If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Tan Tock Seng Hospital, 11 Jalan Tan Tock Seng, Singapore 308433Saw Swee Hock School of Public Health, National University of Singapore and National University Health System; 12 Science Drive 2, Singapore 117649
Ministry of Health Singapore, College of Medicine Building, 16 College Road, Singapore 169854Saw Swee Hock School of Public Health, National University of Singapore and National University Health System; 12 Science Drive 2, Singapore 117649
First published attempt to screen for TB among all residents in an apartment block.
No clear transmission patterns among 3 of 6 phylogenetically similar MDR-TB cases.
In densely populated cities, whole genome sequencing is useful to link cases.
Mass screening for TB is limited by participation and interpretation of results.
Between February 2012 and May 2016, six residents of an 11-storey apartment block were diagnosed with MDR-TB. Based on initial tests, all isolates had similar genotypic profiles, although there were no identifiable epidemiological transmission patterns between three cases. We present findings from the cluster investigation and results of a mass screening exercise.
Free voluntary TB screening was offered to past and current residents of the apartment block, comprising an interview, Chest X-Ray, and Interferon Gamma Release Assay or Tuberculin skin test. Expected latent TB proportions were calculated using a reference population, and whole genome sequencing (WGS) was performed.
The index case was involved in a separate gaming centre outbreak involving five patrons. 241 current (67.9% of 355 residents) and 18 past residents were screened. The latent TB proportion was 19.9%, which was at the higher end of the expected range. WGS confirmed relatedness of cases’ MDR-TB isolates– eight of 10 isolates were genetically identical, while the remaining two were one Single Nucleotide Polymorphism apart.
With WGS, TB clusters not apparent through regular activity-based contact tracing may be detected. Mass screening may help inform the extent of transmission, but is limited by participation and difficulties in interpretation.
Multidrug-resistant tuberculosis (MDR-TB) causes substantial morbidity and mortality globally. Driven by inappropriate and inadequate antimicrobial treatment, the incidence of MDR-TB has been on the rise. Being more difficult and costly to treat, there are increasing concerns about its continued spread and negative impact on population health and health systems (
). However, there have not been reports of phylogenetically-linked MDR-TB cases within high-rise apartments.
TB outbreaks are usually addressed through contact investigations using concentric circles of exposure. The World Health Organisation (WHO) has published principles and recommendations on systematic screening for TB, discouraging indiscriminate mass screening and emphasising the need to prioritise at-risk groups after careful situational assessment (
). Nevertheless, mass screenings during outbreaks in homeless shelters, long-term care facilities, schools and even entire communities have been described and were often implemented due to evidence of unabated transmission (
This paper presents the first report of a mass screening exercise for residents of a high-rise apartment block in Singapore in response to a cluster of six phylogenetically-linked MDR-TB cases, three of whom did not have identifiable interactions with one another.
Study population and methods
Singapore is a densely populated tropical city-state in Southeast Asia, with about 8,000 people/km2. More than 80% of the population live in public high-rise apartment blocks. These extend up to 50 stories high and each forms a vertical community of up to 1,000 individuals, with shared elevators, common corridors, and communal areas for social activities (
). Doctors and laboratories are required to legally notify all new cases of active TB to the Ministry of Health. Once a case is confirmed, the National TB Control Unit initiates contact investigations, supported by DNA fingerprinting for drug resistant TB strains.
Apartment block cluster investigation
Between February 2012 and May 2016, six residents of an 11-storey apartment block with 160 one-bedroom units and two elevators were diagnosed with MDR-TB. The building had a ‘slab block’ design – on every level, units were arranged in a row and faced a naturally ventilated common corridor exposed to the elements [Figure 1]. The isolates were determined in May 2016 to have identical spoligotype (Ocimum Biosolutions, Hyderabad) and 24-loci Mycobacterial Interspersed Repetitive Unit – Variable Number of Tandem Repeat (MIRU-VNTR: Genoscreen, France) profiles (
Contact tracing and epidemiological investigations were conducted to identify links between cases. Trained nurses conducted in-depth interviews and cases detailed their past and present daily routines, workplaces, as well as formal and informal social activities. An investigation of the apartment block was conducted by public health officers to understand its structure and surrounding amenities and the number of registered occupants was obtained with the help of local housing authorities.
Despite these steps, clear transmission patterns between three of the six cases could not be found. As a precautionary measure, the Ministry of Health offered voluntary active and latent TB screening to all past and current residents of the apartment block. On-site screening was held on the ground floor of the block from 16 to 19 June, 2016. This comprised of an interview on their duration of residence, history of previous TB infection, as well as any clinical symptoms. It was followed by a Chest X-Ray (CXR) conducted in a mobile X-Ray bus, and interferon gamma-release assay test (IGRA; Quantiferon-TB Gold, Qiagen, Germany) or tuberculin skin test (Mantoux) for children under five years using 10 mm as the cut-off.
Subsequently, door-to-door interviews and on-site blood draws for IGRA was offered for another three days, and offered at nearby primary care clinics for a month. Those with positive IGRA cases were contacted for further assessment, including a CXR. Regardless of location, all tests were free of charge. The assistance of grassroots volunteers, distribution of circulars in the four national languages, and media reports helped inform residents of the mass screening exercise.
Illumina sequencing platforms was used for whole genome sequencing (WGS), generating 76 base-pair paired-end reads that were subsequently mapped to the H37Rv reference genome (GenBank accession no. NC000962.3) using BWA, after the sequence reads were adaptors-removed and quality-trimmed using Trimmomatic software. Single nucleotide polymorphisms (SNPs) were then identified for each sample, where a homozygous variant call was made at a position having a sequencing depth greater than 10% of the mean genome sequencing depth, and a consensus of more than 75% of the reads. SNPs identified within the PE/PPE and repeat regions were excluded. Positions with more than 10% of the samples having no reads covered, i.e. N base, and heterozygous genotype calls were removed from analysis.
Latent TB proportions were compared against a reference population of similar demographics from a local prevalence study using a multivariable logistic regression model developed from residents of similar residence type, which included age, sex and ethnic group (Cook, personal communications). Coefficients were applied to screened individuals to estimate individual risk, which was used to resample binary infection statuses and generate the distribution of total latent cases among those screened. To obtain confidence intervals, bootstrap was used on the algorithm. 95% prediction intervals were obtained using the upper (lower) confidence intervals for the upper (lower) resampled number of latent cases. Statistical tests were performed using the R statistical environment version 3.3.1 (
). He was diagnosed with MDR-TB in February 2012, defaulted treatment, and stayed at the home of a friend (“B”) and the friend’s mother (“C”); both of whom developed MDR-TB in 2012 and 2015 respectively. In 2014, a neighbour (“D”) living two apartments away on the same floor developed MDR-TB, followed in 2015 by a resident (“E”) who stayed in an apartment three floors directly above the index case’s apartment. The last case (“F”) stayed a substantial distance away in the same apartment block and developed MDR-TB in 2016. Apart from cases A, B and C, the remaining cases were not identified as contacts of cases A to C or to each other despite detailed interviews. There was no known convergence apart from the elevators, which had good ventilation (air volume replaced 20 times an hour in accordance to international standards).
There were 241 current (67.9% of 355 recorded current residents living in the apartment block) and 18 past residents were screened through the mass screening exercise – 28 residents refused CXR and seven refused IGRA tests. Demographic data are shown in Table 1 and results of the screening exercise are shown in Table 2.
Table 1Demographic profile of high-rise apartment block residents in the mass screening for tuberculosis.
Age (n = 259)*
Median: 52.8 (Range 3 mths – 95 y.o.)
- Less than 10
- 10 to 29
- 30 to 49
- 50 to 69
- More than 71
Gender (n = 259)*
Ethnicity (n = 259)*
Country of birth (n = 247)*
- South East Asian Countries
Years lived in apartment block (n = 233)
- Less than 5
- 5 to 9
- 10 to 19
- 20 to 30
- Above 30
The variables marked with an asterisk (*) were those also used in the multivariable analysis for latent TB rates.
Of the 259 total residents screened, three had prolonged cough but TB was excluded after testing. Five had CXR findings compatible with TB. Of these, three had previously-treated pan-sensitive TB, while the remaining two had radiographic findings compatible with minimal pulmonary TB. As no bacilli were isolated from cultures, they were treated with standard first-line anti-TB drugs. The residents have completed treatment uneventfully and remain under surveillance.
There were 46 residents that had positive IGRA or Mantoux but normal CXRs, suggestive of a latent TB proportion of 19.9%. Another six residents had positive IGRA or Mantoux but no CXR. These 52 residents were distributed all across the apartment block. The expected proportion of latent TB based on the reference population was determined to be 13.9% (95% prediction interval [6.7%–22.6%]). The latent TB rate observed in the block (19.9%) was therefore at the higher end of the expected range.
An attempt was made to use WGS results to elucidate transmission patterns between cases. Eight of 10 isolates were found be to genetically identical, while the remaining two were one SNP apart. As such, although WGS confirmed isolates’ relatedness, the differences between isolates were not significant enough to infer transmission patterns between cases within the apartment block and those from the gaming centre [Figure 2].
There have not been previous reports of phylogenetically linked MDR-TB cases among residents of high-rise apartments. This cluster highlights the threat posed by MDR-TB to densely populated areas, and the difficulties in identifying clusters due to the long incubation period of TB. It also shows the utility of WGS in linking cases, without which cases might have been simply assumed as being due to chance. However, even with the additional resolution by WGS, the sequence of transmission could not be conclusively determined – such difficulties have also been described in other TB outbreaks (
). At that time, none of his neighbours were screened, as they were not identified as contacts, in keeping with the principle that TB spread is through prolonged and close contact. The absence of identifiable common activities highlights the difficulties faced in TB contact tracing in urban settings. While recall bias and refusal to declare contacts are possible, casual contact especially with those living in close geographical proximity cannot be ruled out (
). During this period of high infectiousness he had been staying at the apartment block and this may have been the window of highest exposure risk for other cases.
This cluster thus also highlights the public health challenges in MDR-TB control associated with persons who default treatment. This may be precipitated by factors such long duration of treatment, side effects from medications, or social stigma (
). Enforcement of supervised treatment continues to generate much discussion around the ethics of infringement on patients’ rights in the interest of protecting public health, especially during the intensive phase of treatment where there may be substantial risk of spread to others. In Singapore, all persons diagnosed with MDR-TB are placed on a national Directly Observed Therapy (DOT) programme which provides medications free of charge to residents. Nevertheless, enforcement can be problematic, as shown by the delay in locating the index case.
Outbreaks of TB have been traditionally associated with closed or semi-closed settings, such as healthcare facilities (including hospitals and step down care), homeless shelters, schools, and correctional facilities (
). Public high-rise apartments do not fit conventional definitions, although such emerging clusters would challenge these assumptions. Previous studies have shown that public housing residents are at increased risk of TB as compared to other housing types (
Hong Kong has a similar housing model as Singapore and TB rates had been found to be higher among those living on lower floors. Environmental factors such as poorer air quality from street level pollution and less exposure to sunlight were postulated as the cause (
). Although five of six cases in this Singapore cluster also lived on lower floors, its significance remains inconclusive because the index case happened to lodge on a lower floor. Previous environmental studies on apartment blocks in Singapore showed that mean wind speeds increased with height of buildings, and that ‘slab block’ designs (as per this cluster) tended to slow down approaching wind as compared to ‘point block’ designs that comprise units placed at the corners of each floor (
). Although the well-ventilated elevator was less likely to have been the conduit for disease spread, the ventilation efficacy for the common corridors and communal areas were not measured. Further studies are needed to better understand the contribution of environmental factors, including physical structure and dynamic airflow patterns, to the spread of TB.
This is also the first known attempt to screen for active and latent TB among all residents in a high-rise apartment block en-masse. WHO generally discourages indiscriminate mass screening as there should be prioritisation of risk groups based on risk-benefit, feasibility, acceptability, numbers needed to screen, and cost-effectiveness. However, their guidelines also mentioned that as a conditional recommendation, systemic screening may be considered for geographically-defined subpopulations with extremely high levels of undetected TB (1% prevalence or higher) (
). This screening exercise was primarily to determine the extent of transmission, as preliminary investigations failed to reveal common activities or risk factors among three of the six cases. It was also uncertain as to whether the six cases (out of the current 355 residents) indicated ongoing spread and other undiagnosed residents. This was especially since the last known case was only recently diagnosed and stayed a distance away from the first five cases.
As interactions with cases were still likely to have been casual, screening was not made compulsory and turnout for current residents was moderate, at 67.9%. Participation was aided by media interest and grassroots volunteers, and attempts were made to reduce barriers to access. These included making tests free (to address cost concerns), providing alternate screening locations (to address stigma) as well as offering door-to-door Quantiferon screening (to address mobility difficulties). Despite these efforts, there were likely other underlying barriers. There was a skew in those attending screening towards women and older persons. This was somewhat different from a previous meta-analysis that found that while male gender was a determinant for delay in TB diagnosis, age was not (
). While it may be simply a reflection of the demographic composition of residents, anecdotal feedback from non-participants included apathy and lack of time.
Latent TB rates of 19.9% were on the higher end of the expected range and concerns of future MDR-TB cases in the apartment block remain. This is especially with the long latency of TB, cross-sectional nature and incomplete uptake of screening. Unfortunately, the status for the unscreened 32% of residents remains unknown. As most exposure was assumed to be beyond the window period, a follow-up screen for seroconversion was not conducted. It was also not possible to determine if those with latent TB were infected with pan-sensitive or MDR-TB and any prophylactic treatment would not be able to eliminate MDR-TB strains. Nevertheless, an awareness of their latent TB status, coupled with the knowledge of possible past exposure to an MDR-TB case, may encourage future cases to come forward for testing in a timelier fashion. As TB is endemic in Singapore, the discovery of the two newly diagnosed cases of active TB was not unexpected. Furthermore, both cases were only picked up radiologically and had minimal symptoms.
The screening exercise was still useful to allay residents’ anxieties and provided local health authorities with a better understanding of the situation. However, the exercise was also relatively resource intensive in terms of manpower and logistics required, and its cost effectiveness remains unclear. An unintended effect due to its widespread publicity was a rise in public expectations, with subsequent requests for on-site mass screening for single TB cases that were not warranted. Similar screening exercises in the future, and the manner to which they should be conducted, thus still need to be carefully considered and tailored after assessing the risk (and harm) versus benefit to residents on a situation-by-situation basis and aligned with the WHO manual on systematic screening for TB (
With increased global travel and urbanization, MDR-TB will continue to be a challenge for cities, especially those that are densely populated. With WGS, TB clusters that were not apparent through regular activity-based contact tracing may be detected. Mass screening may help inform the extent of transmission, but is limited by participation and difficulties in interpretation of findings. Nevertheless, it provides an option in the varied toolbox of TB detection and control.
Factors associated with patient and provider delays for tuberculosis diagnosis and treatment in Asia: a systematic review and meta-analysis.