Surgical interventions for the prevention or treatment of lymphoedema after breast cancer treatment
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
1. To assess and compare the efficacy of surgical interventions in treating established lymphoedema (LE) of the arm after breast cancer treatment.
2. To assess and compare the efficacy of surgical interventions at preventing the development of LE of the arm after breast cancer treatment.
Background
Description of the condition
Breast cancer is the most common type of cancer amongst women worldwide (Globocan 2012). Advances in treatment for breast cancer mean that more women are surviving and living with the consequences of the cancer and its treatment. One common and distressing sequelae of breast cancer treatment is lymphoedema (LE) (Hayes 2008). LE describes a condition where an excess of fluid builds up in the subcutaneous tissue of the arms, usually because of impairment to the outflow of lymphatic fluid from an affected area. The aetiology of this imbalance between production and drainage can be either congenital or acquired, so‐called primary or secondary LE, respectively (Rockson 2008). Secondary LE is more common and has a variety of causes, including infection, cancer, radiation therapy and surgery. Worldwide the most common cause of LE is filariasis, caused by infection with the parasite Wuchereria bancrofti, whereas in the developed world the most common cause is malignancy and its treatment (Doscher 2012). As breast cancer is very common, its treatment is a major cause of secondary LE.
LE may occur after breast cancer treatment due to damage or destruction of the upper limb lymphatic drainage system, by either surgery or radiotherapy, or both, or owing to the destructive effects of the cancer itself (Brennan 1996). Estimates of the incidence of LE after breast cancer treatment vary widely depending on the type of treatment undergone and the diagnostic criteria for LE. At five years post‐operation, 5% of patients who have had sentinel node biopsy alone develop measurable LE compared to 16% of patients who have had an axillary dissection (McLaughlin 2008). Obstruction to lymphatic drainage and accumulation of lymph fluid in the tissues leads to swelling, cutaneous and subcutaneous thickening, fibrosis and eventually to elephantiasis (Doscher 2012). The International Society of Lymphology staging system is the most universally accepted system for assessing the degree of LE (ISL 2009). In stage 0 disease LE is subclinical, with no swelling despite evidence of impaired lymphatic transport; in stage 1 disease swelling is evident but goes down with limb elevation; in stage 2 the swelling is persistent and fibrosis of the subcutaneous tissues becomes evident; and in stage 3 there is thickening of the skin with fatty deposits and lymphostatic warts, so called elephantiasis (ISL 2009).
Patients with LE may experience significant morbidity related to their condition including loss of limb function, discomfort and disfigurement (Brennan 1996). LE may also be complicated by cellulitis and lymphangitis (Shih 2009). Although not a life‐threatening condition, LE can cause great distress for breast cancer survivors, both through the physical effects, described above, and its psychological consequences (Bulley 2013). It can act as a constant reminder of previous disease and attract unwanted attention from others (Petrek 1998). Until recently the effects that LE can have on patient quality of life have been underestimated. This is being addressed by the development of condition‐specific quality of life assessment tools (Keeley 2010).
Measurement of limb circumference is the most commonly used method to evaluate LE as it is simple and widely available (Yamamoto 2013). Other tests are also used including lymphoscintigraphy, indocyanine green (ICG) lymphography, and volumetry based on computed tomography (CT), magnetic resonance imaging (MRI), measurement of water displacement and laser scanning (Bulley 2013;Yamamoto 2013). It is widely considered to be important to diagnose LE as early as possible in the natural history described above. This allows early intervention, including the provision of information and psychological support, and may prevent symptomatic LE formation and progression to severe disease (Yamamoto 2011). Objective evaluation of existing LE is important both to evaluate its extent and to assess interventions to reduce it.
Description of the intervention
Various conservative strategies are used in women with breast cancer to decrease the risk of developing LE (Stuiver 2012) and to manage it once developed (Preston 2008). These include exercise, patient education, monitoring, compression therapy, manual lymph drainage and lymph taping (kinesiotaping). There is no consensus as to the optimal conservative management.
Surgery is traditionally the option of last resort, when conservative measures have failed (Vignes 2002). Indications for surgery include insufficient LE reduction by well performed medical and physical therapy, recurrent episodes of lymphangitis, intractable pain, decreasing limb function, patient dissatisfaction with conservative methods and the patient's wish to proceed to surgery (Campisi 2010). However, patients and surgeons are increasingly undertaking modern surgical interventions early in the disease process in the hope of preventing or reversing the lymphoedematous changes (Boccardo 2011).
These interventions include liposuction, lymph node transfer and lymphaticovenular anastomosis (LVA) (Becker 2006; Brorson 1997; Campisi 2004; Nagase 2005). The techniques aim to either directly reduce the volume of the arm by removing fat and fibrous tissue (liposuction), bypass the obstruction to lymphatic outflow by creating a direct route for lymphatic fluid to re‐enter the circulation in the arm (LVA), or improve lymphatic drainage from the arm possibly by the creation of new lymphatic channels (lymph node transfer).
How the intervention might work
Chronic LE causes hypertrophy of subcutaneous adipose tissue, which leads to deformity and limb swelling. This is the rationale behind liposuction for LE where the hypertrophied adipose tissue is removed in a circumferential fashion through cannulae inserted into small incisions, usually under general anaesthesia (Brorson 2002;Damstra 2009). Lifelong compression garments are then needed to prevent post‐operative recurrence of the lymphoedema (Brorson 2002; Damstra 2009).
Microsurgical lymphatic reconstruction is another surgical intervention for LE, most often in the form of LVA, though occasionally an unaffected lymphatic channel is used to bypass the area of lymphatic damage, a method known as a lymphatico‐lymphatic bypass (Lee 2011). LVA consists of anastomosing lymphatic vessels to a subcutaneous venule. This returns the lymph fluid to the circulation peripherally, meaning it no longer has to pass through the area of lymphatic damage to return to the circulation. Dye is used to find healthy, functioning lymphatics and to confirm the patency of the anastomosis. Lymphoscintigraphy or ICG lymphography can be used during follow‐up to confirm ongoing anastomosis patency (Campisi 2010).
Lymph node transfers take healthy lymph nodes from an unaffected limb and move them to either the axilla or antecubital fossa of the affected arm (Becker 2006). The aim of the technique is for the transplanted lymph nodes to take over lymphatic drainage from local lymph nodes which have been damaged or removed by surgery (Becker 2006). Notably, the precise mechanism of action of lymph node transplant is complex and incompletely understood.
Why it is important to do this review
LE as a consequence of breast cancer treatment causes severe physical and psychosocial morbidity amongst this patient group. There is no consensus as to which of the available surgical options described above provides the best outcome in this patient group, both in terms of prevention and treatment of established disease. We aim to clarify the indications, benefits and uncertainties in the field.
Objectives
1. To assess and compare the efficacy of surgical interventions in treating established lymphoedema (LE) of the arm after breast cancer treatment.
2. To assess and compare the efficacy of surgical interventions at preventing the development of LE of the arm after breast cancer treatment.
Methods
Criteria for considering studies for this review
Types of studies
We will consider all types of randomised controlled trials (RCTs) that compare a surgical intervention for the treatment or prevention of LE of the arm after breast cancer treatment to either standard intervention (conservative measures such as compression garments, lymphatic massage, bandaging and intermittent pneumatic compression), placebo intervention (surgery performed without the critical surgical step), or another surgical intervention that is included in this review (see types of intervention).
All studies that present predefined, objective criteria for diagnosing or assessing LE, or both, will be considered eligible. These criteria (Bulley 2013;Yamamoto 2013) may extend to, but are not limited to:
-
measurements of tissue volume and distribution, such as water displacement, tape measure, and perometry;
-
determination of limb composition, such as tonometry, skinfold thickness, dual energy x‐ray absorptiometry, and bioelectrical impedance analysis;
-
imaging tests, such as lymphoscintigraphy, lymphangiography, MRI, ultrasound, and CT scanning.
Types of participants
We will include patients of both sexes and all ages who have had treatment for their breast cancer. This includes any type of surgery, radiotherapy, chemotherapy or combination of these.
Types of interventions
We will consider any surgical intervention for the treatment or prevention of secondary LE of the arm after breast cancer treatment. Both reductive and reconstructive techniques will be considered, including but not limited to:
-
liposuction;
-
LVA;
-
lymphatico‐lymphatic bypass;
-
lymph node transfer.
Types of outcome measures
Primary outcomes
All outcome measures will be assessed at the latest timepoint reported in the trial.
Prevention
-
The development of LE at the latest timepoint specified in the study
Treatment
-
The percentage reduction of LE, as measured by a validated system for limb volume measurement
Secondary outcomes
Secondary outcome measures will be as follows.
-
Patient reported outcomes using validated questionnaires, such as the condition‐specific quality of life measure for limb lymphoedema (LYMQOL) which accounts for not only physical symptoms but also social function and psychological effects (Keeley 2010). Other similar validated questionnaires include the Upper Limb Lymphedema 27 (ULL‐27) (Launois 2002) and Freiburg Life Quality Assessment‐Lymphedema (FLQA‐L) (Augustin 2005).
-
Ability to discontinue further interventions for LE; for example, this may be non‐adherence to or the number of dropouts from continued use of compression therapy, manual therapy or lymph taping.
-
Surgical complications, such as seroma, lymphorrhoea, wound dehiscence, wound infections.
-
Long‐term complications, such as rate of cellulitis, lymphangitis, lymphadenitis, skin ulcers, and lymphorrhoea.
Search methods for identification of studies
We will not impose any language or publication date restrictions.
Electronic searches
For published trials, we will search the following.
a) The Cochrane Breast Cancer Group (CBCG) Specialised Register. Details of the search strategies used by the CBCG for the identification of studies and the procedure used to code references are outlined in the Group's module (www.mrw.interscience.wiley.com/cochrane/clabout/articles/BREASTCA/frame.html). Trials with the key words 'lymphoedema', 'lymphedema', 'upper‐limb lymphoedema', 'breast cancer', 'axillary clearance', 'axillary lymphadenectomy', 'axillary dissection', 'lymphaticovenular anastomosis', 'lymphaticovenous anastomosis', 'LYMPHA', 'lymph node transfer', 'lymph node flap', 'liposuction', 'lymphatic bypass', 'prevention', and 'treatment' were extracted and considered for inclusion in the review.
b) Cochrane Central Register of Controlled Trials (CENTRAL). See Appendix 1.
c) MEDLINE via OvidSP. See Appendix 2.
d) EMBASE via EMBASE.com. See Appendix 3 for the search strategy.
e) The Cumulative Index to Nursing and Allied Health Literature (CINAHL) through EBSCO (1980 to May 2013). See Appendix 4 for the search strategy.
For prospectively registered and ongoing trials, we will search:
a) the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) search portal (http://apps.who.int/trialsearch/Default.aspx). See Appendix 5 for the search strategy;
b) ClinicalTrials.gov (http://clinicaltrials.gov/).
Searching other resources
Three review authors will independently handsearch the reference lists of included studies for additional eligible trials.
Data collection and analysis
Selection of studies
Three review authors (PM, NL, VA) will independently determine the eligibility of each study. Three authors will analyse the titles and abstracts of all citations found through the search strategy previously described. A copy of the full article for each citation reporting a potentially eligible trial will be obtained, and three review authors (PM, NL, VA) will independently apply the eligibility criteria; any discrepancies will be resolved by consensus through discussion with a fourth review author (DF). Where necessary and possible, additional information will be sought from the principal investigator of the trial concerned. Any exclusions from the review of a potentially eligible trial will be justified in the final report.
Data extraction and management
Three review authors (PM, NL, VA) will independently extract the data for each included study using a pre‐designed data extraction pro forma. Data will be extracted according to the details of the trial (first author, year of publication, journal, publication status, period and country of study, sources of funding, study design, sample size); patient characteristics (age, sex, type of disease, stage of disease, type of surgery, and prior treatment status); quality of the study; details of the intervention; duration of follow‐up; and the primary and secondary outcomes. We will attempt to contact study authors to acquire any data of relevance that are not fully reported in the published manuscript. A fourth review author (DF) will be used to resolve any discrepancies regarding data extraction, and consensus will be reached. The most complete data set feasible will be assembled.
Assessment of risk of bias in included studies
Three review authors (PM, NL, VA) will independently assess each study for risk of bias using the Cochrane Collaboration's risk of bias tool (Higgins 2011). We will consider selection bias, performance bias, detection bias, attrition bias, and reporting bias for each individual study. Where there are unclear risks of bias due to inadequate descriptions within the manuscript, we will attempt to contact the authors in order to clarify the risk. All efforts made to obtain additional information will be reported in the completed review. Any discrepancies will be resolved by consensus through discussion with a fourth author (DF). We will describe the support for our judgement on the seven risk of bias domains within each study in the final review.
Measures of treatment effect
Hazard ratios (HR) with 95% confidence intervals (CI) will evaluate the comparative times to development of LE after preventative LE surgery and also time to the development of complications related to surgery or to LE itself. Risk ratios (RR) with 95% CI will analyse the risk of developing LE, surgical complications and LE complications. Continuous variables, such as those measuring the percentage reduction of LE after treatment or the psychosocial well‐being of patients with LE, will be analysed as mean difference (MD).
Unit of analysis issues
It is likely that most trials will be addressing patients with unilateral LE after breast cancer. However, in the rare instance that patients have both arms affected by LE or at risk of LE, and the arms have been randomised separately, we will analyse those outcomes that are not dependent on the other arm (such as limb volume measurements, development of LE, surgical complications, long‐term complications, and ability to discontinue other treatments) on a per arm basis. Patient reported outcomes will depend on both arms, and such patients will therefore be removed from the analysis of these outcomes.
If surgical interventions are done after a period of observation or non‐surgical treatment, in the analysis we will consider only the post‐operative intervention period.
Dealing with missing data
If the results of an RCT have been published but information on the outcome of interest has not been reported, an attempt will be made, whenever possible, to contact the trial authors for the missing information. All efforts made to obtain additional information will be reported in the completed review. Where possible, all analyses will be by intention to treat (Hollis 1999). If patients were allocated to one intervention (for example LVA) but after randomisation underwent a different intervention (for example lymphatico‐lymphatic bypass), they will be analysed according to their randomisation allocation. If the results for dichotomous variables are not reported for some participants, we will analyse the results based on both a worst possible outcome (for example surgical complications occurred in all non‐reported cases) and a best possible outcome (for example surgical complications did not occur in any non‐reported cases).
Assessment of heterogeneity
We will explore heterogeneity using the Chi2 test with significance set at a P value of 0.10, and measure the quantity of heterogeneity using the I2 statistic (Higgins 2002).
Thresholds for the interpretation of the I2 statistic can be misleading. A rough guide to interpretation is:.
-
0% to 40%, might not be important;
-
30% to 60%, may represent moderate heterogeneity;
-
50% to 90%, may represent substantial heterogeneity;
-
75% to 100%, considerable heterogeneity.
We will take factors such as clinical and methodological heterogeneity, along with whether the heterogeneity is in the magnitude of effect or in the direction of effect, into account while interpreting the I2 statistic, particularly where there is an overlapping of ranges.
Assessment of reporting biases
We will compare the reported outcomes with those stated in the methods of the studies and also those listed in clinical trials registries as both primary and secondary outcomes (for example http://www.clinicaltrials.gov/). If sufficient studies, at least 10, are identified and the studies are not of similar sizes, that is not similar in standard errors of intervention effect estimates, we will assess publication bias by creating a funnel plot using software within Review Manager 5.2 (RevMan). Where there is evidence of small‐study effects, publication bias will be considered in light of the fact that it may only be one of a number of possible explanations for funnel plot asymmetry. Additionally, because tests for funnel plot asymmetry typically have relatively low power, it is important to note that funnel plot asymmetry does not exclude bias, including publication bias.
Data synthesis
We will present a narrative overview of the included trials. Where appropriate, we will present meta‐analyses of outcome data using Review Manager software (RevMan). The decision to pool data in a meta‐analysis will depend on the availability of outcome data and the assessment of between‐trial heterogeneity. For comparisons where there is no apparent clinical heterogeneity and the I² value is ≤ 40%, we will apply a fixed‐effect model (Mantel‐Haenszel). Where there is no apparent clinical heterogeneity and the I² value is > 40%, we will apply a random‐effects model (DerSimonian). However, we will not pool data where heterogeneity is very high (I² values ≥ 75%).
For time to event data, such as time to development of LE in a prevention trial, we plan to plot (and if appropriate pool) estimates of the HR and 95% CI as presented in the trial reports using the generic inverse variance methods in the Review Manager software (RevMan). For dichotomous outcomes, such as surgical wound infection, we will present the summary estimate as a RR with 95% CI. Where continuous outcomes are measured in the same way across trials, for example measurements of limb volume, we will present a MD with 95% CI. We will present a standardised mean difference (SMD) where trials measure the same outcome using different methods. We recognise that it may not be possible or appropriate to perform meta‐analysis on the available data.
To grade the quality of evidence, GRADE profiler software (GRADE) will be used and summary of findings tables produced for the five pre‐specified outcomes for the prevention and treatment of LE.
Subgroup analysis and investigation of heterogeneity
If there are sufficient trials of adequate size, it may be possible to conduct subgroup analyses. Ability to conduct the analyses will also depend on whether or not the required information is recorded in the trial publications. If data are not included, we will attempt to contact trial authors to provide the data. The following will be considered for possible subgroup analysis:
-
patients with sentinel lymph node biopsy and complete lymph node clearance;
-
patients treated with and without radiotherapy to the axilla;
-
patients treated with and without radiotherapy to any area (axilla, breast, chest wall);
-
patients treated with and without chemotherapy;
-
patients with a body mass index above 30 or below 30.
Sensitivity analysis
A sensitivity analysis, including excluding trials with a high risk of bias, will be performed if sufficient high quality trials are identified to make this feasible. We acknowledge that there is no accepted definition of high risk of bias, so we have chosen the threshold as those trials with greater than or equal to three of seven domains (random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, other bias) with high risk of bias when analysed with the Cochrane Collaboration's risk of bias tool to represent high risk of bias studies (Higgins 2011).
