Trachoma and Ocular Chlamydia trachomatis Were Not Eliminated Three Years after Two Rounds of Mass Treatment in a Trachoma Hyperendemic Village

Abstract

purpose. The World Health Organization recommends mass treatment of trachoma-hyperendemic communities, but there are scant empiric data on the number of rounds of treatment that are necessary for sustainable reductions. The rates of active trachoma and infection with C. trachomatis were determined in a community 3.5 years after two rounds of mass treatment with azithromycin.

methods. Maindi village in Tanzania received a first round of mass treatment with azithromycin after a baseline survey for trachoma and infection. All residents aged 6 months and older were offered single-dose treatment with azithromycin (excluding pregnant women with no clinical trachoma, who were offered topical tetracycline). The residents were followed over an 18-month period, and, according to similar treatment criteria, were offered retreatment at 18 months. Five years after baseline (3.5 years after the second round of mass treatment), a new census and survey of current residents for trachoma and infection was conducted. Children are the sentinel markers of infection and trachoma in communities, so data are presented specifically for ages 0 to 7 years (preschool age) and 8 to 16 years.

results. Treatment coverage was above 80% for all ages in the first round, and highest (90%) in preschool-aged children. Second-round coverage was lower, <70%, and 70% in preschool-aged children. At 5 years, trachoma rates were still lower than baseline, ranging from 45% in those aged 0 to 3 years to 8% in those aged 11 to 15 years (compared with 81% and 39% at baseline, respectively). Infection rates at baseline ranged from 71% to 57%, but were 27% to 17% at 5 years after two rounds of mass treatment. At 5 years, there were no differences in trachoma or infection rates, when comparing new residents who came after the second mass treatment with those who had been resident in the village during both rounds (P > 0.05). Infection rates were lower in those who had been treated twice or at 18 months than in those treated only at baseline or never treated.

conclusions. Although mass treatment appears to be associated with lower disease and infection rates in the long term, trachoma and C. trachomatis infection were not eliminated in this trachoma hyperendemic village 3.5 years after two rounds of mass treatment. Continued implementation of the SAFE strategy in this environment is needed.

Trachoma, an eye disease caused by repeated and/or prolonged episodes of Chlamydia trachomatis, is the leading infectious cause of blindness world-wide. ¹ In recognition of this major public health problem, the World Health Organization (WHO) has created the alliance for the Global Elimination of Blinding Trachoma by the year 2020. The cornerstone of the approach to controlling trachoma is the SAFE strategy: Surgery (for late complications), Antibiotics (to reduce the pool of infection in the community), Face-washing and Environmental improvements (to inhibit transmission). For the antibiotic component, there are scant data to suggest the number of rounds of mass treatment needed to reduce infection and disease. 

Studies of community-based mass treatment with a single dose of azithromycin suggest that one round with high coverage is effective in lowering infection rates. ^{2 3 4 5 6 7} A study in small villages from a trachoma-hypoendemic area in the Gambia suggested that a single round of mass treatment would be sufficient for the virtual elimination of infection, unless the infection was introduced from outside exposure. ⁴ In a mesoendemic village in Tanzania, a single round of mass treatment plus provision of tetracycline ointment to active cases at 6, 12, and 18 months eliminated infection. ³ However, in a hyperendemic community in Tanzania (which is the subject of these analyses), a single round of mass treatment reduced infection rates but did not eliminate infection and there was evidence of re-emergence at 18 months. ⁶ Similarly, in a study of twice-yearly targeted treatments (only school children with disease and their families were treated) with poor infection coverage in Vietnam, infection and clinical trachoma appeared to decline; but once no more antibiotic was provided, there was some evidence of re-emergence. ⁸ In hyperendemic villages in Ethiopia, a single round of mass treatment was effective in reducing infection in eight villages, with an average coverage of more than 90% (individual village’s coverage was not provided) with evidence for return of infection by 24 months in seven of the eight villages. ⁹ In a study in Nepal where more than 30% of children had active trachoma, three rounds of treatment directed at children age 1 to 10 years were successful in reducing infection and disease in these children up to 6 months after the last treatment. ¹⁰ However, it is not clear whether 6 months is long enough to determine whether infection and disease would re-emerge. 

A WHO working group on trachoma control recommended three rounds of yearly mass treatment in communities where the trachoma rate in children age 0 to 10 is greater than 10%. ⁷ However, there are few published data on the effect of more than one round of mass treatment on infection or clinical trachoma. The purpose of this study was to determine, after two rounds of mass treatment with azithromycin spaced 18 months apart, the rate of trachoma and infection at 5 years after baseline in Maindi. The population of this village was studied at 6, 12, and 18 months after one round of mass treatment and was shown to have a decline in infection, but at 18 months, infection and infection load was starting to increase with evidence of spreading across households. ^{5 6} At 18 months, mass treatment was again provided to the village, and the purpose of this study is to evaluate infection and clinical trachoma at 3.5 years after two rounds of mass treatment. 

Methods

The village of Maindi in Kongwa district of Central Tanzania, was originally studied in 2000 for the effect, up to 18 months, on infection and trachoma of a single round of mass treatment with azithromycin, and the methods and results have been reported at baseline and follow-up. ^{5 6 11} Briefly, all residents of the village were eligible for the study, and 873 (86%) of 1017 participated in the baseline examination, performed in April to June of 2000. After baseline examination, the participants received mass treatment with a single dose of azithromycin of 20 mg/kg up to 1 g. Fifty-seven women who self-reported being pregnant without signs of active trachoma were provided topical tetracycline, and children less than age 6 months were not treated, in accordance with the approved use of azithromycin in Tanzania; of the 10 children in that age group, one was offered topical tetracycline. During the 18-month course of the study, census updates were performed every 2 weeks, and new village members were added and those who had moved or died were removed from the census. At the end of the 18-month survey (performed in November–January of 2001–2002), the village residents were again offered mass treatment with the same approach as before. One child aged <6 months of 32 in that age group was offered topical tetracycline, 8 were treated with azithromycin (as there was some question that they might be older than 6 months), and 27 pregnant women with no clinical signs were offered topical tetracycline. 

No further treatment with antibiotic for trachoma was provided to the community once the previous study had ended. The Tanzania National Trachoma Control Program uses infrequent radio spots to broadcast messages about face washing and latrine construction, but no specific health education programs within Maindi were provided. At the final 18-month examination, our survey team reminded mothers of the importance of face washing to reduce trachoma. No additional water sources for the community were built in the interim between the last survey and the present study. 

In 2005, the village underwent an additional census by our trained team to determine the current status of previous residents and to add persons who had moved into the village or were born since the time of the last census update 3.5 years previously. Transient visitors, defined as those who had not lived in the village for at least 1 month and who did not plan to stay for another 6 months, were not added. 

After the census, a trachoma survey was conducted in May to July, targeting everyone in the current census aged 6 months and older. An experienced trachoma grader (HK) with 2.5× loupes assessed the presence of clinical trachoma using the WHO simplified grading system for presence and absence of follicular trachoma (TF) and intense trachoma (TI). ¹² Active trachoma is defined as anyone with either of those signs. Photographs were taken of the left eye and graded for trachoma. The comparison of photogrades and clinical grades showed high agreement (κ > 0.65 for both signs). A Dacron swab was rubbed across the upper left conjunctiva and placed in a sterile, dry, vial and frozen to be analyzed for C. trachomatis. Careful techniques were used in the field to avoid contamination. At the laboratory station in the field, only the laboratory supervisor was allowed to open the vial and open the swab. He was the only one to touch the swab and take the specimen. He wore gloves that were changed between subjects, but he did not touch anyone during the entire session, not even the subject. The person who flipped lids changed gloves between persons. Once new gloves were on, he was not permitted to touch anything other than the subject’s upper eye lid. If there was inadvertent touching of surfaces or other subjects by either person, the gloves were removed and new gloves put on. 

The swabs were stored in a −20° freezer until they were shipped to the International Chlamydia laboratory at Johns Hopkins for processing. 

A C. trachomatis qualitative PCR assay (Amplicor; Roche Molecular Systems, Indianapolis, IN) was used to identify samples positive for organism DNA. In essence, PCR was performed on eluted samples according to the manufacturer’s instructions. Testing for PCR inhibition was also performed according to the manufacturer’s directions. If there was evidence of inhibition, samples were retested by diluting 1:5 with a 50:50 mixture of lysis buffer (Amplicor CT/NG; Roche Molecular Systems) and specimen diluent. A positive test was defined, as per the manufacturer, as an optical density >0.7. Equivocal samples were those in which the optical density was between 0.2 and 0.7 in two replicates. 

For the purposes of this study, we show data from children aged less than 15 years, because there was very little trachoma or infection in persons aged 16 and older at either the 18-month time point (trachoma rate of 10% and infection rate of 5%) or the 5-year follow-up (3% and 9%). We concentrated on children as the sentinel markers of disease and infection in communities, similar to other studies and to WHO Guidelines for Ultimate Intervention Goals. However, we show antibiotic coverage data for the entire community, consistent with national treatment reporting guidelines. The data analyses considered both the treatment and an age effect. That is, the residents of the village at baseline who were exposed to both rounds of treatment were all 5 years older by the time of this follow-up, and it is particularly children 3 years of age and younger who have high rates of trachoma and infection. ⁶ Thus, it was important to compare children by age groups at the time of the various surveys, for proper interpretation. At each survey time—baseline, 18 months, and 5 years—there are different denominators, depending on the changing census. For this reason, we did not try to compare the same people over time, but rather compared the cross-sectional prevalences in the relevant groups, testing significance by using probabilities adjusted for age, gender, and clustering within children in the same household. 

The research adhered to the tenets of the Declaration of Helsinki. All procedures received ethics approval by the Tanzania National Institute for Medical Research and the Institutional Review Board of Johns Hopkins University. Pfizer International had no role in the study design, protocols, data collection, or data analyses or in the preparation of this report. 

Results

Of the 1282 residents of the village in the current survey (2005), 342 were new arrivals, and 216 were born after the 18-month visit in 2001. Since baseline, 583 of the original sample had either died or moved away (Fig. 1) . The new arrivals were spread evenly throughout the village, and did not cluster in any one neighborhood (data not shown). The trachoma and infection rates at 18 months were higher in the people who stayed in the village than in those who died or moved away after the 18-month survey (trachoma rates of 35.9% vs. 27.1% [age-, gender-adjusted P = 0.09] and infection rates of 13.8% vs. 9.0% [age-, gender-adjusted P = 0.036]). 

The coverage rate for the first round of mass treatment (azithromycin and topical tetracycline) was high, ranging from 91% (282/311) in the age group 6 months to 7 years, to a low of 79% (157/199) in the age group 16 to 25 years. Compliance with the second round of treatment was lower, ranging from 70% (258/370) in the age group 6 months to 7 years to a low of 55% (110/201) in the age group 16 to 25 years (Table 1) . For those in the village at both rounds, a total of 60% (484/814) received treatment at both rounds. However, most received at least one round, as only 6% (46/814) were not treated at either visit. 

The prevalence of active trachoma and infection, based on the cross-sectional surveys at baseline, 18 months, and 5 years after baseline in the children younger than 15 years are shown in Table 2 . At baseline, the prevalence of active trachoma varied by age group, from 81% in children younger than 4 years, to 39% in the 11- to 15-year-olds. Eighteen months after the first mass treatment, the prevalence of active disease was lower than at baseline. Five years after baseline treatment and 3.5 years after the second treatment, the prevalence of active trachoma was even lower than at the 18-month follow-up. Of note, at baseline, 81% (108/134) of children younger than 4 years had active trachoma, compared with 45% (91/204) in those younger than 4 years at the 5-year follow-up; the latter were born into the village since the 18-month treatment and had never received treatment. 

Disease prevalences were stratified by number of times treated (Fig. 2) . Of note, disease prevalences were similar within the age group 4 to 7 years, irrespective of the number of times that they have been treated. Children aged 8 years and older appeared to derive the most benefit for trachoma from two rounds of mass treatment. 

After the first mass treatment, the rates of infection declined and were low at 18 months. Unlike the steady decline in active trachoma observed over time, at 3.5 years after the second mass treatment, the prevalence of infection in the children was higher than at the 18-month follow-up, ranging from 28% in the <4-year-olds to 17% in the 10- to 15-year-olds. However, the infection rates were significantly lower for those treated both times and those treated at 18 months only, compared with those who had never been treated or were only treated at baseline (prevalence infection of 28% versus 17%; age-, gender-, and cluster-adjusted P = 0.02; Fig. 3 ). At baseline, children younger than 4 years had a prevalence of infection of 71% (94/133), compared with the 5-year survey results of 27% (55/202) for children younger than 4 years born into the village since the 18-month treatment. Neither of these cohorts had received previous treatment at the time of the survey, but the latter group was born 5 years later and born into a village that had lower pressure of infection compared with the earlier cohort. 

We also compared those who were new arrivals to the village between 18 months and 5 years from baseline (and thus had not received antibiotic treatment) with those who had been in the village during at least one round of mass treatment (Table 3) . The prevalence of active disease in those ages 4 to 15 years and new to the village at the 5-year follow-up was similar to the prevalence of those that had been residents (prevalence 27% vs. 26%; age-, gender-, and cluster-adjusted P = 0.71). Similarly, there was no difference in infection rates between those new to the village and those who were previous residents (infection rates of 26% versus 20%; age-, gender-, and cluster-adjusted P = 0.10). 

Discussion

Although trachoma and C. trachomatis infection have not been eliminated from this village after two rounds of mass treatment, the data suggest that rates were substantially lower at 3.5 years after the last round compared with baseline. The rates had clearly not returned to pretreatment levels, although infection rates were higher than at 18 months after the first round of mass treatment. These rates could reflect several factors. First, infection rates may be slowly returning to pretreatment levels, but will take much more time to return. As this was only a single survey, we do not have data to model a trajectory of re-emergent infection. Second, coverage rates were lower for the second round. Another factor is that, although no directed health education or infrastructure development occurred over the 5 years, the village is part of a district with other villages enrolled in the National Trachoma Control Program of Tanzania. The program includes mass treatment and national media attention to face-washing and sanitation through radio announcements. This suggests another possibility: After treatment lowered the community infectious pool, a change occurred in the stability of transmission within the study village, reflecting less intense transmission as well as pressure of infection, which could decrease trachoma and keep infection rates from returning quickly. Thus, this village, in the absence of further treatment or intervention, may be stabilizing at a lower rate of disease and infection. The fact that trachoma rates continued to decline steadily over the course of the 5 years supports this possibility. We do not feel that our results could be due to field contamination of laboratory specimens, as we strictly followed previously standardized field protocols at each visit. Even if there was some contamination, the rates of infection would be even lower and again not at pretreatment levels. Finally, we cannot exclude other secular trends, apart from antibiotic treatment and the National Program, resulting in a decline in trachoma as well. This study focused on a population cohort in a single village, and ideally, secular trends would be studied by comparison with populations in other villages that have had no intervention. Chidambaram et al. ⁹ observed, in hyperendemic areas of Ethiopia, that the infection rates were about half in children aged 1 to 5 years in the villages enrolled in the study 12 months after the baseline compared with rates in the baseline villages. They attributed the difference to secular trends of declining disease, but acknowledged that the findings could be due to selection of very different villages at 12 months than at baseline. In our years of performing studies in the hyperendemic villages in Tanzania, we have not observed such dramatic declines in 12 months as Chidambaram et al. ⁹ reported with no intervention. Other researchers, observing secular trends, describe a much longer time scale. In Malawi, for example, Hoechsmann et al. ¹³ observed a halving of disease prevalence, but after a 16-year period. 

We expected, compared to residents, that new arrivals to the village after the last round of mass treatment would have higher rates of infection and disease as they did not have the benefit of previous treatment. However, new arrivals did not have disease or infection rates that were significantly different from those who had been in the village. Possible explanations for this finding include the following: First, we do not have data on the actual arrival dates of the new persons, except the births. Those “new” to the village may have in fact been there for, on average, 2.5 years since baseline. Thus, it is not unexpected that the trachoma and infection rates in this group would equilibrate to those persons already in the village. We have previously demonstrated that spread of infection across households occurs within 12 months. ¹¹ This explanation is supported by the finding at 5 years that the 0- to 3-year-olds, who were born into the village after the last treatment, were not protected from incident infection but rather have rates of disease and infection similar to the 4- to 7-year-olds with whom they doubtlessly mingle. Second, we do not know from which villages the new arrivals came—whether it was from other program villages—the new arrivals may have had equally low or lower rates of infection and trachoma when they arrived, comparable to those in the village. Finally, the village had high rates of in and out migration, and we have shown that those who left the village after the 18-month survey were less likely to have trachoma and infection compared with those who stayed. Thus, if there had been less migration, the rates in the village residents may have been lower than were observed at the 5-year survey. However, migration is a factor that national trachoma control programs will have to contend with in planning coverage of villages and districts. 

We feel our findings of sustained low rates of infection likely reflect the antibiotic mass treatment coverage. As we have reported previously, the first round of mass treatment had very high antibiotic coverage, which resulted in a dramatic drop in infection, but did not eliminate either trachoma or infection by 18 months. ⁶ In hyperendemic villages, there will be persons with very high loads for whom a single dose may be insufficient to clear infection. We previously reported re-emergent infection within families by 6 months, from family members who had not cleared infection despite treatment. ⁶ The second round of mass treatment had much lower coverage, although more than 70% of the highest risk group, preschool-age children, was treated. These coverage rates are lower than ideal, but may be more typical of what programs providing multiple rounds of mass treatment are able to achieve. We can speculate that if coverage of the second round had been especially high, the rates observed at 5 years may have even been lower. Solomon et al., ³ in another village with lower rates of trachoma and infection at baseline, achieved virtual elimination of infection 2 years after mass treatment, with very high coverage rates (98%) plus interim (2, 6, 12, and 18 months) treatment of trachoma cases with antibiotic ointment. Even though our coverage for the second round was less than the 80% target, at 3.5 years after treatment, the infection and trachoma rates were not back to pretreatment levels, suggesting a long-term benefit of two rounds of mass treatment even at <80% in the second round. 

However, to reach the Ultimate Intervention Goal of WHO of trachoma in less than 10% of children younger than 10 years and to be certain the low rate of infection will be sustainable, more must be done. Attaining the goal may require more than two rounds of mass treatment, probably spaced closer together, as projected by a model of infection elimination, ¹⁴ and more improvements in the environment and hygiene, as recommended by the WHO SAFE strategy approach. 

 Supported by National Eye Institute Grant EY001765 and Pfizer International. SW received an unrestricted grant as a Research to Prevent Blindness Senior Scientific Investigator.

 Submitted for publication June 8, 2006; revised November 10 and December 8, 2006; accepted February 20, 2007.

 Disclosure: S.K. West, None; B. Munoz, None; H. Mkocha, None; C. Gaydos, None; T. Quin n, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Village residents at the three survey times.

Table 1.

 

Single-Dose Azithromycin Coverage of Maindi Village at Baseline and 18 Months by Age and Gender

Age and Gender	Overall Baseline Coverage*			Overall 18-mo Coverage*			Present at Both Times N	Treated at Baseline Only %	Treated at 18 mo Only %	Treated at both Rounds, † %
		N	%	N	%
Male
6 mo–7 y	142	90.9	175	69.1	118	17.8	5.1	73.7
8–15 y	106	79.3	117	61.5	94	30.9	7.4	50.0
16–25 y	92	73.9	91	45.0	61	41.0	11.5	34.4
26–35 y	59	78.0	58	62.1	54	25.9	14.8	51.9
>35 y	65	87.7	67	67.2	51	29.4	11.8	54.9
Female
6 mo–7 y	169	90.5	191	71.7	133	18.8	9.8	69.9
8–15 y	115	91.3	123	67.5	91	26.4	3.3	65.9
16–25 y	107	83.2	110	62.7	76	31.6	13.2	52.6
26–35 y	66	89.4	66	60.6	56	32.1	8.9	55.4
>35 y	96	87.5	102	63.7	80	26.3	3.7	61.3
Total
6 mo–7 y	311	90.7	370	69.7	251	18.3	7.6	71.7
8–15 y	221	85.5	240	64.6	185	28.7	5.4	57.8
16–25 y	199	78.9	201	54.7	137	35.8	12.4	44.5
26–35 y	125	84.0	124	61.3	110	29.1	11.8	53.6
>35 y	161	87.6	169	65.1	131	27.5	6.9	58.8

 Gender is missing for four children in the 6 mo–7 y age category.

*  Based on age at time of survey.

†  Based on age at baseline.

Table 2.

 

Prevalence of Active Trachoma^* and C. trachomatis Infection over Time in Maindi Village, by Age

Survey Time	Age Groups
		6 mo–3 y	4–7 y	8–10 y	11–15 y
Baseline Prevalence*
% TF alone	53.0 (71/134)	53.3 (73/137)	43.0 (34/79)	31.2 (35/112)
% TI (w or w/o TF)	27.6 (37/134)	19.7 (27/137)	13.9 (11/79)	8.0 (9/112)
% Infection	70.7 (94/133)	66.4 (91/137)	63.3 (50/79)	56.8 (63/111)
18 months after baseline, †
% TF alone	46.3 (75/162)	52.9 (74/140)	40.3 (27/67)	16.1 (19/118)
% TI (w or w/o TF)	9.9 (16/162)	6.4 (9/140)	3.0 (2/67)	7.6 (9/118)
% Infection	13.7 (22/161)	24.3 (34/140)	16.4 (11/67)	9.6 (11/115)
5 years after baseline, †
TF alone	33.8 (69/204)	36.4 (60/165)	11.6 (11/95)	5.6 (8/143)
TI (w or w/o TF)	10.8 (22/204)	12.7 (21/165)	6.3 (6/95)	2.1 (3/143)
% Infection	27.2 (55/202)	28.2 (46/164)	20.2 (19/95)	16.9 (24/142)

*  Active trachoma is defined as having either TF alone or TI (with or without TF).

†  Based on age at the time of the survey.

Figure 2.

 

Rates of trachoma in children younger than 16 years, 3.5 years after the second round of mass treatment, by number of rounds of treatment.

Figure 3.

 

Rates of C. trachomatis in children younger than 16 years, 3.5 years after the second round of mass treatment, by number of rounds of treatment.

Table 3.

 

Prevalence of Trachoma^* and Infection with C. trachomatis at 5-Year Survey, Comparing Those New to the Village at 5 Years with Those Present at Baseline and/or 18 Months

Groups	Age Groups
		6 mo–3 y	4–7 y	8–10 y	11–15 y
Prevalence of trachoma in those new to the village, †
TF alone	33.8 (69/204)	38.0 (19/50)	12.5 (3/24)	7.5 (3/40)
TI (w or w/o TF)	10.8 (22/204)	16.0 (8/50)	0.0 (0/24)	0.0 (0/40)
Infection	27.2 (55/202)	34.0 (17/50)	12.5 (3/24)	27.5 (11/40)
Prevalence of trachoma in those in the village at baseline and/or 18 months, †
TF alone	N/A	35.7 (41/115)	11.3 (8/71)	4.9 (5/103)
TI (w or w/o TF)	N/A	11.3 (13/115)	8.5 (6/71)	2.9 (3/103)
Infection	N/A	25.4 (29/114)	22.5 (16/71)	12.8 (13/102)

*  Active trachoma is defined as having either TF alone or TI (with or without TF).

†  Age is based on age at 5-year survey: Those children aged <1 year at baseline were 5 years old at the 5-year survey; so children age 0–3 can only be new to the village at the 5-year survey.

The authors thank Billie Jo Wood for careful laboratory assistance and the Kongwa Trachoma Project team for their field efforts.