The design and reporting of this study were supported by the following guidelines: PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses; Multimedia Appendix 1) and PRISMA-P (Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols) [16,17]. In accordance with PRISMA-P guidelines, the protocol of the systematic review was registered on the PROSPERO (CRD42022339764) and disseminated through JMIR Research Protocols (PRR1-10.2196/4233), with the protocol being published before any data were collected [12]. There were no deviations from the protocol.

The Population, Intervention, Comparison, Outcome (PICO) framework was used to formulate inclusion criteria.

Studies that do not differ in surgical timing (staged vs same day), nonhuman or adolescent patient populations, reviews, conference abstracts, single-case studies, or technical notes were excluded from the analysis. Further, only studies originally published in English were considered.

Databases explored included PubMed, MEDLINE, Embase, Cochrane CENTRAL, Web of Science, and Scopus. A literature search was conducted in accordance with the PRISMA guidelines on August 2, 2023. We used a complex search string that was modified and fitted to the unique search functions of each queried database (Multimedia Appendix 2). Additional searching through gray literature and reference lists was conducted to identify studies not initially captured by the database query.

Studies and full text were screened, data were extracted using Covidence (Veritas Health Innovation), and duplicates were automatically removed by the software [18]. Titles and abstracts were first screened independently by 2 reviewers (FCO and ME). Next, the full text of each paper was assessed by 2 reviewers (FCO and ME) to determine the eligibility of the studies. At both stages, a third reviewer (MMD) resolved any conflicts. The following data were extracted by two authors (FCO and ME) with a third (MMD) resolving any conflicts: authors; publication year; location; number of patients; age; study type; population details; surgery details; and results, including intraoperative adverse events, postoperative adverse events, postoperative infection, perioperative adverse events, hospital length of stay (LOS), intensive care unit (ICU) LOS, reoperation, readmission, and patient-reported outcomes.

The ROBINS-I (Cochrane) tool was used to assess the risk of bias in the included nonrandomized studies, covering bias due to confounding variables, patient selection, classification of interventions, deviations from intended interventions, missing data, measurement of outcomes, and selection of reported results [19]. Two reviewers (FCO and JG) independently scored all domains, with a third reviewer (MMD) resolving any conflicts. Robvis was used for figure generation [20].

Studies with comparable outcomes were merged for the quantitative meta-analysis and split based on observed measures. After conducting a qualitative evaluation, we determined that there were sufficient data to perform a meta-analysis. RevMan (version 8.4; Cochrane) using random-effects modeling was used for all quantitative analyses. Mean differences for continuous variables (surgical time, EBL, hospital LOS, and ICU LOS) and odds ratios for categorical variables (intraoperative adverse events, postoperative adverse events, postoperative infection, any adverse events, readmission, and reoperation) were the end points of the meta-analysis.

In our search (Figure 1), we identified 5199 unique studies by searching PubMed, MEDLINE, Embase, Cochrane CENTRAL, Web of Science, and Scopus, which were included for abstract screening, of which 64 were forwarded for full-text screening. After full-text review, 7 original studies were included in the data extraction process and 4 of them were in the quantitative analysis [21-27]. Studies were excluded during the full-text review for the following reasons: wrong comparator (n=19, 29.7%), wrong patient population (n=19, 29.7%), pediatric population (n=11, 17.2%), wrong outcomes reported (n=5, 7.8%), and wrong study design (n=4, 6.3%).

Six of the studies included were conducted in the United States [21-25,27], while 1 study was conducted in Japan [26]. The included studies describe a total of 741 patients undergoing either staged (n=331, 44.7%) or same-day (n=410, 55.3%) CF for ASD: 297 (40.1%) patients underwent anterior lumbar interbody fusion, 408 (55.1%) patients underwent lateral interbody fusion (extreme lateral interbody fusion, direct lateral interbody fusion, and lateral lumbar interbody fusion), 54 (7%) patients underwent either lateral lumbar interbody fusion or transforaminal lumbar interbody fusion, and all 741 patients underwent posterior spinal fixation. The average number of fused vertebrae across studies ranged from 4.4 to 10 (SD 3.9). The largest variation between groups within a study ranged from 7.3 (SD 3.1) in the same-day group versus 10 (SD 3.9) in the staged group. The combined mean vertebrae fused for staged and same-day CF was 7.54 (SD 2.41) and 6.62 (SD 2.40), respectively. The follow-up period over all included studies ranged from 1 to more than 36 months, and the average age of patients ranged from 58.8 (SD 9.0) to 72.3 years. The patients included in the study of Masuda et al [26] and Albayar et al [21] were inverse probability weighted to facilitate the comparison of spinal deformity and control for any differences between the groups (Table 1).

Of the included studies, 4 studies compared EBL in the relevant groups, with 3 studies presenting a lower EBL (mean between 412-1127, SD 954 mL) in same-day surgery compared to staged surgery (mean between 642, SD 550, and 1351, SD 869 mL) [21-24]. The meta-analysis shows a nonsignificant advantage for same-day surgery (Figure 2A [21,24,26]). Only 2 studies that measured EBL were included in the quantitative analysis because of inconsistencies in reporting, where some did not report variables as measures of variance, which made pooling in these instances not feasible. Three studies compared the operative time between staged and same-day CF, with all of them reporting a lower mean operative time for same-day CF (mean between 291-479 minutes) compared to staged CF (mean between 426-541 minutes) [22-24]. Just 1 group reported mean and SD for odds ratio time, thus restricting the potential for a quantitative analysis.

Three studies comparatively evaluated the hospital LOS [21,22,25]. All three studies consistently found that the mean LOS was less for same-day CF (mean between 6-34.1 days) in comparison to staged CF (mean between 9-42 days). The meta-analysis clearly presented a shorter LOS in patients undergoing same-day CF compared to staged CF (Figure 2B).

Two of the 7 studies compared intraoperative and postoperative adverse events between staged and same-day CF procedures [21,26]. Both Masuda et al [26] and Albayar et al [21] reported more intraoperative adverse events in staged CF (10.9% and 13.6%, respectively) compared to same-day CF (9.1% and 3.6%, respectively); however, the meta-analysis failed to show a statistically significant difference between the groups (Figure 3A [21-27]). Masuda et al [26] reported fewer postoperative adverse events in staged CF (10.9 vs 12.4%), while Albayar et al [21] presented a lower incidence in same-day CF (53.6 vs 68.2%), without a significant difference in the meta-analysis (Figure 3B). Four studies measured any perioperative adverse events [21,22,26,27]. The overarching analysis showed a higher incidence of adverse events in Staged CF over all studies; however, the meta-analysis did not show significance (Figure 3D).

Four of the included studies compared reoperation and postoperative infection rates between patient groups [21,22,26,27]. Albayar et al [21] and Masuda et al [26] showed a slightly lower postoperative infection in same-day CF versus staged CF (4.4 vs 6.4%, 2.6 vs 4%, and 1.7 vs 11.4% respectively), while Than et al [27] showed a lower postoperative infection in staged CF (0% vs 3.7%; Figure 3C). While Arzeno et al [24] and Than et al [27] reported a lower reoperation in patients undergoing staged CF (11.1 vs 14.9% and 14.8 vs 25.9%, respectively). Masuda et al [26] and Albayar et al [21] reported less reoperation in patients undergoing same-day CF versus staged CF (10.2 vs 10.9% and 14.2 vs 22.7% respectively). However, none of these differences in reoperation reached statistical significance in either the original respective studies or our meta-analysis (Figure 3F).

Readmission rates reported by Arzeno et al [24] and Albayar et al [21] demonstrated diverging results. Arzeno et al [24] found a lower readmission for staged CF versus same-day CF (2.2 vs 12.8%), while Albayar et al [21] reported a higher readmission in staged CF versus same-day CF (22.7 vs 14.3%). The meta-analysis did not indicate a conclusive result (Figure 3E). Harris et al [25] are the only authors reporting on either the Oswestry Disability Index (ODI) or the Scoliosis Research Society Score, showing a better outcome measured by the ODI in same-day CF and no difference in Scoliosis Research Society Score.

There was a moderate degree of bias for all included studies (Figure 4 [21-27]).

In this systematic review and meta-analysis, we compared differences in perioperative outcome variables for individuals undergoing staged or same-day CF in 7 included studies. The meta-analysis revealed a statistically significant increase in hospital LOS in the staged group when compared to the same-day group. There were no statistically significant differences in EBL, intraoperative adverse event, postoperative adverse event, postoperative infection, perioperative adverse event, 30-day readmission, and reoperation rates between patients who underwent staged or same-day CF for ASD. There were marked differences between patient populations and the subsequent clinical outcomes that were reported in each observational study, which most did not adjust for. Our quantitative results are paradoxical as there were inconsistencies between studies in the staged and same-day subgroups for certain reported outcomes, while other findings had stronger conclusive evidence. Due to the apparent differences in consensus, the research included in the following discussion is organized according to the consistency of results.

Heterogeneity in meta-analyses grants the opportunity to examine variable factors that may be leading to the results. There were several reported outcomes that yielded inconclusive evidence in this study. Adverse event rates were reported by authors in several ways, with some splitting between intraoperative or postoperative adverse events, while others reported any adverse event throughout the perioperative course. Two studies reported more intraoperative adverse events in staged CF, but the meta-analysis failed to show a statistically significant difference. This can indicate that there is an increased operative risk when staging, but it can also be due to the generally sicker patients with more spinal deformity that is leading to increased intraoperative adverse events. Further, 1 study presented a lower postoperative adverse events rate during staged CF, while a different group found a higher rate in staged, indicating no difference in postoperative complications between the 2 approaches. When all perioperative adverse events were instead considered, every study reported a higher incidence in staged CF, and this result trended toward significance. Vastly different conclusions depending on the stage of surgery in which the adverse events are considered illuminate the need for greater consistency in reporting outcomes. However, the increased perioperative adverse events seen in staged CF, although not statistically significant, could be clinically important in decision-making.

The Centers for Medicare & Medicaid Services targeted 30-day readmission rates as a source of unnecessary costs [28]. A study conducted by McCarthy et al [29] found the total hospital costs to surgically treat ASD averaged US $120,394, with primary surgery averaging US $103,143, and total readmission costs of US $67,262 for their cohort. The high costs associated with spine surgery make patient readmission and reoperation lucrative targets. There is profound heterogeneity in reoperation and readmission rates that depend on several factors, such as patient demographics, procedure types, and institutional factors. A recent systematic review found the 30-day readmission rate in spine surgery to be 4.2% and 7.4% [30]. In this review, 2 studies reported readmission with diverging results. Arzeno et al [24] found a lower readmission for staged CF, while Albayar et al [21] reported a higher readmission for same-day CF. Likewise, 4 studies reported reoperation with a split in consensus [21,24,26,27]. Given the small number of studies that measured reoperation and readmission, the results from our meta-analysis were inconclusive. The lack of statistically significant differences in the quantitative analysis combined with the heterogeneity in the qualitative review indicates that no differences in readmission and reoperation rates exist between the 2 groups.

It was also uncommon for the included studies to report specific reasons that reoperation or readmission occurred. As readmission depends on many variables including patient comorbidities, initial risk, or operative adverse events, it is crucial for authors to include these measures in future comparative studies for stronger subgroup analyses. Postoperative infection is also a driver leading to reoperation or readmission, as well as additional health care costs [31]. A recent meta-analysis found the pooled incidence of surgical site infection in 22,475 patients to be 3.1% [32]. Our included studies also reported a low incidence of postoperative infection; however, there were inconsistencies as 3 out of the 4 groups found a lower postoperative infection in same-day CF [21,24,26,27]. There are several risk factors for postoperative infection that may lead to heterogeneity. Farshad et al [33] found that intraoperative EBL was a risk factor for postoperative infection. Some studies found BMIs to be predictors of postoperative infection, although this result remains controversial [34-36]. The discrepancies seen in this review may stem from variations in surgical techniques, perioperative antibiotic prophylaxis, institutional differences, and patient-specific risk factors. The results of this study’s meta-analysis do not support any difference between the 2 techniques in postoperation infection risk.

For highly invasive spine surgeries, such as long-segmented CF for ASD, patients may require extended resource use [37]. Surgical operative time and LOS are critical metrics in evaluating the efficiency and resource use of staging in CF. Shorter surgical times are generally associated with reduced intraoperative adverse events, lower anesthesia-related risks, and decreased blood loss, thus contributing to improved patient safety and outcomes [15,38]. Peng et al [39] conducted a meta-analysis on correlations between operative time and postoperative infection, and they concluded that there was a 4-fold increase in postoperative infection risk for operations greater than 3 hours [39]. The studies included in our review all found a lower mean operative time for same-day CF.

A shorter LOS is also desirable as it minimizes health care resource use and lowers costs [40]. Further, early mobilization and discharge are associated with improved patient satisfaction and reduced psychological stress [41]. However, reducing LOS has also been associated with an increasing readmission rate at the population level [42]. Our meta-analysis found a significantly lower LOS for same-day CF compared to staged CF. This difference in LOS may be attributed to the cumulative effect of multiple hospital admissions, prolonged recovery periods between surgeries, and the need for additional preoperative preparation in staged CF. While the lower operative time and LOS results in patients undergoing same-day CF seem intuitive, it is important to interpret them cautiously, considering potential factors such as patient selection, discharge criteria, and institutional practices that may differ between studies. Moreover, while shorter operative time and LOS may reduce health care costs and improve resource allocation, they should not compromise patient safety or postoperative care quality.

The results of shorter LOS and operative time combined with the lack of conclusive differences between groups for adverse events, readmission, reoperation, and infection rates indicate that patient prognosis is similar when between staging preferences. The shorter time spent in the hospital and operating room is clinically important and contributes to improved patient safety and outcomes while reducing hospital resource use. However, the potential advantages of staging may outweigh these benefits and thus improve patient prognosis for more complex cases with a higher degree of relative deformity. Therefore, the results from this review and meta-analysis show preference toward same-day CF, while staging may largely depend on a patient’s condition. Further studies investigating the impact of patient-specific factors, for example, varying degrees of deformity severity, on treatment selection and subsequent outcomes are required.

There are several limitations to this study and the papers included in this review. First, the limited literature comparing staged and same-day CF resulted in only 7 studies being included, which reduced the statistical power of the meta-analysis as only 4 studies qualified for quantitative analysis. All the included studies are retrospective cohort studies with a level of evidence of III or II, and only 2 studies used inverse probability weighting to control for between-patient differences. Consequently, there was a moderate potential for bias, which limited our ability to reach generalizable conclusions. To elucidate the differences in surgical treatment options for CF, it will be necessary to investigate staging in level I or II randomized controlled trials (RCTs). Without RCTs, it is difficult to adjust for surgeon preference bias and the complexity of the case. Each study also demonstrated heterogeneity with respect to the patient outcomes reported, further limiting the ability for robust statistical analysis. Continuous variables were generally poorly reported since many did not report measures of variance such as mean and SD, but rather mean or median and range (min-max or IQR). This made pooling in those instances not feasible, limiting the generalizability of results. Further, only 1 group presented patient-reported outcomes, such as ODI, which is a standardized clinical variable that can be useful for measuring subjective pain and disability. Novel studies should aim to investigate patient-reported outcomes and standardize the variables being reported, so future meta-analyses will have wider samples to yield more conclusive evidence for staging differences in CF. In doing so, the mixed results presented in this study can be further assessed to support wider implementation of either staged or same-day CF while integrating clinically significant metrics to evaluate patient outcomes.

Here, we present the first systematic review and quantitative meta-analysis on staging in CF to treat patients with ASD. Operative time and hospital LOS were significantly lower in same-day CF surgery, with EBL and perioperative adverse events also trending toward significance. However, there was heterogeneity in additional operative measures, such as intra- or postoperative adverse event rates, reoperation, and readmission, and no differences were found in our meta-analysis. Based on our results, it is suggested that same-day CF is advantageous as a potential time-save with patients spending less time in the operating room and hospital, potentially saving costs. However, it is still unclear whether either same-day or staged CF provides a clinical advantage for patient outcomes. Additional level I and II RCTs should be conducted to elucidate the associations between these variables and provide stronger evidence in favor of either approach.