Alternative Diagnoses in Patients About to Start Empiric Tuberculosis Treatment

Background and justification: Tuberculosis remains a common disease worldwide, with an estimated 10 million new cases in 2020. It is one of the leading causes of death worldwide, causing an estimated 1,3 million deaths in 2020.1 In Sub-Saharan Africa in particular, the combined burden of tuberculosis and human immunodeficiency virus (HIV) infection leads to increased tuberculosis prevalence and mortality. Tuberculosis is a treatable disease, and most, if not all, deaths could be prevented by timely diagnosis and appropriate treatment. Unfortunately, diagnosing tuberculosis can be challenging.2 Several diagnostic strategies have been developed, but the most impactful development in recent years has been the widespread implementation of the GeneXpert MTB/RIF test. This is a cartridge-based nucleic acid amplification test, which can be used with minimal supporting laboratory structure, and has an estimated sensitivity and specificity for diagnosing pulmonary tuberculosis in adults of 85% and 98%, respectively.1,3 The GeneXpert MTB/RIF test replaced sputum smears for acid-fast bacilli as a first-line test, which had much lower sensitivity for diagnosing pulmonary tuberculosis. Historically, this lower sensitivity meant that clinicians also had to rely on clinical factors to determine whether to start tuberculosis treatment. Currently, the practice of giving empiric tuberculosis treatment to patients who had negative microbiology diagnostics is still widespread.4-7 However, in the GeneXpert MTB/RIF era, this might no longer be appropriate.8-11 For example, in Zambia, there were an estimated 40000 new tuberculosis cases in 2020 (with 92% pulmonary tuberculosis), 100% of whom were tested with GeneXpert MTB/RIF (or other rapid molecular diagnostics). However, only 51% of all patients who were treated for pulmonary tuberculosis were bacteriologically confirmed.1 This is not consistent with the supposed 85% sensitivity of the GeneXpert MTB/RIF test. This might mean that a significant proportion of patients, perhaps in the order of 30-40% of all patients, and 60-80% of all empirically treated patients, receives tuberculosis treatment inappropriately. Standard tuberculosis treatment can be burdensome, as the usual treatment duration is 6 months, and 5-10% of patients are estimated to experience serious side effects.2 In some cases, a self-limiting condition such as a viral respiratory tract infection might be the underlying diagnosis, which would not need to be followed up for six months. More importantly, empiric tuberculosis treatment can delay the diagnosis of another severe underlying condition such as malignancy, pulmonary aspergillosis, or interstitial lung disease. It has been shown that mortality is higher in patients that are empirically treated for tuberculosis, and this may be in part to missed other diagnoses.12 Over the past years, the departments of pulmonology and radiology from St. Antonius Hospital have been holding bimonthly videoconferences with the department of pulmonology at University Teaching Hospital in Lusaka, Zambia. We have mainly discussed patients with suspected interstitial lung disease. Our experience has taught us that interstitial lung diseases, which have barely been reported in Sub-Saharan Africa, can indeed be identified when you look well enough, and that these patients have oftentimes received inappropriate empiric tuberculosis treatment prior to the diagnosis of interstitial lung disease.26,27 In recent years, specialty services, such as a pulmonology clinic, as well as computed tomography (CT)-scanning and novel microbiological tests have become available in Zambia, as well as other developing countries. This, in combination with the widespread availability of the high-sensitivity GeneXpert MTB/RIF test, leads us to believe that common practice of giving empiric tuberculosis treatment after a negative initial microbiological test for tuberculosis is no longer appropriate. In this study we want to explore the feasibility of an enhanced diagnostic workup for patients about to start empiric tuberculosis treatment and the impact this has on patient diagnoses and outcomes. This workup will include discussing patients in international multidisciplinary videoconferences when needed.

Societal relevance: No specific societal relevance. Economic relevance: Might be associated with cost savings if tuberculosis treatment can be avoided, if long follow-up is avoided for self-limiting conditions and if severe conditions are diagnosed and treated in a timelier manner.

Medical relevance: Appropriate diagnosis and treatment are the cornerstone of medicine. This is in accordance with the principles of beneficence and non-maleficence. If the enhanced diagnostic workup does indeed lead to more appropriate diagnosis and treatment, this likely impacts patient quality of life and might even impact patient survival.

Other relevance: If the study findings are positive, this will justify the further building of specialty services as an addition to program-based tuberculosis services in Zambia and other developing countries.

Inclusion criteria

• adult patient
• about to start TB treatment for suspected pulmonary tuberculosis, but with sputum GeneXpert negative

Exclusion criteria

• suspected drug-resistant tuberculosis
• inpatient (because of suspected tuberculosis)
• unable to visit central study site for review or additional investigations