Long-term patient-reported outcomes and mechanical complications following spinal fusion in patients with surgical site infection managed with retained instrumentation versus non-infected patients: a systematic review and meta-analysis

Introduction

Instrumentation in spine fusion surgeries is on the rise over time across a wide range of pathologies, and therefore, postoperative complications are growing (1). One of these consequential complications is surgical site infection (SSI) that leads to increased reoperation, prolonged antibiotic exposure, and delayed functional recovery (2). SSI incidence varies across different cases regarding their surveillance, but contemporary summaries consistently place elective spine SSI in a clinically meaningful range (1).Its impact can extend beyond short-term morbidity, affecting cost and quality-of-life outcomes (3). Risk of SSI in spine surgeries is related to patient and procedural factors; a meta-analysis indicated that diabetes, obesity, longer operative time, and need for transfusion correlate with postoperative SSI (4,5). These risk profiles often overlap with fusion cohorts, where maintaining stability and achieving successful arthrodesis are clinical goals.

Once SSI develops in an implant, it facilitates the persistence and recurrence of infection (6). While culture-based antibiotics and debridement are the core strategy, implant retention is often pursued in order to preserve stability and alignment (7). As a recent systematic review stated, “debridement, antibiotic and implant retention” in acute presentation can be a feasible pathway (6). Antimicrobial treatment can be important in prognosis; in a cohort analysis, biofilm-active antibiotics improved treatment outcomes and reduced postoperative pain (8,9). In long-term follow-up, relevant outcomes are relapse and mortality (10); altogether, these studies show infection eradication can be part of the postoperative course in instrumented fusion complicated by SSI.

For patients, recovery is defined by pain relief and functional restoration; patient-related outcome measure (PROM) studies show that SSI can adversely affect treatment success and PROM outcomes after lumbar instrumentation (11). These outcomes are inconsistent across infection series, and controlled comparisons against non-infected patients are not uniformly available (12).

Quantifying PROM outcomes and mechanical sequelae in SSI managed with instrumentation retention versus non-infected controls in spinal fusion can strengthen prognosis and clinical decision-making, and help clarify expected recovery. In this systematic review and meta-analysis, we aim to evaluate pain and disability outcomes besides mechanical postoperative complications between these two groups to have a better view of the impact of SSI in instrumented fusion surgeries. We present this article in accordance with the PRISMA reporting checklist (available at https://jss.amegroups.com/article/view/10.21037/jss-2026-0132/rc) (13).

Methods

Protocol and registration

The review protocol has been registered in the PROSPERO database (registration number: CRD420261295638).

Search strategy, databases, eligibility criteria, and selection process

We performed a comprehensive search across the four main databases, including PubMed, Web of Science, Scopus, and Embase from inception through December 2025. The search strategy and eligibility criteria are summarized in Table 1. A comprehensive literature search was performed across PubMed, Web of Science, Scopus, and Embase from inception through December 2025. The full search strategies for each database, incorporating database-specific controlled vocabulary and synonyms without title-only restrictions, are provided in Table S1. SB conducted the search and performed the paper duplication removal process using the CADIMA tool (14). Two independent researchers conducted the screening process utilizing the CADIMA web tool. Any discrepancies that arose were resolved through consultation and discussion with a senior researcher.

Extracted outcomes and data collection process

Two authors independently extracted the data for analysis, resolving any discrepancies through discussion or, if required, adjudication by a senior reviewer. The extracted variables encompassed demographics (age, gender, sample size, and study design), SSI characteristics (deep versus mixed), and fusion details (short versus long, with long fusion defined as ≥4 levels). Evaluated outcomes were categorized as clinical [Visual Analog Scale (VAS) for back and leg pain, and the Oswestry Disability Index (ODI)] or mechanical [unplanned readmission, revision surgery excluding debridement, pseudarthrosis, proximal junctional kyphosis (PJK), and rod fracture]. In instances where individual patient data were inaccessible, standard methodological approaches were employed to compute summary statistics from group-level reports. For paired pre- and post-operative datasets reporting only baseline and follow-up means and standard deviations, the mean change (Δ) was computed as the difference between the group means.

Risk of bias assessment process and sensitivity analysis

Two independent reviewers assessed the methodological quality and internal validity of each study using the Newcastle-Ottawa Scale (NOS) (15), resolving any disagreements through a third reviewer. We used NOS-TLPlot (v2.0.3) to synthesize and visualize the domain-level scores. To test the robustness of our findings, we conducted sensitivity analyses by excluding statistical outliers and studies relying on imputed data. Based on our eligibility criteria, the primary analysis included only studies with a low to moderate risk of bias.

Statistical analysis

Software and effect size measures

All meta-analyses were performed using Comprehensive Meta-Analysis (CMA) software (version 4.0). Quantitative synthesis was restricted to factors reported by at least two studies. For continuous variables, effect sizes were calculated using the standardized mean difference (SMD) with 95% confidence intervals (CIs). Statistical significance was defined as P<0.05. Significant variables reported by only a single study were classified as ‘Single Study Factors’; these were excluded from the meta-analysis and reviewed narratively. For binary outcomes with zero events in one or both arms, the software applied a standard continuity correction to compute odds ratios (ORs).

Publication bias assessment

We evaluated publication bias using Egger’s test (16) for any outcome comprising at least three studies. However, because the statistical power of Egger’s test is inherently limited when analyzing fewer than ten studies, these results should be interpreted with caution (17,18).

Heterogeneity assessment

We evaluated statistical heterogeneity using the I² statistics (19), which represents the proportion of variance in observed effects attributable to true study differences rather than sampling error. Heterogeneity levels were categorized as unimportant (0%to40%), moderate (30% to 60%), substantial (50%to90%), or considerable (75%to100%). An I² of zero indicates no observed heterogeneity, while results demonstrating high heterogeneity (I²>50%) were interpreted with caution. Because our analysis included a limited number of studies (<10), prediction intervals were not calculated; instead, 95% CIs were provided for each variable to assess heterogeneity.

Results

Study selection

Our initial database search yielded 2,530 records, comprising 1,363 from Scopus, 444 from Embase, 407 from PubMed, and 316 from Web of Science. Following the removal of 1,442 duplicate entries, we screened the titles and abstracts of the remaining 1,088 records, leading to the exclusion of 946. We sought to retrieve the remaining 142 reports, though 30 were unavailable. A full-text review of the 112 retrieved articles resulted in 101 exclusions (55 failed to meet inclusion criteria, 30 were additional duplicates, 10 lacked sufficient data, and 6 were conference abstracts). Consequently, 11 studies were retained for the final qualitative and quantitative synthesis. The full study selection process is detailed in Figure 1.

Figure 1 PRISMA 2020 flow diagram illustrating the selection process for studies included in the systematic review.

Study characteristics

Eleven met the inclusion criteria and were conducted in the United States, France, Brazil, Canada, Japan, India, and Spain between 2008 and 2025. Across all studies, 2,681 patients were included, of whom 38.3% (n=1,027) were male, and 61.7% (n=1,654) were female, with a mean age of 57.90±6.93 years (range of study means: 43.69–66.70 years). Individual study sample sizes ranged from 32 to 956 patients, with mean follow-up durations ranging from 22 to 70.7 months (overall mean: 33.9 months). Most studies (73%) focus on deep SSI, while three studies included mixed SSI definitions. Long-segment fusion (≥4 levels) was common, with the remaining studies reporting short-segment (27%) or mixed construct lengths (27%). All studies scored seven or higher on NOS, meeting the inclusion criteria (Table 2).

Risk of bias assessment

Eight studies were rated low risk and three at moderate risk, with generally low risk in the selection domain, supporting the reliability of the synthesized findings (Figure 2). Most concerns were noted in the compatibility and outcome domains, reflecting limitations inherent to the observational design and incomplete adjustment for confounders in some cohorts.

Figure 2 NOS risk of bias assessment for included cohort studies. NOS, Newcastle-Ottawa Scale; RoB, risk of bias.

Synthesis of results

Five studies were pooled to assess change in back pain VAS. Compared with non-infected controls, SSI managed with instrument retention was not significantly associated with differences in back pain improvement (SMD −0.58, 95% CI: −1.85 to 0.68). Three studies reported changes in leg pain VAS that were not significantly different between groups (SMD −0.71, 95% CI: −2.50 to 1.07). For functional outcomes, 5 studies reported changes in ODI. Improvement in ODI was comparable between the two groups (SMD 0.07, 95% CI: −0.125 to 0.26), with minimal heterogeneity (I²=2.6%).

SSI with retained instrumentation was significantly associated with higher odds of unplanned readmission (OR 36.75; I²=93%; P=0.02) in three studies and with higher odds of revision surgery, excluding debridement (OR 3.34; I²=55.8%; P<0.001) in five studies. In addition, five studies reported pseudarthrosis more frequently in the SSI group (OR 2.95; I²=71%; P=0.03). PJK did not differ significantly between groups. Only one study by Prost (20) reported a rod fracture rate of 28% in the SSI group, and 14% in the non-SSI group (Figure 3). Egger’s tests did not suggest the presence of small-study effects for any of the analyzed variables, and publication-bias testing was not performed for outcomes with very few contributing studies. All the analyzed data are summarized in Table 3.

Figure 3 Meta-analysis result and forest plot for pseudarthrosis, unplanned readmission, need for revision surgery. CI, confidence interval; SSI, surgical site infection.

Sensitivity analysis results

Sensitivity analyses were performed to evaluate the robustness of our findings for readmission rates by removing statistical outliers. The study by Karamian et al. [2022] was identified as an extreme outlier (SSI readmission rate of 100%). When this outlier was excluded, the pooled estimate for readmission remained statistically significant, demonstrating a robustly higher risk of readmission in the SSI cohort compared to the non-SSI cohort (OR 12.52, 95% CI: 4.15 to 37.75, P<0.001), confirming that while the magnitude of the effect was influenced by the outlier, the overall direction and significance of the primary findings were not entirely driven by it.

The sensitivity analysis demonstrates that the findings for the revision rate are highly robust. Excluding any individual study from the analysis slightly shifts the pooled odds ratio (ranging from 2.40 to 4.08), but the result remains highly statistically significant (P<0.001) in every scenario. This confirms a consistently higher odd of revision in the SSI cohort compared to the non-SSI cohort.

The results of the sensitivity analysis for pseudarthrosis showed that the pooled estimate remained consistently positive across all iterations; however, the magnitude of the OR exhibited sensitivity to the inclusion of specific studies. Notably, the exclusion of Boishardy et al. [2022] resulted in the most substantial decrease in the pooled OR to 1.99, likely attributable to the study’s larger sample size. Conversely, the exclusion of Urquhart et al. [2021] resulted in an increase of the pooled OR to 4.49, indicating that this study exerted a downward influence on the collective estimate. The exclusion of the remaining studies (Mok, 2008; Fanous, 2019; Haddad, 2018) resulted in only minor fluctuations in the pooled OR (range, 2.52–2.94), suggesting that our findings are relatively stable across the majority of the included literature.

Discussion

In this comparative analysis across 11 studies of adult spinal fusion, patients who developed SSI that were managed with retained instrumentation achieved similar ODI improvement and no statistically significant change in back or leg pain VAS compared to non-infected controls. In contrast, the SSI group had substantially higher unplanned readmission, revision surgery excluding debridement, and risk of arthrosis, while evidence regarding PJK and rod fracture was limited.

Patient reported outcomes

Longer-term functional recovery can approximate that of non-infected controls, consistent with comparative PROM literature reporting that infections impose an early outcome penalty that may attenuate by time in selected fusion cohorts (21,22). In this context, our pooled ODI data with minimal heterogeneity supports this concept that once stability is preserved and infection is adequately controlled, global disability may normalize even following a major complication (23).

An important point here is that although long-term PROM recovery is seen, it does not necessarily mean equivalent fusion biology (21). When the implant is retained, local inflammatory and microbial environment could still be favorable for nonunion risk while mechanical stability is preserved and symptoms have improved (7). Chronic implant infections and the biofilm formed disrupt bone healing and cause delayed fusion maturation (24).

The substantial heterogeneity observed in pain change may reflect clinically meaningful differences across studies in factors including index pathology, construct length, depth, and timing of infection, pathogen, surgical washouts, and follow-up duration; all of which can influence pain recovery independent of disability improvement (21). Despite comparable symptom relief, SSI triggers repeated clinical evaluations, extended wound care, and additional procedures. That is why PROM recovery is similar between groups, but healthcare use is higher in the SSI group (25).

Higher readmissions and non-debridement revisions observed in our study are concordant with established treatment for spinal implant infection, in which implant retention is often used to preserve alignment, yet the treatment course may still need repeated operative management or escalation of therapies when infection control is incomplete (26).

Pseudarthrosis

Pseudarthrosis is one of the mechanical indicators in our pooled analysis. Some deformity cohorts similarly found that SSI is a major risk factor for pseudarthrosis, supporting our results (27). Even when infection resolves and function improves, surgeons should view SSI as fundamentally altering fusion prognosis, not just as a short-term complication (28).

There is also a significant chance that occult infections can appear as “aseptic” pseudarthrosis (29). In about one out of ten revision surgeries for pseudarthrosis, an occult infection is found even when there were no signs before surgery, with Cutibacterium acnes often involved (29,30). These results support taking multiple deep cultures and using longer incubation times during revision surgery for suspected pseudarthrosis, especially in patients with a history of SSI or unexplained hardware loosening (31).

Overall, these data show that infection eradication and PROM recovery can be achieved in many patients with retained implants, but SSI is still a potent factor in mechanical vulnerability and reintervention risk after fusion surgery (32).

Clinical implications

In preoperative and postoperative counselling, these data imply that despite clinically meaningful recovery in patients with SSI treated with implant retention, we should anticipate a higher probability of readmission and later revisions. From a surgical view, SSI should be treated as a modifier in fusion, warranting closer radiographic surveillance for bony union and earlier clinical evaluation when radiographic loosening, persistent pain, or delayed recovery is noticed. In SSI managed by implant retention, clinicians should set a low threshold for CT imaging when symptoms persist or radiographs show unclear findings, as this population has elevated pseudarthrosis risk. When evaluating suspected nonunions, especially in patients with prior SSI, it is better to consider obtaining intraoperative cultures, since infection may remain subclinical despite unremarkable laboratory markers. Furthermore, the increased healthcare utilization and further substantial cost and source burden support prevention, early diagnosis, and standardized care as priority goals.

Limitations

Since our included studies are observational, confounding is likely, given that SSI groups often differ from controls in baseline characteristics and operative complexity. These factors can independently affect readmission, reoperation, and nonunion risk. Outcome definitions and follow-up duration varied across studies. The lack of standard fusion criteria across studies introduced heterogeneity in pseudarthrosis rates and hindered meaningful comparison. Given that both imaging techniques and timing affect pseudarthrosis detection, future comparative investigations of SSI should specify a priori fusion definitions and assessment intervals to minimize outcome misclassification. The heterogeneity in pain change outcomes suggests that these pooled data should be interpreted cautiously and further explored with future research that uses standardized PROM assessment windows. Also, a small number of studies across several outcomes limit precision despite large point estimates. We compared patients with SSI managed with retained instrumentation to non-infected controls, rather than comparing implant retention against implant removal or exchange. Therefore, our analysis evaluates the impact of an SSI in a highly selected group of patients rather than the comparative effectiveness or safety of the retention technique itself. There is an inherent selection bias, as patients deemed suitable for implant retention may have presented with less severe infections, earlier onset, or more stable spinal constructs than those requiring removal. Consequently, our findings cannot be used to conclude that implant retention is superior or equal to removal; they only demonstrate the clinical trajectory of patients who were already selected for retention. Finally, outcome definitions and follow-up durations varied across the included studies. The mean follow-up duration was 33.9 months (range, 22 to 70.7 months); however, the lack of standardized clinical and radiographic criteria across institutions, particularly regarding the definition of pseudarthrosis and successful fusion, introduced heterogeneity into these pooled estimates and warrants cautious interpretation.

Conclusions

In adult spinal fusion, retention of the implant in SSI patients is compatible with disability improvement in comparison to non-infected controls, but associated with substantially higher rates of readmission and revision, as well as risk of pseudarthrosis. Future work should prioritize SSI definitions, standardized PROM time points, and risk-adjusted study designs to clarify better treatment pathways.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://jss.amegroups.com/article/view/10.21037/jss-2026-0132/rc

Peer Review File: Available at https://jss.amegroups.com/article/view/10.21037/jss-2026-0132/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jss.amegroups.com/article/view/10.21037/jss-2026-0132/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Stewart KE, Terada R, Windrix C, et al. Trends and Prediction of Surgical Site Infection After Elective Spine Surgery: An Analysis of the American College of Surgeons National Surgical Quality Improvement Project Database. Surg Infect (Larchmt) 2023;24:506-13. [Crossref] [PubMed]
2. Zhang T, Lian X, Chen Y, et al. Clinical outcome of postoperative surgical site infections in patients with posterior thoracolumbar and lumbar instrumentation. J Hosp Infect 2022;128:26-35. [Crossref] [PubMed]
3. Pennington Z, Sundar SJ, Lubelski D, et al. Cost and quality of life outcome analysis of postoperative infections after posterior lumbar decompression and fusion. J Clin Neurosci 2019;68:105-10. [Crossref] [PubMed]
4. Zhang X, Liu P, You J. Risk factors for surgical site infection following spinal surgery: A meta-analysis. Medicine (Baltimore) 2022;101:e28836. [Crossref] [PubMed]
5. Peng XQ, Sun CG, Fei ZG, et al. Risk Factors for Surgical Site Infection After Spinal Surgery: A Systematic Review and Meta-Analysis Based on Twenty-Seven Studies. World Neurosurg 2019;123:e318-29. [Crossref] [PubMed]
6. Karczewski D, Schnake KJ, Osterhoff G, et al. Postoperative Spinal Implant Infections (PSII)-A Systematic Review: What Do We Know So Far and What is Critical About It?. Global Spine J 2022;12:1231-46. [Crossref] [PubMed]
7. Prinz V, Vajkoczy P. Surgical revision strategies for postoperative spinal implant infections (PSII). J Spine Surg 2020;6:777-84. [Crossref] [PubMed]
8. Köder K, Hardt S, Gellert MS, et al. Outcome of spinal implant-associated infections treated with or without biofilm-active antibiotics: results from a 10-year cohort study. Infection 2020;48:559-68. [Crossref] [PubMed]
9. Bagherzadeh S, Moghadam N, Rostami-Ravari M, et al. Comparative effectiveness of gram-positive versus mixed antibiotic prophylaxis on surgical site infection rates following three-column osteotomy for adult spinal deformity. Clin Neurol Neurosurg 2025;258:109173. [Crossref] [PubMed]
10. Stheme De Jubecourt A, Seng P, Delarbre D, et al. Post-spinal surgery infections: a two-year follow-up study on risks, resistance, and recovery. BMC Infect Dis 2025;25:1492. [Crossref] [PubMed]
11. Lazary A, Klemencsics I, Szoverfi Z, et al. Global Treatment Outcome after Surgical Site Infection in Elective Degenerative Lumbar Spinal Operations. Surg Infect (Larchmt) 2021;22:193-9. [Crossref] [PubMed]
12. Nakamoto H, Nakajima K, Miyahara J, et al. Does surgical site infection affect patient-reported outcomes after spinal surgery? A multicenter cohort study. J Orthop Sci 2024;29:1370-5. [Crossref] [PubMed]
13. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71. [Crossref] [PubMed]
14. Kohl C, McIntosh EJ, Unger S, et al. Online tools supporting the conduct and reporting of systematic reviews and systematic maps: a case study on CADIMA and review of existing tools. Environmental Evidence 2018;7:8.
15. Stang A. Critical evaluation of the Newcastle-Ottawa scale for the assessment of the quality of nonrandomized studies in meta-analyses. Eur J Epidemiol 2010;25:603-5. [Crossref] [PubMed]
16. Egger M, Davey Smith G, Schneider M, et al. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997;315:629-34. [Crossref] [PubMed]
17. Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics 1994;50:1088-101. [PubMed]
18. Page MJ, Sterne JAC, Higgins JPT. Assessing risk of bias due to missing results in a synthesis. Cochrane Handbook for Systematic Reviews of Interventions 2019:349-74.
19. Hemming K, Hughes JP, McKenzie JE, et al. Extending the I-squared statistic to describe treatment effect heterogeneity in cluster, multi-centre randomized trials and individual patient data meta-analysis. Stat Methods Med Res 2021;30:376-95. [Crossref] [PubMed]
20. Prost S, Ambrosino R, Fuentes S, et al. Long-term clinical and biomechanical impact of surgical site infection after thoracolumbar fusion for adult spinal deformity: Retrospective analysis of 124 patients with more than 3 years of follow-up. Orthop Traumatol Surg Res 2026;112:104461. [Crossref] [PubMed]
21. Urquhart JC, Gurr KR, Siddiqi F, et al. The Impact of Surgical Site Infection on Patient Outcomes After Open Posterior Instrumented Thoracolumbar Surgery for Degenerative Disorders. J Bone Joint Surg Am 2021;103:2105-14. [Crossref] [PubMed]
22. Karamian BA, Mao J, Toci GR, et al. Clinical Outcomes at One-year Follow-up for Patients With Surgical Site Infection After Spinal Fusion. Spine (Phila Pa 1976) 2022;47:1055-61. [Crossref] [PubMed]
23. Tobert DG. The Outcome Horizon Is Optimistic After Infection: Commentary on "Clinical Outcomes at 1 Year Follow-up for Patients With Surgical Site Infection After Spinal Fusion". Spine (Phila Pa 1976) 2022;47:1062. [Crossref] [PubMed]
24. Tai DBG, Patel R, Lovecchio F, et al. State-of-the-Art Review: Diagnosis and Management of Spinal Implant Infections. Clin Infect Dis 2024;79:e65-71. [Crossref] [PubMed]
25. Khanna K, Janghala A, Sing D, et al. An Analysis of Implant Retention and Antibiotic Suppression in Instrumented Spine Infections: A Preliminary Data Set of 67 Patients. Int J Spine Surg 2018;12:490-7. [Crossref] [PubMed]
26. Kowalski TJ, Berbari EF, Huddleston PM, et al. The management and outcome of spinal implant infections: contemporary retrospective cohort study. Clin Infect Dis 2007;44:913-20. [Crossref] [PubMed]
27. Boishardy A, Bouyer B, Boissière L, et al. Surgical site infection is a major risk factor of pseudarthrosis in adult spinal deformity surgery. Spine J 2022;22:2059-65. [Crossref] [PubMed]
28. Fanous AA, Kolcun JPG, Brusko GD, et al. Surgical Site Infection as a Risk Factor for Long-Term Instrumentation Failure in Patients with Spinal Deformity: A Retrospective Cohort Study. World Neurosurg 2019;132:e514-9. [Crossref] [PubMed]
29. Burkhard MD, Loretz R, Uçkay I, et al. Occult infection in pseudarthrosis revision after spinal fusion. Spine J 2021;21:370-6. [Crossref] [PubMed]
30. Hu X, Lieberman IH. Revision spine surgery in patients without clinical signs of infection: How often are there occult infections in removed hardware?. Eur Spine J 2018;27:2491-5. [Crossref] [PubMed]
31. Gelderman SJ, Faber C, Kampinga GA, et al. A high prevalence of Cutibacterium acnes infections in scoliosis revision surgery, a diagnostic and therapeutic dilemma. Spine Deform 2023;11:319-27. [Crossref] [PubMed]
32. Hollern DA, Woods BI, Shah NV, et al. Risk Factors for Pseudarthrosis After Surgical Site Infection of the Spine. Int J Spine Surg 2019;13:507-14. [Crossref] [PubMed]