Systemic lupus erythematosus and tuberculosis: A review of complex interactions of complicated diseasesVNN Prabu, S Agrawal
Department of Rheumatology, Nizam's Institute of Medical Sciences, Hyderabad, India
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0022-3859.68653
Source of Support: None, Conflict of Interest: None
Infections, renal failure and cardiovascular disease account for the majority of mortality in systemic lupus erythematosus (SLE). Although most infections are caused by Gram-positive or Gram-negative bacteria, there is an increase in the incidence of Mycobacterium tuberculosis and other opportunistic infections that also account for increased mortality. The higher prevalence of tuberculous infections in SLE is attributed to multiple immune abnormalities seen in these patients. SLE and tuberculosis (TB) interact in complicated ways - they may have similar presentation and may mimic each other. In an individual patient, it becomes important to differentiate one from the other. In this review, we have highlighted the complex interactions of these diseases, the impact of one on the other and the various modalities available for the evaluation and management and their shortcomings. Considering the high prevalence of TB in India, it becomes all the more important for us to be aware of this interaction in our population.
Keywords: Extrapulmonary tuberculosis, lupus, Mycobacterium tuberculosis, pulmonary tuberculosis, systemic lupus erythematosus, tuberculosis
Systemic lupus erythematosus (SLE) is a systemic disease of unknown etiology, characterized by autoantibodies against self-antigens, resulting in inflammation-mediated multiorgan damage. Infections, renal failure and cardiovascular diseases account the for majority of deaths seen in SLE patients.  Although most infections are caused by Gram-positive or Gram-negative bacteria, there is an increase in the incidence of Mycobacterium tuberculosis (MTB) and other opportunistic infections in SLE that also account for increased mortality. 
Higher prevalence of tuberculous infections in SLE is attributed to multiple immune abnormalities that occur in these patients as well as to immunosuppressive therapy. ,, High doses of corticosteroids are also a major risk factor. Also, uncontrolled hyperactivity of the immune system actually makes SLE patients an immunocompromised host. Resistance to MTB, which is mediated by cellular immunity, is deficient in SLE patients both due to the nature of the disease and the immunosuppressive therapy. ,,,
SLE and tuberculosis (TB) interact in complicated ways - they may have similar presentations and may mimic each other. In an individual patient, it becomes important to differentiate one from the other. In this review, we have highlighted the complex interactions of these diseases, the impact of one on the other and the various modalities available for the evaluation and management. This review is based on a Medline search of the literature using a combination of terms such as "SLE," "Systemic Lupus Erythematosus" and "Lupus" with "Tuberculosis," "Pulmonary Tuberculosis," "Extrapulmonary Tuberculosis," "Mycobacteria" and "Mycobacterium tuberculosis."
The incidence of TB is increasing every year, with 9.4 million cases reported in 2008 as compared with 9.3 million cases in 2007. India ranks first in terms of the highest number of incident cases (1.6-2.4 million) in 2008. 
Expectedly, the likelihood of development of TB in lupus patients depends on the local prevalence and incidence of TB. In a systematic review, the incidence of TB in patients on immunosuppressive therapy was found to vary between 11.6 per 100 in India and 0 person years in the USA.  Although this review included all patients on immunosuppressive therapy and was not exclusively for SLE, it did underscore an important point in that the occurrence of TB in SLE is dependent on multiple factors and that the population prevalence of the disease has an important role. There are studies that have more specifically addressed the risk of TB in SLE patients in comparison to the general population. A Spanish study  found that the incidence of TB was six-fold higher in the SLE group compared to the general population. Similarly, Hong Kong , reported a five- to 15-fold higher risk and from India, a 10-60-fold higher risk was reported [Table 1]. ,
The tuberculous infection in SLE can be caused by both MTB and non-tuberculous mycobacterium (NTM) species. MTB infection however tends to occur earlier in the clinical course of lupus than NTM infection.  It is usually due to reactivation of latent infection or to reinfection. In contrast, NTM infections tend to occur in the latter course of the disease. Local implantation of the organism from skin abrasions is the likely route of transmission because most of these Mycobacteria are found in water and soil. Mycobacterium avium complex (MAC), M. chelonae and M. haemophilum cause skin infections in SLE. Acute monoarthritis is mainly caused by M. fortitum and MAC. 
The following characteristics of NTM infections may help in distinguishing them from MTB infections in lupus: insidious presentation, they tend to occur in heavily immunesuppressed SLE patients as compared with MTB infection, tends to have multiple sites of involvement (lung, skin, soft tissue, lymph node, bone and, rarely, disseminated). However, their clinical manifestations are not pathognomonic and may mimic other infections. Therefore, a tissue diagnosis is often required. 
Mok et al. studied the clinical manifestation of NTM and MTB infections in 725 lupus patients and reported 11 cases of NTM infection. They found that NTM infections mostly involve the extrapulmonary sites (skin and soft tissue) and are usually seen in patients with longer disease duration. Also, these patients had received a greater cumulative dose of prednisolone than those developing MTB infections. In this study, the disease duration was found to be the only independent predictive factor for NTM infections; severity of SLE and steroid dosage correlated positively with severity of TB and mortality. 
In SLE patients, extrapulmonary involvement is more frequent. , In a study by Zhang et al., of the 452 SLE patients on steroid/other immunosuppressive therapy, 42 were diagnosed to have tuberculous infection, of which 11 (23.8%) had exudative pulmonary TB and 31 patients (73.8%) had extrapulmonary disease.  Of these 31 patients, eight (19.05%) had hematogenous disseminated pulmonary TB, six (14.29%) had tuberculous meningitis, two (4.76%) had thoracic cavity TB, two (4.76%) had abdominal TB and one (2.38%) each had bone TB and nephronophthisis. The focus of infection could not be found in 10 patients. This study concluded that the incidence of TB is higher in SLE patients compared to the general population. Extrapulmonary TB and serious infection occur more likely in SLE patients. Also, those who had lupus nephritis or past history of TB infection were more susceptible to TB. 
In a retrospective analysis of the medical records of more than 3,000 SLE patients by Hou et al.,  TB was documented in 19 lupus patients with 21 episodes. Ten of the 21 episodes (47.6%) were pulmonary while the other 11 episodes (52.4%) were extrapulmonary TB. Among patients with extrapulmonary TB, five had joint and cutaneous involvement, two had miliary, two had TB of the spine, one had peritoneum involvement and one had spleen involvement. Fever and cough were found to be the most common manifestations of TB. 
Tuberculous infection thrives under conditions of immunosuppression and may either be secondary to the disease itself or to its treatment. Several studies from different countries have documented increased susceptibility for TB among SLE patients, especially where TB is endemic. ,,,,,,,,, The incidence of TB in lupus patients in these studies varied from 150 to 2,450/100,000 patient-years (among Turkish and Indian SLE patients, respectively). Not only is the incidence increased but reports from endemic countries also show the increased severity of tuberculous infection in SLE patients, with a significant percentage having extrapulmonary involvement (up to 66% in a Hong Kong cohort of 526 SLE patients).  Reactivation of TB has been reported in SLE patients treated with corticosteroids. 
Patients with SLE are also at increased risk of dissemination. This may either be due to their disease or to the use of immunosuppressive medications. In a study from Northern California, disseminated TB was found in 50% of the patients who developed TB after a diagnosis of SLE was made while none of the patients with active TB infection had disseminated disease before their diagnosis of SLE. 
In most cases, pulmonary TB presented as fever and cough, while extrapulmonary TB presented with arthralgia, skin nodules and weight loss. Sayarlioglu et al., in their analysis, compared 96 lupus patients without TB with 20 patients who developed TB and found that arthritis and renal disease were significantly higher in the latter group. Nephritis has been reported to be a risk factor for TB. ,, A recent Brazilian study reported nine cases of SLE and renal TB; all cases had nephritis.  This increased prevalence of arthritis and renal disease in SLE patients who developed TB might be a reflection of a more severe disease course, demanding a more aggressive therapeutic approach. ,,,,
TB and SLE have overlapping pulmonary and neurological manifestations; symptoms of fever, malaise and weight loss are also common in both lupus and TB.  Feng et al. found concurrent TB in 16 of 311 patients with SLE (5%) seen in Singapore between 1963 and 1979  and Tam et al. reported on TB in 11% of 526 Hong Kong SLE patients and 3.6% of 556 Turkish SLE patients. ,
The occurrence of TB also correlates with steroid dosing. Several studies have demonstrated a higher cumulative dose and/or a higher mean daily dose of prednisone in SLE patients before development of TB. ,,, One study even suggested that for each gram of prednisolone dose there was a 23% increment in the chance of developing TB.  To more convincingly prove the role of steroids in increasing the risk of TB, the effects of cortisol on the immune response to MTB antigens were studied in vitro.  It was found that cortisol, within physiological concentrations, can inhibit mycobacterial antigen-driven proliferation of cells as well as the production of intereferon-γ from healthy controls and TB patients. The mortality rate in several series varied from 0% to 31%, the highest reported for TB-affected SLE patients in Singapore. 
There is growing evidence, both clinical and experimental, which supports the pivotal role of infections in the induction and exacerbation of SLE. ,,,,,
Microbes may trigger autoimmune reactions by several ways. Firstly, microbial antigens may get associated with self-antigens to form immunogenic strains and bypass the T-cell tolerance. Secondly, certain bacterial and viral products are non-specific polyclonal B-cell mitogens and may induce the formation of autoantibodies. Thirdly, infection may induce the suppression of T-cell functions. Autoantibodies against heat shock protein (HSP) have been described in a variety of autoimmune conditions  and the available literature links mycobacterial infections with humoral autoimmunity. Stress proteins of bacteria are homologous with that of the host. This poses a constant threat for development of autoimmune disease in a genetically susceptible host due to the failure of mechanisms for self/non-self discrimination through molecular mimicry. Studies indicate the possible involvement of mycobacterial intracellular 70kDa HSP acting as a trigger for autoantibody production.  Patients with TB are found to have positivity for rheumatoid factor and antinuclear antibodies, the latter of which are characteristically seen in lupus.
Various mouse models of mycobacterial-induced arthritis have demonstrated classical manifestations of autoimmune diseases. , In a study by Ghosh et al., of 70 SLE patients, 14 patients had confirmed antecedent tuberculous infection (20%), which was 40-times higher than the prevalence of TB seen in the local population.  Thirty confirmed pulmonary TB without SLE were also studied simultaneously, which demonstrated a high incidence of autoantibodies. This suggests that prior tuberculous infection may have a role in precipitating SLE in genetically predisposed individuals.  This was supported by a Romanian study that also reported a higher incidence of antecedent TB infection (19% of 280 cases). 
With growing evidence in the form of increased prevalence of antecedent TB noted in various studies, , and the evidence from experimental autoimmune arthritis along with the demonstration of various antibodies in TB patients.  and the cross-reactivity of the monoclonal antibodies against MTB to various components of DNA, it can be postulated that MTB being a immunomodulatory agent can precipitate SLE, especially in endemic areas.
Confirmation of clinical suspicion of TB is hindered by several factors. TB presenting in a miliary pattern or with mediastinal lymphadenopathy, or as extra-pulmonary disease, poses a great diagnostic challenge because these presentations may point to a bacterial etiology or to other diseases such as lymphomas. Also, extrapulmonary TB usually presents with symptoms like persistent fever, arthralgia, arthritis, anemia and pleural and pericardial effusion, which are common in other diseases and may mimic SLE flares as well.  Because extrapulmonary involvement is more common in SLE patients, it often requires tissue and body fluid analysis for diagnosing TB and thus may take a longer period to establish definitive diagnosis. All these factors may contribute to the delay in diagnosis and institution of specific treatment. A high index of suspicion is hence required, especially in the countries with high TB prevalence to evaluate and diagnose TB in lupus patients. Studies have shown that the interval between TB onset and diagnosis may vary from 1 month to up to 1 year. ,,,,, Sputum examination for tubercular bacilli and chest radiographs are valuable tools in the diagnosis of TB, as in any other case [Table 2].
Limitations of Tuberculin Skin Testing (TST) has several limitations, some of a technical nature (low reproducibility due to the injection and reading technique), some immunological (booster effect, i.e. increase in size of the test results if repeated at short intervals), low sensitivity in patients with immunosuppressive conditions and low specificity due to cross-reactivity with NTM, among which prior vaccination with M. bovis strains, Bacillus Calmette Guerin (BCG), plays an important role. Furthermore, it needs two visits 72 h apart for placing and reading the test.
TST may be considered in areas with a high prevalence of TB for patients who may receive long-term prednisone equal to or greater than the equivalent of 15 mg/day of prednisone.  In such cases, a 5-mm tuberculin reaction represents a realistic cut-off.  For patients living in low-prevalence countries, current data do not support the routine use of TST before immunosuppressive therapy.
Relying on the result of TST may induce an overestimation of latent TB infection and the corresponding overtreatment in healthy contacts and underdiagnosis and undertreatment in immunosuppressed patients. In a TST-positive population, "classic" TST is not a suitable test to identify acute or latent TB in SLE patients and the test has to be modified as in human immunodeficiency virus (HIV) infection or malignancy.  Rivera et al. studied 112 SLE and 165 rheumatoid arthritis (RA) patients and found that TST was positive in one-third of the control patients and in less than 20% of the RA and SLE patients. 
The report presented in the American College of Rheumatology meeting of 2004  suggests that the usual screening combination of TST and chest radiograph are not adequate for detecting asymptomatic TB in a population where TB is endemic or BCG vaccination is widely used. It is preferred to test Adenosine aminase (ADA) levels in the serum/plasma or MTB polymerase chain reaction with computed tomography/magnetic resonance imaging in SLE patients.
ADA, a key enzyme in the purine salvage pathway, plays a critical role in the proliferation and maturation of lymphoid cells. This enzyme has three isoforms - ADA1 with its two forms and ADA2. Estimation of ADA levels in body fluids is a valuable tool in establishing the diagnosis of TB. ADA levels ≥42 IU/L are considered to be highly suggestive of TB, with one study reporting a sensitivity of 100%. 
Elevated ADA levels can be seen in other conditions like parainfective effusions, empyema, malignancy and autoimmune diseases like RA and SLE. Pettersson et al. compared ADA levels in 14 patients with TB, six RA and three SLE patients and found that the mean ADA levels in RA is comparable with that in TB but is significantly higher than in patients with SLE. Thus, high ADA levels do not help to differentiate between tuberculous and rheumatoid effusion but may help to differentiate them from that caused by SLE. 
Determination of isoforms of ADA activity is helpful in differentiating these conditions from the predominant ADA2 activity seen in TB and ADA1 in empyema and parainfective effusions,  but in another study it was noted that the serum ADA levels are elevated in SLE and the isoform was ADA2, similar to that seen in TB. 
Interferon-gamma Release Assays (IGRAs) are one of the recent innovations for the identification of latent tuberculous infection. These assays measure the release of interferon-γ from sensitized T lymphocytes after stimulation with antigens from MTB.
The IGRAs are performed in two ways: Quantiferon gold and T-SPOT TB. The antigens used in the commercial assays include ESAT-6 and CFP-10 supplemented by TB7.7. These antigens are not expressed in any of the attenuated M. bovis strains thus annulling the effect of previous vaccination producing false-positive results. Both tests need only one visit (for blood sampling). They are independent of the reader, are repeatable without booster effect and are not influenced by prior BCG vaccination. Numerous studies are available comparing the performance of the IGRAs with the TST. In healthy adults, the sensitivity of both tests is usually similar, whereas in immunosuppressed conditions like SLE patients, in whom the TST may be falsely negative, IGRAs seem to be superior to TST. 
In general, the diagnosis of TB in SLE patients can be established by conventional methods, including the presence of a suggestive clinical presentation, a positive TST and microbiological confirmation (by the identification of bacilli in the smear or positive cultures). In selected patients, when these measures are non-diagnostic, typical histological abnormalities such as granulomas or proper clinical response to specific therapy may help in the diagnosis. The last criteria, although not ideal, is still a useful tool in underdeveloped countries, where specific laboratory tests may not be widely available. In addition, due to the extended time required for final test results and the risk associated with this time lapse, a therapeutic trial with specific TB drugs may be entailed, especially in countries with increased TB-associated mortality. The advent of the T-cell interferon-γ release assays, a more specific test than the tuberculin skin test, may improve the detection of latent TB. 
Choice of treatment: Combination therapy
The treatment of TB in patients with SLE follows the recommendations as for other patients with this disease. After categorizing the patient, combination therapy should be started. The choice of regimen may depend on patterns of resistance, sputum culture results after the initial phase of therapy, patient characteristics and contraindications to particular medication. Pending sensitivities, this regimen usually includes combination therapy with isoniazid, rifampicin, ethambutol and pyrazinamide for 2 months followed by at least 4 months of isoniazid and rifampicin. Directly observed treatment short course(antitubercular therapy) (DOTS) is recommended for the majority of TB patients worldwide, which significantly improves the chances of cure by way of increased compliance. In most cases, there is a good response to conventional TB treatment, without many complications.
Duration of treatment
This depends on the clinical condition of the patient and response to the treatment. The decision to extend therapy from 6 to 9 months might be indicated in the presence of cavitary lung disease, extensive disease on the chest radiograph or a positive sputum culture at 2 months.  TB meningitis requires 9-12 months of therapy. TB osteomyelitis or arthritis may also require extension of therapy if there is a slow clinical response.
Renal insufficiency and end-stage renal disease
Nephritis is one of the common manifestations of SLE and patients with nephritis are more prone to developing TB. , For patients undergoing hemodialysis, administration of all drugs after dialysis is preferred to facilitate DOTS and to avoid premature removal of drugs such as pyrizinamide and cycloserine, if used. In persons with renal failure who are taking cycloserine or ethambutol, it is important to monitor serum drug concentrations to avoid toxicity. There is little information concerning the effects of peritoneal dialysis on clearance of anti-TB drugs.  The dose adjustments in patients with renal insufficiency have been explained in the American Thoracic Society guidelines for the treatment of TB, which the reader might refer to for detailed reports. 
Isoniazide, rifampicin and pyrazinamide can all cause hepatitis that may result in additional liver damage in SLE patients with pre-existing liver disease. However, because of the effectiveness of these drugs (particularly isoniazide and rifampicin), one should always try to accommodate these in the regimen even in the presence of pre-existing liver disease.  If serum aspartate aminotransferase (AST) is more than three times normal before the initiation of treatment (and the abnormalities are not thought to be caused by TB), several treatment options exist. One option is to treat with rifampicin, ethambutol and pyrazinamide for 6 months, avoiding isoniazide. The second option is to treat with isoniazide and rifampicin for 9 months, supplemented by ethambutol until isoniazide and rifampicin susceptibility is demonstrated, thereby avoiding pyrazinamide. For patients with severe liver disease, a regimen with only one hepatotoxic agent, generally rifampicin plus ethambutol, could be given for 12 months, preferably with another agent such as a fluoroquinolone, for the first 2 months; however, there are no data to support this recommendation.  In all patients with pre-existing liver disease, frequent clinical and laboratory monitoring should be performed to detect drug-induced hepatic injury.
Rifampicin induces a variety of enzymes in metabolic pathways, particularly those involving the cytochrome-P450 system. By inducing the activity of metabolic enzymes, rifampicin therapy results in a decrease in the serum concentrations of many drugs, including the various corticosteroids, sometimes to levels that are subtherapeutic. Lupus, being a disease with frequent exacerbations, whether this interaction leads to disease flare is not known, but a general statement is that the dose of steroids may have to be raised by two- to three-folds. Rifabutin, as an alternative, is supposed to have a lesser interaction and topical steroids, which may be required by lupus patients for various reasons such as certain cutaneous manifestations, usually do not have much interaction. 
Second-line drugs are used only for the treatment of TB that is resistant to first-line drugs. They are often required in the treatment of atypical mycobacterial infections as these latter infections are generally resistant to standard drugs.
Lupus-like syndromes are known to be induced by certain anti-tuberculous drugs such as isoniazide, rifampicin and para-amino salicylic acid. Approximately 20% of the patients receiving isoniazide develop antinuclear antibodies.  Fortunately, however, less than 1% develop clinical lupus necessitating drug discontinuation. Patients may present with fever, erythematous malar rash, lymphadenopathy and pleural effusions up to 1 year after initiation of therapy. Laboratory features include anemia, leucopenia, deranged LFT, positive ANA and antihistone antibodies.
As such, the development of a cutaneous rash, arthralgia and other common symptoms during treatment might be attributed either to a SLE disease flare or represent a drug side-effect. In spite of the possible major side-effects of these drugs, isoniazide and rifampicin are recommended as a first-line TB therapy.
Gartonde et al. studied the role of isoniazide prophylaxis in SLE patients receiving long-term steroid therapy and reported that isoniazide prophylaxis reduces the incidence of TB from 11% to 2%. But, there are certain limitations in the study. One is the lack of generalizability of the benefits of isoniazide prophylaxis in the areas where the prevalence of TB differs. Another concern is the occurrence of hepatotoxicity in these patients as a part of their disease process. Another study conducted in Greece, where the prevalence of TB is low, concluded that isoniazide prophylaxis may not be necessary. 
Lupus is a disease that has frequent exacerbations and infections are one of the most common reasons for flare of disease. Tuberculous infections in the setting of SLE are one of the most difficult conditions to manage as clinical features and laboratory investigations can coexist in both diseases and the presentation may be variable. The prevalence of the disease in the community is also an important deciding factor. Unlike in HIV where the immunodeficiency is quantifiable and there are appropriate guidelines available for the management of coinfections, there are no clear guidelines available for the management of TB in the setting of SLE. As the understanding of the disease becomes better in future and with the advancement of the investigations, more light may be thrown in this grey area.
[Table 1], [Table 2]