Sepsis is a serious life-threatening condition that commonly manifests itself in the cancer patients. Although there are studies that have been conducted in Africa on cancer patients presenting with sepsis (10,13), limited data regarding the profiles of the organisms implicated and antibiotic susceptibility data exist. In this study we report the isolation rate of bacterial and fungal pathogen from blood cultures of cancer patients (both adults and paediatric) presenting with sepsis, as well as the effectiveness of commonly used antibiotics in our setting. We also demonstrate that there is a rise in the level of resistance among pathogen causing sepsis in our setting.
The majority (90.8%) of the study participants were on at least one prophylactic antimicrobial agent before blood culture collection and this was as a consequence of their immunosuppression being caused by the cancer. However, it was also observed that 82 (57.7%) were on a cocktail of 3 to 6 broad spectrum antimicrobial agents contrary to the standard empirical treatment of sepsis stipulated in the local EDLIZ (9). Ceftriaxone and gentamicin were the major empirical antibiotics used despite the recommendations that ceftriaxone should only be used in complicated cases (9). Overuse of ceftriaxone and others antimicrobials could be due to limited knowledge on the implications such as antimicrobial resistance and presumed resistance to prescribed empiric treatment.
The microbial pathogen isolation rate was 35.2% and this was lower than expected as most of our participants had prior exposure to antimicrobial therapy. Other studies from high income countries have, on the contrary, reported lower prevalence of sepsis among patients with cancer including studies in Oman (5.0%) and Europe (17%) (14,15). Among the isolates identified, Gram positive to Gram negative percentage ratio was 57:43 which was comparable with the median ratio of 60:40 (range 85:15 to 26:76) obtained in Europe (16,17). This reflects a similarity in the distribution of organisms despite geographical differences although minor difference can be encountered, like a study in Sudan where the ratio was 83%:17% (10). Most of the isolates (78.0%) came from patients with haematological malignancies, a finding comparable to other earlier studies (4,15). The major aetiological agents of sepsis obtained from patients with haematological cancers were CoNS, E. coli, E. faecalis and K. pneumoniae. Similarly, other studies from Europe have reported the same organisms as the causative agents of sepsis but with some minor variations in proportions (15,16). Most studies had not stratified aetiological agents with cancer type but a study in Europe with the same stratification showed similar aetiological agents between the two major cancer groups (15).
Amikacin and meropenem were the most potent drugs against gram negative isolates with more than 80.0% of the isolates being sensitive, similar to findings from a study in the USA (7). Conversely, more than 60.0% of the isolates were resistant to third generation cephalosporins, in contrast with the USA and an earlier study in Zimbabwe where 80-100% were sensitive (7,18). This difference could be due to the wide availability and uncontrolled use of ceftriaxone as first line treatment, as was found in this study. As also shown in this and other studies (16,19–21), the increase in the emergence of ESBL producing isolates has also lead to this high level of resistance to the third generation cephalosporins. Gentamicin, the most commonly used empirical aminoglycoside, also had a low activity against these gram negative isolates as noted with more than 60.0% of the isolates being resistant. Resistance to the third generation cephalosporins and gentamicin has been reported in earlier studies to be rising in low-income countries (20,22). Cefoxitin and piperacillin-tazobactum were effective against 66.7% and 71.4% of all the gram negative isolates. However, more than 90.0% of the isolates were resistant to trimethoprim-sulfamethoxazole and ampicillin, a finding similar to most studies worldwide (18,22,23). The resistance to trimethoprim-sulfamethoxazole has been attributed to overuse of the drug as prophylaxis against Pneumocystis jirovecii pneumonia in HIV endemic regions such as Zimbabwe.
Unsurprisingly, due to their limited use locally, minocycline, chloramphenicol, linezolid and vancomycin showed to be effective against more than 80.0% of the Staphylococcus isolates. A moderately high activity was displayed by gentamicin, clindamycin and erythromycin. These results were partly in agreement with findings from Ghana and India (22,24). The low activity observed in penicillin was previously reported in Ghana, India and Zimbabwe (18,22,24). Enterococcus and Streptococcus species in our study were sensitive to fosfomycin, vancomycin and linezolid with the latter being the most effective (isolates were 100.0% sensitive) antibiotic. Contrary to findings in India where they found 50% of Enterococcus species to be susceptible to vancomycin, all our isolates were sensitive vancomycin (25). These isolates also displayed a moderate sensitivity to gentamicin, ampicillin and penicillin. Only one fungal isolate was obtained in this study and antimicrobial susceptibility patterns could not be fully established for statistical inference.
Some isolates phenotypically showed multidrug resistance capabilities. Our methicillin resistance carriage was comparable to USA isolates where MRSA was 50.0% in our current study versus 41.0% in USA while that of MR-CoNS was 75.0% versus 72.0% respectively (7). In Ghana, a low proportion of MRSA (5.8%) was reported in contrast to our findings (22). This high level methicillin resistance limits the choices of antimicrobial treatment since it also implies that these isolates will also be clinically resistant to most if not all commonly used beta-lactam antibiotics. We also found a high proportion of ESBL producers among E. coli and K. pneumoniae isolates and this was in agreement with some studies around the world (16,19,20,26). However, of note was a higher proportion of ESBL producing E. coli (75.0%) than K. pneumoniae (57.1%), a different finding from most reports in other parts of the world where ESBL production is predominantly found in K. pneumoniae isolates (19,20).
In summary, sepsis remains a leading cause of morbidity and mortality among patients with cancer with the major aetiological agents being CoNS, E. coli, K. pneumoniae, E. faecalis and S. aureus. Similar aetiological pathogens were present in both haematological and solid cancers in the Zimbabwean population. Most of the microbial aetiological agents of sepsis showed high levels of resistance to commonly used antimicrobial drugs. Resistance to gentamicin, penicillin and third generation cephalosporins is a major cause for concern as these are the major empirical antibiotics in resource limited settings. Apart from vancomycin, linezolid was shown to be another better option to be considered in the treatment of serious and non-responsive gram positive infections while amikacin and meropenem can also be considered in gram negative infections. The emergence of multidrug resistance mechanisms like ESBL carriage and methicillin resistance among isolates is disturbing. This indicates the need for surveillance to reduce their transmission with a goal to mitigate mortality and morbidity among patients.