Complete decongestive therapy for lymphedema following breast cancer treatment
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
The primary objective of this review is to assess the effects of complete decongestive physiotherapy or manual lymphatic drainage alone on the primary outcome of arm volume reduction or the secondary outcomes of chest wall volume reduction, functional improvement, and pain reduction for women with secondary lymphedema as a result of treatment of breast cancer.
Background
Lymphedema (LE) in breast cancer patients is due to damage incurred to, or removal of, the axillary lymphatic system by surgery or radiation therapy, resulting in fluid accumulation in the subcutaneous tissues of the chest wall or arm (Brennan 1996). LE, secondary to breast cancer treatment, can result in chronic and frequently debilitating swelling of the affected arm and/or chest wall. LE can cause discomfort, pain, heaviness, limited motion, unsatisfactory appearance, altered lifestyle, and impacts upon quality of life (Brennan 1992; Tobin 1993). Lack of recognition of LE symptoms and/or inadequate treatment may lead to specific complications such as cellulitis, infection, and lymphangitis (Petrek 1998).
LE severity is frequently assessed by grade. Grade I is characterized by soft pitting edema that reduces with elevation and is without clinical fibrosis. Grade II is characterized by non‐pitting edema, which does not reduce with elevation and demonstrates the presence of fibrosis. Grade III is characterized by hardening and hypertrophy of the subcutaneous tissues as well as thickening and changes in the skin. For untreated arm LE, the duration of symptoms is associated with disease severity (Casley‐Smith 1995).
The reported prevalence of LE varies from study to study depending on the predisposing factors. Women receiving axillary radiation and axillary lymph node resection are at the highest risk for LE. For example, among those receiving both axillary surgery plus axillary radiation, LE prevalence estimates are 12%‐60%; with most reports suggesting more than one‐third of the women experienced LE (Colette 1997). Axillary surgery without axillary radiation shows a lower prevalence (0‐42%) with most studies estimating 20‐30% after one year (Colette 1997). It is important to note that post‐operative irradiation to the breast had no impact on the occurrence of LE among women who had already had axillary dissection (Liljegren 1997). It is estimated that 400,000 of the 2 million breast cancer survivors alive today cope daily with the disfigurement, discomfort, and disability of arm and hand LE (Petrek 1998).
Conservative treatment is the most widely accepted method for managing LE symptoms. Conservative treatment can include complete decongestive physiotherapy (CDP), manual lymphatic drainage (MLD) alone, or sequential pneumatic compression. CDP is a fourfold approach which includes MLD, compression, skin care, and remedial exercises (Foldi 1998). Compression (bandaging the affected limb) is frequently done after MLD because the elastic fibers of the skin have often been damaged and require compression to prevent the reaccumulation of fluid. Skin care regimens teach LE patients how to support the acidity of the skin. Skin care regimens are an important part of CDP because LE is protein‐rich swelling and therefore LE patients are at risk for fungal infections. Remedial exercises are designed for the patient to do at home to support the joint/muscle pump.
MLD is a type of specialized massage, which is administered as part of CDP or alone. MLD is designed to mobilize and reroute lymphatic fluid from areas of stasis, due to damage to the lymphatic vessels, into healthy lymphatic pathways (Leduc 1998). MLD is performed by practitioners educated in the anatomy and function of the lymphatic system, and trained in massage techniques for moving lymphatic fluid. MLD is performed slowly to prevent lymphangiospasm, and lightly to target the superficial tissues where most LE occurs.
The CDP or MLD alone is usually performed in two phases. The goals of the first phase are to mobilize the accumulation of the protein‐rich fluid and initiate the reduction of fibrosclerotic tissue. In this initial phase, commonly called the therapeutic phase, which is typically the phase assessed in clinical trials, MLD is generally done once or twice a day for usually five days a week for approximately four weeks. The number of weeks of MLD treatment varies from one to six depending on the severity of disease. If MLD is being performed as part of CDP, then patients are instructed in the other aspects of the therapy (exercise, compressive therapy, skin care) during this initial phase. The second phase, commonly called the maintenance phase, is a phase in which MLD is performed only as needed. If MLD is performed as part of CDP, then during the maintenance phase, patients are instructed to continue the other aspects of CDP (compressive therapy, skin care and remedial exercises) on their own. Therefore, although the MLD component of CDP is highly utilized during the therapeutic or initial phase of therapy, making CDP more expensive in the beginning of therapy; CDP is ultimately designed for patients to become more self‐sufficient over time, with MLD ultimately used on an as‐needed basis.
The effectiveness of CDP has been reported in a number of case studies with varying reports on arm volume reduction from study to study (Bunce 1994; Casley‐Smith 1994). Megans (Megans 1998) analyzed the research literature to determine the effectiveness of physical therapy management of LE. This included an analysis of the strength of the evidence for MLD. Overall, reports appear to be positive; however, a systematic assessment of the clinical trials of CDP or MLD alone for LE secondary to breast cancer treatment has not been performed.
Objectives
The primary objective of this review is to assess the effects of complete decongestive physiotherapy or manual lymphatic drainage alone on the primary outcome of arm volume reduction or the secondary outcomes of chest wall volume reduction, functional improvement, and pain reduction for women with secondary lymphedema as a result of treatment of breast cancer.
Methods
Criteria for considering studies for this review
Types of studies
Studies of any language will be included which have patients with LE secondary to breast cancer treatment, use a randomized or quasi‐randomized (i.e., by alternate assignment or date of birth) clinical trial design, and evaluate the effects of CDP or MLD alone on reduction in limb volume as measured by circumference, water displacement or electronic volometer (i.e., Perometer).
Types of participants
All female patients with LE secondary to a breast cancer treatment (i.e., axillary node resection and/or radiation therapy).
Types of interventions
Interventions include CDP or MLD alone.
Types of outcome measures
Primary outcome is reduced arm volume as measured by circumference, water displacement, or electronic volometer (i.e., Perometer) in female patients with chest wall lymphedema. Secondary outcomes are chest wall volume reduction, pain, function, sensations of heaviness, sensations of tension, range of motion, tonometry, infection rates, quality of life, body image assessments, and cost assessments.
Search methods for identification of studies
A computerized search of the medical literature will be undertaken. Medline, EMBASE, Cinahl, and Health Star will be searched from 1950 to present. The text term "physiotherapy" will be searched. Also, the text terms "lymphedema" or "lymphoedema" and "drainage" will be combined. Additionally, "lymphedema" or "lymphoedema" and "physical therapy," will be combined as well as "lymphedema" or "lymphoedema" and "massage." References from these studies will also be searched for relevant clinical trials. Additionally, these databases will be searched on text words "Complete Decongestive Physiotherapy," "Complex Physical Therapy," "Complex Lymphatic Therapy," "Complex Lymphedema Therapy," and "Complex Decongestive Physical Therapy," because these are all terms that have been used to describe decongestive therapy. We will also search on "Foldi," "Vodder," "Casley‐Smith," "Lerner," and "Leduc," which are proper nouns commonly associated with brand names of the intervention. The Cochrane Library CENTRAL/CCTR will also be searched. Additionally the Bodywork Knowledge Base, a database specifically housing citations of the various aspects of massage, will be searched on keywords "lymphedema," "lymphoedema," "edema," "swelling," and "manual lymphatic drainage." Additionally, the Specialised Register maintained by the Cochrane Breast Cancer Group will be searched. (Details of search strategies used by the group for the identification of studies are outlined in the group's module).
Handsearching of journals with the highest number of studies published on the treatment of lymphedema using CDP will be handsearched. This will involve a handsearching of "Lymphology." (Any relevant trials identified will be forwarded to CENTRAL/CCTR if these trials are not already included in the Cochrane Library).
Lymphedema‐associated websites which will be searched for additional references are:
The British Lymphology Society www.lymphoedema.org/bls
The National Lymphedema Network www.lymphnet.org
Lymphoedema Association of Australia www.lymphoedema.org.au
International Lymphology Association http://www.u.arizona.edu/˜witte/ISL.htm
Investigators who have completed research on LE in the past will be contacted to determine if there are any studies that are currently under review for publication and to identify any additional randomized studies not previously identified in the systematic literature search.
Data collection and analysis
a. Selection of Trials
Two reviewers will review the titles (and abstracts when available) of the articles identified for possible inclusion. Those believed to meet inclusion criteria will be obtained for a hard‐copy review by both reviewers. Disagreements will be resolved by discussion in order to decide which studies meet inclusion criteria. Duplicate publications of the same study will be excluded and documented in the final report.
b.1. Assessment of study quality
All included trials will be assessed by two reviewers blinded to each other's assessments for trial quality and disagreements on scoring will be resolved by discussion. The following criteria, modified from the Jadad scale (Jadad 1996) and the Cochrane Handbook will be used to assess trial quality:
1. Was the trial reported as randomized? Yes, No
2. Was the randomization scheme adequate? Yes, No, Don't know (Adequate will include coin tosses, table of random numbers, or computer generated schemes. Inadequate will include assignment by date of birth, hospital number or alternate assignment).
3. Was the method of allocation concealment adequate? Yes, No, Don't know (Adequate will include calling a central office once eligibility has been determined, or using opaque envelopes sequentially numbered.
Inadequate will include using an open list, drawing numbers from a hat, or using envelopes not sequentially numbered).
4. Was the outcomes assessor reported as blinded? Yes, No
5. Were the numbers and reason for dropping out reported for each group? Yes, No, Don't know
6. How would you rate the overall chance that this trial is influenced by bias? Low chance of bias, high chance of bias, or don't know.
b. 2. Assessment of the quality of CDP procedure
Two physical therapists trained in CDP and having at least two years of experience using this technique in treating LE secondary to breast cancer will assess the adequacy of the manual lymphatic drainage treatment. They will be provided with a description of the study population and CDP technique and asked, 'Given your clinical experience, would you rate the treatment as 'adequate' or 'inadequate' or 'not sufficient information provided' for this population?' Kappa statistics will be calculated between panelists. Differences will be resolved by discussion. In sensitivity analysis we will compare the 'adequate' treatments vs. all others (inadequate, insufficient information). When data are able to be pooled, sensitivity analysis will be done by pooling 'adequate' interventions vs. 'not adequate.' When data cannot be pooled, the sensitivity analysis will be performed using a chi square analysis of two variables: result of study x adequacy of treatment.
c. Data extraction
Two reviewers will extract data on the following characteristics of the included trials:
1. Demographic and study information (age and gender of participants, and recruitment site information, year of study, country of study).
2. Clinically relevant information (details of breast cancer treatment, grade of LE, and duration of LE).
3. Experimental treatment (CDP or MLD alone) (total number of sessions, number of sessions per day and per week, number of weeks of treatment, duration of each treatment session, whether compression bandaging was or was not used, and any details about the credentials or experience of the therapist).
4. Control treatment (including a description of the control intervention, the total number of sessions, number of sessions per day and per week, number of weeks of treatment, duration of each treatment session, whether compression bandaging was or was not used, whether exercise was used, and any details about the credentials or experience of the therapist).
5. Outcomes (the outcomes that have been measured including the proportions of responses for dichotomous data and the means and standard deviations for continuous data). The data will dictate what we extract. For example, with the primary outcome, volumetric changes are continuous variables and will be extracted as means and SDs. Occasionally, a trial will define an a priori volumetric change as a 'clinical response,' and report the number of persons who are responders/nonresponders. These outcomes will be treated as proportions.
6. Any adverse effects reported. The most common adverse effects are: increased pain, increased swelling, skin infection, decreased mobility. We will record these and other adverse effects that are described in the papers.
7. Drop‐outs and withdrawals (including the number and reason for dropping in each group).
d. Data Analysis
All trials included in the systematic review will be entered into Review manger 4.1 (RevMan). The results for each group will be weighted by the sample size of the group. For dichotomous results, we will treat dropouts and withdrawals in both groups as nonresponders (intention‐to‐treat, worst case analysis). Dichotomous data will be summarized as relative risks (RRs) with 95% confidence intervals. For continuous data (i.e., volumetric changes) the standardized mean difference will be calculated.
Because overall effects will only be estimated for groups of trials using the same type of control group (e.g., medication, pneumonic compression device) and same outcome, it is likely that more than one meta‐analysis will be performed. A fixed effects model will be used unless there is significant heterogeneity, in which case a random effects model will be used.
Arm volume is usually the primary outcome in LE clinical trials; however, there is little correlation between the common measures of arm volume (i.e., displacement vs. arm circumference). Therefore, we will report on the method used and only pool those trials using the same arm‐volume measurement technique.
As mentioned, there are two phases of LE treatment using CDP or MLD alone. These are the initial therapeutic phase and the maintenance phase. We will analyze studies examining the initial phase (the phase in which the greatest decreases in arm volume are expected) separately from trials assessing the maintenance phase.
A sensitivity analysis will be conducted to evaluate the robustness of the meta‐analysis. This analysis will examine the effects of methodological quality on study outcomes by assessing for associations between individual items in the methodological quality checklist and the study outcomes. When data are able to be pooled, sensitivity analysis will be done by pooling the 'yes' vs 'no' responses to each methodological quality item. When data cannot be pooled, the sensitivity analysis will be performed using a chi square analysis.
