Retrospective cohort evaluation on risk of pneumonia in patients with pulmonary tuberculosis

Early identification of suspected serious infection among patients afebrile at initial presentation using neural network models and natural language processing: A development and external validation study in the emergency department

OBJECTIVE

To develop and externally validate models based on neural networks and natural language processing (NLP) to identify suspected serious infections in emergency department (ED) patients afebrile at initial presentation.

METHODS

This retrospective study included adults who visited the ED afebrile at initial presentation. We developed four models based on artificial neural networks to identify suspected serious infection. Patient demographics, vital signs, laboratory test results and information extracted from initial ED physician notes using term frequency-inverse document frequency were used as model variables. Models were trained and internally validated with data from one hospital and externally validated using data from a different hospital. Model discrimination was evaluated using area under the receiver operating characteristic curve (AUC) and 95% confidence intervals (CIs).

RESULTS

The training, internal validation, and external validation datasets comprised 150,699, 37,675, and 85,098 patients, respectively. The AUCs (95% CIs) for Models 1 (demographics + vital signs), 2 (demographics + vital signs + initial ED physician note), 3 (demographics + vital signs + laboratory tests), and 4 (demographics + vital signs + laboratory tests + initial ED physician note) in the internal validation dataset were 0.789 (0.782-0.796), 0.867 (0.862-0.872), 0.881 (0.876-0.887), and 0.911 (0.906-0.915), respectively. In the external validation dataset, the AUCs (95% CIs) of Models 1, 2, 3, and 4 were 0.824 (0.817-0.830), 0.895 (0.890-0.899), 0.879 (0.873-0.884), and 0.913 (0.909-0.917), respectively. Model 1 can be utilized immediately after ED triage, Model 2 can be utilized after the initial physician notes are recorded (median time from ED triage: 28 min), and Models 3 and 4 can be utilized after the initial laboratory tests are reported (median time from ED triage: 68 min).

CONCLUSIONS

We developed and validated models to identify suspected serious infection in the ED. Extracted information from initial ED physician notes using NLP contributed to increased model performance, permitting identification of suspected serious infection at early stages of ED visits.

Asthma Therapies on Pulmonary Tuberculosis Pneumonia in Predominant Bronchiectasis-Asthma Combination

Background: It is sometimes difficult to distinguish between asthma and bronchiectasis as their symptoms overlap, and these two diseases are associated with pulmonary tuberculosis (PTB) or pneumonia.

Objective: The purpose of this study is to determine the effects of bronchodilator drugs, steroids, antidepressants drugs, and antianxiety drugs on the risks of PTB or pneumonia in patients with bronchiectasis–asthma combination or bronchiectasis–asthma–chronic obstructive pulmonary disease combination—BCAS cohort.

Methods: After propensity score matching, we retrospectively studied patients with BCAS (N = 620) and without BCAS (N = 2,314) through an analysis. The cumulative incidence of PTB or pneumonia was analyzed through Cox proportional regression. After adjustment for sex, age, comorbidities, and medications [including long-acting beta2 agonist/muscarinic antagonists (LABAs/LAMAs), short-acting beta2 agonist/muscarinic antagonists (SABAs/SAMAs), leukotriene receptor antagonist, montelukast, steroids (inhaled corticosteroids, ICSs; oral steroids, OSs), anti-depressants (fluoxetine), and anti-anxiety drugs (benzodiazepines, BZDs)], we calculated the adjusted hazard ratios (aHR) and their 95% confidence intervals (95% CI) for these risks. Similar to OSs, ICSs are associated with an increased risk of PTB or pneumonia, lumping these two as steroids (ICSs/OSs).

Results: For the aHR (95% CI), with non-LABAs/non-OSs as the reference 1, the use of LABAs [0.70 (0.52–0.94)]/OSs [0.35 (0.29–0.44)] was associated with a lower risk of PTB or pneumonia. However, the current use of LABAs [2.39 (1.31–4.34)]/SABAs [1.61 (1.31–1.96)], steroids [ICSs 3.23 (1.96–5.29)]/OSs 1.76 (1.45–2.14)], and BZDs [alprazolam 1.73 (1.08–2.75)/fludiazepam 7.48 (1.93–28.9)] was associated with these risks. The current use of LAMAs [0.52 (0.14–1.84)]/SAMAs [1.45 (0.99–2.11)] was not associated with these risks.

Conclusion: The current use of LAMAs/SAMAs is relatively safe with respect to PTB or pneumonia risks, but LABAs/SABAs, steroids, and BZDs could be used after evaluation of the benefit for the BCAS cohort. However, we must take the possible protopathic bias into account.

Predominance of Other Pathogenic Bacteria among Presumptive Tuberculosis Cases Attending Tuberculosis Clinics in Mwanza, Tanzania: A Cross-Sectional Laboratory-Based Study

This study was designed to determine the prevalence and co-infection of Mycobacterium tuberculosis and other pathogenic bacteria among presumptive cases of tuberculosis (TB) at selected hospitals in Mwanza, Tanzania. GeneXpert and conventional bacteriological culture and sensitivity were used for the detection of TB and other pathogenic bacteria, respectively. STATA version 13.0 was used for data analysis. The median (IQR) age of participants was 33 (19−51) years with males forming more than half (i.e., 59% (158/264)) of the participants. Microscopically, 29.5% (78/264) of the patients had polymorphonuclear leucocytes in the sputum samples. Approximately 7.2% (19/264), 16.3% (43/264), and 1.1% (3/264) of participants had TB, other pathogenic bacteria, and co-infections, respectively. One sample had growth of two other bacteria, resulting in a total of 44 isolated bacteria with the predominance of Gram-negative bacteria at 75.0% (33/44). The predominant species isolated was the Klebsiella pneumoniae complex at 52.3% (23/44). Overall, 27.3% (9/33) of GNB were resistant to third-generation cephalosporins, while Gram-positive bacteria were more resistant to erythromycin at 63.6% (7/11). Good quality sputa had a significantly higher yield of pathogenic bacteria than poor quality sputa (37.2% vs. 7.5%, p < 0.001). Presumptive TB cases were predominantly infected with other pathogenic bacteria than M. tuberculosis. Therefore, other pathogenic bacteria should be considered when attending presumptive TB cases to ensure favorable treatment outcomes.

The prevalence and risks of major comorbidities among inpatients with pulmonary tuberculosis in China from a gender and age perspective: a large-scale multicenter observational study

In clinical practice, PTB patients have concurrent many types of comorbidities such as pneumonia, liver disorder, diabetes mellitus, hematological disorder, and malnutrition. Detecting and treating specific comorbidities and preventing their development are important for PTB patients. However, the prevalence of most comorbid conditions in patients with PTB is not well described. We conducted a large-scale, multicenter, observational study to elucidate and illustrate the prevalence rates of major comorbidities in inpatients at 21 hospitals in China. The 19 specific comorbidities were selected for analysis in this patient cohort, and stratified the inpatient cohort according to age and gender. A total of 355,929 PTB inpatients were included, with a male:female ratio of 1.98 and the proportion of ≥ 65 years PTB inpatients was the most. Approximately 70% of PTB inpatients had at least one defined type of comorbidity. The prevalence of 19 specific comorbidities in inpatients with PTB was analyzed, with pneumonia being the most common comorbidity. The prevalence of most comorbidities was higher in males with PTB except thyroid disorders, mental health disorders, etc. The prevalence of defined most comorbidities in patients with PTB tended to increase with increasing age, although some specific comorbidities tended to increase initially then decrease with increasing age. Our study describes multiple clinically important comorbidities among PTB inpatients, and their prevalence between different gender and age groups. The results will enhance the clinical aptitude of physicians who treat patients with PTB to recognize, diagnose, and treat PTB comorbidities early.

Pneumonia caused by Mycobacterium tuberculosis

Tuberculosis (TB) is one of the top ten leading causes of death worldwide. The causative agent of TB is Mycobacterium tuberculosis. Acute tuberculous pneumonia (TP) is an acute form of pulmonary TB. However, acute TP and non-tuberculous community-acquired pneumonia can be easily confused, resulting in deterioration of TP due to delayed treatment. Therefore, rapid and accurate diagnosis of acute TP is crucial in order to stop the transmission of TB. Moreover, development of new diagnostic tools (technologies and approaches), and flexible application of different therapy schemes will help to reduce the incidence of TP and promote the goal of ending the TB epidemic.

Risk of contracting pneumonia among patients with predialysis chronic kidney disease: a population-based cohort study in Taiwan

OBJECTIVES

The objective of the study was to investigate the association between predialysis chronic kidney disease and contracting pneumonia in Taiwan.

METHODS

We employed a population-based, retrospective cohort design using the database of the Taiwan National Health Insurance (NHI) Program. There were 18807 subjects aged 20-84 years who were newly diagnosed with predialysis chronic kidney disease between 2000 to 2012 as the predialysis chronic kidney disease group and 18807 randomly selected subjects without chronic kidney disease as the non-chronic kidney disease group. The predialysis chronic kidney disease and non-chronic kidney disease groups were matched according to sex, age, comorbidities, and the year of index date. The incidence of contracting pneumonia among both groups at the end of 2013 was calculated. The multivariable Cox proportional hazards regression model was used to calculate the hazard ratio (HR) and 95% confidence interval (CI) for contracting pneumonia being associated with predialysis chronic kidney disease.

RESULTS

The overall incidence of contracting pneumonia was 1.47-fold higher in the predialysis chronic kidney disease group than that in the non-chronic kidney disease group (24.6 vs. 16.7 per 1, 000 person-years, 95% CI 1.40, 1.55). After adjusting for co-variables, the HR of contracting pneumonia became 1.52 for subjects with predialysis chronic kidney disease (95% CI 1.43, 1.60) compared to subjects without chronic kidney disease. With even further analysis, in the absence of any comorbidity, the adjusted HR of contracting pneumonia was 1.53 for subjects with predialysis chronic kidney disease alone (95% CI 1.32, 1.76).

CONCLUSIONS

Patients with predialysis chronic kidney disease have a 1.52-fold increased risk of contracting pneumonia as compared to those with non-chronic kidney disease. Even in the absence of any comorbidity, a greater than average risk of contracting pneumonia remains present.