Tuberculosis in Takayasu arteritis: a retrospective study in 1105 Chinese patients.

BACKGROUND: Tuberculosis (TB) infection has been reported to have a possible relationship with the occurrence and clinical course of Takayasu arteritis (TA). We aimed to describe the characteristics of TB in a large population of TA patients.
METHODS: We included a total of 1105 patients with TA, who were hospitalized between January 1992 and December 2017. Comparisons of clinical features were made according to the presence of TB.
RESULTS: Among the 1105 patients, 109 (9.9%) had TB, including 53 patients (48.6%) diagnosed with TB before the onset of TA, 23 (21.1%) with a concurrent diagnosis of TB and TA, and 24 patients (22.0%) who developed TB after TA. Pulmonary TB was the most frequently identified (97 patients, 89.0%). Patients with TB had more frequent involvement of the pulmonary artery and experienced more chest discomfort and constitutional symptoms but had less interventional treatment. Demographic characteristics, comorbid diseases, and use of steroids were similar between patients with and without TB.
CONCLUSIONS: The proportion of Chinese TA patients with TB was not low, and about half of the patients had TB before TA. Pulmonary TB was the most common. Pulmonary artery involvement and pulmonary hypertension was more frequent in TA patients with TB.

Introduction

Takayasu arteritis (TA) is a rare, chronic vasculitis, which mainly involves the aorta and its main branches. Coronary arteries and pulmonary arteries can also be involved. Initially, patients with TA present constitutional symptoms, including low-grade fever, fatigue, night sweats, and weight loss. As inflammation of the artery walls progresses, symptoms of organic ischemia appear, such as weak pulse, exertion fatigue of the limbs, renovascular hypertension, and even ischemic stroke.[1],[2] Prednisone and immunosuppressive agents are essential anti-inflammatory medications and can achieve and maintain remission in most patients.[3] TA is more common in Asia area, Turkey, Mexico, and South Africa.[4]–[6] The exact etiology of TA remains to be elucidated. Genetic, environmental, and autoimmune factors have been suggested to have important roles.[7]

Infection with Mycobacterium tuberculosis (Mtb) is very common, affecting one-third of the world's population, although only a small proportion of people with Mtb infection (nine million new cases annually) develop tuberculosis (TB).[8] The prevalence of TB is relatively high in China, although the prevalence of TB is continuously decreasing in the Chinese population.[9] Previous studies have shown that the proportion of the general Chinese population with a history of TB is between 1.36% and 1.69%.[10],[11] China remains a country with one of the highest TB burdens and the rate of decline is slower than expected, according to the Global Tuberculosis Report of 2017 published by the World Health Organization.[12] TB is reportedly related to TA, based on observational data. Compared with healthy Brazilian children, the ratio of those with a positive tuberculin skin test (TST) was higher in children with TA.[13] A Korean study showed that the incidence of TB in patients with TA (17%) was higher than that of the general population (5.5%–5.8%).[14] Pathological characteristics of both TA and TB are granulomas and caseous necrosis.[15],[16] Some studies have reported the concurrence of TA and TB, and they have suggested that anti-inflammation therapy should be carefully initiated after the control of active TB.[17]–[19] Despite these study findings, how TB manifests in patients with TA remains to be elucidated in a large population.

In this retrospective study, we aimed to describe the proportion of TB and clinical features of patients with TA who have TB, and to determine whether TB affects the clinical course of patients with TA in a large patient population.

Methods

Study population

This study protocol was approved by the Ethics Committee of Fuwai hospital, and was conducted in accordance with the 1964 Helsinki declaration and its later amendments. All patients were given their informed consent for access to their medical records during hospitalization. We retrospectively reviewed the electronic medical records of 1105 patients with TA admitted to Fuwai Hospital from January 1992 to December 2017. All patients fulfilled the TA diagnostic criteria established by the American College of Rheumatology in 1990.[20] Angiographic classification was made according to the criteria set out by Hata and Numano.[21]

Disease activity was assessed using the modified National Institutes of Health criteria.[22] Patient was defined as having active disease if the patient met two or more of the followings: (1) systemic symptoms without another cause; (2) elevated erythrocyte sedimentation rate (ESR) and/or C-reactive protein (CRP) without infection, anemia, or other cause; (3) new onset or deterioration of vascular ischemia or inflammation; and (4) typical angiographic features.

TB was defined according to past medical history and the first discharge diagnosis at our hospital. We calculated the age of the first TB infection and TA symptom onset, respectively. Active TB was diagnosed based on the combination of TB symptoms (cough, sputum, hemoptysis, fever, and chest pain for more than two weeks), chest radiography, and bacteriological examination (sputum smear microscopy or culture).[9] The TST was defined as strongly positive if the diameter of induration was more than 20 mm or with the presence of blistering, hemorrhage, necrosis, or lymphangitis. TSTs and interferon-gamma release assays (IGRAs) are not done routinely in patients with TA. They were performed if a patient was suspected with TB by physicians. People with latent tuberculosis infection (LTBI) were infected with and had immune sensitization to Mtb and were asymptomatic. We did not describe LTBI in the present study.

Data collection

Artery lesions (including lesions in the pulmonary artery) were determined with catheter angiography, digital subtraction angiography, computed tomography angiography, and magnetic resonance angiography. For carotid and subclavian arteries, sonographic results were acceptable. We defined results of 18F-fluorodeoxyglucose-positron emission tomography/computed tomography (18F-FDG PET/CT) as PET-active if the maximum standardized uptake value (SUV) of the region of interest was higher than the average SUV in liver.

Laboratory findings, such as hemoglobin, white blood cell count, serum creatinine, blood urea nitrogen, ESR, and CRP at the time of admission were noted and compared between patients with and without TB.

Concomitant diseases, such as hepatitis virus B infection, stroke, hypertension, hyperlipidemia, mellitus diabetes, aortic regurgitation, and pulmonary hypertension were also recorded. Pulmonary hypertension was diagnosed according to the results of right heart catheterization (RHC) or estimated pulmonary arterial systolic pressure (PASP). Pulmonary hypertension was defined as a mean pulmonary arterial pressure > 25 mmHg at rest by RHC or PASP > 40 mmHg by transthoracic echocardiography in patients without RHC.[23]

Statistical analysis

Continuous variables were expressed as mean ± SD for normally distributed data and as median (interquartile rank) for data with a skewed distribution. Classification variables were described as percentages. Differences were examined using the independent t-test, chi-square test, Fisher's exact test, or Mann–Whitney U test, as appropriate. Analysis was performed using IBM SPSS software, version 19.0 (IBM Corp., Armonk, NY, USA). We analyzed the proportion of TB in TA on a five-yearly basis, as follows: 1992.01–1996.12, 1997.01–2001.12, 2002.01–2006.12, 2007.01–2011.12, and 2012.01–2017.12. The time trend in the proportion of patients with TA and TB was analyzed using the Cochran–Armitage test (R version 3.5.1). Statistical differences were set at P-value < 0.05.

Results

Demographic data

Of the 1105 patients with TA, 109 patients (9.9%) had TB infection. The proportion of patients with TA and TB decreased from 13.7% to 8.3% over time. However, the decrease was not significant (Z = 1.132, P = 0.190). The sex ratio was 1 : 3.8 (23 men and 86 women). The mean age at TA onset was 26.7 ± 10.0 years, and the duration from symptom onset to first hospitalization was 56.3 (20.1–143.4) months. These data were similar between patients with and without TB (Table 1).

Table 1.

Demographic information and clinical features of the 1105 TA patients regarding TB.

Variables	Without TB	With TB	P-value
Total	996 (90.1%)	109 (9.9%)	0.523
Female	809 (73.2%)	86 (7.8%)
Male	187 (16.9%)	23 (2.1%)
Age at symptom onset, yrs	26.5 ± 10.2	27.6 ± 10.0	0.281
Duration of clinical course, months	57.3 (18.1–156.0)	56.3 (20.1–143.4)	0.848
Symptoms and signs
Fever or fatigue	87 (8.7%)	17 (15.6%)	0.020*
Carotidynia	14 (1.4%)	1 (0.9%)	1.000
Dizziness	262 (26.3%)	30 (27.5%)	0.784
Headache	109 (10.9%)	13 (11.9%)	0.756
Syncope	48 (4.8%)	9 (8.4%)	0.123
Amaurosis	39 (3.9%)	1 (0.9%)	0.171
Visual disorders	74 (7.4%)	6 (5.5%)	0.462
Weak pulse	164 (16.5%)	17 (15.6%)	0.815
BP discrepancies between arms	39 (3.9%)	7 (6.4%)	0.207
Easy fatigability	136 (13.7%)	14 (12.8%)	0.815
Intermittent claudication	75 (7.5%)	6 (5.5%)	0.441
Exertional dyspnea	173 (17.4%)	36 (33.0%)	< 0.001*
Chest tightness	218 (21.9%)	27 (24.8%)	0.491
Chest pain	81 (8.1%)	4 (3.7%)	0.097
Hemoptysis	14 (1.4%)	8 (7.3%)	0.001*
Cough	25 (2.5%)	4 (3.7%)	0.520
Comorbid diseases
Hypertension	544 (54.6%)	60 (55.0%)	0.932
Dyslipidemia	110 (11.0%)	8 (7.3%)	0.234
Diabetes mellitus	21 (2.1%)	0 (0.0%)	0.255
Stroke	51 (5.1%)	4 (3.7%)	0.508
Angina	74 (7.4%)	5 (4.6%)	0.274
Renal dysfunction	39 (3.9%)	3 (2.8%)	0.791
Heart failure	63 (6.3%)	11 (10.1%)	0.135
Hepatitis B infection	17 (1.7%)	4 (3.7%)	0.145
Aortic regurgitation	240 (24.1%)	21 (19.3%)	0.788
Pulmonary hypertension	107 (10.7%)	21 (19.3%)	0.008*
NIH active	482 (48.4%)	67 (61.5%)	0.010*

Data are presented as means ± SD, median (interquartile range) or n (%). *P < 0.05. BP: blood pressure; TA: Takayasu arteritis; TB: tuberculosis.

TB in patients with TA

Pulmonary TB was most common in 97 patients (89.0%), followed by TB lymphadenitis in eight patients (7.3%), cutaneous TB in two patients (1.8%), joint TB in one patient (0.9%), intestinal TB in one patient (0.9%), and pericardial TB in one patient (0.9%). Among the 97 patients with pulmonary TB, fourteen patients had TB pleurisy. Of the 109 patients with TA and TB, 53 patients (48.6%) were diagnosed with TB before symptom onset of TA, and the interval was 8.0 (3.0–14.5) years. Twenty-four patients (22.0%) developed TB during the follow-up period of TA, and the period between onset of TA and TB was 4.5 (2.0–8.8) years. Nine patients (8.3%) did not have a confirmed time of TB infection. Twenty-three patients (21.1%) were found to have TB concurrently with TA. Among them, ten patients were diagnosed in our hospital.

Clinical manifestations

The proportion of active TA in patients with TB was 61.5% versus 48.4% in patients without TB (P = 0.010). Pulmonary artery involvement was more common in patients with TB (31.2% vs. 17.3%, P = 0.001). There were no significant differences between patients with and without TB regarding involvement of the aorta and its major branches, except for involvement of the aortic arch (P = 0.019). The angiographic classification was similar between patients with and without TB. The ratio of occlusion (19.3% vs. 23.9%) and dilation (18.3% vs. 16.2%) of arteries was similar in patients with and without TB (P > 0.05) (Table 2).

Table 2.

Hata's angiographic classification, detailed artery lesions, and interventions in TA patients according to TB.

Variables	Without TB	With TB	P-value
Angiographic type (n = 1062)			0.606
I	269 (28.1%)	28 (27.2%)
IIa	45 (4.7%)	7 (6.8%)
IIb	62 (6.5%)	4 (3.9%)
III	131 (13.7%)	10 (9.7%)
IV	92 (9.6%)	10 (9.7%)
V	360 (37.5%)	44 (42.7%)
Lesions
Ascending aorta	109 (10.9%)	12 (11.0%)	0.899
Thoacic aorta	188 (18.9%)	16 (14.7%)	0.502
Abdominal aorta	297 (29.8%)	29 (26.6%)	0.741
Aortic arch	37 (3.7%)	10 (9.2%)	0.019*
Left subclavian artery	497 (49.9%)	50 (45.9%)	0.529
Right subclavian artery	284 (28.5%)	29 (26.6%)	0.734
Left vertebral artery	81 (8.1%)	8 (7.3%)	0.947
Right vertebral artery	60 (6.0%)	5 (4.6%)	0.707
Left carotid artery	343 (34.4%)	30 (27.6%)	0.182
Right carotid artery	232 (23.3%)	26 (23.9%)	0.814
Branchiocephalic artery	83 (8.3%)	10 (9.2%)	0.986
Axillary artery	60 (6.0%)	11 (10.1%)	0.090
Left renal artery	329 (33.0%)	28 (25.7%)	0.142
Right renal artery	312 (31.3%)	33 (30.3%)	0.475
Superior mesenteric artery	107 (10.7%)	8 (7.3%)	0.199
Coeliac trunk	68 (6.8%)	4 (3.7%)	0.126
Iliacofemoral artery	103 (10.3%)	9 (8.3%)	0.522
Pulmonary artery	172 (17.3%)	34 (31.2%)	0.001*
Coronary artery	51 (5.1%)	5 (4.6%)	0.810
Occlusion	238 (23.9%)	21 (19.3%)	0.279
Dilatation	161 (16.2%)	20 (18.3%)	0.559
Stent implantation	235 (23.6%)	16 (14.7%)	0.035*
PTA	427 (42.9%)	36 (33.0%)	0.048*

Data are presented as n (%). *P < 0.05. PTA: percutaneous transluminal angiography; TA: Takayasu arteritis; TB: tuberculosis.

More patients with TB experienced constitutional symptoms and hemoptysis (P < 0.05). Two patients had thin-walled cavity, which did not cause hemoptysis and respiratory complications. Dizziness, chest tightness, and dyspnea were the top three symptoms reported at first hospitalization by all patients whereas dyspnea was more common among patients with TB (P < 0.05). The most common sign was weak pulse in patients with and without TB (15.6% vs. 16.5%, P > 0.05) (Table 1).

Pulmonary hypertension occurred in 21 patients (19.3%) with TB and in 107 patients (10.7%) without TB (P = 0.008). Other concomitant diseases showed similar prevalence between patients with and without TB (P > 0.05) (Table 1).

Laboratory findings

The level of ESR was significantly higher in patients with TB. Other relevant laboratory results did not show significant differences (Table 3).

Table 3.

Laboratory findings in TA patients according to TB.

Variables	Without TB	n	With TB	n	P-value
C-reactive protein, mg/L	4.0 (2.1–12.0)	646	4.37 (2.42–16.5)	63	0.233
Erythrocyte sedimentation rate, mm/h	13.0 (6.0–30.0)	729	17.0 (9.0–45.0)	87	0.010*
White blood cells, ×109/L	7.12 (5.85–8.68)	510	7.39 (6.40–8.40)	54	0.328
Hemoglobin, g/L	125.0 (113.0–140.0)	591	124.0 (111.5–142.0)	61	0.733
Serum creatinine, umol/L	63.75 (53.71–74.89)	483	65.21 (53.78–75.93)	51	0.723
Blood urea nitrogen, mmol/L	5.0 (4.0–6.26)	494	5.0 (4.5–6.5)	54	0.426

Data are presented as median (interquartile range). *P < 0.05. TA: Takayasu arteritis; TB: tuberculosis.

18F-FDG PET/CT with elevated SUV in lymph nodes

Six patients had 18F-FDG PET/CT referred to the SUV of lymph nodes. Elevated SUV of the mediastinal and cervical lymph nodes was found in three patients and of the hilar and supraclavicular lymph nodes in two patients. The diameter of lymph nodes ranged from 1.2 to 2.0 cm, and their SUVs ranged from 3.9 to 8.7. TB infection was ruled out in three patients. One patient had pulmonary TB fourteen years earlier, and two patients had active pulmonary TB. No patients underwent lymph node aspiration.

Treatment

The initiation of prednisone at discharge was similar with respect to TB (56.0% vs. 57.0%). The most frequent dose at discharge was 20 mg per day, followed by 30 mg per day. Intervention treatment for revascularization, percutaneous transluminal angioplasty and stent implantation was performed more frequently in patients without TB than in those with TB (both P < 0.05) (Table 2).

Ten patients had active TB. Elevated ESR was found in eight patients and elevated CRP was found in three patients. They were treated with anti-TB drugs (isoniazid, rifampicin and pyrazinamide), none of them had TB diffusion during follow-up (Table 4).

Table 4.

Detailed information for TA patients with active TB.

Number/Gender/Age	Type	Clinical duration, months	Symptoms & Signs	ESR, mm/h	CRP, mg/L	Hb, g/L	WBC, ×109/L	Evidence for active TB	Dosage of prednisone per day, milligram
1/Female /19	III	84	Extertional chest tightness and dyspnea	75	29.1	-	-	Imaging: cavitary pulmonary tuberculosis	0
2/Female /21	V	48	Weak pulse, cardiac murmurs	26	-	-	-	TST (++++)	0
3/Female /21	VI	12	Fatigue, HTN	40	4.3	109	11.3	Triple anti-tuberculosis regimen on admission	0
4/Female /54	V	216	Dizziness, headache, CHF	75	-	73	-	Imaging: millary pulmonary tuberculosis	20
5/Female /23	V	2	HTN	28	2.7	112	11.1	Triple anti-tuberculosis regimen on admission	20
6/Female /33	III	246	Headache, CHF, CRF	32	2.1	111	7.8	Triple anti-tuberculosis regimen on admission	0
7/Female /39	VI	156	HTN, dizziness	67	26.7	108	10.9	Elevated SUV of carotid lymph nodes, T-SPOT.TB(+)	0
8/Female /45	I	188	Fever, stroke	27	2.6	119	16.0	TST (++++)	0
9/Female /13	V	18	Fever, cough, carotidynia, discrepancy of blood pressure	6	14.8	139	11.0	Triple anti-tuberculosis regimen on admission	15
10/Female/55	VI	18	HTN, stroke, cough, homeptysis	19	2.0	128	7.5	T-SPOT.TB(+); PET/CT: Patchy shadow in right lower lobe	0

All of the patients received triple anti-tuberculosis medication and prednisone was used in three patients (Number: 4, 5, 9). Type refers to angiographic classification according to Hata criteria. CHF: chronic heart failure; CRF: chronic renal failure; CRP: C-reactive protein; ESR: erythrocyte sedimentation rate; Hb: hemoglobin; HTN: hypertension; TA: Takayasu arteritis; TB: tuberculosis; TST: tuberculin skin test; WBC: white blood cells counts.

Discussion

In this retrospective study, we found that TB was common in patients with TA, most patients had TB before the onset of TA, and pulmonary TB was predominant. These findings were consistent with results of studies conducted in other populations.[14],[24] Interestingly, compared with patients who did not have TB infection, the ratio of pulmonary artery involvement was higher in patients with TB infection. Furthermore, symptoms related to pulmonary artery involvement, such as pulmonary hypertension, exertional dyspnea, and hemoptysis, were also more common in patients with TB. In addition, we found that more patients with a history of TB were in active disease stage, indicating that TB infection may affect the clinical course and treatment of TA.

Several explanations for how TB triggers TA have been proposed, but the exact mechanism and relationship between TA and TB remain to be elucidated. One possible hypothesis involves superantigens. Mtb widely activates T cells, dendritic cells, and macrophages, mainly via the lipoarabinomannan on its wall structure. IL-12, IL-18, and resuscitation-promoting factor promote the differentiation of CD4+ T cells into Th1 and Th17 lymphocytes.[25]–[27] A variety of cytokines, including tumor necrosis factor alpha (TNF-α), IL-1, IL-6 and MMP-1, are upregulated in patients with TA and TB.[28],[29] Autoimmune disorder is another a possible explanation. The 65 kDa heat shock protein of Mtb, mHSP65, is implicated in various rheumatic diseases, indicating a relationship between Mtb infection and autoimmune disorders.[30] Cross-reactivity of mHSP65 and human HSP60 (hHSP60) has also been proved in patients with TA. Compared with healthy controls, levels of anti-hHSP60 antibodies were found to be substantially elevated in patients with TA.[31],[32] No studies to date have detected Mtb in artery walls using acid-fast staining.[33],[34] Furthermore, polymerase chain reaction results for the specific gene sequence IS6110 of Mtb are controversial. A previous study reported detection of the IS6110 sequence of Mtb by polymerase chain reaction in 70% of 33 tissue samples from patients with TA.[35] A recent study from Brazil reported that the IS6110 sequence could not be amplified in peripheral blood from 32 patients with TA or in artery tissue samples from 10 patients with TA.[34] This indicates that persistent Mtb infection is not necessary for maintaining inflammation in TA.

Elevated ESR or CRP, fever, and night sweats are indications of active TA, according to the modified National Institutes of Health criteria; however, these could well be owing to active TB. These characteristics are strong indications for initiating anti-inflammatory regimens in patients with TA. Physicians should be attentive for active TB in patients with TA, as treatment with glucocorticoids could cause multifold increased risk of Mtb infection and activation.[36]

Some reports stated that anti-inflammatory therapy should be used in patients with active TA and active TB,[17]–[19],[37],[38] but the appropriate time for initiating anti-inflammatory therapy was not determined. Anti-TB therapy should be completed in patients with active TB, according to the World Health Organization guidelines and global recommendations.[39] Prednisone, immunosuppressive drugs and biologics are three types of anti-inflammatory therapy used in active TA. Prednisone, the first-line anti-inflammatory drug in TA, is inexpensive and effectively controls inflammation. It has been suggested that active TB should be routinely screened in patients taking high dose steroids for a long period (over 15 mg per day for more than two weeks).[40] Azathioprine, the most common immunosuppressive agent used in TA, is contraindicated in active TB.[41] Methotrexate is safe in active TB if it is used with isoniazid,[42] but cyclophosphamide should not be used with rifampicin because rifampicin increases the toxic metabolites of cyclophosphamide.[41] TNF-α inhibitor, a kind of biologic agent used in refractory TA, is contraindicated in active TB; it should be used after the control of active TB.[43] Like in TNF-α inhibitor, TB should also be ruled out before initiating tocilizumab, which blocks IL-6 receptor.[44] Rituximab, a monoclonal antibody leading to the depletion of B cells, was proved to be safe in a patient with active TB.[37]

In the present study, prednisone was the predominant anti-inflammatory therapy. The use of prednisone is sufficient and long-term, with slow tapering. In our hospital, we use half-dose prednisone, which has similar efficiency as full-dose therapy.[45] Inflammation and active TB were well treated in our patients with active TA and TB by triple (isoniazid, rifampicin, and pyrazinamide) anti-TB therapy and half-dose prednisone.

Some studies suggest that during immunomodulatory treatment, annual screening of TB should be done in patients with TB exposure.[46],[47] The diagnosis of active TB is very important but remains difficult in patients with active TA. Constitutional symptoms of the two diseases overlap. The sensitivity of sputum smear examination varies and culture requires several weeks for the long generation time of Mtb.[48] Results of IGRAs (T-SPOT.TB and QuantiFERON-TB Gold In-Tube), TST, and elevated ESR should be interpreted carefully and considered with clinical manifestations and radiography.[49] 18F-FDG PET/CT is a new method for evaluating active disease in TA, but it cannot differentiate the causes of elevated SUV. 18F-FDG PET/CT is particularly useful in detecting the effect of anti-TB regimens when the diagnosis of active TB is already known.[50] We suggest that during treatment with a high dose of prednisone, if the patient presents recurrence or worsening of constitutional symptoms such as weight loss, night sweats, and low-grade fever, then chest radiography, IGRA, and sputum fast-acid staining should be performed to exclude TB infection. It is advisable to consult an infectious disease specialist for the diagnosis of active TB infection.

Limitations

Regarding study limitations, this was a retrospective study based on data from a single center for cardiovascular diseases; thus, there might be selection bias. Another shortcoming was that we had no results of lymph node or artery tissue biopsy. Nevertheless, considering the large sample size, comprehensive clinical and imaging evaluation of each patient, the clinical features of TA patients with TB could be well elucidated.

Conclusions

The proportion of Chinese patients with TA and TB was not low. About half of the patients had TB before TA symptom onset. Pulmonary TB was most common. Pulmonary artery involvement and pulmonary hypertension was more common in TA patients with TB. Active TB should be ruled out before and during anti-inflammatory therapy.