Prevalence of Extended-Spectrum Beta-Lactamase–Producing Enterobacteriaceae and Associated Clinical Implications at the University Teaching Hospital of Kigali in Rwanda

ABSTRACT.

Extended spectrum beta-lactamase (ESBL)-producing Enterobacteriaceae poses a global problem and complicates therapeutic choices. The paucity of data in resource-poor countries undermines the understanding of the problem’s extent, and cases of antimicrobial treatment failure continue to accumulate. This study aimed to determine the prevalence and clinical implications of ESBL-producers at the University Teaching Hospital of Kigali in Rwanda. A 1-year cross-sectional retrospective study was conducted on Escherichia coli and Klebsiella pneumoniae isolated in blood and urine from January 1 to December 31, 2022. In total, 1,283 isolates were recorded. The results showed an overall prevalence of ESBL phenotypes at 300/1,283 (23.4%). Extended spectrum beta-lactamase–positive E. coli was more frequently detected than K. pneumoniae in both urine (20.6% versus 10.1%) and blood (8.8% versus 6.2%). These isolates were 100% resistant to amoxicillin-clavulanic acid, third-generation cephalosporins, piperacillin, sulbactam ampicillin, ampicillin, cefuroxime, and cefoxitin. The least resistance was observed to amikacin (18%), meropenem (10%), and polymyxin B (3%). Hospital stays ranging from 8 to 21 days were the most frequent, and the mortality rate was 10.3% in patients with ESBL cases, which was more than double the general hospital mortality rate in the same period. In conclusion, our findings indicate a high prevalence of ESBL phenotypes, high antibiotic resistance rates, prolonged hospital stays, and an increased mortality rate. These findings suggest the need for continued surveillance, planning appropriate interventions, and caution during empirical therapy.

INTRODUCTION

Extended-spectrum beta-lactamase (ESBL)-producing Enterobacteriaceae remain a global health challenge. Infections due to these bacteria are difficult to treat, extend hospital stays,¹ and are associated with increased medical care costs,^2–4 as well as poor treatment outcomes.⁵ Extended-spectrum beta-lactamase production is more prevalent in Escherichia coli and Klebsiella pneumoniae,^6,7 and its incidence continues to rise in both healthcare and community settings.^8,9 Recent evidence indicates a more than 8-fold increase in their intestinal carriage,¹⁰ complicating public health prevention strategies.

Producers of EBSL display high resistance to commonly used beta-lactam antibiotics,^11,12 constituting approximately 65% of all antibiotics used globally for bacterial infection treatment.¹³ Carbapenems, considered an alternative treatment of ESBL producers,¹⁴ are becoming less effective due to their excessive use, leading to emergence of carbapenem-resistant Enterobacteriaceae, particularly in sub-Sahara African countries.^15,16

The emerging antibiotic resistance in sub-Saharan Africa is a looming crisis, lacking appropriate intervention and prevention measure, despite high resistance rate, particularly in healthcare settings.¹⁷ Estimating the extent of the problem is challenging due to inadequate capacity for detection and surveillance. Poor hygiene, ineffective infection prevention and control measures, and an increase in immunocompromised individuals further exacerbate the issue.⁵

In East Africa, Gram-negative bacteria show relatively high resistance (50–100%) to routinely prescribed antibiotics such as ampicillin, gentamicin, cotrimoxazole, and cephalosporins, causing various disease like wound infection, lower respiratory tract infections, urinary tract infections, and bloodstream infections difficulties to treat.¹⁶

Although Rwanda lacks surveillance and regular report on ESBL producers, clinical evidence shows increased rates of treatment failures, high empirical therapy practices, and prolonged length of hospital stays (LHS), possibly linked to antimicrobial resistance.^18,19 This study aimed to determine the magnitude of ESBL producers and the associated clinical implications at the University Teaching Hospital of Kigali in Rwanda.

MATERIALS AND METHODS

Study site.

This study was conducted at the University Teaching Hospital of Kigali, known as CHUK (Center Hospitalier Universitaire de Kigali), a public teaching tertiary referral hospital in Kigali, Rwanda, with ≈500 beds and approximately 12,000 admissions annually.

Study design and period.

A 1-year cross-sectional retrospective study was used, reviewing patients’ files and microbiology logbooks of urine and blood cultures. Data from January 1 to December 31, 2022 were recorded.

Study population.

The study focused on urine- and blood culture–positive samples. Included in the analysis were ESBL-positive E. coli and K. pneumoniae isolated from either urine or blood. Samples were collected from eight departments: internal medicine, pediatrics, neonatology, intensive care unit (ICU), emergency department (ED), outpatient department, surgery, and obstetrics and gynecology.

Inclusion requirements.

Positive urine and blood cultures were checked for possible isolation of Enterobacteriaceae strains. Antimicrobial sensitivity testing results for K. pneumoniae and E. coli were examined to confirm the ESBL phenotype.

Laboratory isolation and identification of bacteria.

Urine and blood cultures followed standard operating procedures at CHUK. Briefly, urine samples, after wet mount examination, were cultured on cysteine lactose electrolyte deficient. The colonies were counted to rule out significant growth and ≥10⁵ colony-forming units per milliliter was considered significant growth, then species identification and antimicrobial susceptibility testing were performed. For blood culture, bottles were directly incubated at 37°C in BD BACTEC, an automated blood culture machine (BD Biosciences, Franklin Lakes, NJ). The subculture of positive bottles was done on sheep blood agar and incubated overnight at 37°C, in 5% CO₂ incubator. Furthermore, bacterial species were identified manually with Gram stain reaction (Gram negative bacilli) and biochemical tests known as small gallery. The colony morphology and small gallery reactions including triple sugar iron, motility indole urease, and citrate tests were used for species identification. In addition, K. pneumoniae was easily identified with the citrate positive reaction, whereas E. coli was easily identified with indole reaction together with other biochemical reactions.

Antimicrobial susceptibility testing.

The disk diffusion method was used with various antibiotic disks: amikacin, 30 µg; ampicillin,10 µg; ceftazidime, 30 µg; cefotaxime, 30 µg; cefoxitin, 30 µg; cefuroxime, 30 µg; ciprofloxacin, 5 µg; fosfomycin, 50 µg; imipenem, 10 µg; gentamicin, 10 µg; trimethoprim/sulfamethoxazole, 1.25/23.75 µg; amoxicillin-clavulanic acid, 20/10 µg; nitrofurantoin, 300 µg; chloramphenicol, 30 µg; piperacillin, 10 µg; piperacillin tazobactam, 30/6 µg; sulbactam ampicillin, 10/10 µg; and tigecycline, 15 µg (BD Biosciences).

The saline suspension from a pure culture plate was prepared by adding bacterial colonies into sterile saline at 0.9% until the MacFarland turbidity standard of 0.5 is attained. The resulting suspension was inoculated on Mueller–Hinton agar by using a sterile cotton swab. After this procedure, the antibiotic disks were added to the plate with at least 25 mm between each disk and subsequently incubated at 37°C for 18–24 hours. Interpretation of inhibition zone diameters followed Clinical and Laboratory Standards Institute (CLSI) breakpoints. The ESBL phenotypes were determined according to CLSI guidelines, observing the synergistic phenomenon between two cephalosporins (ceftazidime and ceftriaxone or cefotaxime) and amoxicillin-clavulanic acid. Quality control followed internal standard operating procedures and used reference strains from the ATCC.

Data collection.

Extended-spectrum beta-lactamase-positive cases and their antimicrobial susceptibility patterns were recorded in an Excel worksheet. Patient files with ESBL positivity were reviewed for demographic and clinical information, including age, gender, LHS, and outcomes.

STATISTICAL ANALYSES

Data were analyzed using SPSS version 28. Frequencies, percentages, median, mean LHS, and patient outcomes were calculated and are presented in tables and figures.

RESULTS

Demographic characteristics of patients with ESBL-positive infections are presented in Table 1. Males and females were 58% and 42%, respectively, and a higher frequency of ESBL-positive phenotypes were more distributed in the ≥18 years age group (59%).

Table 1.

Demographics of studied subjects (N = 300)

Categories	n (%)
Sex
Male	174 (58)
Female	126 (42)
Age Categories
Neonate	50 (16.6)
<5 years	52 (17.3)
5–18 years	21 (7)
>18 years	177 (59)

The neonate category included children aged 0–28 days, and the <5 years category included children from 29 days up to 59 months.

Supplemental Table 1 presents the prevalence of ESBL-positive K. pneumoniae and E. coli that were isolated from urine and blood samples. In 2022 at CHUK, the rate of urine culture growth was 683/2,621 (26%), and the rate of blood culture growth was 600/1,649 (36.4%). Among 683 urine-positive cultures, 141 (20.6%) were E. coli ESBL positive, and 69 (10.1%) were K. pneumoniae ESBL positive. In addition, among 600 blood culture–positive cases, 53 (8.8%) were E. coli ESBL positive, and 37 (6.2%) were K. pneumoniae ESBL positive. Moreover, among all ESBL-positive cases (300), E. coli was more prevalent 194/300 (64.67%) than K. pneumoniae 106/300 (35.33%). The ESBL-positive bacteria were isolated more frequently from urine cultures 210/300 (70%) than blood cultures 90/300 (30%). Generally, the prevalence of ESBL-positive in urine isolates was 30.7%, whereas the prevalence of ESBL positivity in blood isolates was 15%, and the overall prevalence of ESBL producers was 23.4% (Supplemental Table 1). In Figure 1, we show the total positive cultures according to the departments. Generally, more ESBL positive cases were from surgery, internal medicine, neonatology and ICU.

Figure 1.

Distribution of extended-spectrum beta-lactamase according to the departments. Data are presented as frequency. G.O = gynecology and obstetrics; I.M = internal medicine; ICU = intensive care unit; OPD = outpatient department.

Supplemental Figure 1 presents the overall resistance rate among E. coli and K. pneumoniae ESBL-positive isolates. The isolates were 100% resistant to third-generation cephalosporins, amoxicillin-clavulanic acid, piperacillin, sulbactam ampicillin, ampicillin, cefuroxime, and cefoxitin, and 95% resistant to sulfamethoxazole trimethoprim. Moreover, the isolates were least resistant to amikacin (18%), imipenem (10%), and polymyxin B (3%).

Supplemental Table 2 presents the resistance rate of E. coli and K. pneumoniae on various antibiotics. First, E. coli had a high resistance rate (100%) to cefotaxime, ceftriaxone, ceftazidime, amoxicillin-clavulanic acid, piperacillin, cefuroxime, cefoxitin, and sulfamethoxazole-trimethoprim and a low resistance rate to polymyxin B (2.7%), imipenem (4.4%), and amikacin (12.1%). Second, K. pneumoniae had a high resistance rate to amoxicillin-clavulanic acid (100%), sulbactam ampicillin (100%), cefoxitin (100%), and piperacillin (100%), and a low resistance rate to imipenem (21.1%), amikacin (28.9%), and polymyxin B (3.1%). Overall, both E. coli and K. pneumoniae exhibited a high resistance rate to routinely used antimicrobial agents and a low resistance rate to polymyxin B, imipenem, and amikacin.

Supplemental Table 3 captures information on the LHS of the patients with an infection due to ESBL producers. The LHS ranging from 8 to 21 days was the most frequent, with a mean LHS of 12 (± 10.3) days, which was 2-fold higher the mean LHS of the hospital in 2022. The median hospital stay in the studied group was 9.5 days, and unfortunately, we could not calculate the standard deviation and median of LHS for the hospital in general in 2022 (Supplemental Table 4).

Supplemental Table 4 shows that most patients with infections due to ESBL producers were discharged (89.7%), but there was a notable death rate (10.33%). Additionally, death rates varied across departments, with higher rates observed in neonatology (26.4%), ED (20%), pediatrics (14.3%), and the ICU (10%). Furthermore, the death rate was higher for infections caused by K. pneumoniae (18%) compared with E. coli (6.2%).

DISCUSSION

Our study contributes significant insights into the prevalence of ESBL-producing Enterobacteriaceae (E. coli and K. pneumoniae) and associated clinical implications at the University Teaching Hospital of Kigali. The investigation also sheds light on the resistance patterns to various antibiotics, emphasizing the need to understand the extent of the ESBL problem and its impact on patient outcomes. This is crucial for guiding treatment strategies and formulating appropriate interventions.

In our study, the overall prevalence of ESBL was found to be 23.4%, which is lower than reported rates in various studies in Rwanda and Ethiopia.^11,20,21 The varying prevalence rates highlight the dynamic nature of ESBL phenotypes, with differences observed between sample types, departments, and bacterial species. The higher prevalence in urine samples compared with blood samples underscores the importance of E. coli, particularly in urinary tract infections, and emphasizes the ease of spread in different settings.

The predominant presence of E. coli (64.67%) compared with K. pneumoniae (35.33%) among ESBL-positive cases aligns with findings from other studies in Tanzania²² and Rwanda.^11,12 The higher prevalence in urine samples further underscores the significance of E. coli in urinary tract infections. The rise in intestinal carriage of ESBL-producing E. coli poses a considerable challenge in both community and healthcare settings.

Our study also explores the resistance patterns of ESBL-producing E. coli and K. pneumoniae to different antibiotics. A high resistance rate (100%) was observed for several commonly used antibiotics, emphasizing the urgent need for effective intervention. The results of our study align with a study conducted in Rwanda, which reported resistance rates to amoxicillin-clavulanic acid (100%), ampicillin (90.6%), and ceftazidime (78.6%).²⁰ Similar findings were reported in another study, which showed resistance to ceftazidime at 58%, ceftriaxone at 55%, cefotaxime at 44%, and gentamicin at 53%.¹² However, that study did not exclusively focus on ESBL-positive bacteria. Although some resistance rates differ in various studies, likely because of differences in sample types and tested bacterial species, all suggest that the resistance rate to commonly used antimicrobial agents is high in Rwanda, posing a significant health concern.

The lower resistance rates to amikacin, imipenem, and polymyxin B noted in our study align with previous studies that reported an overall isolate resistance rate of 4.1% to meropenem,²⁰ 7% to imipenem, and 19.1% to amikacin.^3,21 Our findings suggest that these antibiotics could serve as potential therapeutic alternatives, but caution is warranted to prevent future resistance challenges.

The LHS ranging from 8 to 21 days was the most frequent, and mean LHS was high, being 2-fold higher than the mean LHS of the hospital in 2022 (12 days versus 6 days). Additionally, the mortality rate in ESBL-producing K. pneumoniae and/or E. coli infections was notably high at 10.33%, exceeding the overall hospital mortality rate of 4% in 2022. This underscores the severity of the problem and the urgency for intervention.

Although our study provides valuable insights into resistance rates, LHS, and mortality rates associated with ESBL-producing Enterobacteriaceae infections, limitations exist. Future studies should focus on determining the cause of death in these patients and evaluating the specific factors contributing to mortality.

CONCLUSION

In conclusion, our findings emphasize the high prevalence of ESBL-producing Enterobacteriaceae, with ESBL-positive E. coli being more prevalent than K. pneumoniae. High antibiotic resistance rates, prolonged hospital stays, and increased mortality rates further highlight the gravity of infections due to ESBL-positive phenotypes. Ongoing surveillance, appropriate interventions, and caution during empirical therapy are essential for addressing this significant public health concern.

Note: Supplemental materials appear at www.ajtmh.org.