Rectus Plication Does Not Increase Risk of Thromboembolic Events Following Abdominal Body Contouring: A Matched Case-Control Analysis

Abstract

Background

Rectus abdominis plication increases intra-abdominal pressure and lower-extremity venous stasis, which may increase the incidence of venous thromboembolism (VTE) events.

Objectives

The aim of this study was to investigate the potential association between VTE and rectus abdominis muscle plication during surgery.

Methods

A retrospective review of all patients who underwent abdominal body contouring at the authors’ institution between 2010 and 2020 was completed. Cases were those with postoperative VTEs and were matched to controls (1:4) via potential confounders. Variables collected include demographic data, operative details, comorbidities, and postoperative complications. Statistical analysis was performed with parametric, nonparametric, and multivariable regression modeling.

Results

Overall, 1198 patients were included; 19 (1.59%) experienced a postoperative VTE and were matched to 76 controls. The overall cohort was 92.7% female with an average age of 44 years, an average Charlson Comorbidity Index of 1 point, and an average BMI of 30.1 kg/m². History of cerebrovascular events (14.5% vs 36.8%, P = 0.026) differed significantly between cohorts, but no significant associations were noted in all other baseline demographics. Additionally, VTE cases were more likely to have received intraoperative blood transfusions (odds ratio = 8.4, P = 0.04). Bivariate analysis demonstrated cases were significantly more likely to experience concurrent complications, including delayed wound healing (0% vs 5.3%, P = 0.044), seroma formation (5.3% vs 21.1%, P = 0.027), and fat necrosis (0% vs 5.3%, P = 0.044). However, these findings were not significant in a multivariable regression model. Plication was not associated with VTE outcomes.

Conclusions

Rectus plication does not increase the risk of VTE. However, the odds of VTE are significantly increased in patients who received intraoperative blood products compared with those who did not.

Level of Evidence: 3

Abdominal contouring procedures, such as cosmetic abdominoplasty and panniculectomy, are routine operations for the removal of excess abdominal fat and skin and to strengthen the abdominal wall musculature.¹ About 100,000 abdominoplasties were performed in the United States in 2020—a 56% increase since 2000—making it one of the most commonly performed plastic and reconstructive procedures.² Unfortunately, abdominoplasties and panniculectomies are associated with higher complication rates than other surgeries, resulting in hematomas, infections, wound-healing problems, and, rarely, venous thromboembolism (VTE) events.^3,4 Previous studies have shown that the risk of complications increases in patients who are smokers, have multiple medical comorbidities, and are undergoing multiple concurrent aesthetic procedures.^4,5 VTEs, such as deep venous thrombosis (DVT) and pulmonary embolism (PE), are rare but potentially devastating complications of particular concern in this patient population. Although the risk of VTE ranges between 0.2% to 1.5% in abdominal body contouring procedures, the risk is markedly higher than that of other aesthetic procedures and VTEs remain a significant cause of morbidity and mortality.^6-11

Considerable effort has been employed in reducing the incidence of VTE. Risk factors, prophylaxis regimens, and procedural recommendations for thromboembolic events have been well characterized in other areas of surgery with guidelines revised by the American College of Chest Physicians.^12-14 However, these guidelines have yet to achieve consensus in plastic surgery because they contain minimal evidence from the plastic surgery literature and no plastic surgery procedures are included in the recommendations.^15,16 In the arena of abdominal contouring specifically, there are few studies examining preoperative risk factors, and none to date have investigated the association between procedural techniques and VTE outcomes.^10,17,18

Rectus abdominis plication, a commonly utilized method to reduce diastasis width and enhance abdominal contour, has been associated with increased intra-abdominal pressure (IAP) which may subsequently increase the risk of VTEs occurring.¹⁹ Previous investigations have demonstrated that increased IAP, via either plication or compression garments, may lead to lower-extremity venous stasis that persists for up to 48 hours after surgery.^20,21 Momeni et al utilized real-time ultrasound to demonstrate the occurrence of common femoral stasis with tight abdominal wall fascial closures irrespective of intraoperative hydration.^22,23 Some surgeons consider these findings potentially worrying as venous stasis is a key component of Virchow’s triad, a proposed mechanism of thrombosis formation.^24,25 On the other hand, other surgeons believe this risk to be more of a theoretical one, pointing to the fact that patients with epidural anesthesia or paraplegia (both with vessel dilation and venous stasis) are not at an elevated risk of VTE. Despite these concerns, there is a lack of literature investigating direct associations between rectus abdominis plication and VTE. To our knowledge, only 1 case report has examined thrombosis formation after rectus plication with significant findings that point to increased venous stasis.²¹ We aim to bridge this gap in knowledge by characterizing the effect of rectus abdominis plication, as well as other risk factors, on VTEs in patients undergoing abdominal contouring procedures.

METHODS

Approval was obtained from the IRB at our institution, Montefiore Medical Center (Bronx, NY), an academic hospital and tertiary referral center, before performing a retrospective review of all patients who underwent cosmetic abdominoplasty or panniculectomy at our institution between January 2010 and March 2020. Patients were identified via Common Procedural Terminology codes (CPT codes 15847, 17999, and 15830) and confirmed with the surgeon’s operative notes. Exclusion criteria include patients under the age of 18 years, emergent surgeries, and those undergoing concurrent major procedures such as resections of intra-abdominal malignancies. Electronic medical records were reviewed by all investigators to ensure inclusion criteria were met and for collection of all patient data.

Demographic data, including age, sex, race/ethnicity, smoking status, BMI, history of prior bariatric procedures, and medical comorbidities, were collected. The Charlson Comorbidity Index (CCI) and Caprini Risk Score were calculated from comorbidities existing at the time of surgery. The CCI is a previously validated score that is associated with postoperative complications, mortality, and readmission in patients undergoing surgery in a variety of disciplines.^26-29 The CCI predicts the 1-year mortality risk for patients with specific comorbid conditions. A total of 19 conditions are included in the CCI, including: age, cardiac risk, neurologic disease, respiratory disease, liver disease and renal disease, among others. Similarly, the Caprini Risk Score is a previously validated and widely used measure of VTE risk that takes into account both current comorbidities and previous health histories.^30,31 Operative details such as plication of the rectus abdominus muscles during the operation, concurrent procedures (liposuction, mastopexy, breast reduction mammoplasty, brachioplasty, and hernia repair), use of a sequential compression device (SCD), American Society of Anesthesiologists (ASA) score, operative time, and type/duration of chemophylaxis were noted.

We reviewed all consecutive patients and identified those who had experienced a VTE event such as a DVT or a PE within 90 days of their operation. Event status was confirmed via duplex Doppler ultrasound, ventilation-perfusion scans, computed tomography pulmonary angiography scans, pulmonary angiograms, and/or MRI scans. Cases were matched to controls who did not experience a VTE in a 1:4 fashion while controlling for potential confounders.^26,28,32 A 1:4 matching ratio was utilized to maximize statistical power in this rare disease scenario. Case-controls were matched on variables that differed significantly between case-control patients, including age, BMI, and preoperative comorbidities via the CCI.^26,28,32 Due to the limitations of sample size, case-controls were matched on potentially confounding characteristics instead of utilizing a propensity-scored model. Other outcomes of interest include operative characteristics, length of stay, unplanned 30-day readmission/reoperation, and 30-day postoperative complications. Specific complications of interest include hematoma formation, intraoperative transfusions, surgical site infection, hernia/bulge, delayed wound healing, wound dehiscence, seroma, umbilical necrosis, and fat necrosis.

Statistical analysis was performed with Stata v. 17.0 (StataCorp, College Station, TX). Univariate analysis and descriptive statistics were utilized to characterize the study population and to compare baseline cohort characteristics. Findings are presented as mean values with standard deviations (SDs) for variables meeting parametric assumptions, and as median values with interquartile ranges (IQRs) for variables that were nonparametric. Matching was completed with the “calipmatch” package with age, BMI, and CCI caliper widths set as 10, 3, and 3, respectively.³³ Bivariate analysis was performed with t-tests and chi-square analysis with Fisher’s exact test to identify associations between the case variable and other outcomes of interest, such as intraoperative transfusions, concurrent procedures, complications rates, length of stay, and duration of chemophylaxis. An unconditional logistic regression model was used to determine the odds of plication and other variables significant through bivariate analysis after adjusting for confounding variables such as varying baseline comorbidities via the CCI and varying intraoperative factors such as SCD usage. Conditional logistic regression modeling was employed to adjust for the impact of matching in the cohort with an adjusted odds ratio (OR) subsequently calculated for the case variable and other variables significant through bivariate analysis after adjusting for confounding variables such as varying baseline comorbidities via the CCI and varying intraoperative factors such as SCD usage. The CCI was included in the regression model instead of the Caprini Risk Score due to the wider breadth of medical conditions covered by it and because of the issue of multicollinearity between the indexes due to overlap of comorbid conditions encapsulated in each index measure. Conditional logistic regression is an extension of unconditional logistic regression modeling that allows one to take into account stratification and matching. In our model, utilizing a conditional logistic regression allowed for consideration of how the included participants were dispersed along the matching variables (age, BMI, CCI score) because any matching distorts how participants are naturally distributed among these variables in the unmatched cohort. We decided to include both the conditional and unconditional model in the results as debate remains regarding which is the superior model for matched data.³⁴ Statistical significance was set a priori at P < 0.05.

RESULTS

A total of 1265 patients underwent either abdominoplasty or panniculectomy in the 10-year period. After excluding those who did not meet inclusion criteria, 1198 patients were enrolled in our study: 1107 women (92.7%) and 87 men (7.3%) with an overall mean [SD] age of 44 [11.2] years (range, 18-75 years). The median CCI for all patents was 1 (IQR = 0-3) and the median Caprini Risk Score was 4 (IQR = 4-6). The baseline characteristics and demographic data of all subjects enrolled in the study are presented in Table 1.

Seventy-six patients without VTEs were matched to 19 case patients. The matched cohorts were similar in their demographic characteristics with no statistically significant difference in their age (47.9 vs 47.4 years, P = 0.864), BMI (31.6 vs 32.1 kg/m², P = 0.631), CCI (1 vs 1, P = 0.416), Caprini Risk Scores (4 vs 6, P = 0.135), duration of chemophylaxis (0 vs 0, P = 132), or operative time (280.8 vs 266.7 minutes, P = 0.684). There were no significant sex (P = 0.196) or race/ethnic (P = 0.131) differences between cohorts. Further analysis of patient demographics and previous medical history revealed no significant associations with VTE occurrence and previous history of smoking (2.6% vs 5.3%, P = 0.557), diabetes mellitus (22.4% vs 26.3%, P = 0.715), hypertension (36.8% vs 47.4%, P = 0.400), chronic obstruction pulmonary disease (2% vs 0%, P = 0.638), coronary artery disease (3.9% vs 5.3%, P = 0.597), chronic kidney disease (9.2% vs 5.3%, P = 0.597), steroid use (2.6% vs 0%, P = 0.638), immunosuppressant use (1.3% vs 0%, P = 0.800), radiation therapy (2.6% vs 5.3%, P = 0.492), chemotherapy (3.9% vs 10.5%, P = 0.260), and previous bariatric surgery (78.9% vs 68.2%, P = 0.151). Neither group included patients with concurrent hernia repair intraoperatively. Interestingly, a significant association was seen between VTE occurrence and a previous history of cerebrovascular accidents (14.5% vs 36.8%, P = 0.026) (Table 2).

Comparison of intraoperative characteristics demonstrated no significant associations between VTE occurrence and intraoperative plication (42.1% vs 42.1%, P = 0.990), location of surgery (40.8% vs 57.9% outpatient, P = 0.179), and presence of concurrent intraoperative procedures (65.8% vs 73.7%, P = 0.512). Importantly, there was a significant association between VTE occurrence and SCD use during surgery (88.2% vs 68.4%, P = 0.035). Postsurgical outcomes between cohorts were analyzed. No significant differences were observed in the majority of complications, including: hematoma formation (6.6% vs 10.5%, P = 0.556), unplanned return to the operating room (5.3% vs 10.5%, P = 0.399), scheduled revision surgery (1.3% vs 0%, P = 0.615), infection (3.9% vs 15.8%, P = 0.058), and wound dehiscence (7.9% vs 10.5%, P = 0.712). Bivariate analysis showed that patients with VTEs had a significantly higher incidence of readmission (6.6% vs 26.3%, P = 0.012), intraoperative blood transfusion (2.6% vs 15.8%, P = 0.022), delayed wound healing (0% vs 5.3%, P = 0.044), seroma formation (5.3% vs 21.1%, P = 0.027), and fat necrosis (0% vs 5.3%, P = 0.044) (Table 3).

Multivariable regression modeling was used to further delineate the associations between VTE occurrence and perioperative complications while adjusting for potential confounders. The unconditional model noted significantly higher odds of readmission (OR = 5.3, 95% CI = 1.2-23.5, P = 0.03), and intraoperative blood transfusion (OR = 8.4, 95% CI = 1.1-62.4, P = 0.04) in VTE-positive patients. However, the increased odds of blood transfusions did not reach a priori statistical significance when analyzed via the conditional regression model (P = 0.061). A significant association between readmission and VTE status was noted in the conditional model as well, with odds of readmission in VTE patients being 15.5 times that of non-VTE patients (95% CI = 1.5-158.9, P = 0.021). No significant differences were noted in either unconditional or conditional analysis between groups regarding hematoma formation, unplanned return to the operating room, infection, wound dehiscence, seroma formation, and fat necrosis. Plication was not significantly associated with VTE status on bivariate analysis or multivariable regression modeling (Table 4).

DISCUSSION

VTE events such as DVT and PE are an unfortunate but well-documented complication of abdominal contouring procedures such as abdominoplasty and panniculectomy^6-10. Although the risk is about 1%, VTEs represent a considerate cause of morbidity and mortality given the rapidly increasing prevalence and already common presence of these procedures.² There is sparse literature examining VTEs in abdominal contouring procedures and the role that rectus plication may play in increased VTE risk via increased IAP. Several authors have suggested that no increase in the IAP is associated with plication of the rectus abdominus muscle during abdominoplasty.^35,36 Other authors disagree with this point, and some believe that external compression postoperatively, such as with the use of compression garments, may actually increase the IAP more than muscle plication. To our knowledge, this study is the largest examination of VTE and its associated factors in abdominal contouring procedures in a cohort generalizable to the cosmetic abdominoplasty and panniculectomy patient population at large.³

The incidence of VTEs noted our study population was 1.59%, aligning with general trends and rates noted in the existing literature.^6-10 Similar to previous studies, pre-existing comorbidities and higher baseline risk were more prevalent in patients undergoing panniculectomies than in those undergoing abdominoplasties.³ These potential confounders were addressed via post-hoc matching of cases to controls, controlling for age, BMI, and comorbidity indices to generate comparison groups with similar baseline risk profiles. Post-match descriptive statistics demonstrated no significant differences between cohorts with the exception of previous cerebrovascular accidents—a known risk of VTEs which was adjusted for in subsequent analysis.³⁷ It seems likely that the clotting mechanism responsible for cerebrovascular accidents in some patients might also put them at risk for VTEs, which was supported by the results of the analysis.

SCDs, a commonly used tool for prevention of DVTs in surgery, were less likely to be utilized in the VTE cohort. This finding is expected and supports the validity of our study. In their widely cited meta-analysis, Urbankova et al found significant reductions in the number of DVTs in patients who utilized intermittent pneumatic compression devices compared with those who did not.^38-40 Current guidelines for SCD usage are derived mostly from examination of SCDs plus anticoagulation vs anticoagulation alone in surgical patients.^41-43 These results further emphasize the importance of SCD usage, a relatively easy and low-risk treatment that is a validated method for decreasing the risk of VTE in surgical patients.

This study did not find significant associations between rectus abdominis plication and VTEs despite proof of increased venous stasis.^22,23 As rectus abdominis plication is commonly performed in abdominal contouring procedures, it is imperative to establish the technique’s safety profile. Because increased IAP may lead to lower-extremity venous stasis that persists for up to 48 hours after surgery, rectus plication remains a possible contributor in the multifactorial nature of VTE genesis.^19-21,24 However, our results indicate that although a possible factor, plication alone is not a significant cause of VTEs in abdominal contouring patients. Unsurprisingly, 30-day readmission rates and duration of anticoagulation were significantly higher in the VTE group. Given the serious nature of these complications, patients often require extensive work-up, treatment, and monitoring that require careful inpatient medical care.⁴⁴ Unsurprisingly, the majority of admissions occurred after the VTE was noted for diagnosis and anticoagulation management, which likely contributed to this association.

Interestingly, unconditional analysis showed that patients who received intraoperative transfusions had an 8.4-fold increase in the odds of having a VTE compared to those who did not receive transfusions. Although generally considered safe and elective procedures, the risk of bleeding remains of special concern and patients who experience severe intraoperative blood loss may require transfusions. Blood products are lifesaving therapeutics with potential adverse effects including, but not limited to, febrile transfusion reactions, hemolytic transfusion reactions, transfusion-related acute lung injury, and transfusion-associated circulatory overload.^44-46 There are numerous proinflammatory factors and labile blood components that contribute to these overt syndromes and create an environment of heightened inflammation. These proinflammatory factors include cellular blood components, infectious pathogens, undesirable antibodies, as well as foreign leukocytes with high burdens of proinflammatory cytokines and chemokines.⁴⁷ Elevated inflammation has been linked as both a cause and consequence of VTEs and may be a main perpetrator of the pathology. Unlike other mechanisms of thrombosis involving vascular endothelial injury and subsequent exposure of procoagulant materials, vascular inflammation may initiate thrombosis of an intact vessel.^48-50 Additionally, blood products may cause hematologic abnormalities such as thrombocytopenia and coagulation factor derangements that may cause imbalances leading to thrombus formation and subsequent DVT/PE.⁴⁷ Given the potential link between intraoperative transfusions and postoperative VTEs, surgeons may consider incorporating anti-inflammatory or thromboprophylaxis strategies into the perioperative care of patients who require blood products.

There are several limitations to our study beyond its retrospective design. Due to the investigation being carried out at a single institution and the low incidence of VTEs, our case size was limited despite an impressive overall sample size. Thus, we may not have had sufficient power to detect associations between VTE occurrence and plication status as well as other related complications. Statistical power was improved via 1:4 case-control matching. Despite the significant association between intraoperative blood transfusion and VTE occurrence in the unconditional analysis, the P-value via the conditional analysis was slightly above the threshold for significance. We believe that this is likely the result of a lack of statistical power due to the rarity of the complication compounded with the decreased power inherent in conditional matched regressions. Additionally, detailed reasons for transfusions or the specific types of blood products used were not documented and could not be accounted for during analysis. Typically, when patients required multiple units of packed red cells, they required additional blood products such as plasma or platelets, which likely contributed to the statistically significant increase in blood products seen in the VTE group. The severity of VTEs could not be further explored or stratified for analysis due to sample size limitations. Future multicenter studies with larger cohorts should be pursued to enhance statistical power and allow for more nuanced analysis of related complications and the interplay between intraoperative transfusions and VTEs.

CONCLUSIONS

Our findings establish the safety profile of rectus abdominis plication, a commonly used technique in abdominal contouring procedures, after adjusting for potentially confounding factors via case-control match and regression analysis. However, the odds of VTEs are significantly increased in patients who received intraoperative blood products compared with those who did not. Physicians may consider the addition of anti-inflammatory and/or thromboprophylaxis strategies in patients receiving transfusions to decrease the risk of VTEs.

Disclosures

The authors declared no potential conflicts of interest with respect to the research, authorship, and publication of this article.

Funding

The authors received no financial support for the research, authorship, and publication of this article.

REFERENCES