Risk factors of epidural hematoma in patients undergoing spinal surgery: a meta-analysis of 29 cohort studies

Abstract

Objective:

The authors conducted this meta-analysis to identify risk factors for spinal epidural haematoma (SEH) among patients following spinal surgery.

Methods:

The authors systematically searched Pub: Med, Embase, and the Cochrane Library for articles that reported risk factors associated with the development of SEH in patients undergoing spinal surgery from inception to 2 July 2022. The pooled odds ratio (OR) was estimated using a random-effects model for each investigated factor. The evidence of observational studies was classified as high quality (Class I), moderate quality (Class II or III) and low quality (Class IV) based on sample size, Egger’s P value and between-study heterogeneity. In addition, subgroup analyses stratified by study baseline characteristics and leave-one-out sensitivity analyses were performed to explore the potential sources of heterogeneity and the stability of the results.

Results:

Of 21 791 articles screened, 29 unique cohort studies comprising 150 252 patients were included in the data synthesis. Studies with high-quality evidence showed that older patients (≥60 years) (OR, 1.35; 95% CI, 1.03–1.77) were at higher risk for SEH. Studies with moderate-quality evidence suggested that patients with a BMI greater than or equal to 25 kg/m² (OR, 1.39; 95% CI, 1.10–1.76), hypertension (OR, 1.67; 95% CI, 1.28–2.17), and diabetes (OR, 1.25; 95% CI, 1.01–1.55) and those undergoing revision surgery (OR, 1.92; 95% CI, 1.15–3.25) and multilevel procedures (OR, 5.20; 95% CI, 2.89–9.37) were at higher risk for SEH. Meta-analysis revealed no association between tobacco use, operative time, anticoagulant use or American Society of Anesthesiologists (ASA) classification and SEH.

Conclusions:

Obvious risk factors for SEH include four patient-related risk factors, including older age, obesity, hypertension and diabetes, and two surgery-related risk factors, including revision surgery and multilevel procedures. These findings, however, must be interpreted with caution because most of these risk factors had small effect sizes. Nonetheless, they may help clinicians identify high-risk patients to improve prognosis.

Introduction

Highlights

• The related risk factors for spinal epidural haematoma following spinal surgery are still controversial.

• We found that in patients undergoing spinal surgery, the occurrence of spinal epidural haematoma was significantly correlated with older age, obesity, hypertension, diabetes, revision surgery and multilevel procedure.

• The purpose of this study is to determine the related risk factors of spinal epidural haematoma after spinal surgery, and to classify the risk factors according to the level of evidence, so as to remind clinicians to take effective intervention measures for the high-risk population of spinal epidural haematoma.

Degenerative spinal disease (DSD) is one of the most common diseases that reduces the quality of human life and includes spinal stenosis, lumbar disc herniation, degenerative scoliosis and lumbar spondylolisthesis¹. The incidence of DSD has increased with global population aging in recent decades^2,3. The latest studies have shown that an estimated 103 million people worldwide are affected by DSD, while in the US, ~600 000 patients with lumbar spinal stenosis undergo operations each year^4–6. For severe DSD, surgical resection and/or fusion are the primary treatment, especially in patients with neurological symptoms^7–9. However, although surgery can effectively alleviate the symptoms of patients in a timely manner, some postoperative adverse events are inevitable. Among these adverse events, spinal epidural haematoma (SEH) is a rare complication caused by nerve root compression after spinal surgery^10–15. Previous studies have reported inconsistent results, and there is no consensus on the risk factors for SEH after spinal surgery.

Multiple factors have been reported to influence the incidence of SEH in patients following spinal surgery, including patient-related risk factors (e.g. sex, older age, obesity, smoking, alcohol consumption, diabetes or hypertension)^16–21, surgery-related risk factors (e.g. operative time, revision surgery, blood loss or multilevel procedure)^16,20–23 and drug-related risk factors (e.g. anticoagulant use)^21,24–26.

To the best of our knowledge, there are no published systematic reviews of risk factors for SEH following spinal surgery. To obtain a sufficient sample size, we conducted a meta-analysis to assess the risk factors for SEH based on the current literature.

Methods

Standard protocol approvals, registrations and patient consent

This study was conducted and reported in accordance with the Cochrane handbook and the PRISMA, Supplemental Digital Content 1, http://links.lww.com/JS9/A707, Supplemental Digital Content 2, http://links.lww.com/JS9/A708 (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 statement²⁷, MOOSE (Meta-Analysis of Observational Studies in Epidemiology)²⁸, and AMSTAR, Supplemental Digital Content 3, http://links.lww.com/JS9/A709 (Assessing the methodological quality of systematic reviews) guidelines²⁹. The MOOSE checklist is reported in eTable 1 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710. The review scheme was registered at PROSPERO, which can be accessed on the official website (https://www.crd.york.ac.uk/prospero).

Table 1.

Characteristics of the studies included in the meta-analysis.

First author	Year	Study design	Region	Observation period	Sample size	Female (%)	Surgery site	Measurements of haematoma
Kou et al.30	2002	Retrospective analysis	USA	1990–2000	416	NR	Lumbar	MRI or CT
Awad et al.14	2005	Retrospective study	USA	1984–2002	14932	48.50	Spinal	NR
Amiri et al.31	2013	Retrospective study	UK	1999–2006	30	36.67	Spine	Frankel grade
O’Neill et al.21	2014	Retrospective study	USA	1995–2012	2392	49.79	Cervical	NR
Yin, G.32	2014	Retrospective study	China	2004–2008	36	50.00	Spine	MRI
Goldstein24	2015	Retrospective study	Canada	2002–2011	529	32.70	Cervical	MRI or CT
Kao et al.19	2015	Retrospective study	China	2002–2010	100	70.00	Lumbar	MRI
Yamada et al.33	2015	Matched case-control study	Japan	1998–2014	8250	46.88	Spine	The surgical database was searched for the terms “evacuation of haematoma” and “wound washout.”
Kotil26	2016	Prospective study	Turkey	2012–2014	115	50.43	Lumbar	MRI
Fujiwara et al.17	2017	Retrospective study	Japan	2002–2015	61	50.82	Lumbar	MRI
Liu et al.34	2017	Retrospective study	China	2010–2016	124	38.71	Lumbar	MRI
Park35	2017	Retrospective study	USA	2012–2013	5280	56.60	Lumbar	NR
Izeki et al.22	2018	Retrospective study	Japan	2012–2016	182	32.42	Lumbar	MRI
Miao et al.20	2018	Retrospective study	China	2006–2012	1258	42.77	Cervical	MRI
Fujita et al.36	2019	Retrospective study	Japan	2008–2017	159	52.20	Lumbar	MRI
Gao et al.37	2019	Retrospective study	China	2012–2017	64	43.75	Spine	Doctor diagnosed and MRI
Tsuge et al.38	2019	Retrospective study	Japan	2000–2017	2611	40.67	Spine	NR
Knusel et al39	2020	Retrospective study	USA	2012–2016	75878	NR	Lumbar	NR
Hohenberger et al.40	2020	Retrospective study	Germany	2002–2016	168	52.38	Spine	MRI or CT
Kim et al.25	2020	Retrospective study	Korea	2015–2018	206	51.90	Lumbar	MRI
Wang et al.41	2020	Retrospective cohort study	China	2013–2020	9258	NR	Lumbar	The database search keywords were“spinal epidural haematoma”and“delayed–onset spinal epidural haematoma”
Masuda et al.23	2020	Retrospective study	Japan	2002–2012	10680	NR	Spine	MRI
Park35	2020	Retrospective study	Korea	2014.1–2014.12	17549	49.74	Spine	NR
Abola et al.16	2021	Retrospective study	USA	2012–2016	53233	48.40	Cervical	NR
Ahn et al.10	2021	Retrospective case-controlled study	Republic of Korea	2015–2019	236	55.70	Spinal	MRI
Snopko et al.42	2021	Prospective analysis	Martin	2016–2018	371	49.60	Lumbar	CT
Aikeremu43	2021	Retrospective study	China	2010–2020	3717	50.00	Lumbar	MRI
Wang et al.44	2022	Retrospective study	China	2010–2019	75	41.30	Thoracic	MRI
Xia et al.45	2022	Retrospective study	China	2009–2019	18220	33.33	Cervical	Radiological evidence
First author	Haematoma definition	Outcomes	Follow-up period	Funding source	Adjusted variables
Kou et al.30	NR	Multilevel procedures, presence of a preoperative coagulopathy.	120 months	NR	NR
Awad et al.14	NR	Male, age, tobacco use, hypertension, diabetes mellitus, anticoagulation use, revision surgery	NR	NR	Age, body mass index, perioperative durotomy and the use of drains
Amiri et al.31	NR	Regular alcohol use, multilevel surgery, and revision surgery	NR	NR	NR
O’Neill et al.21	NR	Male, age, comorbidity, diabetes, tobacco use, spondylosis, disc herniation, deformity, DISH, OPLL, radiculopathy, myeloradiculopathy, prior surgery, anterior/posterior, operative time, ACDF, corpectomy, inst. placement, inst. removal, levels, autograft, DBM, BMP	12 months	NR	NR
Yin32	NR	Hypertension, Tobacco use, diabetes	14 months	NR	NR
Goldstein24	NR	Increased CCI and postoperative NSAID use.	1 month	No funding	Cardiac arrest, stroke. deep vein thrombosis, surgical site infection, pneumonia
Kao et al.19	NR	Diastolic pressure, postoperative drainage tube output, gelatine sponge	NR	NR	Age and sex
Yamada et al.33	NR	A 50 mm Hg or greater increase in systolic blood pressure after extubation and high body mass index and hypotensive anaesthesia	9.2 months	NR	Age, sex, segment level, herniation type, or disease duration
Kotil26	The maximum thickness of the SEH was measured on the axial slice using MRI and graded as none (<1 mm), minimal (1–1.9 mm), moderate (2–2.9 mm), or prominent (>3 mm).	CSD	NR	NR	Demographics, preoperative laboratory values, ASA classification, medical comorbidities, and surgical characteristics.
Fujiwara et al.17	NR	Hypertension, age, sex, BMI, coagulation status, anticoagulant drugs, intraoperative blood loss, and operation time	NR	NR	NR
Liu et al.34	NR	Female, age, hypertension, diabetes mellitus, white blood cells, platelet, total protein, haemoglobin, albumin, globulin, serum calcium, serum potassium, serum glucose, glycated haemoglobin, prothrombin time, APTT, TT, FIB, blood type	3 months	NR	Ticlopidine, aspirin and surgery type
Park35	NR	Female, male, BMI, wound classification	NR	National Research Foundation of Korea	Local kyphosis angle, occupying rate of cross-sectional, cerebrospinal fluid leakage
Izeki et al.22	The characteristic MRI findings of an SEH are areas of abnormal signal intensity within the dural sac.	Lumbar decompression surgery, anticoagulant therapy, anticoagulant therapy	<12 months	NR	NR
Miao et al.20	NR	Male, Age, BMI, Spondylosis, OPLL, Disc herniation, Operative duration, Levels, Intraoperative blood loss, Operation interval, Drainage, mJOA	NR	NR	NR
Fujita et al.36	NR	Lumbar hypolordosis or multilevel stenosis, patient characteristics, coagulation status, preoperative radiographic parameters, JOA score, VAS, BP, perioperative factors	12 months	NR	Sex, age, body weight, diabetes, hypertension, blood type, use of anticoagulants, INR, platelet count, DBP, SBP and ASA
Gao et al.37	NR	Medical comorbidities, Frankel grade,	NR	No funding	NR
Tsuge et al.38	NR	Sharp elevation of systolic blood pressure at extubation	NR	No funding	NR
Knusel et al39	NR	Age, BMI, dural repair	NR	NR	NR
Hohenberger et al.40	NR	Anticoagulants, impaired coagulation, confirmed by routine preoperative laboratory testing, and smoking	6 months	NR	Age, sex, BMI, comorbidities, anticoagulant therapy anticoagulant and/or antiplatelet medication), and haematological coagulation parameters
Kim et al.25	PSEH was defined as haematoma compressing the dural sac in MRI T2-weighted axial images.	GTMS	NR	NR	Age, anticoagulant therapy, or surgery location.
Wang et al.41	NR	Postoperative SBP and previous spinal surgery at the same level	NR	NR	All the mentioned patient-related, tumour-related, and treatment-related factors
Masuda et al.23	NR	Laminoplasty/laminectomy, Posterior decompression and fusion, anterior decompression and fusion, posterior decompression, posterior decompression and fusion, posterior decompression, posterior decompression and fusion	1 month	NR	Age, sex, and diagnosis between SEH and control groups
Park35	NR	Age, sex, total medical payment, infections, diabetes, hypertension.	NR	National Research Foundation of Korea	NR
Abola et al.16	NR	Age, Male, ASA, hypertension, respiratory, bleeding disorder, prolonged operative time, number of levels, posterior segmental instrumentation, revision surgery, dural repair, perioperative transfusion	NR	No funding	Age, sex and antiplatelet
Ahn et al.10	POSEH compresses the sheath sac on T2 axis images.	BESS	114 months	NR	NR
Snopko et al.42	NR	Obesity	40 months	NR	NR
Aikeremu43	NR	Revision surgery, Use of haemostatic material, Age	NR	No funding	Factors associated with reoperations due to SEH underwent univariate analysis for comparisons between the SEH group and the control group.
Wang et al.44	NR	Local kyphosis angle, occupying rate of cross-sectional, cerebrospinal fluid leakage	3 months	No funding	NR
Xia et al.45	NR	Male, Age, BMI, Hypertension, Diabetes mellitus Smoking, Revision surgery	10.2 months	No funding	NR

ACDF, anterior cervical discectomy and fusion; APTT, activated partial thromboplastin time; ASA, American Society of Anesthesiologists; BESS, bipolar endoscopic spine surgery; BMP, bone morphogenetic protein; BP, blood pressure; CCI, Charlson Comorbidity Index; CSD, closed suction drains; CSS, conventional spine surgery; CT, computed tomography; DBM, demineralized bone matrix; DBP, diastolic blood pressure; EH, epidural haematomas; DISH, diffuse idiopathic skeletal hyperostosis; FIB, fibrinogen; GTMS, gelatin-thrombin matrix sealant; INR, international normalized ratio; mJOA, modified Japanese Orthopedics Association scores; MRI, magnetic resonance imaging; NR, not reported; NSAID, nonsteroidal anti-inflammatory drug; ODI, Oswestry Disability Index; OPLL, ossification of the posterior longitudinal ligament; PSEH, postoperative spinal epidural haematoma; SBP, systolic blood pressure; SEDH, spinal epidural haematoma; SEH, spinal epidural haematoma; SSEH, symptomatic spinal epidural haematoma; SSH, spinal subdural haematoma; TT, thromboplastin Time; VAS, visual analog scale.

Search strategy

Two independent authors conducted a systematic literature search of articles published in PubMed, Embase and the Cochrane Library from inception to 2 July 2022. These three databases were inclusive of most of the possible articles related to our research topic, regardless of language and reported risk factors for postoperative SEH in spinal surgery patients; there were no language or publication data constraints. Medical Subject Heading (MeSH) terms were used in PubMed and the Cochrane Library, and Subject Headings (Emtree) combined with free text words (including synonyms and closely related words) related to spinal surgery and haematoma were used in the Embase database. The search strategy included the search terms ‘spinal disease’ or scoliosis or ‘spinal fusion’ and (haematoma or ‘treatment failure’ or ‘treatment outcome’) (eTable 2 in Supplement 1, Supplemental Digital Content 4, http://links.lww.com/JS9/A710). We also manually searched the references of relevant systematic reviews for potential additional qualifying studies.

Selection criteria

Two authors independently reviewed the title and abstract of each article and then cross-checked to determine whether studies met the PECOS criteria (participant, exposure, comparator, outcome and study design) for inclusion. Any disagreement was resolved through discussion with a senior author.

1. Participants: Patients aged older than or equal to 18 years undergoing spinal surgery.

2. Exposure: Risk factors associated with the incidence of SEH following spinal surgery, such as sex, age, obesity, tobacco use, diabetes, hypertension, operative time, revision surgery, blood loss or multilevel procedure.

3. Comparator: comparison group with lower exposure or no exposure to a modifiable risk factor.

4. Outcome: risk of SEH after spinal surgery, presented as the odds ratio (OR) with the corresponding 95% CI.

5. Study design: prospective or retrospective cohort study.

We excluded conference abstracts, conference papers, reviews and meta-analyses, and duplicate reports. We did not search grey literature or any unpublished materials. We also excluded studies that did not report risk estimates or had insufficient data to assess the risk of SEH following spinal surgery.

Data extraction

Two authors used a predesigned Excel data sheet (Microsoft Corp.) to screen and extract data from the relevant studies. Conflicts were resolved through discussion and consensus. Data on the following characteristics were extracted: first author, publication year, study design, geographical region, observation period, sample size, percentage of female participants, surgery site, measurement and definition of haematoma, risk factors, reported OR and 95% CI, follow-up period and adjusted variables.

Quality assessment

Each qualified study was independently evaluated by the two authors using the Newcastle–Ottawa Scale (NOS)⁴⁶, which includes three domains, including patient representation, exposure and outcome determination, and follow-up adequacy, with an overall score of 9 for each study. NOS scores of 0–5, 6–7 and 8–9 (low risk of bias) indicated low, moderate and high quality⁴⁷.

Evaluation of the strength of evidence

The strength of the evidence in the identified associations was graded using a set of modified criteria for observational cohort studies⁴⁸ (eTable 4 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710). When Egger’s P value was greater than 0.1, the total sample size was greater than 1000, and between-study heterogeneity was low (I²<50%), we considered the study to have Class I (high quality) evidence. When two of the three conditions were met, the study was considered to have Class II (moderate quality) evidence. When only one of these three conditions was met, Class III (moderate quality) evidence was indicated. When none of these three conditions were met, Class IV (low quality) evidence was suggested.

Statistical analysis

The meta-analysis was conducted in accordance with the Cochrane handbook for systematic reviews of interventions. The ORs and 95% CIs were calculated for each potential risk factor for SEH among patients following spinal surgery. All analyses were performed using Stata software (Stata version 12.0). A random-effects model was applied for studies in which I² exceeded 50%⁴⁹, given the expected interstudy heterogeneity in participant characteristics (e.g. age, geographical region, surgical site or follow-up period), exposure variables, outcome measures and definitions (e.g. the definition of spinal epidural haematoma and its measurement), as well as study design (prospective or retrospective). Forest plots were used to display individual-study ORs and the pooled OR. There were some differences in study baseline data, leading to heterogeneity between studies. Therefore, the Cochrane Q test and I² test were used to assess heterogeneity between studies, and when I² was greater than or equal to 50% or P was less than 0.05, the heterogeneity was considered statistically significant⁵⁰. To explore the sources of between-study heterogeneity, we conducted multiple subgroup analyses of the outcomes; if I² was greater than or equal to 50%, a sufficient number of included studies (≥15) was provided, and the number of studies in the group was greater than or equal to3. Subgroup analyses were conducted and stratified by average participant age at surgery (≤60 years or > 60 years), study quality (low or high) and study region (USA, China or Japan); participants were matched by age and sex (yes or no), average follow-up period (≤3 months or > 3 months), surgical site (lumbar spine or spine) and surgery type (decompression or other). We used the change in the pooled OR and 95% CI to evaluate the stability of our results and the decrease in I² in the subgroups (≥30%) to determine the potential source of heterogeneity. Sensitivity analyses were performed to assess the stability of the results by sequentially omitting each study and meta-analyzing the estimates of the other studies. We examined publication bias using Egger’s test for each risk factor to determine the correlation between the effect estimates and their variances, with a P value of less than 0.1 indicating a significant difference⁵¹. For all statistical tests, P less than 0.05 was considered statistically significant.

Results

Literature search

A total of 21 791 studies were identified through a systematic literature search, of which 4303 duplicate records were excluded, and 17 430 irrelevant studies were excluded after screening their titles and abstracts. Next, 58 potentially relevant studies were selected for a full-text review, and we excluded 29 studies that did not report patient outcome data, nonpopulation-based cohorts, meta-analyses, and case reports. Ultimately, 29 cohort studies involving 150 252 participants (mean sample size 5181) met the inclusion criteria for the meta-analysis (Fig. 1 and eTable 5 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710).

Figure 1.

Flowchart of study selection.

Study characteristics

The baseline characteristics of the included studies are shown in Table 1. The incidence of SEH following spinal surgery ranged from 1 to 5%. All studies were published between 2002 and 2022 and were conducted in China (n=9), Japan (n=6), the United States (n=6), Korea (n=3), Britain (n=1), Canada (n=1), Germany (n=1) and Turkey (n=1). A total of 65.5% (19/29) of the studies had an NOS score of greater than or equal to 8 (Fig. 2 and eTable 3 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710).

Figure 2.

Methodological quality assessment of the included studies based on the Newcastle–Ottawa Scale (NOS) tool.

The median sample size of the included studies was 7169. Fifteen studies adjusted for major confounding factors such as age, sex and obesity for multivariate analysis.

The recurrence rates of SEH ranged from 0.2 to 39.3%, and the pooled recurrence rate was 0.7% (95% CI, 0.5–0.9%), with significant heterogeneity across studies (I²=95.4%, P<0.001) (eFigure. 1 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710). Additionally, recurrence rates were significantly different when stratified by some baseline study-level factors (almost all P<0.001) (eTable 6 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710).

Risk factors and strength of evidence for SEH

Our study included the effects of patient-related risk factors and surgery-related risk factors (Fig. 3 and eFigure 2–11 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710) on the risk of SEH following spinal surgery. Studies with high-quality (Class I) evidence showed that older patients (≥60 years) were at higher risk for SEH. Studies with medium-quality (Class II) evidence showed that patients with obesity and hypertension and those undergoing diabetes revision surgery and multilevel procedures were at higher risk for SEH. Studies with high-quality or moderate-quality (Class I or II) evidence revealed no association between tobacco use, ASA classification, operative time or anticoagulant use and SEH (Table 2 and eTable 7 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710).

Figure 3.

Meta-analyses of the association between patient-related risk factors and surgery-related risk factors. ASA, American Society of Anesthesiologists; OR, odds ratio.

Table 2.

Nonsignificant and significant risk factors associated with postoperative spinal epidural haematoma.

Significant factors	No. studies	No. patients	SEH OR (95% CI)	I2,%	P	Egger’s test P
Age (years)
<50			Ref.
>50	10	58083	1.35 (1.03–1.77)	50.30	0.034	0.03
Hypertension
No			Ref.
Yes	18	34244	1.67 (1.28–2.17)	71.30	0.000	0.00
Diabetes
No			Ref.
Yes	14	11684	1.25 (1.01–1.55)	0.00	0.038	0.05
Obesity (kg/m²)
<24–25			Ref.
>24–25	7	31174	1.39 (1.10–1.76)	52.00	0.006	0.13
Revision surgery
No			Ref.
Yes	10	4284	1.92 (1.15–3.25)	51.00	0.013	0.27
Multilevel procedure
No			Ref.
Yes	5	330	5.20 (2.89–9.37)	5.20	0.000	0.01
Non-significant factors	No. studies	No. patients	SEH OR (95% CI)	I², %	P value	Egger’s test P value
Anticoagulants
No			Ref.
Yes	12	35260	2.57 (0.59–11.14)	96.20	0.209	0.49
Tobacco use
No			Ref.
Yes	9	16271	1.43 (0.99–2.08)	42.20	0.057	0.69
ASA classification
1			Ref.
2–5	4	28816	1.19 (0.86–1.66)	9.70	0.291	0.07
Operative time (min)
<120			Ref.
>120	4	4867	1.43 (0.86–1.40)	36.80	0.171	0.95

ASA, American Society of Anesthesiologists; OR, odds ratio; Ref., reference group; SEH, spinal epidural haematoma.

Patient-related risk factors

Hypertension

A total of 18 studies^{14,16,17,19,22,31–38,40,41,45,52} showed that having hypertension was associated with a higher risk for postoperative SEH. Patients with hypertension had a 1.67-fold higher risk of postoperative SEH than those without hypertension (OR, 1.67; 95% CI, 1.28–2.17, P<0.001). Significant heterogeneity was found (I²=71.30%, P<0.001). Heterogeneity was found to be significantly reduced in subgroup analyses stratified by region, surgery site and average age, indicating that these factors could be potential sources of heterogeneity. The results of the subgroup analysis showed that the OR of hypertensive patients under 60 years old was 1.54 (95% CI: 1.18–2.01, I²=67.0%, P<0.001), while that of patients over 60 years old was 4.11 (95% CI: 1.79–9.44, I²=44.8%, P=0.164). However, the subgroup analysis based on surgery site showed that the OR was 2.23 (95% CI: 1.21–4.12, I²=82.4%, P<0.001) for patients undergoing lumbar surgery, 1.71 (95% CI: 1.04–2.83, I²=59.9%, P=0.015) for those undergoing spinal surgery and 1.13 (95% CI: 0.85–1.51, I²=0.0%, P=0.347) for those undergoing cervical surgery (eTable 8, Supplemental Digital Content 4, http://links.lww.com/JS9/A710 and eFigure 3 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710).

Diabetes

A total of fourteen studies^{14,16,19,21,22,31,32,34–36,38,44,45,52} reported an association between diabetes and postoperative SEH. In four studies^21,36,44,45, symptomatic SEH occurred 1–10 days after surgery; in five studies^{16,19,22,31,32}, SEH occurred between 1 and 6 months after surgery; and the remaining five studies^{14,34,35,38,52} did not specify when SEH occurred. All eleven articles^{14,16,19,22,32,34–36,38,44,45,52} were multivariate logistic regression analyses. Our results showed that diabetes is a risk factor for postoperative SEH (OR, 1.25; 95% CI, 1.01–1.55, P=0.038). Heterogeneity was considered insignificant (I²=0.00%, P=0.944).

Age

A total of ten studies^{16,22,31,33,38–40,42,43} reported an association between older age (≥60 years) and postoperative SEH. In four studies^33,39,40,43, symptomatic SEH occurred 1–15 days after surgery; in three studies^16,22,31, SEH occurred between 1 and 6 months after surgery; and the remaining three studies^38,42,52 did not specify when SEH occurred. Multivariate logistic regression analysis was performed in all 8 studies^{16,22,33,38–40,43,52}. Meta-analysis showed that older age was an important risk factor for postoperative SEH (OR, 1.35; 95% CI, 1.03–1.77, P=0.034). Heterogeneity was considered moderately insignificant (I²=50.30%, P=0.034) (eFigure 2 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710).

Obesity

Seven studies^{16,22,33,35,38–40} noted an association between obesity and postoperative SEH. In three studies^33,39,40, symptomatic SEH occurred within 1–15 days after surgery; in three studies^16,22,35, SEH occurred within 1–6 months after surgery; and the study by Tsuge et al. ³⁸. did not specify when SEH occurred. Only one study³⁸ did not involve a multivariate logistic regression analysis. Our results showed that obesity was an important risk factor for postoperative SEH (OR, 1.39; 95% CI, 1.10–1.76, P=0.006). Heterogeneity was considered moderately insignificant (I²=52.0%, P=0.052) (eFigure 6 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710).

Surgery-related risk factors

Revision surgery

Ten studies^{14,16,19,21,22,31,33,37,43,45} reported a relationship between revision surgery and postoperative SEH. In four studies^{21,33,39,43,45}, symptomatic SEH occurred 1–15 days after surgery; in three studies^22,31,43, SEH occurred between 1 and 6 months after surgery; and the studies by Kao et al. ¹⁹ and O’Neill et al. ²¹ did not specify when SEH occurred. Multivariate logistic regression analyses were performed in all six articles^{22,31,33,39,43,45}. Our study showed that revision surgery was closely related to the occurrence of SEH after spinal surgery (OR, 1.92; 95% CI, 1.15–3.25, P=0.013). Heterogeneity was considered moderately insignificant (I²=51.0%, P=0.031) (eFigure 9 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710).

Multilevel procedures

Five articles^{22,30,40,41,43} described the relationship between multilevel procedures and postoperative SEH (OR, 5.20; 95% CI, 2.89–9.37, P=0.006). In four studies^22,40,41,43, symptomatic SEH occurred within 10 days, but the study by Kou et al. ³⁰. did not describe the time of SEH occurrence. All studies were subjected to multiple logistic regression analyses. The heterogeneity was insignificant (I²=27.9%, P=0.236) (eFigure 10 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710).

Publication bias and sensitivity analyses

As mentioned in the previous statistical analysis, potential publication bias was assessed regarding the mentioned factors in more than 10 studies. Therefore, we performed tests for bias regarding hypertension, diabetes, older age and revision surgery. The results of both Begg’s test and Egger’s test are shown in eTable 7 in the supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710, and the funnel plot is shown in eFigure 11–21, Supplemental Digital Content 4, http://links.lww.com/JS9/A710 in the supplement. We used leave-one-out sensitivity analysis to evaluate the stability of the results for each investigated factor. The results showed that the pooled ORs all remained similar across analyses for both patient-related risk factors and surgery-related risk factors (eTable 9 in the Supplement, Supplemental Digital Content 4, http://links.lww.com/JS9/A710).

Discussion

Principal findings

The purpose of this meta-analysis was to provide evidence for predictors of SEH risk in patients undergoing spinal surgery. By pooling 29 cohort studies, we identified four patient-related risk factors, including older age (OR, 1.35; 95% CI, 1.03–1.77), obesity (OR, 1.39; 95% CI, 1.10–1.76), hypertension (OR, 1.67; 95% CI, 1.28–2.17) and diabetes (OR, 1.25; 95% CI, 1.01–1.55), and two surgery-related risk factors, including revision surgery (OR, 1.92; 95% CI, 1.15–3.25) and multilevel procedures (OR, 5.20; 95% CI, 2.89–9.37), with different levels of evidence.

The meta-analysis also analyzed patient-related risk factors, including tobacco use and ASA classification, which had no significant correlation with the risk of SEH following spinal surgery. Despite the above results, our study cannot rule out these potential factors, which have been revealed to be associated with the risk of postoperative SEH in a number of studies^20–25. Therefore, future large prospective studies are still required to confirm the current findings.

Comparisons with previous literature and potential mechanisms

Our study identified multiple patient-related risk factors for SEH, including older age, obesity, hypertension and diabetes. In previous studies, Usubiaga et al. ⁵³ showed that SEH was more likely to occur in elderly patients with the same haematoma type due to higher pressure and more severe compression in the epidural space. Both older age and obesity can reduce haemostasis through mechanisms such as malnutrition or diminished physiological function^54,55. Obesity is an important risk factor for postoperative symptomatic SEH. Obese patients usually require a special posture or a more extensive surgical approach, leading to more soft tissue injury, increased blood loss and a prolonged operation time⁴².

In addition, another study showed that hypertension is more likely to lead to rebleeding after surgical incision¹⁷. However, surgeons fail to act promptly, resulting in the formation of a haematoma. Kao and colleagues found that diastolic blood pressure is related to the formation of an SEH and speculated that excessive diastolic blood pressure is an important factor leading to an increase in whole blood viscosity^56–58, which can easily lead to thrombosis and drainage dysfunction^19,41, causing venous blood to ooze and accumulate to form a haematoma. Careful haemostasis during the operation and the placement of a negative pressure suction device after the operation can reduce the incidence of SEH. The mechanism for the observed increase in the diabetes mellitus–associated risk for postoperative SEH is not clear and may be caused by multiple coexisting diseases. However, our study revealed that tobacco use (OR, 1.43; 95% CI, 0.99–2.08) had no significant relationship with the risk of SEH. In contrast, Hohenberger et al. ⁴⁰ found that smoking is an independent risk factor for the development of SEH, and smoking can generate a hypoxic tissue environment, leading to delayed wound healing, cellular dysfunction and thrombosis⁵⁶. Therefore, large prospective studies are needed to further verify these findings.

Among the surgery-related risk factors, we found that both revision surgery and multilevel procedures led to a significantly increased incidence of postoperative SEH. Aono et al. ⁵⁹ found that the incidence of SEH after revision surgery was 0.41%, and the loss of normal anatomy due to previous surgery and scar tissue formation affected the evaluation and management of active bleeding^60,61. Therefore, spinal surgeons should minimize bleeding and ensure timely haemostasis during surgery to reduce the occurrence of SEH. In the study by Fujita et al. ³⁶, patients undergoing multilevel procedures had a higher incidence of symptomatic SEH (3.0%) than those reported in previous studies (0.5–1%). This may be due to the increased risk of bleeding and haematoma associated with more vascular damage during multilevel fusions⁴³. The use of negative pressure drainage devices to remove blood and other fluids that may accumulate in the surgical area can greatly reduce the incidence of postoperative SEH¹⁹.

Implications for clinical practice and future studies

The current study revealed significant future clinical implications regarding risk factors for SEH in patients following spinal surgery. Risk factors for SEH should not be overlooked because these specific variables can help to identify patients at higher risk of developing SEH, and early interventional strategies should be taken to reduce their risk. Fujita et al. ³⁶ showed that preoperative blood pressure control may help to reduce the incidence of SEH. In the long term, these findings will benefit clinicians in improving the preoperative risk assessment for SEH, and large prospective cohort studies are warranted to confirm these results.

Strengths

The current study has the following strengths. First, to the best of our knowledge, this is the first meta-analysis related to this topic. It provides the latest and most comprehensive evidence of risk factors for SEH following spinal surgery, including older age, hypertension, diabetes, obesity, revision surgery and multilevel procedures. Second, we used MeSH/Emtree terms and free text words to conduct a comprehensive literature search of the three main databases, including PubMed, the Cochrane Library and Embase, and to formulate a comprehensive database search strategy without date and language restrictions. In this way, original literature meeting the inclusion criteria could be found as much as possible, avoiding the influence of publication bias on the combined results and improving the reproducibility of the results. Third, we evaluated the correlation intensity of each risk factor (from Class I to Class IV) based on the sample size, Egger’s test, P values and interstudy heterogeneity, which may be helpful for surgeons in providing early clinical intervention. Finally, we used the trim-and-fill technique to adjust the combined estimation according to the publication bias, and the results remained consistent with our analysis.

Limitations

Several potential limitations to the present study should also be considered. First, we found some heterogeneity in the hypertension results across studies, as we expected, possibly due to the nonstandardization of how hypertension was defined and measured and differences in the baseline characteristics of the study cohorts. However, to explore the sources of heterogeneity, we conducted multiple subgroup analyses and sensitivity analyses, and the adjusted results were consistent with the original results. Second, to our knowledge, there is no uniform international definition of SEH, and the current meta-analysis included studies of radiological SEH or SEH with symptoms of spinal cord or nerve root compression. Therefore, the heterogeneity between studies may have been increased, and the accuracy of the results may have been affected. Third, our data sources are based on cohort studies, and thus, we cannot infer a causal relationship between epidural haematoma after spinal surgery and hypertension, diabetes, revision surgery, blood loss or multilevel procedures. We also found that the effects of certain risk factors were estimated near the border with confidence intervals between 0.90 and 1.10 (e.g. smoking), and large prospective cohort studies are needed to validate these findings.

Conclusions

The current meta-analysis revealed obvious risk factors for SEH, including four patient-related risk factors (older age, obesity, hypertension and diabetes) and two surgery-related risk factors (including revision surgery and multilevel procedures). These findings, however, must be interpreted with caution because most of these risk factors had small effect sizes. Nonetheless, they may help clinicians identify high-risk patients to improve prognosis.

Ethical approval

Not applicable.

Source of funding

This work was supported by Scientific Research Project of Hunan Provincial Health Commission (no. 202204074707), Scientific Research Project of Wuhan City Health Commission (no.WX18C29), and the Natural Science Foundation of Hunan Province (no. 2022JJ30516), Hubei Provincial Natural Science Foundation of China (no. 2022CFB002).

Role of the funder/sponsor

The funder of the study had no role in the study design, data collection, data analysis, data interpretation or writing of the manuscript. The corresponding author had full access to all the data in the study and has final responsibility for the decision to submit for publication.

Author contribution

Study concept and design: Z.X. and Z.M.; Acquisition of data: M.L., Q.C., Z.Z., Q.Y., X.Z., Z.W. and Y.S.; Analysis and interpretation of data: H.W., Y.Y., J.C., B.Z. and G.Y.; Drafting of the manuscript: M.L. and Z.M.; Critical revision of the manuscript for important intellectual content: all authors; Study supervision: M.L., Z.X. and Z.M.

Conflicts of interest disclosure

The authors declare no potential conflicts of interest.

Research registration unique identifying number (UIN)

The review protocol was registered in PROSPERO. Unique Identifying Number (UIN) is “CRD42022343842”. Hyperlink to the specific registration (must be publicly accessible and will be checked): “https://www.crd.york.ac.uk/PROSPERO/display_record.php?RecordID=343842”.

Guarantor

Zhihong Xiao, Zubing Mei.

Date statement

This is a summary design study. Data used for meta-analysis were extracted from previously published papers.

Provenance and peer review

Not commissioned; externally peer-reviewed