Portal de Periódicos da CAPES

Epidemiology and control of trachoma: systematic review

2010; Wiley; Volume: 15; Issue: 6 Linguagem: Inglês

10.1111/j.1365-3156.2010.02521.x

1365-3156

Victor H. Hu, Emma M. Harding‐Esch, Matthew J. Burton, Robin L. Bailey, Julbert Kadimpeul, David Mabey,

Urinary Tract Infections Management

Trachoma is the commonest infectious cause of blindness. Recurrent episodes of infection with serovars A–C of Chlamydia trachomatis cause conjunctival inflammation in children who go on to develop scarring and blindness as adults. It was estimated that in 2002 at least 1.3 million people were blind from trachoma, and currently 40 million people are thought to have active disease and 8.2 million to have trichiasis. The disease is largely found in poor, rural communities in developing countries, particularly in sub-Saharan Africa. The WHO promotes trachoma control through a multifaceted approach involving surgery, mass antibiotic distribution, encouraging facial cleanliness and environmental improvements. This has been associated with significant reductions in the prevalence of active disease over the past 20 years, but there remain a large number of people with trichiasis who are at risk of blindness. Epidémiologie et lutte contre le trachome: Revue systématique Le trachome est la plus fréquente cause infectieuse de cécité. Des épisodes récurrents d’infection par Chlamydia trachomatis sérovars A-C provoquent une inflammation de la conjonctive chez les enfants qui plus tard développent des cicatrices et la cécitéà l’âge adulte. Il a été estimé qu’en 2002, au moins 1,3 millions de personnes étaient aveugles à cause du trachome et actuellement 40 millions de personnes souffriraient de la maladie active dont 8,2 millions avec trichiasis. La maladie sévit en grande partie dans les communautés rurales pauvres dans les pays en développement, particulièrement en Afrique sub-saharienne. L’OMS encourage la lutte contre le trachome à travers une approche à multiples facettes comprenant la chirurgie, la distribution massive d’antibiotiques, la promotion de la propreté du visage et des améliorations de l’environnement. Cela a été associéà une réduction significative de la prévalence de la maladie active au cours des 20 dernières années, mais il reste un grand nombre de personnes atteintes de trichiasis qui sont à risque de cécité. Epidemiología y Control del Tracoma: Revisión sistemática El tracoma es la causa infecciosa más común de ceguera. Los episodios recurrentes de infección con los serovares A-C de Chlamydia trachomatis causan la inflamación de la conjuntiva en niños que posteriormente desarrollan cicatrices y ceguera como adultos. Se ha estimado que en el 2002 al menos 1.3 millones de personas quedaron ciegas por tracoma, y actualmente se calcula que 40 millones de personas sufren una infección activa y 8.2 millones padecen triquiasis. La enfermedad es particularmente común entre los pobres, en comunidades rurales de países en vías de desarrollo, particularmente África sub-Sahariana. La OMS promueve el control del tracoma mediante un enfoque multifacético que incluye cirugía, distribución masiva de antibióticos, educación para la limpieza facial y mejoras medioambientales. Esto se ha asociado con una reducción significativa en la prevalencia de la enfermedad activa durante los últimos 20 años, pero aún continua habiendo un gran número de personas con triquiasis que están en riesgo de ceguera. Trachoma is the leading infectious cause of blindness worldwide. It is caused by infection with Chlamydia trachomatis and is characterised by inflammatory changes in the conjunctiva in children with subsequent scarring, corneal opacity and blindness in adults. The World Health Organization (WHO) estimated in 2002 that 1.3 million people were blind from trachoma (Resnikoff et al. 2004) and it is likely that a further 1.8 million were suffering from low vision (Frick et al. 2003a). Many of the additional 1.9 million cases of blindness from ‘corneal opacities’ were also likely to be because of trachoma in areas where it is endemic (Resnikoff et al. 2004). The number of people with active disease is estimated to be 40 million, and the number with trichiasis, 8.2 million (Mariotti et al. 2009). Trachoma is an ancient disease and has previously been a significant public health problem in many areas of the world including parts of Europe and North America. Today, however, trachoma is largely found in poor, rural communities in low-income countries, particularly in sub-Saharan Africa. In 1998, the WHO established the Alliance for the Global Elimination of Blinding Trachoma by 2020 (GET2020). This promotes trachoma control through the SAFE Strategy: surgery for trichiasis, antibiotics for C. trachomatis infection, facial cleanliness and environmental improvement. Where control measures have been implemented encouraging reductions in the prevalence of trachoma have been found. The earliest references to trachoma come from China in the 27th century BC (Al-Rifai 1988). Features of trachoma were also described in the Ebers papyrus from Egypt, 15th century BC, and epilation forceps discovered in tombs from the 19th century BC (Maccallan 1931, Hirschberg 1982). Trachoma became a major public health problem in Europe at the beginning of the 19th century, when the disease was believed to have been brought back by troops returning from the Napoleonic wars in Egypt. So great was the burden of the disease at that time that many of the major ophthalmic hospitals founded in the 19th century were established to treat trachoma, including Moorfields Eye Hospital and Massachusetts Eye and Ear Infirmary. By the end of the 19th century, immigrants to the United States were routinely screened for trachoma and sent home if they had signs of the disease. Trachoma has now disappeared from developed countries (with the exception of Aboriginal communities in outback Australia (Tellis et al. 2007), probably as a result of general improvements in living and hygiene standards. Trachoma is a chronic keratoconjunctivitis caused by recurrent infection with serovars A, B, Ba and C of C. trachomatis. Infection is most commonly found in children. With repeated reinfection, some people go on to develop scarring complications and blindness in later life. The clinical manifestations of trachoma are subdivided into those associated with ‘active’ disease, usually seen in childhood, and those associated the cicatricial or scarring complications, seen in late childhood and adults (Figure 1). Active disease is characterised by recurrent episodes of chronic, follicular conjunctivitis. Follicles are subepithelial collections of lymphoid cells and appear as small, yellow-white elevations on the conjunctiva of the everted upper lid. Papillary hypertrophy (engorgement of small vessels with surrounding oedema) also occurs and can obscure the deep tarsal vessels if severe enough. Vascular infiltration of the upper cornea (pannus) may also develop in active disease, but this rarely affects vision. Individuals are frequently asymptomatic or have only mild symptoms even if marked signs of inflammation are evident. If present, symptoms are similar to those associated with any chronic conjunctivitis: redness, discomfort, tearing, photophobia and scant muco-purulent discharge. Conjunctival follicles at the upper margin of the cornea leave shallow depressions after they resolve known as ‘Herbert’s pits’ which, unlike follicles and papillae, are a pathognomonic sign of trachoma. Clinical features of trachoma. (a) Active trachoma in a child, characterised by a mixed papillary (TI) and follicular response (TF). (b) Tarsal conjunctival scarring (TS). (c) Entropion and trichiasis (TT). (d) Blinding corneal opacification (CO) with entropion and trichiasis (TT). Repeated and prolonged episodes of infection and inflammation can result in the scarring complications of trachoma. Initially, conjunctival scarring is seen in the subtarsal conjunctiva, which can range from a few linear or stellate scars to thick, distorting bands of fibrosis. Contraction of this scar tissue causes entropion (in-turning of the eyelids) and trichiasis (eyelashes touching the eyeball) which is often painful. Eventually, corneal opacification develops the blinding end-stage of the disease. This is probably a result of multiple insults to the cornea: mechanical trauma from lashes, secondary bacterial or fungal infection and a dry ocular surface. Over the years, various grading systems for trachoma have been proposed. The one which is currently used by trachoma control programmes is the 1987 WHO simplified grading system (Table 1) (Thylefors et al. 1987). The prevalence of active disease is highest in pre-school children and declines to low levels in adulthood (Dawson et al. 1976; West et al. 1991b; Dolin et al. 1998). This parallels the distribution of C. trachomatis infection, with up to half of the community bacterial load being found in children under the age of 1 year in some studies (Solomon et al. 2003; Melese et al. 2004b). Adult bacterial loads are usually lower than those of children, and the duration of infection and disease also declines with age, presumably as the result of an acquired immune response (Bailey et al. 1999; Grassly et al. 2008). This is in contrast to the scarring features of trachoma, the prevalence of which increase with age, reflecting the cumulative nature of the damage. Where the prevalence of active disease is very high, cicatricial complications may be seen at an early age; trichiasis was reported in 2–3% of children under the age of 15 years in southern Sudan where the prevalence of active disease was 70–80% (Ngondi et al. 2006a; King et al. 2008). Cohort studies in trachoma-endemic communities in The Gambia and Tanzania have looked at the progression of the scarring process: Worsening of conjunctival scarring was seen in nearly 50% of scarred subjects over 5 years (Tanzania) (Wolle et al. 2009). Progression from conjunctival scarring to trichiasis was seen in 10% after 7 years and 6% after 12 years (Tanzania and The Gambia) (Munoz et al. 1999; Bowman et al. 2001). Minor trichiasis (<5 lashes touching the eye) progressed to major trichiasis (five or more lashes touching the eye) in 33% after 1 year and in 37% after 4 years; and unilateral progressed to bilateral trichiasis in 46% after 1 year (The Gambia) (Bowman et al. 2002b; Burton et al. 2006). Trichiasis is associated with the development of corneal scarring: 8% of people with trichiasis developed incident corneal scarring after 4 years, and there was worsening of established corneal scarring in 34% after 1 year (The Gambia) (Bowman et al. 2002b; Burton et al. 2006). The first study from Tanzania had a standardised, prospective design but the others did not. There is considerable variation in the reported rates of progression, which may reflect both variation in progression rates in different populations and methodology. A key determinant of the rate of disease progression is probably the burden of C. trachomatis infection in a community over time, although the direct evidence for this is limited. Several studies found that the risk of developing scarring complications is greater in those with recurrent or persistent severe inflammatory trachoma (Dawson et al. 1990; Munoz et al. 1999; West et al. 2001; Burton et al. 2006). There is little doubt that C. trachomatis is the cause of trachoma; Koch’s postulates were largely fulfilled shortly after the first isolation of C. trachomatis in 1957 (Tang et al. 1957; Collier et al. 1958). However, C. trachomatis cannot be detected in all cases of active disease, even using highly sensitive nucleic acid amplification tests (NAAT) (Baral et al. 1999; Lietman et al. 2000; Burton et al. 2003; Miller et al. 2004b). In low prevalence communities, especially those that have received mass antibiotic treatment, C. trachomatis is only found in a minority of those with active disease. Those with intense trachomatous inflammation are more likely to be infected and have higher bacterial loads than those with follicular disease (Burton et al. 2003; Solomon et al. 2004b; Wright & Taylor 2005). In endemic communities infection is sometimes detected in those who do not fulfil the WHO criteria for active disease. Part of the explanation for this poor correlation is likely to be the kinetics of the disease with a short latent phase (infection before clinical signs with the incubation period for disease), a patent phase (infection and clinical signs) and a recovery phase (infection cleared but clinical signs persist, which can last for many months) (Bailey et al. 1994; Wright et al. 2008). The mismatch between the presence of infection and clinical findings is also partly explained by use of the simplified WHO grading system, which excludes those with fewer than five follicles in the subtarsal conjunctiva (Ward et al. 1990). Chlamydia trachomatis is probably transmitted between individuals by a variety of mechanisms, including: Direct spread from eye to eye during close contact such as during play or sleep. Spread of infected ocular or nasal secretions on fingers. Indirect spread by fomites such as infected face-cloths. Transmission by eye-seeking flies. Possible spread from nasopharyngeal infection by aerosol. A combination of these and other transmission mechanisms probably operates in most environments, although their relative importance may vary. For example, in some environments eye-seeking flies probably contribute to the transmission of infection. Chlamydia trachomatis has been detected by polymerase chain reaction in around 20% of Musca sorbens caught on the faces of children in Ethiopia (Jones 1975; Miller et al. 2004a; Lee et al. 2007) and intervention trials to reduce fly density have been associated with a reduction in active trachoma in The Gambia (Emerson et al. 1999, 2004). However, in other locations, the density of eye-seeking flies is insignificant and does not appear to contribute towards transmission (Taylor et al. 1985). Genital strains of C. trachomatis do not cause endemic trachoma, although occasionally they cause a self-limiting conjunctivitis (Brunham et al. 1990). Trachoma is a focal disease and has been found to cluster at the level of the community, the household and within bedrooms, reflecting the infectious nature of the disease and suggesting that prolonged intimate contact is necessary for the transmission of infection (Dawson et al. 1976; Katz et al. 1988; Bailey et al. 1989; West et al. 1991b; Burton et al. 2003). This is particularly important for trachoma control programmes, as it significantly increases the sample size necessary for estimating the prevalence within a region (Katz et al. 1988). Most transmission events occur within the household, and a failure to treat all infected household members during mass antibiotic distribution may result in rapid re-infection of that family followed by more gradual spread across the community (Blake et al. 2009). No non-human reservoir of infection has been found, with flies only acting as passive vectors. The importance of extra-ocular sites of infection has been debated. Chlamydia trachomatis can be detected in secretions from the nasopharynx, and a recent study also showed that infected nasal discharge in children at baseline was associated with an increased risk of active disease and conjunctival infection 2 months after systemic treatment (Malaty et al. 1981; West et al. 1993; Gower et al. 2006). However, nasal swabs were taken only from children with visible discharge and were of the discharge rather than from nasal epithelium. Positive results may simply have been a reflection of severe ocular infection which was not cleared with one dose of antibiotic, with infected secretions passing through the nasolacrimal ducts. An earlier study using nasal swabs on all children showed that new ocular infection after treatment was not related to a positive or negative nasal specimen at baseline (West et al. 1993). In addition, genotyping of conjunctival and nasal samples from individuals with concurrent infection showed different genotypes to be present, suggesting that auto-infection was not an important factor (Andreasen et al. 2008). Trachoma is a major cause of blindness in many less-developed countries, especially in poor, rural areas. Blinding trachoma is believed to be endemic in over 50 countries, with the highest prevalence of active disease and trichiasis in Africa, predominantly in the savannah areas of East and Central Africa and the Sahel of West Africa (Figure 2). It is also endemic in a number of countries in the Middle East, Asia, Latin America and the Western Pacific (Polack et al. 2005). Current WHO estimates for the prevalence of active disease, trichiasis and blindness are significantly lower than previous ones and declines in the prevalence have been noted in several countries, but there is considerable uncertainty around these estimates, as little recent information is available from India and China. Map of trachoma endemic countries in 2009. Reproduced with permission from Dr Silvio P. Mariotti, WHO/NMH/. About half of the global burden of active trachoma is concentrated in five countries: Ethiopia, India, Nigeria, Sudan and Guinea; while half of the global burden of trichiasis is concentrated in three countries: China, Ethiopia and Sudan (Mariotti et al. 2009). Recent studies from southern Sudan, previously inaccessible during the civil war, have shown very high levels of trachoma: up to 80% of children had active disease and one-fifth of adults had trichiasis (Ngondi et al. 2006a; King et al. 2008). Trachoma was shown to account for 35% of blindness, with 5% of the entire population (including children) suffering from low vision or blindness associated with trachoma (Ngondi et al. 2006b, 2007). Some caution is required in the interpretation of global estimates of trachoma prevalence (Burton & Mabey 2009). These have generally been produced with models that have relied on the results of a limited number of surveys conducted in a few endemic countries. Various assumptions and extrapolations are then made, which have considerable potential for error, such as extrapolating data from a single survey within a district to give the district-level prevalence, and national averages being generated from available district prevalence data. The six million people estimated by the WHO to be blind from trachoma in the 1990s was probably a substantial overestimate as results were based on questionnaires reporting numbers of people who might become blind without treatment (Thylefors et al. 1995). More recent estimates have used more reliable survey data. Notwithstanding the aforementioned limitations of the available data, there does appear to be a downward trend in the number of people affected by trachoma. Improved living standards in many countries probably account for at least part of this trend, as was the case with the disappearance of trachoma from industrialised countries a century ago (Dolin et al. 1997; Hoechsmann et al. 2001). The establishment of trachoma control programmes has probably played a major role, although this is difficult to quantify. Worryingly, the number of people estimated to have trichiasis has shown little decline since 1991, with a slight increase estimated between 2003 and 2008. This suggests that progressive conjunctival scarring can occur even when there has been a marked reduction in active disease and C. trachomatis infection, which has long-term implications for control programmes. The most recent estimate from the WHO places the burden of trachoma at 1.3 million disability-adjusted life years. This measures the gap between a normal, healthy population and the ‘cost’ of a disease from premature mortality and disability (WHO 2008). The economic cost of trachoma has been estimated at between US$ 3 billion – 8 billion in lost productivity (Frick et al. 2003a,b). Estimates of the global burden of trachoma, however, are faced with several problems including a lack of robust prevalence data and the decision over inclusion of different disease manifestations (Burton & Mabey 2009). Trichiasis without visual impairment, for example, causes a level of disability comparable to that caused by visual impairment from non-trachomatous causes, yet it has not always been included in disease burden calculations (Frick et al. 2001b). Many studies have examined potential risk factors for trachoma, which have been previously reviewed (Emerson et al. 2000; West 2004; Haylor 2008). Studies examining the relationship between trachoma and various environmental, socio-economic and behavioural factors are difficult to interpret as they often lack adequate controls and are potentially confounded with many factors being closely interrelated. For example, establishing what contribution a dirty face makes to trachoma, or vice versa, is difficult, as active disease may cause ocular/nasal discharge, but discharge may be an important route for transmission. In addition, variability in survey methodology and questionnaires may not allow reliable comparisons between studies (Emerson et al. 2000). Trachoma is currently more common in dry areas, and the relationship between water and trachoma has been studied in several settings, with some conflicting results. It is plausible that better access to water would improve hygiene levels and reduce the transmission of infection. Several studies have indeed found an association between increased distance to water and the prevalence of active disease (Mathur & Sharma 1970; Tielsch et al. 1988; Taylor et al. 1989; West et al. 1989; Schemann et al. 2002). However, other studies have not supported this and the association appears to be absent when the distance to water is small (West et al. 1991b; Zerihun 1997; Kuper et al. 2003). This may be explained by the presence of a ‘water use plateau’ in which per capita water consumption between households often seems to be constant when the round trip to collect water is below a threshold of around 30 min (Cairncross & Feachem 1993). The quantity of water brought into a household may be more important than the distance to water. Indeed, one study found the quantity to be independent of distance and that children from households with a greater quantity of water had less active disease (Kupka et al. 1968). However, other studies have shown that after controlling for distance the total quantity of water used had no effect on the prevalence of disease (West et al. 1989; Bailey et al. 1991). The second of these two studies may unlock the key issue with regard to water and trachoma: the authors actually measured how much water was brought into the house and also observed how the water was used. After controlling for family size, distance to water and other socio-economic factors, families with trachoma used less water for washing children than did control families without trachoma, regardless of the amount of water available for consumption (Bailey et al. 1991). The association between frequent face washing and reduced trachoma has been reported in some, but not all, studies (Taylor et al. 1985; Tielsch et al. 1988; Bailey et al. 1991; Luna et al. 1992). Self-reporting may have compromised the results, as washing may be perceived as a desirable activity and hence over-reported. A large-scale randomised trial of an intensive educational intervention to encourage face washing in Tanzania showed that children with a clean face were less likely to have severe inflammatory trachoma (TI). However, there was no reduction in the overall prevalence of active trachoma and intensive behavioural intervention was required (West et al. 1991a, 1995; Schemann et al. 2002). As discussed previously, flies are also a risk factor for trachoma by facilitating transmission. M. sorbens, the fly most commonly found in contact with eyes, preferentially breeds in human faeces. Latrine access is associated with a lower risk of trachoma. This has been attributed to the removal of faecal material from the environment leading to a smaller fly population (Emerson et al. 2004). Crowding is probably a risk factor for trachoma, especially living in close proximity to children with active disease (Bailey et al. 1989; Sahlu & Larson 1992). Women tend to have a higher rate of the scarring complications of trachoma and this is generally considered to be a result of their increased contact with young children, the main reservoir of infection (Turner et al. 1993). Migration between communities may also be important in the re-introduction of C. trachomatis (Burton et al. 2005b). Trachoma as a public health problem is defined by the WHO as a prevalence of TF of at least 10% in children aged 1–9, or a prevalence of TT of at least 1% in those aged 15 or more. Trachoma is no longer considered a public health problem when the TF prevalence in children falls below 5% and the prevalence of TT is <0.1% (WHO, 1997; Kuper et al. 2003). No specific guidelines are provided for areas where the prevalence falls between these thresholds. To determine where trachoma is a public health problem, WHO recommends cluster random sampling (Ngondi et al. 2009b). Districts likely to be trachoma-endemic are identified using information from previous surveys, written reports, hospital eye surgery records and interviewing people with local experience. A list of all clusters within the districts identified is made. Clusters are preferably areas of approximately the same population size, so that the cluster selection is with probability of selection proportional to size. A random sample of clusters is then selected, which is sufficiently large such that the sample prevalence of TF in 1–9 year olds, or TT in those aged 15 or more, reflects the prevalence in the whole population (WHO, 2006). A two-stage design can be employed, whereby villages (clusters) are selected in the first stage, and households are selected in the second. If household lists are not available, other methods for selecting households are by random walk and compact segment sampling. Reports should present standardisation of the examiners’ grading, the sample size parameters, confidence intervals of the estimate, and adjustment for clustering (Ngondi et al. 2009b). As well as obtaining accurate estimates of TF and TT prevalence, surveys should collect data on the number of public access and surface water points in the district, and the proportion of households that have access to latrines and that are within 15 min walk of the nearest water source available during the dry season. These data allow planning, monitoring and evaluation of control interventions (WHO, 2006). Population-based prevalence surveys provide comprehensive prevalence data and are rightly considered the ‘gold standard’ for trachoma surveys (Wright et al. 2005). Although they can be designed to provide precise prevalence estimates over wide areas, they generally do not give accurate estimates at the cluster level, and the sampling needs to incorporate large design effects (four or more) arising from the focal nature of active trachoma and use large numbers of clusters if they are not to overlook hyperendemic clusters of disease (WHO-ITI, 2004). Moreover, they are time consuming and expensive because of the large sample sizes needed. Two alternative methods have been proposed: trachoma rapid assessment (TRA) and acceptance sampling TRA (ASTRA). Trachoma rapid assessment was designed to allow simple, fast and cost-effective assessment of active disease, trichiasis and environmental risk factors. Existing data are first used to identify areas that are likely to be trachoma-endemic. The burden of trichiasis, active disease and associated risk factors is then assessed in these areas (Negrel & Mariotti 1999). At least three, but no more than seven, villages are selected per district, with priority given to those areas ‘deemed most socio-economically disadvantaged’ (Wright et al. 2005). In these communities, individuals with TT are identified, leading to a crude estimate of TT prevalence. Fifty children aged 1–9 from at least 15 households that ‘appear to have the lowest socio-economic status’ are then assessed for TF and/or TI. Finally, a survey is performed to determine household level trachoma risk factors. Trachoma rapid assessment provides rankings rather than prevalence estimates, and the method of selection of areas, communities and households outlined previously will generally be subjective. This may lead to overestimated and/or inconsistent prevalence data, with the possible extrapolation of biased data to the whole village and district (Negrel et al. 2001; Myatt et al. 2003; Solomon et al. 2004b). Evaluations of TRA rankings in comparison to PBPS in Tanzania and China found comparable ranking of communities, but TRA performed worse in low prevalence settings (Paxton 2001; Liu et al. 2002). However, PBPS does not itself provide reliable estimates or rankings for individual clusters, so these comparisons are flawed (Ngondi et al. 2009b). In The Gambia, a study comparing two TRA surveys found that active disease prevalence estimates and rankings were inconsistent, indicating that it is not a reliable method (Limburg et al. 2001). Acceptance sampling TRA, based on the principle of sequential sampling methods, such as lot quality assurance sampling (LQAS), has been proposed as an alternative to TRA. A maximum sample size and an acceptable number of TF cases are set and sampling stops when one of these is met. Villages are classified as high prevalence if sampling is stopped because the set number of TF cases was exceeded, or as low prevalence where sampling is stopped because the maximum sample size was reached (Myatt et al. 2003). Thus, there is no fixed sample size. ASTRA was evaluated in Malawi (Myatt et al. 2003) and Vietnam (Myatt et al. 2005) and found to be more reliable than TRA for the prioritisation of communities with active disease. The advantages of ASTRA are its speed and low cost as a result of smaller sample sizes than are required for PBPS. Sample sizes may, however, become larg