Systematic Review and Meta-Analysis Medicine ® OPEN Thromboprophylaxis in elective spinal surgery A protocol for systematic review ∗ María J. Colomina, MD, PhDa, , Joan Bagó, MD, PhDb, Javier Pérez-Bracchiglione, MDc, Maria Betina Nishishinya Aquino, MD, MScd, Karla R. Salas, MD, MScd, Carolina Requeijo, MD, MScd, Gerard Urrutia, MD, PhDe Abstract Background: Venous thromboembolism (VTE) is a serious, sometimes life-threatening complication that can occur following spine surgery. The incidence of VTE, and the optimal type and timing of thromboprophylaxis for this complication in elective spine surgery is a matter of debate. Objective: To perform a systematic review with the aim of clarifying the efficacy and adverse effects of mechanical and chemical prophylaxis for preventing thromboembolic complications in elective spine surgery for conditions other than trauma and malignant disease. Methods/design: A search strategy of related articles up to March 2018 was designed and executed in Medline and Embase. Patients: adolescents (>10 years) and adults undergoing elective surgery for spinal deformity or degenerative disease (from C1 to S1). Intervention: Perioperative mechanical and chemical thromboprophylaxis. Studies could be randomized controlled trials or observational studies that reported data on any relevant clinical outcomes. Results: In total, 2451 uniquecitations were identified and 35 studies were ultimately included in the systematic review. The overall mean incidence of complications was 3.7% for deep venous thrombosis, 0.0% for pulmonary embolism, and 3.7% for bleeding in chemoprophylaxis group; 2.9% for deep venous thrombosis, 0.4% for pulmonary embolism and 0.0% for bleeding in mechanoprophylaxis; and 0.7% for deep venous thrombosis, 0.1% for pulmonary embolism and 0.2% for bleeding in mixed prophylaxis group with no specific data on these rates for the type of patient and type and location of surgery. None of the articles retrieved provided information on the adolescent population. Discussion and Conclusions: The poor design and high variability among the studies regarding characteristics of study population, details of interventions, and definitions of outcomes, determines a low quality of the available evidence and limits the interpretation of the results. We were unable to identify a clear advantage of one type of thromboprophylaxis over the other, although there was an increased risk of bleeding with chemoprophylaxis, which could favor the use of mechanoprophylaxis in this scenario. Abbreviations: ACCP = American College of Chest Physicians, CS = Compression Stockings, DVT = Deep Venous Thrombosis, EDH = Epidural Hematoma, LDUH = Low-Dose Unfractionated Heparin, LMWH = Low Molecular Weight Heparine, NASS = North American Spine Society, PE = Pulmonary Embolism, RCTs = Randomized Controlled Trials, SPC = Sequential Pneumatic Compression Devices, VTE = Venous Thromboembolism. Keywords: chemoprophylaxis, deep vein thrombosis, epidural hematoma, mechanoprophylaxis, pulmonary embolism, spine surgery, thromboprophylaxis Editor: Johannes Mayr. Compliance with ethical standards: This study does not contain any animal or human participants. Informed consent: Not applicable. The authors have no funding information to disclose. The authors have no conflicts of interest to disclose. Supplemental Digital Content is available for this article. All data generated or analyzed during this study are included in this published article [and its supplementary information files]. a Department of Anaesthesiology and Intensive Care, Hospital Universitari Bellvitge, L’Hospitalet de LLobregat, Barcelona, Universitat de Barcelona, b Spine Surgery Unit, Department of Orthopedic and Trauma Surgery, Hospital Universitari Vall d’Hebron, Barcelona, Universitat Autònoma de Barcelona, Barcelona, Spain, c Interdisciplinary Centre for Health Studies (CIESAL), Universidad de Valparaíso, Chile, d Iberoamerican Cochrane Centre, Sant Pau Biomedical Research Institute (IIB Sant Pau), e Iberoamerican Cochrane Centre, Sant Pau Biomedical Research Institute (IIB Sant Pau), CIBERESP, Barcelona, Spain. ∗ Correspondence: María J. Colomina, Department of Anaesthesiology and Intensive Care, Hospital Universitari Bellvitge, C/Feixa Llarga s/n, L’Hospitalet de Llobregat, 08907, Barcelona, Spain (e-mail: mjcolomina@bellvitgehospital.cat, mjcolomina@gmail.com). Copyright © 2020 the Author(s). Published by Wolters Kluwer Health, Inc. This is an open access article distributed under the Creative Commons Attribution License 4.0 (CCBY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. tia G. Thromboprophylaxis in elective spinal surgery: a How to cite this article: Colomina MJ, Bagó J, Pérez-Bracchiglione J, Aquino MB, Salas KR, Requeijo C, Urru protocol for systematic review. Medicine 2020;99:21(e20127). Received: 8 January 2020 / Received in final form: 31 March 2020 / Accepted: 1 April 2020 http://dx.doi.org/10.1097/MD.0000000000020127 1 Colomina et al. Medicine (2020) 99:21 Medicine included were grouped according to the type of spine disease: Highlights degenerative, deformity, trauma, infection, and malignant disease. There was a higher risk of DVT in spine surgery for  There is no clinically difference between chemoprophy- non-traumatic causes (6%, range 0–19%) and for deformity laxis and mechanoprophylaxis regarding VTE or PE (5.3%, range 2–14%), and a lower risk in surgery for incidence. degenerative conditions (2.3%, range 0–9%). The overall risk  The appropriate time for the start of postoperative of DVT in elective surgery without pharmacological prophylaxis anticoagulation has not yet been determined. was 1% to 2%. The risk of major bleeding ranged from 0% to  The use of chemoprophylaxis is related to the occurrence 4.3% with use of a variety of anticoagulant agents. This study of bleeding complications. was unable to establish the safest time point to initiate  There is a lack of information in the adolescent coagulation in the postoperative period, as none of the studies population (under 18 years of age) on the incidence of had defined this objective, nor the most appropriate treatment VTE or PE or the effectiveness of thromboprophylaxis. option. Some of the conclusions were derived from a small  This systematic review confirms the low quality of number of studies; hence, the authors considered that the quality research in this field. of the evidence for the associated risk of DVT in the different types of spine surgery was of low or very low level, particularly in the case of elective surgery. Despite these limitations, the conclusions suggested that patients undergoing elective spine surgery could undergo the procedure without the use of 1. Introduction pharmacologic prophylaxis. A more recent systematic review Venous thromboembolism (VTE), including deep venous carried out by Oliveira et al[13] reported average overall rates of thrombosis (DVT), and pulmonary embolism (PE), is a serious 3.55% and 1.04% for DVT and PE, respectively, in all types of and potentially fatal complication that can occur in the spine surgery. The authors stated that it was not possible to postoperative period following spine surgery.[1] DVT and PE indicate a standard thromboprophylaxis for spine surgery, as can have both been extensively studied in several types of orthopedic be done for hip and knee surgeries, because of the heterogeneous and trauma surgeries. However, data on the incidence and impact nature of the studies. Additionally, chemoprophylaxis may not be of VTE in spine surgery is more limited, and currently, there is no safe due to the possibility of bleeding complications. Therefore, it clear consensus on what constitutes optimal thromboprophylaxis falls to the spinal surgeon to analyze the risk of DVT/PE in these procedures.[2] This is particularly true for elective spine individually for each specific case. The authors emphasized the surgery performed for conditions other than trauma or a importance of symptomatic EDH, which often shows rapidly malignant disease.[1,3,4] progressing neurological deterioration and can be associated The documented incidence of VTE in spine surgery varies widely with coagulopathy induced by the administration of anti- from 0.3% to 31%, depending on the population studied and the coagulants. Around 4% of patients who received systemic methods used to detect this complication.[5] It is likely that this LMWH prophylaxis developed bleeding complications, com- considerable divergence is the result of including patients with pared with the 0.5% to 2.5% of patients who developed DVT different conditions (e.g., degenerative spinal disease, deformity, following spinal surgery, making this a substantial disincentive traumatic injury, and tumors) affecting different parts of the spine for chemoprophylaxis in these procedures, although this was not and treated with heterogeneous surgical techniques.[6,7] specifically described in the review.[15,16] The risk–benefit relationship of thromboprophylaxis in this On the other hand, the North American Spine Society (NASS) setting should also be considered. When deciding on the type and and American College of Chest Physicians (ACCP) have timing of thromboprophylaxis, the risk of postoperative bleeding published guideline for VTE prophylaxis in non-traumatic, and epidural hematoma (EDH) must be carefully weighed against non-neoplastic elective spine surgery.[17,18] Nevertheless, all these the benefits of averting thromboembolic complications, and this recommendations are based on a small number of studies may be a complex task in clinical practice.[7–11] including a limited number of patients and showing significant Among the systematic reviews on the indications and risks of methodological defects. thromboprophylaxis in general spine surgery[5,12–14] we highlight Since the publication of these guidelines, new data on the the study by Glotzbecker et al,[5] which reported a DVT rate outcomes of different types of thromboprophylaxis in spinal influenced by the method of thromboprophylaxis—no prophy- surgery have been published. As there is considerable uncertainty laxis, 2.7%; compression stockings (CS), 2.7%; sequential regarding the optimal thromboprophylactic regimen to use in pneumatic compression devices (SPC), 4.6%; SPC + CS, 1.3%; elective spine surgery, we carried out a systematic review on this chemical prophylaxis 0.6%; and vena cava filter with/without topic. We aimed to evaluate the efficacy and safety of mechanical another prophylaxis method, 12%—and by the diagnostic prophylaxis, chemical prophylaxis or both for preventing VTE in method used, with a range of 1% to 12.3%. The authors of adolescents and adults undergoing elective spine surgery for this review concluded that the DVT rate was quite low in elective conditions other than trauma or malignant disease. spine surgery and that mechanical prophylaxis could be the method of choice in most cases. Routine chemoprophylaxis was 2. Methods not recommended because of insufficient evidence, except in patients with cord lesions and those undergoing high-risk spinal We conducted a systematic review following the methodology surgery (simultaneous dual approach or malignant disease). The outlined in the Cochrane Handbook for Systematic Reviews of systematic review by Cheng et al[12] was mainly focused on the Interventions.[19] We searched EMBASE and PubMed, from risk of anticoagulant administration in general spine surgery inception to March 2018. Details of search strategy are provided in (mechanical thromboprophylaxis was not assessed). The studies Appendix 1, http://links.lww.com/MD/E224. We also manually 2 Colomina et al. Medicine (2020) 99:21 www.md-journal.com Medline (n=882) EMBASE (n=1569) Arcles 2451 Excluded following reading of tle and abstract (n=2276): – Duplicates – Did not meet selecon criteria Arcles (review Excluded following reading of full of full text) text (n=140) 175 Included (n=35): – RCT (n=7) – Observaonal studies (n=28) Medline (n=649) EMBASE (n=1569) Arcles 2218 Excluded following reading of tle and abstract (n=2070): – Duplicates – Did not meet selecon criteria Arcles (review Excluded following reading of full of full text) text (n=140) 175 Included (n=35): – RCT (n=7) – Observaonal studies (n=28) Figure 1. Flow chart showing eligibility of studies. 3 Colomina et al. Medicine (2020) 99:21 Medicine Table 1 Summary of the studies of comparative effectiveness. Chemical thromboprophylaxis Du 2015 (RCT, n = 665) Rivaroxaban vs Parnaparin DVT: 3 vs 5 PE: 1 vs 2 Bleeding: 21 vs 17 Voth 1992 (RCT, n = 179) LMWH + dihydroergotamine vs sodium heparin + dihydroergotamine DVT: 4 vs 3 PE: 0 vs 0 Bleeding: 8 vs 7 Gruber 1984 (RCT, n = 41) Miniheparindihydroergotamine + sodium heparin vs placebo DVT: 1 vs 0 PE: 0 vs 0 Bleeding: 6 vs 7 Mechanical thromboprophylaxis Rokito 1996 (1) (RCT, n = 110) CS vs CS + SPC DVT: 0 vs 0 PE: 0 vs 0 Bleeding: 0 vs 0 Wood 1997 (RCT, n = 160) CS vs CS + SPC DVT: 1 vs 0 PE: 1 vs 0 Bleeding: 0 vs 0 Hamidi 2015 (RCT, n = 82) CS vs CS + enoxaparin DVT: 2 vs 0 PE: 0 vs 1 Bleeding: 1 vs 0 Ferree 1993 (prospective, n = 185) CS vs CS + SPC DVT: 4 vs 0 PE: 0 vs 0 Takahashi 2012 (retrospective, n = 1975) No prophylaxis (1990–1996 period) vs foot pump + CS postoperatively DVT: NR vs 0 (2000–2011 period) PE: 8 (1.5%)vs 3 (0.2%) Fawi 2017(retrospective, n = 2366 procedures) TED anti-embolic stockings vs TED anti-embolic stockings + subcutaneous See below (mixed) enoxaparin LMWH40 mg Mixed thromboprophylaxis Hamidi 2015 (RCT, n = 82) CS vs CS + Enoxaparin See above (chemical) Rokito 1996 (2) (RCT, n = 110) CS vs CS + Coumadin DVT: 0 vs 0 PE: 0 vs 0 Bleeding: 0 vs 2 Nelson 1996 (RCT, n = 117) CS + buffered aspirin vs CS + SPC + buffered aspirin DVT: 0 vs 0 PE: 0 vs 0 Cox 2014 (before-after, n = 992) No standardized prophylaxis (2005–2007 period) vs protocolled use of DVT: 25 vs 10 (P = .009) 5000 IU heparin + SPC prophylaxis (2008–2010 period) PE: 6 vs 5 Bleeding: 6 vs 4 Nicol 2009 (before-after, n = 1111) No prophylaxis (1985–1994 period) vs aspirin 150 mg + SPC (1996– DVT: 2 vs 1 2003 period) PE: 0 vs 0 Yang 2015a (CCT, n = 807) Mechanical prophylaxis (not specified) vs mechanical prophylaxis + LMWH DVT: 7 (8%) vs 97 (13.5%) (4100 UI/day) PE: 0 vs 0 Bleeding: 0 vs 2 Fawi 2017 (retrospective, n = 2366 procedures) TED anti-embolic stockings vs TED anti-embolic stockings + subcutaneous DVT: 1 vs 0 enoxaparin 40 mg PE: 9 (10%) vs 0 VTE: 10 vs 0 (P = .041) Bleeding: 0 vs 0 checked the reference lists of all relevant studies that were identified conditions, if these represented >50% of the total study by the above mentioned searches, as well. Only studies published as population and specific data for the relevant population was a full report were included. In order to identify other potentially not available. relevant primary studies, we also reviewed the references of other Studies had to evaluate, either alone or in combination, a systematic reviews and clinical practice guidelines referring to mechanical (CS or SPC), or chemical (low molecular weight thromboprophylaxis in spinal surgery. heparin [LMWH], low-dose unfractionated heparin [LDUH], As inclusion criteria, we considered randomized controlled aspirin, or a new direct oral anticoagulants) thromboprophy- trials (RCTs) and observational studies (minimum of 30 laxis. Prophylaxis could be administered either pre, intra, and/or participants), whether controlled or not, that included adoles- postoperatively. We did not consider vena cava filter for purposes cent (>10 years old) and/or adult patients undergoing elective of this systematic review. Outcomes of interest were DVT, PE, surgery for deformities and/or degenerative spinal disease, from and risk of bleeding or EDH. We excluded studies in language C1 to S1. We excluded studies including other spinal other than English or Spanish. 4 Colomina et al. Medicine (2020) 99:21 www.md-journal.com % Author(year) Design n N ES (95% CI) Weight Du (2015) RCT 8 550 1.5 (0.6, 2.8) 14.98 Gruber (1984) RCT 1 25 4.0 (0.1, 20.4) 12.62 Moayer (2016) Cohort 1 120 0.8 (0.0, 4.6) 14.51 Voth (1992) RCT 4 87 4.6 (1.3, 11.4) 14.30 Voth (1992) RCT 3 92 3.3 (0.7, 9.2) 14.34 Yang b (2015) Cohort 223 995 22.4 (19.9, 25.1) 15.05 Zhi-jian (2011) Cohort 0 78 0.0 (0.0, 4.6) 14.21 Overall (I^2 = 97.7%, p = 0.0) 3.7 (0.0, 13.1) 100.00 0 15 30 Percentage A Figure 2. (A) Meta-analysis of incidence of DVT in Chemoprophylaxis studies. (B) Meta-analysis of incidence of PE in Chemoprophylaxis studies. (C) Meta-analysis of incidence of bleeding in Chemoprophylaxis studies. Two authors (GU and BN) independently screened all articles the included studies. We also reviewed the references list from 8 based on title and abstract. Disagreements were resolved through systematic reviews,[5,12–14,51–54] and two clinical practice guide- discussion. Potentially relevant studies were obtained in full text, lines,[17,55] with no additional relevant primary studies identified. and again, were independently assessed for inclusion by the Appendix 3, http://links.lww.com/MD/E226 provides a list of same two authors. A third author was contacted (MJC) when excluded studies. Overall, most of the included RCTs showed an disagreement persisted. unclear risk of selection bias, and a high risk of detection and One author (KS or CR) extracted data from each study using a performance bias (Appendix 4, http://links.lww.com/MD/E227). predefined data extraction form, which included information Regarding the observational studies, we considered that their related to characteristics of the population, interventions and overall quality was low since either they used historical controls or definition of the outcome measures, details of design and results. they were a single arm study, mostly retrospective. In addition, A second author (JPB or GU) cross-checked all the extracted data. there was great clinical variability in terms of study population, Two authors (KS or CR or JBP or GU) assessed the risk of bias of type of surgery, methods of thromboprophylaxis, definition of included RCTs using the Cochrane Risk of Bias tool.[19] events, and/or VTE diagnostic methods among them. For data analysis, we grouped studies into three categories: Considering the high clinical heterogeneity among all the chemoprophylaxis, mechanical, and mixed thromboprophylaxis. included controlled studies, we decided not to perform a meta- We intended to compare the effectiveness between each category analysis of comparative effectiveness. Instead, we performed using random-effects meta-analysis were possible. Otherwise, a pooled analysis of the incidence rate of the events within each results would be presented descriptively. As we anticipated to find group. We provide a narrative synthesis of the comparative many single arm observational studies, we also planned to effects of different methods of thromboprophylaxis in Table 1. combine incidence rates across studies for each category. 4. Effects on VTE and bleeding according to the 3. Results type of prophylaxis The initial search strategy retrieved 2451 unique references. After 4.1. Chemical prophylaxis (n = 6) the screening process, 35 studies met eligibility criteria, including 7 RCTs[11,20–25] and 28 observational studies[3,6,8,26–50] (Fig. 1). In Three RCTs and three single-arm observational studies were Appendix 2, http://links.lww.com/MD/E225 we provide details of included in this group. Two RCTs compared active pharmaco- 5 Colomina et al. Medicine (2020) 99:21 Medicine % Author(year) Design n N ES (95% CI) Weight Du (2015) RCT 1 341 0.3 (0.0, 1.6) 16.53 Du (2015) RCT 2 324 0.6 (0.1, 2.2) 15.71 Gruber (1984) RCT 0 25 0.0 (0.0, 13.7) 1.23 Moayer (2016) Cohort 0 120 0.0 (0.0, 3.0) 5.83 Voth (1992) RCT 0 87 0.0 (0.0, 4.2) 4.24 Voth (1992) RCT 0 92 0.0 (0.0, 3.9) 4.48 Yang b (2015) Cohort 0 995 0.0 (0.0, 0.4) 48.18 Zhi-jian (2011) Cohort 0 78 0.0 (0.0, 4.6) 3.80 Overall (I^2 = 0.0%, p = 0.5) 0.0 (0.0, 0.0) 100.00 0 7.5 15 Percentage B Figure 2. (Continued). % Author(year) Design n N ES (95% CI) Weight Du (2015) RCT 21 341 6.2 (3.9, 9.3) 13.68 Du (2015) RCT 17 324 5.2 (3.1, 8.3) 13.65 Gruber (1984) RCT 6 25 24.0 (9.4, 45.1) 9.17 Moayer (2016) Cohort 0 120 0.0 (0.0, 3.0) 12.75 Voth (1992) RCT 8 87 9.2 (4.1, 17.3) 12.27 Voth (1992) RCT 7 92 7.6 (3.1, 15.1) 12.36 Yang b (2015) Cohort 1 995 0.1 (0.0, 0.6) 14.05 Zhi-jian (2011) Cohort 0 78 0.0 (0.0, 4.6) 12.07 Overall (I^2 = 93.7%, p = 0.0) 3.7 (0.7, 8.6) 100.00 0 25 50 Percentage C Figure 2. (Continued). 6 Colomina et al. Medicine (2020) 99:21 www.md-journal.com % Author(year) Design n N ES (95% CI) Weight Hamidi (2015) RCT 2 49 4.1 (0.5, 14.0) 3.82 Rokito (1996) RCT 0 42 0.0 (0.0, 8.4) 3.68 Rokito (1996) RCT 0 33 0.0 (0.0, 10.6) 3.44 Wood (1997) RCT 1 75 1.3 (0.0, 7.2) 4.15 Wood (1997) CCT 0 59 0.0 (0.0, 6.1) 3.97 Ferree (1993) CCT 4 74 5.4 (1.5, 13.3) 4.14 Ferree (1993) Cohort 0 111 0.0 (0.0, 3.3) 4.38 Takahashi (2012) CCT 0 841 0.0 (0.0, 0.4) 4.88 Weber (2016) Cohort 2 67 3.0 (0.4, 10.4) 4.07 Fawi (2017) Cohort 1 1677 0.1 (0.0, 0.3) 4.92 Ferree (1994) Cohort 3 60 5.0 (1.0, 13.9) 3.99 Smith (1994) Cohort 2 280 0.7 (0.1, 2.6) 4.72 Dearborn (1999A) Cohort 1 116 0.9 (0.0, 4.7) 4.40 Dearborn (1999B) Cohort 1 318 0.3 (0.0, 1.7) 4.75 Epstein (2005) Cohort 8 200 4.0 (1.7, 7.7) 4.62 Epstein (2006) Cohort 4 139 2.9 (0.8, 7.2) 4.49 Hohl (2015) Cohort 38 5766 0.7 (0.5, 0.9) 4.95 Tominaga (2015) Cohort 20 80 25.0 (16.0, 35.9) 4.19 Yoshioka (2015) Cohort 34 459 7.4 (5.2, 10.2) 4.81 Piasecki (2008) Cohort 6 66 9.1 (3.4, 18.7) 4.06 Yamasaki (2017) Cohort 111 491 22.6 (19.0, 26.6) 4.82 Yang (2015A) Cohort 7 86 8.1 (3.3, 16.1) 4.24 Akeda (2014) Cohort 7 147 4.8 (1.9, 9.6) 4.51 Overall (I^2 = 96.0%, p = 0.0) 2.9 (1.2, 5.2) 100.00 0 20 40 Percentage A Figure 3. (A) Meta-analysis of incidence of DVT in Mechanoprophylaxis studies. (B) Meta-analysis of incidence of PE in Mechanoprophylaxis studies. (C) Meta- analysis of incidence of bleeding in Mechanoprophylaxis studies. logic prophylactic agents with each other (rivaroxaban vs 22.4% in contrast with all the rest of the studies, and parnaparin, and LMWH + dihydroergotamine vs sodium heparin bleeding, where in the study by Gruber et al[25] 24% of the + dihydroergotamine),[20,23] while the other compared mini- patients in the chemoprophylaxis arm presented a bleeding heparin dihydroergotamine + sodium heparin vs placebo,[25] event defined as increased intraoperative bleeding, similar to showing no significant differences between the groups for Voth et al[23] (overall rate was 8.4% for both study thromboembolic events (DVT and PE) or for bleeding/EDH. groups). The three observational studies evaluated chemical prophylaxis with dalteparin[3] or LMWH.[41,49] 4.2. Mechanical prophylaxis (n = 19) The overall incidence (pooled effect across all chemoprophy- laxis arms [n = 8]) of DVT in these studies was 3.7% (0.0–13.1), Three RCTs and 16 observational studies (four of them 0.0% (0.0–0.0) for PE, and for bleeding complications 3.7% controlled) provided data for this group, including arms with (0.7–8.6) (Fig. 2A–C). In the study by Du et al[20] (rivaroxaban or CS, CS + SPC, and foot pump during surgery + CS. Three out of parnaparin groups) only 3 events out of 38 reported consisted of the seven controlled studies compared CS vs CS + SPC, with few serious bleeding events. In this study, there was also 1 case (0.3%) more VTE events in the CS alone group.[8,11,24] Two more studies of EDH in the rivaroxaban group and 1 death due to incoercible compared CS vs CS + enoxaparin which found comparable[21] or gastrointestinal bleeding in the parnaparin group. Only one more better[48] efficacy results with no increase in bleeding risk with the case of EDH was observed in the study by Yang et al,[49] which use of enoxaparin. One more study compared no thrombopro- implies an overall EDH rate of 0.10% (2 events in 2.062 patients) phylaxis vs foot pump + CS with better results with the later.[38] in this group. One more study compared mechanical prophylaxis (not speci- The meta-analysis showed high heterogeneity across studies fied) vs mixed,[43] with better results in favor of the mixed for DVT, due to a high incidence reported by Yang et al[49] of thromboprophylaxis. 7 Colomina et al. Medicine (2020) 99:21 Medicine % Author(year) Design n N ES (95% CI) Weight Hamidi (2015) RCT 0 49 0.0 (0.0, 7.3) 1.73 Rokito (1996) RCT 0 42 0.0 (0.0, 8.4) 1.51 Rokito (1996) RCT 0 33 0.0 (0.0, 10.6) 1.22 Wood (1997) RCT 1 75 1.3 (0.0, 7.2) 2.47 Wood (1997) CCT 0 59 0.0 (0.0, 6.1) 2.02 Ferree (1993) CCT 0 74 0.0 (0.0, 4.9) 2.45 Ferree (1993) Cohort 0 111 0.0 (0.0, 3.3) 3.37 Takahashi (2012) CCT 3 1434 0.2 (0.0, 0.6) 11.21 Weber (2016) Cohort 2 67 3.0 (0.4, 10.4) 2.25 Fawi (2017) Cohort 9 1677 0.5 (0.2, 1.0) 11.54 Ferree (1994) Cohort 0 60 0.0 (0.0, 6.0) 2.05 Smith (1994) Cohort 0 280 0.0 (0.0, 1.3) 6.20 Dearborn (1999A) Cohort 3 116 2.6 (0.5, 7.4) 3.48 Dearborn (1999B) Cohort 7 318 2.2 (0.9, 4.5) 6.64 Epstein (2005) Cohort 3 200 1.5 (0.3, 4.3) 5.08 Epstein (2006) Cohort 1 139 0.7 (0.0, 3.9) 3.97 Hohl (2015) Cohort 51 5766 0.9 (0.7, 1.2) 13.16 Tominaga (2015) Cohort 1 80 1.3 (0.0, 6.8) 2.61 Yoshioka (2015) Cohort 4 459 0.9 (0.2, 2.2) 7.92 Piasecki (2008) Cohort 5 66 7.6 (2.5, 16.8) 2.22 Yang (2015A) Cohort 0 86 0.0 (0.0, 4.2) 2.76 Akeda (2014) Cohort 0 147 0.0 (0.0, 2.5) 4.13 Overall (I^2 = 51.9%, p = 0.0) 0.4 (0.1, 0.8) 100.00 0 10 20 Percentage B Figure 3. (Continued). Twelve single arm observational studies assessed a variety of 4.3. Mixed prophylaxis: chemical–mechanical (n = 15) mechanical prophylaxis, including CS + SPC,[6,26,29,32,36,39] Three RCT and 12 observational studies (five of them controlled) SPC,[30,31] CS or SPC,[45] foot pump,[44] and mechanical (not provided data for this group. Two of the RCT compared mixed specified).[43] with mechanical thromboprophylaxis[11,21] with similar results, The overall incidence (pooled effect across all mechanical although in Rokito[11] mechanical prophylaxis proved to be safer prophylaxis arms [n = 23]) of DVT was 2.9% (1.2–5.2%) and than mixed prophylaxis in terms of bleeding. Another RCT 0.4% (0.1–0.8%) for PE. Bleeding complications had a 0.0% compared two modes of same mixed thromboprophylaxis,[22] incidence (Fig. 3A–C). Of note, only 10 studies accounting with similar results. As for the controlled observational studies, for 12 study arms reported this last outcome with very few two studies compared mixed against no prophylaxis, with better events. results for mixed thromboprophylaxis.[28,33] The other three Similar to chemoprophylaxis, the meta-analysis showed high observational studies compared mixed vs mechanical thrombo- heterogeneity across studies for DVT and moderate for PE. DVT prophylaxis,[43,46,48] with better results in favor of mixed rates varied across studies between 0% in 5 study arms[8,11,24,38] thromboprophylaxis. Overall, there was a slight increase in and 23.8% in Tominaga 2015,[39] or 22.6% in Yamasaki bleeding risk in the mixed group. 2017.[39,45] As for PE, reported rates varied between 0% and 3% The other seven single arm studies[27,34,35,37,42,47,50] evaluated in all studies, except for Piasecki 2008, that reported 7.6% a variety of mixed interventions: nadroparin + CS,[42] LMWH + incidence.[44] As for bleeding, the estimates across studies were CS,[27,35] warfarin or LMWH + CS + SPC,[34] LMWH + SPC,[37] consistent, and the overall risk was lower than that with argatroban or LMWH + SPC,[50] and UH or LMWH or warfarin chemoprophylaxis. + SPC.[47] 8 Colomina et al. Medicine (2020) 99:21 www.md-journal.com % Author(year) Design n N ES (95% CI) Weight Hamidi (2015) RCT 1 49 2.0 (0.1, 10.9) 0.94 Rokito (1996) RCT 0 42 0.0 (0.0, 8.4) 0.81 Rokito (1996) RCT 0 33 0.0 (0.0, 10.6) 0.64 Wood (1997) RCT 0 75 0.0 (0.0, 4.8) 1.42 Wood (1997) CCT 0 59 0.0 (0.0, 6.1) 1.12 Weber (2016) Cohort 0 67 0.0 (0.0, 5.4) 1.27 Fawi (2017) Cohort 0 1677 0.0 (0.0, 0.2) 22.05 Hohl (2015) Cohort 0 5766 0.0 (0.0, 0.1) 42.89 Tominaga (2015) Cohort 0 80 0.0 (0.0, 4.5) 1.51 Yoshioka (2015) Cohort 0 459 0.0 (0.0, 0.8) 7.83 McLynn (2017) Cohort 1 1252 0.1 (0.0, 0.4) 17.89 Yang (2015A) Cohort 0 86 0.0 (0.0, 4.2) 1.62 Overall (I^2 = 12.5%, p = 0.3) 0.0 (0.0, 0.0) 100.00 0 7.5 15 Percentage C Figure 3. (Continued). The overall incidence (pooled effect across all mixed In the study by Al-Dujaili et al,[27] 3 cases of EDH occurred prophylaxis arms [n = 14]) of DVT in these studies was 0.7% (1.8%), all in patients who had received pharmacologic (0.0–2.5), 0.1% (0.0–0.6) for PE, and for bleeding complications prophylaxis. One patient required decompression surgery and 0.2% (0.1–0.5) (Fig. 4A–C). hematoma evacuation, with posterior recovery. None of them A high statistical heterogeneity was observed for DVT and PE had residual neurological deficits. estimates across studies, and moderate for bleeding. Regarding Other EDH events were observed in the study by Pateder DVT, most of studies showed an incidence of <3%, except for et al[34] (3 cases detected in patients receiving warfarin), Guo one study using LWMH + mechanical prophylaxis (not specified), et al[50] (1 case receiving LMWH), Schizas et al[35] (two cases of that showed an incidence of 13.5%.[43] As for PE, two single arm postoperative hematomas, without development of neurological studies (one comparing warfarin or LMWH + CS + SPC[34] and symptoms or signs, requiring emergency evacuation), and Yang the other comparing LMWH + SC [35]) and one study arm of a et al[49] (two cases). Overall, the inside of EDH in this group was RCT with enoxaparin + CS[21] showed higher rates compared to of 0.30% (33 events in 10.933 patients). the rest of studies. The studies showed considerable clinical heterogeneity in the 5. Discussion populations included, types of surgery, and other important factors related to potential bias. Our results show that the mean incidence of DVT and PE is 3.7% Regarding EDH risk, the study by Gerlach et al[42] reported 13 and 0%, respectively in the chemoprophylaxis group; 2.9% and (0.7%) cases of postoperative EDH. In 5 of them, the hematoma 0.4% in the mechanoprophylaxis group; and 0.7% and 0.1% in occurred before nadoparin administration. Four patients with the mixed prophylaxis group. In the case of bleeding, the mean EDH died (0.2%) and 4 others had a persistent neurological incidence was 3.7%, 0%, and 0.2% for chemoprophylaxis, deficit. mechanoprophylaxis, and mixed prophylaxis, respectively. We In the study by Cox et al,[28] 4 patients presented symptomatic were unable to obtain specific rates according to the type of HED requiring surgical evacuation after implantation of a patient, type of surgery or spinal area, based on the available thromboprophylaxis protocol with heparin. data. 9 Colomina et al. Medicine (2020) 99:21 Medicine % Author(year) Design n N ES (95% CI) Weight Nelson (1996) RCT 0 60 0.0 (0.0, 6.0) 6.25 Nelson (1996) RCT 0 57 0.0 (0.0, 6.3) 6.18 Hamidi (2015) RCT 0 40 0.0 (0.0, 8.8) 5.59 Rokito (1996) RCT 0 35 0.0 (0.0, 10.0) 5.35 Nicol (2009) CCT 1 414 0.2 (0.0, 1.3) 7.87 Cox (2014) CCT 10 992 1.0 (0.5, 1.8) 8.08 Weber (2016) Cohort 0 40 0.0 (0.0, 8.8) 5.59 Fawi (2017) Cohort 0 689 0.0 (0.0, 0.5) 8.01 Gerlach (2003) Cohort 1 1954 0.1 (0.0, 0.3) 8.15 Pateder (2008) Cohort 6 407 1.5 (0.5, 3.2) 7.86 Al-Dujaili (2012) Cohort 1 158 0.6 (0.0, 3.5) 7.35 Strom (2014) Cohort 10 367 2.7 (1.3, 5.0) 7.83 Yang (2015A) Cohort 97 721 13.5 (11.0, 16.2) 8.02 Guo (2017) Cohort 5 418 1.2 (0.4, 2.8) 7.87 Overall (I^2 = 95.5%, p = 0.0) 0.7 (0.0, 2.5) 100.00 0 10 20 Percentage A Figure 4. (A) Meta-analysis of incidence of DVT in mixed prophylaxis studies. (B) Meta-analysis of incidence of PE in mixed prophylaxis studies. (C) Meta-analysis of incidence of bleeding in mixed prophylaxis studies. The aim of this systematic review was to provide evidence for incidence of DVT and PE in spine surgery (not including clinicians seeking to establish optimum thromboprophylaxis for pediatric, trauma nor neoplasic cases) in patients receiving no the specific population of adolescent and adult patients thromboprophylaxis, mechanical prophylaxis, and chemical undergoing elective surgery of the spine. Contrary to other prophylaxis. They found a higher incidence of these events in reviews on this topic, we did not include studies or study arms the mechanical prophylaxis group (DVT: 1%, PE: 0.81%) with patients who had not received any type of thrombopro- compared to the chemical prophylaxis group (DVT: 0.85%, PE: phylaxis in our analysis. In our opinion, thromboprophylaxis is a 0.58%) but the difference was not statistically significant. The standard of care and the interest lies in determining which is best, incidence of DVT/PE reported in this review is somewhat lower not in comparing thromboprophylaxis with no treatment. than that reported in our study. As the authors recognize, this Glotzbecker et al[5] published a systematic review including may be explained by the fact that they only included symptomatic elective, trauma and tumour surgery of the spine, where DVT DVT studies whereas in our review we included DVT diagnosed incidence was 4.6% for mechanical prophylaxis and 0.6 for with different techniques. In addition, our review included more combined treatment, and PE incidence was 1.1% for mechanical studies and, unlike Moshental et al, we excluded groups of and 0.3% for combined prophylaxis. In our opinion, a major patients who did not receive any thromboprophylaxis. From the limitation of this review is the inclusion of patients with different point of view of safety, it does not currently seem appropriate to diseases and a diversity of surgeries. This adds confusion to the provide any type of thromboembolic prophylaxis in this type of interpretation of the findings. Oliveira et al[13] conducted a surgery. In contrast to the study by Moshental et al,[14] which systematic using as search keywords “spine” and “thrombosis.” contained a meta-analysis only of the incidence of complications, They reported and overall rates of 3.55% and 1.04% for DVT our review also narratively describes the results of comparative and PE, respectively, in all types of spine surgery. This review studies, and provides a more detailed description of each included presents obvious methodological flaws that seriously limit its study. validity. Recently, Moshental et al[14] conducted a systematic Thromboprophylaxis in spine surgery is a standard of care but review and meta-analysis, whose aim was to determine the evidence is lacking to prove which is the best method of 10 Colomina et al. Medicine (2020) 99:21 www.md-journal.com % Author(year) Design n N ES (95% CI) Weight Nelson (1996) RCT 0 60 0.0 (0.0, 6.0) 3.99 Nelson (1996) RCT 0 57 0.0 (0.0, 6.3) 3.86 Hamidi (2015) RCT 1 40 2.5 (0.1, 13.2) 3.02 Rokito (1996) RCT 0 35 0.0 (0.0, 10.0) 2.74 Nicol (2009) CCT 0 414 0.0 (0.0, 0.9) 8.87 Cox (2014) CCT 5 992 0.5 (0.2, 1.2) 10.10 Weber (2016) Cohort 0 40 0.0 (0.0, 8.8) 3.02 Fawi (2017) Cohort 0 689 0.0 (0.0, 0.5) 9.68 Gerlach (2003) Cohort 0 1954 0.0 (0.0, 0.2) 10.62 Pateder (2008) Cohort 10 407 2.5 (1.2, 4.5) 8.84 Schizas (2008) Cohort 6 270 2.2 (0.8, 4.8) 7.98 Strom (2014) Cohort 4 367 1.1 (0.3, 2.8) 8.64 Yang (2015A) Cohort 0 721 0.0 (0.0, 0.5) 9.74 Guo (2017) Cohort 0 418 0.0 (0.0, 0.9) 8.89 Overall (I^2 = 79.3%, p = 0.0) 0.1 (0.0, 0.6) 100.00 0 10 20 Percentage B Figure 4. (Continued). treatment. Since the first systematic review by Glotzbecker et al in the considerable heterogeneity of this population as a risk group. 2009,[5] research on this topic has not been able to provide This document also mentioned that the appropriate time for the relevant information to resolve this question. This opinion start of postoperative anticoagulation and the type (mechanical coincides with the view set forth in the North American Spine or chemical) has not yet been determined. Society guidelines, which pointed out that evidence on the risks or A peculiar feature of spine surgery is the risk of perioperative benefits of chemical prophylaxis for VTE in elective spine surgery bleeding and/or EDH. Many authors have taken a cautious is insufficient to formulate recommendations.[17] In ACCP stance with regard to routine use of chemical prophylaxis in spine 9th edition (2012) guidelines,[18] within the recommendations surgery for conditions other than trauma and malignant for patients undergoing spine surgery with no risk factors, disease[48–52] because a larger incidence of postoperative bleeding mechanical prophylaxis, preferably with SPC is recommended complications has been observed in relation to chemical over no prophylaxis (Grade 2C), unfractionated heparin (Grade prophylaxis (3.7%) than combined prophylaxis (0.2%).[2] In 2C), or low-molecular-weight heparin (Grade 2C), without our review, the overall incidence of postoperative complications specifying the spinal segment treated or the type of surgery. In related to surgical wound bleeding and postoperative blood loss patients at a high risk for VTE undergoing spine surgery was 3.7% in chemoprophylaxis group, 0.0% in mechanical (including those with malignant disease and those treated with a group and 0.2% in mixed group. The incidence of EDH was combined anterior–posterior approach), the authors suggest globally low; differences between the different types of adding chemoprophylaxis medication to mechanical prophylaxis thromboprophylaxis were not striking (0.1 chemical, 0% once adequate hemostasis is established and the risk of bleeding mechanical, and 0.3% combined). Our findings are overlapping decreases (Grade 2C). In this guide, the need to use some kind of with those reported by other authors. Mosenthal et al[14] found thromboprophylaxis is clearly stated. In addition, the NICE an EDH rate of 0.3% but no data on bleeding complications were CG92 guidelines[56] followed this line, stating that patients gathered. Cheng et al[12,14] conducted a systematic review undergoing spine surgery should be evaluated on an individual focused on the risk of anticoagulant administration in spine basis to weight the risk of bleeding against that of VTE because of surgery (mechanical thromboprophylaxis was not assessed). The 11 Colomina et al. Medicine (2020) 99:21 Medicine % Author(year) Design n N ES (95% CI) Weight Hamidi (2015) RCT 0 40 0.0 (0.0, 8.8) 1.31 Rokito (1996) RCT 2 35 5.7 (0.7, 19.2) 1.16 Cox (2014) CCT 4 992 0.4 (0.1, 1.0) 11.92 Weber (2016) Cohort 0 40 0.0 (0.0, 8.8) 1.31 Fawi (2017) Cohort 0 689 0.0 (0.0, 0.5) 10.35 Gerlach (2003) Cohort 13 1954 0.7 (0.4, 1.1) 14.36 Pateder (2008) Cohort 3 407 0.7 (0.2, 2.1) 7.97 Schizas (2008) Cohort 2 270 0.7 (0.1, 2.7) 6.21 Al-Dujaili (2012) Cohort 3 158 1.9 (0.4, 5.4) 4.23 Strom (2014) Cohort 0 367 0.0 (0.0, 1.0) 7.51 McLynn (2017) Cohort 10 1603 0.6 (0.3, 1.1) 13.73 Yang (2015A) Cohort 2 721 0.3 (0.0, 1.0) 10.55 Guo (2017) Cohort 5 556 0.9 (0.3, 2.1) 9.38 Overall (I^2 = 51.6%, p = 0.0) 0.2 (0.1, 0.5) 100.00 0 10 20 Percentage C Figure 4. (Continued). risk of major bleeding ranged from 0% to 4.3% with use of a Consequently, no recommendation is possible. Second, the use of variety of anticoagulant agents. However, they could not chemoprophylaxis is related to the occurrence of bleeding establish a clear relationship between the risk of bleeding and complications. And third, there is no evidence of any clinically the administered agent. relevant difference between chemoprophylaxis and mechano- As a strength of our study, the search strategy included not only prophylaxis regarding VTE or PE incidence. This information databases but also a review of references of previous related SR. must be kept in mind by surgeons when deciding on the most Also, a detailed description of included primary studies, and a appropriate thromboprophylaxis for each patient. These data meta-analysis of incidence of events for each type of thrombo- would probably favor the use of mechanoprophylaxis in prophylaxis. Among the limitations, 18 observational non- uncomplicated, non-traumatic non-tumoral surgery, especially controlled studies were included. The remaining 10 studies when the procedure may require spinal canal opening. included a control group, either randomized or not, but the Nonetheless, there is a need for additional, well-designed heterogeneous comparisons precluded to pool comparative data. comparative studies to establish high-level evidence for the Some of the studies were old (publication range 1984–2018), safety and effectiveness of this practice Appendix 5, http://links. either retrospective and prospective. The inclusion criteria were lww.com/MD/E228. often poorly defined (population, interventions) and the populations were heterogeneous. The baseline risk for VTE Acknowledgments and the timing of the prophylaxis and follow-up (at hospital discharge, 1 week, 1 month) were not clearly specified in some All data generated or analyzed during this study are included in studies and were very diverse in others. this article. This systematic review confirms the low quality of research in this field. However, some findings seem unquestionable. First, Author contributions there is a complete lack of information in the adolescent population (under 18 years of age) on the incidence of VTE or PE Drs MJC, JB, MBNA, and GU contributed to conception, design, or the effectiveness of thromboprophylaxis in this age group. data analysis, and writing the manuscript. Drs GU, MBNA, JPB, 12 Colomina et al. Medicine (2020) 99:21 www.md-journal.com KRS, and CR contributed in literature search, data collection and surgery: a follow-up study of an institutional protocol in 209 patients. Spine 2014;39:791–7. editing the manuscript. Drs GB and KK contributed to the editing [27] Al-Dujaili TM, Majer CN, Madhoun TE, et al. Deep venous thrombosis and revision of the manuscript. in spine surgery patients: incidence and hematoma formation. Int Surg 2012;97:150–4. [28] Cox JB, Bridger Cox J, Weaver KJ, et al. Decreased incidence of venous References thromboembolism after spine surgery with early multimodal prophylax- is. J Neurosurg Spine 2014;4:677–84. [1] Eskildsen SM, Moll S, Lim MR. An algorithmic approach to venous [29] Dearborn JT, Hu SS, Tribus CB, et al. Thromboembolic complications thromboembolism prophylaxis in Spine Surgery. J Spinal Disord Tech after major thoracolumbar spine surgery. Spine 1999;24:1471–6. 2015;28:275–81. [30] Epstein NE. A review of the risks and benefits of differing prophylaxis [2] Bryson DJ, Uzoigwe CE, Braybrooke J. Thromboprophylaxis in spinal regimens for the treatment of deep venous thrombosis and pulmonary surgery: a survey. J Orthop Surg Res 2012;7:14. embolism in neurosurgery. Surg Neurol 2005;64:295–301. [3] Moayer A, Mohebali N, Razmkon A. Incidence of deep vein thrombosis [31] Epstein NE. Efficacy of pneumatic compression stocking prophylaxis in in patients undergoing degenerative spine surgery on prophylactic the prevention of deep venous thrombosis and pulmonary embolism dalteparin; a single center report. Bull Emerg Trauma 2016;4:38–42. following 139 lumbar laminectomies with instrumented fusions. J Spinal [4] Wei J, Li W, Pei Y, et al. Clinical analysis of preoperative risk factors for Disord Tech 2006;19:28–31. the incidence of deep venous thromboembolism in patients undergoing [32] Hohl JB, Lee JY, Rayappa SP, et al. Prevalence of venous thromboem- posterior lumbar interbody fusion. J Orthop Surg Res 2016;11:68. bolic events after elective major thoracolumbar degenerative spine [5] Glotzbecker MP, Bono CM, Wood KB, et al. Thromboembolic disease in surgery. J Spinal Disord Tech 2015;28:E310–5. spinal surgery. Spine 2009;34:291–303. [33] Nicol M, Sun Y, Craig N, et al. Incidence of thromboembolic [6] Ferree BA. Deep venous thrombosis following lumbar laminotomy and complications in lumbar spinal surgery in 1,111 patients. Eur Spine J laminectomy. Orthopedics 1994;17:35–8. 2009;18:1548–52. [7] Ferree BA, Stern PJ, Jolson RS, et al. Deep venous thrombosis after spinal [34] Pateder DB, Gonzales RA, Kebaish KM, et al. Pulmonary embolism after surgery. Spine 1993;18:315–9. adult spinal deformity surgery. Spine 2008;33:301–5. [8] Ferree BA, Wright AM. Deep venous thrombosis following posterior [35] Schizas C, Neumayer F, Kosmopoulos V. Incidence and management of lumbar spinal surgery. Spine 1993;18:1079–82. pulmonary embolism following spinal surgery occurring while under [9] Oda T, Fuji T, Kato Y, et al. Deep venous thrombosis after posterior chemical thromboprophylaxis. Eur Spine J 2008;17:970–4. spinal surgery. Spine 2000;25:2962–7. [36] Smith MD, Bressler EL, Lonstein JE, et al. Deep venous thrombosis and [10] Leon L, Rodriguez H, Tawk RG, et al. The prophylactic use of inferior pulmonary embolism after major reconstructive operations on the spine. vena cava filters in patients undergoing high-risk spinal surgery. Ann A prospective analysis of three hundred and seventeen patients. J Bone Vasc Surg 2005;19:442–7. Joint Surg Am 1994;76:980–5. [11] Rokito SE, Schwartz MC, Neuwirth MG. Deep vein thrombosis after [37] Strom RG, Frempong-Boadu AK. Low-molecular-weight heparin major reconstructive spinal surgery. Spine 1996;21:853–8. prophylaxis 24 to 36 hours after degenerative spine surgery: risk of [12] Cheng JS, Arnold PM, Anderson PA, et al. Anticoagulation risk in spine hemorrhage and venous thromboembolism. Spine 2013;38:E1498–502. surgery. Spine 2010;35:S117–124. [38] Takahashi H, Yokoyama Y, Iida Y, et al. Incidence of venous [13] Oliveira L, Marchi L, Pimenta L. Up-to-date thromboprophylaxis in thromboembolism after spine surgery. J Orthop Sci 2012;17:114–7. elective spinal surgery. A systematic review. Coluna/Columna 2014; [39] Tominaga H, Setoguchi T, Tanabe F, et al. Risk factors for venous 13:143–6. thromboembolism after spine surgery. Medicine 2015;94:e466. [14] Mosenthal WP, Landy DC, Boyajian HH, et al. Thromboprophylaxis in [40] Yoshioka K, Murakami H, Demura S, et al. Prevalence and risk factors spinal surgery. Spine 2018;43:E474–481. for development of venous thromboembolism after degenerative spinal [15] Yi S, Yoon DH, Kim KN, et al. Postoperative spinal epidural hematoma: surgery. Spine 2015;40:E301–6. risk factor and clinical outcome. Yonsei Med J 2006;47:326–32. [41] Zhi-jian S, Yu Z, Giu-xing Q, et al. Efficacy and safety of low molecular [16] Kou J, Fischgrund J, Biddinger A, et al. Risk factors for spinal epidural weight heparin prophylaxis for venous thromboembolism following hematoma after spinal surgery. Spine 2002;27:1670–3. lumbar decompression surgery. Chin Med Sci J 2011;26:221–6. [17] Bono CM, Watters WC3rd, Heggeness MH, et al. An evidence-based [42] Gerlach R, Raabe A, Beck J, et al. Postoperative nadroparin clinical guideline for the use of antithrombotic therapies in spine surgery. administration for prophylaxis of thromboembolic events is not Spine J 2009;9:1046–51. associated with an increased risk of hemorrhage after spinal surgery. [18] Kearon C, Akl EA, Comerota AJ, et al. Antithrombotic therapy for VTE Eur Spine J 2004;13:9–13. disease: antithrombotic therapy and prevention of thrombosis, 9th ed: [43] Yang S-D, Liu H, Sun Y-P, et al. Prevalence and risk factors of deep vein American College of Chest Physicians Evidence-Based Clinical Practice thrombosis in patients after spine surgery: a retrospective case-cohort Guidelines. Chest 2012;141:e419S–96S. study. Sci Rep 2015;5. [19] Higgins JPT, Green S. Cochrane Handbook for Systematic Reviews of [44] Piasecki DP, Poynton AR, Mintz DN, et al. Thromboembolic disease Interventions. Hoboken, NJ: John Wiley & Sons; 2011. after combined anterior/posterior reconstruction for adult spinal [20] Du W, Zhao C, Wang J, et al. Comparison of rivaroxaban and deformity: a prospective cohort study using magnetic resonance parnaparin for preventing venous thromboembolism after lumbar spine venography. Spine 2008;33:668–72. surgery. J Orthop Surg Res 2015;10:78. [45] Yamasaki K, Hoshino M, Omori K, et al. Prevalence and risk factors of [21] Hamidi S, Riazi M. Incidence of venous thromboembolic complications deep vein thrombosis in patients undergoing lumbar spine surgery. J in instrumental spinal surgeries with preoperative chemoprophylaxis. J Orthop Sci 2017;22:1021–5. Korean Neurosurg Soc 2015;57:114–8. [46] Weber B, Seal A, McGirr J, et al. Case series of elective instrumented [22] Nelson LDJr, Montgomery SP, Dameron TBJr, et al. Deep vein posterior lumbar spinal fusions demonstrating a low incidence of venous thrombosis in lumbar spinal fusion: a prospective study of antiembolic thromboembolism. ANZ J Surg 2016;86:796–800. and pneumatic compression stockings. J South Orthop Assoc [47] McLynn RP, Diaz-Collado PJ, Ottesen TD, et al. Risk factors and 1996;5:181–4. pharmacologic prophylaxis for venous thromboembolism in elective [23] Voth D, Schwarz M, Hahn K, et al. Prevention of deep vein thrombosis in spine surgery. Spine J 2018;8:970–8. neurosurgical patients: a prospective double-blind comparison of two [48] Fawi HMT, Saba K, Cunningham A, et al. Venous thromboembolism in prophylactic regimen. Neurosurg Rev 1992;15:289–94. adult elective spinal surgery: a tertiary centre review of 2181 patients. [24] Wood KB, Kos PB, Abnet JK, et al. Prevention of deep-vein thrombosis Bone Joint J 2017;99-B:1204–9. after major spinal surgery: a comparison study of external devices. J [49] Yang S-D, Ding W-Y, Yang D-L, et al. Prevalence and risk factors of deep Spinal Disord 1997;10:209–14. vein thrombosis in patients undergoing lumbar interbody fusion surgery: [25] Gruber UF, Rem J, Meisner C, et al. Prevention of thromboembolic a single-center cross-sectional study. Medicine 2015;94:e2205. complications with miniheparin-dihydroergotamine in patients under- [50] Guo Y, Zou Z, Jia L, et al. Safety and effectiveness of argatroban versus going lumbar disc operations. Eur Arch Psychiatry Neurol Sci heparin for preventing venous thromboembolism after lumbar decom- 1984;234:157–61. pressive surgery. Int J Surg 2017;44:324–8. [26] Akeda K, Matsunaga H, Imanishi T, et al. Prevalence and counter- [51] Catre MG. Anticoagulation in spinal surgery. A critical review of the measures for venous thromboembolic diseases associated with spinal literature. Can J Surg 1997;40:413–9. 13 Colomina et al. Medicine (2020) 99:21 Medicine [52] Glotzbecker MP, Bono CM, Wood KB, et al. Postoperative spinal [55] Geerts WH, Bergqvist D, Pineo GF, et al. Prevention of venous epidural hematoma: a systematic review. Spine 2010;35:E413–20. thromboembolism: American College of Chest Physicians Evidence- [53] Sansone JM, del Rio AM, Anderson PA. The prevalence of and specific Based Clinical Practice Guidelines (8th Edition). Chest 2008;133:381S– risk factors for venous thromboembolic disease following elective spine 453S. surgery. J Bone Joint Surg Am 2010;92:304–13. [56] Venous thromboembolism: reducing the risk: reducing the risk of venous [54] Schuster JM, Fischer D, Dettori JR. Is chemical antithrombotic thromboembolism (deep vein thrombosis and pulmonary embolism) in prophylaxis effective in elective thoracolumbar spine surgery? Results patients admitted to hospital [CG92]. NICE; n.d. https://www.nice.org. of a systematic review. Evid Based Spine Care J 2010;1:40–5. uk/guidance/CG92/uptake [access December 31, 2018]. 14