Impact of abscess type on outcomes following posterior fixation for thoracolumbar pyogenic spondylitis: a multicenter retrospective cohort study
Article information
, Toru Funayama1, Yosuke Ogata1,3, Takane Nakagawa1, Takahiro Sunami1, Kotaro Sakashita1,4, Shun Okuwaki5, Kento Inomata6, Kaishi Ogawa7, Yosuke Shibao8, Hiroshi Kumagai9, Katsuya Nagashima5, Kengo Fujii7, Yosuke Takeuchi5, Masaki Tatsumura5, Itsuo Shiina4, Masafumi Uesugi10, Masao Koda1
Abstract
Study Design
Multicenter retrospective cohort study.
Purpose
To evaluate the impact of abscess presence and type on treatment duration and clinical outcomes in patients undergoing minimally invasive posterior fixation for thoracolumbar pyogenic spondylitis.
Overview of Literature
Surgical management is increasingly favored for pyogenic spondylitis. Abscesses, particularly epidural and iliopsoas, have been linked to poorer prognoses, and empyema, though uncommon, tends to be particularly severe. However, the effect of specific abscess types on surgical outcomes remains unclear.
Methods
This study included 92 patients who underwent minimally invasive posterior fixation across 10 centers between 2014 and 2024. Patients were classified into an abscess group (epidural, iliopsoas, empyema, or other) and a non-abscess group. Clinical outcomes, including total duration of intravenous antibiotics and unplanned additional surgeries, were compared. Subgroup and regression analyses were conducted to assess the impact of specific abscess type.
Results
Abscesses were present in 65 patients (71%): epidural (n=51), iliopsoas (n=38), and empyema (n=3), with some overlap. Compared with the non-abscess group (n=27; 29%), there were no significant differences in antibiotic duration or rates of unplanned additional surgery. However, iliopsoas abscess was associated with longer antibiotic duration (8.1 weeks vs. 6.6 weeks, p=0.044), while all empyema cases required additional surgery for poor infection control (p=0.000). Regression analysis identified iliopsoas abscess and age ≥65 years as independent predictors of prolonged antibiotic use, whereas epidural abscess was associated with shorter antibiotic duration.
Conclusions
Iliopsoas abscesses were associated with longer antibiotic courses, while empyema was linked to poor infection control and a higher likelihood of additional surgery.
Introduction
The incidence of pyogenic spondylitis has been rising over the years [1]. Evidence indicates that surgical treatment provides superior outcomes compared with conservative management, leading to a shift toward surgery-centered strategies [2]. Several surgical techniques have been developed; among these, minimally invasive posterior fixation without debridement or bone grafting is widely practiced to achieve early infection control and shorten treatment duration [3,4].
In pyogenic spondylitis, abscess formation is regarded as a poor prognostic factor [1,5,6]. Epidural abscesses often necessitate surgical intervention [5,6], while iliopsoas abscesses are associated with prolonged treatment [7,8]. Empyema, although rare, represents a challenging complication that typically requires surgical drainage [9]. However, some studies suggest that abscesses associated with pyogenic spondylitis are secondary manifestations and may resolve without drainage if the underlying spinal infection is effectively managed [10,11].
Although abscess formation is generally considered a negative prognostic factor in pyogenic spondylitis [5–8], its impact on treatment outcomes in surgically managed patients remains unclear. To date, no studies have directly compared the effects of specific abscess types (epidural abscess, iliopsoas abscess, and empyema) on treatment outcomes or antibiotic duration. Therefore, the present study aimed to evaluate the effects of the presence and type of abscess on treatment duration and clinical outcomes in patients undergoing posterior fixation for thoracolumbar pyogenic spondylitis. We hypothesized that patients with any type of abscess would experience longer treatment courses and higher rates of treatment failure than those without abscesses.
Materials and Methods
Ethics statement
This study was approved by the Institutional Review Board (IRB) of the University of Tsukuba (protocol number: 06-110; approval date: July 31, 2024). All procedures were conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants, and the consent procedures were approved by the IRB.
Study design
This multicenter retrospective cohort study was approved by the institutional review boards of the 10 participating institutions. We reviewed 126 consecutive patients who underwent minimally invasive posterior fixation for thoracolumbar pyogenic spondylitis (T1/T2–L5/S1) at the 10 affiliated institutions between January 2014 and December 2024. Eligible patients had thoracolumbar pyogenic spondylitis resistant to conservative treatment, including intravenous antibiotics, and were treated with minimally invasive posterior fixation, primarily using percutaneous pedicle screws (PPSs) [1,3,4,8], with a minimum follow-up of 6 months postoperatively [3,4,8]. Patients were excluded if they had undergone planned surgical procedures involving direct debridement or bone grafting of the infected intervertebral space, or if they had tuberculous spondylitis [3,4,8].
Data collection
The following demographic and clinical data were collected for all patients [3–8]: (1) Patient characteristics at admission: age, sex, infection site (thoracic and thoracolumbar: T1/T2–T12/L1; lumbar and lumbosacral: L1/L2–L5/S1), neurological deficits, number of infected levels (single-level versus multilevel), comorbidities (e.g., diabetes mellitus, cancer), concurrent infectious diseases (e.g., urinary tract infection, pneumonia, or infective endocarditis), laboratory findings (white blood cell [WBC] counts, serum C-reactive protein [CRP] level), and causative organisms. (2) Abscess evaluation by imaging [5,10,11]: Preoperative magnetic resonance imaging and computed tomography (CT) scans were used to determine the presence or absence and type (epidural, iliopsoas, empyema, or other). Abscess size was assessed using standardized imaging parameters [11–13]: the maximum diameter of epidural abscesses on sagittal or axial T2-weighted magnetic resonance images [12], the maximum anteroposterior diameter of iliopsoas abscesses on axial CT images [11], and the maximum thickness of empyema on CT [13]. (3) Surgical and postoperative outcomes: operation time, blood loss, range of fixation, combined procedures (decompression or abscess drainage), postoperative follow-up duration, duration of pre- and postoperative intravenous antibiotics, unplanned additional surgeries (for poor infection control, implant failure, or surgical site infection), and infection recurrence.
Imaging assessments were independently performed by three experienced spine surgeons (each with >10 years of clinical practice). Any discrepancies were resolved by consensus.
Surgical procedures and antibiotic protocol
Surgery was indicated for patients with neurological deficits, progressive bone destruction, persistent or enlarging abscesses, spinal instability, or failed conservative treatment [1,3,4,8]. All procedures were performed under general anesthesia with the patient in the prone position [3,4,8]. PPSs were placed two or three levels above and below the infected segment, with institutional discretion regarding screw placement into the infected vertebrae. No bone grafting or debridement of the infected intervertebral space was performed; rods were inserted in situ, completing the posterior fixation.
For abscesses, posterior fixation alone was generally performed. However, in cases with neurological symptoms or large abscesses, decompression or drainage was performed concurrently.
Following surgery, patients received a standard 6-week course of intravenous antibiotics [1,7,8]. The attending physician at each institution selected the antibiotic type according to culture results and the patient’s clinical course. Antibiotic therapy was discontinued at the physician’s discretion, based on imaging evidence of abscess reduction and normalization or marked reduction of CRP levels [1,7,8].
Clinical outcomes
For background analysis, patients were classified into two groups: the abscess group (any type of abscess—epidural, iliopsoas, empyema, or other) and the non-abscess group. Between these groups, we compared the total duration of intravenous antibiotic therapy (pre- and postoperative) and the incidence of unplanned additional surgeries following posterior fixation [3,4,8]. The primary outcomes were (1) the duration of intravenous antibiotic use and (2) the incidence of unplanned additional surgeries stratified by abscess type (epidural, iliopsoas, or empyema). The effect of the abscess type on the duration of intravenous antibiotic use was also assessed.
Statistical analysis
Continuous variables (e.g., age, CRP, and antibiotic duration) were presented as mean±standard deviation and compared using Welch’s t-test. Categorical variables (e.g., occurrence of unplanned additional surgery) were analyzed using Fisher’s exact test. Age and CRP levels were further dichotomized as follows: age (≥65 years and <65 years) and CRP (<10 mg/dL and ≥10 mg/dL) [3,8]. To identify factors influencing antibiotic duration, we performed multiple regression analysis, with total intravenous antibiotic duration as the dependent variable and age, sex, diabetes mellitus, cancer, CRP level at admission, and the presence of epidural abscess, iliopsoas abscess, and empyema as explanatory variables [3,4,8]. A p-value of <0.05 was considered statistically significant. All statistical analyses were performed using JMP ver. 10.0 (SAS Inc., Cary, NC, USA).
Results
Patient background
Of the 126 patients screened, 24 were excluded due to short follow-up (including dropouts), seven due to planned direct manipulation of the infection site, and three due to tuberculous spondylitis. Consequently, 92 patients were included in the final analysis (mean age, 72.6±10.6 years; range, 44–90 years; 64 male and 28 female) (Fig. 1).
Inclusion and exclusion criteria used for sample selection.
Based on abscess presence, 65 patients (71%) were classified into the abscess group and 27 (29%) into the non-abscess group. The abscess distribution (not mutually exclusive) was as follows: epidural abscess in 51 patients (55%), iliopsoas abscess in 38 (41%), and empyema in 3 (3%). Additionally, two patients had abscesses in other locations (lung, n=1; extrapleural, n=1). The most frequent combination was concurrent epidural and iliopsoas abscesses, observed in 24 patients (37%) (Table 1).
Distribution of different types of abscesses in the study population (n=65)
Compared with the non-abscess group, the abscess group had a significantly higher proportion of patients with neurological deficits (25 vs. 4, p=0.029), a higher incidence of prior infectious disease (25 vs. 4, p=0.029), and higher WBC counts on admission (12,401/μL vs. 9,639/μL, p=0.020). No significant differences were observed in the other baseline variables (Table 2).
Preoperative characteristics of patients with and without abscesses
Comparison of surgical data and postoperative outcome between groups
There were no significant differences between the abscess and non-abscess groups with respect to operation time, blood loss, or range of fixation (Table 3). In the abscess group, 13 patients underwent additional procedures beyond posterior fixation: decompression for epidural abscess drainage (n=8), iliopsoas abscess drainage (n=2), empyema drainage (n=2), and extrapleural abscess drainage (n=1). In the non-abscess group, only one patient required decompression for neurological symptoms.
Surgical and postoperative outcomes of the abscess and non-abscess groups
Regarding postoperative clinical outcomes, no significant differences were observed between the two groups in terms of the duration of pre- or postoperative intravenous antibiotic therapy. No cases of infection recurrence occurred during follow-up. Unplanned additional surgeries were required in 19 patients: poor infection control (n=7), implant failure (n=5), and surgical site infection (n=4). The incidence of unplanned additional surgery was comparable between the two groups (abscess group: 13 patients [20%]; non-abscess group: six patients [22%]; p=0.78).
Impact of abscess type on clinical outcomes
Postoperative clinical outcomes
Patients with iliopsoas abscesses required significantly longer total antibiotic treatment than those without (8.1 weeks vs. 6.6 weeks, p=0.044) (Table 4). In terms of unplanned additional surgeries, all three patients with empyema required additional surgery (p=0.0077), primarily due to poor infection control (p=0.0003), reflecting a statistically significant difference compared to patients without empyema.
Postoperative clinical outcomes categorized according to abscess type
Among the three empyema cases, two patients underwent simultaneous empyema drainage at the time of posterior fixation but subsequently required anterior debridement and autologous bone grafting or corpectomy with metal implants due to persistent infection. The third patient developed worsening empyema after posterior fixation, necessitating unplanned drainage (Fig. 2).
Illustrative case of a patient with empyema in the abscess group. (A, B) An 81-year-old man with pyogenic spondylitis at T11/12 had a history of urinary tract infection 1 month earlier and presented with an epidural abscess (A) (arrows) and empyema (B). Despite 4 days of preoperative intravenous antibiotics, his back pain worsened, making continued conservative treatment difficult. (C) Minimally invasive posterior fixation from T10 to L1 was performed. (D) However, 2 weeks after surgery, the empyema worsened, requiring additional drainage surgery. (E) After 2 weeks of drainage, the empyema improved. Intravenous antibiotics were administered for a total of 6.6 weeks. The patient remains under follow-up without infection recurrence.
In the abscess group, the total duration of antibiotic therapy was not significantly different between the 13 patients who underwent drainage simultaneously with posterior fixation (6.8±3.2 weeks) and the 52 patients who did not (7.3±3.0 weeks) (p=0.61). The mean maximum diameters of the abscesses were as follows: epidural abscesses, 20.5±16.6 mm (range, 3.2–68 mm); iliopsoas abscesses, 32.4±19.6 mm (range, 7.8–90 mm); and empyema, 42.0±9.5 mm (range, 32–51 mm). However, the abscess size showed no significant association with the total duration of antibiotic therapy.
Relationship between abscess and intravenous antibiotic duration
Regression analysis showed that older age (≥65 years) and the presence of an iliopsoas abscess were associated with prolonged antibiotic treatment (p=0.018 and 0.035, respectively), whereas the presence of an epidural abscess was associated with shorter treatment duration (p=0.029) (Table 5, Fig. 3). Other factors, namely sex, comorbidities, CRP levels at admission, and the presence of empyema, were not significantly associated with treatment duration.
Results of the multiple regression analysis for the duration of total intravenous antibiotics
Illustrative case of a patient with an epidural abscess in the abscess group. A 68-year-old man developed pyogenic spondylitis at L3/4 (A, B). He had a concurrent diagnosis of infective endocarditis and received 3 weeks of preoperative antibiotics. However, (A, B) an epidural abscess persisted (arrows), and (C, D) minimally invasive posterior fixation from L2 to L5 was performed. (E) Magnetic resonance imaging performed 2 weeks after surgery showed complete resolution of the epidural abscess (arrow), allowing discontinuation of intravenous antibiotics. The total duration of intravenous antibiotic therapy was 5 weeks. The patient remains under follow-up without infection recurrence.
Discussion
This study yielded several key findings. Although patients with abscesses presented with higher WBC counts and a greater incidence of prior infectious diseases at admission, postoperative clinical outcomes, including antibiotic duration and unplanned additional surgeries, did not differ significantly between patients with and without abscesses. However, abscess type influenced outcomes: iliopsoas abscesses were associated with significantly longer antibiotic treatment, epidural abscesses with shorter treatment durations, and empyema with poor infection control and a higher rate of unplanned additional surgeries.
Abscesses, such as epidural, iliopsoas, and empyema, are established indicators of disease severity and are incorporated into classification systems and prognostic scores for pyogenic spondylitis [14–16]. Abscess formation is common in pyogenic spondylitis, with epidural and paravertebral abscesses reported in approximately 55% of cases [6], although the frequency varies across studies. Despite its rarity, empyema is now recognized as a severe complication of pyogenic spondylitis [9,17]. Although the potential severity of untreated abscesses is well known, comprehensive data on the epidemiology of specific abscess types and their differential impact on outcomes remain limited [1,5–9].
To our knowledge, this is the first study to investigate the prevalence of different abscess types and their impact on outcomes in patients undergoing surgery for pyogenic spondylitis. Two key findings emerged. First, the presence of an abscess did not significantly affect the duration of antibiotics or the rate of unplanned additional surgeries. Although abscess prevalence was quite high (71%; epidural 55%, iliopsoas 41%, and empyema 3%), their presence alone did not worsen treatment duration, recurrence rates, or complication rates. It should be emphasized that this study focused exclusively on postoperative outcomes in patients treated with posterior fixation. Surgical approaches to pyogenic spondylitis vary widely, including percutaneous endoscopic drainage (PED) and combined anterior-posterior surgery [18,19]. Compared with posterior fixation, PED is less invasive but has certain limitations in managing large abscesses [18]. In this regard, the minimally invasive posterior fixation approach used in this study provided stable infection control and yielded generally favorable outcomes despite the presence of an abscess [20]. Nonetheless, the inherent risks of posterior fixation, including implant failure (5% in this series) and adjacent segment disease (2%), should be recognized.
Second, the type of abscess was an important determinant of outcome. Iliopsoas abscesses were associated with prolonged antibiotic treatment, empyema was linked to poor infection control and unplanned surgeries, whereas epidural abscesses did not adversely affect clinical outcomes.
The presence of an iliopsoas abscess may indicate greater infection severity and necessitate longer antibiotic therapy, consistent with previous reports [8,9]. In such cases, earlier surgical intervention or combined anterior-posterior surgery may help shorten treatment duration [8,19,21–23]. Interestingly, epidural abscesses in our cohort were associated with shorter antibiotic courses. Although epidural abscesses are often considered surgical indications [5,6], their presence after posterior fixation was not linked to poor outcomes; only eight of 51 patients (16%) in our series required direct drainage. Recent studies have similarly reported reduced mortality in patients with epidural abscesses, possibly reflecting early surgical intervention and aggressive infection control [24]. These findings suggest that, in the absence of severe neurological deficits, posterior fixation alone may be sufficient for managing both pyogenic spondylitis and associated epidural abscesses [10,11]. In our cohort, the incidence of empyema was only 3%; however, unlike epidural and iliopsoas abscesses, all patients with empyema required surgical drainage. Empyema represents a severe infection that extends beyond the parietal pleura into the pleural space [9,17], and evidence indicates that minimally invasive surgery alone is insufficient for its control [25]. Whenever feasible, empyema drainage should be performed concurrently with initial posterior fixation, with surgeons remaining alert to the possible need for additional anterior debridement if infection control is inadequate. However, given the small number of empyema cases in this study (n=3), these findings should be considered exploratory and hypothesis-generating.
This study has several limitations. First, its retrospective, multicenter design introduced inevitable variability in treatment strategies across institutions. The criteria for additional procedures, such as decompression, were not standardized, raising the possibility of selection bias. Second, decisions regarding the duration of antibiotic therapy were made at the discretion of the attending physicians, without a uniform protocol. Third, the sample size was modest (92 patients overall), with only three cases of empyema, and the surgical approach was limited to posterior fixation. As a result, the findings related to empyema should be interpreted as exploratory and hypothesis-generating. Fourth, seven patients who underwent direct manipulation of the infection site were excluded, which may have biased the cohort toward less severe disease.
Fifth, among the 65 patients in the abscess group, 28 had multiple abscesses. Consequently, in analyses other than the multivariate model, these cases could not be treated as fully independent observations, introducing potential confounding. The results should therefore be interpreted with caution.
Sixth, because this study focused exclusively on postoperative outcomes, the findings may not be generalizable to all patients with pyogenic spondylitis, particularly those managed conservatively. In addition, the limited sample size limited the ability to perform multivariate analyses to identify risk factors for treatment failure. Larger studies are needed to validate these findings, compare different surgical approaches, and clarify the impact of causative organisms and pathogen types, which were not evaluated here.
Despite these limitations, this study offers meaningful evidence by reporting outcomes of a single surgical approach in 92 patients with spinal infections. Despite its retrospective design, data were collected across 10 centers with a mean postoperative follow-up exceeding 2 years. While caution is warranted in extrapolating these findings to patients managed conservatively, treated with alternative surgical procedures, or from different racial and geographic populations, a reasonable degree of generalizability can be inferred.
Conclusions
In patients undergoing posterior fixation for thoracolumbar pyogenic spondylitis, the presence of abscesses did not significantly influence antibiotic duration or the rate of unplanned additional surgeries. Epidural abscesses were associated with shorter antibiotic courses, whereas iliopsoas abscesses required prolonged antibiotic therapy. Empyema appeared to be related to poor infection control, requiring additional surgery. Careful postoperative monitoring is advisable for patients with iliopsoas abscesses and empyema.
Key Points
Abscess presence did not significantly affect antibiotic duration or the rate of unplanned additional surgeries following posterior fixation.
Iliopsoas abscesses required significantly longer antibiotic treatment.
Epidural abscesses were associated with shorter treatment durations.
All empyema cases required additional surgery for poor infection control.
Posterior fixation alone often achieved effective infection control, even with abscesses.
Notes
Conflict of Interest
No potential conflict of interest relevant to this article was reported.
Author Contributions
Conceptualization: HG, TF, YO, MK. Data curation: HG, YO, TN, TS, KS, SO, KI, KO, YS, HK, KN, KF, YT, MT, IS, MU. Formal analysis: HG. Investigation: HG, YO, TN, TS, KS, SO, KI, KO, YS, HK, KN, KF, YT, MT, IS, MU. Methodology: HG. Project administration: TF, MK. Supervision: TF, MK. Writing–original draft: HG, YO, TN. Writing–review & editing: TF, MK. Final approval of the manuscript: all authors.