Prevalence, distribution and antimicrobial susceptibility pattern of bacterial isolates from a tertiary Hospital in Malawi

We conducted a retrospective audit of records for bacterial culture and susceptibility testing results between the years 2002 and 2014 from MCH microbiology laboratory department. No inclusion and exclusion criteria were applied. The MCH Laboratory participates in international microbiology quality assessment programs such as the UK External Quality Assessment program (UKNEQAS) and Zimbabwe Quality Assurance Program (ZIMQAP). Ethical approval for the study was obtained from National Health Science Research Committee (NHSRC; approval number NHSRC # 1206).

Specimens from outpatient department and clinical wards were transported to microbiology department laboratory for analysis within 2 h of collection. Conventional microbiological culture methods were employed to isolate and identify bacteria. Media was prepared in-house as per procedures stipulated in Cheesbrough [10].

About 10 mL and 3 mL of blood collected from adults and paediatric patients, respectively were inoculated into BD Bactec plus aerobic culture vials™ for incubation at 35–37 °C in BD Bactec™ 9050 Blood Culture System (BD, USA). Positive cultures were sub-cultured onto Blood Agar Plate (BAP), MacConkey Agar Plate (MAP) or Chocolate Agar Plate (CAP) if any fastidious organism was suspected. Inoculated plates were incubated at 37 °C, 5% CO₂ for up to 72 h.

Swabs collected from wounds, skin and soft tissue infection were Gram stained, cultured on either BAP/CAP or MAP, and, incubated at 37 °C,5% CO₂ for up to 72 h [10]. Swabs collected from cerebrospinal fluid (CSF) and other fluids, were Gram stained, cultured on BAP and CAP for incubation at 37 °C,5% CO₂ for up to 72 h [10]. Urine samples were inoculated on CLED agar and incubated at 37 °C for up to 48 h [10]. A specimen was considered positive for urinary tract infection (UTI) if the number of enumerated colonies were greater than ≥10⁵ CFU/mL [10, 11].

Isolates were identified and processed according to standard techniques summarised in Table 1. Antibiotic susceptibility was determined by disc diffusion technique on Mueller -Hinton agar except for Streptococcus pneumonia and other fastidious organisms, the Mueller-Hinton agar was supplemented with 5% human blood (from blood donors). Data was interpreted according to the CLSI guidelines with quality-controlled strains; Escherichia coli ATCC 25922 and Staphylococcus aureus ATCC 25923 [12]. Major bacterial isolates were evaluated for their susceptibility against the most commonly used antimicrobials during the study period. Antimicrobial screening was performed with the following antibiotics: ciprofloxacin, ceftriaxone, penicillin, tetracycline, ampicillin, chloramphenicol, amoxicillin, and gentamicin.

Data was analysed using Excel (Microsoft office, USA) and GraphPad Prism 7 software (San Diego, CA, USA). Graphs or pie chart were used to show the prevalence and distribution of the isolated bacteria against gender (male and female), age groups (Children and adults) and specimens (Pus, ear swab, V. Blood, Cerebral Spinal Fluids, High Vaginal Swabs, Urine and Other fluids). In addition, a frequency table expressed in percentage and absolute numbers was used to display the susceptibility patterns of the six commonly isolated bacteria against the six commonly used antibiotics.

To understand the demographics of the patient population, clients were characterised according to their sex and age group. Out of 4617 patients who submitted samples for microbial analysis, 2246 (49%) were male and 2371 (51%) were female (Fig. 1). We observed that 84% (n =3889) were adult patients (> 12 years old) while were 11% (n =517) were children (≤12 years old) (Fig. 1).

To analyse the dynamics of bacterial infections, total number of specimens collected and isolates recovered were determined for each hospital department. Most samples were collected in the Out Patient Department (OPD;1885, 43%; Fig. 2a). Pus from wound and ear infections were the most frequent samples and the majority of them were collected in the OPD, male surgical ward, paediatric ward and female surgical ward (Fig. 2a). Corresponding to the high number of samples collected, the OPD had the highest number of bacterial isolates (784, 44.5%; Fig. 2a). TB ward had the least number of isolates (20, 1.1%). Across the departments, S. aureus was the most isolated pathogen (783, 43.5%) followed by Klebsiella species (387, 22%) (Fig. 2b).

To evaluate the prevalence of bacteria pathogens over the study period, we determined the overall numbers of individual isolates and established the comprehensive figure of pathogens per year. Staphylococcus aureus was the most commonly isolated bacterium (783, 34.7%) (Fig. 3a). Other common isolates included Klebsiella species (393, 17.4%), Proteus species (256, 11.4%), Coagulase-Negative Staphylococcus (193, 8.6%), Escherichia coli (169, 7.5%) and Pseudomonas species (131, 5.8%; Fig. 3a). The pattern of number of patients over the years matched the pattern of positive cultures. As the number of patients increased or decreased so did the number of positive cultures (Fig. 3b). Altogether, the culture positivity rate was approximately 48.9% (2258/4611). Assuming that each patient submitted one sample, the highest culture positivity rate was recorded in 2011 (212/367; 57.7%) whereas the lowest rate was registered in 2005 (85/284; 29.9%; Fig. 3c). In 2011, 54.2% (115/212) of the bacteria isolates were gram negative with Klebsiella species being dominant (25/115; 21.7%).

To identify the most susceptible group from bacterial infections, isolates were distributed according to gender and age groups of the patients. The percentage of male and female patients whose samples were positive for bacterial culture was roughly the same i.e. 51 and 49% respectively (Fig. 4a). A significant number of bacterial isolates were isolated from patients aged more than 12 years (Fig. 4b). Largely, the distribution of isolates among male and females was fairly similar with prevalence rates of within 50% except for Pseudomonas species which were isolated more in males than females (63.1, 36.9% respectively; Fig. 4a). Most of the isolates were recovered from patients older than 12 years except S. pneumoniae which was isolated predominantly in clients younger than 12 years (88.2%; Fig. 4b).

To assess antimicrobial susceptibility, most commonly isolated bacteria were tested for their susceptibility to the most commonly used antimicrobials at the hospital during the 13-year period. In comparison to other pathogens, S. pneumoniae demonstrated the highest rate of susceptibility to most of the antibiotics except for gentamicin where 40% of the strains showed resistance (Table 2). In this study, ciprofloxacin followed by gentamicin and cotrimoxazole proved to be the most effective antimicrobials with overall pathogen susceptibility rates of 84, 75 and 70% respectively (Table 2). Nonetheless, worrying levels of antimicrobial resistance was registered in the study. On the whole, Pseudomonas species exhibited relatively the highest resistance to most antibiotics. Overall, high rates of antimicrobial resistance to erythromycin (49%), chloramphenicol (47%), and nalidixic acid (42%) were recorded (Table 1).

Some of the limitations of this study include lack of patient history on previous antimicrobial use. Consequently, this might have an influence on the observed rate of antimicrobial resistance. Secondly, due to limited availability of raw data, we were not able to capture an in-depth analysis of critical epidemiological data such as name of referral hospitals for the patients, onset of disease, occupation, and HIV status. The study was conducted at one tertiary hospital in northern Malawi as such the findings cannot be generalized to the whole country. Lastly, laboratory data on pathogen isolation, identification and susceptibility testing was generated using conventional methods, hence the results may be limited. Nevertheless, this study has provided an awareness on common microorganism isolated in our set up, their distributions and antimicrobial resistance pattern. This data could help practitioners and policy makers to make informed decisions on management of patients.