Musculoskeletal tuberculosis in Bradford

Introduction

Tuberculosis (TB) is the biggest isolated communicable causes of mortality in the modern world (more than HIV/AIDS).¹ Between 2000 and 2017, a global decrease in both incidence and mortality, with deaths falling from 23% to 16%, followed a call for global focus on the disease led by the World Health Organization (WHO) and United Nations (UN).^1,2 Although TB is traditionally thought of as a pulmonary disease, it can manifest throughout the body, and reports of musculoskeletal (MSK) involvement stretch back to the late 18^th century, following Sir Percivall Pott’s description of tuberculous spondylitis in 1779.³ TB is a ‘notifiable’ diagnosis in the UK, a status requiring clinicians to notify statutory authorities, including Public Health England (PHE), of new cases. This enables accurate and timely monitoring of outbreaks or epidemics across the country.

The City and District of Bradford, in West Yorkshire, has a long industrial heritage stemming from iron, coal, and its textile mills in the 18^th and 19^th centuries.⁴ It also has a long modern history of immigration to the District, with settlers from Ireland and Germany in the mid-19^th century, and Poland, Ukraine, India, Pakistan, and Bangladesh in the post-war years of the 20^th century.^{5, 6} Today, Bradford is home to a large and vibrant immigrant South Asian population, one of the largest in the UK.⁷

Bradford Teaching Hospitals NHS Foundation Trust operates from six sites in the Bradford district, including two hospitals with inpatient capacity. It is responsible for hospital services to the city district of Bradford and surrounding communities, to a local population of over 500,000. Bradford Royal Infirmary (BRI) is a major teaching hospital with inpatient capacity of around 780 beds, and provides all the urgent/emergency services to the Trust. Local Clinical Commissioning Groups (CCGs) are responsible for the commissioning of district NHS services across the primary and secondary care groups, and general practitioners in the Bradford area make acute secondary care referrals directly to the BRI.

Previous studies from this department in 1974⁸ and 2007⁹ highlighted the high proportion of MSK TB in the immigrant population of this diverse, growing city.^8,9 The clinical landscape has changed significantly since each of these studies, with the emergence of HIV in the 1980s (a now frequently cited risk factor for coexistent TB infection), the increase in multidrug-resistant strains, as well as modernized screening, diagnostic tools, and pharmacological protocols over the last 15 years.

This study examines the local disease burden of MSK TB in the civic region of Bradford and reports clinical practice, patient outcomes, and common presenting factors to assist timely diagnosis and successful management of this scarcely seen disease. This will be achieved by reporting the demographic trends and local incidence of MSK TB over the preceding decade. These data, in addition to local treatment regimes, treatment outcomes, and association of risk factors within the study population, will be considered against the concurrent national context of declining incidence.

Methods

All cases were identified retrospectively from our TB service database, collating all new diagnoses of TB in Bradford Teaching Hospitals between January 2005 and July 2017. Following the definition detailed below, a new MSK TB database was created. Patient care records, pathology results, and imaging studies were reviewed for all cases. Patient demographic data were collated, including age, country of birth, and date of entry to the UK (where applicable). Further data were collated for site(s) of disease, positive culture/histology results, antibiotic chemotherapy regime employed, evidence of antibiotic resistance, the need for surgical management, and total duration of treatment. Contemporaneous local population data were acquired from the Office of National Statistics (ONS) national mid-year reports.¹⁰

MSK TB was defined as radiological evidence (on plain film radiograph, CT, or MRI)¹¹ of an extrapulmonary bone or soft-tissue focus, with or without specific radiological features of concern for TB, in addition to one or more of the following: biopsy demonstrating positive culture for Mycobacterium tuberculosis, positive histological evidence, or clinical features of concern for TB. Features of clinical concern included recent travel to high burden countries (within last two years), identification through contact tracing and clinical symptoms/signs of TB (persistent cough, night sweats, or weight loss). Cases not involving musculotendinous units or bone, such as nodal disease (cervical chain, axillary, etc.), intracranial/intrapharyngeal, and intrathoracic foci, and visceral and genitourinary pathology, were excluded. MSK TB cases also involving pulmonary/visceral foci were included; however, mention of the non-MSK foci was omitted. It is worth noting that further modern diagnostic tests, such as genetic assays (e.g. polymerase chain reaction) or other molecular techniques, were subsequently introduced to our department, post-dating the study period.

Annual population estimates were collected from the ONS annual mid-year population estimate reports.¹⁰ Estimated annual incidence figures (presented in cases per 100,000 population, per calendar year) were calculated using the ONS mid-year population data and annual number of new diagnoses identified through interrogation of our local TB database. The regional population, as described by the ONS report, was understood to equate to the acute catchment area of BRI and its TB service.

Results

Demographic details and incidence

A complete summary of the data can be found in Supplementary Table i. Over 12.6 consecutive years, between January 2005 and July 2017, 109 cases of MSK TB were diagnosed and treated in Bradford Teaching Hospitals (Table I; Figure 1 and Figure 2).

Table I.

Disease incidence in Bradford for years 2005 to 2016. 2017 data (partial) are not included. Calculated incidence is given as cases per 100,000 population per calendar year.

Year	Total TB diagnoses	MSK TB diagnoses, %	MSK TB diagnoses, %	Local population, n*	Incidence MSK TB (per 100,000)
2005	146	7	4.79	491,361	1.31
2006	152	6	3.95	495,658	1.12
2007	148	8	5.41	501,357	1.50
2008	149	8	5.37	507,323	1.50
2009	196	18	9.18	512,392	3.37
2010	154	7	4.55	518,002	1.31
2011	153	16	10.46	523,115	2.99
2012	158	9	5.70	524,386	1.68
2013	139	8	5.76	525,936	1.50
2014	91	5	5.49	527,567	0.94
2015	97	6	6.19	529,879	1.12
2016	82	8	9.76	532,539	1.50
Total	1,665	106
Mean (SD)	138.75 (20.76)	8.83 (4.00)	6.38 (2.16)	515,792.92 (13,969.87)	1.65 (0.75)

1. *

   Office of National Statistics mid-year population estimates 2001 to 2017.¹⁰

1. MSK, musculoskeletal; SD, standard deviation; TB, tuberculosis.

Fig. 1

Graph showing tuberculosis (TB) incidence 2005 to 2016 for the Bradford locality and the UK. Upper and lower Poisson confidence intervals for local calculated incidences were all < 4 × 10^-4.

Fig. 2

Graph showing local musculoskeletal tuberculosis (MSK TB) incidence by year from 2005 to 2016. Upper and lower Poisson confidence intervals for local calculated incidences were all < 5.5 × 10^-5.

Mean patient age in the dataset was 41.2 years (14 to 86). The observed male:female sex ratio was 62:47. Subsequent loss to follow-up prior to completion of treatment was 4/109 cases (3.7%). Median follow-up/treatment duration was 12 months (interquartile range (IQR) 6 to 12); mode was 12 months. Treatment duration ranged from six to 24 months.

Nationality, immigration status, and ethnic background are routinely recorded at first clinical review for purposes of risk stratification, contact tracing, and disease surveillance. In total, 20/109 patients were confirmed as having been born in the UK. The remaining 89 cases were all members of the local immigrant population (Table II). A chi-squared test for independence was performed using data from the ONS 2011 census (most recent) to determine the relation of MSK TB diagnosis within the immigrant population versus UK-born local populations,⁷ which demonstrated that members of the immigrant population were more likely to be diagnosed with MSK TB than UK-born locals (p < 0.001).

Table II.

New diagnoses of musculoskeletal tuberculosis by patient country of origin in Bradford (2005 to 2017).

Origin	n
Indian subcontinent*	73
India	16
Pakistan	50
Bangladesh	5
UK	20
Sub-Saharan Africa†	11
Eritrea	3
Gambia	2
Somalia	1
Zimbabwe	1
Uganda	1
South Africa	1
Malawi	1
Other
Trinidad	1
Slovakia	1
Afghanistan	1
Yemen	1
Indonesia	1
Total, n	109

1. *

   Two unknown.

1. †

   One unknown.

Of the 89 patients born elsewhere, 83 provided a date of entry to the UK. Mean time from entry to the UK to diagnosis was 13.76 years (0.75 to 54) and median was five years. The data showed a large variance with an IQR of 23 years. There was significant skewing of the dataset (Figure 3).

Fig. 3

Time from patient entry to UK to tuberculosis diagnosis.

Discussion

In the UK during 2011, 8,963 new cases of TB were diagnosed and notified to PHE (an incidence of 14.4/100,000). Of these new cases, 83.6% occurred among a population born outside the UK.¹² Incidence rates have continued to fall in England year on year since 2011 (Figure 1). In 2017, PHE were notified of a total 5,102 new diagnoses.¹³ For the first time, incidence during this period fell below ten cases per 100,000, classifying England as a low-incidence country according to WHO definitions.¹³

Mean annual incidence of TB remained high in Bradford between 2005 and 2017, at 27 cases per 100,000. This figure is similar to a reported 26 per 100,000 between 1999 and 2004.⁹ In 2017, by CCG, Bradford district was reported as the highest incidence (outside of London) at ≥ 40.0/100,000 in 2017,¹³ and has consistently reported a high disease burden for TB year on year. Despite ongoing improvement in national incidence over the last decade, local incidence remains high, although it has more recently displayed a downward trend. Mean annual incidence of MSK TB was found to be 1.65 cases per 100,000, although a large variance in the data precludes confident assertion of change over the time period (Figure 2).

Data indicated a local surge of TB diagnosis in 2009. The majority of these cases were found to be due to reactivation of latent infection. Screening for latent disease has become much more common, particularly in those entering the UK from high-burden countries around the world.

The 2011 national census results for the Bradford district reflected its famously diverse ethnicity, with 24.89% of the population declaring itself as Indian, Pakistani, or Bangladeshi,⁷ three of the WHO’s “high-burden countries” which account for over 80% of the world’s TB cases.¹ Proportion of new TB incidence in patients originating from outside the UK has been falling in recent years; although still a clear majority, incidence in the immigrant population has halved since 2012 and is at its lowest rate since 2000.¹³ Notably, while an increased incidence in the local immigrant population is certainly not unique to Bradford, the extent of its continued prevalence cannot be said to reflect national trends.

While MSK TB is an uncommon presentation to the majority of orthopaedic departments in this country, Bradford has a reported history of high MSK TB prevalence dating back to the mid-1950s.⁸ A more recent local study in 2007 referenced the resurgence of pulmonary TB (and concomitant MSK TB) within an industrialized conurbation.⁹ TB prevalence remains higher in developing countries, and immigration from those countries has long been linked to an increase in disease incidence in destination countries.^14,15

The diagnosis of MSK TB is notoriously challenging and has often been said to require a high index of suspicion, as appearances to admitting clinical staff may mimic more pedestrian infective processes.^16,17

Our data clearly show the overwhelming majority of MSK TB cases to be diagnosed in the immigrant population: 81.7% of all MSK TB diagnoses in the study period (89 patients) were born outside the UK, and 67% of the cohort (73 patients) originated from one of the three countries of risk mentioned above (India, Pakistan, and Bangladesh). As such, 67% of new MSK TB diagnoses came from an ethnic background representing only 25% of the local population⁷. This fact reinforces the importance of proactive investigation and intervention in at-risk subsections of the community.

Of 83 cases in patients born outside the UK, time to diagnosis from entry to the UK was widely varied. Although from the data available it was not possible to establish the proportion of de novo active MSK TB compared to reactivated latent disease, the descriptive statistics (in particular data range of 0.75 to 54 years) suggest a large amount of latent disease being diagnosed many years later subsequent to reactivation. Half of these cases, however, were diagnosed within three years of entry to the UK, allowing for a considerable mix of de novo active disease and reactivated disease. It was not possible to ascertain likely relative ratio of these presentations from the data retrospectively; indeed, this would be challenging in a prospective study. Furthermore, data for these cases regarding recent travel and diagnosis subsequent to contact tracing were, unfortunately, not available.

The most common single focus of TB infection in this country remains pulmonary (54.4% in 2017).¹³ The relative proportion of gross extrapulmonary TB diagnoses in the UK has, however, steadily increased from 40.9% (2004) to 58% (2017). Of extrapulmonary TB cases in 2017, 6% were sited in bone, including the spine. The comparable rate of MSK TB within the local TB data shows an annual mean of 6.38% (SD 2.16; 2 to 10), which is proportionate to national incidence data.¹³

The most common site for MSK TB in our cohort was the thoracolumbar spine, as is frequently quoted elsewhere.^18,19 Although spinal infection in general remains a common diagnosis, a high index of suspicion is important to identify TB infection, despite the characteristic classical descriptions of Potts’ disease on plain film radiographs.³ Diagnosis of TB has historically been, and continues to be, avoidably delayed.^20-23 Advancing spinal infection can lead to significant instability, structural deformity, and significant neurological dysfunction or myelopathy if left untreated, requiring major surgical intervention.²⁴ Objective improvement in motor symptoms was common postoperatively in our cohort; however, some persistent features of altered sensation or hyperalgesia were frequently reported at follow-up. Typically, treatment of simple infective paraspinal collections or myelodiscitis remains non-surgical.²⁴ Nonetheless, radiologically guided aspiration/biopsy and culture for acid-fast bacilli in refractory uncomplicated spinal infections would seem germane.

A total of 38 cases (34.8%) required surgical intervention, principally in the form of debridement. The mainstay of treatment for MSK TB would appear to be pharmacological, with debridement and/or drainage employed only when specifically indicated.

Reports of drug-resistant, or multidrug-resistant strains of TB (MDR TB) have been rising globally for years. In 2017, 3.5% of new TB cases globally demonstrated multidrug resistance, a small increase from the previous year,¹ whereas reports from Europe in the last decade have revealed up to 35.5% of new cases identified as MDR TB.²⁵ Total case numbers of MDR TB on initial diagnosis in the UK from 2016 to 2017 have fallen from 60 to 55, respectively. However, the proportion of overall incidence for these consecutive years remained similar (1.7% and 1.8%).¹³ Overall prevalence of resistant TB cases in our dataset was six from 109 (5.5%). Of note, three of these six cases displayed features of Isoniazid resistance, previously reported as significant in the Bradford region,²⁶ and the remaining three cases demonstrated MDR features. All but one case were subsequently successfully treated, suggesting that within our locality, MDR TB may not be as pertinent an issue as elsewhere.

In developing countries, TB remains one of the leading opportunistic causes of morbidity/mortality in the HIV-positive population.²⁷ Of the estimated ten million individuals diagnosed worldwide with active TB in 2017, approximately 9% are thought to have been HIV-positive, and approximately 300,000 people around the world died from TB coexistent with HIV.¹ There were no identified cases within our dataset of concomitant HIV, and only two patients from 109 with positive Hepatitis B serology. Viral hepatitis, syphilis, and HIV co-infection have long been posited as risk factors for TB; the prevalence of these co-infections has been described repeatedly as linked to cohorts observed in conditions of poverty, incarceration (particularly in developing countries), or areas of poor sanitation.^28,29 High rates of HIV co-infection are typically reported from sub-Saharan Africa where the HIV burden is very high and may not be directly comparable to our local demographic.¹⁸ Although Bradford is in a developed country, it has recently been ranked fifth most income-deprived local authority in England.³⁰ Social deprivation remains an issue frequently reported in the local TB population, and is likely to influence Bradford’s high disease burden; furthermore, endemic drug abuse and poor sanitation are also frequently reported in Bradford, and are known to contribute to the spread of this disease.^28,31

At the beginning of the 2010s, following the highest reported incidence of TB in England since the 1970s, improvement of national data for completion of treatment was urgently mandated.³² Peak incidence in England was reported during 2011 (8,963 cases).¹² The WHO established a minimum target of 85% completion of treatment within 12 months of therapy commencing, and reported national data still fail to achieve this. Treatment completion rate was reported as 84.4% for cases notified in 2016.¹³ Despite a significant level of poor outpatient attendance, only four patients (3.7%) were formally lost to follow-up. Two further patients had evidence of ongoing disease, and so six patients in total (5.5%) had not successfully completed treatment, leaving 94.5% of the study group who successfully completed their treatment regime. As such, our local performance against WHO targets is strong.

The limitations of this study reflect its observational and retrospective nature. Although the local clinical TB database was exhaustive, certain helpful data on recent travel and the proportion of new diagnoses achieved through contract tracing would have added further strength to our conclusions. In addition to this, as stated, the local population was understood to equate to the acute catchment area of BRI. This may not be entirely accurate, due to possible overlap between some CCG boundaries and nearby geographical catchment boundaries between neighbouring Trusts, such as Leeds Teaching Hospitals NHS Trust and Airedale NHS Foundation Trust. Some smaller parts of the granular data were absent from the database, but despite its size and the length of time reflected this was only a small amount; for example, the time between arrival in UK and diagnosis was unavailable in only six patients. One patient had no visible imaging to review due to diagnostic imaging being performed elsewhere, and one patient had no information on sensitivities available. Data on anatomical focus, the incidence of biopsy, and incidence of successful microbial culture were available for all but one patient, who had no clinical data available at all. As such, the strength of our database was felt to be sufficiently robust so as to exert only minimal impact on the validity of our conclusions.

Once again, the data from our unit show a preponderance of MSK TB within the immigrant population compared to the UK-born population, especially within the cohort of the immigrant population originating from the Indian subcontinent. The success rate of pharmacological treatment for MSK TB in Bradford was high.

Although annual TB incidence in the region continues to fall, Bradford remains one of the areas in the country worst affected by the disease. Consequently, the data reinforce the need for a high level of suspicion in the prompt and successful diagnosis of MSK TB in geographical areas with a large proportion of at-risk populations, recently highlighted by Broderick et al.²³ Notably, however, relative MSK TB prevalence in Bradford was comparable to national data. The level of local social deprivation may bear closer scrutiny as to its association with increased MSK TB prevalence in the immigrant population.