https://orcid.org/0000-0002-1388-3901
Figures
Abstract
Background
Polymicrobial bloodstream infections (BSI) are difficult to treat since empiric antibiotics treatment are frequently less effective against multiple pathogens. The study aimed to compare outcomes in patients with polymicrobial and monomicrobial BSIs.
Methods
The study was a retrospective case-control design conducted at Muhimbili National Hospital for data processed between July 2021 and June 2022. Cases were patients with polymicrobial BSI, and controls had monomicrobial BSI. Each case was matched to three controls by age, admitting ward, and duration of admission. Logistic regression was performed to determine independent risk factors for in-hospital and 30-day mortality.
Results
Fifty patients with polymicrobial BSI and 150 with monomicrobial BSI were compared: the two arms had no significant differences in sex and comorbidities. The most frequent bacteria in polymicrobial BSI were Klebsiella pneumoniae 17% (17/100) and Enterobacter species 15% (15/100). In monomicrobial BSI, S. aureus 17.33% (26/150), Klebsiella pneumoniae 16.67% (25/150), and Acinetobacter species 15% (15/150) were more prevalent. Overall, isolates were frequently resistant to multiple antibiotics tested, and 52% (130/250) were multidrug resistance. The 30-day and in-hospital mortality were 33.5% (67/200) and 36% (72/200), respectively. On multivariable analysis, polymicrobial BSIs were independent risk factors for both in-hospital mortality (aOR 2.37, 95%CI 1.20–4.69, p = 0.01) and 30-day mortality (aOR 2.05, 95%CI 1.03–4.08), p = 0.04). In sub-analyses involving only neonates, polymicrobial BSI was an independent risk factor for both 30-day mortality (aOR 3.13, 95%CI 1.07–9.10, p = 0.04) and in-hospital mortality (aOR 5.08, 95%CI 1.60–16.14, p = 0.006). Overall, the median length of hospital stay post-BSIs was numerically longer in patients with polymicrobial BSIs.
Citation: Manyahi J, Joachim A, Msafiri F, Migiro M, Mwingwa A, Kasubi M, et al. (2024) Polymicrobial bloodstream infections a risk factor for mortality in neonates at the national hospital, Tanzania: A case-control study. PLoS ONE 19(4): e0302076. https://doi.org/10.1371/journal.pone.0302076
Editor: Iddya Karunasagar, Nitte University, INDIA
Received: September 13, 2023; Accepted: March 26, 2024; Published: April 16, 2024
Copyright: © 2024 Manyahi et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Minimal data have been included in the submission as a Supporting information.
Funding: The author(s) received no specific funding for this work.
Competing interests: The authors have declared that no competing interests exist.
Background
Bloodstream infections (BSI) in low-middle-income countries (LMICs) are steadily associated with poor treatment outcomes, with the rising concern associated with antibiotic-resistant bacteria. BSI has been reported in 10–13% of inpatients with severe febrile illness in East and West Africa, with an estimated case fatality rate of 12% [1, 2].
Following the reduction of malaria, a frequent cause of febrile illness in Tanzania, BSIs remain a common cause of hospitalization in children with febrile illness. In Tanzania, BSIs is detected in 11–14% of children with severe febrile illness [3, 4]. Studies have shown that BSIs in adults and children result in worse outcomes, such as high mortality and length of hospital stay [5–8]. In addition, increasing evidence of antibiotic-resistant bacteria, including methicillin-resistant Staphylococcus aureus (MRSA), extended-spectrum beta-lactamases (ESBL), and multidrug-resistant (MDR) bacteria in BSIs, has been reported in Tanzania [3–5, 9].
Studies in Tanzania have reported that BSIs due to antibiotic-resistant bacteria had a significantly higher risk of death than those with antibiotic susceptible bacteria [5, 8, 9]. Mortality in children with ESBL BSI was approximately 2-fold higher than in non-ESBL septicemia [9]. Our recent study found that patients with MDR BSIs (aOR = 15) had a higher risk of dying than those with non-MDR BSIs [5]. However, studies still need to evaluate the outcomes of polymicrobial BSIs in Tanzania thoroughly.
Polymicrobial BSIs are frequently difficult to treat, and choosing the appropriate empiric antibiotics effective for more than one pathogen is more challenging. It is, therefore, vital to promptly identify the pathogens causing polymicrobial BSIs and administer the proper empiric antibiotics/drugs. In addition, studies on adults and children from other settings have demonstrated that mortality attributable to polymicrobial BSIs is higher than in monomicrobial BSIs [10–13]. Therefore, we carried out this retrospective case-control study at Muhimbili National Hospital (MNH) to ascertain the impact of polymicrobial BSIs on patient outcomes. A better understanding of the risk factors for worse outcomes in patients with BSIs would improve clinical outcomes for patients.
Materials and methods
Study setting and design
This was a retrospective case-control study for the data processed between July 2021 and June 2022 at MNH, Dar es Salaam, Tanzania. The data for this study were accessed and collected between July 1, 2022, and September 30, 2022. MNH is a tertiary-level hospital in Tanzania and serves as a teaching hospital for Muhimbili University of Health and Allied Sciences (MUHAS). It has a 1,500-bed capacity and attends up to 2,000 outpatients daily. The hospital has an infection control policy and an infection prevention and control (IPC) unit. Furthermore, blood culture sample in all wards is collected under aseptic condition and transferred to the laboratory within one hour of collection. Under ideal situations, the laboratory reporting system allows to flag clinicians with preliminary results immediately when the blood culture flag-positive and final results 2–3 days after the blood culture is positive. In this retrospective case-control study, cases were patients with polymicrobial BSIs, and controls were patients with monomicrobial BSIs. Each case was matched to three controls by age ±5 years (neonates ±5 days), admission date, and admitted wards.
Blood culture results of patients aged 0–86 years with polymicrobial and monomicrobial growth were first retrieved from the laboratory registry. Patients’ information, including ages, sex, admission wards, date of admission, time of diagnosis of BSIs, date of discharge, comorbidity, and clinical outcomes (death or survival), were collected from clinical case notes. All cases and their matched controls with complete demographics and clinical information were included in the analysis. Only the first episode was included in cases where patients had two or more blood cultures positive with similar organisms. Patients with incomplete demographics and medical records were excluded from the analysis.
Microbiological methods for blood culture at MNH
Blood samples for culture were incubated in the BD BACTEC FX 200 at MNH, central pathology laboratory. Only one set of blood cultures was performed in adults and pediatrics. A set of BD BACTEC Plus Aerobic and Anaerobic/F Culture Vials (Becton Dickinson, BD) were processed for adults. For pediatrics, 1–3 mL of blood sample was collected in the BD BACTEC Peds PlusTM/F Culture Vials. Blood culture vials in BD BACTEC 200 were incubated for a maximum of five days before being regarded as a negative culture.
Positive blood cultures were Gram-stained and sub-cultured into sheep blood agar, chocolate agar, and MacConkey agar. Identification of bacteria was based on colonial morphology, Gram stain, and biochemical tests, including conventional API20E (BioMérieux, Marcy, I’Etoile, France) and Staphaurex (Remel Europe Ltd, Dartford, UK). Antimicrobial susceptibility testing was performed and interpreted according to clinical and laboratory standard institute (CLSI) guidelines [14].
Definition of terms
Polymicrobial BSI was defined as bacteremia with at least two different organisms from the same blood culture. The onset of BSI was defined as the day when the blood was drawn for culture that subsequently produced positive pathogenic growth. Laboratory-confirmed BSIs was defined as blood culture with the growth of recognizable pathogens in patients with signs and symptoms of BSIs. Coagulase-negative staphylococci were considered a pathogen in immunosuppressed patients and neonates, in accordance with hospital blood culture standard operational procedure. Micrococcus, Corynebacterium, Viridans Group streptococci, and Bacillus species were considered as contaminants. The 30-day mortality was defined as death from any cause within 30 days of the onset of BSI. In-hospital mortality was defined as death from any cause during index hospitalization following diagnosis of BSI. The length of hospital stay was calculated from the date of diagnosis of BSIs. It excluded all patients with in-hospital mortality.
In addition, appropriate treatment was defined as empirical antimicrobial therapy with an antibiotic that has in vitro activity against the isolated pathogen based on susceptibility testing results.
Statistical analysis
Categorical variables were presented as frequency and percentages. A comparison of categorical variables between cases and controls was performed using Pearson chi-square. The median length of hospital stays among cases, and controls was compared using the Wilcoxon rank-sum test. Binary logistic regression compared outcomes of interest (30-day and in-hospital mortality) between cases and controls. All plausible biological variables (Admission ward, comorbidity, and appropriate treatment) with p-values < 0.20 on the bivariate analysis were included in the logistic regression modal for the multivariable analysis.
Multivariable logistic regression was performed to find an independent association of polymicrobial and monomicrobial BSIs to 30-day and in-hospital mortality. We adjusted for age, sex, admission ward, appropriate treatment, and comorbidity as potential confounders, defining confounders as factors that change the effect size by ≥10%.
We performed a sub-analysis for neonates and non neonates, where bivariate and multivariable logistic regression were performed. Independent associations between polymicrobial BSIs and outcomes of interest (30-day and in-hospital mortality) were adjusted for age, sex, admission ward, and comorbidity. All statistical analysis was performed using STATA version 16 (College Station, TX). A significance level of ≤0.05 was used, and all p-values refer to two-sided tests.
Ethics approval and consent to participate
Ethical approval was granted by the Muhimbili University of Health and Allied Sciences senate and publications committee Ref. No DA.25/111/01C/157. Permission to conduct the study was approved by Muhimbili National Hospital management. As the study was a retrospective study using identified patients, informed consent was waived by Muhimbili University of Health and Allied Sciences senate and publication committee. All methods were carried out under relevant guidelines and regulations in the declaration.
Results
Demographics and clinical characteristics
Between July 2021 and June 2022, 3398 blood cultures were processed at the microbiology laboratory of MNH. Overall, the proportion of BSI was 19% (642/3398). From 642 patients with BSI, we obtained 50 cases with polymicrobial BSI, matched to 150 controls with monomicrobial BSI in a ratio of 1:3. The age distributions among the cases and controls ranged from 1 day to 86 years, with majority of partipants being neonates. There were no significant differences in sex and comorbidities between polymicrobial BSI and monomicrobial BSI (Table 1).
Microbiology of bloodstream infections
Table 2 provides the microbiology for both polymicrobial and monomicrobial BSIs. The bacteria-causing polymicrobial and monomicrobial BSIs were similar, with the predominance of Gram-negatives, although they varied in frequency. The most frequent bacteria in polymicrobial BSIs were Klebsiella pneumoniae 17% (17/100) and Enterobacter species 15% (15/100). Combinations of these two bacteria, as well as S. aureus/Pseudomonas aeruginosa and Acinetobacter species/CoNS, were also frequent, occurring in three instances each. In monomicrobial BSIs, S. aureus 17.33% (26/150), Klebsiella pneumoniae 16.67% (25/150), and Acinetobacter species 10% (15/150) were more prevalent (Table 2).
Antimicrobial susceptibility pattern
Overall, isolates were frequently resistant to multiple antibiotics tested, and 53% (130/243) were multidrug resistance. We observed that 77% (27/35) of S. aureus and 85% (25/29) of coagulase-negative Staphylococcus were resistant to cefoxitin, hence methicillin-resistant strains. All Enterobacterales were frequently resistant to all antibiotics tested except for meropenem and amikacin. Furthermore, S. aureus was less resistant to chloramphenicol (28%), clindamycin (38%), and trimethoprim-sulfamethoxazole (45%) (Table 3).
Outcomes in bloodstream infections
Overall, the 30-day mortality was 33.5% (67/200). The risk of mortality was higher in polymicrobial BSI compared to monomicrobial BSI, with an odd ratio (OR) of 1.83 (95% confidence interval (CI) 0.95–3.54), though the difference was not statistically significant (p = 0.07). However, after adjusting for possible confounders including sex, age, the admission ward, appropriate treatment, and comorbidity in a multivariable model, polymicrobial BSI independently predicted 30-day mortality, aOR 2.05, 95%CI 1.03–4.08), p = 0.04. Additionally, in-hospital mortality was 36% (72/200), with cases dying more frequently than controls (50% vs. 31.3%). We observed that polymicrobial BSI was a risk factor for in-hospital death based on univariable analysis crude odd ratio (cOR), 2.19, 95%CI 1.14–4.21, p = 0.02. We considered sex, age, the admission ward, appropriate treatment, and comorbidity in the final multivariable analysis model as potential confounding variables. After adjusting for these variables, we found that polymicrobial BSIs adjusted odd ratio (aOR) 2.37, 95% CI 1.20–4.69, p = 0.01, had an independent association with in-hospital mortality (Table 4).
In sub-analyses involving only neonates, we found that polymicrobial BSI had an independent association with both 30-day mortality (aOR 4.02, 95% CI 1.32–12.32, p = 0.01) and in-hospital mortality (aOR 6.18, 95% CI 1.89–20.22, p = 0.003) after adjusting for, sex, appropriate treatment, and comorbidities. For all age groups and neonates, the median length of hospital stay post-BSIs was numerically longer in patients with polymicrobial than in those with monomicrobial BSIs, though the difference was not statistically significant (Table 5).
In a sub-analysis of non-neonates, we observed that polymicrobial BSI had no association with both in-hospital and 30-day mortality (Table 5).
Discussion
Our study demonstrates that polymicrobial BSIs are significantly associated with in-hospital mortality compared to monomicrobial BSIs in patients of all age groups. We observed significantly higher 30-day mortality in participants with polymicrobial BSI than those with monomicrobial BSI. We also found that patients with polymicrobial BSIs had a longer median length of hospital stay than patients with monomicrobial BSIs, but the differences were not statistically significant. A similar finding has been reported from previous studies, which found polymicrobial BSIs increase the risk of mortality [11, 15]. The difficulties in selecting appropriate empiric antimicrobials targeting multiple pathogens and delays in reporting blood culture results could impact our study’s observed findings.
Most participants in this study received inappropriate treatment, which was more common in the polymicrobial BSI arm; however, we adjusted for inappropriate treatment in our analysis, and polymicrobial remained the main contributor to mortality. We could not account for the turn-around time of reporting polymicrobial BSI bacteriology results to clinicians as a possible contributor to the delay in initiation of treatment, possibly leading to increased mortality. Subsequently, ceftriaxone was the most commonly prescribed empiric antibiotic on review of clinical case notes, and most bacteria we isolated were resistant to this antibiotic. Our findings call for microbiologists and clinicians to be more vigilant when blood culture reveals polymicrobial BSI. We noted that the blood culture standard operating procedure does not detail how to communicate when polymicrobial BSIs are detected. On the other hand, the guidelines for treating sepsis consider only aerobic and anaerobic polymicrobial BSIs, and consideration is not given to aerobic-aerobic polymicrobial BSIs.
When we examined the neonates’ risk factors for 30-day mortality, we found polymicrobial BSIs were an independent risk for mortality in neonates after adjusting for all possible confounders. The previous study in other settings has observed similar findings in neonates, where polymicrobial bloodstream infections were associated with 3 fold increased risk of mortality compared to monomicrobial bloodstream infection [10]. However, other studies in neonates have found comparable 30-day mortality between polymicrobial and monomicrobial BSIs [16, 17]. Immature immunity in neonates increases vulnerability to serious infections leading to an increased mortality risk. On the other hand, inappropriate antibiotic therapy compounded by overwhelming virulence from more than one pathogen would have led to the severity of illness in neonates, leading to the worse clinical outcomes observed.
Studies have reported that polymicrobial infections are associated with forming polymicrobial biofilms, which tend to increase pathogens’ virulence, host immune evasion, and antibiotic resistance, hence poor treatment outcomes [18–20]. Polymicrobial BSI occurs more in neonates with severe underlying conditions and prolonged indwelling central venous catheters [16]. As central venous catheters were not commonly used at MNH during study time, the observed cases can be attributed to overwhelming illness.
We found that the median length of hospital stay was longer in patients with polymicrobial BSIs than in patients with monomicrobial BSIs. This could be due to the ineffectiveness of empiric antibiotic therapy since it was impossible for the prescribed empiric antibiotics to be effective in two different organisms. Therefore, the likelihood of receiving inappropriate treatment resulting in failure and prolonged hospital stays was obvious. Our findings are similar to those reported in the studies conducted in other settings [13, 21–23]. This finding calls for establishing a local cumulative antibiogram to guide empiric therapy in suspected bacteria infections, as it currently not existing in health facilities in the country.
The microbiology of polymicrobial BSIs was similar to that of monomicrobial infections in this study, with the predominant Gram-negative bacteria. This indicates that the sources for polymicrobial and monomicrobial BSIs are the same. Other studies examining the microbiological profiles of polymicrobial and monomicrobial BSIs have reported similar microbiology patterns in both cases [12, 17, 24]. However, depending on the variability of the population studied, the predominance of either Gram-positive or Gram-negative has been reported. On the other hand, we found that most bacteria causing BSIs in our study were MDR, which limits the available treatment options in our settings. The findings of this study emphasize the need for microbiology laboratories to promptly perform and report antimicrobial susceptibility results to guide evidence-based antibiotics therapy.
One of the important caveats of our study was that it was a retrospective design dependent on recorded information from clinical case notes. Information like the severity of BSIs and the classification of BSIs (community onset versus hospital onset) was not documented in the clinical case notes, and this means missing data could introduce information bias. However, missing end-point outcomes would be non-differentially distributed evenly across polymicrobial and monomicrobial BSIs. Furthermore, we performed multivariable analysis to control for possible confounding variables associated with mortality. Another important limitation of our study is that we could not assess the risk factors for polymicrobial BSIs because of the missing information. Understanding risk factors for polymicrobial BSIs would be critical to controlling and preventing infections before they even occur. However, our findings call for microbiologists and clinicians to manage polymicrobial BSIs because of the associated risk aggressively.
Conclusion
In neonates, polymicrobial BSI is an independent risk for 30-day and in-hospital mortality. Overall, the median length of hospital stay was longer in patients with polymicrobial BSI than those with monomicrobial BSI. These findings highlight the need for clinicians to be more aggressive when polymicrobial BSI is detected. The laboratory should also prompt notify clinician when polymicrobial growth is detected in blood culture.
Acknowledgments
The authors acknowledge the members of the bacteriology section at the central pathology laboratory, MNH.
References
- 1. Prasad N, Murdoch DR, Reyburn H, Crump JA. Etiology of Severe Febrile Illness in Low- and Middle-Income Countries: A Systematic Review. PLoS One. 2015;10(6):e0127962. Epub 20150630. pmid:26126200.
- 2. Marchello CS, Dale AP, Pisharody S, Rubach MP, Crump JA. A Systematic Review and Meta-analysis of the Prevalence of Community-Onset Bloodstream Infections among Hospitalized Patients in Africa and Asia. Antimicrob Agents Chemother. 2019;64(1). Epub 20191220. pmid:31636071.
- 3. Moyo SJ, Manyahi J, Blomberg B, Tellevik MG, Masoud NS, Aboud S, et al. Bacteraemia, Malaria, and Case Fatality Among Children Hospitalized With Fever in Dar es Salaam, Tanzania. Front Microbiol. 2020;11:2118. Epub 20200910. pmid:33013772.
- 4. Seni J, Mwakyoma AA, Mashuda F, Marando R, Ahmed M, DeVinney R, et al. Deciphering risk factors for blood stream infections, bacteria species and antimicrobial resistance profiles among children under five years of age in North-Western Tanzania: a multicentre study in a cascade of referral health care system. BMC Pediatr. 2019;19(1):32. Epub 20190126. pmid:30684964.
- 5. Manyahi J, Kibwana U, Mgimba E, Majigo M. Multi-drug resistant bacteria predict mortality in bloodstream infection in a tertiary setting in Tanzania. PLoS One. 2020;15(3):e0220424. Epub 20200304. pmid:32130227.
- 6. Godfrey E, Majaliwa E, Assenga EN. Aetiology, antimicrobial susceptibility and outcome of children with sepsis, admitted at Muhimbili National Hospital, Dar es Salaam. Pan Afr Med J. 2022;42:167. Epub 20220701. pmid:36187027.
- 7. Kortz TB, Sawe HR, Murray B, Enanoria W, Matthay MA, Reynolds T. Clinical Presentation and Outcomes among Children with Sepsis Presenting to a Public Tertiary Hospital in Tanzania. Front Pediatr. 2017;5:278. Epub 20171222. pmid:29312910.
- 8. Blomberg B, Manji KP, Urassa WK, Tamim BS, Mwakagile DS, Jureen R, et al. Antimicrobial resistance predicts death in Tanzanian children with bloodstream infections: a prospective cohort study. BMC Infect Dis. 2007;7:43. Epub 20070522. pmid:17519011.
- 9. Blomberg B, Jureen R, Manji KP, Tamim BS, Mwakagile DS, Urassa WK, et al. High rate of fatal cases of pediatric septicemia caused by gram-negative bacteria with extended-spectrum beta-lactamases in Dar es Salaam, Tanzania. J Clin Microbiol. 2005;43(2):745–9. pmid:15695674.
- 10. Pammi M, Zhong D, Johnson Y, Revell P, Versalovic J. Polymicrobial bloodstream infections in the neonatal intensive care unit are associated with increased mortality: a case-control study. BMC Infect Dis. 2014;14:390. Epub 20140714. pmid:25022748.
- 11. Zhang L, Liu ZY, Xu YC, Li TS, Yang QW, Wang AX. [The microbiological and clinical analysis of bloodstream infections with identifiable sources]. Zhonghua Nei Ke Za Zhi. 2012;51(5):366–70. pmid:22883335.
- 12. Lin JN, Lai CH, Chen YH, Chang LL, Lu PL, Tsai SS, et al. Characteristics and outcomes of polymicrobial bloodstream infections in the emergency department: A matched case-control study. Acad Emerg Med. 2010;17(10):1072–9. pmid:21040108.
- 13. Zheng C, Zhang S, Chen Q, Zhong L, Huang T, Zhang X, et al. Clinical characteristics and risk factors of polymicrobial Staphylococcus aureus bloodstream infections. Antimicrob Resist Infect Control. 2020;9(1):76. Epub 20200527. pmid:32460851.
- 14.
Clinical and Laboratory Standards Institute (2018) Performance standards for antimicrobial susceptibility testing; twenty-eighth informational supplement CLSI document M100–S20. Clinical and Laboratory Standards Institute W, PA.
- 15. Xu J, Yuan Y, Wang B, Zhang Q, Wang J, Wang S, et al. Microbiological Analysis and Mortality Risk Factors in Patients with Polymicrobial Bloodstream Infections. Infect Drug Resist. 2023;16:3917–27. Epub 20230620. pmid:37361937.
- 16. Bizzarro MJ, Dembry LM, Baltimore RS, Gallagher PG. Matched case-control analysis of polymicrobial bloodstream infection in a neonatal intensive care unit. Infect Control Hosp Epidemiol. 2008;29(10):914–20. pmid:18808341.
- 17. Tsai MH, Chu SM, Hsu JF, Lien R, Huang HR, Chiang MC, et al. Polymicrobial bloodstream infection in neonates: microbiology, clinical characteristics, and risk factors. PLoS One. 2014;9(1):e83082. Epub 20140114. pmid:24454692.
- 18. Nabb DL, Song S, Kluthe KE, Daubert TA, Luedtke BE, Nuxoll AS. Polymicrobial Interactions Induce Multidrug Tolerance in Staphylococcus aureus Through Energy Depletion. Front Microbiol. 2019;10:2803. Epub 20191205. pmid:31866973.
- 19. Little W, Black C, Smith AC. Clinical Implications of Polymicrobial Synergism Effects on Antimicrobial Susceptibility. Pathogens. 2021;10(2). Epub 20210201. pmid:33535562.
- 20. Tay WH, Chong KK, Kline KA. Polymicrobial-Host Interactions during Infection. J Mol Biol. 2016;428(17):3355–71. Epub 20160508. pmid:27170548.
- 21. Sutter D, Stagliano D, Braun L, Williams F, Arnold J, Ottolini M, et al. Polymicrobial bloodstream infection in pediatric patients: risk factors, microbiology, and antimicrobial management. Pediatr Infect Dis J. 2008;27(5):400–5. pmid:18398386.
- 22. Song F, Zhang K, Huang J, Qian Z, Zhou H, Cai J, et al. Clinical Characteristics, Risk Factors, and Outcomes of Patients with Polymicrobial Klebsiella pneumoniae Bloodstream Infections. Biomed Res Int. 2021;2021:6619911. Epub 20210619. pmid:34239928.
- 23. Sancho S, Artero A, Zaragoza R, Camarena JJ, González R, Nogueira JM. Impact of nosocomial polymicrobial bloodstream infections on the outcome in critically ill patients. Eur J Clin Microbiol Infect Dis. 2012;31(8):1791–6. Epub 20111214. pmid:22167257.
- 24. Goldman S, Itshaki O, Shochat T, Gafter-Gvili A, Yahav D, Rubinovitch B, et al. Risk Factors and Outcome of Polymicrobial Bacteremia: A Retrospective Cohort Study. Isr Med Assoc J. 2020;22(5):279–84. pmid:32378818.