This prospective study provides evidence that basic lymphedema management is feasible and acceptable to patients in resource-poor areas where lymphatic filariasis is endemic. In general, patients were able to incorporate basic self-care measures, especially leg washing, into their daily routines. When proper hygiene and skin care were emphasized, the incidence of ADLA decreased rapidly to 31% of earlier levels and these reductions were sustained over time. A follow-up study by Dahl and colleagues in 2000–2001 showed that ADLA incidence in these patients remained low and even decreased further (Dahl, BA, unpublished thesis, Emory University, Atlanta, Georgia).
Recent studies indicate that, in addition to being a critical risk factor for progression of lymphedema 
, ADLA episodes are strongly associated with poor quality of life 
. Thus, reducing ADLA frequency is arguably the most important objective of lymphedema management in resource-poor countries. Accordingly, ADLA has emerged as a key indicator for monitoring lymphedema management programs in filariasis-endemic areas 
In contrast, the usefulness of leg volume as an indicator of clinical improvement in filariasis-endemic areas seems limited. Leg volume, albeit an “objective” measure, varied widely with use of compression garments and, to an extent, with time of day (data not shown). Overall, leg volume decreased in 78% of our patients, but these reductions were generally modest, and statistically significant only for stage 2 lymphedema. Our initial focus on reducing leg volume, although popular with patients, distracted attention from hygiene and skin care; as a result, no significant reduction in ADLA incidence was observed during Phase I of the study. Compression bandages reduced leg volume quickly, but their use was associated with an increase in ADLA incidence. It was difficult to keep the compression bandages clean, and they were prohibitively expensive – equivalent to the annual income of many patients. Frequent visits to the clinic for reapplication of compression bandages in the initial stages of treatment also fostered dependence on clinic staff and undermined key messages of self-reliance and self-care. Thus, although reduction in leg volume is a desirable outcome, our experience suggests that it should not be the primary goal of lymphedema treatment in resource-poor settings.
After receiving training from Dr. Dreyer midway through the study, the clinic staff was better equipped to deliver simple, clear, assertive messages to patients regarding hygiene and skin care. These messages were later reinforced by a simple booklet that was given to each patient, opportunities to participate in support groups, and a radio drama about a young woman with lymphedema that was aired repeatedly on local radio. With these interventions, ADLA incidence dropped dramatically and remained low. The relative contribution of each component of the overall intervention is unclear; we believe that they acted synergistically. Coreil and colleagues showed that patients in Leogane who regularly participated in support groups had a lower incidence of ADLA than those who did not 
. In addition, self-efficacy (belief in one's ability to perform lymphedema self-care) increased significantly following distribution of booklets and broadcasts of the radio dramas (Wendt, JM, unpublished thesis, Emory University, Atlanta, GA).
The number and duration of patient contacts required for patients to understand, become proficient in, and fully committed to life-long lymphedema self-care is not clear. In practice, the intensity of patient training and follow-up varies among filariasis programs 
. Our results indicate that patients who were illiterate and those with advanced disease continued to be at risk of ADLA episodes. These results suggest that more intensive effort may be needed for such patients, such as home visits and support groups. The fact that illiterate persons remained at a 2-fold risk of ADLA throughout the study, even when controlling for other factors, also suggests that illiteracy may be a marker for other factors related to ADLA incidence, such as lower self-efficacy or increased risk of skin lesions that predispose to ADLA 
In addition to reduced ADLA incidence, we observed other improvements in quality of life among patients in this study. These were often dramatic. Patients commonly reported decreased stigma and shame and a return to active life, including school or work [30, Wendt JM, unpublished]
. Those with advanced lymphedema and multiple skin folds and lesions reported, often for the first time in years, the elimination of offensive odor. Serial biopsies in a subset of patients confirmed a decrease in chronic inflammation in lymphedematous legs 
Despite these remarkable changes at the tissue level, most patients did not experience reduction in lymphedema stage. The 4-stage classification that we used is not particularly sensitive to localized changes, and considerable heterogeneity can be found within each stage. Most of the lymphedema classification and staging systems that have been proposed suffer from these limitations 
Studies have reported that the incidence of adenolymphangitis (presumably ADLA or ADLA-like) may decrease following mass treatment with antifilarial drugs 
. The current study was completed almost two years before mass treatment was initiated in Leogane 
; diethylcarbamazine was not readily available. Therefore, it seems unlikely that the clinical improvement and decreased ADLA frequency that we observed can be attributed to reduced transmission of W. bancrofti
Seasonal fluctuations in ADLA, associated with rainfall, have been reported in other studies 
. We observed no consistent seasonal trends in ADLA incidence. The tendency for ADLA incidence to increase between the third and fourth quarters was statistically significant only for 1997. The reason for this is not clear. Rainfall patterns varied from year to year in Haiti during the study period, and the fall of 1997 was relatively dry.
Prospective ADLA monitoring, as was done in this study, is accurate, but expensive. The accuracy of patient recall over longer periods is unknown, but the pain and suffering associated with ADLA make it a memorable event. Among the 127 patients who entered the study during Phase I, the observed incidence of ADLA during Phase I was 1.6 episodes per person-year, compared to 2.4 self-reported for the 12 months before entering the study. It is risky to compare retrospective and prospective data, and it also is possible that messages regarding hygiene and skin care during Phase I, albeit sub-optimal, might have reduced ADLA incidence from baseline levels. However, the data suggest that 12-month recall may be adequate for rapid epidemiologic assessments or program monitoring. Additional research is warranted to assess reliability of 12-month ADLA recall in different filariasis-endemic areas.
The preponderance of women in our study sample reflects the gender distribution among persons with lymphedema in the Leogane population 
. In many areas where bancroftian filariasis is endemic, lymphedema of the leg is more common in women than in men, although this finding is not universal 
This study has several limitations. First, both the incidence of ADLA and compliance with lymphedema self-care were assessed by self-report – a method that, typically, is less than completely reliable. Recall of ADLA episodes during the year before entering the study was facilitated by careful questioning and using memorable public events to help “frame” the previous 12 months. During the study, patients were encouraged to report ADLA and compliance accurately and honestly, and they were assured that clinical care would continue, free of charge, regardless of what they reported. Some 58% of ADLA episodes reported by patients were observed and confirmed by research staff at the clinic. In many of the remaining cases, patients were seen within 1–2 weeks after the episode, when clinical signs (such as peeling of the skin, warmth, or edema 
) were still visible. Skin condition and interdigital lesions were carefully assessed during routine clinic visits, and they provided a check on self-reported compliance. Ad-hoc home visits also encouraged accurate self-reporting. For example, the absence of soap, a clean towel, or an appropriate washing basin would call into question high levels of self-reported compliance.
Second, because of ethical concerns about withholding treatment, we did not collect comparative prospective data on a control group of patients. Further, although we desired to randomize patients with regard to compressive bandaging, we were unable to do so. However, we were able to compare ADLA incidence as a function of the focus of our programmatic intervention (i.e., leg volume during Phase I and ADLA incidence during Phase II).
Third, because of the conditions under which this study was conducted, it cannot be regarded as a study of the efficacy of lymphedema management. We had limited on-site access to specialists in lymphedema management, our staff was comprised mostly of motivated young people rather than health professionals, and the intervention evolved as we gained experience. It is likely that even greater reductions in ADLA frequency would be observed under more optimal conditions.
Despite these limitations, our results suggest that basic lymphedema management is both feasible and effective in filariasis-endemic areas where resources are limited. When provided with the skills and motivation to practice lymphedema self-care, most patients will do so. However, simple messages that unequivocally emphasize the importance of hygiene and skin care are important. Our experience also suggests that compressive bandages, although useful for individual patients, should not be the mainstay of treatment in filariasis-endemic areas, and that volume reduction may not be the optimal measure of success.