Evidence for a bacterial aetiology of ADLA in filariasis-endemic areas comes from the distinctive clinical signs and symptoms, isolation of bacteria at the time of the acute episode, and changes in antibody titres between acute and convalescent serum specimens [21-31]. In India, the bacteria most frequently associated with ADLA are Group A Streptococcus. Other bacteria are often found in cultures, including those that are usually regarded as non-pathogenic [22,24,28].

Available evidence indicates that the immune system may amplify or modulate ADLA. The relative infrequency with which bacteria are isolated from patients with ADLA [22,29,32], as well as from persons with cellulitis in areas not endemic for lymphatic filariasis [33,34], suggests a role for inflammatory mediators [33-35], perhaps even in the absence of bacteria.

Little has been published on the antimicrobial sensitivity of bacteria isolated from persons with ADLA in filariasis-endemic areas. Available experience suggests that the organisms most commonly involved are sensitive to penicillin; thus, penicillin is usually recommended for treatment [36,37].

Clinical descriptions of ADLA in filariasis-endemic areas are remarkably similar to those of erysipelas and cellulitis, about which much has been written in the dermatologic literature [38]. Group A Streptococcus is the classical causative organism for erysipelas, and lymphoedema is a well-recognized risk factor for erysipelas and cellulitis in areas not endemic for lymphatic filariasis [35].

During the early 1990s, the Special Programme for Research and Training in Tropical Diseases (TDR) sponsored a series of population-based studies on the incidence of 'acute attacks' among the general population in filariasis-endemic areas. The case definition – localized pain, lymphadenitis and/or lymphangitis and/or cellulitis and local warmth, with or without systemic manifestations of fever, nausea, and vomiting (in some studies, lasting for at least three days) – is consistent with ADLA. In these studies, the overall incidence of ADLA ranged from 33 per 1000 per year to 97 per 1000 per year [39-41]. A study from Papua New Guinea, which found an incidence of 31 attacks per 1000 population per year, included only cases with fever [42]. One study in an area endemic for Brugia malayi, which also included only cases with fever, found an incidence of 371 episodes per 1000 people per year [43]. Taken as a group, these studies indicate that the rate of ADLA is higher in persons with chronic disease, principally lymphoedema. Among patients in filariasis-endemic areas, the mean annual reported incidence of ADLA ranges from 1.5 to more than 7 episodes per patient [21,32,41,44-51] (Table 1).

The duration of ADLA, primarily based on patient self-reporting, ranges from 1 to 16 days [21,32,39-46,48,49,51,52] (Table 1). Recurrent ADLA episodes result in significant short-term disability, and are of much greater concern to patients than is lymphoedema per se [53]. Studies in Ghana indicate that patients with ADLA are incapacitated for 3 of the 5.1 days of ADLA duration [40]; in Tanzania, patients are incapacitated for 3.7 of the 8.6 days of ADLA [39]. In India, total disability from ADLA lasted no more than 3 days in an area with brugian filariasis [43]. However, preliminary data from Haiti [44] and Togo [49] suggest that the number of workdays lost may exceed the duration of the acute ADLA episode itself.

Among persons with lymphoedema, risk factors for ADLA include increasing patient age [39-41], poor hygiene [54], and illiteracy [47]. Gender, lymphoedema severity, and the presence of entry lesions are additional risk factors. Females tend to experience higher rates than males, although exceptions have been noted [41]. The relationship between lymphoedema stage and incidence of ADLA is not consistent among all studies. It is complicated, in part, by the use of different systems to stage lymphoedema. Most studies show a positive association between lymphoedema stage and observed or patient-reported incidence of ADLA [6,13,32,55-58]. However, other studies – all of which relied on patient recall of ADLA incidence – found no such association [39,40,44,45]. Data from Brazil, India, and Guyana indicate that the presence [54] and number [50,59] of interdigital skin lesions are remarkably strong risk factors for ADLA.

The epidemiologic association between ADLA frequency and stage, as well as extensive clinical experience from both filariasis-endemic and non-endemic areas, strongly suggest that ADLA episodes are a major – likely the most important – factor in lymphoedema progression, particularly in filariasis-endemic areas.

As noted above, among persons born and raised in areas endemic for bancroftian filariasis, episodes of AFL, due to death of the adult worm or 4^th-stage larva, are less severe and have less systemic involvement than ADLA. Systemic involvement may be greater in 'immune-naïve' immigrants to endemic areas. Classical AFL was described extensively in US and European soldiers during World War II [11,82-85]. AFL is commonly observed following individual or mass treatment with DEC [86,87], and this is considered evidence of the drug's macrofilaricidal efficacy [88-90].

Treatment of AFL is supportive. Cold compresses, rest, and analgesics are recommended. Treatment with antifilarial drugs during acute inflammatory episodes used to be recommended, but now is not considered indicated [27,36,91].

The degree to which AFL triggers or hastens the development of hydrocele in bancroftian filariasis has been investigated by several authors. Norões and colleagues reported a 22% incidence of acute hydrocele following a single 'scrotal nodule event', whether spontaneous or induced by DEC [92]. Overall, 5% of men with scrotal nodules (adult worm death) developed hydrocele that persisted for 18 months or longer. Similar findings were observed in Haiti following mass treatment with DEC and albendazole [93]. Hussein and colleagues in Egypt found that 14 of 16 infected men developed detectable fluid in the tunica vaginalis cavity after treatment with DEC and albendazole, of whom three developed chronic hydrocele [94]. It is unclear whether the lifetime risk of acute or chronic hydrocele is increased by DEC treatment, or whether the drug merely synchronizes adult worm death and, therefore, resulting hydrocele.

AFL appears to trigger the onset of lymphoedema less frequently than it does hydrocele, and persistent lymphoedema following AFL is unusual in the absence of other co-factors [5].

Costs to patients for lymphoedema treatment, reported as both per-visit and per-year costs, vary greatly by study. A study in India reported an average of US$ 0.56 per visit, more than half a day's wages [62]. A Ghanaian study reported costs for treatment of chronic disease (both lymphoedema and hydrocele) of US$ 0.87 per visit, equivalent to almost one day's wages [61], and greater than costs incurred by controls with other chronic diseases. In India, the annual cost for lymphoedema treatment ranges from US$ 2.17 to US$ 8.70 per person [108,109]. Average treatment costs are often low, in part because many patients who find potential treatment costs prohibitive either self-treat or do not seek treatment [109-111].

Productivity losses from lymphoedema have been captured as lost working hours and as changes in individual output. Lymphoedema patients in India lose 0.55 to 1.61 hours per day in time at work; 11%–31% of workdays are lost annually [66,109]. These findings are similar to those of another study of both lymphoedema and hydrocele patients, which estimated 1.13 hours lost per day, for a total of 19% of workdays lost per year [108]. In Ghana, female labour input loss due to lymphoedema was estimated at 1.5% per year, using the average percentage of lymphoedema patients unable to complete certain activities and the local prevalence of lymphoedema [61]. In general, many patients report changing to less strenuous occupations or giving up working altogether due to lymphoedema and ADLA [69,72,75,112]. A study of male weavers in India with chronic disease, 26% of whom had lymphoedema, found a 27% decrease in output compared to controls [113].

Several studies have quantified the impact of lymphoedema on quality of life using standardized measures [44,69,70,73,78,114,115]. McPherson, using a 30-point Dermatology Quality of Life Index in Guyana, found a mean baseline score for lymphoedema patients of 10.9 (comparable to patients with psoriasis and atopic eczema in the United Kingdom), with controls scoring 0.5 [83]. Six months after starting regular hygiene treatment, the scores improved significantly by an average of 6.8 points [122]. In Haiti, Kanda compared different ways of measuring quality of life among rural people with lymphoedema [46]. Using the EuroQol scale, he found that no respondents had extreme problems in mobility or self-care, but more than half reported pain or discomfort. On a depression scale, the CES-D, these same patients had a mean score of 13.2 (16 and above indicates depression). On the CDC Healthy Days questionnaire, Kanda found that 88% of patients ranked their health as fair or better; however, they also reported an average of 9.9 physically or mentally unhealthy days during the past month. Advanced age, advanced stage of illness, and low educational level were strongly associated with lower quality-of-life measures [44]. In India, patients with lymphoedema scored from 9.2 to 12.4 on a 28-point scale of 'health state severity' using an extended EuroQol measuring system [69]. Severity was associated with stage of lymphoedema; in higher stages, 'severe or very severe problems' were reported for the domains of usual activities, pain, anxiety/depression, cognition and social participation. Among men, the severity score for lymphoedema was significantly higher than that for hydrocele [70].

Studies in India and Ghana show that 46%–100% of persons with lymphoedema sought treatment from health care centres, local healers, or pharmacies during the previous year [61,62,108,109]. The studies in Ghana show that modern medical care often is avoided due to lack of interest from health care workers and a belief by patients that lymphoedema treatment requires spiritual interventions [74,118]. Although many patients believe lymphoedema progression cannot be prevented, they continue to consult spiritualists and treat themselves with herbal preparations or analgesics [75]. In contrast, in areas of India with networks of public healthcare facilities, most patients seek care from modern medical practitioners, although a minority consult Ayurvedic doctors or use home remedies first [72,111]. Access to care is not necessarily universal, however; young women in India may not seek treatment because of social constraints, such as the paucity of female doctors [116]. Other barriers to care include distance to a health facility, lack of awareness, lack of time, lack of child care, perceived severity of disease, and dissatisfaction with previous treatment [74,116,122,123]. Even when patients seek treatment, health personnel often will prescribe antifilarial or other drugs that are expensive and ineffective. Inadequate knowledge of lymphoedema management by health workers results in suboptimal patient care [105,111,123,125,126].

Relatively few studies have documented the effectiveness or impact of the basic package of lymphoedema management, and most of these have focused on ADLA. The available data indicate that such treatment is associated with a marked reduction in incidence of ADLA [30,32,47,55,56,58,78]. An unpublished study from Haiti reported that risk factors for continued ADLA include more advanced disease, 'negligence', and illiteracy [47].

A few studies have documented changes in leg volume or circumference in response to basic lymphoedema management. Although an 'objective' measurement, leg volume can vary considerably with time of day, exercise, elevation, and other factors. In Orissa, India, Kerketta and colleagues reported significant reductions in leg circumference with all treatment regimens that included basic foot care [79]. Pani and colleagues reported greater volume reductions in patients with oedema of recent onset than in those with lymphoedema of longer duration [51]. An unpublished study from Haiti, which initially included compressive bandaging as one of its modalities, reported that more than 65% of 178 patients had a reduction in leg volume after two years when compared with pre-treatment measurements [47].

It is commonly observed that, with basic lymphoedema management, the prevalence and severity of entry lesions decrease [36].

Reduction in offensive odour is commonly observed with regular hygiene. To our knowledge, there have been no studies focusing on reduction in odour as an outcome of lymphoedema treatment in filariasis-endemic areas, although anecdotally this improvement has an important effect on quality of life.

Few studies have attempted to address the degree to which basic lymphoedema management results in regression of lymphoedema stage or grade. In part, this is because most staging systems have not been developed for this purpose. Thus, considerable improvement in skin condition or even leg volume is possible without regression in stage per se.

Improved flexibility and a feeling of 'lightness' are commonly reported by patients, but few studies have documented the effectiveness of basic lymphoedema management on limb range of motion.

Several studies are currently underway that address the extent to which basic lymphoedema management in filariasis-endemic areas improves quality of life. One study, by McPherson in Guyana, documented highly significant improvement in quality of life as measured by the Dermatology Quality of Life Index [78]. Similar work in non-endemic areas has shown substantial gains in quality of life with lymphoedema treatment. Patients who incorporate regular lymphoedema management into their daily routines have reported satisfaction with the results [53,138].

A study in Haiti collected skin punch biopsy specimens from the lymphoedematous legs of 27 patients before and about 12 months after they initiated basic lymphoedema management [139]. Follow-up biopsies showed significant reductions in perivascular mononuclear infiltrate in the superficial dermis (41% decrease in prevalence), in perivascular fibrosis in the deep dermis (58% decrease), and in periadnexal mononuclear infiltrate (53% decrease).

Although there is general agreement as to the basic elements of lymphoedema management within the GPELF, considerable regional variation exists in the availability of supplies, including soap, water, and topical skin preparations (e.g. antiseptics, antifungal and antibacterial agents). These differences contribute to variation in approaches used in different regions. For example, in some countries, macerated interdigital lesions are treated with Whitfield ointment, an inexpensive antifungal agent, on the presumption that dermatophytes are the primary pathogen. In Guyana, McPherson and colleagues attempted to culture fungi from these lesions and concluded that bacteria probably play a more important role than fungi [50]. McPherson's observations are consistent with studies of intertriginous lesions in non-endemic areas [140,141].

Controlled studies of how best to optimize the effectiveness of treatment, particularly for skin care, have not been published. Some investigators have argued for more widespread adoption of breathing exercises to mobilize lymph fluid, and for emollients to protect and rebuild the skin barrier function [133]. These are issues that are amenable to basic, inexpensive clinical trials.

Although there remains some debate about the optimal package of interventions for basic lymphoedema management in filariasis-endemic areas, the benefits of such treatment are generally recognized, and foci of activity in several countries have demonstrated success. However, relatively few persons with lymphoedema living in filariasis-endemic areas currently have access to treatment. Thus, the key programmatic issue is how best to 'scale up' basic lymphoedema management to state and national levels. The challenges can be considered in four major categories:

• Finding patients and bringing them to treatment (many are reluctant to seek care as discussed above)

• Education of patients and family members on the principles and practice of lymphoedema self-care

• Encouragement and support to sustain daily self-care (this support may include improved access to supplies such as clean water, soap, antiseptics, topical antibacterial and antifungal agents, and oral antibiotics)

• Referral networks for management of ADLA and for patients with advanced lymphoedema or lymphoedema complicated by other diseases.

There is general agreement that most patients can manage their lymphoedema routinely at home, and that this is preferable and less costly than clinic-based care. WHO has developed training packages for 'informal caregivers' to instruct patients on home-based care, and this approach has been adopted by most programmes. However, numerous key programmatic and operational research questions remain unanswered for each of the four major programme components. For example: 1) although McPherson and colleagues have shown that health workers in Guyana with limited training can reliably stage lymphoedema and identify entry lesions [142], the ability of such workers to recognize or diagnose lymphoedema in other settings is unknown; 2) the frequency and intensity of education required for patients to become competent in lymphoedema self-care has not been evaluated; and 3) basic requirements for referral care, provider training, and clinical competency have not been determined. The costs of treatment need to be better understood, as well as the benefits. These are areas in urgent need of investigation if the benefits of lymphoedema management are to reach those who most need it.

Considerable anecdotal evidence suggests that the onset of chronic lymphoedema is triggered by the first or second episode of ADLA. Data from a filariasis-endemic area of Haiti indicate that skin lesions between the toes, which could provide portals of entry for bacteria, are common in children, and are significantly more common in those who test positive for circulating filarial antigenaemia [143]. Similar findings have been observed in northeast Brazil (G. Dreyer, personal communication). The degree to which initial ADLA episodes, and therefore lymphoedema, can be prevented through school-based education programmes focused on hygiene, skin care, and recognition and treatment of entry lesions has not been studied.

Patient expenditures for hydrocele treatment are generally low, as treatment other than surgery is found to be ineffective and most patients cannot afford to pay for surgery. Hydrocele patients in India paid from US$ 1.38 to US$ 4.29 per year for non-surgical treatment; daily wages in the areas studied averaged less than US$ 1.00 [62,108,109]. A Ghanaian study found an average of US$ 0.87 a year (almost one day's wages) spent for treatment of chronic filariasis, which included both hydrocele and lymphoedema – significantly more than was spent by patients with other chronic diseases [61]. In general, treatment costs are difficult to collect accurately as much of treatment is paid in-kind or provided by traditional healers who are members of the extended family.

Published costs to patients for hydrocele surgery range from US$ 5 to US$ 60, depending on the country and source of care. The types of surgery performed and the parameters of costing are not known for all studies. Ramaiah reported costs of US$ 5–14 in government hospitals and US$ 15–47 in private hospitals in India [109], while Babu reported costs of US$ 44 in another Indian study [108]. In Ghana, Gyapong reported surgery costs of US$ 30–35 at local hospitals [61,151] and Ahorlu reported surgery costs of US$ 30–60 for surgery sponsored by non-governmental organizations (NGOs) [152]. Interestingly, the patients in Ahorlu's study estimated that surgery would have cost US$ 75–125 at local hospitals before the NGO programme was put in place. Ahorlu also reported other costs associated with surgery, including transport to hospital and food, estimated by patients at US$ 20–30, with an average hospital stay of 4–12 days.

Early studies on hydrocele differed in their conclusion about the impact on productivity, and reductions in productivity were not quantified [65,117,129,153]. In recent studies, the effect of hydrocele on productivity has been quantified in three different ways:

• Individual working hours. Studies in India have shown that hydrocele patients work approximately one hour less per day than matched controls [66,108,109]. Lu et al. in the Philippines found that 30% of 22 males interviewed lost time from work due to hydrocele [154].

• Individual output. A study of weavers with chronic filarial disease in India, 69% of whom had hydrocele, showed that those with disease produced 27% less cloth than matched controls [113].

• National output. In India, 8% of potential male labour input was estimated lost due to hydrocele and lymphoedema [109] and this loss was valued at US$ 704 million per year [67]. In Ghana this figure was similar, with more than 7% of potential labour lost [61].

There is almost no evidence on the degree to which hydrocelectomy improves productivity, with only one study reporting qualitative data [152].

Early studies described the socially unacceptable nature of hydrocele, but they were vague about the degree of associated stigma, its consistency across communities and cultures, and the psychosocial burden of hydrocele on those affected [65,117,122,155]. To date, research has not been carried out on quality of life in men with hydrocele in filariasis-endemic areas that would allow for comparison with other diseases, or with men who do not have hydrocele.

A wide range (25%–80%) of hydrocele patients seek treatment [61,62,109,151]. Patients seek treatment from local health centres, traditional healers, and through self-medication. In certain countries, the belief that hydrocele has a supernatural cause leads people to seek out traditional healers or sorcerers instead of modern medical care [74]. While most studies describe treatment only during the previous year, patients may have tried various remedies in the past but stopped seeking care after the treatments were ineffective [74,122,151]. Other reasons for not seeking treatment include problems with access, such as cost of surgery or medical treatment, distance from the health centre, inability to take time off work for recovery, and cultural issues such as fear of anaesthesia during surgery and stigma associated with having hydrocele [61,117,152,154,155]. A study in coastal Kenya found that, in highly endemic districts, hydrocelectomies accounted for 23% of all major operations [157]. Similarly, in Tanzania in 1976, 15% of operations in one district hospital were for hydrocele [158].

Ahorlu et al. interviewed hydrocele patients in Ghana 1.5 years after surgery. Patients reported that within three to six months post-surgery, they had experienced significant improvement in self esteem, sexual function, and capacity for work, and they participated more in community activities [152].