Urinary tract infections (UTIs) represent a significant health concern globally, affecting an estimated 404.6 million individuals in 2019, and contributing to 5.2 million disability-adjusted life years.¹ A systematic review of the aetiology and prevalence in sub-Saharan Africa found the prevalence to be 32.1%, with South Africa recording the highest prevalence at 67.1%.² Women have a higher incidence compared with men; in their lifetime, approximately 50% of women will be diagnosed with a UTI, with 20% – 40% experiencing recurrences.³ Globally, the most common causative organisms of community-acquired UTIs are Gram-negative bacteria, predominantly Escherichia coli and Klebsiella pneumoniae.⁴

The high incidence of a positive urinary culture frequently results in the prescription of an antimicrobial, which is often unnecessary, as in the case of asymptomatic bacteriuria. This has the effect of increasing the antimicrobial load in the environment, and exacerbation of the pandemic of antimicrobial resistance (AMR), which has steadily increased internationally because of over- and misuse of antimicrobials, both in humans and animals.⁵ Concurrently, the development of novel antimicrobial therapies has lagged significantly behind the development of AMR, highlighting the need for intensification of antimicrobial stewardship.^5,6

A recent systematic analysis of the burden of AMR has reported the alarming statistic of 4.95 million deaths worldwide related to AMR in 2019 alone.⁷ The highest mortality rates were reported in Western and sub-Saharan Africa, at approximately 27.3 deaths per 100 000 individuals. The need to implement changes that combat and reduce AMR, both globally and in South Africa, is therefore urgent.⁸

Once infection has been confirmed, the use of antibiogram to guide treatment of susceptible organisms is recommended by the Society for Healthcare Epidemiology of America, the Infectious Diseases Society of America, and the Clinical and Laboratory Standards Institute (CLSI). Local susceptibility patterns should be determined such that appropriate empiric therapy can be utilised.⁹

This study aimed to determine the rate of culture positivity for uropathogens and the antimicrobial susceptibility patterns. This is to guide antimicrobial stewardship and prescribing practices at Potchefstroom Hospital; a regional hospital servicing the JB Marks municipality and drainage area in the North West province, South Africa. In addition, the results were compared with the National Department of Health’s standard treatment guidelines and National Essential Medicines List (NEML)¹⁰ for empiric antimicrobial management of confirmed UTIs.

The NEML currently recommends the following for treatment of an uncomplicated UTI¹⁰:

Additionally, the NEML recommends the following for treatment of a complicated UTI¹⁰:

For pregnant women, the following has been recommended¹⁰:

A total of 3561 urine cultures were identified during the 4-year period, 48 of which were excluded from the final analysis because of the presence of more than one organism cultured. Therefore, a total of 3513 urine cultures were included in the data analysis. The number of positive urine cultures per year is depicted in Figure 1. During the study period, the number of samples that were positive for an organism decreased from 1014 samples in 2020 to an average of 833 per year for the following 3 years (832 for 2021; 820 for 2022; and 847 for 2023) (Figure 1).

As outlined in Figure 2, the most frequently cultured organisms were E. coli (n = 1481; 42.2%), K. pneumoniae (n = 568; 16.2%), E. faecalis (n = 249; 7.1%), P. mirabilis (n = 229; 6.5%), Enterobacter cloacae (n = 137; 3.9%), and Candida albicans (n = 119; 3.4%).

Regarding first-line antimicrobials (Table 1 and Figure 3), the aminoglycosides (gentamicin) would have been appropriate for 87.8% of the cultured organisms (p < 0.001), except K. pneumoniae, which was susceptible in only 77.7% of isolates. More than 90% of all organisms isolated were susceptible to amikacin (p = 0.001).

The cumulative susceptibility to nitrofurantoin was 67.5% (p < 0.001). Of the five most commonly isolated organisms, E. coli isolates were the only susceptible organism (92.2%). Other isolates, such as K. pneumoniae (40%), E. cloacae (23.4%) P. mirabilis (1.4%), E. faecalis (0%), all had reduced susceptibility to nitrofurantoin.

A similar trend of emerging resistance to ciprofloxacin was observed in four of the five most frequently isolated pathogens in this study; P. mirabilis remained the only persistently sensitive organism (92.5%), while reduced susceptibility was detected in E. cloacae (78.8%), K. pneumoniae (78.2%), E. faecalis (76.9%), and E. coli (74.3%). Overall, organisms had a cumulative sensitivity of 77.2% (p < 0.001).

Individual sensitivity to trimethoprim sulfamethoxazole were as follows: E. cloacae (60.3%), P. mirabilis (54%), K. pneumoniae (51.5%), E. coli (36.7%), and E. faecalis (0%). Cumulatively, sensitivity to trimethoprim sulfamethoxazole was 42.8% (p < 0.001).

Significant resistance to amoxicillin/clavulanic acid was present in E. cloacae (3.6% susceptible), K. pneumoniae (65.2% susceptible), and E. coli (79.6% susceptible), with 72.3% of all organisms isolated being susceptible (p < 0.001). There was a similar pattern with ampicillin, with an overall susceptibility of 30.6% of all the cultured organisms (p < 0.001). The susceptibility of individual organisms to amoxicillin/clavulanic acid was also greatly reduced, namely, K. pneumoniae (1.6%), E. cloacae (27.3%), E. coli (28.8%), and P. mirabilis (55.1%). Ampicillin showed continued sensitivity of 93.5% for E. faecalis.

For second-line antimicrobials, the cumulative susceptibility of the organisms to cefuroxime was 70.6% (p < 0.001), cefotaxime/ceftriaxone (79.5%; p < 0.001), and cefepime (77.9%; p < 0.001) (Table 1). The third- and fourth generation cephalosporins were effective against P. mirabilis, with sensitivities as high as 96% (third generation) and 95.2% (fourth generation). The sensitivity of E. coli to third generation cephalosporins was 82.2%, and to fourth generation, 81.8% and 66.4% (third generation) and 62.8% (fourth generation) for K. pneumoniae. Similar rates of resistance were seen in E. cloacae, with susceptibilities of 77.6% (third generation) and 63% (fourth generation).

The overall cumulative sensitivity to meropenem was 94.1% (p = 0.004), and 87.5% (p < 0.001) to imipenem. Proteus mirabilis was only sensitive in 96.3% of cases to meropenem, and to imipenem in 21.4% of cases. E. coli, K. pneumoniae and E. cloacae had sensitivities of over 83% to bothimipenem and meropenem.

The cumulative sensitivity to piperacillin–tazobactam was 87.1% (p = 0.017). Escherichia coli was relatively sensitive, at 89.4%, as was P. mirabilis, at 96.4%, whereas K. pneumoniae (76.5%) and E. cloacae (72.7%) both had reduced sensitivity.

Candida albicans isolates were, as expected, relatively sensitive to fluconazole (92.9%; p < 0.001), amphotericin B (92.9%; p < 0.001) and voriconazole (94.9%; p < 0.001) (Table 3).

During the study period, Pseudomonas aeruginosa was isolated on 33 occasions. Susceptibility testing reached the threshold of 30 or more isolates for four antibiotics: amikacin (sensitivity 97%), gentamicin (sensitivity 50%), ciprofloxacin (sensitivity 76%), and piperacillin–tazobactam (sensitivity 94%) (Table 2). Acinetobacter baumannii was isolated 47 times, but only three antibiotics were tested in 30 or more isolates. These included ciprofloxacin (37% sensitivity), gentamicin (42% sensitivity), and piperacillin–tazobactam (28% sensitivity) (Table 2).

There were significant changes in trend noted for second-, third-, and fourth generation cephalosporins, and nitrofurantoin over the study period (Table 4). The organisms demonstrated an increase in sensitivity to cefepime, rising from 73.6% to 78.6% (p = 0.033). In contrast, sensitivity to cefotaxime/ceftriaxone and cefuroxime declined over the study period, decreasing from 79.5% to 75.6% (cefotaxime/ceftriaxone; p = 0.006) and from 68.7% to 63.9% (cefuroxime; p = 0.009). Sensitivity to nitrofurantoin also improved over the study period, from 32.8% in 2020 to 37.2% by 2023 (p = 0.013).

The prevalence of the various uropathogens identified in this study aligns with findings from a recent investigation at RK Khan Hospital in KwaZulu-Natal, South Africa, which employed a similar sample size and methodology. In this study, E. coli was the predominant pathogen, followed by K. pneumoniae and E. faecalis.¹²

The authors of the RK Khan Hospital study found that resistance rates to nitrofurantoin were low in E. coli, at 6.2%, and high in K. pneumoniae, at 61.3%.¹² These figures reflect our findings. A multicentre local study conducted by Lewis et al. in 2013 forms part of the basis of the NEML, in which good cumulative sensitivity to nitrofurantoin (91.7%) was noted.¹³ The markedly reduced sensitivities of K. pneumoniae (40%), E. cloacae (23.4%), and P. mirabilis (1.4%) highlight emerging resistance and make nitrofurantoin less useful in our setting for empiric treatment. Additionally, the CLSI states that P. mirabilis is intrinsically resistant to nitrofurantoin, which reflects our results.¹⁴

Given that resistance patterns tend to differ by geographical location, and the current recommendations by the NEML are based partially on American data, we believe it was necessary to verify the NEML recommendations for the use of gentamicin. The suggested utilisation of gentamicin in the guidelines was based on a systematic review, published in 2018, in Arizona (United States) which showed that 94.5% of pathogens isolated from pooled studies were susceptible to gentamicin.¹⁵ At a local level, Naidoo et al. also noted a lower cumulative susceptibility of 65% in K. pneumoniae.¹² In this study, K. pneumoniae was sensitive to gentamicin in 77.7% of isolates, a 12.7% higher sensitivity rate when compared to Naidoo et al. The use of gentamicin may still be an effective empiric antimicrobial in our setting, as this study revealed an 87.8% cumulative sensitivity overall.

The Infectious Diseases Society of America guidelines recommend that ciprofloxacin be used as empiric therapy for complicated UTIs, only if resistance rates are below 10%.¹⁶ Lewis et al. noted that cumulative sensitivity to fluoroquinolones (94.1%) had been maintained; however, Naidoo et al. published alarming ciprofloxacin sensitivity rates, with only 62% of E. coli and 70% of Klebsiella spp. found to be sensitive.^12,13 The authors highlighted that the reduced sensitivity would lead to high treatment failure rates with empiric ciprofloxacin use.¹² Another study conducted at 3 Military Hospital (Bloemfontein) in 2011 also demonstrated that cultures in complicated UTIs showed a reduced sensitivity to ciprofloxacin (41%),¹⁷ and the authors recommended against its use at that hospital.¹⁷

Ciprofloxacin has been reported to promote resistance through horizontal transmission of resistance genes. contributing collateral damage with higher resistance rates in other classes of antimicrobials.^17,18 Emerging resistance to ciprofloxacin was noted in four of the five common pathogens in the current study, specifically E. cloacae (79%), K. pneumoniae (78%), E. faecalis (77%) and E. coli (74%), with P. mirabilis remaining susceptible, at 92.5%. Ciprofloxacin was once the preferred treatment for UTIs, but this is no longer the case in our setting.

Lewis et al. also reported reduced cumulative sensitivity rates of 82.8% to amoxicillin–clavulanate in 2013,¹³ and similarly, Naidoo et al. found low sensitivity in both K. pneumoniae (58%) and E. coli (70%).¹² A high level of resistance towards amoxicillin clavulanate, with reduced cumulative sensitivity of 72.3%, was also noted in our study, indicating that these agents may not be effective towards the two most prevalent uropathogens, namely E. coli and K. pneumoniae. Whereas Lewis et al. reported a high rate of sensitivity to cefuroxime (93.4%) in 2013,¹³ Naidoo et al. showed that sensitivity to third-generation cephalosporins was only 74% in E. coli, and 58% in K. pneumoniae, which correlated with the reduced sensitivities in this study.¹² The inherent resistance to cephalosporins found in E. faecalis has led the NHLS to forgo susceptibility testing to cephalosporins.¹⁹

Enterobacter cloacae has been reported to be intrinsically resistant to amoxicillin–clavulanate by the CLSI, which is in line with our study (96.4%).¹⁴ Furthermore, this organism is an AmpC-producer and is therefore at high risk of developing AmpC enzyme overproduction, subsequently leading to third-generation cephalosporin resistance.¹⁴ Although we have variable sensitivities in our beta-lactams, CLSI recommend against the use of third-generation cephalosporins and piperacillin–tazobactam because of the risk of treatment failure.¹⁴

Naidoo et al. demonstrated meropenem sensitivities of 95% for K. pneumonia and 100% for E. coli.¹² These findings mirrored the cumulative sensitivity of our study, of 94.1% overall. Similarly, the authors reported the sensitivity to piperacillin–tazobactam to be 76% for K. pneumoniae and 95% for E. coli, which is consistent with our findings of 76.5% for K. pneumoniae and 89.4% for E. coli.¹²

There was a reduced cumulative sensitivity to trimethoprim–sulfamethoxazole at 42.8%. These figures were mirrored by Lewis et al., with a cumulative sensitivity of 44.3%, and Naidoo et al., with 39% E. coli and 52% K. pneumoniae susceptibility.^12,13 The Infectious Diseases Society of America guidelines recommend the use of trimethoprim–sulfamethoxazole for empiric therapy for uncomplicated UTIs only if local uropathogen resistance rates do not exceed 20%.¹⁶ In our setting, and considering sensitivities from other local studies, trimethoprim–sulfamethoxazole would unlikely be suitable as an empiric agent.^12,15,17

Fosfomycin susceptibility testing is not routinely performed by NHLS North West, and thus no analysis was performed.

Future therapies include the possibility of using vaccines to prevent recurrent UTIs. A meta-analysis by Mak et al. in 2024, showed limited evidence for this approach, but that more rigorous research is needed in order to determine the value of vaccines and oral bacterial lysates immunostimulants as an option for prophylactic treatment.²⁰

We recommend that E. coli be considered as the culprit pathogen in both in-and outpatient UTIs and, in compliance with most international guidelines, nitrofurantoin should be used as first-line agent. This does not negate the need for culture and sensitivity to be performed, as well as de-escalation or escalation of therapy, as needed. If parenteral therapy is necessary, most organisms remain susceptible to gentamicin, and although we do not have data for fosfomycin at Potchefstroom Hospital, other local data support its use for empiric therapy.