Case Presentation: Establishing the diagnosis of tuberculosis (TB) can be challenging due to its heterogeneous presentation and limited sensitivity of testing. Pulmonary TB can share overlapping radiological, clinical, and histopathological features with other lung diseases. Physicians often wait for confirmatory testing prior to starting treatment. A 36-year-old male from Haiti presented with a 4-months of fever, dry cough, night sweats, and weight loss. Chest imaging appeared classic for miliary TB with the additional finding of hilar lymphadenopathy. However, QuantiFERON testing, as well as sputum stains, cultures, and PCR were all negative. Transbronchial biopsy showed non-necrotizing granulomas with negative acid-fast staining, and next-generation sequencing (NGS) of the tissue biopsy and blood sample was also negative. Without confirmation of TB, differential diagnosis was expanded to include sarcoidosis. After starting prednisone 60mg to treat possible sarcoidosis, fevers continued, and he developed cytopenias and transaminitis. Secondary hemophagocytic lymphohistiocytosis (HLH) was considered and confirmed with a bone marrow exam and elevated serum CXCL9 and IL-2 receptor levels. Treatment with etoposide and steroids was followed by acute hypoxic respiratory failure requiring mechanical ventilation, hemodynamic instability, and multiorgan system dysfunction. With ongoing high fever and CT scan showing superimposed infiltrates, cefepime and vancomycin were commenced and NGS blood test was repeated. This was positive for Tuberculosis and confirmed on repeat BAL. Despite initiation of RIPE anti-TB regimen, patient continued to deteriorate and ultimately died.
Discussion: There are many questions related to the events of this case. Whether patient had tuberculosis from the start is the most obvious. There is a real limitation in diagnostic testing for MTB. QuantiFERON testing is adequate for screening in areas of low prevalence but in a patient from Haiti where prevalence of latent TB is as high 54.8% [1] there is real risk of false negative. U.S. physicians are trained to rely heavily on diagnostic testing, and with negative results from stains, PCR, NGS, and biopsy, a rare presentation of a rare disease was considered more likely at the time than false-negative testing in a patient with a classic history, exam, and imaging. MTB may take many forms including non-caseating granulomas and tissue biopsy has a diagnostic yield of only 50–75% [2, 3]. In medicine, certainty in diagnosis is often not necessary for initiation of treatment. With this patients’ presentation, pretest probability for TB was high. Anti-tuberculosis therapy, while associated with potential risks, might have been initiated without confirmatory testing in this case, as the benefits were deemed to outweigh the risks. The additional data of clinical response may have supplemented consideration of whether testing was accurate.
Conclusions: This case highlights the real-life challenges in clinical reasoning with the suggestion that pretest probabilities, history and examination are often sufficient to treat and should not be forgotten when testing fails to confirm the initial assessment.