Research and analysis

Tuberculosis diagnosis and microbiology, England, 2023

Published 5 December 2024

Applies to England

Main points

In 2023:

• reducing the time to diagnosis of pulmonary tuberculosis (TB) reduces the risk of transmission to others, and, for all forms of TB, improves outcomes for an individual
• the median time to diagnosis from symptom onset is 75 days, this remains long and unchanged since 2018
• timely notification supports prompt public health action; 59.8% of notifications were made in 2023 within the statutory 3-day window from diagnosis, slightly higher than previous years
• TB infection is optimally confirmed by culture for full information on drug resistance and transmission to be provided; in 2023 61.2% of individuals had positive cultures similar to previous years
• there has been an increase since 2022 in the proportion of people with multidrug-resistant (MDR) or rifampicin-resistant (RR) TB with 2.4% of people with positive cultures having this pattern of resistance in 2023
• the proportion of individuals with culture positive disease in a genomic cluster, suggesting recent transmission, has declined gradually since 2019 (40.7%), and is 34.0% in 2023

Diagnostic delay

The prompt diagnosis and treatment of active TB improves patient outcomes, and for pulmonary TB, reduces the period of infectiousness and onward TB transmission. Diagnostic delay is the time from onset of symptoms to diagnosis and may be affected by symptom recognition, healthcare seeking behaviours, access to services, clinical pathways and awareness, and service capacity.

Median diagnostic delay for pulmonary TB was 75 days in 2023, static since 2018 (see appendix figure A1 and Supplementary Table 1 of the accompanying dataset).

Healthcare-related diagnostic delay covers the period between presentation to a healthcare service and a TB diagnosis being made. In 2023, this was 28 days, a small increase over the past 5 years, this is in part due to the slow growth rate of Mycobacterium tuberculosis complex (MTBC) (see appendix figure A2 and Supplementary Table 1 of the accompanying dataset).

Time to notification

TB disease is notifiable within 3 days of diagnosis under the Health Protection Regulations and prompt notification supports early public health action.

In 2023, timely notification of TB was delayed beyond the statutory period in 40.2% of individuals (1,951 out of 4, 855), with 8.6 % of individuals being notified more than 30 days after diagnosis, a slight improvement since the previous year, when this was 44.2% (1,933 out of 4,375) (see Supplementary Table 2 and appendix figure A3).

There is variation between regions in the timeliness of notification (Supplementary Table 3), with the highest proportion of notifications after 3 days in West Midlands region and the highest proportion after 30 days in the East Midlands region (see appendix figure A4).

Diagnostic confirmation

In 2023, 96.0% (4,659 out of 4.855) of individuals notified with TB had at least one diagnostic test recorded (including radiology, histology and microbiology). This is similar to 2022. The proportion is higher in people with pulmonary (98.5%; 2,627 out of 2,668 individuals) than non-pulmonary disease (92.9%; 2,032 out of 2,187 individuals)(see Supplementary Table 4).

PCR tests

Use of molecular testing, including polymerase chain reaction (PCR), can provide a more rapid diagnosis than culture and reduce delay between symptom onset and start of treatment. NICE recommends that respiratory samples should have a PCR in all cases for diagnosis of TB in children. PCR is recommended in adults if any of the following conditions are met: resistance to rifampicin is suspected, the patient has HIV, the result would change management, or a large contact tracing exercise is being planned. For non-respiratory samples, PCR is indicated if it would change management.

PCR information relies on user entry into the National Tuberculosis Surveillance System (NTBS) and therefore numbers recorded may underestimate diagnostic testing.

In 2023, data on PCR testing was recorded in NTBS for 37.4% of notifications (1,814 of 4,855), which is a small increase compared with 2022 (33%). Only 35.8% (29 of 81) of children (aged less than 15 years) with pulmonary disease, had PCR testing information recorded. Overall, in 2023, 80.7% (1,464 of 1,814) of people with TB who had a PCR result recorded were PCR positive (see Supplementary Table 5).

The recording of PCR results may not reflect actual diagnostic practice as it relies on user entry.

Data on PCR availability and use from a national laboratory audit and from the Getting it Right First-Time review will be available in late 2024 and can be used to support local quality improvement in this area.

Culture confirmation

Culture of M. tuberculosis or MTBC is required to allow drug resistance information to be provided and for typing. In 2023, 61.2% of individuals (2,973 of 4,855) with TB had their diagnosis confirmed on culture. This is similar to previous years. Culture confirmation is higher in pulmonary than non- pulmonary disease due to the greater ease of sampling the respiratory system and often a higher bacterial load. In 2023, 75.1% of individuals (2,006 of 2,668) with pulmonary disease had culture confirmation compared to 44.2% of individuals (967 of 2,187) with non-pulmonary disease. These proportions are in line with previous years, and the proportion of pulmonary disease confirmed by culture remains below the National Action Plan target of 80% (see Supplementary Table 6).

There is regional variation in culture confirmation, details are found in Supplementary Table 7 by UKHSA region (see appendix figure A5), Supplementary Table 8 by integrated care board (ICB). As in previous years, males, people born outside the UK and young adults (aged 15 to 44) are more likely to have diagnoses confirmed by culture than the relevant comparator groups. Children aged 0 to 14 have the lowest culture confirmation proportions (see Figure 1).

Figure 1. Risk ratio of being culture positive, the reference categories are: 15 to 44 years for age, non-pulmonary for site of disease, female for sex, and non-UK born for place of birth these all have a risk ratio of 1 (see Supplementary Table 9)

Determination of Mycobacterium tuberculosis complex (MTBC) species

Determination of species within the Mycobacterium tuberculosis complex (MTBC) has important implications for individual treatment and public health action. Most disease in humans is caused by M. tuberculosis rather than other members of the complex. In 2023, of the 2,973 people with positive cultures, 97.7% had M. tuberculosis, consistent with previous years. Supplementary Table 10 of the accompanying dataset reports the breakdown of all species within the complex that caused human disease from 2011 to 2023.

Identification and classification of drug resistance

Whole-genome sequencing (WGS) is used for all individuals with culture confirmation to make drug susceptibility predictions for the first line agents rifampicin, isoniazid, pyrazinamide, and ethambutol, plus aminoglycosides and fluoroquinolones. Identification and reporting of drug resistance using WGS is more rapid than conventional phenotypic drug susceptibility testing (pDST). However, pDST is still performed for first line drugs if there are incomplete results from WGS. pDST is performed for first-, second- and third-line agents if resistance to first line agents is detected by WGS.

Drug resistance

In 2023, 97.1% of people whose TB was confirmed by culture (2,886 out of 2,973) had results reported for first line drugs (see Supplementary Table 11, Figure 2). This is below the National Action Plan and WHO endTB target of 100%, but consistent with previous years. The 100% target is not attainable due to biological failures in subculture of organisms and contamination of samples with other organisms that cannot always be eliminated.

In 2023, for individual first line antibiotics in the 2,973 people with culture confirmed TB (see Supplementary Table 12):

• 7.5% of people had resistance to isoniazid
• 2.4% of people had resistance to rifampicin
• 1.7% of people had resistance to ethambutol

We are not reporting on the proportion with resistance to pyrazinamide (and therefore the category of any first line agent only includes rifampicin, isoniazid and ethambutol) in 2023 because the laboratory testing was adversely impacted by a problem with quality control in the supply chain for the media used for pDST for this drug. The manufacturer issued a Field Safety Notice in July 2024 stating that there may have been false detection of resistance from June 2023.

Figure 2. Drug resistance proportion of culture positive cases in 2023

Note: resistance to any first line drug excludes pyrazinamide for 2023.

Isoniazid resistant TB

In 2023, 5.4% of individuals (161 of 2,973 culture positive) had isoniazid resistant TB which was not additionally resistant to rifampicin. This proportion is consistent with previous years (see Supplementary Table 13).

Rifampicin resistant (RR) or multidrug resistant (MDR) TB

In 2023, increases were seen in the proportion of both culture-confirmed RR or MDR TB, 71 out of 2,973 (2.4 %), and individuals treated as RR-MDR TB, 108 out of 4855 (2.2%) (see Figure 3). These are now the highest since enhanced surveillance began (see Supplementary tables 13, 14 and 15).

People may be treated for RR or MDR TB in the absence of culture confirmation if the diagnosis was made from a molecular test alone, made abroad or based on a close contact history with an individual known to have RR or MDR TB. In addition, individuals may be treated with an MDR regimen if they are unable to tolerate rifampicin-based combinations because of severe intolerances or critical drug interactions.

Treatment of RR or MDR TB requires involvement of a specialist centre, and may require prolonged admission, especially for pulmonary disease, which places heavy burdens on local services.

69% of individuals with RR or MDR TB in 2023 had pulmonary involvement (Supplementary tables 16 and 17). Contact tracing around individuals with RR or MDR TB is often complex and extensive.

In 2023, 65 individuals (91.5%) with RR or MDR TB were born outside the UK. Of these 55.4% had entered the UK within 5 years (see Supplementary Table 16).

Breakdown of RR or MDR TB case numbers by UKHSA region is in Supplementary Table 18.

Figure 3. Number (a) and proportions (b) of RR-MDR TB and Isoniazid resistance without MDR TB since 2011

Figure 3 (a)

Figure 3 (b)

Testing and resistance to WHO group A and new and repurposed drugs (linezolid, clofazimine, moxifloxacin, levofloxacin, delamanid and bedaquiline)

The WHO defines pre extensively drug resistant (pre-XDR) TB as RR or MDR TB with additional resistance to any quinolone, and XDR as RR or MDR TB which is also resistant to any quinolone and at least one additional group A drug (bedaquiline, linezolid) (see WHO consolidated guidelines on drug resistant TB).

Overall, the proportion of infections with RR or MDR TB which have additional resistance leading to classification as pre-XDR has increased slightly over time and was 21.1% in 2023 (see Supplementary Table 13). There have been 6 individuals with XDR TB in England in the years 2018 to 2023.

Details of resistance to other agents in individuals with RR or MDR TB are found in Supplementary tables 19 and 20.

Results of phenotypic testing for quinolones and new and repurposed agents including linezolid, bedaquiline, clofazimine and delaminid, are reported in this chapter for the first time. In line with the commissioning by NHS England of the new WHO-approved short course regimen for MDR TB (WHO and NHSE) including pretomanid in 2024, susceptibility results for this will be reported from next year.

Quinolone resistance was seen in 2.1% (63 of 2,973) of cultured isolates in 2023. Most of this quinolone resistance is in people with RR or MDR TB (Figure 4) with the highest proportion of quinolone resistance in people with pre-XDR TB from Lithuania (see Supplementary tables 19 and 21).

Figure 4. Proportion of Quinolone resistance in the RR-MDR, isoniazid resistance without MDR and non MDR or isoniazid resistant cases

Phenotypic susceptibility testing for linezolid, bedaquiline, clofazimine and delaminid is performed in all RR or MDR TB and in selected individuals with non RR-MDR TB treated with MDR regimens. The proportion of isolates resistant to any of these agents remains low, but increasing use of these agents and corresponding global emergence of resistant strains means that testing and surveillance need to be maintained. The complexity of the pDST testing for new and repurposed drugs mean that little information is available globally on the baseline resistance for these drugs, or of specific intrinsic resistance against these drugs for the distinct species and lineages (see Supplementary Table 20).

Clustering

WGS also provides information on relatedness and potential transmission. In England, individuals with a positive culture are grouped into genomic clusters if they have at least one other individual within 12 single nucleotide polymorphisms (SNPs). This information is used to guide contact tracing and public health action. More information is found in the WGS handbook. 34% of individuals with a positive TB culture were part of a cluster in 2023, down from 40% in 2019 (see Supplementary Table 22). The reasons for the decrease are likely to include a combination of changes in transmission patterns as a result of more importation of new strains from recent migrants and, possibly, better awareness regarding airborne transmission and how to limit this due to the COVID-19 pandemic.

Drug resistance in genomic clusters

In 2018 to 2023, the most common drug resistance in genomic clusters was isoniazid resistance without RR or MDR TB. This was found in 5.4% of individuals in a cluster.

129 individuals (40.7%) with RR or MDR TB between 2018 to 23 were in a genomic cluster (see Supplementary Table 23), with similar likelihood of being in a cluster as people without RR or MDR disease.

Risk factors for being in a genomic cluster

Males, people aged 0 to 14 and 15 to 34 are more likely to be in genomic clusters, as are people with pulmonary disease and those born in the UK (see Figure 5 and Supplementary Table 23). Individuals with pulmonary disease are the most likely to be infectious and so transmit TB to other people, resulting in clustering. Younger individuals and those born in the UK are more likely to have been exposed to TB in the UK relatively recently, which increases the likelihood that there will be a genomically related individual. The presence of any social risk factor nearly doubles the likelihood of being in a cluster (relative risk 1.88), with drug misuse and history of incarceration having the highest risks. This is related to increased contact between individuals at high risk of having TB.

People born outside the UK who are recorded as having asylum seeker status are 80% more likely to be in a cluster than non-UK born individuals not seeking asylum (relative risk 1.84, confidence intervals 1.86 to 2.02). It is unclear if this represents transmission in the UK, recent transmission en route, or due to recent outbreaks caused by failing healthcare in the countries from where asylum seekers travel. History of incarceration and drug misuse remain the social risk factors with the highest relative risk of clustering (see appendix figure A6).

Figure 5. Risk ratio of being in a cluster

Note: only culture positive cases are considered.

The reference categories are: 25 to 44 years for age, non-pulmonary for site of disease, female for sex, UK born for place of birth, non-RR-MDR and non-isoniazid resistance without RR-MDR for the resistance categories, these all have a risk ratio of 1 (see Supplementary Table 23).

Appendix

Figure A1. Box plot of the number of days between symptom onset and diagnosis, 2018 to 2023

Whiskers show 10th and 90th percentile, box represents 25th and 75th percentile, median is indicated by the number.

Missing data: 154 people notified post mortem, 7 people with diagnostic delay under 0 days and 1,494 people with missing data (all due to missing presentation date) out of 14,692 pulmonary infections, are not reported.

Figure A2. Box plot of the number of days between presenting to healthcare service and diagnosis, 2018 to 2023

Whiskers show 10th and 90th percentile, box represents 25th and 75th percentile, median is indicated by the number.

Figure A3. Duration from tuberculosis (TB) diagnosis to notification by year, 2018 to 2023

Figure A4. Duration from tuberculosis (TB) diagnosis to notification by UKHSA region, 2023 only

Figure A5. Proportion of people with culture confirmation by UKHSA region, 2023

Figure A6. Risk ratio between presence of social risk factors and belonging to a cluster

Risk ratio of social risk factors associated with being in a cluster, of note only culture positive cases are considered. The reference categories for each social risk factor is not possessing the social risk factor, these all have a risk ratio of 1 (see Supplementary Table 23).

Note: the effect of being an asylum seeker is only evaluated in the non-UK born population.