Research and analysis

TB diagnosis and microbiology, England, 2022

Updated 15 February 2024

Main messages

Reducing the time to diagnosis of pulmonary 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 remains long and unchanged since 2017 (median 76 days in 2022).

Timely notification supports prompt public health action; only 53.8% of notifications were made in 2022 within the statutory 3-day window from diagnosis, consistent with previous years.

TB infection is optimally confirmed by culture for full information on drug resistance and transmission to be provided; in 2022 61.6% of individuals had positive cultures similar to previous years.

Culture confirmation proportions are higher for pulmonary (75.6%) than non-pulmonary infection (46.9%) as pulmonary sites are easier to sample. These proportions are similar to previous years Targets of 80% of pulmonary infection and a 5% annual increase for all infections being confirmed on culture are not being met suggesting more work needs to be done to improve sample collection and processing from people in whom TB is suspected.

The proportion of culture confirmed cases with resistance to first line anti-tuberculous agents has risen again in 2022 to 11.9%. This is caused by an increased proportion resistant to pyrazinamide, from improved detection of pyrazinamide resistance, with resistance to the other first line agents remaining stable. Treatment duration is extended to 9 months in individuals with pyrazinamide resistant TB.

There has been an increase in the proportion of people treated as MDR (multidrug-resistant) or RR (rifampicin-resistant) TB with 2.2% of all notified individuals treated for MDR or RR TB in 2022 compared with 1.3% in 2019.

The proportion of individuals in a genomic cluster, suggesting recent transmission, has declined gradually since a peak in 2019, and is 19.7% in 2022.

Diagnostic delay

The prompt diagnosis and treatment of active TB can improve patient outcomes, and for pulmonary TB, reduce the period of infectiousness and onward TB transmission. Diagnostic delay is the delay 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 was 76 days in 2022, static since 2017 and increased compared with earlier years (63 days in 2012; see Appendix Figure A1 and Supplementary Table 1 of the accompanying data set).

Median healthcare-related diagnostic delay covers the period between presentation to a healthcare service and a TB diagnosis being made. In 2022 this was 26 days, increased from 24 days in 2018 when this data was first available. (Appendix Figure A2 and Supplementary Table 1 of the accompanying data set).

Time to notification

TB disease is notifiable within 3 days of diagnosis under the health protection regulations.

In 2022 timely notification of TB was delayed beyond the statutory period in 1,940 out of 4,203 (46.2%) of individuals, with 10.2% of individuals being notified more than 30 days after diagnosis.

Figure 1. Proportion of people notified with TB within 3 days of diagnosis by year, England, 2017 to 2022

The data used in this figure can be found in Supplementary Table 2 of the accompanying data set.

Median time to notification for all years since 2012 and the proportion of individuals notified within 3 days in 2022 and by UKHSA region are found in Appendix Figures 3 and 4 and in Supplementary Tables 1 to 3 in the accompanying data set.

Diagnostic confirmation

The National Institute for Health and Care Excellence (NICE) recommends that all patients in whom TB is suspected should provide samples for microbiological diagnosis.

For tests other than culture confirmation, data was obtained from user-entered variables in National TB Surveillance service (NTBS). Some of these variables have large amounts of missing data, so should be interpreted with caution.

Confirmation of active TB infection through any diagnostic test

In 2022, the proportion of notifications with any diagnostic test reported was 95.0% (95% confidence interval (CI), 94.3% to 95.6%). It was higher in those with pulmonary disease 97.6% (95% CI, 96.9% to 98.2%) compared with 91.8% (95% CI, 90.5% to 93.0%) in those with non-pulmonary disease. The number individuals with any diagnostic test recorded increased slightly compared with 2021 as did those with any diagnostic test positive and culture positive tests (see Supplementary Table 4 of the accompanying data set.

PCR tests

NICE recommends that respiratory samples should have a PCR in all cases for diagnosis of TB in children. Polymerase chain reaction (PCR) is recommended in adults if any of the following are met:

• resistance to rifampicin is suspected
• the patient has HIV
• the result would change management
• a large contact tracing exercise is being planned

For non-respiratory samples, PCR is indicated if it would change management.

In 2022, data on PCR testing was recorded in the NTBS for 33.1% of notifications (1,450 of 4,380). When limited to children (aged less than 15 years) with pulmonary disease, 98 (72.1%) of children had no PCR testing information recorded. Overall, in 2022 26.0% of people with TB (1,139 of 4,380) were recorded as PCR positive and 7.1% (311 of 4,380) were recorded as PCR negative.

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

A national laboratory audit is due to be piloted early in 2024 to gather laboratory-level data and, along with the ‘Getting it right first time’ (GIRFT) review, to identify if diagnostic testing is in line with NICE guidance.

Culture confirmation

In 2022, 62.6% of all notifications were culture confirmed (Appendix Table A1), 75.6% with pulmonary disease and 46.9% in non-pulmonary disease, similar to previous years,

Culture confirmation rates for England by pulmonary and non-pulmonary disease (2000 to 2022) are shown in Supplementary Table 5, by UKHSA Region, 2011 to 2022 in Supplementary Table 6 and Appendix Figure A5, and by integrated care board (2017 to 2022) in Supplementary Table 7 of the accompanying data set.

Predictors of culture confirmation

To better understand factors associated with the probability of culture confirmation, the proportions of people with culture confirmation for the years 2020 to 2022 were compared by potential predictive factors such as age, sex and country of birth and pulmonary versus non-pulmonary TB. The differences were quantified by calculation of risk ratios and, where statistically significant, are shown in Figure 2 and, for all factors assessed, in Supplementary Table 8 in the accompanying data set.

People with pulmonary disease were 63% more likely to be culture-confirmed compared with those with non-pulmonary TB, likely reflecting the greater ease of sampling respiratory sites.

Males were 16% more likely to have culture-confirmed disease than females. Those born in the UK were 9% less likely to be culture confirmed, despite pulmonary disease being more frequent in the UK born population. Culture confirmation was most likely in the 15 to 44 year age group and children under 15 years were 52% less likely to have a positive culture.

Figure 2. Predictors of culture confirmation (all TB notifications), England, 2020 to 2022, aggregated data (univariable analysis)

Note: reference groups for risk ratios in the order presented in the above figure are: female, non-pulmonary TB, born outside of the UK and aged 15 to 44 years old.

Data underlying this figure is available in Supplementary Table 8 in the accompanying data set.

Determination of Mycobacterium tuberculosis complex (MTBC) species

In 2022, 2,742 people (62.6% of notifications) had their TB successfully cultured, sequenced and the species within the Mycobacterium tuberculosis (MTB) complex identified. 97.4% were M. tuberculosis in 2022, consistent with previous years. Supplementary Table 9 of the accompanying data set reports the breakdown of species from 2011 to 2022.

Identification and classification of drug resistance

Whole-genome sequencing (WGS) is used to make drug susceptibility predictions for all culture-confirmed individuals 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 second line drugs if first line resistance is detected by WGS, or if WGS does not give complete results for first line drugs.

Drug resistance

In 2022, 98.5% (2,700 of 2,742) of people with culture-confirmed TB had susceptibility results for all first line drugs (Supplementary Table 10 of the accompanying data set. This remains consistent with previous years, slightly under the target of 100%, from the WHO ‘End TB’ strategy of universal susceptibility testing. Contamination of cultures and failure of isolates to grow for phenotypic testing accounts for the proportion being under 100%.

In 2022, 11.9% (327 of 2,742) of people with culture-confirmed TB had resistance to any first line drug. In 2022, for individual first line antibiotics in the 2,742 people with culture confirmed TB:

• 7.2% of people had resistance to isoniazid
• 1.6% of people had resistance to rifampicin
• 1.1% of people had resistance to ethambutol
• 5.8% of people had resistance to pyrazinamide

Supplementary Table 11 of the accompanying data set shows resistance for the above 4 drugs from 2011 to 2022. Other than for pyrazinamide, the proportion of cultures resistant to each of the first line agents was stable. Pyrazinamide resistance has increased since 2017 corresponding with the staggered introduction of new testing protocols due to the recognition that TB strains from a particular branch (‘lineage 1 – East African-Indian Ocean’) have resistance mechanisms which are not detected by the current genotypic methods.

Figure 3 summarises initial drug resistance for people notified in 2022.

Figure 3. Summary of drug resistance in individuals with culture positive TB in 2022

Accessible text version of Figure 3.

This figure contains arrow boxes from left to right, describing 98.9% (2,711) drug resistance results available, resulting in 11.9% (327) with resistance to any first line drug, 1.6% having MDR or RR TB (43), and 0.2% (6) having pre-XDR and zero with XDR-TB.

Note: denominator for percentages is total with culture confirmation.

Isoniazid resistance and MDR or RR TB

Figure 4a and 4b and Supplementary Table 12 of the accompanying data set report the numbers and proportions of people with initial (at diagnosis) isoniazid monoresistance (INH-R) and MDR or RR TB, pre-extensively drug resistant (pre-XDR) and extensively drug resistant (XDR) TB (following WHO definitions) at diagnosis for the period 2011 to 2022.

The proportion of people with initial INH-R in 2022 was 6.1% (166 out of 2,742) of people; this has been stable over the period. In 2022 the proportion of MDR or RR TB decreased to 1.6% (43 people) from a peak prevalence in 2020 of 2.1% (54 people). However, the number and proportion of people treated for MDR or RR TB in 2022 was the highest since 2012 (Supplementary Table 13 of the accompanying data set.

In 2022, 95 individuals were treated for MDR or RR, 2.2% of all those notified (95% CI, 1.8 to 2.6). In 2022, none of these were due to post-diagnosis acquisition of drug resistance (Supplementary Table 14 of the accompanying data set.

Individuals may be treated for MDR or RR disease 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 MDR or RR infection. 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.

Figure 4a and 4b. Number and proportion of people notified with culture-confirmed TB with initial drug resistance to isoniazid alone and with multidrug-resistant or rifampicin-resistant TB (MDR or RR TB), England, 2012 to 2022

a)

b)

Data underlying these figures is available in Supplementary Table 12 in the accompanying data set.

Pre-XDR and XDR TB

In 2022, of 43 people with culture-confirmed MDR-TB, there were 6 (14.0%) that were confirmed as pre-XDR (resistant to rifampicin and fluoroquinolones). Between 2011 and 2022, there were no individuals diagnosed with XDR-TB and no individuals who acquired MDR or RR disease on treatment (Supplementary Tables 12 and 14 of the accompanying data set.

Drug resistance by site of disease, country of birth, presence of one or more SRFs and previous TB diagnosis

Supplementary Table 15 of the accompanying data set reports the proportions of INH-R and MDR or RR TB aggregated for the last 3 years by site of disease (pulmonary or non-pulmonary), UK born and specific countries of birth, plus the presence of at least one SRF and whether previously diagnosed with TB.

Both INH-R and MDR or RR TB were observed in a greater proportion of individuals born outside the UK  than UK born (6.3% and 5.3% INH-R (p equal to 0.12), 2.0 and 1.2% MDR or RR, (p equal to 0.03) respectively), The country of birth with the highest numbers of individuals with both INH-R and MDR or RR was India (105 and 34 individuals respectively), but the countries of birth from which the highest proportions of individuals had drug-resistant disease were, for INH-R, the Philippines (13.6%, 28 individuals) and China (11.1%, 6 individuals) and for MDR or RR, Lithuania (10.1%, 8 individuals) and Vietnam (9.0%, 6 individuals). INH-R was more common those with a SRF (8.4% versus 5.6%; p equal to or less than 0.001). MDR or RR TB was much more common in those with a history of a previous TB diagnosis (6.1% versus 1.6%, p less than 0.001).

Supplementary Table 16 of the accompanying data set reports the numbers and most common countries of birth for those with MDR-RR TB sub-divided into extended groupings of drug-resistant TB (resistant to all first line drugs, pre-XDR and XDR-TB) for the period 2018 to 2022.

Supplementary Table 17 of the accompanying data set reports the aggregated numbers and proportions of MDR or RR TB by UKHSA region for the period 2018 to 2022. Yorkshire and the Humber had the highest proportion of people with MDR or RR TB at 2.4%. The other UKHSA regions had proportions between 1.1% (East Midlands) and 1.9% (London and the East of England).

Identification of isolates in genomic clusters

In 2018, WGS was introduced for all culture-confirmed TB cases in England and is used to identify people with TB whose isolates cluster together. More information is found in the WGS handbook.

Appendix Table A1 shows the number of people identified whose TB was in a cluster at 12 single nucleotide polymorphisms (SNPs) or fewer, with at least one other person’s samples between 2018 and 2022. The proportion of notifications in a cluster in 2022 was 19.7%. This is the lowest number and proportion since 2018 when country-wide sequencing started and is significantly lower than years before 2020 (Appendix Table A1). The reasons for this are complex and may be related to a combination of changes in transmission patterns as a result of the COVID-19 pandemic, more importation of new strains from recent migrants and potentially less culture confirmation of clustered individuals.

Species in genomic clusters

Between 2018 and 2022, 98.1% of individuals whose isolates clustered had M. tuberculosis (4,906 of 5,003). There were 54 clustered individuals who had M. bovis (1.1% of clustered isolates) and 36 with M. africanum (0.7%). In 2022, there were 842 individuals who clustered who had M. tuberculosis, 12 with M. bovis and 4 with M. africanum.

Resistance in genomic clusters

In 2018 to 2022, isoniazid monoresistance (INH-R) was the most common drug resistance pattern found in 351 (7.0%) of clustered notifications.

Of 1,045 notifications with INH-R between 2018 and 2022, a third were in a cluster. This suggests that while there is ongoing transmission of INH-R disease, there is also frequent generation or importation of new INH-R strains.

Genomic clustering of isolates by UKHSA region

Appendix Figure A6 and Table 18 of the accompanying data set show the number and proportion of notifications in a genomic cluster in each UKHSA region. In 2022, the highest proportion of isolates in a cluster was in the East of England (24.5%) and the lowest in the South East (14.6%). However, the proportion of notifications clustering is also dependent on the numbers and proportions of culture confirmed notifications which varies by UKHSA region.

Predictors of being in a genomic cluster

Figures 5a and 5b and Supplementary Table 19 of the accompanying data set report risk ratios of factors associated with genomic clustering using aggregated data from 2018 to 2022 and univariable analysis. Isolates from people with pulmonary TB were more than 3 times more likely to be clustered. Individuals with pulmonary disease are the most likely to be infectious and so transmit TB to other people, resulting in clustering. Males, individuals born in the UK and people aged 15 to 44 years are more likely to be in clusters. Isolates from people with social risk factors were also more likely to be in a cluster, with the exception of people who were asylum seekers. Clustering is more likely as deprivation score of the postcode of residence increases. Public health interventions to interrupt transmission need to focus on the highest risk groups for clustering.

Figure 5a. Predictors of being in a cluster, England, 2018 to 2022 (aggregated data)

Figure 5b. Predictors of being in a cluster: social risk factors, England, 2018 to 2022 (aggregated data)

Data underlying this figure is available in Supplementary Table 19 in the accompanying data set.

Appendix

Figure A1. Median diagnostic delay for people notified with pulmonary TB in England, 2012 to 2022

Notes
1. The left and right sides of the boxes represent the interquartile range, and the black capped lines represent the minimum and maximum values.
2. The following were excluded in this analysis: those diagnosed postmortem (407); those with missing data (230); those who had diagnostic delays of over 2 years (356); and those with values less than 0 days (7). This resulted in a total of 11.1% of total notifications with pulmonary TB from 2012 to 2022 being excluded.

The data underlying this figure is available in Supplementary Table 1 in the accompanying data set.

Figure A2. Median healthcare-related diagnostic delay for people notified with pulmonary TB in England, 2018 to 2022

Notes
1. The red dotted line indicated the baseline target for 2023 notifications of the median of 2021 and 2022. The left and right sides of the boxes represent the interquartile range, and the black capped lines represent the minimum and maximum values.
2. Data is only available for this variable from 2018.
3. The following were excluded in this analysis: those diagnosed postmortem (130); those with missing data (615); those who had delays of over 2 years (60); and those with values less than 0 days (48). This resulted in a total of 7.1% of total notifications with pulmonary TB from 2018 to 2022 being excluded.

The data underlying this figure is available in Supplementary Table 1 in the accompanying data set.

Figure A3. Median reporting time for people notified with TB in England, 2012 to 2022

Notes
1. The left and right sides of the boxes represent the interquartile range, and the black capped lines represent the minimum and maximum values.
2. The following were excluded in this analysis: those with missing data (4,065) due to missing diagnosis date; those who had delays of over 90 days (1,791); and those with values less than 0 days (1,161). This resulted in a total of 12.5% of total notifications from 2012 to 2022 being excluded.

The data underlying this figure is available in Supplementary Table 1 in the accompanying data set.

Figure A4. Proportion of people notified with TB within 3 days of diagnosis by UKHSA centre, England, 2022 (Supplementary Table 3)

The data underlying this figure is available in Supplementary Table 3 in the accompanying data set.

Figure A5. Proportions of culture confirmation for all people with TB and for people with pulmonary TB by UKHSA region, England, 2022

Note: the line at 80% indicates the target of 80% for culture confirmation of pulmonary TB.

The data underlying this figure is available in Supplementary Table 6 in the accompanying data set.

Table A1. Number of people notified, proportion with culture confirmation and proportion of notifications identified in a cluster with more than one other person in England, 2018 to 2022

Year	Total notified	Total cultured (n)	Total cultured (%) 95% CI	Total identified in a cluster with more than one person (n)	Total identified in a cluster with more than one person (%) 95% CI
2018	4,610	2,866	62.2 (60.8 to 63.6)	1,108	24.0 (22.8 to 25.3)
2019	4,704	2,898	61.6 (60.2 to 63.0)	1,154	24.5 (23.3 to 25.8)
2020	4,122	2,537	61.5 (60.0 to 63.0)	919	22.3 (21.0 to 23.6)
2021	4,411	2,694	61.1 (59.6 to 62.5)	960	21.8 (20.5 to 23.0)
2022	4,380	2,742	62.6 (61.2 to 64.0)	862	19.7 (18.5 to 20.9)

Figure A6. The proportion of people in a cluster by UKHSA region, England 2018 to 2022

The data underlying this figure is available in Supplementary Table 18 in the accompanying data set.