Nationwide Retrospective Analysis of Early Complications After Adult Spinal Deformity Surgery

Abstract

Study Design.

Retrospective nationwide registry-based observational study.

Objective.

To determine the incidence of and factors associated with early complications following surgery for adult spinal deformity (ASD).

Summary of Background Data.

Complications after ASD surgery contribute to morbidity, mortality, and health care costs. Prior studies were limited by small sample sizes and incomplete surgical information. The Japanese Spinal Instrumentation Society database (JSIS-DB) includes detailed instrumentation data, enabling a robust secondary analysis of perioperative factors.

Materials and Methods.

We analyzed JSIS-DB data from patients undergoing ASD surgery involving the fusion of three or more vertebrae between 2018 and 2022. Demographics, comorbidities, and surgical variables [fusion length, osteotomy, interbody fusion type, minimally invasive surgery (MIS), staged procedures, navigation, and neuromonitoring] were assessed. The primary endpoint was ultra-early perioperative complications within two weeks, classified as major (mortality, massive bleeding, neurological, cardiopulmonary, or infectious events) or minor (dural injury, delirium, or urinary tract infection). Multivariable logistic regression was used to identify the associated factors.

Results.

Among 3193 patients (mean age 60.8 yr), the overall complication rate was 28.5%, including 25.4% with major complications and 0.1% with mortality. Increased risk was associated with operative time >5 hours [adjusted odds ratio (aOR) 4.40; 95% CI: 3.60–5.47], fusion >5 levels (aOR: 1.51; 95% CI: 1.20–1.91), posterior/transforaminal interbody fusion (PLIF/TLIF) (aOR: 1.47; 95% CI: 1.20–1.79), and three-column osteotomy (3CO) (aOR: 3.50; 95% CI: 2.21–5.53). Reduced risk was associated with adult idiopathic scoliosis (aOR: 0.48; 95% CI: 0.37–0.63), staged surgery (aOR: 0.24; 95% CI: 0.18–0.32), lateral lumbar interbody fusion (LLIF) (aOR: 0.74; 95% CI: 0.58–0.95), and MIS (aOR: 0.65; 95% CI: 0.49–0.86). Respiratory complications, massive bleeding, neurological deficits, surgical site infections, and proximal junctional kyphosis were associated with prolonged hospitalization (>30 d).

Conclusions.

This nationwide registry analysis demonstrated that prolonged operative time, long fusion, PLIF/TLIF, and 3CO increase the early complication risk after ASD surgery, whereas staged surgery, LLIF, and MIS are protective.

Level of Evidence.

Level III—retrospective cohort study.

As adult spinal deformity (ASD) surgery has increased in tandem with population aging, the burden of perioperative complications has also risen.^1–3 ASD is associated with serious complications and high treatment costs.^4–7 Therefore, it is essential to comprehensively evaluate the factors that may adversely affect the clinical course of patients undergoing ASD surgery.

Multicenter analyses of open, hybrid, and minimally invasive surgery (MIS) techniques have shown that complications reflect the interplay of patient-related factors [e.g., age, comorbidities, and body mass index (BMI)] and procedure-specific factors (e.g., number of fused levels, osteotomy, and operative time).^2,3,8–10 Several reports based on analyses of large spinal surgery databases have shown that most complications occur during or within several weeks after surgery, including systemic events such as thromboembolism and cardiopulmonary disorders.^11–13 However, despite these valuable insights, previous studies have been limited by small sample sizes, heterogeneous inclusion criteria, and incomplete surgical information.

Against this background, the Database Committee of the Japanese Spinal Instrumentation Society (JSIS) established a nationwide registry to systematically collect perioperative information on spinal instrumentation surgeries across Japan.¹⁴ The JSIS database (JSIS-DB) registry records detailed intraoperative and device-related variables—including surgical approach, fused segments, fusion to pelvis, three-column spinal osteotomy (3CO) (i.e., pedicle subtraction osteotomy or vertebral column osteotomy), interbody fusion type, MIS options, assistive navigation, and intraoperative neuromonitoring (IONM)—together with early postoperative events.¹⁴

The present study involved a retrospective secondary analysis of data from the existing JSIS-DB registry. This study aimed to clarify the incidence and factors associated with early complications within two weeks after ASD surgery. By analyzing a large national data set rich in instrumentation details, this study offers insights that may help inform surgical planning, perioperative management, and health care resource allocation.

MATERIALS AND METHODS

Study Design and Data Collection

This study retrospectively analyzed data from the JSIS-DB registry of patients who underwent spinal instrumentation surgery. Data on patient demographics, surgical procedures, and postoperative outcomes were prospectively recorded on standardized data collection sheets from February 2018 to March 2022 at multiple participating institutions across Japan. Data were collected by data managers or surgeons who completed a unified e-learning training program and entered the data postoperatively.¹⁴

This study retrospectively examined patients registered with the International Statistical Classification of Diseases and Related Health Problems (ICD-10) codes M40 (kyphosis and lordosis) and M41 (scoliosis). To focus on cases undergoing spinal deformity surgery, we also included cases registered under the K codes in the Japanese Medical Fee Schedule for scoliosis surgery, spinal fusion (anterior, posterior, or simultaneous anterior-posterior), and spinal osteotomy with fusion of three or more vertebral segments. In accordance with the JSIS-DB exclusion criteria, surgeries that were not related to spinal deformities (e.g., biopsy, kyphoplasty, laminoplasty, and internal fixation for pelvic fractures) were excluded. In addition, 2.1% of the cases registered during the above period with data deficiencies were excluded.

Independent Variables

Independent variables included patient and surgical characteristics. Patient characteristics included age, sex, BMI, diagnosis at the time of surgery in accordance with the ICD-10 codes, neurological status, American Society of Anesthesiologists (ASA) classification, and Osteoporosis Self-assessment Tool for Asians (OSTA).¹⁵ Surgical characteristics including prior spine surgery, number of fused levels, operative time, estimated blood loss, surgical approach (e.g., anterior, posterior, anterior, and posterior), 3CO, type of interbody fusion [i.e., posterior interbody fusion (PLIF), transforaminal interbody fusion (TLIF), or lateral lumbar interbody fusion (LLIF)], MIS options including percutaneous screws, and use of a spinal surgery navigation system and IONM. This information was collected through a mandatory questionnaire survey.

Outcome Variables

The primary outcome variable was complications occurring within two weeks of ASD surgery. In the JSIS-DB registry, complications are categorized as either major or minor. Major complications included mortality, major vascular injury, respiratory complications, cardiovascular complications, cerebral infarction or cerebrovascular complications, massive intraoperative bleeding, surgical site infection, and neurological complications. Minor complications include dural injury, postoperative epidural hematoma, postoperative delirium, urinary tract infection, and transient nerve paralysis. We also investigated the length of the postoperative hospital stay as a secondary outcome variable related to postoperative complications.

Statistical Analysis

Univariate and multivariate analyses were conducted to investigate the effect of patient background characteristics, comorbidities, surgical conditions, and other clinical variables on complications. The Fisher exact probability test was used for the univariate analysis. The significance level was set at P<0.05. Factors with P<0.2 in univariate analysis were included in the multivariate analysis. To ascertain the occurrence of major, minor, and other significant complications, a multivariate logistic regression model was fitted based on the likelihood ratio test and stepwise variable addition. In the final multivariate model, the adjusted odds ratios (aOR) and 95% CIs were reported. The Hosmer-Lemeshow test was used to evaluate the fit of the model. In addition, multivariate analysis was performed to identify complications associated with postoperative hospital stay. Similar multivariate analyses were performed on the group of cases involving fusion of six or more segments. Furthermore, trends in complication incidence rates during the study period were compared. For post hoc analysis, two data sets were created: one containing patients aged 50 years or younger and the other containing those aged 51 years or older. Each data set was analyzed separately. All statistical analyses were performed using the SPSS software (version 30.0.0.0; SPSS; IBM Corp., Armonk, NY). The overall analytical framework and the rationale for focusing on ultra-early perioperative complications are provided in Supplementary Digital Content 1 and 2, http://links.lww.com/BRS/C1000.

RESULTS

Patient Background and Surgical Overview

Between 2018 and 2022, 3193 patients who underwent ASD surgery were identified in the JSIS-DB registry. years (81.1% female; mean age 60.8±20.7 yr). Among them, 69% were aged 61 years or older, and the data set exhibited a bimodal age distribution with two peaks: one in the early 20s and one in the 70s (Table 1, Figure 1). The prevalence of some form of neurological disorder was 34.1%, and 3.5% of the patients had Parkinson disease. The breakdown of ICD-10 codes is shown in Table 1, with M40 kyphosis accounting for 33.7% and M41.2 adult idiopathic scoliosis accounting for 19.5%. The mean number of fused segments was 8.6, with 75% of patients undergoing fusion of six or more segments, and fusion to the pelvis was performed in 54.6% of cases. The distribution of the number of fused spinal segments demonstrated a bimodal pattern, indicating that the cohort included a substantial proportion of long-segment fusion procedures (Figure 2). More patients underwent surgery through the posterior approach (83.3%) than through the anterior approach (6.0%) or the simultaneous anterior-posterior approach (10.6%). Staged surgery was performed in 23.5% of patients, while 3CO was performed in 3.8%. Interbody procedures involved LLIF cages (34.8%) and PLIF cages (32.0%). MIS options, including percutaneous techniques, were used in 20.8% of cases. The assistive technologies included spinal surgical navigation (30.7%) and IONM (77.9%). The mean length of postoperative hospital stay was 31.5 days.

TABLE 1.

Baseline Demographics, Surgical Information, and Postoperative Information

Age in years, n (%)	60.8±20.7
<51	22.9
51–60	8.1
61–70	25.0
71–80	36.4
>80	7.6
Female sex, n (%)	2590 (81.1)
Etiology, n (%)
M40 kyphosis	1077 (33.7)
M41.2 adult idiopathic scoliosis	624 (19.5)
Adult scoliosis other than M41.2 (M41.3, 4, 5, 8, 9)	1492 (46.7)
BMI, kg/m2	22.5±4.0
OSTA	−1.58±4.5
Previous spine surgery, n (%)	538 (16.8)
Preoperative neurological deficit, n (%)	1090 (34.1)
Operative time in minutes	350±163.8
Total estimated blood loss in mL	649.3±810.2
Surgical experience JSSR-certified instructor, n (%)	2910 (91.1)
No. fused levels	8.6±3.5
Anterior intervertebral device (LLIF), n (%)	1113 (34.9)
Posterior intervertebral device (PLIF, TLIF), n (%)	1023 (32.0)
3CO, n (%)	121 (3.8)
MIS, n (%)	663 (20.8)
Assistive technologies
Surgical navigation, n (%)	979 (30.7)
Neuromonitoring, n (%)	2487 (77.9)
ASA grade, n (%)
I	971 (30.4)
II	1937 (60.7)
III	257 (8.0)
IV	12 (0.4)
Hospital stay in days (from the date of surgery)	31.5±30.7

Data are presented as mean±SD, n, or n (%).

3CO indicates three-column spinal osteotomy; ASA, American Society of Anesthesiologists; BMI, body mass index; IONM, intraoperative neuromonitoring; JSSR, Japanese Society for Spine Surgery and Related Research; LLIF, lateral lumbar interbody fusion; MIS, minimally invasive surgery; OSTA, Osteoporosis Self-assessment Tool for Asians; PLIF, posterior interbody fusion; TLIF, transforaminal lumbar interbody fusion.

Figure 1.

Age at surgery by International Statistical Classification of Diseases and Related Health Problems (ICD-10) code. The age distribution peaked in the 20s and 70s. Adult idiopathic scoliosis was more common in younger patients (≤50 yr), while other scoliosis and kyphosis were more common in older patients (>50 yr).

Figure 2.

Distribution of the number of fused spinal segments. The distribution of fused spinal segments in the study cohort demonstrated a bimodal pattern, with prominent peaks at three and 10 segments. This distribution reflects the inclusion of both short-segment and long-segment fusion procedures within this nationwide registry of adult spinal deformity surgery.

Complications and Associated Factors

All complications occurred in 910 cases (28.5%), major complications occurred in 812 cases (25.4%), and death occurred in three cases (Table 2). A trend toward increased surgical procedures was observed from 2018 to 2022, but there was no change in complication frequency. Registration for 2022 was limited to a three-month period from January to March. The breakdown of the major complications is shown in Table 3. The most frequent major complications were massive intraoperative bleeding (>1000 mL) (20.0%), neurological complications (nerve root injury, cauda equina injury, spinal cord injury) (3.1%), and respiratory complications (1.8%). Minor complications were observed in 186 patients (5.8%), with the most frequent being dural tear (2.4%) and postoperative delirium (1.3%).

TABLE 2.

Frequency of Complications in ASD Surgery (3193 Cases From 2018 to 2022)

Complication	Number (Frequency (%))
Total patients with ≥1 complication	910 (28.5)
Major complications	812 (25.4)
Mortality	3 (0.1)
Major vascular injury	3 (0.1)
Intestinal injury	0
Ureteral injury	0
Respiratory complications	57 (1.8)
Pneumonia	18 (0.6)
Pneumothorax, hemothorax, chylothorax	12 (0.4)
Pulmonary embolism	26 (0.8)
Airway obstruction, dysphagia	3 (0.1)
Gastrointestinal complications	24 (0.8)
Cardiovascular complications	14 (0.4)
Cerebral infarction or cerebrovascular complications	3 (0.1)
Extensive blood loss (>1000 mL)	638 (20.0)
Wrong-level surgery	2 (0.1)
Surgical site infection	30 (0.9)
Implant fracture or dislocation	17 (0.5)
Proximal junctional kyphosis (including vertebral fracture)	14 (0.4)
Neurological complications (nerve root injury, cauda equina injury, and spinal cord injury)	98 (3.1)
Minor complications	186 (5.8)
Dural injury	78 (2.4)
Postoperative hematoma	21 (0.6)
Implant displacement or misplacement	18 (0.6)
Postoperative delirium	40 (1.3)
Urinary tract infection	5 (0.2)
Skin complications	5 (0.2)
Transient nerve palsy	16 (0.5)
Others	3 (0.5)

TABLE 3.

Univariate Analysis of Complications for ASD Surgery

Characteristic	Total (n=3193), n (%)	No complication (n=2283), n (%)	P	Minor complication (n=186), n (%)	P	Major complication (n=812), n (%)	P
Demographics
Age in years
<51	22.9	25.8	<.001	8.1	<.001	16.6	<.001
51–60	8.1	7.9	0.391	5.9	0.331	9.0	0.299
61–70	25.0	24.3	0.160	29.6	0.139	26.6	0.222
71–80	36.4	34.2	<.001	48.9	<.001	40.9	0.002
>80	7.6	7.8	0.555	7.5	0.999	6.9	0.4
Female sex	81.1	82.2	0.016	80.1	0.7	78.3	0.02
BMI, kg/m2
No obesity (BMI <30 kg/m2)	95.2	95.8		93.5		93.7
Obesity class I (BMI ≥30 kg/m2)	4.8	4.2	0.017	6.5	0.283	6.3	0.028
OSTA <−4	37.3	36.2	0.047	42.3	0.024	38.9	0.275
ASA grade 3 or 4	8.4	8.2	0.324	8.1	0.999	9.2	0.38
Etiology
M40 kyphosis	33.7	30.6	<.001	47.8	<.001	41.0	<.001
M41.2 adult idiopathic scoliosis	19.5	22.6	<.001	5.9	<.001	12.7	<.001
Adult scoliosis other than M41.2 (M41.3, 4, 5, 8, 9)	46.7	46.9	0.814	44.6	0.348	46.3	0.807
Previous spine surgery	16.8	17.1	0.530	16.2	0.729	16.9	1
Preoperative neurological deficit	34.1	33.6	0.363	45.2	0.001	34.4	0.898
Parkinson disease	3.5	3.0	0.032	5.4	0.150	4.6	0.059
Operative variables
Operative time >5 h	58.3	50.2	<.001	73.7	<.001	81.8	<.001
Procedure length
Long (>5 levels)	75.1	71.7	<.001	75.8	0.862	86.5	<.001
Short (≤5 levels)	24.9	28.3		24.2		13.5
Staged surgery	23.5	28.6	<.001	16.7	0.026	9.4	<.001
Surgical approach
Anterior	6.0	7.4	<.001	2.7	0.055	2.5	<.001
Posterior	83.3	83.1	0.713	84.4	0.761	84.1	0.513
Combined anterior-posterior	10.6	13.5	<.001	12.9	0.325	13.4	0.004
Procedure subtypes
Anterior intervertebral device (LLIF)	34.8	39.8	<.001	24.7	0.003	21.6	<.001
Posterior intervertebral device (PLIF, TLIF)	32.0	24.0	<.001	48.9	<.001	53.6	<.001
3CO	3.8	1.4	<.001	8.6	<.001	10.6	<.001
Fusion to pelvis	54.6	51.6	<.001	67.1	<.001	69.1	<.001
MIS	20.8	25.2	<.001	11.8	<.001	8.9	<.001
Assistive technologies
Surgical navigation	30.7	30.9	0.64	26.9	0.287	30.0	0.692
IONM	77.9	77.0	0.033	78.5	0.927	80.9	0.019

Boldface indicates P<0.05.

3CO indicates three-column spinal osteotomy; ASA, American Society of Anesthesiologists; ASD, adult spinal deformity; BMI, body mass index; IONM, intraoperative neuromonitoring; LLIF, lateral lumbar interbody fusion; MIS, minimally invasive surgery; OSTA, Osteoporosis Self-assessment Tool for Asians; PLIF, posterior interbody fusion; TLIF, transforaminal lumbar interbody fusion.

Univariate analysis of all complications revealed associations with patient characteristics (age, sex, BMI, OSTA, kyphosis, and Parkinson disease) and surgical details (operative time, number of fused vertebrae, staged surgery, LLIF, PLIF or TLIF, 3CO, fusion to pelvis, and MIS). All complications were associated with a prolonged postoperative hospital stay (>30 d) (P<0.001). Similar findings were observed in univariate analyses of major and minor complications (Table 3). These factors were further examined using multivariate analyses (Table 4). Multiple logistic regression analysis identified the following prognostic factors. The factors significantly associated with an increased risk of complications were operative time >5 hours (aOR: 4.40, 95% CI: 3.6–5.47), long fusion (>5 levels) (aOR: 1.51, 95% CI: 1.20–1.91), PLIF or TLIF (aOR: 1.47, 95% CI: 1.20–1.79), and 3CO (aOR: 3.50, 95% CI: 2.21–5.53). The factors associated with reduced risk of complications were adult idiopathic scoliosis (aOR: 0.48, 95% CI: 0.37–0.63), staged surgery (aOR: 0.24, 95% CI: 0.18–0.32), LLIF (aOR: 0.74, 95% CI: 0.58–0.95), and MIS (aOR: 0.65, 95% CI: 0.49–0.86). The associated factors that increased the risk of major complications were operative time >5 hours (aOR: 4.69, 95% CI: 3.78–5.83), PLIF/TLIF (aOR: 1.573, 95% CI: 1.28–1.93), and 3CO (aOR: 3.70, 95% CI: 2.37–5.77), while the factors that decreased the risk of major complications were adult idiopathic scoliosis (aOR: 0.59, 95% CI: 0.45–0.77), staged surgery (aOR: 0.19, 95% CI: 0.14–0.25), and MIS (aOR: 0.61, 95% CI: 0.45–0.81). The only factor that increased the risk of minor complications was operative time >5 hours (aOR: 2.00, 95% CI: 1.41–2.83), while the factors that decreased the risk of minor complications were age <51 years (aOR: 0.24, 95% CI: 0.14–0.42), LLIF (aOR: 0.49, 95% CI: 0.33–0.71), and MIS (aOR: 0.59, 95% CI: 0.36–0.98). The model showed a good fit, with a Hosmer-Lemeshow goodness-of-fit test P-value of 0.697. Because of the small number of deaths, adjustment for mortality in the multivariate analysis was not statistically supported.

TABLE 4.

Independent Risk Factors for Complications Identified By Multivariate Logistic Regression

Spinal fusion of three or more vertebral segments (3193 cases)					Spinal fusion of six or more vertebral segments (2399 cases)
	All complications					All complications
Effect	Estimate	95% confidence limits		P	Effect	Estimate	95% confidence limits		P
M41.2 adult idiopathic scoliosis	0.483	0.371	0.629	<.001	M41.2 adult idiopathic scoliosis	0.549	0.394	0.764	<.001
Operative time >5 h	4.399	3.57	5.422	<.001	Operative time >5 h	5.191	4.033	6.683	<.001
Long fusion (>5 levels)	1.512	1.196	1.912	<.001	Staged surgery	0.212	0.151	0.299	<.001
Staged surgery	0.237	0.177	0.317	<.001	Intervertebral device; anterior (LLIF)	0.57	0.428	0.759	<.001
Intervertebral device; anterior (LLIF)	0.741	0.579	0.949	0.017	Intervertebral device; posterior (PLIF, TLIF)	1.399	1.073	1.823	0.013
Intervertebral device; posterior (PLIF, TLIF)	1.468	1.202	1.793	<.001	3CO	2.83	1.629	4.918	<.001
3CO	3.499	2.212	5.534	<.001	Fusion to pelvis	1.513	1.162	1.971	0.002
MIS	0.649	0.489	0.86	0.003
	Major complications					Major complications
M41.2 adult idiopathic scoliosis	0.591	0.452	0.772	<.001	M41.2 adult idiopathic scoliosis	0.637	0.453	0.897	0.01
Operative time >5 h	4.693	3.776	5.831	<.001	Operative time >5 h	5.124	3.939	6.666	<.001
Long fusion (>5 levels)	1.85	1.441	2.375	<.001	Staged surgery	0.192	0.133	0.277	<.001
Staged surgery	0.186	0.138	0.249	<.001	Intervertebral device; anterior (LLIF)	0.664	0.486	0.908	0.01
Intervertebral device; posterior (PLIF, TLIF)	1.573	1.284	1.927	<.001	Intervertebral device; posterior (PLIF, TLIF)	1.341	1.023	1.756	0.034
3CO	3.698	2.371	5.768	<.001	3CO	2.937	1.706	5.056	<.001
MIS	0.606	0.451	0.814	<.001	Fusion to pelvis	1.642	1.245	2.165	<.001
					MIS	0.677	0.466	0.983	0.04
	Minor complications					Minor complications
Age <51 yr	0.241	0.14	0.417	<.001	OSTA <−4	1.633	1.107	2.411	0.013
Operative time >5 h	2.001	1.414	2.833	<.001	M40 kyphosis	1.515	1.02	2.25	0.039
Intervertebral device; anterior (LLIF)	0.487	0.332	0.714	<.001	M41.2 adult idiopathic scoliosis	0.37	0.159	0.858	0.02
MIS	0.593	0.36	0.976	0.04	Preoperative neurological deficit	1.83	1.243	2.693	0.002
					Operative time >5 h	2.911	1.828	4.636	<.001
					Intervertebral device; anterior (LLIF)	0.358	0.231	0.556	<.001

3CO indicates three-column spinal osteotomy; LLIF, lateral lumbar interbody fusion; MIS, minimally invasive surgery; OSTA, Osteoporosis Self-assessment Tool for Asians; PLIF, posterior interbody fusion; TLIF, transforaminal lumbar interbody fusion.

Furthermore, similar multivariate analyses were performed on 2399 cases involving the fusion of six or more segments (Table 4). The factors significantly associated with an increased risk of complications were operative time >5 hours (aOR: 5.19, 95% CI: 4.03–6.68), PLIF or TLIF (aOR: 1.40, 95% CI: 1.07–1.82), and 3CO (aOR: 2.83, 95% CI: 1.63–4.92), and fusion to pelvis (aOR: 1.51, 95% CI: 1.16–1.97). The factors associated with reduced risk of complications were adult idiopathic scoliosis (aOR: 0.55, 95% CI: 0.39–0.76), staged surgery (aOR: 0.21, 95% CI: 0.15–0.30), and LLIF (aOR: 0.57, 95% CI: 0.43–0.76). The results were largely consistent with those from the analysis of 3194 cases involving the fusion of three or more segments.

Both major and minor complications were associated with prolonged hospitalization (>30 d). The complications associated with prolonged hospitalization were respiratory complications (aOR: 2.0, 95% CI: 1.16–3.47, P=0.013), intraoperative massive bleeding (>1000 mL) (aOR: 2.2, 95% CI: 1.87–2.69, P<0.001), neurological complications (aOR: 3.5, 95% CI: 2.25–5.33, P<0.001), surgical site infection (aOR: 18.9, 95% CI: 5.68–63.1, P<0.001), and proximal junctional kyphosis (PJK) (aOR: 7.69, 95% CI: 2.1–28.2, P=0.002).

Figure 3 shows the annual number of surgeries and overall complication rates from 2018 to 2022. The annual complication rates were 25.9%, 25.9%, 29.5%, 30.2%, and 29.4% from 2018 to 2022, respectively, showing slight fluctuations but no statistically significant differences (P=0.233). Table 5 shows the results of the post hoc analysis of the two data sets (Table 5). In the data set containing patients aged 50 years or younger, where adult idiopathic scoliosis predominated, male sex, ASA grade ≥3, and operative time >5 hours were significantly associated with an increased risk of all complications, while staged surgery significantly decreased the risk of all complications. In the data set containing patients aged 51 years and older, where kyphosis and other types of adult scoliosis predominated, operative time >5 hours, long fusion, PLIF or TLIF, and 3CO were significantly associated with an increased risk of all complications, whereas staged surgery, LLIF, and MIS significantly decreased the risk of all complications.

Figure 3.

Annual case registrations and overall complications. The numbers of cases were low in 2018 and 2022 because the registration periods were less than one year. The complication rate did not significantly fluctuate between years: 2018 (25.9%), 2019 (25.9%), 2020 (29.5%), 2021 (30.2%), and 2022 (29.4%).

TABLE 5.

Frequency of Complications and Independent Risk and Protective Factors Identified By Multiple Logistic Regression Analysis in Two Data Sets Defined By Age Distribution

Data set containing patients aged 50 yr and under (733 cases)					Data set containing patients aged 51 yr and over (2460 cases)
Complication	Number	Frequency				Number	Frequency
Total patients with ≥1 complication	143	19.5%				767	31.2%
Major complications	135	18.5%				677	27.5%
Mortality	1	0.1%				2	0.1%
Minor complications	15	2.0%				171	7.0%
	All complications					All complications
Effect	Estimate	95% confidence limits		P	Effect	Estimate	95% confidence limits		P
Male sex	3.07	1.923	4.902	<.001	Operative time >5 h	4.006	3.147	5.1	<.001
ASA grade ≥3	4.925	1.459	16.629	0.01	Long fusion (>5 levels)	1.508	1.168	1.948	0.002
Operative time >5 h	6.327	4.162	9.619	<.001	Staged surgery	0.247	0.183	0.333	<.001
Staged surgery	0.071	0.015	0.333	<.001	Intervertebral device; anterior (LLIF)	0.731	0.564	0.947	0.018
					Intervertebral device; posterior (PLIF, TLIF)	1.379	1.104	1.722	0.005
					3CO	3.398	2.138	5.4	<.001
					MIS	0.609	0.457	0.811	<.001
	Major complication				Major complication
Effect	Estimate	95% confidence limits		P	Effect	Estimate	95% confidence limits		P
Male sex	3.299	2.06	5.284	<.001	Operative time >5 h	4.438	3.446	5.715	<.001
ASA grade ≥3	5.429	1.599	18.43	0.007	Long fusion (>5 levels)	1.885	1.432	2.481	<.001
Operative time >5 h	5.919	3.854	9.089	<.001	Staged surgery	0.191	0.141	0.26	<.001
Staged surgery	0.067	0.013	0.333	<.001	Intervertebral device; posterior (PLIF, TLIF)	1.541	1.233	1.927	<.001
3CO	14.176	0.691	290.775	0.085	3CO	3.651	2.33	5.723	<.001
	Minor complication				Minor complication
Effect	Estimate	95% confidence limits		P	Effect	Estimate	95% confidence limits		P
BMI >30 kg/m2	4.553	1.019	20.346	0.047	Operative time >5 h	1.751	1.222	2.507	0.002
Operative time >5 h	4.468	1.19	16.776	0.027	Intervertebral device; anterior (LLIF)	0.497	0.338	0.732	<.001
Fusion to pelvis	3.72	1.097	12.62	0.035	MIS	0.559	0.337	0.929	0.025

3CO indicates three-column spinal osteotomy; ASA, American Society of Anesthesiologists; LLIF, lateral lumbar interbody fusion; MIS, minimally invasive surgery; PLIF, posterior interbody fusion; TLIF, transforaminal lumbar interbody fusion.

DISCUSSION

This is the first nationwide analysis of early complications after ASD surgery, based on a secondary analysis of a nationwide spinal instrumentation registry in Japan. Unlike voluntary surveys, the JSIS-DB registry systematically collects perioperative data from multiple institutions using standardized reporting forms, thereby minimizing selection and reporting biases. The present study retrospectively analyzed data originally collected for clinical documentation and quality monitoring. A unique strength of the JSIS-DB registry is its instrumentation-oriented structure that allows for the aggregation of detailed surgical information, including device usage, fusion levels, osteotomies, interbody fusion types, and assistive technologies such as surgical navigation and IONM. Such comprehensive, instrumentation-rich nationwide databases are rare internationally and provide valuable real-world evidence derived from secondary data utilization.^{11–13,16,17}

The overall complication rate in the present study was 28.5%, with no significant variation by registration year and no concerns regarding data registration deficiencies. However, this figure is lower than the complication rates of 30% to 70% reported in previous studies and systematic reviews.^{11,13,16–18} This discrepancy is thought to be caused by differences in the timing of complication occurrence. Many registry reports have focused on complications occurring within four to six weeks postoperatively, whereas the present study investigated complications occurring within two weeks postoperatively, which may have led to an underestimation of the incidence of complications. The two-week postoperative period examined in this study could be considered “ultra-early” compared with previous reports. Another factor influencing the interstudy differences in complication rates is the difference in patient populations. Klineberg et al ¹³ reported that intraoperative complications occurred in 30% and perioperative complications occurred in 48% of 558 patients with severe ASD. The present study classified participants using ICD-10 codes and did not assess the baseline severity of deformity, meaning that it likely included patients with relatively mild deformities. Consequently, perioperative complication rates may have been underestimated. The findings of a study based on the American College of Surgeons National Surgical Quality Improvement Program database reported an early complication rate (within 30 d after surgery) of 31%,¹⁷ which is closer to the rate obtained in the present study.

Comparisons with other Japanese studies have revealed similarities and differences. An analysis of 4822 scoliosis cases reported in the Japanese Scoliosis Society database from 2022 to 2023 showed similar usage rates for IONM (71%), surgical navigation (27%), and LLIF (32%), with an overall complication rate of 11.8%.¹⁹ The slightly higher complication rate in the present study was potentially influenced by differences in case composition and a shorter observation period (2 wks vs. 30 d). Conversely, a questionnaire survey by the Japanese Scoliosis Society reported a complication rate of 26%, similar to our findings.²⁰

The breakdown of complications in the present study, with intraoperative bleeding, respiratory complications, and postoperative delirium being the most common, is consistent with previous studies.^13,17,21 Furthermore, multiple logistic regression identified long fusion (>5 levels), operative time >5 hours, and 3CO as significant risk factors for all complications and severe complications, which is consistent with previous studies.^{2,3,8–10,19,22–26}

Of the 812 major complications in this study, 638 were cases of massive intraoperative bleeding (>1000 mL). While significant intraoperative blood loss is undoubtedly a major complication, there is no clear consensus on whether 1000 mL should be the cutoff value. The Clavien–Dindo classification for general surgery and perioperative databases does not categorize blood loss as a complication unless a transfusion occurs.²⁷ According to existing reports, transfusions are performed in 50% to 60% of ASD surgeries,^28,29 which significantly exceeds the 1000 mL of intraoperative blood loss recorded in 20% of cases in this study (Table 2). Although it is difficult to directly correlate transfusion volume with intraoperative blood loss, the definition of extensive blood loss adopted in this study may be considered a valid indicator of complications.

Incidentally, one challenge in collecting data for this registry is the lack of information on the severity of reported cases. Reports of “infection” do not include details on infection severity, additional procedures, or surgeries. Superficial and severe deep infections were both categorized as major complications, which could lead to overestimation. For “dural injuries,” it was unclear whether patients had symptoms or were asymptomatic. There is also concern that asymptomatic dural injuries were counted as minor complications, which could lead to an overestimation of the number of complications. Thus, complications that cannot be fully captured in a multicenter registry present a major challenge and limitation of this study.

In this study, we conducted a sensitivity analysis restricted to patients who underwent fusion at six or more segments to address the concern that fusion at three segments may not adequately represent adult spinal deformity surgery. Despite the stricter definition and smaller sample size (n=2399), the association between the identified risk factors and early complications remained consistent with that observed in the main cohort. These analyses support the robustness of the study.

A long operative time exceeding five hours is one factor related to complications, but it should not be interpreted as the sole cause of postoperative complications. Rather, it should be viewed as an indicator of surgical complexity and the severity of the underlying disease. In adult spinal deformity surgery, a prolonged operative time is usually associated with extensive fusion, osteotomy, and multilevel decompression—all of which are procedures that are prioritized for patients with more severe deformities or neurological impairments. These results underscore the need to avoid misinterpretation by nonspecialists or administrative personnel who might focus solely on operative time as the problem. As previous studies have noted, advanced age, male gender, and ASA ≥3 are significant risk factors for complications that cannot be ignored.

LLIF and staged surgery for ASD are less invasive than single-stage surgery and are expected to reduce perioperative complications.^30–33 However, some reports indicate that multiple surgeries are associated with increased anesthesia risks and venous thromboembolism.^34–36 The present study demonstrated that LLIF and staged surgery were not associated with pulmonary embolism and were protective factors against all complications. These findings reinforce the advantages of these surgical options for ASD. MIS using percutaneous pedicle screws has been reported to reduce blood loss and myocardial injury, thereby lowering the rate of postoperative complications.^37,38 The present study also found that MIS was a protective factor against all major and minor complications. As demonstrated in the post hoc analysis of the present study, reducing surgical invasiveness through LLIF and MIS may be crucial for avoiding complications, particularly in the population aged 51 years or older, which includes a high prevalence of kyphotic deformities.

The independent risk factors for prolonged hospital stay in the present study were respiratory complications, extensive intraoperative blood loss, surgical site infection, neurological complications, and PJK. Intraoperative and perioperative complications are known to prolong hospital stay,^13,18 and a Latin American multicenter study reported that intraoperative bleeding, surgical site infection, and PJK are associated with prolonged hospitalization and readmission within 30 days.³ While differences in national and regional health care insurance systems influence the length of hospital stay, it is crucial to identify early complications affecting the duration of hospitalization to benefit patients and reduce health care costs.^4–7

The present study had several limitations. First, the JSIS-DB registry did not capture radiographic parameters, such as the sagittal vertical axis or Cobb angle, nor did it record the postoperative correction status, which limits the investigation of correlations between deformity severity, surgical goals, and complications. For this study, a preoperative diagnosis was defined using the nationally standardized ICD-10 classification system. Therefore, the results should be interpreted as reflecting actual surgical practice rather than representing a homogeneous ASD population in terms of radiology. Second, the JSIS-DB registry did not include patient-reported outcomes, which prevented the assessment of the functional impact of perioperative complications. Third, the JSIS-DB registry lacks data on some complications (e.g., chronic obstructive pulmonary disorder, diabetes, hypertension, ischemic heart disease, heart failure, rheumatoid arthritis, smoking history, etc.) that are established risk factors for adverse outcomes.^2,3,8–10 Finally, the JSIS-DB registry focused on short-term complications (within 14 d), whereas data on complications occurring weeks to months postoperatively were not collected. Therefore, high-frequency and severe mechanical complications that occur after ASD surgery have not been sufficiently examined. While the findings of the present study generate hypotheses for managing the postoperative complications of ASD surgery, further research is needed before they can be adopted as best practices in clinical settings.

In conclusion, this nationwide registry analysis provides a comprehensive overview of early complications following ASD surgery in Japan. The overall complication rate was 28.5%. The factors significantly associated with an increased risk of complications were long fusion (five levels), operative time >5 hours, and 3CO, while staged surgery, LLIF, adult idiopathic scoliosis, and MIS demonstrated protective effects. These findings are valuable for surgical planning and perioperative patient counseling, potentially contributing to patient safety and cost reduction through complication mitigation.

Key Points

• We analyzed the nationwide JSIS-DB registry to identify the risk and protective factors of complications after adult spinal deformity surgery.

• The overall complication, major complication, and mortality rates were 28.5%, 25.4%, and 0.1%, respectively.

• Risk factors were operative time >5 hours, fusion >5 levels, posterior/transforaminal interbody fusion, and three-column osteotomy.

• The protective factors included adult idiopathic scoliosis, staged surgery, lateral lumbar interbody fusion, and minimally invasive surgery.

Supplementary Material

brs-51-e519-s001.pptx ^{(47.2KB, pptx)}