software version 1.0.143 (R Foundation for Statistical Computing, Vienna, Austria). Propensity score matching was carried out using the nearest neighbour method in order to reduce treatment assignment bias and simulate randomization between the ALIF and LDA cohorts. Baseline demographics, comorbidities, and 30-day complications were compared using Student’s 2-tailed t-test and chi-squared analysis for continuous and categorical variables, respectively. Although no formal testing for normality was performed, these tests are relatively robust to minor deviations from normality, especially when sample sizes are large. Lastly, multivariable logistic regression with robust error variance identified risk factors for complications after ALIF and LDA. The variable(s) showing significance (p < 0.05) or a p-value of less than 0.2 from the bivariate comparisons were used in the multivariable logistic regression analysis [20]. Statistical significance was defined a priori as a p-value of less than 0.05.

Although the numerous comparisons in this study give the potential for a Type I error, we chose not to apply a Bonferroni correction as it can be too conservative, increasing the probability of a Type II error. Given that our dataset had less significant differences between cohorts compared to other similar studies, it was decided that forgoing a Bonferroni correction would maximize this study’s ability to detect statistical differences between groups. Additionally, this preserved comparability to similar studies that also chose not to perform a Bonferroni correction.

3. Results

A total of 2620 patients were identified, with 2011 (76.76%) receiving ALIF and 609 (23.24%) receiving LDA. After propensity score matching, there were 609 subjects in each cohort, with 1218 total subjects included in the matched analysis. Prior to matching, there were statistically significant differences between groups in sex, outpatient status, ASA class, length of stay, steroid use, and history of hypertension, diabetes, and COPD (

Table 1. Demographic and comorbidity characteristics for patients undergoing anterior lumbar interbody fusion (ALIF) or lumbar disc arthroplasty (LDA).

	ALIF Unmatched	LDA Unmatched	p-Value (Unmatched)	ALIF Matched	LDA Matched	p-Value (Matched)
Patients, N (%)	2011 (76.8%)	609 (23.2%)	NA	609 (50%)	609 (50%)	NA
Age (years)	56.4 ± 14.0	43.7 ± 13.1	<0.001 *	44.7 ± 12.6	43.7 ± 13.1	0.168
BMI (kg/m)	30.6 ± 6.1	29.2 ± 6.0	<0.001 *	29.6 ± 5.6	29.2 ± 6.2	0.312
Male sex	1047 (52.1%)	396 (65%)	<0.001 *	415 (68.1%)	396 (65%)	0.274
Outpatient status	122 (6.1%)	94 (15.4%)	<0.001 *	70 (11.5%)	94 (15.4%)	0.054
Operative time (minutes)	186.7 ± 117.2	118.1 ± 54.8	<0.001 *	180.8 ± 108.2	118.1 ± 54.2	<0.001 *
Average Length of Stay (days)	3.30 ± 3.88	1.92 ± 3.04	<0.001 *	2.94 ± 3.38	1.92 ± 3.04	<0.001 *
Morbidity Probability	0.045 ± 0.020	0.026 ± 0.012	<0.001 *	0.035 ± 0.015	0.026 ± 0.012	<0.001 *
Mortality Probability	0.002 ± 0.004	0.001 ± 0.002	<0.001 *	0.001 ± 0.002	0.001 ± 0.002	0.005 *
ASA class (mean ± SD)	2.4 ± 0.6	2.1 ± 0.6	<0.001 *	2.2 ± 0.6	2.1 ± 0.6	0.023 *
One (no disturbance)	83 (4.1%)	79 (13%)	NA	56 (9.2%)	79 (13%)	NA
Two (mild disturbance)	1082 (53.8%)	396 (65%)	NA	399 (65.5%)	396 (65%)	NA
Three (severe disturbance)	804 (40%)	132 (21.7%)	NA	148 (24.3%)	132 (21.7%)	NA
Four (life-threatening disturbance)	42 (2.1%)	2 (0.3%)	NA	6 (1%)	2 (0.3%)	NA
Five (moribund)	0 (0%)	0 (0%)	NA	0 (0%)	0 (0%)	NA
Race
White	1583 (78.7%)	386 (63.4%)	NA	445 (73.1%)	386 (63.4%)	NA
Black	176 (8.8%)	50 (8.2%)	NA	52 (8.5%)	50 (8.2%)	NA
Asian	38 (1.9%)	9 (1.5%)	NA	14 (2.3%)	9 (1.5%)	NA
Other	214 (10.6%)	164 (26.9%)	NA	98 (16.1%)	164 (26.9%)	NA
Dependent functional status (partial or total)	41 (2%)	6 (1%)	0.123	6 (1%)	6 (1%)	1.000
Current smoker	408 (20.3%)	131 (21.5%)	0.551	138 (22.7%)	131 (21.5%)	0.679
Comorbidities N (%)
Congestive heart failure	8 (0.4%)	2 (0.3%)	1.000	2 (0.3%)	2 (0.3%)	1.000
Dialysis	7 (0.3%)	0 (0%)	0.313	3 (0.5%)	0 (0%)	0.248
Steroid use	64 (3.2%)	4 (0.7%)	0.001 *	12 (2%)	4 (0.7%)	0.078
Hypertension	956 (47.5%)	164 (26.9%)	<0.001 *	171 (28.1%)	164 (26.9%)	0.700
Bleeding disorder	23 (1.1%)	0 (0%)	0.016 *	1 (0.2%)	0 (0%)	1.000
Ascites	2 (0.1%)	0 (0%)	1.000	0 (0%)	0 (0%)	NA
Pre-operative transfusion	5 (0.2%)	0 (0%)	0.483	3 (0.5%)	0 (0%)	0.248
Diabetes	319 (15.9%)	56 (9.2%)	<0.001 *	61 (10%)	56 (9.2%)	0.697
IDDM	104 (5.2%)	25 (4.1%)	NA	32 (5.3%)	25 (4.1%)	NA
NIDDM	215 (10.7%)	31 (5.1%)	NA	29 (4.8%)	31 (5.1%)	NA
COPD	90 (4.5%)	9 (1.5%)	0.001 *	8 (1.3%)	9 (1.5%)	1.000

BMI: body mass index; COPD: chronic obstructive pulmonary disease; DOE: dyspnea on exertion; Renal failure: wherein renal function has been compromised within 24 h prior to surgery; Dialysis: acute or chronic renal failure requiring dialysis within 2 weeks of indexed procedure; Weight loss: considered as greater than 10% decrease in body weight in the six-month interval preceding surgery; IDDM: insulin-dependent diabetes mellitus; ASA: American Society of Anesthesiology score. * p-value is <0.05.

). After propensity matching, only ASA class (p = 0.023), length of stay (p < 0.001), and operative time (p < 0.001) remained different between groups (Table 1).

After matching, the ALIF cohort had a mean age of 44.7 ± 12.6 years, of which 68.1% (n = 415) were male, and the mean BMI was 29.6 ± 5.6 kg/m². The respective values for the LDA cohort were 43.7 ± 13.1 years, 65% (n = 396) male, and 29.2 ± 6.2 kg/m². Following matching, the baseline demographics, aside from ASA class, were not statistically different. The top 10 most frequently listed preoperative diagnoses for each procedure are outlined in

Table 2. Top 10 most common ICD-9 and ICD-10 diagnosis codes in the ALIF cohort.

	Diagnosis	Percentage of Total Cohort
1	M43.16 (spondylolisthesis, lumbar region)	8.83%
2	M51.16 (intervertebral disc disorders with radiculopathy, lumbar region)	6.55%
3	722.52 (degeneration of lumbar or lumbosacral intervertebral disc)	6.42%
4	M48.061 (spinal stenosis, lumbar region without neurogenic claudication)	5.58%
5	M48.062 (spinal stenosis, lumbar region with neurogenic claudication)	5.45%
6	M48.06 (spinal stenosis, lumbar region)	5.24%
7	M51.36 (other intervertebral disc degeneration, lumbar region)	4.86%
8	M51.17 (intervertebral disc disorders with radiculopathy in the lumbosacral region)	3.93%
9	M51.37 (other intervertebral disc degeneration, lumbosacral region)	3.25%
10	M43.17 (lumbar spondylolisthesis)	3.08%

and

Table 3. Top 10 most common ICD-9 and ICD-10 diagnosis codes in the LDA cohort.

	Diagnosis	Percentage of Total Cohort
1	722.52 (degeneration of lumbar or lumbosacral intervertebral disc)	12.44
2	M51.26 (other intervertebral disc displacements in the lumbar region)	12.29
3	722.10 (displacement of a lumbar intervertebral disc without myelopathy)	11.83
4	M51.37 (other intervertebral disc degeneration in the lumbosacral region)	7.37
5	M51.16 (intervertebral disc disorder with radiculopathy in the lumbar region)	6.76
6	M51.06 (intervertebral disc disorder with myelopathy, lumbar region)	6.45
7	M51.36 (other intervertebral disc degeneration in the lumbar region)	5.68
8	M51.17 (intervertebral disc disorder with radiculopathy in the lumbosacral region)	5.53
9	M51.27 (Other intervertebral disc displacements, lumbosacral region)	5.07
10	M54.16 (radiculopathy, lumbar region)	2.92

.

The frequency of extended LOS was significantly higher in the ALIF group than in the LDA group (14.6% (n = 89) vs. 4.76% (n = 29); p < 0.001) (

Table 4. Incidence of adverse events for patients undergoing ALIF or LDA in matched cohorts.

	ALIF		LDA
	No.	Rate (%)	No.	Rate (%)	p-Value
Overall	609		609
Any adverse event	35	5.7	22	3.61	0.104
Renal complication	1	0.2	0	0	1.000
SSI	10	1.6	9	1.48	1.000
Intubation issues	3	0.5	1	0.16	0.616
Sepsis	3	0.5	1	0.16	0.616
Transfusion	17	2.8	7	1.15	0.064
DVT	3	0.5	3	0.49	1.000
Pneumonia	2	0.3	2	0.33	1.000
UTI	3	0.5	3	0.49	1.000
MI	2	0.3	0	0	0.16
Return to the OR	8	1.3	7	1.15	1.000
Cardiac Arrest	1	0.2	0	0	1.000
Stroke	0	0	0	0	NA
Pulmonary embolism	1	0.2	2	0.33	1.000
Death	1	0.2	0	0	1.000
Extended LOS	89	14.6	29	4.76	<0.001 *
Wound dehiscence	1	0.2	1	0.16	1.000
Readmission	14	2.3	12	1.97	0.843

Any adverse event (AAE): superficial and deep surgical site infection, organ space infection, renal failure or insufficiency, intubation (fail to wean or reintubation), post-operative transfusion, pneumonia, DVT, PE, UTI, stroke, cardiac arrest, MI, return to the OR; DVT: deep vein thrombosis; UTI: urinary tract infection; SSI: surgical site infection; LOS, length of stay (extended: greater than the 75% quartile [>4 d]); OR: operating room; Intubation issues: re-intubation or failure to wean from intubation; Renal complication: progressive renal insufficiency or renal failure; Inf: infinity. * p-value is <0.05.

). No other complication assessed reached statistical significance in terms of differences between cohorts. However, the incidence of AAE (p = 0.104), MI (p = 0.16), and postoperative transfusion (p = 0.064) met the threshold (p < 0.2) for inclusion in the multivariate analysis as defined previously.

When accounting for other variables, the LDA cohort had lower odds of extended LOS (odds ratio [OR] 0.457, 95% confidence interval [CI] 0.283–0.738, p = 0.001) than the ALIF cohort (

Table 5. Odds of specific post-surgery complication development as related to patient demographics and comorbidities when comparing LDA to ALIF.

			Statistical Comparisons
	Adverse Events		Bivariate	Multivariable **
	Number	Percent (%)	p-Value	OR Coef.	95% CI	p-Value
Overall	1218
Any adverse event	57	1.97	0.104	0.973	0.525–1.802	0.931
Transfusion	24	1.19	0.064	1.302	0.436–3.889	0.637
MI	2	<0.01	0.16	<0.001	0.000-Inf	0.997
Extended LOS	118	9.69	<0.001	0.457	0.283–0.738	0.001 *

* p-value is <0.05. ** Variables are adjusted for all baseline characteristics OR: Odds Ratio; Coef: coefficient; 95% CI: 95% Confidence Interval; Any adverse event (AAE): superficial and deep surgical site infection, organ space infection, renal failure or insufficiency, intubation (fail to wean or reintubation), post-op transfusion, pneumonia, deep vein thrombosis, pulmonary embolism, urinary tract infection, stroke, cardiac arrest, myocardial infarction, return to the operating room; MI: myocardial infarction; Inf: infinity.

). None of the other variables included in the multivariate analysis (AAE, transfusion, MI) demonstrated statistically different odds ratios as they relate to the surgery performed (LDA to ALIF).

In a second multivariate analysis, longer operative times (1 min intervals) were found to increase the odds of extended LOS in both ALIF (OR: 1.006; 95% CI: 1.004–1.008; p < 0.001) and LDA (OR: 1.007; 95% CI: 1.002–1.012; p = 0.011) (

Table 6. Multivariate analysis of risk factors for extended length of stay (LOS) after ALIF or LDA as related to patient demographics and comorbidities.

	ALIF				LDA
	OR Coef.	2.5% OR	97.5% OR	p-Value	OR Coef.	2.5% OR	97.5% OR	p-Value
Male sex	0.736	0.432	1.253	0.258	0.634	0.280	1.432	0.314
Age	1.022	0.999	1.045	0.258	1.031	0.995	1.067	0.072
BMI	0.946	0.902	0.993	0.023 *	0.986	0.919	1.057	0.777
Operative time (1 min. intervals)	1.006	1.004	1.008	<0.001 *	1.007	1.002	1.012	0.011 *
ASA Class	1.480	0.929	2.354	0.098	1.872	0.868	4.039	0.070
Dependent functional status (total or partial)	1.689	0.203	14.066	0.628	3.965	0.354	44.371	0.216
Current smoker	0.856	0.461	1.588	0.622	0.765	0.275	2.126	0.564
Congestive heart failure	5.392	0.315	92.248	0.245	<0.001	0.000	Inf	0.995
Dialysis	4.441	0.286	68.881	0.286	n/a	n/a	n/a	n/a
Steroid use	1.383	0.273	6.995	0.695	<0.001	0.000	Inf	0.994
Hypertension	1.542	0.857	2.775	0.149	0.899	0.337	2.394	0.950
Bleeding disorder	<0.001	0.000	Inf	0.986	n/a	n/a	n/a	n/a
Pre-operative transfusion	8.016	0.527	121.93	0.134	n/a	n/a	n/a	n/a
Diabetes	1.383	0.660	2.898	0.390	1.699	0.551	5.234	0.356
COPD	2.497	1.000	1.013	0.304	<0.001	0.000	Inf	0.991

BMI: body mass index; COPD: chronic obstructive pulmonary disease; Dialysis: acute or chronic renal failure requiring dialysis within 2 weeks of indexed procedure; ASA: American Society of Anesthesiology score (numeric). * p-value is <0.05.

). Lastly, lower BMI increased the odds of extended LOS in ALIF (OR: 0.946; 95% CI: 0.902–0.993; p = 0.023), but not in the LDA cohort (p = 0.777).

4. Discussion

Prior studies have investigated short-term postoperative complications after ALIF and LDA [14,15]. To the authors’ knowledge, this is the largest study comparing the two surgeries, using NSQIP data spanning the most recent 10 years. This study demonstrates a significantly higher incidence of extended LOS in propensity-matched subjects who received ALIF compared to LDA (p < 0.001). There was no statistically significant difference between cohorts in any other outcome. The multivariate analysis revealed that patients who received ALIF had higher odds of experiencing an extended LOS compared to patients who received LDA while showing no difference in odds of developing MI, AAE, or requiring more transfusions. A second multivariate analysis revealed that operative time was predictive of extended LOS in both cohorts while low BMI was predictive of extended LOS in the ALIF cohort.

Our findings regarding differences in LOS are in line with prior findings by Shultz et al. [14]. They demonstrated a significantly longer LOS in an ALIF cohort compared to an ALIF cohort after propensity matching (+0.28 days, p< 0.001) [14]. In the present study, the ALIF cohort stayed 1.02 days longer (p < 0.001), on average, than the LDA cohort. Additionally, the LDA cohort demonstrated a significantly lower incidence of extended LOS (>4 days in the hospital) compared to the ALIF cohort (14.6% vs. 4.76%, p < 0.001). In a cohort of subjects who received single-level ALIF in a single-surgeon database, Patel et al. demonstrated worse Oswestry Disability Index scores at 6 weeks postoperatively in patients who had a longer LOS (≥45 h) compared to patients with a shorter LOS (<45 h) [21]. Given these trends in LOS, it comes as no surprise that prior studies have demonstrated that ALIF is associated with increased hospital costs [22,23]. Further, our study found that operative time was significantly different between cohorts after matching (ALIF [180.8 ± 108.2] vs. LDA [118.1 ± 54.8], p < 0.001). Using the NSQIP database, Ansari et al. found that operative times longer than 120 min were associated with a LOS of ≥3 days in a cohort of subjects who received LDA. Further, they found that LOS ≥3 days were more likely to have venous thromboembolisms, pneumonia, surgical site infections, and reoperations [24].

Song et al. demonstrated higher rates of 30-day morbidity, DVT, and blood transfusions in an ALIF cohort compared to an LDA cohort after propensity matching using the NSQIP dataset from the years 2011–2019 [15]. Our study failed to demonstrate any significant differences in these particular outcomes between cohorts. This may be due to the difference in included years (2011–2019 versus 2012–2021). Although the utilization of lumbar disc arthroplasty decreased for a time (2010–2017), its popularity was shown to increase in the years 2018–2019 [25]. As the LDA becomes more and more commonly used, it is possible that patient selection is improving, and more patients who were once recommended ALIF procedure may now be recommended LDA instead. Changes in LDA versus ALIF utilization since 2019 could account for the difference between Song et al.’s findings and the present study.

An interesting finding in our study is that BMI was inversely associated with extended LOS in the ALIF cohort. This may be due to our decision to control for multiple comorbidities, many of which are linked to BMI (COPD, CHF, smoking status, age, diabetes). Additionally, high BMI patients may be less likely to be selected for ALIF due to complicated surgical approaches in these patients. Increased BMI has been found to be associated with increased hospital costs and surgical complications for patients who receive ALIF [23]. Mortazavi et al. demonstrated that BMI was associated with a longer length of stay [26]. Findings by Kuo et al. demonstrate that age, preoperative benzodiazepine use, higher intraoperative blood loss, delayed mobilization, and lower 12-item Short Form mental component score were correlated with increased LOS (≥3 days) [27]. Obesity classes II and III have been shown to be associated with postoperative wound infections after ALIF [28].

This study is not without limitations. The NSQIP data records only report certain patient characteristics and complications. It does not include patient-reported outcome scores, functional outcomes, or radiographic analyses. Therefore, our study is unable to analyze these outcomes that are relevant to the practice of orthopedic surgeons. Secondly, the NSQIP dataset does not describe the degree of pathology, level of spinal disease, or severity of radiographic abnormalities at baseline in these patients. Differences in preoperative characteristics likely influence whether a surgeon chooses to perform ALIF vs. LDA, and therefore, may introduce bias to this study. Third, LDA is performed less frequently than ALIF. Hence, our matched cohort size is limited to the number of LDA subjects available. For this reason, the investigators chose to include data from 10 years to provide a large, yet feasible sample size for the analyses performed in this study. Fourth, although propensity score matching is practical and effective, it is imperfect, as demonstrated by some differences between the matched cohorts in this study. Lastly, the database is limited to 30 days postoperatively. Many of the most important complications, however, that lead to readmission and significantly increased health care costs to the system occur early.

5. Conclusions

Based on NSQIP data from a span of 10 years (2012–2021), patients who receive ALIF are more likely to have an extended LOS compared to patients who receive LDA. Further, multivariate analysis demonstrated that operative time increases the odds of extended LOS (>4 days) in both ALIF and LDA cohorts while BMI is inversely associated with increased LOS in the ALIF cohort only. Thirty-day complications are otherwise similar between ALIF and LDA cohorts after controlling for baseline characteristics such as age, gender, BMI, ASA status, diabetes mellitus (DM), hypertension requiring medication, congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), inpatient/outpatient status, smoking status, and bleeding disorders. Further studies are needed to compare outcome measures more specific to orthopedic surgery practice, such as patient-reported outcome measures and radiographic outcomes.