Formulas based on red blood cell indices have been used to differentiate between iron deficiency anemia (IDA) and thalassemia (Thal). However, they exhibit varying efficiencies. In this study, we aimed to develop a tool for discriminating between IDA and Thal by using the random forest (RF) and gradient boosting (GB) algorithms. Complete blood count data from 1143 patients with anemia and low mean corpuscular volume were collected (382 patients with IDA, 635 with Thal, and 126 with IDA and Thal). The data were randomly divided into the training and testing datasets in a ratio of 80:20. The RF and GB models had good diagnostic performances for predicting IDA and Thal in the training and testing datasets. In the testing dataset for predicting binary outcomes, GB and RF both had an accuracy of 90.7%, and an area under the receiver operating characteristic curve (AUC-ROC) of 0.953. A lower diagnostic performance was observed when patients with IDA and Thal were included. GB and RF showed accuracies of 80.4% and 82.2%, respectively, and AUC-ROC values of 0.910 and 0.899, respectively. In conclusion, we developed a machine learning approach using GB algorithm. This tool is potentially useful in Thal- and IDA-endemic regions.

Saelue P, et al. Infect Chemother. 2024 Aug;56:e39. https://doi.org/10.3947/ic.2024.0045

Bleeding from the affected organs is a common manifestation of amyloidosis. The risk for perioperative bleeding in patients with amyloidosis remains controversial. In this study, we aimed to compare the bleeding risk of tissue biopsies for a definitive diagnosis between patients with and without amyloidosis, identify risk factors, and generate a prediction model for bleeding risk in these patients. We enrolled patients aged >15 years who had amyloidosis as part of their differential diagnosis before tissue biopsy. After obtaining histopathological reports, we randomly selected patients with and without amyloidosis at a ratio of 4:1. A total of 360 patients were enrolled before tissue biopsy. Bleeding complications were observed in 5.6% and 4.2% of patients with and without amyloidosis, respectively. Amyloidosis was not associated with an increased perioperative bleeding risk (adjusted odds ratio 1.19; 95% confidence interval 0.17-8.41, p = 0.859). Kidney biopsy was a significant risk factor for perioperative bleeding in tissue biopsies. A KiHPL model was generated to predict the bleeding risk. The area under the curve was 0.87, with a good calibration plot for this model. Amyloidosis is not associated with an increased risk of bleeding in tissue biopsies. The KiHPL model can predict the bleeding risk of tissue biopsies for a definitive diagnosis in patients with suspected amyloidosis.

PURPOSE: To evaluate anticoagulant trends and clinical outcomes in the management of cancer-associated thrombosis (CAT) within Thailand, an upper-middle–income country (UMIC). METHODS: This multicenter retrospective cohort study included adult patients with cancer diagnosed with venous thromboembolism (VTE) hospitalized in Thailand from 2017 to 2021. Anticoagulants were classified as low-molecular-weight heparin (LMWH), direct oral anticoagulants (DOACs), and warfarin. Prescription trends were assessed, and patients were followed for 1 year, or until 2022 to evaluate outcomes. The primary effectiveness outcome was recurrent VTE, whereas the primary safety outcome was major bleeding. Secondary outcomes included net clinical benefit and all-cause mortality. Treatment effects were examined using inverse probability of treatment weighting (IPTW) Cox proportional hazards models. RESULTS: Among 1,611 patients (61.3% female; mean age, 58.8 years; standard deviation, 13.1 years), 86% received LMWH, 10% warfarin, and 4% DOACs. In the study cohort, LMWH prescriptions remained consistent, warfarin use declined, and DOAC prescriptions notably increased. In IPTW analysis, DOACs showed comparable rates of VTE recurrence (weighted hazard ratio [HR], 0.77 [95% CI, 0.22 to 2.70]; P = .679) and major bleeding (weighted HR, 0.62 [95% CI, 0.15 to 2.55]; P = .506) with LMWH. Warfarin had a higher risk of major bleeding (weighted HR, 2.74 [95% CI, 1.12 to 6.72]; P = .028) but a similar rate of VTE recurrence (weighted HR, 1.46 [95% CI, 0.75 to 2.84]; P = .271) compared with LMWH. Secondary outcomes were consistent across groups. CONCLUSION: LMWH remains the primary treatment for CAT, in line with current guidelines. The study highlights the challenges faced in these settings with the continuous use of warfarin. The comparable efficacy and safety of DOACs with LMWH suggest a potential shift in CAT management within UMICs.

Objective: Bone marrow examination is valuable for identifying the cause of fever of unknown origin (FUO) in HIV-infected patients. Based on the outcomes of bone marrow examination of patients with FUO, we aimed to develop a predictive model for identifying the factors that can increase the diagnostic yield of bone marrow examination. Design: For this retrospective cohort study, we enrolled HIV-infected patients, aged more than 15 years and diagnosed with FUO, at Songklanakarind Hospital in Southern Thailand, between January 2009 and December 2019. Methods: Evaluations were based on bone marrow aspiration, biopsy, and culture; any missing data were imputed with regression imputation. Results: Among the final 108 included patients, 44 (40.74%) showed positive bone marrow results. The diagnoses mainly comprised histoplasmosis, penicilliosis, and tuberculosis. Bone marrow examination led to treatment modifications in approximately 33% patients. Platelet count less than 150 000 cells/μl, alkaline phosphatase (ALP) level at least 200 U/l, and no previous antibiotic treatment were significantly associated with higher diagnostic yields. The HIV bone marrow (HIVBM) model, comprising of spleen size, hematocrit (Hct), platelet count before bone marrow examination, ALP level at admission, and previous antibiotic treatment, was generated as a nomogram to predict the diagnostic yield of bone marrow examination in HIV-infected patients with FUO. Conclusion: The results of this study indicate that the HIVBM model can be used to predict the diagnostic yield of bone marrow examination, and therefore assist in clinical decision-making regarding bone marrow procedures, to be performed for identifying the origin of fever in HIV-infected patients.