COVID-19 on the Nile: a cross-sectional investigation of COVID-19 among Nile River cruise travellers returning to the United States, February-March 2020.

BACKGROUND: Early in the pandemic, cruise travel exacerbated the global spread of SARS-CoV-2. We report epidemiologic and molecular findings from an investigation of a cluster of travellers with confirmed COVID-19 returning to the USA from Nile River cruises in Egypt. METHODS: State health departments reported data on real-time reverse transcription-polymerase chain reaction-confirmed COVID-19 cases with a history of Nile River cruise travel during February-March 2020 to the Centers for Disease Control and Prevention (CDC). Demographic and epidemiologic data were collected through routine surveillance channels. Sequences were obtained either from state health departments or from the Global Initiative on Sharing Avian Flu Data (GISAID). We conducted descriptive analyses of epidemiologic data and explored phylogenetic relationships between sequences. RESULTS: We identified 149 Nile River cruise travellers with confirmed COVID-19 who returned to 67 different US counties in 27 states: among those with complete data, 4.7% (6/128) died and 28.1% (38/135) were hospitalized. These individuals travelled on 20 different Nile River cruise voyages (12 unique vessels). Fifteen community transmission events were identified in four states, with 73.3% (11/15) of these occurring in Wisconsin (as the result of a more detailed contact investigation in that state). Phylogenetic analyses supported the hypothesis that travellers were most likely infected in Egypt, with most sequences in Nextstrain clade 20A 93% (87/94). We observed genetic clustering by Nile River cruise voyage and vessel. CONCLUSIONS: Nile River cruise travellers with COVID-19 introduced SARS-CoV-2 over a very large geographic range, facilitating transmission across the USA early in the pandemic. Travellers who participate in cruises, even on small river vessels as investigated in this study, are at increased risk of SARS-CoV-2 exposure. Therefore, history of river cruise travel should be considered in contact tracing and outbreak investigations.

Molecular surveillance for large outbreaks of tuberculosis in the United States, 2014-2018.

OBJECTIVE: This study describes characteristics of large tuberculosis (TB) outbreaks in the United States detected using novel molecular surveillance methods during 2014-2016 and followed for 2 years through 2018. METHODS: We developed 4 genotype-based detection algorithms to identify large TB outbreaks of ≥10 cases related by recent transmission during a 3-year period. We used whole-genome sequencing and epidemiologic data to assess evidence of recent transmission among cases. RESULTS: There were 24 large outbreaks involving 518 cases; patients were primarily U.S.-born (85.1%) racial/ethnic minorities (84.1%). Compared with all other TB patients, patients associated with large outbreaks were more likely to report substance use, homelessness, and having been diagnosed while incarcerated. Most large outbreaks primarily occurred within residences among families and nonfamilial social contacts. A source case with a prolonged infectious period and difficulties in eliciting contacts were commonly reported contributors to transmission. CONCLUSION: Large outbreak surveillance can inform targeted interventions to decrease outbreak-associated TB morbidity.

Using Machine Learning Techniques and National Tuberculosis Surveillance Data to Predict Excess Growth in Genotyped Tuberculosis Clusters.

The early identification of clusters of persons with tuberculosis (TB) that will grow to become outbreaks creates an opportunity for intervention in preventing future TB cases. We used surveillance data (2009-2018) from the United States, statistically derived definitions of unexpected growth, and machine-learning techniques to predict which clusters of genotype-matched TB cases are most likely to continue accumulating cases above expected growth within a 1-year follow-up period. We developed a model to predict which clusters are likely to grow on a training and testing data set that was generalizable to a validation data set. Our model showed that characteristics of clusters were more important than the social, demographic, and clinical characteristics of the patients in those clusters. For instance, the time between cases before unexpected growth was identified as the most important of our predictors. A faster accumulation of cases increased the probability of excess growth being predicted during the follow-up period. We have demonstrated that combining the characteristics of clusters and cases with machine learning can add to existing tools to help prioritize which clusters may benefit most from public health interventions. For example, consideration of an entire cluster, not only an individual patient, may assist in interrupting ongoing transmission.

Genomic Surveillance for SARS-CoV-2 Variants: Predominance of the Delta (B.1.617.2) and Omicron (B.1.1.529) Variants - United States, June 2021-January 2022.

Genomic surveillance is a critical tool for tracking emerging variants of SARS-CoV-2 (the virus that causes COVID-19), which can exhibit characteristics that potentially affect public health and clinical interventions, including increased transmissibility, illness severity, and capacity for immune escape. During June 2021-January 2022, CDC expanded genomic surveillance data sources to incorporate sequence data from public repositories to produce weighted estimates of variant proportions at the jurisdiction level and refined analytic methods to enhance the timeliness and accuracy of national and regional variant proportion estimates. These changes also allowed for more comprehensive variant proportion estimation at the jurisdictional level (i.e., U.S. state, district, territory, and freely associated state). The data in this report are a summary of findings of recent proportions of circulating variants that are updated weekly on CDC's COVID Data Tracker website to enable timely public health action.† The SARS-CoV-2 Delta (B.1.617.2 and AY sublineages) variant rose from 1% to >50% of viral lineages circulating nationally during 8 weeks, from May 1-June 26, 2021. Delta-associated infections remained predominant until being rapidly overtaken by infections associated with the Omicron (B.1.1.529 and BA sublineages) variant in December 2021, when Omicron increased from 1% to >50% of circulating viral lineages during a 2-week period. As of the week ending January 22, 2022, Omicron was estimated to account for 99.2% (95% CI = 99.0%-99.5%) of SARS-CoV-2 infections nationwide, and Delta for 0.7% (95% CI = 0.5%-1.0%). The dynamic landscape of SARS-CoV-2 variants in 2021, including Delta- and Omicron-driven resurgences of SARS-CoV-2 transmission across the United States, underscores the importance of robust genomic surveillance efforts to inform public health planning and practice.

Logically Inferred Tuberculosis Transmission (LITT): A Data Integration Algorithm to Rank Potential Source Cases.

Understanding tuberculosis (TB) transmission chains can help public health staff target their resources to prevent further transmission, but currently there are few tools to automate this process. We have developed the Logically Inferred Tuberculosis Transmission (LITT) algorithm to systematize the integration and analysis of whole-genome sequencing, clinical, and epidemiological data. Based on the work typically performed by hand during a cluster investigation, LITT identifies and ranks potential source cases for each case in a TB cluster. We evaluated LITT using a diverse dataset of 534 cases in 56 clusters (size range: 2-69 cases), which were investigated locally in three different U.S. jurisdictions. Investigators and LITT agreed on the most likely source case for 145 (80%) of 181 cases. By reviewing discrepancies, we found that many of the remaining differences resulted from errors in the dataset used for the LITT algorithm. In addition, we developed a graphical user interface, user's manual, and training resources to improve LITT accessibility for frontline staff. While LITT cannot replace thorough field investigation, the algorithm can help investigators systematically analyze and interpret complex data over the course of a TB cluster investigation. Code available at: https://github.com/CDCgov/TB_molecular_epidemiology/tree/1.0; https://zenodo.org/badge/latestdoi/166261171.

Mutation of Mycobacterium tuberculosis and Implications for Using Whole-Genome Sequencing for Investigating Recent Tuberculosis Transmission.

Tuberculosis (TB) control programs use whole-genome sequencing (WGS) of Mycobacterium tuberculosis (Mtb) for detecting and investigating TB case clusters. Existence of few genomic differences between Mtb isolates might indicate TB cases are the result of recent transmission. However, the variable and sometimes long duration of latent infection, combined with uncertainty in the Mtb mutation rate during latency, can complicate interpretation of WGS results. To estimate the association between infection duration and single nucleotide polymorphism (SNP) accumulation in the Mtb genome, we first analyzed pairwise SNP differences among TB cases from Los Angeles County, California, with strong epidemiologic links. We found that SNP distance alone was insufficient for concluding that cases are linked through recent transmission. Second, we describe a well-characterized cluster of TB cases in California to illustrate the role of genomic data in conclusions regarding recent transmission. Longer presumed latent periods were inconsistently associated with larger SNP differences. Our analyses suggest that WGS alone cannot be used to definitively determine that a case is attributable to recent transmission. Methods for integrating clinical, epidemiologic, and genomic data can guide conclusions regarding the likelihood of recent transmission, providing local public health practitioners with better tools for monitoring and investigating TB transmission.

Increased H. pylori stool shedding and EPIYA-D cagA alleles are associated with gastric cancer in an East Asian hospital.

BACKGROUND: Helicobacter pylori infection increases risk for gastric cancer. Geographic variation in gastric cancer risk has been attributed to variation in carriage and type of the H. pylori oncogene cagA. Colonization density may also influence disease and cagA has been associated with higher shedding in stool. However, the relationship between H. pylori load in the stool and in the stomach is not clear. METHODS: To investigate possible differences in H. pylori load in the stomach and shedding in stool, H. pylori load and cagA genotype were assessed using droplet digital PCR assays on gastric mucosa and stool samples from 49 urea breath test-positive individuals, including 25 gastric cancer and 24 non-cancer subjects at Henan Cancer Hospital, Henan, China. RESULTS: Quantitation of H. pylori DNA indicated similar gastric loads among cancer and non-cancer cases, but the gastric cancer group had a median H. pylori load in the stool that was six times higher than that of the non-cancer subjects. While the cagA gene was uniformly present among study subjects, only 70% had the East Asian cagA allele, which was significantly associated with gastric cancer (Fisher's Exact Test, p = 0.03). CONCLUSION: H. pylori persists in a subset of gastric cancer cases and thus may contribute to cancer progression. In this East Asian population with a high prevalence of the cagA gene, the East Asian allele could still provide a marker for gastric cancer risk. IMPACT: This study contributes to our understanding of H. pylori dynamics in the context of pathological changes.

Droplet Digital PCR-Based Detection of Clarithromycin Resistance in Helicobacter pylori Isolates Reveals Frequent Heteroresistance.

Chronic infection with Helicobacter pylori causes peptic ulcers and stomach cancer in a subset of infected individuals. While standard eradication therapy includes multiple antibiotics, treatment failure due to resistance is an increasing clinical problem. Accurate assessment of H. pylori antimicrobial resistance has been limited by slow growth and sampling of few isolates per subject. We established a method to simultaneously quantify H. pylori clarithromycin-resistant (mutant) and -susceptible (wild-type) 23S rRNA gene alleles in both stomach and stool samples using droplet digital PCR (ddPCR). In 49 subjects, we assessed the performance of these assays alongside clarithromycin MIC testing of up to 16 H. pylori isolates per subject and included both cancer (25 subjects) and noncancer (24 subjects) cases. Gastric ddPCR and H. pylori culture showed agreement with urea breath test (UBT) detection of infection in 94% and 88% of subjects, respectively, while stool ddPCR showed agreement with UBT in 92% of subjects. Based on MIC testing of 43 culture-positive cases, 20 subjects had only susceptible isolates, 14 had a mix of susceptible and resistant isolates, and 9 had only resistant isolates. ddPCR of gastric samples indicated that 21 subjects had only wild-type alleles, 13 had a mixed genotype, and 9 had only mutant alleles. Stool ddPCR detected mutant alleles in four subjects for which mutant alleles were not detected by stomach ddPCR, and no resistant isolates were cultured. Our results indicate that ddPCR detects H. pylori clarithromycin resistance-associated genotypes, especially in the context of heteroresistance.

High prevalence of Helicobacter pylori clarithromycin resistance mutations among Seattle patients measured by droplet digital PCR.

BACKGROUND: Treatment of Helicobacter pylori infection is often empiric; however, current guidelines for management of Helicobacter pylori infection advise against the use of standard triple therapy (clarithromycin, amoxicillin, and proton-pump inhibitor) when clarithromycin resistance exceeds 20%. We developed and tested a new culture-free assay to detect clarithromycin resistance-conferring mutations to determine the prevalence of H. pylori clarithromycin resistance in patients from the United States Pacific Northwest. MATERIALS AND METHODS: Droplet digital PCR (ddPCR) was used to detect the H. pylori 23S rRNA gene, and resistance-conferring mutations, in archived, formalin-fixed, paraffin-embedded (FFPE) gastric tissue and to retrospectively determine the prevalence of clarithromycin-resistant H. pylori among 110 patients at an academic medical center in the Northwest United States between 2012 and 2014. RESULTS: Of 102 patients with the H. pylori 23S rRNA gene detected by the ddPCR assay, 45 (44%) had clarithromycin resistance mutations. Thirty-three of the 45 patients with clarithromycin resistance mutations had a mix of wild-type and resistance alleles. Prevalence of clarithromycin resistance mutations differed among racial groups and was highest among Asians, with mutations detected in 14 (67%) of the 21 patient samples. CONCLUSIONS: The prevalence of clarithromycin resistance detected in this region exceeds 20%, indicating that standard triple therapy should not be the first-line antibiotic treatment for H. pylori infection. Culture-free assays for detecting clarithromycin resistance mutations can be performed on archived tissue samples and will aid in informing tailored treatment for effective H. pylori eradication.

Quantitative Detection and Genotyping of Helicobacter pylori from Stool using Droplet Digital PCR Reveals Variation in Bacterial Loads that Correlates with cagA Virulence Gene Carriage.

BACKGROUND: Epidemiologic studies of the carcinogenic stomach bacterium Helicobacter pylori have been limited by the lack of noninvasive detection and genotyping methods. We developed a new stool-based method for detection, quantification, and partial genotyping of H. pylori using droplet digital PCR (ddPCR), which allows for increased sensitivity and absolute quantification by PCR partitioning. MATERIALS AND METHODS: Stool-based ddPCR assays for H. pylori 16S gene detection and cagA virulence gene typing were tested using a collection of 50 matched stool and serum samples from Costa Rican volunteers and 29 H. pylori stool antigen-tested stool samples collected at a US hospital. RESULTS: The stool-based H. pylori 16S ddPCR assay had a sensitivity of 84% and 100% and a specificity of 100% and 71% compared to serology and stool antigen tests, respectively. The stool-based cagA genotyping assay detected cagA in 22 (88%) of 25 stools from CagA antibody-positive individuals and four (16%) of 25 stools from CagA antibody-negative individuals from Costa Rica. All 26 of these samples had a Western-type cagA allele. Presence of serum CagA antibodies was correlated with a significantly higher load of H. pylori in the stool. CONCLUSIONS: The stool-based ddPCR assays are a sensitive, noninvasive method for detection, quantification, and partial genotyping of H. pylori. The quantitative nature of ddPCR-based H. pylori detection revealed significant variation in bacterial load among individuals that correlates with presence of the cagA virulence gene. These stool-based ddPCR assays will facilitate future population-based epidemiologic studies of this important human pathogen.

Regulation of Helicobacter pylori adherence by gene conversion.

Genetic diversification of Helicobacter pylori adhesin genes may allow adaptation of adherence properties to facilitate persistence despite host defences. The sabA gene encodes an adhesin that binds sialyl-Lewis antigens on inflamed gastric tissue. We found variability in the copy number and locus of the sabA gene and the closely related sabB and omp27 genes due to gene conversion among 51 North American paediatric H. pylori strains. We determined that sabB to sabA gene conversion is predominantly the result of intra-genomic recombination and RecA, RecG and AddA influence the rate at which it occurs. Although all clinical strains had at least one sabA gene copy, sabA and sabB were lost due to gene conversion at similar rates in vitro, suggesting host selection to maintain the sabA gene. sabA gene duplication resulted in increased SabA protein production and increased adherence to sialyl-Lewis antigens and mouse gastric tissue. In conclusion, gene conversion is a mechanism for H. pylori to regulate sabA expression level and adherence.

Identification of factors for tuberculosis transmission via an integrated multidisciplinary approach.

It was reported previously that the major fraction of the recent decrease of tuberculosis incident cases in Arkansas had been due to a decrease in the reactivated infections. Preventing transmission of Mycobacterium tuberculosis is the key to a continued decline in tuberculosis cases. In this study, we integrated epidemiological data analysis and comparative genomics to identify host and microbial factors important to tuberculosis transmission. A significantly higher proportion of cases in large clusters (containing >10 cases) were non-Hispanic black, homeless, less than 65 years old, male sex, smear-positive sputum, excessive use of alcohol, and HIV sero-positive, compared to cases in small clusters (containing 2-5 cases) diagnosed within one year. However, being non-Hispanic black and homeless within the past year were the only two host characteristics that were identified as independent risk factors for being in large clusters. This finding suggests that social behavioral factors have a more important role in transmission of tuberculosis than does the infectiousness of the source. Comparing the genomic content of one of the large cluster strains to that of a non-clustered strain from the same community identified 25 genes that differed between the two strains, potentially contributing to the observed differences in transmission.

How dormant is Mycobacterium tuberculosis during latency? A study integrating genomics and molecular epidemiology.

Mycobacterium tuberculosis may survive for decades in the human body in a state termed latent tuberculosis infection (LTBI). We investigated the occurrence during LTBI of insertion/deletion events in a selected set of mononucleotide simple sequence repeats, DNA sequence changes in four M. tuberculosis genes, and large sequence variations in 4750 M. tuberculosis open reading frames. We studied 13 paired M. tuberculosis clinical isolates, with each pair representing a reactivation of LTBI more than three decades after primary infection. Absence of sequence variations between paired isolates in nearly all investigated loci suggests a low likelihood of bacterial replication during LTBI.

Helicobacter pylori's unconventional role in health and disease.

The discovery of a bacterium, Helicobacter pylori, that is resident in the human stomach and causes chronic disease (peptic ulcer and gastric cancer) was radical on many levels. Whereas the mouth and the colon were both known to host a large number of microorganisms, collectively referred to as the microbiome, the stomach was thought to be a virtual Sahara desert for microbes because of its high acidity. We now know that H. pylori is one of many species of bacteria that live in the stomach, although H. pylori seems to dominate this community. H. pylori does not behave as a classical bacterial pathogen: disease is not solely mediated by production of toxins, although certain H. pylori genes, including those that encode exotoxins, increase the risk of disease development. Instead, disease seems to result from a complex interaction between the bacterium, the host, and the environment. Furthermore, H. pylori was the first bacterium observed to behave as a carcinogen. The innate and adaptive immune defenses of the host, combined with factors in the environment of the stomach, apparently drive a continuously high rate of genomic variation in H. pylori. Studies of this genetic diversity in strains isolated from various locations across the globe show that H. pylori has coevolved with humans throughout our history. This long association has given rise not only to disease, but also to possible protective effects, particularly with respect to diseases of the esophagus. Given this complex relationship with human health, eradication of H. pylori in nonsymptomatic individuals may not be the best course of action. The story of H. pylori teaches us to look more deeply at our resident microbiome and the complexity of its interactions, both in this complex population and within our own tissues, to gain a better understanding of health and disease.

Pediatric Helicobacter pylori isolates display distinct gene coding capacities and virulence gene marker profiles.

Helicobacter pylori strains display remarkable genetic diversity, and the presence of strains bearing the toxigenic vacA s1 allele, a complete cag pathogenicity island (PAI), cagA alleles containing multiple EPIYA phosphorylation sites, and expressing the BabA adhesin correlates with development of gastroduodenal disease in adults. To better understand the genetic variability present among pediatric strains and its relationship to disease, we characterized H. pylori strains infecting 47 pediatric North American patients. Prevalence of mixed infection was assessed by random amplified polymorphic DNA analysis of multiple H. pylori clones from each patient. Microarray-based comparative genomic hybridization was used to examine the genomic content of the pediatric strains. The cagA and vacA alleles were further characterized by allele-specific PCR. A range of EPIYA motif configurations were observed for the cagA gene, which was present in strains from 22 patients (47%), but only 19 (41%) patients contained a complete cag PAI. Thirty patients (64%) were infected with a strain having the vacA s1 allele, and 28 patients (60%) had the babA gene. The presence of a functional cag PAI was correlated with ulcer disease (P = 0.0095). In spite of declining rates of H. pylori infection in North America, at least 11% of patients had mixed infection. Pediatric strains differ in their spectrum of strain-variable genes and percentage of absent genes in comparison to adult strains. Most children were infected with H. pylori strains lacking the cag PAI, but the presence of a complete cag PAI, in contrast to other virulence markers, was associated with more severe gastroduodenal disease.

Does the lipR gene of tubercle bacilli have a role in tuberculosis transmission and pathogenesis?

Mycobacterium tuberculosis lipases, a diverse class of enzymes involved in lipid metabolism, may have an important role in tuberculosis (TB) pathogenesis. We explored the association of large sequence polymorphism (LSP) in one of the M. tuberculosis lipase-encoding genes, lipR (Rv3084), with patient characteristics using a population-based sample of clinical isolates to elucidate the potential role of lipR in TB pathogenesis. LSP in lipR was found in 104 (15.6%) of 665 isolates, of which 96% belonged to principal genetic group 3. When linkage by molecular type and epidemiologic evidence were compared, molecularly clustered cases infected with a lipR LSP isolate were more often epidemiologically linked than clustered cases infected with a lipR wild-type isolate. Further epidemiologic and functional studies are necessary to determine if the association between this lipR LSP and recent transmission we identified in this population reflects a functional role of lipR in TB transmission and pathogenesis or other unidentified mechanisms.

Mycobacterium tuberculosis PE_PGRS16 and PE_PGRS26 genetic polymorphism among clinical isolates.

The Mycobacterium tuberculosis PE_PGRS multigene family is thought to be involved in antigenic variation, which can be generated by differential regulation of expression and a high frequency of genetic polymorphism. PE_PGRS16 and PE_PGRS26 are inversely regulated during persistent M. tuberculosis infection, suggesting that differential regulation of the expression of these two PE_PGRS genes may have a role in latency. To understand how genetic diversity, in addition to differential regulation, contributes to antigenic variability, we investigated the sequence variations in the PE_PGRS16 and PE_PGRS26 genes among 200 clinical M. tuberculosis strains, in comparison to the sequenced laboratory strain H37Rv, using PCR and DNA sequencing. Among the 200 strains, 102 (51%) and 100 (50%) had sequence variations within the PE_PGRS16 gene and the PE_PGRS26 gene, respectively. In-frame insertions and deletions, frameshifts, and SNPs were observed in both the PE_PGRS16 gene and the PE_PGRS26 gene. However, the frequency of frameshifts and in-frame deletions differed between the two PE_PGRS genes. Examining the profile of the PE_PGRS16, PE_PGRS26, and the previously investigated PE_PGRS33 amino acid sequences for each of the 200 strains, 72 different profiles were observed with frequencies ranging from 0.5% to 13%. In conclusion, a remarkable level of genetic diversity exists in the PE_PGRS16 and PE_PGRS26 genes of M. tuberculosis clinical strains. The significant sequence variations in the two PE_PGRS genes observed in this study could impact the function of these two PE_PGRS proteins and be associated with differences in the ability of the tubercle bacilli to remain persistent within the host.

DNA polymorphisms in the pepA and PPE18 genes among clinical strains of Mycobacterium tuberculosis: implications for vaccine efficacy.

Tuberculosis continues to be a leading cause of death worldwide. Development of an effective vaccine against Mycobacterium tuberculosis is necessary to reduce the global burden of this disease. Mtb72F, consisting of the protein products of the pepA and PPE18 genes, is the first subunit tuberculosis vaccine to undergo phase I clinical trials. To obtain insight into the ability of Mtb72F to induce an immune response capable of recognizing different strains of M. tuberculosis, we investigated the genomic diversity of the pepA and PPE18 genes among 225 clinical strains of M. tuberculosis from two different geographical locations, Arkansas and Turkey, representing a broad range of genotypes of M. tuberculosis. A combination of single nucleotide polymorphisms (SNPs) and insertion/deletions resulting in amino acid changes in the PPE18 protein occurred in 47 (20.9%) of the 225 study strains, whereas SNPs resulted in amino acid changes in the PepA protein in 14 (6.2%) of the 225 study strains. Of the 122 Arkansas study strains and the 103 Turkey study strains, 32 (26.2%) and 15 (14.6%), respectively, had at least one genetic change leading to an alteration of the amino acid sequence of the PPE18 protein, and many of the changes occurred in regions previously reported to be potential T-cell epitopes. Thus, immunity induced by Mtb72F may not recognize a proportion of M. tuberculosis clinical strains.

Association of Mycobacterium tuberculosis PE PGRS33 polymorphism with clinical and epidemiological characteristics.

There is evidence that some members of the Mycobacterium tuberculosis PE PGRS gene subfamily, including PE PGRS33, may have a specific function in M. tuberculosis persistence. The impact of naturally-occurring PE PGRS33 genetic variations on the virulence and transmissibility of clinical M. tuberculosis isolates is not known. We used PCR and DNA sequencing to identify genetic variations in the PE PGRS33 gene in comparison with the sequenced laboratory strain, H37Rv, among 649 isolates from a population-based sample. The PE PGRS33 alleles were placed into two groups, based on the effect of the sequence variations on the PE PGRS33 protein, and their associations with clinical and epidemiological characteristics were assessed using multivariate logistic regression to control for potential confounding of host-related factors. Of the 639 isolates for which sequence data were obtained, 139 (21.8%) had PE PGRS33 alleles that would result in a significant change to the PE PGRS33 protein due to large insertions/deletions or frameshift mutations. These isolates were significantly associated with clustering based on genotype and absence of cavitations in the lungs, compared to isolates having PE PGRS33 alleles that would result in no or minimal change to the PE PGRS33 protein. The association of significant changes to PE PGRS33 with clinical and epidemiological characteristics suggests that PE PGRS33 may have an important role in M. tuberculosis persistence.