A Systematic Review of the Evidence for Complete Decongestive Therapy in the Treatment of Lymphedema From 2004 to 2011

Transcription

1 Critical Review A Systematic Review of the Evidence for Complete Decongestive Therapy in the Treatment of Lymphedema From 2004 to 2011 Bonnie B. Lasinski, MA, PT, Kathryn McKillip Thrift, BS, DeCourcy Squire, PT, Melanie K. Austin, MPH, Kandis M. Smith, PhD, Ausanee Wanchai, RN, PhD, Jason M. Green, PhD, Bob R. Stewart, EdD, Janice N. Cormier, MD, MPH, Jane M. Armer, RN, PhD, FAAN Objective: To critically analyze the contemporary published research that pertains to the individual components of complete decongestive therapy (CDT), as well as CDT as a bundled intervention in the treatment of lymphedema. Data Sources: Publications were retrieved from 11 major medical indices for articles published from by using search terms for lymphedema and management approaches. Literature archives of the authors and reference lists were examined through Study Selection: A research librarian assisted with initial literature searches by using search terms used in the Best Practice for the Management of Lymphoedema, plus expanded terms, for literature related to lymphedema. Authors sorted relevant literature for inclusion and exclusion; included articles were sorted into topical areas for data extraction and assessment of level of evidence by using a published grading system and consensus process. The authors reviewed 99 articles, of which 26 met inclusion criteria for individual studies and 1 case study did not meet strict inclusion criteria. In addition, 14 review articles and 2 consensus articles were reviewed. Data Extraction: Information on study design and/or objectives, participants, outcomes, intervention, results, and study strengths and weaknesses was extracted from each article. Study evidence was categorized according to the Oncology Nursing Society Putting Evidence into Practice level of evidence guidelines after achieving consensus among authors. Data Synthesis: Levels of evidence were only moderately strong, because there were few randomized controlled trials with control groups, well-controlled interventions, and precise measurements of volume, mobility and/or function, and quality of life. Treatment interventions were often bundled, which makes it difficult to determine the contribution of each individual component of treatment to the outcomes achieved. Conclusions: CDT is seen to be effective in reducing lymphedema. This review focuses on original research about CDT as a bundled intervention and 2 individual components, manual lymph drainage and compression bandages. Additional studies are needed to determine the value and efficacy of the other individual components of CDT. PM R 2012;4: INTRODUCTION Lymphedema (LE) is a swelling of the soft tissues that results from the accumulation of protein-rich fluid in the extracellular spaces. It is caused by decreased lymphatic transport capacity and/or increased lymphatic load. The severity of LE is usually graded by using the scale (Table 1) from the International Society of Lymphology [1]. LE is a chronic, progressive condition that impairs mobility and joint movement as the swollen areas increase in size and weight, often causing postural alterations and pain as the individual struggles to B.B.L. Lymphedema Therapy and The Boris- Lasinski School, Woodbury, NY Disclosure: nothing to disclose K.M.T. The Dr. Vodder Schools, Victoria, British Columbia, Canada/Walchsee, Austria Disclosure: nothing to disclose D.S. Complex Lymphatic Therapy Courses and Piedmont Hospital Lymphedema Program, Atlanta, GA Disclosure: 3, Emory University Physical Therapy School, Lighthouse Lymphedema Network, Thomas Hovatter Lymphedema Awareness day; 9, Lymphology Association of North America Board member M.K.A. University of Missouri Sinclair School of Nursing, Columbia, MO Disclosure: nothing to disclose K.M.S. University of Missouri Sinclair School of Nursing, Columbia, MO Disclosure: nothing to disclose A.W. University of Missouri Sinclair School of Nursing, Columbia, MO Disclosure: nothing to disclose J.M.G. University of Missouri Sinclair School of Nursing, Columbia, MO; Department of Computer Science, College of Engineering, University of Missouri, Columbia, MO Disclosure: nothing to disclose B.R.S. University of Missouri Sinclair School of Nursing, Columbia, MO Disclosure: nothing to disclose J.N.C. MD Anderson Cancer Center, University of Texas, Houston, TX Disclosure: nothing to disclose J.M.A. Ellis Fischel Cancer Center, DC Suite 415, 115 Business Loop 70 West, University of Missouri Sinclair School of Nursing, Columbia, MO Address correspondence to: J.M.A.; armer@ missouri.edu Disclosure: nothing to disclose Disclosure Key can be found on the Table of Contents and at Submitted for publication March 27, 2012; accepted May 8, PM&R 2012 by the American Academy of Physical Medicine and Rehabilitation /12/$36.00 Vol. 4, , August 2012 Printed in U.S.A.

2 PM&R Vol. 4, Iss. 8, Table 1. Stages of lymphedema* Stage 0, latency: Lymph stasis is present limb at No clinical signs of lymphedema risk Individual may experience intermittent ache, fatigue, heaviness in limb at risk 1 Edema pits on pressure Swelling reduces completely on elevation No clinical fibrosis present Variable time to progression to stage 2 2 Edema no longer pits Swelling does not completely reduce on elevation Clinical fibrosis is present Can be combination of pitting and nonpitting edema: mixed stage I II 3 Skin becomes thickened and leathery Hypertrophy of subcutaneous tissues develop Papillomas and warts develop on skin Tissue bulges and skin folds develop Less common in upper extremity *Reproduced with permission from [1]. perform activities of daily living. Even a relatively minimal LE can alter body image and cause significant impairments in physical and psychosocial function and quality of life (QOL). Complete decongestive therapy (CDT) is currently recognized as the standard of care in LE treatment [1-3]. Ideally, the initial intensive phase of CDT is performed daily until maximal volume reduction and normalization of tissue texture is achieved [3]. CDT consists of (1) an average of 60 minutes of manual lymph drainage (MLD), a specialized gentle massage to stimulate the lymphatic system; (2) multilayer, short-stretch compression bandaging (CB) with foam or layers of fabric padding of the affected limbs, also known as multilayer limb bandages (MLLB) or CB; (3) exercises to enhance lymphatic pumping; (4) meticulous skin care of the affected areas; and (5) fitting of appropriate compression garments (CG) to maintain the reductions achieved through treatment. Patient education in self-care is critical for successful long-term outcomes. Unfortunately, reimbursement for CDT has been hampered by a lack of rigorous research evidence. Boris et al [4] and Ko et al [5] conducted seminal research in the 1990s that demonstrated the value of CDT in patients who adhered to the self-care regimen after intensive therapy over longer periods of follow-up [4] and with larger samples than those generally found [5]. Boris et al [4] conducted a retrospective study of a 4-week intensive course of CDT in 119 patients (56 unilateral upper extremity, 38 unilateral lower extremity, 23 bilateral lower extremity, and 2 bilateral upper extremity) with 36-month follow-up. Initial volume reductions averaged 62.6% in patients with unilateral upper extremity LE and 68.6% in patients with 1 affected leg. After 36 months follow-up, the average reduction increased to 63.8% for unilateral upper and remained at 62.7% for unilateral lower extremity LE. Patients with bilateral LE maintained their initial reductions. Ko et al [5] conducted a prospective study of 299 patients with LE (149 upper extremity and 150 lower extremity) treated with CDT for an average of 15.7 days. LE reductions averaged 59.1% for the patients with upper and 62.7% for patients with lower extremity LE. The average follow-up was 9 months, and 86% of the patients were adherent and maintained 90% of their initial reductions, whereas nonadherent subjects lost 33%. Patient adherence with compression and decongestive exercises correlated with maintaining and improving treatment results [4,5]. CDT is criticized as time consuming, costly, and not well researched. Although identified as the criterion standard management approach for LE, the level of evidence rating by the International Lymphoedema Framework (www. lympho.org) for MLD was C: experienced common sense judgment (Oncology Nursing Society Putting Evidence into Practice [PEP] level: effectiveness not established (Oncology Nursing Society, Pittsburgh, PA), and for CB was B: limited supporting research evidence (PEP level: likely to be effective ). Stakeholders require evidence to support the intensity of initial treatment as well as the importance of each component of the regimen in achieving outcomes that lead to improved function and QOL. Therefore, the American Lymphedema Framework Project (ALFP) ( undertook an extensive review of the literature that pertains to the management of LE from 2004 to 2011 with the following aims: (1) to provide evidence for the second edition of Best Practice Document for the Management of Lymphoedema, a document published in 2006, based on literature reviewed through 2004 (Tables 2 and 3) [6]; and (2) to inform efforts to elevate the level of evidence for LE treatment. In 2009, the Medicare Evidence Development and Coverage Advisory Committee (MEDCAC) met to review the current evidence for a variety of LE treatment modalities. The MEDCAC concluded that the existing evidence gives reasonable confidence that these commonly used protocols for the treatment for LE are effective: (1) CDT (including MLD, compression bandaging and/or garments, exercise, skin care), (2) CDT in conjunction with a sequential pump, and (3) CB systems, CGs, and sequential pumps. However, it was recognized that there remains a paucity of high-level evidence that supports best practices in the management of LE. The Centers for Medicaid and Medicare Coverage and Analysis Group will use the 2009 MEDCAC analysis [7] in revising policies to coordinate coverage for LE treatment. Therefore, it is imperative to document the levels of evidence in the contemporary literature, , and beyond, both for best practices

3 582 Lasinski et al EVIDENCE FOR COMPLETE DECONGESTIVE THERAPY IN LYMPHEDEMA Table 2. Review articles and consensus document included in CDT systematic review Study Topic Time Frame No. Reviewed Conclusions Cheifetz and Haley [52], 2010 Devoogdt et al [31], 2009 Karki et al [53], 2009 Kligman et al [51], 2004 Review of articles on management of BCRL by using the Sackett levels of evidence Review of treatment modalities for BCRL Effectiveness and cost of current treatment practices for treating BCRL in Finland Evidence base for treatment of BCRL articles (11 on CDT and 10 on exercise; also 9 on risk factors) Several treatment options exist for LE management. Although the literature does not identify which treatment is superior, education, exercise, and MLD are important components of a successful treatment program for reducing edema and increasing function. The article did not review the role of compression articles CDT is an effective therapy for LE, but the role of each component is unclear; higher quality studies are needed. Five of the studies covered IPC; several studies included arm elevation, but none focused on it. The researchers concluded from the literature available that the long-term effect of IPC and the benefits of elevation on LE are not yet proved. One RCT reviewed showed higher reduction of LE by the MLD group than the group without MLD and similar reduction in subjective symptoms between the groups RCTs The main problems that the researchers found with the studies that they reviewed were poor methodology, including small sample sizes and high risk of bias. They found that CB had some effectiveness but that the studies did not establish the effectiveness of other modalities, such as laser (2 studies; decrease of LE was not analyzed). IPC (4 studies): no significant difference between IPC and no intervention; no significant difference between MLD and CG vs IPC and CG; no significant difference between IPC and CG vs SLD and CG; greater decrease between IPC with CG vs IPC without CG. Electrical stimulation (1 study): no significant difference between CG with vs CG without electrical stimulation RCTs Articles on physical therapies and on medical interventions were reviewed as well as questionnaires to oncologists and therapists for their recommendations and/or opinions. On the basis of the studies and the survey of health care practitioners, the researchers recommended the use of compression garments with an added benefit of MLD. In 1 study reviewed, IPC compared with no intervention did not show a significant difference. Another study examined MLD as a strategy to prevent the onset of LE and showed a nonstatistically significant advantage to the use of MLD vs CG. The review researchers found most of the studies inconclusive because of poor study design and/or small sample size and recommended further research.

4 PM&R Vol. 4, Iss. 8, Table 2. Continued Study Topic Time Frame No. Reviewed Conclusions Leal et al [32], 2009 McCallin et al [30], 2005 Megens and Harris [54], 1998 Meneses and McNees [55], 2007 Review of treatment modalities for BCRL Effectiveness of treatment for BCRL Effectiveness of physical therapy in the management of BCRL Upper extremity LE after treatment for breast cancer: a review of literature studies Within the therapeutic modalities used for LE treatment, CDT undoubtedly has the strongest scientific support. Its application, together with IPC, has demonstrated efficacy. New techniques with satisfactory results are being studied, such as HVES and laser therapy. Combined techniques produce the most beneficial effects, depending on the physiopathology of the LE. Selection of best combination of therapies is based on a detailed assessment of individual cases. MLD was not found to be an effective stand-alone treatment; CB is needed in combination with MLD for optimal effectiveness articles The researchers reviewed studies on CDT and on individual components (MLD, CB, exercise, CG) and on adjuncts to treatment (laser, IPC). They found the studies methodologically limited. The weight of the evidence showed effectiveness of CDT but failed to strongly demonstrate the role of the individual components. The researchers concluded that studies of MLD, exercise, skin care, and laser were limited in number and suggest that individuals with stage II or III LE may benefit from the addition of MLD to their treatment regimen. They concurred with other researchers who conclude that larger studies with longer follow-up examining individual components of CDT are needed articles Researchers modified Sackett rules of evidence to assess their study results. Level 1 recommendation: CG reduced limb size after 6 mo use; Level c recommendations: elevation alone is not useful; CDT does tend to reduce limbs; therapies with combined modalities (eg, CDT plus IPC) also showed promise articles cited Although LE is not a curable condition, ways to prevent skin breakdown and relieve suffering are available. Risk factors for LE include axillary node dissection (vs smaller risk in sentinel node biopsy), weight gain, infection, and injury. Patients aware of LE can take precautions to lower their risk. Results of studies show reduced QOL with LE. Although methodology in studies is sometimes weak, CDT appears to be helpful in reducing lymphedema.

5 584 Lasinski et al EVIDENCE FOR COMPLETE DECONGESTIVE THERAPY IN LYMPHEDEMA Table 2. Continued Study Topic Time Frame No. Reviewed Conclusions Moseley et al [28], 2007 Badger et al [50], 2004 Partsch et al [39], 2008 Review of conservative therapies for BCRL RCTs that tested physical therapies with a follow-up period of at least 6 mo studies Studies on CDT and on MLD, compression, laser, oral drugs, exercise, and elevation were reviewed for quality of evidence and effectiveness of treatment. Treatments by health care professional had better results than self-care, but self-care was better than no care and important for maintenance of results. MLD and compression had greatest benefit, followed by pumps, MLD alone, oral drugs, laser, compression, exercise, elevation, and self-mld. Level of evidence was poor, and there is a need for larger, betterdesigned clinical trials (updated through 2007 with no changes in conclusion) 36 articles reviewed (3 RCTs met criteria) Only 3 (representing N 150) of 36 reviewed studies met established criteria. Each had limitations: different interventions were not replicated, and findings are unable to be compared. One crossover study of MLD followed by self-administered massage vs no treatment concluded that improvements seen in both groups were attributable to the use of compression sleeves and that MLD provided no extra benefit at any point during the trial. Another trial looked at hosiery vs no treatment and had a very high dropout rate, with only 3 of 14 participants in the intervention group finishing the trial and only 1 of 11 in the control group. The researchers concluded that wearing a compression sleeve is beneficial. The bandage-plus-hosiery vs hosiery-alone trial concluded that, in this mixed group of participants, bandage-plushosiery resulted in a greater reduction in excess limb volume than hosiery alone and that this difference in reduction was maintained long term. There is a great need for well-designed RCTs to determine the best approaches to managing LE. CVI and LE references A low level of compression (10-30 mm Hg) is effective in managing venous leg edema. Moderate compression (30-40 mm Hg) is effective in healing ulcers, preventing postthrombotic syndrome, and managing LE. There is little evidence on the differential effect of elastic and short stretch CB. Sustained pressure of 60 mm Hg in supine position decreased arterial blood flow. If pressure 40 mm Hg is needed, then inelastic material provides greater working pressure and lower resting pressure on the limb.

6 PM&R Vol. 4, Iss. 8, Table 2. Continued Study Topic Time Frame No. Reviewed Conclusions Poage et al [2], 2008 Shrubb and Mason [40], 2006 Level of evidence for LE management Review of literature on DVT in the general population to consider applicability to the population with LE: DVT and leg LE Consensus document ISL [1], 2009 Consensus document of the ISL ILF Framework [6], 2006 Consensus document of the ILF articles Common limitations were found in many of the studies reviewed, including small sample sizes, lack of randomized control groups, imprecise intervention standardization within and across groups, imprecise or diverse measurement approaches, and limited follow-up. Longer follow-up time frames are needed before definitive conclusions can be reached regarding interventions for patients at risk for or with LE articles Difficulties with research: heterogeneous populations, lack of knowledge about LE, care based on symptom management. Some of the studies reviewed were of patients receiving palliative care, some of them bedbound, and so extrapolating conclusions from them to the ambulatory LE population was problematic. Compression stockings are effective in managing DVT in ambulatory patients. MLLB is superior to off-the-shelf CG, with a better fit and comfort for patients. Patients with LE may need higher pressure than patients with DVT No references cited in document (developed by group consensus panel from literature review and consensus of subject matter experts) The document attempts to amalgamate the broad spectrum of protocols advocated worldwide for the diagnosis and treatment of peripheral LE into a coordinated proclamation that represents the international community. No treatment method has really undergone a satisfactory meta-analysis (let alone rigorous, randomized, stratified, longterm, controlled study). With this understanding, the absence of definitive answers and optimally conducted clinical trials and with emerging technologies and new approaches and discoveries on the horizon, some degree of uncertainty, ambiguity, and flexibility along with dissatisfaction with current LE evaluation and management is to be expected citations The BPD was derived from a U.K. National Consensus on Standards of Practice for people who are at risk of or who have LE. The consensus process was launched in 2002 by the ILF with input from national patient support groups, patients with LE, national professional LE groups, clinical experts, and industry. The document was reviewed by an international panel of experts and endorsed by key national LE organizations. Recommendations were derived by consensus with classification according to the U.K. NHS Health Technology Assessment model for guideline development. CDT manual lymphatic drainage, compression bandaging; BCRL breast cancer treatment related lymphedema; LE lymphedema; MLD manual lymphatic drainage; IPC intermittent pneumatic compression, pump; RCT randomized controlled trial; CB compression bandaging using padding and short-stretch bandages and garments, exercise, skin care; CG compression garment; SLD simplified/self-mld; HVES high-voltage electrical stimulation; QOL quality of life; CVI chronic venous insufficiency; DVT deep vein thrombosis; MLLB multilayer limb bandaging; ISL International Society of Lymphology; ILF International Lymphoedema Framework; BPD best practices document; NHS National Health Service.

7 586 Lasinski et al EVIDENCE FOR COMPLETE DECONGESTIVE THERAPY IN LYMPHEDEMA Table 3. Articles detailed in CDT systematic review Study Design Sample Define LE Measurements Likely to be effective Andersen et al Prospective randomized [17], 2000 study 42 (22 standard therapy; 20 standard therapy MLD) Volume difference between arms of at least 200 ml Volume calculation using numerical integration by piecewise quadratic approximating the Simpson rule of integration Badger et al [38], 2000 Bernas et al [60], 2005 RCT 83 (34: CB CG; 49: CG) Retrospective 82: BCRL; 40: MLD; 42: MLD CB 20% larger than other limb 20% or less increase Perometry or circumference; no standard TM every 4 cm, TC Damstra et al [36], 2008 Case control 29 (20 with LE; 9 without LE) ND Volumetry Damstra Partsch [37], 2009 RCT 36 (18 low pressure CB; 18 high pressure CB) ND Volumetry and pressure changes under bandages Karadibak et al [15], 2005 RCT 53 BCRL;27 CDT; 26 CDT without MLD 2 cm Volumetry and TM Hamner and Retrospective 119 BCRL 10% LV change Volumetry Fleming [21], 2007 Jeffs [29], 2006 Retrospective 168 of 263 BCRL ND Perometer Kasseroller and Brenner [27] 2009 Prospective RCT 61 BCRL, 2 groups: CB alginate LVF classification U/E TM every 4 cm, cyl

8 PM&R Vol. 4, Iss. 8, Table 3. Continued Intervention F/U Results Strengths Weakness Standard therapy: CG (32-40 mm Hg); ex instruction; education in skin care and precautions; standard therapy plus MLD group: MLD 8 times in 2 wk and instruction in daily SLD CB CG vs CG alone Only 38 patients had f/u at 12 mo No significant difference in the reduction in LE over time between the 2 groups; results from pooled data showed that LE reduction after 1 mo was statistically significant compared with the baseline; over time, further slight improvement but not significant 24 wk CB CG was 2 times as effective as CG alone Repeated measurements; control group; power of 90% sample SI; 6 mo f/u SS; NB; MLD 8 times in 2 wk and was part of 1hTx, including SLD teaching MNS Vod Var MLD 5 times/wk for 2 wk, with or without CB CB to L/E of subjects with and without LE CB to U/E, measured subbandage pressure over time; measured LV change over time CDT, including Vod Var MLD vs CDT without MLD 3 times/wk for 4 wk CDT with Vod Var MLD, 2/wk for mean 7.5 wk CDT and/or selfcare, varied according to degree of swelling, needs, and life expectancy CDT with Vod MLD 2 times/d for 5 d for 3 wk; 3 weekends: IPC CB or IPC CB alginate No funding for f/u No significant difference in % reduction of swelling between groups: MLD only: LV loss 80.6% 0.2% vs MLD CB: LV loss 72.6% 0.2%, mean (SD) 24 h Pressure under CB reduced 37% in normal subjects and 48% in subjects with LE; LV loss was 2.6% (60 ml) in normal subjects and 1.2% (50 ml) in subjects with LE 24 h 9.2% volume reduction with low pressure CB; 4.8% reduction with high pressure CB; low pressure tolerated better; all CB lost pressure over time None listed Reduced excess volume: 55.7% in MLD group, 36% in non-mld group None listed Reduced excess volume 41.7% on average 12 mo 100% experienced reduction: 70% reduction in patients with limb edema; 40% reduction in patients with complex conditions; 30% reduction in patients with mild swelling doing home care None listed Subjective parameter of patients comfort and tolerance clearly more advantageous by group given alginate CB No dropouts in either group SI accurate measurements SI addressed patient comfort over time; compared with high or low pressure CB; QOL measured Large sample; perometer increased; reliability IP; no dropout; HS SS, NR, NB, NLTF SS; HS; only 24-h f/u SS, only 24-h f/u SS, NB Measured to only 2 inches above elbow NC, NR NB

9 588 Lasinski et al EVIDENCE FOR COMPLETE DECONGESTIVE THERAPY IN LYMPHEDEMA Table 3. Continued Study Design Sample Define LE Measurements McNeely et al [18], 2004 Prospective RCT 50 BCRL (25 CB MLD; 25 CB) 150 ml Volumetry TM every 4 cm, TC at baseline, end of wk 1, 2, 3, 4 Pinell et al [46], 2008 Retrospective 72 (16 with active tumors) 2 cm TM, volume formula Vignes et al [26], 2006 Prospective 357 BCRL ND TM every 5 cm, TC Vignes et al [14], 2007 Prospective 537 BCRL Difference between limbs TM every 5 cm: TC Williams et al [16], 2002 RCT with crossover 31 BCRL (15 patients in group A; 16 patients in group B) 2 consistent LV measurements of 10% excess volume TM every 4 cm, cyl, calipers, ultrasound Yamamoto et al [45], 2008 Prospective 31 U/E, 52 L/E ND TM at 7 points, TC Benefits balanced with harms Johansson et al [10], 2005 RCT 31 BRCL (16 CG ex; 15 ex only) 10% LV change Volumetry, bioimpedance Kim and Park [11], 2008 Prospective 57 unilateral LE, gynecologic cancer related ND TM at 6 points

10 PM&R Vol. 4, Iss. 8, Table 3. Continued Intervention F/U Results Strengths Weakness CDT: 45 min Vod MLD 5/wk for 4 wk CB; or CB only min CDT, 5 times/wk until plateau CDT, 30 min Vod Var MLD, 5 times/wk, mean 11.8 d CDT: 30 min Vod Var MLD Tx 11 d; then HP of daily CG, CB 3 nights/ wk, MLD 1-3 times/wk Group A: 45 min Vod MLD, 5 times/wk for 3 wk then a 6-wk nontreatment period, then SLD daily for 3 wk; group B: SLD daily for 3 wk, then 6-wk nontreatment period, then 45 min Vod MLD 0.5/wk for 3 wk; both groups: CG. skin care, and information about LE 60 min CDT, 6/wk until plateau None listed MLD group reduced 44.1%- 46.1%; CB only group reduced 37.2%-38.6%; mild LE subjects in MLD group reduced the most None listed Reduction of excess LV for both groups: 22%, nontumor: range 21% to 164% in median of 12 visits. Similar reductions took longer for patients with active tumors: range of 23% to 72% in median of 15 visits None listed More swelling associated with higher BMI, longer duration, but all groups reduced 36%- 38% 426 patients (80%) at 6 mo; 356 patients (66%) at 12 mo None listed Volume decreased significantly with mean absolute volume reduction of 407 (range, ml); CB and CG factor in continued reduction 10% in 28% of patients who were available for follow-up MLD led to significant decrease of excess LV and reduced dermal thickness, improved emotional function, dyspnea, and sleep disturbance, but other QOL subscales did not achieve statistical significance; SLD did not result in significant changes in any outcome measurements None listed U/E lost 54% after first day, 8% more after second day, plateaued at 59.1% in 4 d (range, 2-10 d); L/E lost 55.5% after first day, 11% more after second day, plateaued at 73.5% in 5 d (range, 2-17 d) BA; sample size chosen for 80% power Included patients on chemotherapy Large sample size Repeated and LTE; large sample size; HS and Tx; IP; QT Control group; repeated measures; included truncal SS, NLTF, 10% dropout NB, NC; LE not defined NR; 34% of patients lost to f/u at 12 mo SS, NB Japanese subjects had lower BMI than subjects in U.S. studies; NC, NB Ex only vs ex and CG min MLD 5 times/wk, 2-4 wk CB, ex, fit with CG; home program SLD, ex, CG; CB 3 nights/wk 24 h Immediate increase in volume after ex; at 24 h, no significant difference from baseline or between the groups 1 mo 43.6% average volume decrease; trend toward decreased QOL with increased LV and statistically significant increase QOL with LV decrease HS, multi-site SS MNS; NC; bias; researchers scored QOL questionnaire; 8 subjects lost in f/u

11 590 Lasinski et al EVIDENCE FOR COMPLETE DECONGESTIVE THERAPY IN LYMPHEDEMA Table 3. Continued Study Design Sample Define LE Measurements Effectiveness not established Hwang et al [48], 2007 Prospective 7 U/E; 13 L/E ND Lymphoscintigram Johansson et al [20], 1998 Prospective RCT 24 BCRL (12 MLD; 12 IPC) More than 10% difference between arms Volumetry, arm mobility, strength, and subjective assessment Johnstone et al [22], 2006 Retrospective 11 patients (no. of patients who returned for f/u of 82 seen) ND F/U TM every 4 cm, cyl Karadibak et al [23], 2008 Prospective 62 BCRL 2 cm TM every 3 cm, volumetry Kim et al [12], 2007 Prospective 53 BCRL ND TM at 6 sites, TC Liao et al [24], 2004 Prospective 18 U/E, 12 L/E 2 cm TM, every 10 cm, TC Mondry et al [13], 2004 Prospective 20 BCRL 2 cm, 200 ml TM every 2 inches in both directions from olecranon

12 PM&R Vol. 4, Iss. 8, Table 3. Continued Intervention F/U Results Strengths Weakness Lymphoscintigram at baseline, CDT, then 1-y followup on results Part I: 2 wk CG, patients were randomized; part II: group 1, 45 min MLD for 2 wk; group 2, 2 h IPC at mm Hg for 2 wk min CDT for 5 d/wk, to plateau; then maintenance phase: CG by day, CB by night, ex, SLD CDT, 60 min Vod Var MLD, 5 times/wk for 3 wk, then HP of CG (10-20 mm Hg) and ex CDT with min Vod MLD 5/wk for 2-4 wk; HP: SLD, ex, CG daily, with CB 3 nights/wk CDT, 30 min C-S MLD, 5/wk, 4-21 d min CDT 5 times/wk for 2-4 wk 12 mo Those with good results from CDT after 1 y (skin condition, decrease in reported symptoms and a volume reduction of at least 15%) had had lymphoscintigrams that showed a lack of collateralization and/or a main lymph vessel before treatment None listed During part II, significant reduction in MLD group 15% (P.001); and, in the IPC group, 7% (P.03); only the MLD group reported a further decrease of tension and heaviness; both groups significantly decreased arm volume but no significant difference between the 2 groups Median 7.5 mo (range, 6-13 mo) 7 patients reporting compliance had significantly better results at f/u (loss of 3.4% edema [range, 2.8% to 10.9%]) than did 4 patients reporting noncompliance (gain of 3.4% [range, 0.05% to 6.22%]) 12 wk for QOL Reduced 55% in intensive phase; QOL and activity level increased at follow-up 6 mo QOL and functional scores measured by SF-36; increased after treatment; reduction of limb after treatment: 42%; at 6 mo, regressed to 15% reduction from baseline None listed Pretreatment swelling correlated with age and duration of swelling; percentage of reduction was not significantly affected by these factors: U/E reduced 67.6%; L/E reduced 68.1% 3 mo, 6 mo, 1 y Decrease 138 ml after intensive phase; longer treatment resulted in a greater decrease but less adherence during Tx course; QOL gradually increased during Tx and f/u; pain gradually decreased; limb size began increasing during f/u Blinded Control group HS; full CDT protocol used Full CDT protocol used Full CDT protocol used Full CDT protocol used, LTE SS, LE not defined, not stated if 15% reduction is of limb volume or edema volume SS, NLTF; MLD group older and LE longer duration SS, NC; majority lost to f/u; LE not defined NC, LV reduction not assessed at follow-up SS, NB, NC, adherence not measured SS, NC NB, no reasons given for variation in endpoint SS; NB; NC; did not measure higher than 6 inches proximal to the olecranon; only 13 of 20 patients left at final f/u

13 592 Lasinski et al EVIDENCE FOR COMPLETE DECONGESTIVE THERAPY IN LYMPHEDEMA Table 3. Continued Study Design Sample Define LE Measurements O Neill and Beatus [25], 2006 Prospective 17 BCRL Symptoms Volumetry Yamamoto and Yamamoto [44], 2007 Retrospective 27 U/E, 55 L/E ND TM at 7 points; TC; ultrasound Effectiveness not established: case study Whitaker [43], 2007 Single case study 1 ND Circumference CDT manual lymphatic drainage, compression bandaging and garments, exercise, skin care; LE lymphedema; f/u follow-up; MLD manual lymphatic drainage; CG compression garment; ex exercise; SLD simplified/self-mld; SS small sample size; plateau no further significant reduction; NB nonblinded; Tx treatment; RCT randomized controlled trial; CB compression bandaging by using padding and short-stretch bandages; SI standardized intervention; MNS measurements not standardized; BCRL breast cancer treatment related lymphedema; TM tape measurement; Vod Var MLD variation of the Vodder method; Vod MLD Vodder MLD; LV limb volume; NR not randomized; NLTF no long-term follow-up; ND not defined; volumetry water displacement; L/E lower extremity; HS homogenous sample; U/E upper extremity; QOL quality of life; LVF location, volume difference, skin-fold index; cyl cylinder formula used to figure volume; IPC intermittent pneumatic compression, pump; IP inpatient; TC truncated cone formula used to figure volume; BA blinded assessors; BMI body mass index; NC no controls; HP home program; LTE long-term evaluation; QT qualified therapist noted; SF-36 Short-form health survey with 36 questions; C-S Casley-Smith method of MLD. in LE management and to guide future research efforts to support the evidence base for reimbursement. The goal of this article is to critically analyze the research literature from 2004 to 2011 that pertains to the individual components of MLD and CB as well as CDT as a multicomponent LE treatment regimen. This review will examine the methods, outcome measures, strengths and limitations, and results of the research studies. The findings will then be discussed in terms of the effectiveness of CDT treatment for LE, with recommendations for clinical practice and future research. Review of alternative therapies outside CDT is under development elsewhere; this review focuses on original research about CDT as a bundled intervention and 2 individual components, MLD and CB. Due to the selected time frame ( ) and the desire to build upon seminal research from prior years, published reviews were separately evaluated. International non-english literature also was included if a translation was available. METHODS A research librarian assisted with initial searches with the terms used in the Best Practice for the Management of Lymphoedema [6] plus expanded terms for all literature related to LE ( ). The databases searched included PubMed, CINAHL (CINAHL Information Systems, Glendale, CA), the Cochrane Database of Systematic Reviews (The Cochrane Collaboration, Oxford, England), Cochrane Controlled Trials Register (The Cochrane Collaboration, Oxford, England), PapersFirst (OCLC Online Computer Library Center Inc, Dublin, OH), Proceedings- First (OCLC Online Computer Library Center, Inc., Dublin, OH), Worldcat (OCLC Online Computer Library Center, Inc., Dublin, OH), PEDro (Centre for Evidence-Based Physiotherapy, Sydney, Australia), National Guideline Clearing house (Agency for Healthcare Research and Quality, U.S. Department of Health and Human Services, Rockville, MD), ACP Journal Club (American College of Physicians, Philadelphia, PA), and DARE (Centre for Reviews and Dissemination, University of York, York, UK), as well as from our archives. The literature was sorted for inclusion or exclusion (see Figure 1). Selected references were sent to ALFP research associates who reviewed abstracts to apply broad inclusion and exclusion criteria of LE/non-LE (Screening 1). ALFP authors (K.M.S., M.K.A., B.R.S., J.N.C., J.M.A.) sorted relevant literature by passing Screening 1 into included and excluded articles and topical areas (Screening 2). Full articles were requested for selected abstracts for further review. The studies were grouped by topic and assigned to authors for data extraction and to evaluate the quality of the evidence by using a consensus process (Screening 3). Articles from 2011 were identified by the coauthors, screened, and included, as appropriate. Topic authors (B.B.L., K.M.T., D.S.) reviewed the relevant included library for verification of the appropriate-

14 PM&R Vol. 4, Iss. 8, Table 3. Continued Intervention F/U Results Strengths Weakness CDT, 30 min MLD, 3-5/wk for 3-5 wk vs CB only; no ex component None listed Volume decreased with CDT approximately 48.%; increased QOL, increased grip strength, decreased pain SS, NB, NC min CDT, 3-6/wk for 3-26 d (average, 6 d) U/E, 2-35 d (average, 10 d) L/E None listed L/E reduced 73.4%; U/E reduced 58.7% Measured tissue changes with ultrasound SS, NC, NB SLD and applied Whitaker pouch 3 wk Scrotum circumference decreased in 2 places (5.5 and 8.7 cm) Unique measurement tool for genital L/E SS, NC, standard measure; short f/u; insufficient evidence, single case ness of inclusion, and possible omissions, with rationale provided for each. Studies that used case series designs were included if the number of cases in the study was more than or equal to 10. In addition, 1 individual case report on genital LE was included due to the paucity of literature in this area. No gray literature was included: nonrefereed articles, abstracts, and dissertations were excluded. In total, 56 articles were excluded by the topic authors due to duplication, inability to obtain English translation, inadequate sample size, insufficient level of evidence due to study design, and failure to meet inclusion criteria. The level of evidence for each study was assessed by using the research grading system from the PEP level of evidence guidelines (Table 4), a system judged to be both reliable and accessible [2,8,9]. Each topic author rated all articles and the final rating was agreed upon by consensus. RESULTS Ninety-nine articles related to LE treatment were reviewed, of which 26 studies met inclusion criteria (Table 3). In addition, 1 article that did not meet study size, along with 14 review articles and 2 consensus documents were included (see Table 2). Fifteen articles were scored as likely to be effective (according to PEP classification). Seven were randomized controlled trials (RCT) (2 with crossover), 4 retrospective, 3 prospective, and 1 casecontrol study. Twelve studies focused on breast cancer related LE (BCRL) exclusively, whereas 3 included both the upper and lower extremities. Nine used circumference measures, 2 perometry, 5 volumetry or a combination of volumetry and circumference, 1 measured tissue thickness with ultrasound and calipers, and 1 measured pressure under the CB. The number of subjects ranged from 29 to 537. Nine studies reported no follow-up (60%), 2 had 24-hour follow-up (13.3%), 1 had 6-month (6%) followup, and 3 had 12-month follow-up (20%). Assessors were not blinded to the intervention in 14 of the 15 studies, and 2 measured the intervention impact on the subjects quality of life (QOL). Two articles were scored as benefits balanced with harm. One RCT of 31 subjects with BCRL studied volume change achieved with exercise combined with wearing CGs versus exercise alone. Volume was measured by volumetry and bioimpedance, with only a 24-hour follow-up [10]. The other was a prospective study of 57 subjects with lower extremity LE who received 2-4 weeks of MLD plus CB and exercise. Volume change was measured by circumference, with 1-month follow-up [11]. Both studies assessed the impact of intervention on subjects QOL. Assessors were not blinded in either study. Ten articles were scored as effectiveness not established. Seven were prospective (1 RCT), 2 retrospective, and 1 single case study. Six focused exclusively on BCRL, 3 included lower extremity LE, and a single case study on genital LE. Eight studies measured volume reduction by circumference, 1 by lymphoscintigram, and 1 by volumetry. QOL was assessed in 3 studies. Sample size ranged from 1 to 82. Four studies reported no follow-up (44.4%), 1 followed-up subjects for 3 months (11.1%), and 2 followed-up subjects for 6 and 12 months, respectively (22.2%). Assessors were not blinded in 9 of the 10 studies. Individual study details are summarized in Tables 2 and 3. The following are syntheses of studies by topics of CDT as a bundled intervention, MLD alone, and CB alone as well as QOL, special cases, lymphovenous disease, and other differential impacts of CDT. CDT CDT is currently recognized as the criterion standard in LE treatment [1-3], with the intensive phase of CDT, involving (1) MLD, (2) MLLB or CB, (3) exercise, (4) skin care, and (5)

15 594 Lasinski et al EVIDENCE FOR COMPLETE DECONGESTIVE THERAPY IN LYMPHEDEMA Literature Search (n = 5927) Research Librarian Hand Search Screening 1 (n = 1303) Review of titles and abstracts for inclusion Screening 2 (n = 659) Review of titles, abstracts, and full text if necessary for inclusion and sorting into topics Screening 3a CDT & LE Articles Only (n = 99) Review of full text of CDT and LE articles by authors for inclusion Screening 3b (n = 27 studies + 14 reviews + 2 consensus) Final number of articles included by authors after review Figure 1. Literature review process for complete decongestive therapy (CDT) and lymphedema systematic review ( ). CGs. The majority of studies have considered cumulative outcomes of CDT as a bundled intervention, with varying techniques and dosages, and relatively few have considered the contribution of the individual components. Acute and Long-term Management of LE with CDT Several studies looked at what happens after the initial intensive treatment is completed. Kim et al [12] conducted a prospective study with 53 subjects with BCRL, observing them 1 and 6 months after the initiation of an intensive CDT course. At 1 month, subjects limb volume had decreased. At the 6-month follow-up, QOL had improved significantly, whereas limb measurements had started to increase, although not above baseline. Mondry et al [13] conducted a similar prospective study of 20 subjects, with 3-, 6-, and 12-month follow-ups after an intensive phase of CDT. Swelling and pain decreased with treatment, and QOL increased; longer length of treatment was associated with increased girth loss but decreased patient adherence. During followup, girth increased slightly, but QOL improved. Vignes et al [14] followed up 537 consecutive patients prospectively after completion of intensive treatment for 6 months (426 patients) and 12 months (356 patients). The results correlated with adherence in using compression sleeves by day and/or bandaging at least 3 nights a week. A number of subjects were lost to follow-up (34%). In addition, there was no random assignment to the follow-up components of self-mld, sleeves, and bandages. MLD Unbundled Karadibak et al [15] conducted an RCT of arm edema in breast cancer survivors in which the intervention group (n 27) received standard CDT (including MLD) and the control group (n 26) had standard therapy without MLD. Their trial produced arm edema reduction of 36% in the control group without MLD and a 56% edema reduction in the MLD group. Although this was a small trial (n 53), it showed the MLD contribution within standard CDT treatment. Williams et al [16] conducted an RCT crossover study with 31 (16 intervention group, 15 control group) subjects with more severe chronic LE who received MLD daily compared with a control group of subjects who performed a simplified form of self-mld; both groups wore compression sleeves. After 6 weeks, the alternate intervention was assigned. The study found that MLD led to a significant reduction in swelling in both the arm and the trunk. Andersen et al [17] conducted an RCT crossover study of 42 women with mild or early onset BCRL and found that MLD was not a significant factor that contributed to further volume reduction compared with compression alone, whereas McNeely et al [18] conducted an RCT study of 50 women with BCRL and found that MLD followed by compression bandaging was more beneficial than compression bandaging alone. The group who benefited most from compression bandaging plus MLD were women with mild BCRL. Contradictory findings in the various studies may be due in part to differences in study protocols (Table 3), which points to the need for further research with consistent protocols, standard definitions, and longer follow-up. There were inconsistencies in defining and measuring LE [19], a problem pervasive in this body of literature (Table 3) and in wide variations of how the MLD component of CDT was applied by both technique and dose. For example, the MLD might be the Vodder technique [12,16,17,20], a modification of the Vodder technique [15,21-23], or the Casley-Smith technique [24]. In the studies, MLD was performed 2 days a week (8 weeks) [21], 3 days a week (4 weeks) [15], 3 days a week (12 weeks) [23], 3-5 days a week (3-5 weeks) [25], 5 days a week (2-4 weeks) [13,18,20,22,26], or twice a day for 22 days

16 PM&R Vol. 4, Iss. 8, Table 4. Review guidelines:matrices of levels of evidence* PEP Card BPD Technology Assessment Green Go. The evidence supports the consideration of these interventions in practice. Recommended for practice: interventions for which effectiveness has been demonstrated by strong evidence from rigorously designed studies, metaanalysis, or systematic reviews, and for which expectation of harm is small compared with the benefits: At least 2 multi-site, well-conducted RCTs with at least 100 subjects. A panel of expert recommendation derived from explicit literature search strategy; includes thorough analysis, quality rating, and synthesis of evidence. Likely to be effective: Interventions for which effectiveness has been demonstrated from a single rigorously conducted controlled trial, consistent supportive evidence form well-designed controlled trials by using small samples, or guidelines developed from evidence and supported by expert opinion. One well-conducted RCT with fewer than 100 patients or at 1 or more study sites. Guidelines developed by consensus or expert opinion without synthesis or quality rating. Yellow Caution. There is not sufficient evidence to say whether these interventions are effective. Benefits balanced with harms: Interventions for which clinicians and patients should weigh the beneficial and harmful effects according to individual circumstances and priorities. RCTs, meta-analyses, or systematic reviews with documented adverse effects in certain populations. Effectiveness not established: Interventions for which insufficient or conflicting data or data of inadequate quality currently exist, with no clear indication of harm. Well-conducted case control study or poorly controlled RCT. Conflicting evidence or statistically insignificant results. Red Stop. Evidence indicates these interventions either are ineffective or may cause harm. Effectiveness unlikely: Interventions for which insufficient or conflicting data or data of inadequate quality currently exist, with no clear indication of harm. A single RCT with at least 100 subjects that showed no benefit. No benefit and unacceptable toxicities found in observational or experimental studies. Not recommended for practice: Interventions for which lack of effectiveness or harmfulness has been demonstrated by strong evidence from rigorously conducted studies, meta-analyses, or systematic reviews, or interventions in which the costs, burden, or harm associated with the intervention exceed anticipated benefit. No benefit or excess costs or burden from at least 2 multi-site, well-conducted RCTs with at least 100 subjects. Discouraged by expert recommendation derived from explicit literature search strategy; includes thorough analysis, quality rating, and synthesis of evidence. A: Clear research evidence. B: Limited supporting research evidence. C: Experienced common judgment. PEP Putting Evidence into Practice; BPD best practices document; RCT randomized controlled trial. *Reprinted with permission from [56-59] Good (low risk of bias). These studies have the least bias and results are considered valid. A study that adheres mostly to the commonly held concepts of high quality, including the following: a formal randomized controlled study; clear description of the population, setting, interventions, and comparison groups; appropriate measurement of outcomes; appropriate statistical and analytic methods and reporting; no reporting errors; low dropout rate; and clear reporting of dropouts. Fair. These studies are susceptible to some bias, but it is not sufficient to invalidate the results. They do not meet all the criteria required for a rating of good quality because they have some deficiencies, but no flaw is likely to cause major bias. The study may be missing information, which makes it difficult to assess limitations and potential problems. Poor (high risk of bias). These studies have significant flaws that imply biases of various types that may invalidate the results. They have serious errors in design, analysis, or reporting; large amounts of missing information; or discrepancies in reporting.