Systematic Review
Venous Thromboembolism Events After Hip Arthroscopy: A Systematic Review Chloe E. Haldane, M.D.(Cand), Seper Ekhtiari, M.D., Darren de Sa, M.D., F.R.C.S.C., Nicole Simunovic, M.Sc., Marc Safran, M.D., Filippo Randelli, M.D., Andrew Duong, M.Sc., Forough Farrokhyar, M.Phil., Ph.D., and Olufemi R. Ayeni, M.D., M.Sc., F.R.C.S.C.
Purpose: The purpose of this systematic literature review focused on hip arthroscopy was to (1) report the venous thromboembolism (VTE) event incidence in patients who receive VTE prophylaxis and those who do not, (2) report how VTE prophylaxis is currently being administered, and (3) report operative and patient-related risk factors for VTE identified in the literature. Methods: The electronic databases MEDLINE, Embase, and PubMed were searched from database inception to October 10, 2016, and screened in duplicate for relevant studies. Data were collected regarding VTE prophylaxis, traction use, surgical time, VTE incidence, patient and operative factors, and postoperative weight bearing and rehabilitation. Study quality was assessed in duplicate with the Methodological Index for Non-Randomized Studies criteria. Results: Outcome analyses included 14 studies that involved 2,850 patients (2,985 hips). The weighted mean follow-up period was 19 8 months, ranging from 7 days to 103 months. The weighted mean age was 40.7 7 years, ranging from 6 to 82 years, and 39.6% of patients were male patients. The overall weighted proportion of VTE events after hip arthroscopy found in 14 included studies was 2.0% (95% confidence interval, 0.01%-4.1%), with 25 VTE events. Several studies reported patient risk factors, which included increased age, increased body mass index, prolonged traction time, and use of oral contraceptives. Conclusions: The use and efficacy of VTE prophylaxis are highly under-reported within hip arthroscopy. The low incidence of VTE events found in this review (2.0%) suggests that prophylaxis may not be necessary in low-risk patients undergoing hip arthroscopy; however, the true rate may be under-reported. Current literature suggests that prophylaxis is typically not prescribed. Early mobility and postoperative rehabilitation may also help to further mitigate the risk of VTE events, but use of these strategies needs further prospective evaluation. Level of Evidence: Level IV, systematic review of Level II through IV studies.
ower limb arthroscopic procedures have long been perceived as procedures with a relatively low risk of complications, particularly in comparison with open procedures such as knee and hip arthroplasty.1 However, arthroscopic procedures have also been associated with the development of venous thromboembolism (VTE) events including pulmonary embolism (PE) and
L
deep vein thrombosis (DVT).2,3 Within knee arthroscopy, the reported incidence of VTE with prophylaxis ranges from 0.1% to 11.9%.4 This is significantly lower than that of more invasive procedures such as total knee arthroplasty, for which the VTE event rate has been reported to range from 40% to 84%.5 Although prophylaxis guidelines have been developed for hip
From the Michael G. DeGroote School of Medicine (C.E.H., S.E.), Division of Orthopaedic Surgery, Department of Surgery (D.d.S., A.D., F.F., O.R.A.), and Department of Methods, Evidence and Impact (N.S., F.F.), McMaster University, Hamilton, Ontario, Canada; Department of Orthopaedic Surgery, Stanford University (M.S.), Redwood City, California, U.S.A.; and Dipartimento di Ortopedia e Traumatologia V, Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS) Policlinico San Donato (F.R.), Milan, Italy. The authors report the following potential conflict of interest or source of funding: M.S. receives support from ISAKOS, ISHA. Paid board membership. Medacta, ConMed, Smith & Nephew, Biomimedica. Consultant. Smith & Nephew, ConMed. Fellowship grant. Medacta, Smith & Nephew, ConMed. Education consultant. Hip brace. Patent submitted. Hip distraction device. Patent submitted with Stanford. Williams and Wilkins, WB Saunders. Book royalties. Howmedica/Stryker. Shoulder anchor. Smith & Nephew. Hip anchor. DJO. Shoulder sling. F.R. receives support from Stryker. Consultancy for
Course Organization on Anterior Hip Approach in Italy and Europe 2016. Smith & Nephew. Lecture at the “Hip Revision Master Class,” Bologna, October 21, 2016. Bayer. Presentation at SIOT Symposia on Risk Management after Orthopedic Surgery, Rome, November 24, 2014. Full ICMJE author disclosure forms are available for this article online, as supplementary material. Received January 30, 2017; accepted July 5, 2017. Address correspondence to Olufemi R. Ayeni, M.D., M.Sc., F.R.C.S.C., Division of Orthopaedic Surgery, Department of Surgery, McMaster University, 1200 Main St W, Ste 4E17, Hamilton, Ontario, Canada L8N3Z5. E-mail: [email protected] Ó 2017 by the Arthroscopy Association of North America 0749-8063/1791/$36.00 http://dx.doi.org/10.1016/j.arthro.2017.07.006
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol
-,
No
-
(Month), 2017: pp 1-10
1
2
C. E. HALDANE ET AL.
arthroplasty and knee arthroplasty, thus far, none have been developed for lower limb arthroscopic procedures, despite their increasing use. As a result, several systematic reviews have sought to inform optimal prophylaxis in knee arthroscopy and have examined whether prophylaxis is necessary given a low reported incidence.4 A recent review on DVT within knee arthroscopy by Sun et al.4 concluded that prophylaxis may in fact reduce total DVT occurrence. Within hip arthroscopy, for which the reported VTE incidence rate is similar to that of knee arthroscopy, there has yet to be a similar investigation into how prophylaxis is used and its influence on the VTE event rate. DVTs and PEs can generally be attributed to 3 factors identified in the Virchow triad: hypercoagulability, endothelial dysfunction, and/or hemodynamic changes (stasis).6,7 These factors may be further influenced by operative, postoperative, and patientrelated factors. Operative factors can include prolonged surgical time and traction time, whereas patient factors may include obesity, prothrombotic conditions, smoking, use of oral contraceptives, malignancy, and metabolic or cardiovascular disturbances. Finally, postoperative factors can include weight-bearing and mobility restrictions.7 Current literature on VTE events related to hip arthroscopy suggests a low reported incidence of VTE events. However, many previous studies have relied on symptomatic presentations in the reporting of the VTE incidence, with reported rates ranging from 1.4% to 3.7%.8,9 A recent prospective study by Mohtadi et al.,10 which used ultrasound, postoperatively found an overall higher incidence of DVT, at 4.4%, which included asymptomatic presentations. Within orthopaedic surgery, the risk of PE from asymptomatic proximal venous thrombosis (of popliteal or more proximal veins) has been reported to be about 25%, wherein a fatal PE occurs in 1% to 2% of patients in this group.11,12 Given the increasing use of hip arthroscopy, as well as the morbidities and deaths associated with both asymptomatic and symptomatic VTE, it is important to elucidate the incidence of these events and the role of prophylaxis in their prevention. Furthermore, improved identification of surgical and patient-related risk factors that may lead to VTE events may also aid in informing prophylaxis management. The purpose of this systematic literature review focused on hip arthroscopy was to (1) report the VTE event incidence in patients who receive VTE prophylaxis and those who do not, (2) report how VTE prophylaxis is currently being administered, and (3) report operative and patient-related risk factors for VTE identified in the literature. The hypothesis of this article was that VTE events would be rare reported
complications after hip arthroscopy that would occur in patients with identifiable risk factors.
Methods Design and Eligibility Criteria The research question and study eligibility criteria for this systematic review were established a priori. The patient population included male and female patients of all ages who underwent hip arthroscopy. The primary outcome measure was a VTE event in patients who received prophylaxis versus those who did not. The inclusion criteria were (1) all levels of evidence; (2) studies published in English, (3) studies conducted in humans that indicated whether VTE prophylaxis was or was not used, and (4) studies that reported whether VTE events occurred or did not occur. The exclusion criteria were (1) nonsurgical treatment studies (conservative treatment, technique articles without outcomes, and so on) and (2) studies in which the outcomes for the exact same patient population were reported in multiple articles (the most recent article was included). Search Strategy A systematic search strategy previously described by us was used.13,14 One reviewer (S.E.) searched 3 online databases (Embase, MEDLINE, and PubMed) for literature related to hip arthroscopy. The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines were followed in the development of this study. The search was conducted on October 10, 2016, and retrieved articles from database inception to search date. To be more inclusive, the following key terms were used in a broad-based search: “hip” and “arthroscopy.” The search results were merged, and duplicates were removed. The search strategy is outlined in Appendix Table 1 (available at www.arthroscopyjournal.org). Study Screening Two reviewers (C.E.H. and S.E.) independently screened titles, abstracts, and full texts of retrieved studies. Discrepancies during the title and abstract screening stages were resolved by automatic inclusion to ensure thoroughness. Discrepancies during full-text screening were resolved through consensus between the reviewers. If a consensus could not be reached, a senior reviewer (D.D.) assisted to resolve the discrepancy. In addition, abstracts from American Orthopaedic Society for Sports Medicine; International Society of Arthroscopy, Knee Surgery & Orthopaedic Sports Medicine; European Society of Sports Traumatology, Knee Surgery and Arthroscopy; and American Academy of Orthopaedic Surgeons proceedings from 2011 to 2016 were also screened for relevant information. The references of included
3
VENOUS THROMBOEMBOLISM AFTER HIP ARTHROSCOPY
studies were searched to identify articles that warranted inclusion in this review. Quality Assessment of Included Studies A quality assessment of included studies was completed with the Methodological Index for Non-Randomized Studies (MINORS) criteria. MINORS is a validated scoring tool for nonrandomized studies. Each of the 12 items in the MINORS criteria is given a score of 0, 1, or 2, with maximum scores of 16 and 24 for noncomparative and comparative studies, respectively.15 Data Abstraction Two reviewers (C.E.H. and S.E.) independently abstracted relevant data from included articles and recorded the data in a Microsoft Excel spreadsheet (2013) designed a priori. Demographic information including author, year of publication, region, sample size, study design, level of evidence, patient demographic data (age, sex), and length of follow-up was recorded. Regardless of the study design being comparative or a single-group cohort, data on the use of preoperative VTE prophylaxis, procedures performed, traction use and duration, surgical time, VTE incidence, VTE risk factors (body mass index, use of anticoagulants, and so on), and postoperative rehabilitation were abstracted. Statistical Analysis A k value was calculated at each stage of article screening to evaluate inter-reviewer agreement. Agreement was categorized a priori as follows: k greater than 0.60 indicates substantial agreement; k of 0.21 to 0.60, moderate agreement; and k less than 0.21, slight agreement.16,17 An intraclass correlation coefficient was used to evaluate the inter-reviewer agreement of the MINORS scores. Weighted mean values with standard deviations are reported for patients’ baseline characteristics. For the purpose of meaningful quantitative pooling of the data, only studies with sample sizes of 30 or larger were included in the analyses. The number of hips was used in reporting pooled data from all included studies. To calculate the weighted proportions of VTE, the randomeffects DerSimonian-Laird model was used because of the inherent heterogeneity in the observational studies. The I2 value was calculated to quantify between-study heterogeneity. The weighted proportions and I2 values with 95% confidence intervals (CIs) are reported. An I2 of less than 0.25 is considered low heterogeneity; 0.25 to 0.5, moderate heterogeneity; and greater than 0.5, high heterogeneity.18 The StatsDirect website (www. statsdirect.com) was used in the statistical analysis.
Results Eligibility Assessment and Study Selection Our search yielded 5,746 studies, of which 21 met the eligibility criteria for this review (Fig 1). A list of
6683 Studies Iden fied Medline: 1088 Studies Embase: 2332 Studies Removal of duplicates
3334 Studies Title
2042 Studies Abstract
Removed: 3349
Removed: 1292 - Animal/cadaver/in vitro: 113 - Different/combined procedures: 111 - Non-opera ve: 215 Removed: 964 - Animal/cadaver/in vitro: 39 - Different/combined procedures: 139 - Non-opera ve: 71
1078 Studies Full Text Review
Removed: - No DVT prophylaxis data: 1058
20 Studies Hand Search of Literature and Full Text
Addi onal Studies Iden fied: 1
21 Studies Included
Fig 1. Screening strategy. (DVT, deep vein thrombosis.)
included studies can be found in Appendix Table 2 (available at www.arthroscopyjournal.org). Of the 21 included studies, all were published between 1998 and 2016 (Table 1). Most studies were based in North America (n ¼ 12), followed by Europe (n ¼ 6), with 1 each in Australia, South America, and Asia. No additional abstracts were found in a gray-literature search of American Orthopaedic Society for Sports Medicine; International Society of Arthroscopy, Knee Surgery & Orthopaedic Sports Medicine; European Society of Sports Traumatology, Knee Surgery and Arthroscopy; and American Academy of Orthopaedic Surgeons conferences from 2011 to 2016. There was substantial agreement between reviewers at the title (k ¼ 0.74; 95% CI, 0.71-0.76), abstract (k ¼ 0.78; 95% CI, 0.76-0.81), and full-text (k ¼ 0.97; 95% CI, 0.95-0.99) screening phases. Study Characteristics and Patient Demographic Data Most included studies were of Level IV evidence (n ¼ 14); there were 5 studies of Level III evidence and 2 prospective cohort studies (Level II evidence). There was substantial agreement between the reviewers for the quality assessment (intraclass correlation coefficient, 0.88; 95% CI, 0.84-0.94). The included studies had a mean MINORS score of 10.4 3.2, which indicates fair-quality evidence. Specifically, the nonecase report studies had a mean MINORS score of 11.2 3.4,
Authors, Year Alaia et al.,9 2014
Beutel et al.,19 2015
Domb et al.,22 2016 Fabricant et al.,23 2012 Fukushima et al.,24 2016
Horisberger et al.,25 2010
No. of Hip No. of Arthroscopies Patients Performed Mean Mean at Final During Age SD Follow-up Lost to Prophylaxis Follow-up Male Study Period (No. of Hips) Patients (Range), yr SD (Range) Follow-up Used (Type) d 139 (139) 39% 37.7 12.0 16.12 d None Yes (NSAID not specified)
d
1,054
d
1,155
d
98 (101)
150
18 (18)
28%
56 (41-68)
1,054 (1,054) 43%
37 (6-80)
39 (39)
29.4 mo None (18-54 mo)
6 wk
None
Yes (Aspirin 325 mg, daily for 2 weeks)
No
40%
39.5 (21-64)
23 mo None (12-43 mo)
Yes (Aspirin 325 mg, daily for 2 weeks)
921 (1,038) 26%
36.3 (13-76)
28.8 mo (23.569 mo)
18 mo None (12-30 mo)
Yes (Aspirin 325 mg, twice daily for 2 weeks) No
21 (27)
57%
17.6 1.6 (14.5-19.9)
72 (72)
40.3%
46.3 1.7
19 (19)
80%
47.3 (22-65)
7d
8.8%
None
No in 71, yes in 1
Yes (LMWH, 36 mo (18- 5%; 1 daily for 2 49.2 mo) patient weeks) died of cause unrelated to topic at hand (continued)
C. E. HALDANE ET AL.
Clarke et al.,20 2003 Collins et al.,21 2015
Average Study Level of MINORS Inclusion Exclusion Criteria Criteria Design Evidence Score Retrospective III 7.5 Hip arthroscopy with Patients who cohort postoperative were deemed ultrasound screen at high risk of VTE for DVT and given prophylaxis Prior surgery on Case control III 18 Patients undergoing ipsilateral side hip arthroscopy with traction with or without previous hip or knee arthroplasty None stated Retrospective III 9 Patients undergoing cohort hip arthroscopy between 1989-2001 Case control IV 19 Patients undergoing Previous hip surgery hip arthroscopy; (arthroscopic or BMI >30 for obese open) group Prospective II 11 Patients undergoing Previous hip cohort hip arthroscopy conditions (AVN, with minimum LCPD, EDS, PVNS) follow-up of 2 yr Case series IV 10 Athletically active Revision patients aged 19 arthroscopy yr undergoing hip arthroscopy for FAI Prospective III 9 Patients who Patients who were cohort underwent hip not able to undergo arthroscopy between preoperative and September 2012 and postoperative December 2015 ultrasound, who underwent combined open procedures, who were aged 18 yr Involvement in undergoing elective litigation and/or hip arthroscopy patients with acute fracture, dislocation or septic arthritis, drug addiction, underlying risk of thromboembolic event, or prior surgery on same hip Patients undergoing None stated arthroscopic FAI management Consecutive patients None stated undergoing hip arthroscopy Patients undergoing None stated hip arthroscopy
Lost to Follow-up
Prophylaxis Used (Type)
d
88 (105)
68%
40.9 (17-66) 27.6 mo (15.6- None 49.2 mo)
1,693
41 (42)
66%
65 (60-82) 30 mo None (12-54 mo)
145
115 (115)
48%
d
42 (42)
60%
35.1 (16-52) 28.2 mo None (10-72 mo)
Yes (not specified)
d
81 (81)
40%
32.2 (14-59) 8 d
None
No
d
194 (194)
59%
36.2 12.2 39.5 mo None (7-78) (4-103 mo)
No
39.15 (18-59)
3 mo
None
Yes (LMWH, daily for 2 weeks) Yes (Subcutaneous LMWH, 20 mg, at induction of anesthesia)
No
VENOUS THROMBOEMBOLISM AFTER HIP ARTHROSCOPY
Authors, Year
No. of Hip No. of Arthroscopies Patients Performed Mean Mean at Final During Age SD Follow-up Follow-up Male Study Period (No. of Hips) Patients (Range), yr SD (Range)
NOTE. The follow-up data presented for Fukushima et al.24 (7 d) and for Mohtadi et al.10 (3 mo) were both specific follow-up times and not means. AVN, avascular necrosis; BMI, body mass index; DVT, deep vein thrombosis; EDS, Ehlers-Danlos syndrome; FAI, femoroacetabular impingement; LCPD, Legg-Calves Perthes disease; LMWH, lowemolecular weight heparin; MINORS, Methodological Index for Non-Randomized Studies; NSAID, nonsteroidal anti-inflammatory drug; OA, osteoarthritis; PVNS, pigmented villonodular synovitis; SD, standard deviation; VTE, venous thromboembolism.
5
6
C. E. HALDANE ET AL.
whereas the case report studies had a mean MINORS score of 8.33 1.3. The nonecase report studies mostly lacked control groups, prospective sample size calculation, and unbiased assessment of endpoints. The case report studies also lacked appropriate endpoints and follow-up. Among the 21 included studies, there were 7 singlepatient case reports, 7 case series of 19 to 88 patients, and 7 comparative studies of 18 to 1,054 patients. The 7 single-patient case reports were excluded, and the remaining case series and comparative studies were included in the outcome analysis (Table 1). Among the 14 studies included in the outcome analysis, a total of 2,850 patients (2,985 hips) were available at final follow-up. The weighted mean follow-up period was 19 8 months, ranging from 7 days to 103 months. The weighted mean age was 40.7 7 years, ranging from 6 to 82 years, and 39.6% of patients were male patients (Table 1). VTE Incidence Table 2 presents the studies that measured VTE. The overall pooled proportion of VTE events after hip arthroscopy in 2,985 hips was 2.0% (95% CI, 0.01%-4.1%), with a total of 25 VTE events and between-study heterogeneity of 85.9% (95% CI, 73.1%-91.1%). There were 3 PE and 22 DVT events. To address the large heterogeneity, the studies that did not indicate the use of ultrasound for the diagnosis of VTE were removed from analysis, wherein the pooled VTE event rate increased to 4.2% (95% CI, 2.4%-6.5%) and heterogeneity decreased to 17.6% (95% CI, 0%-70%), which is considered low heterogeneity. A single study that reported that no VTE events occurred did not use ultrasound imaging for confirmation and thus relied on a lack of presentation of clinical symptoms in reporting the VTE event incidence.22 The pooled proportion of VTE events was 2.5% (95% CI, 0.6%-5.9%) in 1,947 patients after removal of this study, which also had 11% loss to follow-up. Diagnoses of DVT were made or confirmed using duplex Doppler ultrasound for 17 of the 25 VTE events. In 3 studies involving 326 hips, ultrasound was used after surgery to prospectively detect VTE events.10,24 Within these studies, 6 asymptomatic DVTs were detected out of the 12 total VTE events that occurred in this population. Thus, of the 25 VTE events in the included studies, 24% were asymptomatic. Time to diagnosis of VTE was highly under-reported and is presented, when reported, in Table 2. The incidence of VTE was 2.0% (95% CI, 0.3%-5.0%) in 9 studies with 1,443 hips in which VTE prophylaxis was used, whereas it was 2.3% (95% CI, 0.02%-6.6%) in 6 studies with 1,542 hips in which VTE prophylaxis was not used. In 1 study, prophylaxis was used for 1 hip with the remainder receiving prophylaxis.22 Most of the hips in the studies that did not use prophylaxis were from a single large study of 1,054 hips in which there
were no reported VTE events. With this study being excluded, the incidence of VTE in studies that did not use VTE prophylaxis (488 hips) was 3.6% (95% CI, 1.1%-7.6%). All studies reported that patients who had VTE events were treated with various forms of anticoagulation. VTE Prophylaxis The weighted proportion of DVT-VTE prophylaxis used in 8 of 14 studies encompassing 1,443 hips was 48.3% (Table 1). Within this group, 325 mg of aspirin was used in 1,095 hips, lowemolecular weight heparin was used in 166 hips; and nonsteroidal antiinflammatory drugs were used in 139 hips. The remaining studies, encompassing 43 hips, did not specify the type of prophylaxis used. The duration of prophylaxis was reported for 1,219 hips, with most studies reporting the duration of prophylaxis as 2 weeks. Patient and Operative Factors Procedures performed included femoroacetabular impingement resection, labral debridement and/or repair, and loose body removal. Within the 14 included studies, patient factors were identified in 15 of the 25 total VTE events (Table 2). The use of oral contraceptives was an identified factor in 5 patients. In addition, noneweight bearing postoperatively was an identified factor in 2 patients and partial weight bearing was a factor in 5. It is known that hypercoagulability and immobility may increase the risk of development of VTE.7 Traction was reportedly used in 1,734 hips. In studies that reported traction duration, the mean duration ranged from 28.6 to 58.9 minutes. When reported, the mean traction time in studies in which VTE events occurred (398 hips, 9 VTE events) varied from 46.5 to 58.9 minutes. A single study with 39 hips and no VTE events reported a mean traction duration of 28.6 minutes. The duration of surgery was also a highly underreported factor and was only reported in 3 studies (195 hips). In studies with no VTE events (123 hips), the duration of surgery varied from 54 to 130.7 minutes. In a single study with 5 VTE events (72 patients), the mean duration of surgery was 81.6 minutes. Postoperative Weight Bearing and Rehabilitation Weight-bearing status was reported for 2,442 hips (84.3%) in 12 studies. In the remaining studies, weight bearing was reported at the discretion of the surgeon or not specified. A total of 1,054 patients (1,054 hips) were deemed noneweight bearing for 72 hours after surgery, after which they were deemed weight bearing as tolerated. In 3 studies encompassing 166 hips, weightbearing parameters were provided for patients with microfracture, in whom partial weight bearing was
Table 2. VTE Events and Risk Factors Identified
39 921
2 DVTs (6.9%) 5 DVTs and 2 PEs (0.76%)
Obese (i.e., BMI >30) (2) d
72
5 DVTs (6.9%)
Increased mean age compared with non-DVT patients (5)
115
5 DVTs (4.4%)
No
81
3 DVTs (3.7%)
Souza et al.,29 2010 No
194
1 DVT (0.5%)
Use of oral contraceptive (2) PWB (5) Prolonged mean traction time and increased mean age compared with non-DVT patients (5) NWB (2) Use of oral contraceptive (1) No risk factors (1) d
Collins et al.,21 2015 Yes Domb et al.,22 2016 Yes
Fukushima et al.,24 2016
Mohtadi et al.,10 2016
Salvo et al.,8 2010
No in 71 patients, yes in 1 patient No
WB Status d
Traction Time to PE/DVT (Duration, min) Diagnosis Diagnosis Yes (mean, 58.9) Ultrasound screen (58.3%), Mean, 16.12 d clinical (symptomatic) (SD, 10.72 d)
72 h NWB, followed by WBAT d 8 wk PWB (20 lb), fitted brace limiting hip adduction and flexion d
Yes
At discretion of surgeon
Yes (mean, 49.5) Clinical (1 asymptomatic), physical examination, modified Wells criteria, ultrasound screen
d
d
Yes (46.5) d
Ultrasound screen d
Yes (45)
Ultrasound screen, CT Within 7 d screen, asymptomatic (5) postoperatively
d
Yes (
Venous Thromboembolism Events After Hip Arthroscopy: A Systematic Review Chloe E. Haldane, M.D.(Cand), Seper Ekhtiari, M.D., Darren de Sa, M.D., F.R.C.S.C., Nicole Simunovic, M.Sc., Marc Safran, M.D., Filippo Randelli, M.D., Andrew Duong, M.Sc., Forough Farrokhyar, M.Phil., Ph.D., and Olufemi R. Ayeni, M.D., M.Sc., F.R.C.S.C.
Purpose: The purpose of this systematic literature review focused on hip arthroscopy was to (1) report the venous thromboembolism (VTE) event incidence in patients who receive VTE prophylaxis and those who do not, (2) report how VTE prophylaxis is currently being administered, and (3) report operative and patient-related risk factors for VTE identified in the literature. Methods: The electronic databases MEDLINE, Embase, and PubMed were searched from database inception to October 10, 2016, and screened in duplicate for relevant studies. Data were collected regarding VTE prophylaxis, traction use, surgical time, VTE incidence, patient and operative factors, and postoperative weight bearing and rehabilitation. Study quality was assessed in duplicate with the Methodological Index for Non-Randomized Studies criteria. Results: Outcome analyses included 14 studies that involved 2,850 patients (2,985 hips). The weighted mean follow-up period was 19 8 months, ranging from 7 days to 103 months. The weighted mean age was 40.7 7 years, ranging from 6 to 82 years, and 39.6% of patients were male patients. The overall weighted proportion of VTE events after hip arthroscopy found in 14 included studies was 2.0% (95% confidence interval, 0.01%-4.1%), with 25 VTE events. Several studies reported patient risk factors, which included increased age, increased body mass index, prolonged traction time, and use of oral contraceptives. Conclusions: The use and efficacy of VTE prophylaxis are highly under-reported within hip arthroscopy. The low incidence of VTE events found in this review (2.0%) suggests that prophylaxis may not be necessary in low-risk patients undergoing hip arthroscopy; however, the true rate may be under-reported. Current literature suggests that prophylaxis is typically not prescribed. Early mobility and postoperative rehabilitation may also help to further mitigate the risk of VTE events, but use of these strategies needs further prospective evaluation. Level of Evidence: Level IV, systematic review of Level II through IV studies.
ower limb arthroscopic procedures have long been perceived as procedures with a relatively low risk of complications, particularly in comparison with open procedures such as knee and hip arthroplasty.1 However, arthroscopic procedures have also been associated with the development of venous thromboembolism (VTE) events including pulmonary embolism (PE) and
L
deep vein thrombosis (DVT).2,3 Within knee arthroscopy, the reported incidence of VTE with prophylaxis ranges from 0.1% to 11.9%.4 This is significantly lower than that of more invasive procedures such as total knee arthroplasty, for which the VTE event rate has been reported to range from 40% to 84%.5 Although prophylaxis guidelines have been developed for hip
From the Michael G. DeGroote School of Medicine (C.E.H., S.E.), Division of Orthopaedic Surgery, Department of Surgery (D.d.S., A.D., F.F., O.R.A.), and Department of Methods, Evidence and Impact (N.S., F.F.), McMaster University, Hamilton, Ontario, Canada; Department of Orthopaedic Surgery, Stanford University (M.S.), Redwood City, California, U.S.A.; and Dipartimento di Ortopedia e Traumatologia V, Istituto di Ricovero e Cura a Carattere Scientifico (IRCCS) Policlinico San Donato (F.R.), Milan, Italy. The authors report the following potential conflict of interest or source of funding: M.S. receives support from ISAKOS, ISHA. Paid board membership. Medacta, ConMed, Smith & Nephew, Biomimedica. Consultant. Smith & Nephew, ConMed. Fellowship grant. Medacta, Smith & Nephew, ConMed. Education consultant. Hip brace. Patent submitted. Hip distraction device. Patent submitted with Stanford. Williams and Wilkins, WB Saunders. Book royalties. Howmedica/Stryker. Shoulder anchor. Smith & Nephew. Hip anchor. DJO. Shoulder sling. F.R. receives support from Stryker. Consultancy for
Course Organization on Anterior Hip Approach in Italy and Europe 2016. Smith & Nephew. Lecture at the “Hip Revision Master Class,” Bologna, October 21, 2016. Bayer. Presentation at SIOT Symposia on Risk Management after Orthopedic Surgery, Rome, November 24, 2014. Full ICMJE author disclosure forms are available for this article online, as supplementary material. Received January 30, 2017; accepted July 5, 2017. Address correspondence to Olufemi R. Ayeni, M.D., M.Sc., F.R.C.S.C., Division of Orthopaedic Surgery, Department of Surgery, McMaster University, 1200 Main St W, Ste 4E17, Hamilton, Ontario, Canada L8N3Z5. E-mail: [email protected] Ó 2017 by the Arthroscopy Association of North America 0749-8063/1791/$36.00 http://dx.doi.org/10.1016/j.arthro.2017.07.006
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol
-,
No
-
(Month), 2017: pp 1-10
1
2
C. E. HALDANE ET AL.
arthroplasty and knee arthroplasty, thus far, none have been developed for lower limb arthroscopic procedures, despite their increasing use. As a result, several systematic reviews have sought to inform optimal prophylaxis in knee arthroscopy and have examined whether prophylaxis is necessary given a low reported incidence.4 A recent review on DVT within knee arthroscopy by Sun et al.4 concluded that prophylaxis may in fact reduce total DVT occurrence. Within hip arthroscopy, for which the reported VTE incidence rate is similar to that of knee arthroscopy, there has yet to be a similar investigation into how prophylaxis is used and its influence on the VTE event rate. DVTs and PEs can generally be attributed to 3 factors identified in the Virchow triad: hypercoagulability, endothelial dysfunction, and/or hemodynamic changes (stasis).6,7 These factors may be further influenced by operative, postoperative, and patientrelated factors. Operative factors can include prolonged surgical time and traction time, whereas patient factors may include obesity, prothrombotic conditions, smoking, use of oral contraceptives, malignancy, and metabolic or cardiovascular disturbances. Finally, postoperative factors can include weight-bearing and mobility restrictions.7 Current literature on VTE events related to hip arthroscopy suggests a low reported incidence of VTE events. However, many previous studies have relied on symptomatic presentations in the reporting of the VTE incidence, with reported rates ranging from 1.4% to 3.7%.8,9 A recent prospective study by Mohtadi et al.,10 which used ultrasound, postoperatively found an overall higher incidence of DVT, at 4.4%, which included asymptomatic presentations. Within orthopaedic surgery, the risk of PE from asymptomatic proximal venous thrombosis (of popliteal or more proximal veins) has been reported to be about 25%, wherein a fatal PE occurs in 1% to 2% of patients in this group.11,12 Given the increasing use of hip arthroscopy, as well as the morbidities and deaths associated with both asymptomatic and symptomatic VTE, it is important to elucidate the incidence of these events and the role of prophylaxis in their prevention. Furthermore, improved identification of surgical and patient-related risk factors that may lead to VTE events may also aid in informing prophylaxis management. The purpose of this systematic literature review focused on hip arthroscopy was to (1) report the VTE event incidence in patients who receive VTE prophylaxis and those who do not, (2) report how VTE prophylaxis is currently being administered, and (3) report operative and patient-related risk factors for VTE identified in the literature. The hypothesis of this article was that VTE events would be rare reported
complications after hip arthroscopy that would occur in patients with identifiable risk factors.
Methods Design and Eligibility Criteria The research question and study eligibility criteria for this systematic review were established a priori. The patient population included male and female patients of all ages who underwent hip arthroscopy. The primary outcome measure was a VTE event in patients who received prophylaxis versus those who did not. The inclusion criteria were (1) all levels of evidence; (2) studies published in English, (3) studies conducted in humans that indicated whether VTE prophylaxis was or was not used, and (4) studies that reported whether VTE events occurred or did not occur. The exclusion criteria were (1) nonsurgical treatment studies (conservative treatment, technique articles without outcomes, and so on) and (2) studies in which the outcomes for the exact same patient population were reported in multiple articles (the most recent article was included). Search Strategy A systematic search strategy previously described by us was used.13,14 One reviewer (S.E.) searched 3 online databases (Embase, MEDLINE, and PubMed) for literature related to hip arthroscopy. The Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines were followed in the development of this study. The search was conducted on October 10, 2016, and retrieved articles from database inception to search date. To be more inclusive, the following key terms were used in a broad-based search: “hip” and “arthroscopy.” The search results were merged, and duplicates were removed. The search strategy is outlined in Appendix Table 1 (available at www.arthroscopyjournal.org). Study Screening Two reviewers (C.E.H. and S.E.) independently screened titles, abstracts, and full texts of retrieved studies. Discrepancies during the title and abstract screening stages were resolved by automatic inclusion to ensure thoroughness. Discrepancies during full-text screening were resolved through consensus between the reviewers. If a consensus could not be reached, a senior reviewer (D.D.) assisted to resolve the discrepancy. In addition, abstracts from American Orthopaedic Society for Sports Medicine; International Society of Arthroscopy, Knee Surgery & Orthopaedic Sports Medicine; European Society of Sports Traumatology, Knee Surgery and Arthroscopy; and American Academy of Orthopaedic Surgeons proceedings from 2011 to 2016 were also screened for relevant information. The references of included
3
VENOUS THROMBOEMBOLISM AFTER HIP ARTHROSCOPY
studies were searched to identify articles that warranted inclusion in this review. Quality Assessment of Included Studies A quality assessment of included studies was completed with the Methodological Index for Non-Randomized Studies (MINORS) criteria. MINORS is a validated scoring tool for nonrandomized studies. Each of the 12 items in the MINORS criteria is given a score of 0, 1, or 2, with maximum scores of 16 and 24 for noncomparative and comparative studies, respectively.15 Data Abstraction Two reviewers (C.E.H. and S.E.) independently abstracted relevant data from included articles and recorded the data in a Microsoft Excel spreadsheet (2013) designed a priori. Demographic information including author, year of publication, region, sample size, study design, level of evidence, patient demographic data (age, sex), and length of follow-up was recorded. Regardless of the study design being comparative or a single-group cohort, data on the use of preoperative VTE prophylaxis, procedures performed, traction use and duration, surgical time, VTE incidence, VTE risk factors (body mass index, use of anticoagulants, and so on), and postoperative rehabilitation were abstracted. Statistical Analysis A k value was calculated at each stage of article screening to evaluate inter-reviewer agreement. Agreement was categorized a priori as follows: k greater than 0.60 indicates substantial agreement; k of 0.21 to 0.60, moderate agreement; and k less than 0.21, slight agreement.16,17 An intraclass correlation coefficient was used to evaluate the inter-reviewer agreement of the MINORS scores. Weighted mean values with standard deviations are reported for patients’ baseline characteristics. For the purpose of meaningful quantitative pooling of the data, only studies with sample sizes of 30 or larger were included in the analyses. The number of hips was used in reporting pooled data from all included studies. To calculate the weighted proportions of VTE, the randomeffects DerSimonian-Laird model was used because of the inherent heterogeneity in the observational studies. The I2 value was calculated to quantify between-study heterogeneity. The weighted proportions and I2 values with 95% confidence intervals (CIs) are reported. An I2 of less than 0.25 is considered low heterogeneity; 0.25 to 0.5, moderate heterogeneity; and greater than 0.5, high heterogeneity.18 The StatsDirect website (www. statsdirect.com) was used in the statistical analysis.
Results Eligibility Assessment and Study Selection Our search yielded 5,746 studies, of which 21 met the eligibility criteria for this review (Fig 1). A list of
6683 Studies Iden fied Medline: 1088 Studies Embase: 2332 Studies Removal of duplicates
3334 Studies Title
2042 Studies Abstract
Removed: 3349
Removed: 1292 - Animal/cadaver/in vitro: 113 - Different/combined procedures: 111 - Non-opera ve: 215 Removed: 964 - Animal/cadaver/in vitro: 39 - Different/combined procedures: 139 - Non-opera ve: 71
1078 Studies Full Text Review
Removed: - No DVT prophylaxis data: 1058
20 Studies Hand Search of Literature and Full Text
Addi onal Studies Iden fied: 1
21 Studies Included
Fig 1. Screening strategy. (DVT, deep vein thrombosis.)
included studies can be found in Appendix Table 2 (available at www.arthroscopyjournal.org). Of the 21 included studies, all were published between 1998 and 2016 (Table 1). Most studies were based in North America (n ¼ 12), followed by Europe (n ¼ 6), with 1 each in Australia, South America, and Asia. No additional abstracts were found in a gray-literature search of American Orthopaedic Society for Sports Medicine; International Society of Arthroscopy, Knee Surgery & Orthopaedic Sports Medicine; European Society of Sports Traumatology, Knee Surgery and Arthroscopy; and American Academy of Orthopaedic Surgeons conferences from 2011 to 2016. There was substantial agreement between reviewers at the title (k ¼ 0.74; 95% CI, 0.71-0.76), abstract (k ¼ 0.78; 95% CI, 0.76-0.81), and full-text (k ¼ 0.97; 95% CI, 0.95-0.99) screening phases. Study Characteristics and Patient Demographic Data Most included studies were of Level IV evidence (n ¼ 14); there were 5 studies of Level III evidence and 2 prospective cohort studies (Level II evidence). There was substantial agreement between the reviewers for the quality assessment (intraclass correlation coefficient, 0.88; 95% CI, 0.84-0.94). The included studies had a mean MINORS score of 10.4 3.2, which indicates fair-quality evidence. Specifically, the nonecase report studies had a mean MINORS score of 11.2 3.4,
Authors, Year Alaia et al.,9 2014
Beutel et al.,19 2015
Domb et al.,22 2016 Fabricant et al.,23 2012 Fukushima et al.,24 2016
Horisberger et al.,25 2010
No. of Hip No. of Arthroscopies Patients Performed Mean Mean at Final During Age SD Follow-up Lost to Prophylaxis Follow-up Male Study Period (No. of Hips) Patients (Range), yr SD (Range) Follow-up Used (Type) d 139 (139) 39% 37.7 12.0 16.12 d None Yes (NSAID not specified)
d
1,054
d
1,155
d
98 (101)
150
18 (18)
28%
56 (41-68)
1,054 (1,054) 43%
37 (6-80)
39 (39)
29.4 mo None (18-54 mo)
6 wk
None
Yes (Aspirin 325 mg, daily for 2 weeks)
No
40%
39.5 (21-64)
23 mo None (12-43 mo)
Yes (Aspirin 325 mg, daily for 2 weeks)
921 (1,038) 26%
36.3 (13-76)
28.8 mo (23.569 mo)
18 mo None (12-30 mo)
Yes (Aspirin 325 mg, twice daily for 2 weeks) No
21 (27)
57%
17.6 1.6 (14.5-19.9)
72 (72)
40.3%
46.3 1.7
19 (19)
80%
47.3 (22-65)
7d
8.8%
None
No in 71, yes in 1
Yes (LMWH, 36 mo (18- 5%; 1 daily for 2 49.2 mo) patient weeks) died of cause unrelated to topic at hand (continued)
C. E. HALDANE ET AL.
Clarke et al.,20 2003 Collins et al.,21 2015
Average Study Level of MINORS Inclusion Exclusion Criteria Criteria Design Evidence Score Retrospective III 7.5 Hip arthroscopy with Patients who cohort postoperative were deemed ultrasound screen at high risk of VTE for DVT and given prophylaxis Prior surgery on Case control III 18 Patients undergoing ipsilateral side hip arthroscopy with traction with or without previous hip or knee arthroplasty None stated Retrospective III 9 Patients undergoing cohort hip arthroscopy between 1989-2001 Case control IV 19 Patients undergoing Previous hip surgery hip arthroscopy; (arthroscopic or BMI >30 for obese open) group Prospective II 11 Patients undergoing Previous hip cohort hip arthroscopy conditions (AVN, with minimum LCPD, EDS, PVNS) follow-up of 2 yr Case series IV 10 Athletically active Revision patients aged 19 arthroscopy yr undergoing hip arthroscopy for FAI Prospective III 9 Patients who Patients who were cohort underwent hip not able to undergo arthroscopy between preoperative and September 2012 and postoperative December 2015 ultrasound, who underwent combined open procedures, who were aged 18 yr Involvement in undergoing elective litigation and/or hip arthroscopy patients with acute fracture, dislocation or septic arthritis, drug addiction, underlying risk of thromboembolic event, or prior surgery on same hip Patients undergoing None stated arthroscopic FAI management Consecutive patients None stated undergoing hip arthroscopy Patients undergoing None stated hip arthroscopy
Lost to Follow-up
Prophylaxis Used (Type)
d
88 (105)
68%
40.9 (17-66) 27.6 mo (15.6- None 49.2 mo)
1,693
41 (42)
66%
65 (60-82) 30 mo None (12-54 mo)
145
115 (115)
48%
d
42 (42)
60%
35.1 (16-52) 28.2 mo None (10-72 mo)
Yes (not specified)
d
81 (81)
40%
32.2 (14-59) 8 d
None
No
d
194 (194)
59%
36.2 12.2 39.5 mo None (7-78) (4-103 mo)
No
39.15 (18-59)
3 mo
None
Yes (LMWH, daily for 2 weeks) Yes (Subcutaneous LMWH, 20 mg, at induction of anesthesia)
No
VENOUS THROMBOEMBOLISM AFTER HIP ARTHROSCOPY
Authors, Year
No. of Hip No. of Arthroscopies Patients Performed Mean Mean at Final During Age SD Follow-up Follow-up Male Study Period (No. of Hips) Patients (Range), yr SD (Range)
NOTE. The follow-up data presented for Fukushima et al.24 (7 d) and for Mohtadi et al.10 (3 mo) were both specific follow-up times and not means. AVN, avascular necrosis; BMI, body mass index; DVT, deep vein thrombosis; EDS, Ehlers-Danlos syndrome; FAI, femoroacetabular impingement; LCPD, Legg-Calves Perthes disease; LMWH, lowemolecular weight heparin; MINORS, Methodological Index for Non-Randomized Studies; NSAID, nonsteroidal anti-inflammatory drug; OA, osteoarthritis; PVNS, pigmented villonodular synovitis; SD, standard deviation; VTE, venous thromboembolism.
5
6
C. E. HALDANE ET AL.
whereas the case report studies had a mean MINORS score of 8.33 1.3. The nonecase report studies mostly lacked control groups, prospective sample size calculation, and unbiased assessment of endpoints. The case report studies also lacked appropriate endpoints and follow-up. Among the 21 included studies, there were 7 singlepatient case reports, 7 case series of 19 to 88 patients, and 7 comparative studies of 18 to 1,054 patients. The 7 single-patient case reports were excluded, and the remaining case series and comparative studies were included in the outcome analysis (Table 1). Among the 14 studies included in the outcome analysis, a total of 2,850 patients (2,985 hips) were available at final follow-up. The weighted mean follow-up period was 19 8 months, ranging from 7 days to 103 months. The weighted mean age was 40.7 7 years, ranging from 6 to 82 years, and 39.6% of patients were male patients (Table 1). VTE Incidence Table 2 presents the studies that measured VTE. The overall pooled proportion of VTE events after hip arthroscopy in 2,985 hips was 2.0% (95% CI, 0.01%-4.1%), with a total of 25 VTE events and between-study heterogeneity of 85.9% (95% CI, 73.1%-91.1%). There were 3 PE and 22 DVT events. To address the large heterogeneity, the studies that did not indicate the use of ultrasound for the diagnosis of VTE were removed from analysis, wherein the pooled VTE event rate increased to 4.2% (95% CI, 2.4%-6.5%) and heterogeneity decreased to 17.6% (95% CI, 0%-70%), which is considered low heterogeneity. A single study that reported that no VTE events occurred did not use ultrasound imaging for confirmation and thus relied on a lack of presentation of clinical symptoms in reporting the VTE event incidence.22 The pooled proportion of VTE events was 2.5% (95% CI, 0.6%-5.9%) in 1,947 patients after removal of this study, which also had 11% loss to follow-up. Diagnoses of DVT were made or confirmed using duplex Doppler ultrasound for 17 of the 25 VTE events. In 3 studies involving 326 hips, ultrasound was used after surgery to prospectively detect VTE events.10,24 Within these studies, 6 asymptomatic DVTs were detected out of the 12 total VTE events that occurred in this population. Thus, of the 25 VTE events in the included studies, 24% were asymptomatic. Time to diagnosis of VTE was highly under-reported and is presented, when reported, in Table 2. The incidence of VTE was 2.0% (95% CI, 0.3%-5.0%) in 9 studies with 1,443 hips in which VTE prophylaxis was used, whereas it was 2.3% (95% CI, 0.02%-6.6%) in 6 studies with 1,542 hips in which VTE prophylaxis was not used. In 1 study, prophylaxis was used for 1 hip with the remainder receiving prophylaxis.22 Most of the hips in the studies that did not use prophylaxis were from a single large study of 1,054 hips in which there
were no reported VTE events. With this study being excluded, the incidence of VTE in studies that did not use VTE prophylaxis (488 hips) was 3.6% (95% CI, 1.1%-7.6%). All studies reported that patients who had VTE events were treated with various forms of anticoagulation. VTE Prophylaxis The weighted proportion of DVT-VTE prophylaxis used in 8 of 14 studies encompassing 1,443 hips was 48.3% (Table 1). Within this group, 325 mg of aspirin was used in 1,095 hips, lowemolecular weight heparin was used in 166 hips; and nonsteroidal antiinflammatory drugs were used in 139 hips. The remaining studies, encompassing 43 hips, did not specify the type of prophylaxis used. The duration of prophylaxis was reported for 1,219 hips, with most studies reporting the duration of prophylaxis as 2 weeks. Patient and Operative Factors Procedures performed included femoroacetabular impingement resection, labral debridement and/or repair, and loose body removal. Within the 14 included studies, patient factors were identified in 15 of the 25 total VTE events (Table 2). The use of oral contraceptives was an identified factor in 5 patients. In addition, noneweight bearing postoperatively was an identified factor in 2 patients and partial weight bearing was a factor in 5. It is known that hypercoagulability and immobility may increase the risk of development of VTE.7 Traction was reportedly used in 1,734 hips. In studies that reported traction duration, the mean duration ranged from 28.6 to 58.9 minutes. When reported, the mean traction time in studies in which VTE events occurred (398 hips, 9 VTE events) varied from 46.5 to 58.9 minutes. A single study with 39 hips and no VTE events reported a mean traction duration of 28.6 minutes. The duration of surgery was also a highly underreported factor and was only reported in 3 studies (195 hips). In studies with no VTE events (123 hips), the duration of surgery varied from 54 to 130.7 minutes. In a single study with 5 VTE events (72 patients), the mean duration of surgery was 81.6 minutes. Postoperative Weight Bearing and Rehabilitation Weight-bearing status was reported for 2,442 hips (84.3%) in 12 studies. In the remaining studies, weight bearing was reported at the discretion of the surgeon or not specified. A total of 1,054 patients (1,054 hips) were deemed noneweight bearing for 72 hours after surgery, after which they were deemed weight bearing as tolerated. In 3 studies encompassing 166 hips, weightbearing parameters were provided for patients with microfracture, in whom partial weight bearing was
Table 2. VTE Events and Risk Factors Identified
39 921
2 DVTs (6.9%) 5 DVTs and 2 PEs (0.76%)
Obese (i.e., BMI >30) (2) d
72
5 DVTs (6.9%)
Increased mean age compared with non-DVT patients (5)
115
5 DVTs (4.4%)
No
81
3 DVTs (3.7%)
Souza et al.,29 2010 No
194
1 DVT (0.5%)
Use of oral contraceptive (2) PWB (5) Prolonged mean traction time and increased mean age compared with non-DVT patients (5) NWB (2) Use of oral contraceptive (1) No risk factors (1) d
Collins et al.,21 2015 Yes Domb et al.,22 2016 Yes
Fukushima et al.,24 2016
Mohtadi et al.,10 2016
Salvo et al.,8 2010
No in 71 patients, yes in 1 patient No
WB Status d
Traction Time to PE/DVT (Duration, min) Diagnosis Diagnosis Yes (mean, 58.9) Ultrasound screen (58.3%), Mean, 16.12 d clinical (symptomatic) (SD, 10.72 d)
72 h NWB, followed by WBAT d 8 wk PWB (20 lb), fitted brace limiting hip adduction and flexion d
Yes
At discretion of surgeon
Yes (mean, 49.5) Clinical (1 asymptomatic), physical examination, modified Wells criteria, ultrasound screen
d
d
Yes (46.5) d
Ultrasound screen d
Yes (45)
Ultrasound screen, CT Within 7 d screen, asymptomatic (5) postoperatively
d
Yes (
