Surgical Lymphedema Treatment: A Meta-Analysis and Recommendations

Lymphedema is a chronic disease marked by lymphatic fluid accumulation that causes swelling and tissue changes, skin discoloration, limb heaviness, altered sensation, and pain.¹ Lymphedema can be classified into either primary (genetic) or secondary (acquired) lymphedema. Secondary lymphedema is caused by an injury, insult, or obstruction to the lymphatic system, usually due to surgical excision of lymph nodes or some form of medical therapy.¹ Upper extremity lymphedema is commonly seen in patients with breast cancer as a postoperative complication of mastectomy, axillary lymph node dissections, and, to a lesser degree, sentinel lymph node dissections.² Up to 65% of women undergoing treatment for breast carcinoma experience lymphedema.^3-5

Recent advances in treating lymphedema have sought to reduce the volume of lymphedematous limbs and improve patient satisfaction.⁶ These newer treatments include but are not limited to lymphovenous bypass, vascularized lymph node transplant (VLNT), lymphaticovenous anastomosis (LVA), vascularized groin lymph node transfer, lymph node flap transfer (LNFT), microsurgical lymphaticovenous implantation, and lymphovenous shunt.

In 1990, Baumeister and Siuda found that autologous lymph vessel transplantation was a fundamental step toward microsurgical treatment of lymphedema.⁷ Since then, many lymphatic reconstruction procedures have been considered for managing the condition. However, there are still limitations to treatment options. Garza et al concluded that there is a lack of consistency in the literature for the design of personalized surgical management strategies in patients with lymphedema.⁸ Raju and Chang concluded that further exploration into standardized protocols is needed for the diagnosis and treatment of lymphedema to improve patient outcomes.⁹

There are no official guidelines from any health care organization regarding a recommended protocol for diagnosing, monitoring, and treating lymphedema. The 3 specific objectives of this review are to (1) analyze the current treatment options and the data being reported, (2) report short- and long-term treatment outcome data as well as patient satisfaction outcomes, and (3) propose evidence-based guidelines to improve and enhance the consistency with which the treatment for lymphedema is delivered.

Search Strategy

This paper’s systematic search was conducted according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.¹⁰ The PubMed and Embase databases were searched for all publications through June 1, 2020 that included the following keywords and/or MeSH terms: secondary AND lymphedema AND neoplasm metastasis AND surgical procedures AND treatment OR plastic surgery.

Study Selection

All English-language studies published before June 1, 2020, were included. We excluded nonsurgical interventions, literature reviews, letters, commentaries, nonhuman or cadaver studies, and studies with inadequate sample size (N < 20). In addition, we excluded all studies that had patients with both upper and lower extremity lymphedema concurrently, as well as those studies of patients who underwent both LVA and VLNT performed together. In this way, we prevented any inaccurate data analysis that was not central to our study.

Data Extraction

Two reviewers (FS, CJ) independently screened and extracted data in 2 steps: (1) review of titles and abstracts and (2) analysis of full-text articles. Whenever there was a conflict between the 2 reviewers, the senior author (CH) resolved the conflict by making the final decision. Data from each relevant study was extracted into a standardized form that included: (1) article author, (2) year of publication, (3) study design, (4) sample size, (5) demographics, (6) surgical treatment and techniques, (7) study outcomes, and (8) length of follow-up time. Studies were grouped into upper extremity treatment (LVA and VLNT) or lower extremity treatment (LVA and LNFT).

Statistical Analysis

It is important to note that statistical analysis of data from upper extremity lymphedema studies was conducted independently of the analysis of data from lower extremity studies. Both groups yielded vastly different surgical outcomes and hence were not compared directly to each other.

Statistical analyses were processed with OpenMeta [Analyst] (Brown University) and STATA 16.0 (StataCorp). Results were considered significant at a P value < .05. Corresponding events/means and standard deviations (SDs) were extracted. Hozo’s method was used to convert data with ranges to SD.¹¹ Meta-analysis was first performed using a continuous random-effects model and metafor package for R determination.¹² The pooled mean effect size and arcsine of square root proportion difference were estimated for quantitative and binary data, respectively, and plotted as forest plots. Heterogeneity was assessed using Cochran’s Q statistic and quantified using I2 statistics. Articles were considered to have significant heterogeneity between studies when the P value was less than .1 or the I2 greater than 50%.¹³ Subgroup analyses by treatment method, comorbidities, quality of life (QOL), postoperative complications, and extremity were carried out. Random-effects meta-regression was performed to identify the influence of potential effect modifiers on the pooled results. Publication bias was evaluated by Egger’s regression test.¹⁴ Comparison between 2 rates was estimated using OpenEpi 3.01 software (Emory University) for epidemiologic statistics.

The search process yielded a total of 1255 articles. Of these, 15 studies (583 total patients) were included for meta-analysis. Eleven studies included patients undergoing upper extremity lymphedema treatment (LVA or VLNT), and 4 studies included cohorts who underwent lower extremity treatment (LVA or LNFT) (Figure 1). As stated earlier, we excluded all studies that had patients with both upper and lower extremity lymphedema concurrently as well as those with patients who underwent both LVA and VLNT performed together. Of the 15 included studies, 3 were retrospective cohort studies, 11 had prospective cohort design, and 1 was a clinical trial. Included studies were published between 2009 and 2019: 4 from the US, 5 from Italy, 2 from France, and 2 from Japan. Patient demographics and characteristics of included studies are described in Table 1 and Table 2.

Volume measurements (volumetric, lymphoscintigraphy, and limb circumference) were taken at average follow-up. The pooled overall reduction rate of both upper and lower extremity treatments was 41.5% (95% CI, 31.7%-51.3%) (Figure 2). The volume reduction rate was lower in the upper extremity lymphedema subgroup (38.0%; 95% CI, 25.9%-50.2%) than the lower extremity subgroup (49.5%; 95% CI, 32.6%-66.3%) (Figure 2). When compared with volumetric measurements taken on surgery day, only 2.0% (95% CI, 1.0%-3.5%) of patients experienced worsening symptoms. About 2.6% (95% CI, 1.0%–4.9%) of the upper extremity and 1.2% (95% CI, 0.1%-3.1%) of the lower extremity treatment cohorts had worsening lymphedema (Figure 3).

Table 3 shows the average percentage of volume reduction for LVA and VLNT in the upper extremity treatment subgroup on surgery day and at average follow-up after surgery. The volume was reduced by 34.6% (95% CI, 20.0%-49.2%) in the LVA subgroup and by 44.8% (95% CI, 32.1%-57.6%) in the VLNT subgroup. Table 3 shows the average percentage of volume reduction for LNFT and LVA in the lower extremity treatment subgroup on surgery day and at average follow-up after lymphedema surgery. The volume was reduced by 50.4% (95% CI, 30.5%-70.3%) in the LVA subgroup and by 46.3% (95% CI, 35.3%-57.3%) in the VLNT subgroup.

Table 4 shows the specific postoperative complications for each subgroup. For upper extremity lymphedema treatment, cellulitis was the most common postoperative complication (4.5%; 95% CI, 0.9%-10.6%). In lower extremity treatment, seromas were the most common (4.6%; 95% CI, 0%-17.8%).

The correlation between average follow-up time and average volume reduction was not significant (r = 0.16; P = .62). Longer average follow-up time increased the frequency of postoperative complications for both upper and lower extremity treatments; however, this was not statistically significant (r = 0.32; P = .26). The correlation between total operation time with frequency of complication was not significant (r = -0.38; P = .40). There was a statistically significant direct correlation between longer total operation time and increased average volume reduction (r = 0.90; P = .01) and a statistically significant direct correlation between body mass index and average volume reduction (r = 0.83; P = .04). However, there was no statistically significant correlation between lymphedema duration before treatment and average volume reduction (r = -0.131; P = .8) or frequency of complication (r = -0.621; P = 1.0).

Analysis of Lymphedema Treatment Subgroups for Upper Extremity

The two subgroups analyzed in upper extremity lymphedema treatment were LVA and VLNT. Figure 4 shows that patients who underwent VLNT experienced more postoperative complications (12.9%; 95% CI, 0.2%-40.1%) than patients who underwent LVA (8.8%; 95% CI, 1.7%-20.7%) at average follow-up after surgery. However, compared with initial volume measurements on the day of surgery, more patients who underwent LVA had worsening lymphedema (2.6%; 95% CI, 0.3%-4.9%) compared with the VLNT cohort (0.9%; 95% CI, -0.4% to 2.2%) at average follow-up after surgery (Figure 5).

Furthermore, 9 selected studies (N = 326) conducted patient QOL measurements (Table 5). Average QOL measures across all studies improved by 52.2% (95% CI, 25.1%-79.2%) for patients who underwent upper extremity lymphedema treatment. Patients who underwent VLNT reported a greater QOL improvement at 66.0% (95% CI, 9.2%-123.0%) than LVA (45.2%; 95% CI, 18.4%-72.0%). Table 6 shows the postoperative complications by upper extremity and lower extremity treatment subgroup. For both LVA and VLNT, cellulitis was the most common postoperative complication (4.5%; 95% CI, 0.9%-10.6%).

Publication Bias

For average volume reduction, Egger’s linear regression test showed evidence of potential publication bias (P = .001) (Figure 6).

Our study aims to consolidate diverse staging systems and ascertain which method of lymphedema surgery is most effective and associated with the fewest complications.

Staging

There are several different staging tools described in the literature: International Society of Lymphology (ISL), Cheng, Campisi, indocyanine green (ICGN) staging scale, and the MD Anderson Classification.^15-19 The variety of classification systems causes confusion across the literature and impedes the validity of pooling together data in a standardized method to accurately determine treatment efficacy. Staging of lymphedema has been shown to be important in determining treatment groups.^18-20

Use of an accurate staging system should be the first step in lymphedema management. Given the information provided, we recommend implementing the ICGN staging scale described by Schaverian and Coroneos,¹⁹ which is a modified version of the MD Anderson classification system described by Chang et al.¹⁸ The ICGN staging scale seems to give the best objective information regarding lymphatic vessel function and thus is most accurate for selecting the most efficacious treatment option.

In accordance with the ICGN staging scale, a 3- to 6-month trial of complete decongestive therapy (CDT) is usually the initial implemented treatment. If lymphedema persists, surgical intervention may be required.¹⁹ As defined by ICGN, LVA is a minimally invasive surgical option of choice for patients with grade I and early grade II lymphedema. These patients are believed to still have patent lymphatic vessels, an important requirement for the success of the LVA procedure.²¹ In more severe cases of lymphedema, fibrosclerotic changes and loss of vessel patency prevent success of LVA.²² Consequently, for patients with late grade II, III, and IV disease, vascularized lymph node transfer is the recommended surgical option.²¹ The continued analysis of the various staging scales is important, as an agreed-upon staging system and postoperative management care will help to streamline treatment of secondary lymphedema.

Lymphedema Surgical Treatments: A Comparison Between LVA and VLNT

The 2 surgical intervention subgroups analyzed in upper extremity lymphedema treatment were LVA and VLNT. Of the 11 studies included in the upper extremity analysis, 4 utilized VLNT as the main surgical intevention.^21,23-25 Evidence regarding the location of lymph node transfer was reviewed from these studies, which involved 143 patients. The donor site lymph nodes were harvested from the superficial inguinal lymph nodes based on branches from the superficial circumflex iliac artery and vein (SCIA/V) or superficial inferior epigastric artery and vein (SIEA/V) in 108 patients. Five patients had donor lymph nodes harvested from the lateral thoracic artery and vein if they lacked either the SCIA/V or SIEA/V lymph nodes.²⁴ Three patients who underwent a bilateral mastectomy and lacked both SCIA/V and SIEA/V utilized a third, less favorable choice for transfer due to the small size of the flap, based on the transverse cervical artery and vein.²⁴ The final location of transfer is described in the study by Engel et al, in which 27 patients underwent a submental lymph node transfer, with the flap being dissected around the submandibular gland.²¹

Currently, there are 2 prevailing surgical approaches employed in the management of lymphedema after failed decongestive therapy: bypassing the lymphatic drainage via LVA or transferring lymphoid tissue to the area via VLNT or LNFT. The former acts to divert flow into the venous system and therefore lessens the load of extremity lymphatic drainage, whereas the latter has a few proposed mechanisms, including promoting lymph angiogenesis via growth factors (VEGF-C) and new lymphaticovenous drainage via perfusion gradients (pumping mechanism).^26-28 Scaglioni et al have described the objective and subjective benefits of both LVA and VLNT procedures independently; however, no studies have analyzed the  procedures comparatively.²⁹

Overall, both LVA and VLNT procedures appeared to yield large volume reductions in both upper extremity and lower extremity lymphedema at average postoperative follow-up (Figure 2). Although the overall pooled reduction in both the upper extremity and lower extremity groups showed a statistically significant volume reduction, that significance seemed to disappear when stratifying by LVA and VLNT. This may be explained by the aforementioned staging classification problem, which may lead to some patients not being paired with the most appropriate procedure. One reason for the success of the LVA procedure may be patent lymphatic channels, which can be visualized and classified using indocyanine green, thus possibly mitigating the reduction rate if performed on candidates with less than desirable lymphatic patency.³⁰

Interestingly, 7 of the 15 included studies (46.6%) incorporated patients with lymphedema in late stages 3 and 4, which are more severe and have been shown to be more applicable for management with VLNT. Furthermore, 6 of these 7 studies (85.7%) in late-stage populations used the LVA procedure. Thus, the results in decreased circumferential reduction may possibly have been greater if the patients with late stage 3 or 4 disease underwent VLNT instead, thus creating a more accurate significant difference in reduction through both LVA and VLNT procedures.

In addition, the number of anastomoses completed during the surgical procedure in patients undergoing LVA has been posited in the literature as being a prognostic factor in volume reduction.¹⁷ This positive correlation may support the idea that additional time creates more anastomoses in the LVA group, thereby leading to increased lymphatic outflow. This surgical technique may also be explained more simply by a methodical, careful dissection to ensure patency and decrease collateral damage.

A Discussion on Debulking and Liposuction

Excisional surgery for lymphedema currently consists of debulking and liposuction.³¹ Of the 15 studies included in our analysis, none utilized a traditional debulking technique to reduce limb volume before anastomosis or lymph node transfer. The decision to exclude debulking from the primary method of treating secondary lymphedema derives from it causing extensive scarring and significant morbidities, which outweighs the simplicity of this surgical approach.²⁰  In addition, liposuction can effectively reduce the volume of the hypertrophic adipose tissue; however, it increases the risk of damage to any remaining lymphatic vessels, potentially worsening the preexisting lymphedema.²⁰ Thus, debulking and liposuction procedures are considered only in the most severe cases of lymphedema.

The Impact of Conservative Management

The impact of conservative management on lymphedema must be addressed in the context of therapeutic approaches. CDT is one such strategy and consists of 4 conservative treatments including manual lymphatic drainage (MLD), compression therapy (ie, compression bandages, compression sleeves), skin care, and lymph-reducing exercises. MLD is a form of massage therapy based on the lymphatic system’s anatomy. The massage performed stimulates movement of lymph while reducing the hydrostatic resistance to lymph flow, effectively minimizing lymph stasis and the formation of lymphedema. Similarly, use of compression garments helps to reduce interstitial fluid buildup by preventing lymph backflow into the interstitial space.³² Although the first-line treatment option for lymphedema is CDT, which provides symptomatic relief, it relies heavily on patient compliance and may be less effective for more severe cases.³³

In essence, continued postoperative use of compression garments must be analyzed. Studies have demonstrated a high discontinuation rate in the use of compressive garments and other forms of CDT after surgery. The cessation of CDT could be explained by the immediate improvement of the lymphedema after surgery and hence the patient’s perception that there was no further benefit in continuing CDT. Patients reported that they felt the rigorous CDT was no longer needed to maintain the level of reduced volume achieved by surgical intervention.^33,34 Studies have found that surgery combined with decongestive therapy yields better outcomes.³³ Reinforcement of the benefits of CDT with patients is an important part of patient counseling and improved treatment outcomes in patients who have undergone LVA and VLNT.

There are several limitations to our study that may alter the findings in either direction. As previously mentioned, these novel techniques for lymphedema treatment carry a certain level of heterogeneity among the selection, treatment, and data collection. We attempted to select studies with large sample sizes (N > 20) and specific populations and outcomes to make a more reliable assumption. Due to our strict inclusion criteria, we cannot eliminate the possibility of selection bias or any influence the excluded studies may have had on our results. Additionally, the degree of fibrosis secondary to lymphedema with prolonged inflammation varies by race and ethnicity. Unfortunately, the data we have used in this meta-analysis does not account for all patient demographics, specifically race. Further studies and analysis are needed to examine the degree of fibrosis across racial groups.

It has been postulated that there lies a steep learning curve in performing LVA and VLNT because they are novel procedures, and the studies in our analysis may be biased toward more experienced surgeons. However, Pereira et al found no evidence of surgeon experience correlating to outcome, at least for VLA. In their analysis, only the timeliness in which the procedure was completed was affected (P = .017).³⁰ More studies will be needed to confirm this interesting finding.

Surgical management of lymphedema continues to grow in popularity. Although the techniques are relatively new and literature surrounding these surgical treatments is scarce, our data suggest these approaches show great promise for overall volume reduction and improved QOL for some patients with lymphedema. Our findings indicate that VLA is neither superior nor inferior to VLNT, but the stage of disease may dictate which procedure is more appropriate.

Affiliations: ¹Division of Plastic and Reconstructive Surgery, Tulane University School of Medicine, New Orleans, LA; ²Department of Surgery, Florida State University, Tallahassee, FL;

³Department of Surgery, Grand Strand Medical Center, Myrtle Beach, SC; ⁴Division of Plastic and Reconstructive Surgery, Tufts Medical Center, Boston, MA

Correspondence: Christopher Homsy, MD; christopherhomsy@gmail.com

Disclosures: None of the authors has declared a financial interest in any of the products, devices, or drugs mentioned in this manuscript.