Lymphocele or Seroma After Modied Radical Mastectomy for Breast Cancer: Systematic Review and Meta-analysis.

Seroma or lymphocele remains the most common complication after mastectomy and lymphadenectomy for breast cancer. Many different techniques are available to prevent this complication. We searched MEDLINE, clinicaltrials.gov, Cochrane Library, and Web of Science databases for publications addressing the issue of prevention of lymphocele or seroma after mastectomy and axillary lymphadenectomy. Quality was assessed using a standardised tool. Incidence of seroma or lymphocele were collected. Fifteen randomized controlled trials including a total of 1,766 patients undergoing radical mastectomy and axillary lymphadenectomy for breast cancer were retrieved. The incidence of lymphocele or seroma in the study population was 24.2% (411/1698): 25.2% (232/920) in the test groups and 23.0% (179/778) in the control groups. Neither modication of surgical technique (RR= 0.86; 95%CI [0.72, 1.03]) nor application of a medical treatment (RR= 0.96; 95%CI [0.72, 1.29]) was effective in preventing lymphocele. On the contrary, decreasing the drainage time increased the risk of lymphocele (RR= 1.88; 95%CI [1.43, 2.48). To conclude, despite the heterogeneity of study designs, drainage appears to be the most effective technique. all techniques tested to reduce seroma or lymphocele after breast surgery are based on the different physiological theories. Six studies tested a drug for prevention of seromas. These drugs inhibit the inammatory or immunopathological response, which is considered to play a role in seroma formation 36 . Four studies evaluated a specic surgical procedure. A French multicenter, superiority, randomized controlled trial, compared seroma formation using quilting suture versus conventional closure with drainage in 320 patients undergoing mastectomy 37 , results have not yet been published. A meta-analysis by Sajid et al. studied application of brin glue under skin aps to prevent seroma-related morbidity following breast and axillary surgery 8 , but this technique failed to reduce the incidence of postoperative seroma (RR: 1.02; 95%CI: 0.90 - 1.16, P value = 0.73). Four studies included in our meta-analysis evaluated modication of the drainage technique. Since 1947 and the rst description of drainage after axillary dissection for breast cancer by Murphey 38 , drainage is the technique most commonly used to prevent lymphocele or seroma after radical mastectomy and axillary lymphadenectomy. In 2013, a Cochrane meta-analysis by Thomson et al. 3 compared wound drainage versus no wound drainage after axillary lymphadenectomy for breast carcinoma. Seven RCTs including 960 participants were identied. The quality of trials was generally low, with several studies at risk of selection bias, and no studies used blinding during treatment or outcome assessment. There was a high level of statistical variation between studies, which therefore reduces the reliability of the evidence. The R for seroma formation was 0.46 ([95%CI: 0.23 to 0.91], P = 0.03) in favor of a reduced incidence of seroma in participants with drains inserted.


Introduction
Axillary lymph node dissection (ALND) and mastectomy are performed as part of the surgical management of breast cancer and are associated with signi cant morbidity, as 70% of patients experience complications 1,2 .
Seromas or lymphoceles are the most common complication of these procedures and can delay local healing and initiation of adjuvant therapy. They are also a source of discomfort for patients. Many techniques have been developed to decrease the risk of seroma formation: wound drainage 3 , reduction of the dead space by ap xation 4 , use of various types of energy 5 , external compression dressings 6 , shoulder immobilization or physical activity 7 , as well as numerous drugs and glues [8][9][10][11] .
Two previous Cochrane meta-analyses have evaluated brin glues and wound drainage and concluded on the ine cacy of brin glues and the limited e cacy of wound drainage 3,8 . To our knowledge, no meta-analysis has compared all proposed techniques for seroma prevention after mastectomy and axillary lymphadenectomy.
The objective of this meta-analysis was to determine the incidence of lymphocele or seroma after mastectomy and axillary lymphadenectomy based on a review of the published data by taking all techniques into account and to evaluate the e cacy of each technique.

Materials And Methods
This meta-analysis was performed in accordance with the 2009 Preferred Reporting Items for Systematic Reviews and Metaanalysis (PRISMA) guidelines and the Cochrane Collaboration recommendations 12 . The "Prevention of seroma after breast cancer surgery" trial was registered on the Open Science Framework (OSF) platform DOI 10.17605/OSF.IO/RFVG6.

Literature search
We searched MEDLINE, clinicaltrials.gov, Cochrane Library, and Web of Science databases for publications of randomized controlled trials (RCT) and clinical trials addressing the issue of prevention of lymphocele or seroma after mastectomy and axillary lymphadenectomy. Various combinations of the following terms were searched: "lymphocele", "lymphorrhea", "seroma", "breast cancer", "breast surgery".

Eligibility criteria
Three authors independently conducted the initial research to evaluate eligibility criteria (AC, MLB, KM). We selected randomized controlled trials and clinical trials published after January 2000 in English, including more than 50 participants, reporting the incidence of lymphocele or seroma after mastectomy and axillary lymphadenectomy for breast cancer. The latest search was performed in March 2021.
The following publications were excluded: retrospective studies, case reports, letters to the editor, publications concerning plastic surgery, brachytherapy or radiation therapy.

Data collection process and outcome measures
Three authors independently performed data collection using a standardized data extraction table (AC, MLB, KM). The following data were extracted: author, year and country of publication, study characteristics, prevention technique, inclusion and exclusion criteria, number of patients, data necessary to build 2 x 2 contingency tables.

Statistical analysis
Publication bias A funnel plot was used to visualize publication bias. The estimate of the difference between groups was pooled, depending upon the effect weights of the variance estimate determined in each trial. Egger's test was used to assed asymmetry of the funnel plot 13 .

Outcomes
For dichotomous outcomes, the Mantel-Haenszel method was used for calculation of relative risk (RR) under the xed-effect and random-effects models 13 . The Forest plot was used for graphic display of the results of the meta-analysis. The heterogeneity of studies was calculated using the I 2 index. The I 2 value was interpreted by balancing the direction and magnitude of I 2 with its statistical signi cance, using the values in the Cochrane Handbook for Systematic Reviews of Interventions as a guide 14 : 0% to 40%: might not be important; 30% to 60%: may represent moderate heterogeneity; 50% to 90%: may represent substantial heterogeneity; 75% to 100%: represents considerable heterogeneity. Meta-analyses with insigni cant heterogeneity were calculated using the xed-effects model 15 . For meta-analyses with low or moderate heterogeneity, the random-effects model was used 16 . The square around the estimate represents the accuracy of the estimation (sample size) and the horizontal line represents the 95% con dence interval (95%CI).
Data were entered in an Excel le and all statistical analyses were performed using Rstudio software (RStudio, PBC, Boston, U.S.A.). A P value < 0.05 was considered to be statistically signi cant.

Quality assessment of the studies included
We used a quality assessment tool elaborated by Hawker et al. 17 in 2002 for systematic review of qualitative evidence. The scale contains nine items assessing abstract/title, introduction/aims, method/data, sampling, data analysis, ethics/bias, results, transferability and implications. Each item is rated as "good", "fair", "poor" and "very poor". Lorenc et al. 18 added a graduation to this scale by assigning answers from 1 point (very poor) to 4 points (good), to provide a nal score for each study (9 to 36 points). The overall quality grades were de ned by the following description: grade A (high quality) 30-36 points; grade B (medium quality), 24-29 points and grade C (low quality), 9-24 points. Each of the three readers assessed the studies independently. When differences were observed, a majority agreement was reached.

Study selection
The PRISMA ow diagram explaining the literature search strategy and trial selection is presented in Figure 1. Fifteen randomized controlled trials including a total of 1,766 patients undergoing radical mastectomy and axillary lymphadenectomy for breast cancer were retrieved from the electronic databases. Analysis was based on 822 patients in the test groups and 778 patients in control groups. The characteristics of the trials included in this meta-analysis are provided in Table 1. The technique used in each article is described in Table 1. The incidence of lymphocele or seroma in the study population was 24.2% (411/1698): 25.2% (232/920) in the test groups and 23.0% (179/778) in the control groups.
As the study by Dalberg et al. 19 compared two different techniques in two separate groups of patients, we decided to divide this study into one group treated by drainage and the other group treated by the fascia preservation surgical technique.

Study characteristics
Study characteristics are described in Table 1. Two of the 15 studies concerned lymphoceles 20,21 , while all of the other studies concerned seromas. Six studies did not specify their de nition of seroma, 9 studies reported a clinical de nition of seroma or lymphocele (palpation, clinical examination, needle aspiration) and one study used an ultrasound de nition. Five studies reported statistically signi cant results 19,[22][23][24][25] .

Publication bias
The funnel plot did not show any asymmetry (Supplemental Figure 1). Egger's test did not reveal any publication bias (p=0.36). Other techniques: One study that investigated prevention of lymphocele or seroma using brin glue 33  One study that investigated prevention of lymphocele or seroma using physical activity and manual lymphatic drainage 34 found these technique to be statistically ineffective (RR=1.67; 95%CI [0.44; 6.29]).

Study quality
The results of the quality assessment are described in Supplemental table 1. One study was considered to present high quality (Grade A), 8 studies were considered to present medium quality (Grade B), and 6 studies were considered to present low quality (Grade C).

Discussion
This work represents the rst meta-analysis of all techniques proposed for the prevention of lymphocele formation after mastectomy and axillary lymphadenectomy in prospective randomized controlled trials and clinical trials. Global analysis of all of the various techniques showed that they were not effective to prevent lymphocele formation (RR=1. 23 . The overall quality of these items was moderate with 8 items presenting average quality, 6 items presenting low quality, and only one item presenting high quality. In this study, we chose to restrict our analysis to the population at high risk of lymphocele or seroma 35 . In our meta-analysis, regardless of the de nitions and techniques used to prevent seroma or lymphocele, the overall incidence of these complications was 24.2% (411/1698): 25.2% (232/920) in the test groups and 23.0% (179/778) in the control groups. The reported seroma or lymphocele incidence is dependent on the author's de nition of seroma or lymphocele and the method of detection used. Risk factors for seroma formation include age, body mass index (BMI), tumor size, use of neoadjuvant chemotherapy, type of surgery (MRM versus breast-conserving surgery) 35 , axillary lymph node status, axillary lymph nodes sampled or removed, and subsequently the extent of surgical dead space produced 36 . In our meta-analysis, only one article 33 considered neoadjuvant chemotherapy to be an exclusion criterion, while most of other studies did not mention neoadjuvant chemotherapy. Other risk factors, except for the type of surgery, were not well documented. This lack of information on risk factors may result in an incidence bias.
The various techniques tested to reduce seroma or lymphocele after breast surgery are based on the different physiological theories. Six studies tested a drug for prevention of seromas. These drugs inhibit the in ammatory or immunopathological response, which is considered to play a role in seroma formation 36 . Four studies evaluated a speci c surgical procedure. A French multicenter, superiority, randomized controlled trial, compared seroma formation using quilting suture versus conventional closure with drainage in 320 patients undergoing mastectomy 37  Four studies included in our meta-analysis evaluated modi cation of the drainage technique. Since 1947 and the rst description of drainage after axillary dissection for breast cancer by Murphey 38 , drainage is the technique most commonly used to prevent lymphocele or seroma after radical mastectomy and axillary lymphadenectomy. In 2013, a Cochrane meta-analysis by Thomson et al. 3 compared wound drainage versus no wound drainage after axillary lymphadenectomy for breast carcinoma. Seven RCTs including 960 participants were identi ed. The quality of trials was generally low, with several studies at risk of selection bias, and no studies used blinding during treatment or outcome assessment. There was a high level of statistical variation between studies, which therefore reduces the reliability of the evidence. The R for seroma formation was 0.46 ([95%CI: 0.23 to 0.91], P = 0.03) in favor of a reduced incidence of seroma in participants with drains inserted.
Finally, wound drainage appears to be the most effective way to prevent seroma, although no consensus has been reached concerning the optimal duration of drainage. However, persistence of foreign devices under the skin could predispose to surgical site infection. Surgical site infection is one of the possible complications after breast cancer surgery, causing signi cant morbidity, additional costs and which can delay initiation of adjuvant therapy. In Reiffel's review 39 of the potential association between closed-suction drains and surgical site infection, few studies suggested an increased risk of surgical site infection associated with drain placement and no studies attributed a decreased incidence of surgical site infection (including organ/space surgical site infection) with drain placement.

Conclusion
The lack of consensus concerning the de nition of lymphocele or seroma is probably responsible for the heterogeneity of seroma incidence reported in the literature and the ine cacy of the techniques proposed for seroma prevention after breast cancer surgery. However, drainage is the most effective technique currently available. Forest plot for seroma incidence following application of a treatment designed to prevent lymphocele after mastectomy with axillary lymphadenectomy. Risk ratios are shown with 95% con dence intervals.

Figure 3
Forest plot for seroma incidence following application of a medical treatment designed to preventing lymphocele after mastectomy with axillary lymphadenectomy. Risk ratios are shown with 95% con dence intervals.

Figure 4
Forest plot for seroma incidence following application of a surgical technique designed to prevent lymphocele after mastectomy with axillary lymphadenectomy. Risk ratios are shown with 95% con dence intervals