Pulmonary tuberculosis presenting as acute respiratory distress syndrome

Abstract

Pulmonary tuberculosis (TB) may present in the form of parenchymal disease or extraparenchymal disease. Patients with TB as a primary cause of respiratory failure requiring mechanical ventilation have been reported to have mortality rates ranging between 47% and 80%. However, acute respiratory distress syndrome (ARDS) as a presentation of TB is rarely reported. We describe two cases of immunocompetent women presenting with ARDS. They were initially worked up for viral aetiologies in view of the ongoing COVID-19 pandemic but were later diagnosed to have microbiologically proven parenchymal pulmonary TB. One of our patients succumbed to nosocomial pneumonia, while the other was discharged to follow-up.

Background

Pulmonary tuberculosis (TB) may present in the form of parenchymal or extraparenchymal disease. Parenchymal disease includes radiographic evidence of focal or diffuse opacities, reticulonodular opacities, cavities and nodules. Extraparenchymal disease includes miliary TB, intrathoracic lymphadenopathy, pleural effusion and endobronchial disease.¹ While most patients present with prolonged symptoms (fever, cough, weight loss, loss of appetite and so on) on an outpatient basis, a significant proportion may require hospitalisation. Patients with TB as a primary cause of respiratory failure requiring mechanical ventilation have been reported to have mortality rates ranging between 47% and 80%.^2–4 However, acute respiratory distress syndrome (ARDS) as a presentation of TB, as seen in our cases, is rarely reported.

We describe two cases of immunocompetent women presenting with ARDS. They were initially worked up for viral aetiologies in view of the ongoing COVID-19 pandemic but were later diagnosed to have microbiologically proven parenchymal pulmonary TB. One of our patients succumbed to nosocomial pneumonia, while the other was discharged to follow-up. We believe that it is important to consider TB as a differential for severe acute respiratory illness in high burden countries to avoid delays in treatment initiation.

Case presentation

Case 1

A 31-year-old woman presented to the hospital in March 2020 with an 8-day history of fever, 4-day history of dry cough and 2-day history of shortness of breath. History was significant for an operated traumatic cervical spine fracture 3 years prior. Her postoperative period was complicated by an abscess, which was determined to be of tubercular origin for which she received adequate antitubercular therapy, and was deemed cured. At presentation, she was tachycardic (heart rate=116 bpm), blood pressure of 118/72 mm Hg, respiratory rate of 36 breaths per minute and room air oxygen saturation of 76%. Chest auscultation revealed bilateral basal crepitations with bronchial breath sounds in the left lower lung fields. Her chest radiogram showed diffuse bilateral alveolar opacities with left lower lobe consolidation. She was managed in the intensive care unit after intubation for type I respiratory failure. Her blood gas showed a partial pressure of oxygen (PaO₂)/fractional inspired oxygen (FiO₂) ratio of 110 (FiO₂ 70%) at a positive end-expiratory pressure of 10 cmH₂O. She was mechanically ventilated as per the ARDSnet protocol with a provisional diagnosis of viral pneumonia with ARDS.

Case 2

A 22-year-old woman with no prior comorbidities presented to the hospital in April 2020 with a 10-day history of fever and a 2-day history of shortness of breath. At presentation, she had a heart rate of 124 bpm, blood pressure of 132/76 mm Hg, respiratory rate of 46 breaths per minute and room air saturation of 62%. Her chest radiogram showed diffuse bilateral alveolar opacities suggestive of ARDS (figure 1). Her blood gas was significant for type I respiratory failure with a PaO₂/FiO₂ ratio of 90. She was mechanically ventilated and managed in the intensive care unit.

Figure 1.

Chest radiograph of case 2 at admission. Diffuse bilateral alveolar opacities suggestive of acute respiratory distress syndrome.

Investigations

Case 1

Baseline haemogram and chemistry is depicted in table 1. Mini-BAL PCR for COVID-19 (repeated twice), H1N1, RSV and bacterial cultures were negative. In view of her significant history, her mini-BAL was tested for Mycobacterium tuberculosis by cartridge-based nucleic acid amplification test (CBNAAT), which was positive (sensitive to rifampicin) though smear for acid-fast bacilli (AFB) was negative. Further processing yielded M. tuberculosis on culture after 2 weeks. She was tested negative for HIV. A contrast-enhanced CT of the chest (figure 2) was reported to have consolidation in bilateral lung fields predominantly in the dependent location, cavities within the consolidation in the apical segments of the right upper lobe, centrilobular nodules in bilateral lungs, mediastinal lymphadenopathy and bilateral pleural effusion; the final impression was ARDS with infective aetiology, likely tubercular.

Table 1.

Haemogram and basic metabolic panel of the described cases

Laboratory parameter	Case 1	Case 2
Haemogram
Haemoglobin (g/dL)	9.8	9.3
Haematocrit (%)	34.7	30.1
Total leucocyte count (cells/mm3)	8310	15 870
Absolute neutrophil count (cells/mm3)	7795	13 886
Absolute lymphocyte count (cells/mm3)	300	1220
Platelet count (cells/mm3)	353 000	277 000
Basic metabolic panel
Total bilirubin (mg/dL)	0.1	2.0
Conjugated bilirubin (mg/dL)	–	1.9
Aspartate transaminase (IU/L) (ULN=40)	35	116
Alanine transaminase (IU/L) (ULN=45)	07	63
Alkaline phosphatase (IU/L) (ULN=240)	285	2127
Serum creatinine (mg/dL)	0.4	0.5
Serum urea (mg/dL)	49	35
Albumin (g/dL)	2.7	2.5
Serum sodium (mEq/L)	131	130
Serum potassium (mEq/L)	5.1	4.5
Serum ferritin (ng/mL)	164.4	204.9
C reactive protein (mg/dL)	9.8	20.2

ULN, upper limit of normal.

Figure 2.

Contrast-enhanced CT of case 1. (A) Lung window showing consolidation of bilateral lower lung fields depicted by the black asterisks. (B) Lung window showing cavitation within the consolidation in the right upper lobe depicted by black arrows. (C) Mediastinal window showing mediastinal lymphadenopathy depicted by white arrows. (D) Mediastinal window showing bilateral pleural effusion depicted by white arrows.

Case 2

Initial haemogram and chemistry are depicted in table 1. Her mini-BAL PCR was negative for COVID-19 (repeated twice), H1N1, RSV and bacterial cultures. CBNAAT for M. tuberculosis was positive in her mini-BAL (sensitive to rifampicin), though smear for AFB was negative. The patient was tested negative for HIV. Inflammatory markers, though non-specific, were also sent for both patients in view of the on-going COVID-19 pandemic.

Differential diagnosis

Both cases presented to us with a severe acute respiratory illness and ARDS. In view of their age and lack of comorbidities, our initial diagnosis consisted of viral pneumonia with ARDS (with a high suspicion of influenza). Both presented to us in March and April 2020, respectively. This was in the midst of the COVID-19 pandemic fitting into the clinical picture as well. Hence, COVID-19 was also on our differential list.

Case 1

Her history of tubercular collection, normal initial total leucocyte count at presentation with lymphopaenia alerted us to a possibility of pulmonary TB. This was however lower down in our differentials due to the acute presentation with ARDS, which is quite rare for pulmonary TB.

Case 2

Her initial elevated total leucocyte count with neutrophilia hinted at the possibility of bacterial pneumonia superimposed on a viral pneumonia with ARDS. She was covered with appropriate empirical antibiotics. The initial hepatic dysfunction was also attributed to sepsis. However, with repeated negative cultures, it was decided to look for other causes of the acute presentation. Thus, a decision to test for TB was taken. A lack of prior history of TB, TB contact, her acute history and neutrophilia led to the delay in diagnosis and initiation of appropriate treatment.

Treatment

Case 1

She was started on weight based antitubercular therapy (isoniazid, rifampicin, pyrazinamide and ethambutol).

Case 2

She was started on weight-based antitubercular therapy (isoniazid, rifampicin, pyrazinamide and ethambutol).

Outcome and follow-up

Case 1

Her ventilatory parameters improved; she was extubated successfully and discharged after a further uneventful hospital course. In view of the national lockdown, she was followed up telephonically after 1 month and was ascertained to be improving. She claimed to have subjectively gained weight with a significant reduction in cough.

Case 2

After a period of initial improvement, the patient developed increasing secretions with new onset fever and shock. Endotracheal tube aspirate sample was sent for culture which grew Acinetobacter baumanii sensitive only to colistin. Despite culture guided antibiotics, the patient succumbed to the ventilator-associated pneumonia.

Discussion

Pulmonary TB is characterised by its subacute to chronic symptoms but has been increasingly implicated with acute presentations including ARDS. Existing literature only describes ARDS as a presentation of TB in case reports and case series. The largest report is a series of 29 cases. This study analysed predictors of presentation that were likely to present/progress to ARDS. They showed a higher likelihood of development of ARDS in miliary TB compared with parenchymal TB. Patients with ARDS were more likely to be febrile, and dyspnoeic, but were less likely to have haemoptysis. They concluded that prolonged illness (duration of more than 30 days), absolute lymphocytopaenia (absolute lymphocyte count <1625 cells/mm³) and elevated alanine transaminase (>100 IU/L) were independently associated with ARDS development. The second series reported nine cases with eight presenting with miliary TB. They express a median duration of 45 days of symptoms (fever, cough and anorexia) at presentation. None of their cases complained of haemoptysis. The transition to ARDS was marked by an acute onset of shortness of breath of less than 5 days’ duration. All the patients reported in the above studies were negative for HIV.^{3 5}

Our cases reflect some of the reported results of the above series. Both cases had a relatively short history of about 10 days, but their deterioration to ARDS was marked by dyspnoea that was of shorter duration of about 2 days. The presentation was not marked with the typical features of weight loss, anorexia, night sweats and evening rise of temperature, which led us to initially consider alternate aetiologies. Both patients had lymphocytopaenia (absolute lymphocyte count=300 cells/mm³ and 1220 cells/mm³) on investigation and were negative for HIV. It is interesting to note that our patients were additionally anaemic, hypoalbuminaemic and hyponatraemic (table 1). In a study published in 2017, TB presenting as severe respiratory illness with symptoms less than 14 days was found in 6.3% of the 2097 individuals who were tested for TB. They also highlight that these individuals may not present with cough and night sweats.⁶ Due to the acute presentation, in the midst of a pandemic, we initially considered viral aetiologies. We urge those living in high burden nations like ours to consider TB in cases of severe acute respiratory illness as they are curable if treatment is initiated timely. A prior history of TB and a history of contact with a case of TB in our experience are strong factors that may influence the list of differentials.

Tubercular pneumonia requiring mechanical ventilation is reported to have a higher mortality than pneumonia of non-tubercular origin. Two case series from Canada and Taiwan report an in-hospital mortality rate of 69% and 65.9%, respectively.^{2 4} The two Indian case series report lower mortality rates of 41% and 22%.^{3 5} Delay in diagnosis and treatment has been associated with a worse prognosis and contributes to higher mortality in TB patients with ARDS. This may be due to lack of clinical suspicion, difficulty in obtaining lower respiratory tract samples due to non-permissive ventilator parameters or misinterpretation of radiographic findings.² It is interesting to note that the mini-BAL smear for AFB was negative in both our cases. CBNAAT, being highly specific in smear-negative samples, helped confirm the diagnosis.⁷ Higher suspicion of tubercular aetiology due to its high prevalence might reflect the lower mortality rates in the Indian series. High clinical suspicion of TB is thus advised in the setting of ARDS, in patients with prolonged symptoms or suggestive radiological findings. Another potential explanation for high mortality in tubercular ARDS may be due to poor enteral absorption of oral antitubercular therapy in patients with multiorgan dysfunction syndrome and subsequent gut intolerance, making conversion to intravenous formulations an important consideration.⁴

The pathogenesis of ARDS in TB is not clear. It has been postulated that miliary TB, being a diffuse disease, is more likely to cause diffuse lung injury, compared with the local injury of tubercular pneumonia.⁴ A potential mechanism may be the release of a large burden of bacilli into the pulmonary circulation, causing obliterative endarteritis and damage to the alveolo-capillary membranes.⁸ Lipoarabinomannan in the mycobacterial cell wall activates macrophages to release tumor necrosis factor- α (TNF-α) and interleukin-1β (IL-1β).⁹ Intravenous infusions of TNF-α, which can cause dose-dependent endothelial injury, have been demonstrated to reproduce acute lung injury seen in sepsis.¹⁰ M. tuberculosis has been shown to render endothelial cells more susceptible to the toxic effects of TNF-α. M. tuberculosis also increases the expression of intercellular adhesion molecule-1 (ICAM-1) on endothelial cells, which promotes binding of activated neutrophils.⁴ It is not yet known whether the development of ARDS is related to variations in host immune responses or different virulence of strains of the bacillus.

An interesting aspect of this presentation of TB is the utility of steroids in treatment. Corticosteroids inhibit the release of lymphokines and cytokines that are responsible for tissue damage and allow antitubercular drugs to penetrate granulomas, disrupting them. Steroids are established in the treatment of severe extrapulmonary TB, such as meningeal or pericardial disease. A South Korean study retrospectively analysed the effects of steroids in TB patients with acute respiratory failure. There was no uniformity in the indications for steroid use or the protocol for administration. The unadjusted mortality rate was not different in those receiving steroids. However, a post hoc analysis demonstrated a significant reduction in 30-day and 90-day mortality.¹¹ We used methylprednisolone in the second case, with little success. It is possible that earlier initiation of immunosuppression may have helped. There is a paucity of randomised data in this area and may be a potential area of interest for research.

Learning points.

• Pulmonary tuberculosis may have an acute presentation with a clinical picture similar to bacterial pneumonia or acute respiratory distress syndrome.

• Such a drastic presentation of tuberculosis carries high morbidity and mortality. High index of suspicion must be kept in countries with a high burden of tuberculosis.

• Previous history of tuberculosis, history of contact with a patient of active tuberculosis and normal total leucocyte count at presentation (especially with lymphopaenia) should alert the physician to a possibility of tuberculosis.

• Microbiological evidence (culture or nucleic acid tests) is the gold standard for confirming the diagnosis of pulmonary tuberculosis. Radiological investigations may play an ancillary role when the clinical history is not typical of tuberculosis.