Ethics statement
The study protocol was approved by the Ethics Committee of the Institute of Quzhou City Center for Disease Control and Prevention (Approval NO:IRB-2020-R-NO.001). All participants provided informed consent for both the active screening and subsequent follow-ups. For those who were illiterate or could not read, the research staff would read the content of the informed consent to him/her, and then asked the family member to sign the consent form as an agent.
Study setting and population
Quzhou City is located in the western region of Zhejiang Province and comprises six counties (cities, districts) with 100 towns. In December 2019, there were 377,205 permanent residents aged ≥65 years, of whom 311,385 were rural residents. The reported incidence rate of pulmonary tuberculosis in the city was 67 cases per 100,000 person-years in 2019; among those aged ≥65 years, it was 194 cases per 100,000 person-years. The proportion of all pulmonary tuberculosis patients that are elderly in the city reached 41%25. Outmigration is low in Quzhou, especially among the elderly population.
Quzhou City has six designated tuberculosis laboratories (one at the prefectural level and five at the county level), all of which are BSL-2 level facilities capable of performing sputum smear microscopy, molecular testing, MGIT-320 culture, and drug susceptibility testing. Quality control in these laboratories is in accordance with guideline requirements, with a Mycobacterium tuberculosis culture contamination rate of less than 10%. Since 2015, all laboratories have successfully participated in 10 rounds of proficiency testing organized by the National Tuberculosis Reference Laboratory of the Chinese Center for Disease Control and Prevention for TB molecular diagnostics and drug susceptibility testing, with all results being satisfactory.
Study design
Active screening period
From March to October 2020, Quzhou City conducted a large-scale campaign involving active screening of all residents aged 65 years or older in the rural areas. The active tuberculosis screening intervention among the elderly was led by health administrative departments at the county level, organized and launched by township people’s governments, and supervised by county-level medical institutions. This mass screening involved a comprehensive physical examination and chest X-rays on all enrollees26. The majority of chest X-ray radiographic examinations for tuberculosis were scheduled at township health centers, as most of these centers have radiographic equipment. Among the 89 township health centers, seven that lacked on-site chest digital radiography (DR) equipment utilized mobile chest DR units for screening. Screening for tuberculosis was scheduled to coincide with the residents’ annual physical examination. At township health centers, targeted individuals underwent general health examinations and chest X-ray screenings27; images were uploaded to the local county hospital through the Picture Archiving and Communication System (PACS) for radiological interpretation, with the results subsequently fed back to the township health centers. A diagnostic panel composed of three members (two infectious disease specialists and one radiologist) was responsible for identifying abnormalities in participants’ chest radiographs and classifying the radiographic diagnoses into normal or abnormal28. Abnormal signs were further categorized into active lesions, stable lesions, and indeterminate lesions. Stable signs were subclassified into fibrotic lesions, sclerotic lesions, and calcified lesions29,30,31, while uncertain signs were subclassified into patchy and non-patchy lesions30,31. For abnormal classification, diagnostic results consensus from at least two specialists; disagreements were resolved through discussion until consensus was reached. The diagnostic panel included qualified licensed physicians, with at least one being a certified attending hospital physician (or more senior in terms of experience). All panelists received training from the Quzhou City Quality Control Center for Tuberculosis Diagnosis and Treatment. During on-site quality control, experts randomly reviewed a certain percentage of X-ray images; a concordance rate of over 90% between the evaluated image decisions and the quality control team was deemed acceptable. For individuals with active lesions, referral to designated tuberculosis hospitals for TB diagnostic tests was made to confirm or rule out active pulmonary tuberculosis. Those without active lesions were advised to seek tuberculosis diagnosis at designated hospitals if tuberculosis-related symptoms developed. In asymptomatic individuals with a chest X-ray result indicating an abnormality, referral to an appropriate healthcare facility for further evaluation and management was recommended to evaluate for other serious conditions such as lung cancer or pulmonary fibrosis. Pulmonary tuberculosis diagnosis primarily relied on pathogen detection (including bacteriological and molecular biological methods), combined with epidemiological history, clinical presentation, chest imaging, relevant ancillary tests, and differential diagnosis for comprehensive analysis. Bacteriological examinations included three sputum samples for Mycobacterium tuberculosis smear microscopy (one immediate, one morning, and one night sample), liquid culture of one sputum sample, and one GeneXpert MTB/RIF test24. When sputum volume or quality was inadequate, artificial sputum induction was performed, and bronchoalveolar lavage was used when necessary to obtain sputum samples.
Screening costs were fully covered by the government as part of the Quzhou Municipal Government’s Civil Welfare Project from 2020 to 2022. A specific tuberculosis screening subsidy of 55 yuan per person (1 U.S. dollar is equivalent to 6.5 RMB at the time of the study) was provided, with 35 yuan allocated for chest X-ray costs, 15 yuan for promotional activities, organizational expenses, transportation, and follow-up, and 5 yuan reserved for subsequent diagnostic tests for those identified with active pulmonary lesions. The costs for residents’ health examinations were covered by the urban and rural health insurance funds.
Follow-up period
Follow-up was conducted quarterly by trained community health professionals through on-site visits during health service provision or via telephone interviews. Follow-up visits consisted of inquiries about tuberculosis-related symptoms, educational activities on prevention and treatment, and general health services. An end visit of 2 years (from 1 November 2020 to 31 October 2022) was scheduled for all participants unless they developed TB during follow-up, refused follow-up, or died. Individuals with active lesions during the baseline radiological assessment who were not diagnosed with pulmonary tuberculosis, were instructed to participate in further annual health check-ups and were referred to designated hospitals for tuberculosis diagnostic testing to confirm or rule out tuberculosis. For participants with baseline X-ray findings not suggestive of tuberculosis, were instructed to self-refer to designated hospitals for tuberculosis diagnostic testing if they developed tuberculosis-related symptoms during follow-up. In addition to field visits and telephone interviews, we utilized the local Tuberculosis Management Information System (TBIMS) to capture data on confirmed tuberculosis individuals, thereby supplementing information on unvisited individuals. Confirmed patients were treated according to the standard protocol for active tuberculosis. Participants not diagnosed with tuberculosis were not prescribed preventive treatment, regardless of chest X-ray status, following Zhejiang Provincial and Chinese national guidelines. All tuberculosis patients are managed through the TBIMS maintained by the National Center for Tuberculosis Control and Prevention10. Information regarding death (including timing) was collected from the Cause of Death Registration Reporting Information System maintained by the Zhejiang Provincial Center for Disease Control and Prevention.
Relevant definitions
Pulmonary tuberculosis, according to the People’s Republic of China Health Industry Standard WS 288-2017 for Diagnosis of Pulmonary Tuberculosis24, refers to tuberculous lesions in the lung tissue, trachea, bronchi, and pleura. This includes suspected individuals with tuberculosis, clinically diagnosed patients, and confirmed patients. The diagnosis of pulmonary tuberculosis primarily relies on pathogen testing (including bacteriology and molecular biology), combined with epidemiological history, clinical presentation, chest imaging, related ancillary tests, and differential diagnosis, for a comprehensive analysis leading to a diagnosis. Pathogen and pathology results serve as the basis for confirmation. Confirmed patients tested microbiologically positive, which includes sputum smear positivity, and/or culture positivity, GeneXpert MTB/RIF positivity, and/or pathology positivity in lung tissue specimens. Clinically diagnosed patients were microbiologically-negative and diagnosed by a group of clinicians based on several relevant clinical characteristics which may include (but are not limited to) epidemiological history, clinical presentation infection testing results, tuberculosis exposure history, and chest imaging results. Subclinical tuberculosis was a condition caused by viable Mycobacterium tuberculosis that did not produce the symptoms associated with clinical tuberculosis but could lead to other abnormalities detectable by current radiological or microbiological testing methods. Tuberculosis-related symptoms included cough, sputum production, hemoptysis, fever, night sweats, weight loss, fatigue, or chest pain.
Active signs: including multiple nodular lesions, patchy, cloudy flocculent and lobar lung consolidations, mass-like opacities, and enlarged hilar or mediastinal lymph nodes. These lesions are featured by heterogeneous density, high central density, peripheral low density, irregular distribution, and infiltrative changes. They may be accompanied by thick-walled, thin-walled, tension cavities, and multiple moth-eaten cavities, as well as satellite lesions, bronchial dissemination, bronchial dilation, lymphangitis, and pleural effusion. Stable signs: including dense nodules, plaques, calcified nodules, fibrous strands and residual purification cavities after treatment. These lesions have clear and sharp margins and may accompany pleural and/or mediastinal lymph node calcifications. Dense nodules and plaques are classified as sclerosing lesions, fibrous striated lesions are classified as fibrotic lesions. Calcified nodules and residual purgative cavities after treatment are classified as calcified lesions. Uncertain signs: including patchy lesions and non-patchy lesions (destroyed lungs, atelectasis, tuberculomas) and other lesions that have not yet been fully calcified, cannot make a diagnosis of inactive pulmonary tuberculosis based on these signs alone; further analysis after obtaining computed tomography scans is required28.
Prior tuberculosis was defined as participants with tuberculosis treatment history registered through TBIMS from January 1, 2005 to December 31, 2019. Prevalence cases were defined as those identified either actively or passively among participants during the screening period. Individuals with tuberculosis found before the screening period were not included in this study. A one-month period was allowed for referral and diagnosis following screening, thus including prevalent cases found within one month after the last screening date. Therefore, screening took place from March to September, and all cases found during the time period of March to October were defined as prevalent. Incident cases were those identified and detected either actively or passively during the follow-up period. The follow-up was during the two years post-completion of the screening period. This was from 1 November 2020 to 31 October 2022. Tuberculosis patients occurred exclusively among those who participated in the baseline screening; data from follow-up years (year 1, year 2) did not include individuals who did not partake in the baseline screening.
Elderly people were participants 65 years of age or over. Body mass index (BMI) was categorized into subgroups relevant to Asian populations, including 32,33.
Statistical analysis
Data was described in terms of proportions and case rates. Tuberculosis risk was analyzed by specified time periods as well as chest X-ray classification from the baseline screening visit. Time periods were divided based on initial screening (March 1 to October 31, 2020), first year of follow-up (November 1, 2020 to October 31, 2021), and second year of follow-up (November 1, 2021, to October 31, 2022). The second year of follow-up was specified to account for any unconfirmed early tuberculosis patients that may have been diagnosed in year 1 of follow-up. This time period began > 1 year after completion of screening. All analyses, except for dividing the follow-up time period into the first and second year, were pre-specified.
R version 4.2.0 was used for all analyses. Binomial logistic regression models were used to estimate the prevalence of tuberculosis among participants with different characteristics during the screening process. For the two distinct follow-up time period after screening, time-to-event data were constructed between the first date of the follow-up time period and the date of development of tuberculosis. Follow-up was censored at death, development of tuberculosis, or end of follow-up. We compared tuberculosis incidence in groups with differing chest X-ray status using hazard ratios (HRs) and 95% confidence intervals (CIs) obtained from Cox proportional hazard models14,34. We used Kaplan-Meier analyses to present the results of participants with distinct chest X-ray classifications developing tuberculosis at different follow-up stages. A two-sample likelihood ratio test was used.
We calculated confirmed active tuberculosis incidence in cases per 100,000 person-years for participants with distinct baseline chest X-ray findings. 95% Poisson CIs were calculated around these estimates and tuberculosis incidence rates were compared using two-sample Poisson tests.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
