Profile of bacterial infectious disease and antimicrobial choices in an urban hospital at Maputo, Mozambique: a prospective observational study

Abstract

Background: Few antimicrobial resistance profile data exist for Mozambique. Limited antimicrobial stewardship and a lack of guidelines favour near anarchic antibiotic use. We aimed to study antibiotic selection rationale, empirical therapy, and use of cultures in Maputo Central Hospital, Maputo, Mozambique. Methods: We recruited adults admitted to Maputo Central Hospital's medical wards between February and April, 2018, and who were prescribed antibiotics. We collected data from patient charts and prescription sheets and studied the profile of acute bacterial infections, antibiotic treatment choices, and possible treatment errors. Findings: We included 232 patients (12·5% of admissions). Of these, 143 (61·6%) were HIV infected, more than half (64%) with a CD4 count of less than 200 cells/μL. Diagnosed infections included: community-acquired pneumonia (106 patients [44·1%]), bacterial meningitis (50 [15·1%]), aspiration pneumonia (42 [12·7%]), acute enteritis (38 [11·5%]), chronic enteritis (7 [2·1%]), urinary tract infection (19 [5·7%]), soft tissue infections (10 [3%]), sepsis (11 [3·3%]), brain abscess (3 [0·9%]), neutropenic fever (3 [0·9%]), and otitis with effusion (1 [0·9%]). Antibiotics were prescribed for 52 patients (22·4%) who did not have a recorded infection on the medical chart. Eight prescriptions were meant for infection prophylaxis in patients with hepatic encephalopathy, stroke, diabetic ketoacidosis, or haematological malignancies. Of the included inpatients, 12·1% (28) died, 77·2% (179) were discharged home improved, 10·8% (25) transferred to lower acuity hospitals, and 4·3% (10) transferred to a tuberculosis hospital. We identified 347 prescriptions for the 232 patients. One antibiotic was prescribed in 56·9% (132) of cases and two or more were prescribed to 37·1% (86) and 6% (14), respectively. Ceftriaxone (176 prescriptions [50·7%]) and metronidazole (57 [16·4%]) were most frequently prescribed. We identified 46 prescriptions (12·7%) for patients with no indication recorded on their chart. In 42 cases (12%), the antibiotic choice was not consistent with international guidelines. Incorrect duration of therapy was specified in 8 cases (2%). Glomerular filtration rate estimations were seldomly used to adjust dosage. Blood cultures were rarely ordered (2 [0·86%]). CSF was cultured in 38 patients, with no positive findings. Interpretation: A scarcity of antimicrobial drugs, shortages of qualified staff, a lack of local stewardship of antimicrobial programmes, and poor laboratory capacity encourage the overuse of empirical treatment rather than treatment decisions based on evidence and laboratory findings. Findings suggest the need for continued education, and accurate and updated hospital antimicrobial resistance algorithms for best clinical practices. Funding: Fogarty International Center from the NIH D43TW009343 and the UCGHI.