Is Enoxaparin Associated With a Higher Risk of Persistent Wound Drainage Than Aspirin? A Secondary Analysis of Data From the CRISTAL Randomized Trial

Abstract

Background

Owing to its association with prosthetic joint infection, persistent wound drainage has become an important clinical entity after THA or TKA. The association between venous thromboembolism (VTE) prophylaxis and persistent wound drainage has not been extensively reported before but has potentially important clinical implications.

Questions/purposes

(1) Is the type of VTE prophylaxis (enoxaparin or aspirin) used after hip arthroplasty or knee arthroplasty associated with a higher risk of persistent wound drainage? (2) In patients who experience persistent wound drainage, is the type of VTE prophylaxis associated with a longer time taken to achieve a dry wound? (3) Is type of VTE prophylaxis associated with a higher risk of joint-related reoperation within 6 months?

Methods

This was a secondary analysis of data from an earlier cluster-randomized trial conducted through the Australian Orthopaedic Association National Joint Replacement Registry; data were drawn from two participating hospitals from that study. According to the trial’s allocation sequence, the two participating hospitals were randomized to administer aspirin (100 mg daily) or enoxaparin (40 mg daily) as VTE prophylaxis to all patients undergoing hip arthroplasty for 35 days after the procedure and for all patients undergoing knee arthroplasty for 14 days afterwards. Crossover to the alternate prophylaxis group occurred after the patient enrollment target had been met for the first arm. Between April 2019 and December 2020, 1339 of 1679 eligible patients were included in this study; 82% (707 of 861) of eligible patients were allocated to the enoxaparin group and 77% (632 of 818) of eligible patients we allocated to the aspirin group. The mean age in both groups was 67 ± 10 years and the mean BMI was 32 ± 7 kg/m². There was a higher proportion of male patients (43% [302 of 707] versus 36% [227 of 632]; p = 0.01), hip arthroplasties (36% [254 of 707] versus 29% [182 of 632]; p = 0.006), and patients receiving subcuticular closure (62% [441 of 707] versus 33% [208 of 631]; p < 0.001) in the enoxaparin group than in the aspirin group. Patients were monitored for wound drainage on each postoperative day until discharge, and this was recorded in the medical record once per day. Assessors were not blinded to the type of prophylaxis each patient received. Persistent wound drainage was defined as any wound drainage beyond Postoperative Day 3. For patients who experienced persistent wound drainage, the time taken to achieve a dry wound was defined as the number of days beyond Postoperative Day 3 for the wound to become dry. Logistic regression was used to determine whether the prophylaxis type was associated with persistent wound drainage. For patients with persistent wound drainage, the median time of drainage was compared between groups using the Kruskal-Wallis test. The number of patients undergoing a joint-related reoperation within 6 months was identified through data linkage to the Australian Orthopaedic Association National Joint Replacement Registry and electronic record review, and was compared using a Fisher exact test.

Results

We found no difference between the enoxaparin and aspirin groups in terms of the percentage of patients who had persistent wound drainage (9% [65 of 707] versus 8% [49 of 632], odds ratio 1.2 [95% confidence interval 0.8 to 1.8]; p = 0.40). For patients receiving subcuticular closure, after controlling for other potentially confounding variables, including age, sex, BMI, preoperative anticoagulant use, and type of arthroplasty, enoxaparin was associated with a higher risk of persistent wound drainage than aspirin (OR 3.6 [95% CI 1.5 to 10.6]; p = 0.009). For patients receiving a skin staple closure, after controlling for the same variables above, we found enoxaparin was not associated with a higher risk of persistent wound drainage (OR 1.1 [95% CI 0.7 to 1.9]; p = 0.66). For patients who experienced persistent wound drainage patients (114: 65 in the enoxaparin group and 49 in the aspirin group), there was no difference in the median (interquartile range) time taken to achieve a dry wound (enoxaparin: 1 day [IQR 1 to 2 days], aspirin: 1 day [IQR 1 to 3 days]; p = 0.22). There was no difference in the risk of joint-related reoperation within 6 months between enoxaparin (2.4% [17 of 707]) and aspirin (2.2% [14 of 632], OR 1.1 [95% CI 0.5 to 2.4]; p = 0.86).

Conclusion

Enoxaparin was not associated with an increased risk of persistent wound drainage compared with aspirin for all patients included in this study. Enoxaparin may be associated with a higher risk of drainage for patients receiving subcuticular closure. However, this finding should be interpreted cautiously, given the small sample size in this analysis. The duration of drainage was short regardless of the prophylaxis used, and enoxaparin was not associated with an increased risk of joint-related reoperation. These findings should not deter clinicians from using enoxaparin for VTE prophylaxis after hip or knee arthroplasty.

Level of Evidence

Level III, therapeutic study.

Introduction

Persistent wound drainage has become an important clinical entity after THA and TKA owing to its association with an increased risk of prosthetic joint infection [13] and increased length of stay [11]. However, the evidence describing the risk factors for and most effective treatment of persistent wound drainage is heterogeneous, owing to the lack of studies focusing on persistent wound drainage as a distinct entity from general postoperative wound complications and major bleeding events [2, 10, 18]. Additionally, few recent studies have investigated the relationship between the type of venous thromboembolism (VTE) chemoprophylaxis and persistent wound drainage [6, 11, 16]. The largest study compared aspirin and warfarin using a retrospective chart study and concluded that aspirin use was associated with a lower likelihood of persistent wound drainage; however, important confounders such as the method of wound closure were omitted from the analysis [16].

Secondary analyses of available randomized trial data offer a potential way around this problem. A cluster randomized, crossover, noninferiority trial of aspirin compared with low-molecular-weight heparin for VTE prophylaxis in hip or knee arthroplasty, a registry-nested study (CRISTAL) [4, 13], compared aspirin and enoxaparin in preventing symptomatic VTE after hip or knee arthroplasty [4]. This was conducted across 31 institutions in Australia, nested in the Australian Orthopaedic Association National Joint Replacement Registry. We believed this trial was a good venue to explore whether either prophylaxis agent was associated with an increased risk of persistent wound drainage after hip or knee arthroplasty because it allowed enrollment of a large number of patients across more than one site and the crossover design would theoretically allow each hospital to serve as its own control, reducing confounding from unmeasured variables.

We therefore asked: (1) Is the type of VTE prophylaxis (enoxaparin or aspirin) used after hip arthroplasty or knee arthroplasty associated with a higher risk of persistent wound drainage? (2) In patients who experience persistent wound drainage, is the type of VTE prophylaxis associated with a longer time taken to achieve a dry wound? (3) Is type of VTE prophylaxis associated with a higher risk of joint-related reoperation within 6 months?

Patients and Methods

Study Design and Setting

Before the commencement of patient enrollment into the CRISTAL trial, two participating institutions were approached and agreed to be included in this study [14, 15]. These centers were chosen because they are high volume (perform more than 500 THAs and TKAs per year), have excellent resources for collecting preoperative and postoperative data, have uniform methods of wound closure in their institutions, and use dedicated postoperative orthopaedic wards that allow wound monitoring.

Participants, Eligibility Criteria, and Randomization

At the patient level, all patients undergoing a hip or knee arthroplasty during the CRISTAL trial at these two institutions were included, provided they had correct documentation of their wound (Supplemental Digital Content 1; http://links.lww.com/CORR/B19). Patients were divided into two groups: those who were allocated to receive enoxaparin and those who were allocated to receive aspirin for VTE prophylaxis.

Each hospital was allocated to consecutive periods of a standard protocol of enoxaparin or aspirin, with the initial treatment order being randomized. Hospitals were the unit of randomization, and the two participating institutions were randomized according to the sequence generated by the CRISTAL trial. Crossover to the alternate study drug occurred once the sample size for the first arm had been met.

Preoperative data collected included patient characteristics (age, sex, BMI, American Society of Anesthesiologists classification), comorbidities (diabetes, smoking status, history of inflammatory arthropathy, liver disease [diagnosed chronic hepatitis or cirrhosis], cardiac disease [diagnosed ischemic heart disease, cardiac failure, or atrial fibrillation]), and preoperative anticoagulant or antiplatelet use (type, indication, and day ceased preoperatively). Potent preoperative anticoagulation was defined as a dual antiplatelet, direct oral anticoagulant; therapeutic heparin; therapeutic low-molecular-weight heparin; or warfarin. Operative and postoperative data included the type of arthroplasty (knee, hip), side, revision surgery, indication for surgery, tourniquet use, use of tranexamic acid, method of wound closure, and length of stay.

There were no important differences in terms of surgeon experience levels or volume. Both sites were considered high volume and all surgeons were considered high volume. Surgical approach and implant type were left to the discretion of the treating surgeon. Variables and data on operative time and estimated blood loss were not recorded. All patients received 1 g of cephazolin 30 minutes before skin incision and for an additional two doses 8 hours apart postoperatively. This was increased to 2 g for patients weighing greater than 120 kg. For patients who were allergic to cephalosporins or penicillin, vancomycin was administered as a single preoperative dose of 1 g. All wounds were closed in layers according to surgeon preference, with the main difference being the method of skin closure. Five of six surgeons at the first institution used skin staples for wound closure, and the remaining surgeon at the first institution along with all seven surgeons at the second institution used a buried, nonbarbed, running subcuticular suture for wound closure. No surgeons at either site used tissue adhesive (wound glue) for closure.

Patient Characteristics

The study was conducted between April 15, 2019, and December 20, 2020; during this period, 1679 patients underwent hip or knee arthroplasty at one of the two participating institutions. A total of 861 were allocated to receive enoxaparin and 818 were allocated to receive aspirin. Of these, 1339 patients were included in this study, with 340 excluded (186 allocated to aspirin and 154 allocated to enoxaparin). One patient in the aspirin group died before discharge, and 185 and 154 in the aspirin and enoxaparin groups, respectively, did not have wound documentation available for data analysis. There were 707 patients in the enoxaparin group and 632 patients in the aspirin group available for analysis at follow-up (Fig. 1).

Fig. 1.

This Consolidated Standards of Reporting Trials flowchart shows the trial enrollment and analysis.

There were higher proportions of female sex, preoperative aspirin use, knee arthroplasty, tourniquet use, and closure with staples in the aspirin group than in the enoxaparin group and higher proportions of revision arthroplasty in the enoxaparin group (Table 1).

Table 1.

Patient characteristics by prophylaxis type allocated

Characteristic	Enoxaparin (n = 707)	Aspirin (n = 632)	p value
Age in years, mean ± SD	67 ± 10	67 ± 10	0.58
Male sex, % (n)	43 (302)	36 (227)	0.01
BMI in kg/m2, mean ± SD	32 ± 7	32 ± 7	0.48
ASA classification, % (n)a
1	6 (42)	4 (25)	0.20
2	47 (334)	50 (313)
3	45 (319)	45 (289)
4	0.3 (2)	1 (5)
Diabetes, n (%)	16 (110)	19 (118)	0.15
Current smoker, % (n)	4 (30)	4 (24)	0.78
Inflammatory arthritis, % (n)	3 (18)	3 (20)	0.61
Liver disease, % (n)	3 (21)	3 (19)	1.0
Heart disease, % (n)	18 (130)	20 (128)	0.43
Preoperative anticoagulation, % (n)
Aspirin	12 (83)	16 (101)	0.008
Other single antiplatelet	2 (17)	1 (9)
Dual antiplatelet	1 (4)	2 (12)
Direct oral anticoagulant	5 (38)	7 (47)
Warfarin	0.4 (3)	1 (4)
Type of arthroplasty, % (n)
THA	36 (255)	29 (182)	0.006
TKA	64 (452)	71 (450)
Indication for primary arthroplasty, % (n)
Osteoarthritis	89 (628)	94 (593)	0.07
Avascular necrosis	2 (17)	2 (10)
Inflammatory arthropathy	2 (11)	2 (11)
Other	7 (51)	3 (18)
Revision, % (n)	4 (30)	2 (13)	0.03
Tourniquet use, % (n)	57 (405)	58 (369)	0.01
Tourniquet time in minutes (TKA only), mean ± SD	31 ± 33	37 ± 35	0.002
Tranexamic acid, % (n)	90 (638)	91 (574)	0.49
Closure, % (n)
Staples	38 (266)	67 (423)	< 0.001
Subcuticular	62 (441)	33 (209)

^aData missing for 10 patients within the enoxaparin group.

ASA = American Society of Anesthesiologists.

Interventions

Postoperatively, patients received either aspirin or enoxaparin as allocated by the CRISTAL trial. Both drugs were initiated within 24 hours of surgery. Aspirin was given at 100 mg daily (PO) and enoxaparin via subcutaneous injection at a dose of 40 mg daily, which was reduced to 20 mg for patients weighing less than 50 kg and for those with an estimated glomerular filtration rate of less than 30 mL/minute who were not on dialysis. Both drugs were continued for 14 days after TKA and for 35 days after THA. Patients not eligible to receive either study drug were those using preoperative anticoagulants (specifically a dual antiplatelet, direct oral anticoagulant; warfarin; or dual-antiplatelet therapy) or those with a medical contraindication (allergy or bleeding disorder precluding anticoagulation). Data on whether patients received either drug were collected for as-treated analyses.

Ward nursing staff, surgeons, and residents at each site were educated about the study and the definition used to determine persistent wound drainage before commencement. Patients had their dressings assessed daily by ward nursing staff, who documented the presence of persistent wound drainage in the patient record, using a form specific to this study (Supplemental Digital Content 1; http://links.lww.com/CORR/B19). Nursing staff were not blinded to treatment allocation. In patients who experienced persistent wound drainage, the time taken from Postoperative Day 3 to achieve a dry wound was recorded.

Outcome Measures

Previous studies have used the Southampton Wound Assessment Scale [6] or a definition of persistent wound drainage as a greater than 2 x 2-cm area on postoperative dressings beyond Day 3 (72 hours) postoperatively [18]. However, the Southampton Wound Assessment Scale considers other measures, such as wound bruising and erythema, rather than focusing solely on wound drainage, and the use of the 2 x 2-cm definition relies on accurate daily wound measurement. We defined persistent wound drainage as any wound drainage (regardless of amount) beyond Postoperative Day 3 (72 hours). We chose this definition after a discussion between a panel of orthopaedic surgeons and nursing unit managers. It retained the requirement of duration beyond Day 3 postoperatively; allowed an objective, clinically relevant measurement of persistent wound drainage; and avoided burdening nursing staff with measuring the drainage amount. For patients who experienced persistent wound drainage, the time taken to achieve a dry wound was defined as the time taken (in days) beyond Postoperative Day 3 for the wound to have no drainage. The treatment of patients with persistent wound drainage (dressings or surgical intervention) was left at the discretion of the treating surgeon, and surgeons were encouraged not to discontinue prophylaxis. Patients undergoing a joint-related reoperation within 6 months were identified via data linkage through the Australian Orthopaedic Association National Joint Replacement Registry and through electronic record review.

Ethical Approval

Ethical approval for this study was obtained from the South West Sydney Local Health District and the Sydney Local Health District (Reference X18-0424 and HREC/18/RPAH/603). The study was registered with the Australian and New Zealand Clinical Trials Registry (trial number: ACTRN12618001879257) before the study commenced.

Statistical Analysis

The proportions of patients experiencing persistent wound drainage in published studies vary considerably [1, 11, 17]. A recent review reported proportions varying between 0.2% and 20.1% [18]. Using an estimated proportion of patients with persistent wound drainage of 10% in the enoxaparin group, an 80% power level to detect an absolute between-group difference of 5% (10% in the enoxaparin group and 5% in the aspirin group), 434 patients in each group would be needed. The clinicians involved in the study considered this to be a clinically relevant difference.

Statistical analyses were performed using R (R Foundation for Statistical Computing Platform), version 4.1.2. We first conducted bivariable analyses between all collected variables and persistent wound drainage (the primary outcome) to identify which variables were predictors of persistent wound drainage. We used a Fisher exact test for categorical variables, and logistic regression and a t-test for continuous variables. The outcomes are reported using a point estimate (odds ratio) with a 95% confidence interval.

We then performed multivariable analyses to identify whether VTE prophylaxis or other collected variables were independent predictors of persistent wound drainage. Prophylaxis type and variables with a p value of less than 0.25 (on bivariable analyses) were entered into a backward stepwise multivariable regression model using the Akaike information criterion to select the final model. As the exposure of interest, prophylaxis type was forced into all models. Because each prophylaxis group had uneven proportions of patients receiving subcuticular closure and undergoing hip arthroplasty, potential interaction terms for prophylaxis type, method of closure, and arthroplasty type (hip or knee arthroplasty) were tested for significance. If significant, these interaction terms were retained in the model.

Multiple imputation using chained equations was used to address missing values, and model averaging from each individual imputed dataset was used for the multivariable analyses.

To determine whether prophylaxis type was associated with an increased time taken to achieve a dry wound for patients who experienced persistent wound drainage, the median time to achieve a dry wound was compared between groups using the Kruskal-Wallis test.

To determine whether prophylaxis type was associated with an increased risk of joint-related reoperation within 6 months, a Fisher exact test was used to test for a difference between groups. The outcomes are reported using a point estimate (OR) and 95% CI.

The primary analyses were performed according to allocation group. Sensitivity analyses included a comparison of allocation group and as-treated analyses, and a comparison of complete case and imputed analyses.

Results

Association Between Prophylaxis Type and Persistent Wound Drainage

We found no difference between the enoxaparin and aspirin groups in terms of the percentage of patients who had persistent wound drainage (9% [65 of 707] versus 8% [49 of 632]; OR 1.2 [95% CI 0.8 to 1.8]; p = 0.40). Because of significant interaction terms between the closure method and prophylaxis type as well as between closure method and joint type, we could not interpret the association between prophylaxis type and persistent wound drainage for the full model (Table 2). Therefore, we constructed separate models stratified by the method of wound closure that was common to both significant interaction terms. The interaction term between allocation and joint type was not significant.

Table 2.

Results of the multivariable regression analysis for persistent wound drainage

Parameter	OR (95% CI)	p value
Allocation group
Aspirin	Reference
Enoxaparin	1.1 (0.7 to 1.9)	0.63
BMI per kg/m2 increase	1.04 (1.02 to 1.08)	0.003
Heart disease	0.6 (0.3 to 1.1)	0.07
Preoperative potent anticoagulation	4.0 (1.8 to 8.8)	0.001
Preoperative single antiplatelet therapy	2.9 (1.6 to 5.1)	< 0.001
Skin closure
Staples	Reference
Subcuticular	0.1 (0.02 to 0.2)	< 0.001
Arthroplasty type
THA	Reference
TKA	0.4 (0.2 to 0.6)	< 0.001
Allocation*closure	3.1 (1.1 to 10.2)	0.04
Joint type*closure	5.4 (2.2 to 14.6)	0.001

The terms “allocation*closure” and “joint type*closure” refer to the interaction terms between these variables included in the model.

For patients who had subcuticular closure (n = 649), the proportion of patients experiencing persistent wound drainage in the enoxaparin group was 7.7% (34 of 441) and was 2.4% (five of 209) in the aspirin group. Controlling for other potentially confounding variables including age, sex, BMI, preoperative anticoagulant use, and type of arthroplasty, we found that enoxaparin was associated with higher odds of persistent wound drainage (OR 3.6 [95% CI 1.5 to 10.6]; p = 0.009). Knee arthroplasty (compared with hip arthroplasty) was the only other factor associated with higher odds of persistent wound drainage in this stratum (Table 3).

Table 3.

Results of the multivariable regression analysis for persistent wound drainage for patients by method of closure

Parameter	OR (95% CI)	p value
Subcuticular closure
Allocation group
Aspirin	Reference
Enoxaparin	3.6 (1.5 to 10.6)	0.009
Female sex	0.6 (0.3 to 1.2)	0.13
Arthroplasty type
THA	Reference
TKA	2.6 (1.2 to 6.1)	0.02
Skin staple closure
Allocation group
Aspirin	Reference
Enoxaparin	1.1 (0.7 to 1.9)	0.66
Age per year increase	1.04 (1.01 to 1.07)	0.02
BMI per unit kg/m2	1.08 (1.04 to 1.12)	< 0.001
Heart disease	0.5 (0.2 to 1.1)	0.07
Preoperative potent anticoagulation	5.7 (2.1 to 14.9)	< 0.001
Preoperative single antiplatelet therapy	3.2 (1.6 to 6.3)	0.001
Arthroplasty type
THA	Reference
TKA	0.3 (0.2 to 0.5)	< 0.001

For patients who had skin staple closure (n = 689), the proportion of patients experiencing persistent wound drainage was 12% (31 of 266) in the enoxaparin group and 10% (44 of 423) in the aspirin group. Controlling for the same variables above, enoxaparin was not associated with a higher risk of persistent wound drainage (OR 1.1 [95% CI 0.7 to 1.9]; p = 0.66). Hip arthroplasty, age (per year), BMI (per kg/m²), and preoperative single antiplatelet or potent anticoagulant therapy were associated with higher odds of persistent wound drainage in this stratum (Table 3).

The as-treated and complete case analyses (nonimputed dataset) showed similar findings to the main analysis (data not shown).

Time Taken to Achieve a Dry Wound

For the patients who experienced persistent wound drainage (114 in total: 65 in the enoxaparin group and 49 in the aspirin group), there was no difference in the median (interquartile range) time taken to achieve a dry wound (enoxaparin: 1 day [IQR 1 to 2 days], aspirin: 1 day [IQR 1 to 3 days]; p = 0.22).

Joint-related Reoperation Within 6 Months

We found no difference between the enoxaparin and aspirin groups in the percentage of patients who had joint-related reoperation within 6 months (2.4% [17 of 707] versus 2.2% [14 of 632], OR 1.1 [95% CI 0.5 to 2.4]; p = 0.86). In the enoxaparin group, 1.6% of patients (11) underwent manipulation under anesthesia for arthrofibrosis, 0.6% (four patients) underwent debridement and liner exchange (two for prolonged wound oozing and two for acute deep wound dehiscence), 0.1% (one patient) underwent debridement for a superficial wound collection, and 0.1% (one) underwent debridement to remove a retained drain. In the aspirin group, 1.4% of patients (nine) underwent manipulation under anesthesia for arthrofibrosis, 0.6% (four) underwent a debridement and liner exchange (three for prolonged wound oozing and one for an acute infection on Day 28), and 0.2% (one) underwent debridement for a superficial wound infection.

Discussion

Persistent wound drainage after hip and knee arthroplasty is recognized as an important postoperative complication because of its association with prosthetic joint infection [13, 18]. This study demonstrated that enoxaparin was not associated with an increased risk of persistent wound drainage compared with aspirin. In a secondary analysis, however, we found enoxaparin was associated with an increased risk of persistent wound drainage in patients who had subcuticular closure. The duration of wound drainage was short in both groups and enoxaparin did not result in an increased risk of joint-related reoperation. Given the benefits of enoxaparin in reducing symptomatic VTE after THA and TKA [4], these findings should not deter clinicians from using enoxaparin for VTE prophylaxis.

Limitations

First, an interpretation of the analysis between prophylaxis type and persistent wound drainage was not possible because of the unexpected effect modification caused by the method of wound closure. This required a stratified analysis, which reduced the effective sample size of the study. Because of this, the results should be interpreted cautiously. We feel the finding of effect modification is of clinical interest, because it suggests that the effect of prophylaxis type depends on the type of wound closure used. Second, some variables (such as preoperative anticoagulation, type of arthroplasty, and method of closure) were not balanced between prophylaxis groups. However, the multivariable analyses adjusted for these differences.

Third, data were not available on whether VTE prophylaxis was withheld for patients who experienced persistent wound drainage. Withholding prophylaxis may have affected the time taken to achieve a dry wound. Fourth, conducting this study at only two of the 31 participating sites may limit the generalizability of the results. Neither institution used tissue glue or a barbed, subcuticular running suture in their method of wound closure, both of which have been associated with a decreased risk of persistent wound drainage [12]. Further studies are required to determine the effect of this method of skin closure on the association between prophylaxis type and persistent wound drainage.

Fifth, the method of measuring persistent wound drainage in this study was not validated. Our results rely on a specific definition of wound ooze, and the findings may have been different if other measures had been used. However, we consider our definition to be clinically appropriate because of the limitations of the other classifications, as described in the Methods section of this article.

Finally, there is an increased risk of persistent wound drainage after revision arthroplasty [18]. Because of the small number of revision procedures included in this study, the effect of revision arthroplasty on wound drainage could not be determined.

Association Between Prophylaxis Type and Persistent Wound Drainage

We found no increased risk of persistent wound drainage for patients receiving enoxaparin, but an increased risk for patients receiving enoxaparin and subcuticular closure. Three previous studies have investigated the association between persistent wound drainage and VTE prophylaxis in patients undergoing hip or knee arthroplasty [6, 11, 16]. Jones et al. [6] published a prospective study of 327 patients, comparing patients who received enoxaparin or aspirin with patients who received no chemoprophylaxis. The authors found no difference in the risk of any postoperative wound drainage when comparing enoxaparin and aspirin. No comparative analysis of wound drainage risk beyond Day 3 was performed. The remaining two studies were retrospective. The first study [11] collected data on patients between 1997 and 2004, before the routine use of tranexamic acid, and compared aspirin with warfarin or enoxaparin. That study reported that enoxaparin was associated with a higher risk of drainage until Postoperative Day 5 (in 23% of patients), compared with 13% and 17% in the aspirin and warfarin groups, respectively. The second study compared aspirin with warfarin [16] and reported that the risk of wound drainage beyond Day 3 postoperatively was higher with warfarin than with aspirin (8.5% versus 3.2%). However, in each these studies, the administered prophylaxis was at the discretion of the treating surgeon, and important confounders such as the method of wound closure were not included. Our study is the most recent to compare enoxaparin and aspirin for persistent drainage beyond Day 3 and includes adjustment for important confounders [3, 5, 7, 9, 13].

Time Taken to Achieve a Dry Wound

For patients who experienced persistent wound drainage, we found a short median time taken to achieve a dry wound for both prophylaxis groups, regardless of closure method. Most of these patients in both groups had a dry wound by Postoperative Day 5. This is clinically important, because some have recommended consideration of surgical intervention (debridement with or without liner exchange) if wound drainage persists beyond Postoperative Day 7 [8, 18]. Our findings signify that any drainage beyond Day 3 associated with enoxaparin is likely to cease with nonoperative care and does not increase the risk of joint-related reoperation within 6 months. Clinicians should not be deterred from using enoxaparin for VTE prophylaxis, given its benefits in reducing symptomatic VTE within 90 days [4]. The time taken to achieve a dry wound is an outcome that has rarely been reported in studies investigating persistent wound drainage. Patel et al. [11] reported that enoxaparin was associated with a longer time to achieve a dry wound by Postoperative Day 5; however, this association was not seen by Postoperative Day 8. This study was conducted prior to the routine use of tranexamic acid and when the standard dosage of enoxaparin was 30 mg twice daily. A second study reported that warfarin was associated with a higher risk of persistent wound drainage than aspirin but was not a risk factor for a longer time to achieve a dry wound [13]. Neither study included the method of closure as a confounder. Both studies are consistent with the results of this study; although more potent forms of anticoagulation may increase the risk of persistent wound drainage, they were not found to be associated with a longer time to achieve a dry wound. Future studies should aim to determine whether the duration of drainage is associated with an increased risk of joint-related reoperation.

Conclusion

We demonstrated that enoxaparin was not associated with a higher risk of persistent wound drainage than aspirin for all patients included in this study. The secondary analysis demonstrated patients receiving enoxaparin and subcuticular closure experienced a higher risk of wound drainage; however, these findings should be interpreted cautiously because of the small sample size in this analysis. The duration of wound drainage beyond Day 3 was short regardless of the VTE prophylaxis that was used, and there was no association between prophylaxis and joint-related reoperation within 6 months. In conjunction with the results of the CRISTAL study [4], these findings should not deter clinicians from using enoxaparin for VTE prophylaxis after hip and knee arthroplasty. Future studies should aim to control the method of wound closure to reduce confounding.