Bacterial and host determinants of cough aerosol culture positivity in patients with drug-resistant versus drug-susceptible tuberculosis

Abstract

A burgeoning epidemic of drug-resistant tuberculosis (TB) threatens to derail global control efforts. Although the mechanisms remain poorly clarified, drug-resistant strains are widely believed to be less infectious than drug-susceptible strains. Consequently, we hypothesized that lower proportions of patients with drug-resistant TB would have culturable Mycobacterium tuberculosis from respirable, cough-generated aerosols compared to patients with drug-susceptible TB, and that multiple factors, including mycobacterial genomic variation, would predict culturable cough aerosol production. We enumerated the colony forming units in aerosols (≤10 µm) from 452 patients with TB (227 with drug resistance), compared clinical characteristics, and performed mycobacterial whole-genome sequencing, dormancy phenotyping and drug-susceptibility analyses on M. tuberculosis from sputum. After considering treatment duration, we found that almost half of the patients with drug-resistant TB were cough aerosol culture-positive. Surprisingly, neither mycobacterial genomic variants, lineage, nor dormancy status predicted cough aerosol culture positivity. However, mycobacterial sputum bacillary load and clinical characteristics, including a lower symptom score and stronger cough, were strongly predictive, thereby supporting targeted transmission-limiting interventions. Effective treatment largely abrogated cough aerosol culture positivity; however, this was not always rapid. These data question current paradigms, inform public health strategies and suggest the need to redirect TB transmission-associated research efforts toward host-pathogen interactions.

Extended Data Fig. 1. Study profile and participant overview

Abbreviations: CASS, cough aerosol sampling system; DS, drug-susceptible; INH, isoniazid; MDR, multidrug-resistant; MGIT, mycobacterial growth indicator tube; RIF, rifampicin; TB, tuberculosis; XDR, extensively drug-resistant; VL, viral load; WGS, whole genome sequencing. *Done using Sensititre MYCOTB plates for isoniazid, rifampicin, ethambutol, ethionamide, kanamycin, ofloxacin, p-aminosalicylic acid, rifabutin, streptomycin, cycloserine, amikacin, and moxifloxacin.

Extended Data Fig. 2. M. tuberculosis CFU from cough aerosol particles at recruitment as a function of days on treatment before CASS stratified by regimen type

Beyond eight days, no patients receiving the first-line regimen were CASS-positive, whereas patients receiving second-line regimens had CFU in their aerosol for months. The y-axis is logarithmic and one was added to CFU counts. Abbreviations: ACI, Andersen Cascade Impactor; CLF, clofazimine; FQ, fluoroquinolone; IQR, interquartile range; SLID, second-line injectable drug.

Extended Data Fig. 3

Relationships between sputum bacillary load (liquid culture time-to-positivity) and disease extent (A and B) and cavitation score (C), and between CASS-status and cavitation score (D) assessed by chest radiography using the standardised reporting system in Supplementary Table 2.

Extended Data Fig. 4

Maximum likelihood phylogenetic tree of baseline CASS isolates (n=318) rooted to Mycobacterium bovis BCG. Heatmaps to the right of the tree denote from left to right CASS status (also denoted on the branch nodes), treatment duration, drug resistance group and M. tuberculosis lineage

Extended Data Fig. 5. A Manhattan plot of multivariate genome wide association study (GWAS) using nsSNPs per gene for CASS-positivity including two principal components, patient age, TB symptom score, HIV-status, duration of treatment less than 48h, PCF, culture time to positivity, and drug-resistance category did not detect significant associations

The horizontal dashed line represents the threshold of statistical significance.

Figure 1.

A cough aerosol sampling system (CASS) was used to measure culturable M. tuberculosis in cough aerosol droplets. This system (A) consisted of (i) a six-stage Anderson Cascade Impactor [median expected droplet diameter (μm) for each stage is shown] that (ii) lies horizontally in a 10 l autoclavable chamber that (iii) patients cough into (iv) permitting CFU from individual aerosol droplets to be isolated in a size-dependent manner. (B) Most patients (n=310) were CASS-negative and a minority of CASS-positives with varied CFU readouts (1-310 CFU) were identified. (C) CFU counts were highest in the 2.1-4.7 μm particle size range (stages 3-4) (circles, left y-axis), and, on average, most patients’ total CFU came from particles 2.1-3.3 μm in diameter (stage 4) [bars (standard deviation, right y-axis]. CFU in aerosol correlated with (D) sputum smear microscopy grade and (E) Xpert MTB/RIF semi-quantitation grade [both D and E have the proportion of CASS-positive patients and median (IQR) aerosol CFU in CASS-positives indicated], and (F) these translated into moderate-to-high receiver operator characteristic curve areas under the curves (AUCs) in analyses for CASS-positivity (shown for sputum Xpert MTB/RIF-positive patients) that were improved by a clinical prediction rule. (G) CASS-positivity is shown by patient group and treatment status (all patients, treatment ≤48 h, treatment >48h). Black, grey, and white indicate the proportions of patients with >10, 1-9, or 0 aerosol CFU. P-values are for comparisons of Group 1 vs. Groups 2 and 3 within patients of the same treatment status. After adjustment for treatment status, no differences in CASS-readouts by drug susceptibility were detected (Supplementary Table 4). B, D-E: black indicates no treatment or on treatment ≤48 h, grey indicates treatment >48 h. C and D have logarithmic axes and one was added. A is adapted from . Abbreviations: AUC, area under the curve; CASS, cough aerosol sampling system; CFU, colony forming units; CI, confidence interval; IQR, interquartile range; Rif^R, rifampicin resistant. Missing data: Smear microscopy grade (n=2); Xpert MTB/RIF grade (n=46).

Figure 2.

Comparison of mycobacterial factors in CASS-positive and negative patients. (A) Manhattan plot showing that whole genome sequencing of sputum culture isolates identified no variants associated with CASS-status that met significance thresholds (n=115 CASS-positives, n=203 CASS-negatives). The top variant was in rpoB (a S531L mutation known to cause rifampicin resistance). The dashed lines above panel represent the p-value threshold for significant associations. (B) Analyses sub-stratified by patient group, treatment status, and HIV status reached similar conclusions [Supplementary Table 11 shows the top three SNPs (with and without population correction) in each stratum]. (C) A representative image of Auramine and Nile-Red strained sputum showing a mixture of Auramine-positive bacilli (green) with or without Nile-Red co-localisation (red; left panel) and a dot plot (mean, standard deviation) of the proportion of Auramine-positive bacilli also Nile-Red positive (Nile-Red results are also in the Supplement).

Figure 3.

Repeat cough aerosol sampling results in patients initially CASS-positive who were re-sampled until CASS-negative. Days on treatment for the current episode are shown for all patients except the one, six and three in Groups 1, 2, and 3, respectively who were not on any TB treatment between CASS readings (dashed lines). Most Group 1 (A) and all adherent Group 2 (B) patients were CASS-negative at repeat sampling after treatment and many Group 3 patients were repeatedly CASS-positive (C), however, the Group 3 patients they were on a low number of likely effective drugs (Supplementary Tables 16-17). Each line represents a patient and each symbol a sampling event (CASS and sputum microbiology). Microbiology (microscopy, Xpert MTB/RIF, culture) from paired sputa, shown below each graph, and continued to have poor diagnostic accuracy for CASS-positivity. All patients on treatment, irrespective of the treatment’s likely effectiveness, are shown. Patients CASS-positive more than once are shown as thick black lines (lines in grey represent patients only CASS-positive at baseline). Diamonds in (C) indicate patients discharged with programmatically uncured drug-resistant TB (three baseline CASS-positives of this type, indicated by a dashed line, who were not on any treatment were CASS-negative at re-testing). Vertical dashed lines indicate the median days to the second and third CASS. Y-axes are logarithmic and one was added to CFU counts.