Antibiotic use among hospitalized patients in northern Nigeria: a multicenter point-prevalence survey

The point-prevalence survey was conducted in three acute care hospitals across two states in Northern Nigeria, between April and May 2019 using the point-prevalence survey of healthcare-associated infections and antimicrobial use in European acute care hospitals protocol version 5.3 [13]. The survey was conducted in one teaching (tertiary) and two general (secondary) hospitals in the region. The participating hospitals had approximately 850 beds at the time of the survey and provided acute care services including internal medicine, surgery, obstetrics and gynecology, pediatrics, neonatal, psychiatry, intensive care and ear, nose and throat care. The wards in each specialty in the participating hospitals were divided into male and female wards, the wards were further subdivided into ‘A’ and ‘B’ respectively. None of the wards of the participating hospital had more than 30 patients on admission at the time of the survey.

Hospitalized patients in all the wards including pediatric emergency (where patients were often admitted for more than 24 h) and neonatal wards of the participating hospitals were considered for inclusion. However, those admitted to psychiatric, adult accident and emergency, and outpatient units were excluded. All hospitalized patients in the ward before or at 8.00 a.m. on the day of the survey were included, and those who were discharged from the wards before the time of the survey and daycare patients were excluded.

The data was collected using the European protocol for point-prevalence survey of healthcare-associated infections and antimicrobial use in European acute care hospitals version 5.3 [13]. In addition to the prevalence of antibiotic use, the current study collected data regarding the presence and types of redundant antibiotic therapy. The data was collected by a clinical pharmacist with expertise in antimicrobial stewardship and infectious diseases. The patients’ records were reviewed by the clinical pharmacist with the help of the attending physician, nurse and pharmacist.

The data collected includes: hospital type, ward specialty, patient’s demographic information and the presence of antibiotic prescription on the day or at the time of the survey. For those with at least one antibiotic prescription, the following information was collected for each antibiotic: name of drug, number of doses per day, strength of one dose, route of administration (oral, parenteral, inhalation), indication for prescription (community and hospital acquired infections, surgical prophylaxis, medical prophylaxis and unknown indication), diagnosis, documentation of reason for antibiotic prescription, date antibiotic was started, and whether antibiotic was changed during the current infection episode. For patients receiving surgical antibiotic prophylaxis, their medical records were reviewed to 24 h before the time of the survey to ascertain whether it was prescribed as a single dose, for less than 24 h or for more than 24 h. Medical records of those with potential redundant antibiotic combinations on the day of the survey were reviewed to 48 h prior to the time of the survey. All the patients in a single ward were surveyed on the same day. The data collection process lasted for 2 weeks in one of the participating hospitals. Ethics approval was obtained from the Human Research and Ethics Committee at the three hospitals with a request for waiver of patient informed consent (reference number: ABUTHZ/HREC/G23/2019, HMB/GHM/136/VOL.III/541 and MOH/ADM/744/VOL.1/716).

Redundant antibiotic therapy was defined as the prescription of two or more antibiotics with overlapping spectra of activity for 48 h or more [14, 15]. Dual anaerobic antibiotic prescriptions including oral metronidazole or vancomycin used for the treatment of Clostridium difficile infection or biliary tract infection were not classified as redundant [14]. Human immunodeficiency virus cases who received co-trimoxazole for the prevention or treatment of pneumocystis pneumonia in combination with another redundant antibiotic were excluded [15]. In addition, overlapping antibiotic spectra that involved erythromycin used in women with premature rupture of membrane was not included [15]. Those with redundant antibiotic therapy which was discontinued before the time of survey and those who received redundant antibiotic therapy for less than 48 h at the time of the survey were also excluded. Patient medical and nursing records and medication chart were reviewed by a clinical pharmacist, and potential redundant antibiotic therapy was discussed with the attending physician and/or nurse. The following antibiotics were classified as broad spectrum agents: piperacillin and beta-lactamase inhibitor, third- and fourth-generation cephalosporins, monobactams, carbapenems, fluoroquinolones, glycopeptides, polymyxins, daptomycin and oxazolidinones [6]. The World Health Organization’s Anatomical Therapeutic Chemical/Defined Daily Dose (ATC/DDD) index was used to classify the prescribed antibiotics into groups. The prevalence of antibiotic use and redundant antibiotic therapy was calculated as the percentage of patients who received one or more antibiotic and redundant antibiotic therapy on the day of the survey, respectively.

The data was analyzed using Statistical Package for the Social Sciences (SPSS) version 23. The data was de-identified before analysis. Categorical data was reported as frequency with percentage while continuous data was presented as mean (with standard deviation) or median. Bivariate and multivariate regression analyses were used to determine the factors associated with redundant antibiotic therapy. Only variables that demonstrated statistical significance (P < 0.05) in the bivariate analysis were included in the multivariate model. P value less than 0.05 was considered as statistical significance.

Of the 321 patients surveyed, 80.1% had at least one antibiotic prescription on the day of the survey. The prevalence of antibiotic use was higher in pediatric medical specialty (94.6%) followed by neonatal (92.9%) and pediatric surgical (90.0%) specialties. Obstetrics and gynecology had the lowest prevalence (72.9%) (Table 1). Most of the patients (59.1%) had two antibiotic prescriptions while 33.5% had one prescription. Overall, 449 antibiotics were prescribed on the day of the survey and approximately 56% were parenteral prescriptions. The reason for antibiotic prescription was documented in the medical note in 75.8% of the cases.

The most common indication for antibiotic prescription was community acquired infection (38.7%) followed by surgical antibiotic prophylaxis (22.5%) and hospital acquired infection (16.3%). Medical prophylaxis and unknown indication represented 14.9 and 7.6% of all the indications, respectively. Bloodstream infection (22.8), pneumonia (17.3%), skin and soft tissue infections (13.4%) and surgical site infections (11.0%) were the most common diagnoses with antibiotic prescription. Other diagnoses included gastro-intestinal infections (9.4%), intra-abdominal sepsis (7.9%), pyelonephritis (7.1%), obstetric or gynecological infection (3.9%) and central nervous system infection (3.1%). All the prescriptions for surgical antibiotic prophylaxis were prolonged beyond a day.

Of the 446 prescriptions with data regarding change of an initial antibiotic, 170 (38.1%) were changed during the infection episode. The reasons for change of an initial antibiotic include switch from parenteral to oral therapy (n = 127, 28.5%), escalation (n = 20, 4.5%), de-escalation (n = 16, 3.6%) and unknown reason (n = 7, 1.6%). The initial antibiotic was not changed during the infection episode in 61.9% of the patients.

Twenty two cases of redundant antibiotic therapy were identified in 20 patients corresponding to a point-prevalence of 6.2% in all patients (and 11.7% among patients who received two or more antibiotics). The most common types of redundancy were dual anaerobic (77.3%) and dual beta-lactam (13.6%) therapy. Other types of redundancy include dual anti-pseudomonal (4.5%) and dual anti-streptococcal (4.5%). Redundancy was higher in oral (54.5%) compared to parenteral route (36.4%). Of the 22 redundant antibiotic therapy, 40.9% were prescribed for community acquired infections while 27.3 and 22.7% were used for hospital acquired infections and surgical antibiotic prophylaxis, respectively. Overall, the combination of amoxicillin-clavulanate with nitro-imidazoles (59%) was the highest type of redundant antibiotic therapy. Multivariate regression analysis showed that the number of antibiotics prescribed (AOR: 4.724, 95% CI: 1.595–13.991, P = 0.005) and oral antibiotics (AOR: 4.075, 95% CI: 1.340–12.390, P = 0.013) were independent predictors of redundant antibiotic therapy.