Introduction

Trachoma is the leading infectious cause of blindness worldwide.[1] It is caused by repeated re-infection with the ocular serotypes (A, B, Ba and C) of the bacterium Chlamydia trachomatis, and is predominantly found in the poorest countries in the world. Active trachoma, characterised by the presence of subepithelial follicles (trachomatous inflammation, follicular (TF)) and/or inflammation (trachomatous inflammation, intense (TI)), is usually found in children. After years of repeated re-infection scarring may occur, which can lead to distortion of the eyelid, causing the eyelashes to turn inwards (trichiasis (TT)) and scratch the cornea, resulting in corneal opacity and blindness.[2]

The World Health Organization (WHO) strategy for Global Elimination of Blinding Trachoma by the year 2020 (GET2020) is through employment of the SAFE strategy (Surgery for trichiasis, Antibiotics for active trachoma, Facial cleanliness, and Environmental improvement).[3] The Gambian National Eye Care Programme (NECP), established in 1986, expanded its national intervention programme to cover the whole country by 1996. NECP activities include the training of health workers in primary eye care, surgery for trichiasis cases, recognition and treatment of conjunctivitis, school screening, and face-washing promotion.[4] The NECP also has a network of nyateros (Friends of the Eye) who are non-health professionals identified by their own communities, and have been trained to promote good eye health practices in the community. The programme policy is to treat active trachoma cases and household contacts with tetracycline eye ointment. At the time the survey presented here was conducted (January to March 2006) The Gambia was yet to receive a donation of the antibiotic azithromycin from the International Trachoma Initiative (ITI) for mass treatment.

Evidence from two national surveys carried out in 1986 and 1996 suggested that the prevalence of trachoma fell in The Gambia, with age standardised prevalences of blinding trachomatous corneal opacities falling from 0.10% to 0.02%. In the same time span, the prevalence of active trachoma in 0–14 year-olds fell from 10.4% to 4.9%.[5] Disease prevalence data collected in surveys is of importance to any national control programmes seeking to eliminate blinding trachoma in accordance with the WHO definitions (prevalence of TT less than 1 case per 1000 total population, and prevalence of TF in 1–9 year-old children less than 5%).[3] Data collected in surveys also allow control efforts to be directed to trachoma endemic areas. However, many studies have demonstrated that the prevalence of detected ocular C. trachomatis infection is lower than the prevalence of active trachoma, especially in mass treated and low prevalence settings.[6],[7],[8],[9] The Gambia has not collected infection data in its national surveys but these data may lead to a better understanding of the disease's epidemiology with potential implications for the introduction of trachoma control interventions. Mass antibiotic treatment for trachoma is likely to be effective in treating communities with C. trachomatis infection, but of questionable value where no infection can be demonstrated. [10],[11]

We aimed to estimate the prevalence of active trachoma and ocular C. trachomatis infection in children aged less than 10 years, a decade after the last national survey, in Lower River Region (LRR) and North Bank Region (NBR). These two regions had the highest prevalence of active trachoma in children aged 0–9 years in 1996 (11.5% and 7.7%, respectively).[5] This study reports the results of this cluster-randomised cross-sectional survey in The Gambia, conducted in 2006.

Methods

Results

Of the 61 randomly selected EAs, 60 consented to participate. A total of 2990 children aged 0–9 years (950 from LRR and 2040 from NBR) were examined from an estimated population of 75,000. Five children from three EAs were excluded from analyses because they were aged over 9 years, and in five EAs only 49 children were screened.

A total of 338 children were enumerated but not examined (Table 1). The age distribution in those examined was significantly different from those not examined (p<0.001), with a higher proportion examined aged 1–5 years and a lower proportion aged 6–9. The sex distribution was similar (p = 0.954). The examined population also differed from the underlying populations of LRR and NBR established in the 2003 census,[12] but in the opposite direction, with a lower proportion examined aged 3–6 and a higher proportion aged 7–9 years. Adjusting for age against the populations established by census led to a TF prevalence in children aged 1–9 years of 13.2% in LRR and 11.3% in NBR, versus 12.3% and 10.0% when unadjusted, respectively. This reflects the peak TF prevalence in children aged 1–5 years.[2] However, since trachoma control programmes usually do not calculate adjusted prevalence rates before making treatment decisions, we have presented unadjusted prevalences.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Age distribution of children enumerated and children examined, and prevalence of TF, TI and TS in those examined.

https://doi.org/10.1371/journal.pntd.0000573.t001

Prevalence of clinical signs

Overall, TF was found in 310 (10.4%) children aged 0–9 years (Table 1). Children aged 3–5 years had the highest prevalence of TF (17.0%), followed by those aged 1–2 years (13.1%). Children under one year old and those aged 6–9 years had TF prevalences of 5.6% and 6.3%, respectively. Only three (0.1%) children had TI (two in NBR and 1 in LRR), and two of these also had TF. The left and right eye were concordant for TF (both clinically normal or both with TF) in 2933 (98.1%) children.

Considering the 1–9 year age group, the overall TF prevalence (the WHO indicator) was 10.7% (95% CI 8.7–13.1). The prevalence of TF in LRR and NBR, respectively, was 12.3% and 10.0% (Table 2). The design effect for heterogeneity among EAs was 3.474 overall; 3.322 for LRR and 3.561 for NBR. The district prevalence varied significantly in LRR (p = 0.006) from 5.7% to 18.2%, and in NBR (p = 0.001) from 5.6% to 15.1%. There was a trend towards higher rates in the eastern districts compared to those in the west in LRR (χ² test for trend p<0.001) but not in NBR (p = 0.287) (Figure 1). Prevalence also appeared to vary dramatically between EAs in the same district. For example, in Jarra West in LRR, prevalence ranged from 0% to 36.7%, and in Central Baddibu in NBR, prevalence ranged from 6.3% to 34.0%. Comparing the results with those from 1996 indicates little change overall: LRR increased slightly from 11.5% to 11.9%, and NBR increased from 7.7% to 9.7%. [5].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Spatial distribution of TF prevalence in 1–9 year-olds in districts and sampled villages in North Bank and Lower River Regions.

https://doi.org/10.1371/journal.pntd.0000573.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Prevalence of trachomatous inflammation, follicular (TF) in 1–9 year-olds by region, district, and enumeration area (EA).

https://doi.org/10.1371/journal.pntd.0000573.t002

There were three urban EAs in LRR and nine in NBR. Overall, 53 of 552 (9.6%, 95% CI 7.3–12.4) urban children aged 1–9 years had TF, compared with 245 of 2225 rural children (11.0%, 95% CI 9.7–12.4). There was no evidence of variation in TF prevalence between urban and rural EAs in either LRR (p = 0.395) or NBR (p = 0.854).

Discussion

The results of this survey, covering two regions of The Gambia with a population of approximately 245,000 people, suggest that the overall prevalence of active trachoma in children aged 0–9 years has not noticeably changed since the 1996 national survey. There is some evidence of a secular trend in trachoma prevalence in The Gambia prior to 1996: in one rural village active disease prevalence reduced from the 65.7% documented in 1959 to 2.4% in 1996 in the absence of direct control interventions [19]. However there is no evidence that such a trend continued in LRR and NBR between 1996 and 2006 despite trachoma control efforts being in place in these areas. In contrast to TF, the prevalence of C. trachomatis infection was extremely low in this survey, with only 3 of the 940 samples from LRR that were positive for human mtDNA also testing positive for C. trachomatis. No infection data were collected in the previous Gambian national surveys so no trends over time can be estimated. In addition, other countries have not collected data on ocular C. trachomatis infection when conducting population-based surveys of active trachoma, so between-country comparisons cannot be made. However, our results show a lower prevalence of infection compared with research studies conducted in The Gambia over the last two decades.[10],[20],[21],[22] In 1991, the prevalence of infection among all inhabitants in two Gambian villages was 17.2%.[20] In 1999, the overall prevalence in eight Gambian villages was 35.9%.[21] Burton et al. reported an overall prevalence of 7.2% in 14 Gambian villages in 2003,[10] and 19.8% in two villages in 2006.[22] It is important to note that these projects were focused on villages with notable trachoma public health problems usually identified by finding a high prevalence of active trachoma on prior screening, with later laboratory demonstration of high rates of ocular C. trachomatis infection. In these villages, therefore, higher prevalence rates may have been recorded than would be estimated in the national Gambian population.

A disparity between the prevalence of active trachoma and of ocular C. trachomatis infection has been reported by others, especially in communities where the prevalence of disease is declining, or in communities that have received mass treatment with antibiotics.[6],[9],[10],[20],[23],[24],[25],[26],[27],[28] Solomon et al. observed a lag time of several years between infection elimination and the prevalence of TF falling below 5% in a mass-treated Tanzanian community.[7],[26] In a low prevalence region of Nepal, 6.3% of children aged 1–10 years had clinical signs but no C. trachomatis was detected.[6] Our results suggest that trachoma may be in decline in The Gambia, with the clearance of clinical signs lagging behind that of infection. That TF cases remain clustered by EA with a design effect of 3.474 may suggest that previous transmission is being captured, or that undetected infections may have existed in close contacts. Despite diagnostic assays within the last decade becoming increasingly more sensitive, the prevalence of detectable C. trachomatis infection is now very low. Our observation that the vast majority of cases were TF, with only three cases of TI, supports this hypothesis, as TI cases are more likely to have detectable C. trachomatis infection and with a higher organism load.[11],[21],[29],[30] In the one TI case tested here, C. trachomatis infection was not detected. Clinically, our impression was that many of the TF cases were rather mild, in the senses that cases with very large numbers of follicles were rarely present, and also that ‘minor signs’ of trachoma such as active pannus and limbal follicles[31], were rare. Possibly therefore increasing the 10% TF threshold for intervention, increasing the number of central zone follicles required to diagnose TF, and including minor signs of trachoma in the assessment, might all improve the ability of clinical signs to predict communities or districts with significant ocular C. trachomatis infection. Much more data would however be needed to justify any such departure from the current simplified survey methods.

This association between prevalence of disease, severity of disease, and load of C. trachomatis infection could provide another explanation for the observed disparity between prevalence of clinical signs and detected ocular C. trachomatis infection. The bacterial load may have been too low for detection by Amplicor PCR. However, the detection limit of Amplicor PCR has been placed at about 2–4 elementary bodies (or 20–40 plasmid copies) per 100 µl,[32] and thus even a few bacteria collected on the swab should have yielded a positive result. The use of a Dacron polyester-tipped swab may have lowered the PCR detection level as it absorbs approximately 200 µl of Amplicor lysis buffer and therefore a higher volume is required to elute the sample. The use of a flocked swab could help overcome this.[33]

In terms of clinical grading, the two examiners were experienced and had fulfilled the criteria of the validation sessions. Swabs were taken according to a standardised protocol [11] by experienced ophthalmic nurses, and 99.5% (937/942) of the samples tested positive for human-specific mtDNA. Thus, we believe that the observed prevalence of active trachoma, as well as the low prevalence of ocular C. trachomatis infection, is genuine.

The follicular conjunctivitis (TF) observed in this study could alternatively be due to organisms other than C. trachomatis. Potential causes include adenovirus, Herpes simplex virus, Epstein Barr virus, Molluscum contagiosum, Moraxella spp. and other species of Chlamydia, such as C. pneumoniae, which was first isolated from the eyes of children with trachoma.[6],[34],[35],[36],[37],[38] Further laboratory investigations to test the C. trachomatis negative swabs from clinically active children for these other organisms may be warranted.

Our results have two important implications. The first relates to policy decisions regarding control efforts, particularly now that The Gambia has for the first time received a donation of azithromycin for trachoma control. Currently, according to WHO criteria, districts and communities with a TF prevalence greater than or equal to 10% in children aged 1–9 years should receive mass treatment annually, in addition to the ‘F’ and ‘E’ components of the SAFE strategy, for at least three years, until the prevalence of TF falls below 10%.[39] According to these criteria, 3 districts and 10 EAs in LRR, and 3 districts and 17 EAs in NBR, should be treated. However, our sampling strategy was designed to have an accurate estimate of TF prevalence at the region level and not at the community or district level. The size of Gambian regions is similar to the average size of a district in many countries (100,000 to 150,000 people). On this basis, our results suggest that the whole of LRR and NBR should undergo mass treatment according to WHO criteria, as their respective TF prevalence in 1–9 year-olds is 12.3% and 10.0%.

Conversely, the very low prevalence of C. trachomatis infection indicates that treatment with antibiotics is almost certainly not necessary in many of these communities, since transmission of C. trachomatis is now presumed to be rare. Treatment decisions based on clinical signs of trachoma may lead to unnecessary mass treatment of whole regions or districts, wasting the scarce resources available, although there may be benefits beyond trachoma control to the mass distribution of azithromycin, such as a decline on overall mortality as shown in Ethiopia.[40] A cheap, rapid, point-of-care test that can detect ocular C. trachomatis infection could allow policy makers to devise treatment strategies based on the prevalence of infection. At present, no such test is available despite previous encouraging results [41]. Alternatively, the use of laboratory tests, such as PCR, has been suggested as a means of detecting ocular C. trachomatis infection. Despite reduced costs through specimen pooling[42], national programmes are unlikely to adopt these assays as they require expensive reagents, electricity-dependent equipment, and highly trained technicians.[43] In the absence of a test for infection for use by national programmes, our observation that TF clustered by EA supports the strategy of treating households containing someone with active trachoma in order to reach infected but clinically normal individuals.[10]

The second implication concerns itself with the question of whether children with clinical signs of trachoma in the absence of C. trachomatis infection will develop blinding sequelae. The epidemiology of trachoma is not sufficiently understood to enable this question to be answered for certain, although the evidence that repeated infection is required for disease progression suggests that it is unlikely.[44],[45]

This study shows a large discrepancy between the prevalence of clinical signs and of C. trachomatis infection in two Gambian regions in the absence of a national mass treatment programme. The prevalence of TF in children aged less than 10 years in the two Gambian regions that were sampled remains largely unchanged from the prevalence 10 years ago, at around 10%. In contrast, the prevalence of C. trachomatis infection is low (0.3%). Our results indicate that The Gambia may not only be on course for certification of blinding trachoma elimination according to the WHO definition (prevalence of TT less than 1 case per 1000 total population, and prevalence of TF in 1–9 year-old children less than 5%)[3], but also on course to eliminate the ocular strains of C. trachomatis prior to any mass azithromycin distribution.

Acknowledgments

We would like to thank the village leaders and villagers for participating in this study. We thank Bakary Ceesay, Jawo Colley, Bakary Gibba, Abdoulie Jallow, Hassan Joof, Mass Laye, Pateh Makalo and Binta Sonko for their work in the field. Thanks to Haddy Kanyi and Kaddy Conteh for help with data entry and to Aura Aguirre-Andreasen for laboratory support.