A Trial of Three Rounds of Mass Drug Administration with Azithromycin for Yaws

Background Treponema pallidum subsp. pertenue causes yaws. Strategies to better control and hopefully eliminate yaws are needed. Methods We conducted an open-label cluster-randomized community trial in a yaws-endemic area of Papua New Guinea. Thirty-eight wards were randomized to receive either one mass drug administration (MDA) round followed by two target treatment of active cases rounds (control arm) or three MDA rounds (experimental arm) at 6-month intervals. The difference in the prevalence of active and latent yaws were measured at 18-month surveys. Results Nineteen wards (30,438 individuals) were randomized to the control arm and 19 (26,238 individuals) to the experimental arm. 24,848 azithromycin doses were administered in the control arm (22,033 at baseline, 207 participants with yaws-like lesions and 2,608 contacts at 6-month and 12-month), compared to 59,852 doses in the experimental arm. At 18 months, the prevalence of active yaws had decreased from 0.46% (102/22,033) to 0.16% (47/29,954) in the control arm and from 0.43% (87/20,331) to 0.04% (10/25,987) in the experimental arm (RR 3.54; 95%CI 1.72–7.27). The prevalence of other infectious ulcers decreased to a similar extent in the two study arms. The prevalence of latent yaws at 18 months, assessed in 994 and 945 children in the control and experimental arms, was 6.54% (5.00–8.08) and 3.28% (2.14–4.42), respectively (RR 2.03; 1.12–3.7). Three cases with resistance to macrolides were found in the experimental arm. Conclusions These data show that three rounds of azithromycin MDA 6 months apart are better than one round of azithromycin MDA with two rounds of targeted treatment for decreasing the community prevalence of yaws. Monitoring for the emergence and spread of antimicrobial resistance is needed. (ClinicalTrials.gov number, NCT03490123.)


Setting
Before study commencement, teams underwent standardized training in clinical diagnosis of yaws, rapid diagnostic testing, collection of lesion samples, management of adverse events, and completion of the documentation in line with the Good Clinical Practice guidelines. Each team was equipped with a box containing all material required for the trial and mobile phones for communication with study supervisors in the field.

Clinical surveys for active yaws
Clinical surveys for active yaws prevalence were undertaken in the entire resident population. The clinical definition of active yaws was an ulcerative or nodular skin lesion of more than 1 2 cm in diameter (the size of 5 toea PNG coin). Demographic and clinical data were collected for every case of suspected yaws, including age, gender, number of ulcers, size and duration, and travel history in the preceding 6 months.

Molecular characterization of lesion swab samples
Swab samples were stored in a transport medium (100mM TRIS, 100mM EDTA, 1% SDS) at -20 °C until DNA isolation. DNA was isolated using QIAamp DNA Blood Mini Kit (Qiagen) directly from transport medium according to manufacturer´s instructions and eluted into 100 microliters of AE buffer (Qiagen).
Molecular characterization of samples was performed by amplification and sequencing of 3 different gene loci of T. p. pertenue identified during whole genome comparisons (primers used are listed in the Table  S1). The PCR amplification was performed using nested-PCR protocol and the PCR products of the second step were sequenced by dideoxyterminator method [1].
Allelic profiles based on sequences of TP0548, TP0488 and TP0858 genes were assigned to individual samples. The targets, all of which are putative or bona fide outer membrane proteins, each contain small (300-600 nt) readily amplifiable regions with sequence heterogeneity among strains. The adopted nomenclature for different T. p. pertenue strain types is expressed as one letter, representing tp0548 types, followed by two numbers, representing tp0488 and tp0858 types.
Altogether, three different allelic profiles were discovered: J11, S22 and T13 (corresponding to JG8, SE7 and TD6 described previously) [2]). The presence of point mutations in the 23S rRNA genes causing macrolide resistance (A2058G and A2059G) were identified by sequencing. All samples were screened by qPCR targeting H. ducreyi as described previously [3].
To determine genetic diversity of T. p. pertenue isolates at each timepoint we estimated the Mean Evolutionary Diversity by calculating the number of base substitutions per site across the concatamers of the 3 loci from each round using a Kimura 2 parameter (K80/K2P) model using the dist.dna function from ape v5.4.1 [4]. We evaluated differences between baseline diversity and later timepoints using AMOVA (1000 permutations), implemented in the pegas v0.13 [5] package in R v3.6.0 [6]. Sample distributions were plotted using ggplot2 v3.3.3 [7].

Latent yaws surveys
A 10 μL of capillary blood was obtained from all randomly selected children aged 1-15 years to perform a rapid quantitative serological test (dual-path platform [DPP] syphilis screen and confirm assay; Chembio Diagnostics, Medford, NY, USA) [8]. Naked eye and optical density microreader DPP values were recorded. A participant was considered to have positive yaws serology if the density of the treponemal line was greater than 12 (as established by the manufacturer) and the density of the non-treponemal line was ≥30 (equivalent to rapid plasma reagin [RPR] ≥1:4) [9] . A participant was considered to have hightiter serology if the non-treponemal line had a value ≥90 (equivalent to RPR ≥1:16).

4
Table S1. Primers used for the nested-PCR amplification of typing loci and the 23S rRNA gene of T. p. pertenue.

Locus
External primers (5-  Figure S1. Prevalence of ulcers from each source over time.
Results are presented as population prevalence with error bars showing the 95% confidence interval.