Introduction

Blood stream infection (BSI) is a leading cause of mortality and morbidity worldwide. The pathogen in the blood, quickly spreads to various organs in the body, resulting in septic shock and organ dysfunction. This makes it a formidable medical emergency, requiring prompt diagnosis and treatment. The diagnosis and management of the BSIs can be very challenging, as it requires high degree of suspicion, prompt investigations in the form of blood cultures and other laboratory parameters and immediate administration of appropriate empirical antibiotics followed by continuous assessment with escalation or de-escalation of the antibiotics depending upon the patient’s clinical response and culture reports. The prognosis of the patient also depends on finding the source of the infection and appropriate source control of the infection focus. (1) As a result, the overall crude mortality rates of BSIs are high and may vary between 15% and 30%. BSIs also have considerable effect on the admission to intensive care units (ICUs), length of hospital stay and cost of health care. The problem gets further complicated by antibiotic resistance. This makes bloodstream infection a very important burden to the health care systems. (2)

The burden of BSIs in the Lower- and Middle-income countries (LMIC) is higher than the developed countries, especially due to higher proportions of antimicrobial resistance (AMR). Unfortunately, there is a scarcity of studies which describe the epidemiology and outcomes of the patients with blood stream infections from these regions. (2) There is an urgent need for the knowledge on the causative agents and outcomes of the blood stream infections from different parts of India. It is also equally important to monitor the antibiotic resistance of the pathogens causing BSIs and their epidemiology. This will ensure improvement of the health care systems and thus better healthcare service will be provided to community. (3) Gram negative bacterial pathogens are shown to be common cause of blood stream infections, with high level of resistance against the common empiric antibiotics. (4–7) In this study we try to fill the gaps in the knowledge and present the epidemiology of the Gram-negative bacterial blood stream infection (GNBSI), prevalence of carbapenem resistance, patient management and outcome in an Indian tertiary care hospital.

Methodology

This retrospective study was conducted, after obtaining the ethical clearance from the institute ethics committee. (FMMC/FMIEC/455/2022) The patients ≥ 18 years of age, with Gram negative bacteria (GNB) growth in blood culture, during the study period August 2020 to May 2022 were included in the study. The selection of the patients for the study is described in the Figure 1.

Figure 1: Flowchart of the study population

Only monomicrobial Gram negative bacterial blood stream infections (GNBSI) were included. Patients with ≥2 bacteria in blood were excluded from the study. Patients with ≥2 episodes of BSIs with different bacteria were also excluded from the study.

Definitions

GNBSI was defined as at least one blood culture positive for growth of GNB and any one or more of following features of blood stream infections like fever, hypothermia, tachypnoea, altered mental status, multiple organ dysfunction syndrome, total leucocyte count >11.0 ×10⁹/L. Source of the GNBSI was determined microbiologically by the documented isolation of the same GNB from another site of infection by culture of the corresponding sample.(8) The electronic case records were screened for the presence of co-morbidities. The electronic case records were evaluated for the documented administration of antibiotics, ceftriaxone, piperacillin-tazobactam, meropenem, tigecycline and polymyxin B during the hospital stay. The event of intensive care unit (ICU) admission, central venous catheter (CVC) insertion, mechanical ventilation and vasopressor administration anytime during the hospital stay were noted. The crude in hospital mortality was noted from the electronic case records. The death at 7 days or 30 days was calculated from the day of collection of the positive blood culture sample. (2)

Microbiology

BACTEC^TM aerobic system (Becton Dickinson, Maryland, USA) was used to process blood cultures. The positive blood cultures were processed further by Gram stain and subcultures on 5% sheep blood agar (HiMedia Laboratories Pvt. Ltd, Mumbai, India) and MacConkey agar (HiMedia Laboratories Pvt. Ltd, Mumbai, India). The growth was identified by the standard biochemical methods and appropriate antibiotic susceptibility testing was done by modified Kirby Bauer method, using the appropriate antibiotic discs (HiMedia Laboratories Pvt. Ltd, Mumbai, India). Reporting of susceptibility was done according to the Clinical Laboratory Standards Institute (CLSI) guidelines.(9) Ampicillin 10 µg, amoxicillin clavulanic acid 20/10 µg, cefuroxime 30 µg, ceftriaxone 30 µg, cefotaxime 30 µg, piperacillin tazobactam 100/10 µg, imipenem 10 µg, meropenem 10 µg, amikacin 30 µg, co-trimoxazole 1.25/23.75 µg and ciprofloxacin 5µg discs were used for antibiotic susceptibility testing. A GNB was considered to be carbapenem resistant if it was resistant to either meropenem or imipenem, on antibiotic susceptibility testing. The analysis of antibiogram was done using WHONET software.

Statistical analysis

The data was processed for statistical analysis using the IBM SPSS Statistics (version 23). The continuous variables were checked for normality of distribution for each subset. The continuous variables with normal distribution were expressed as mean with standard deviation (SD), whereas the continuous variables with non-normal distribution were expressed as median with interquartile range (IQR). The categorical variables were expressed as absolute counts and frequency percentages.

A comparison was made between the carbapenem susceptible and carbapenem resistant GNBSI by Chi square test, Fisher’s Exact test, Student t test or Mann-Whitney U test, as applicable according to the type of variable. (table 1)

The survival analysis was performed using the in-hospital crude mortality rates at days 7 and 30, from the day of collecting the blood culture sample positive for the GNB. The length of stay after the positive GNBSI was used for the calculation of hazard ratio (HR), 95% confidence interval by the univariate analysis of the possible independent variables by Cox proportional hazards model. The variables with P values < 0.1 for HR, during the univariate analysis were used for calculating the adjusted hazard ratio (aHR) with 95% confidence interval (CI) by multivariate survival analysis.

A similar mortality analysis at days 7 and 30 was performed among the subset of patients who had an event of ICU admission during the hospital stay.

Results

During the study period of 22 months, from August 2020 to May 2022, monomicrobial GNBSIs were detected in 375 patients. Among the patients, 204 (54.4%) were males, median age was 60 (IQR 48-69) years. The GNBSIs were caused commonly by Escherichia coli, 105 (28.0%), Klebsiella pneumoniae, 100 (26.7%) and Acinetobacter species, 65 (17.3%), as shown in figure 2.

Figure 2: Bacterial epidemiology of monomicrobial GNBSIs (n=375) Abbreviations: GNBSI, Gram negative bacterial blood stream infections

As shown in Figure 3, Acinetobacter species was the most resistant common pathogen, followed by Klebsiella pneumoniae compared to Escherichia coli.

Figure 3: Antibiogram of the common pathogens causing GNBSIs

The characteristics of carbapenem susceptible and carbapenem resistant Gram-negative bacterial blood stream infections (CR-GNBSI) are shown in Table 1. The total duration of hospital stay was longer among patients with CR-GNBSIs (p value, 0.000). The length of hospital stay after the positive blood culture was not different between the 2 groups. CR-GNBSIs were associated with COVID-19 infection (p value, 0.000) and use of immunosuppressive drugs (p value, 0.005). CR-GNBSIs were associated more with the lung infections as source (p value, 0.000). The CR-GNBSIs were associated with associated with administration of the higher end antibiotics like meropenem, tigecycline and polymyxin B (p value, 0.000). They were also associated with the insertion of central venous catheter, mechanical ventilation, vasopressor administration and ICU admission (p value, 0.000). There was an association between mortality and the CR-GNBSIs (p value, 0.000).

Mortality analyses

Overall, in-hospital crude mortality among the patients with monomicrobial GNBSI in our study was 40.3%. Mortality at 7 days and 30 days were 33.9% and 40.3% respectively. As shown in table 2, ICU admission, vasopressor use, immunosuppressive drug administration were independent risk factors for mortality.

Table 3 describes a sub analysis for mortality, among the 232 patients who had event of ICU admission during hospital stay. When compared to patients with carbapenem susceptible-GNBSI (CS-GNBSI), carbapenem resistant Acinetobacter species- GNBSI (CRA-GNBSI) and carbapenem resistant Klebsiella pneumoniae GNBSI (CRKP-GNBSI) were independent risk factors for mortality at 7 days and 30 days. Vasopressor administration was an independent risk factor for mortality at both 7 days and 30 days.

Discussion

Blood stream infections are medical emergencies, requiring immediate diagnosis and treatment and continuous monitoring of the patient for the response to treatment.

Escherichia coli was the most common pathogen, followed by Klebsiella pneumoniae and Acinetobacter species in the present study. Segala et al, in a study in Italy, also reported Escherichia coli, as the most common GNB pathogen, followed by Klebsiella pneumoniae. (10) Klebsiella pneumoniae and Acinetobacter baumanii were the most common pathogens that were resistant to carbapenems, in the current study. Similarly, in a study at Delhi, by P et al, the most common difficult to treat GNBSI were due to Klebsiella pneumoniae and Acinetobacter baumanii. (11)

We recorded in hospital crude mortality of 40.8% among the patients with GNBSI during the study period. Similar mortality rates have been reported in many studies all over the world and India, during the period of COVID-19 pandemic. Unterberg et al, reported an increase in hospital crude mortality rate from 33% to 52% among the patients with sepsis during the pandemic period. He demonstrated a high initial sequential organ failure assessment (SOFA) score at admission among the patients during the pandemic, indicating delayed presentations of the patients to the hospital. (12) In an Indian study, during the same time, P et al, recorded an overall mortality rate of 50.8% among the patients with GNBSI episodes. (11) This high mortality could be due to late presentation of patient to the hospital, in critical stage with comorbidities and infection with multidrug resistant (MDR) bacteria leading to failure of treatment.

On multivariate mortality analysis, we found that ICU admission was the strongest independent risk factor associated with mortality at both 7 days (aHR,6.714, p value, 0.000, 95% CI, 2.824-15.961) and 30 days (aHR, 8.300, p value- 0.000, 95% CI,3.383-20.362). Use of vasopressors was also independent risk factor for the mortality at days 7(aHR, 2.496, p value, 0.001, 95% CI, 1.495-4.167) and 30 (aHR, 2.569, p value, 0.000, 95% CI,1.612-4.093), in all the patients (table 2) and the subset of patients with event of ICU admission (Table 3). These findings indicate that the mortality was highest among the patients who were critically ill. These findings are consistent with similar studies, where the patients with critical conditions in ICU, indicated by high Pitt score, Charlson comorbidity index, SOFA score or APACHE score at the time of BSI were largest risk factor among patients with sepsis. (8,12–18)

We found that the use of immunosuppressive drugs was associated with higher mortality among the patients at 30 days (aHR, 1.706, p value, 0.019, 95% CI, 1.092-2.665) after the episode of GNBSI. Interestingly we found that fever was present in only 68.0% of the patients with episode of GNBSI. (Table 1) Also, it was a protective factor against mortality, on univariate analysis. (Supplementary Table 1) The protective effect of fever in patients with sepsis has been documented by Rumbus et al and others. (19,20) The decreased rate of fever in GNBSIs may be due to the increased administration of the immunosuppressive drugs during the pandemic, which led to reduced fever and inflammatory markers, making it difficult to objectively assess the response to therapy, as reported by Segala et al.(10) This is also demonstrated in our study, where we find that Covid 19 infection was a risk factor for mortality on univariate analysis, but it was not an independent risk factor, during multivariate analysis. (Table 2 and supplementary Table 1). Palanisamy et al, reported 64 patients with Covid-19 infections in ICU who developed BSI. All the patients were on corticosteroids and succumbed to the BSI. (17)

CR-GNBSI was detected in 36.53% of the patients, in the present study. There are few studies from India, presenting the epidemiology of carbapenem resistance in GNBSIs. In one such study during the similar period, Naveenraj P et al reported 111/240 cases of GNBSIs to be caused by carbapenem resistant bacteria. (11) Kaur et al, in a multicentric longitudinal study from 2017 to 2022, reports that there has been significant increase in resistance to carbapenem resistance among cases of GNBSIs.(21) We find in our study that carbapenem resistance was not an independent risk factor for mortality, among all the patients during the study period (Table 2), we found that, among the subset of patients with ICU admission, CRA-GNBSI and CRKP GNBSIs were independent risk factors for mortality at 7 and 30 days, when, compared to patients with CS-GNBSIs.(Table 2) Palanisamy et al also has reported Acinetobacter baumannii as the most common GNB, followed by Klebsiella pneumoniae and non-fermenters causing BSIs among the COVID-19 patients in ICU. (17) Most of the countries worldwide have reported increased hospital acquired infections and AMR during the pandemic, due to the unprecedented workload, beyond the capacity of the healthcare and laboratory systems. There was a breakdown of HAI surveillance and breach in the infection control practices. (10,11,15,17,22–29)

It is important to note that in 65.9% of GNBSI cases in our study, microbiological confirmation of the source was not available. Though microbiological source determination may not be possible in every case of GNBSI, it would certainly help to effectively treat the infection and reduce the AMR rates. (1,2,8) The culture utilization rates need to be studied in context of empirical antimicrobial prescription and development of AMR. Lim et al, in a study, recommends including blood culture utilization rates in addition to AMR surveillance reports. (30) There is a lack of studies regarding this from India. Also, there is a risk of overestimating the severity of BSIs and over-reporting of MDR-BSIs, if the blood cultures are not prescribed timely. The lack of culture testing may also promote AMR due to prolonged use of inappropriate empirical antibiotics. The cost and mortality differences need to be studied with respect to the utility of microbiological cultures for infections. (2)

This study has relatively large sample size of 375 consecutive monomicrobial GNBSIs among the patients admitted to the hospital. The study provides valuable insights into the problems associated with GNBSIs, especially during the pandemic. The study identifies the risk of immunosuppressive drugs associated with mortality.

The study has some limitations. The study is retrospective and may be influenced by the errors of data documentation. We were not able to collect the data on source control and appropriateness of empirical antibiotic therapy, which may influence the mortality among the patients. Since the study was conducted in a single hospital, it may not fully reflect the situation in other healthcare settings. The blood cultures were sent according to the discretion of the clinicians, which if inadequate, may have resulted in overestimation of severe GNBSIs and the rates of CR-GNBSIs.

Conclusion

Our study explores the epidemiology and clinical outcomes of 375 monomicrobial GNBSIs among the patients admitted in an Indian tertiary care hospital from August 2020 to May 2022. Escherichia coli, Klebsiella pneumoniae and Acinetobacter species were the most common bacteria causing GNBSIs. Carbapenem resistance was observed in 36.56% of the isolates. Overall mortality was 40.8% among the patients with GNBSIs. Events of ICU admission and vasopressor administration were independent risk factors for death. Use of immunosuppressive drugs was an independent risk factor for mortality. Among the patients with ICU admission, carbapenem resistant Acinetobacter species GNBSIs and carbapenem resistant Klebsiella pneumoniae GNBSIs were independent risk factors for mortality. Our study indicates that there is a need for monitoring the carbapenem resistance and measures need to be taken to reduce it, as there are very few antibiotics available for the treatment of carbapenem resistant pathogens. Similarly, measures also need to be taken for early identification and effective treatment of GNBSIs to facilitate better outcomes. Hence there is need to empower the antibiotic stewardship practices in the healthcare systems.

References

1. Aslan AT, Tabah A, Köylü B et al. Epidemiology and risk factors of 28-day mortality of hospital-acquired bloodstream infection in Turkish intensive care units: a prospective observational cohort study. J Antimicrob Chemother. 2023;78(7):1757–68.
2. Allel K, Stone J, Undurraga EA et al. The impact of inpatient bloodstream infections caused by antibiotic-resistant bacteria in low- and middle-income countries: A systematic review and meta-analysis. PLOS Med. 2003;20(6):e1004199.
3. Zeng Q, Xiang B, Liu Z. Profile and Antibiotic Pattern of Blood Stream Infections of Patients Receiving Hematopoietic Stem Cell Transplants in Southwest China. Infect Drug Resist 2022;15:2045–54.
4. Singh AK, Venkatesh V, Singh RP, Singh M. Bacterial and antimicrobial resistance profile of bloodstream infections: A hospital-based study. CHRISMED Journal of Health and Research. 2014;1(3):140-4.
5. Prabhash K, Medhekar A, Ghadyalpatil N et al. Blood stream infections in cancer patients: A single center experience of isolates and sensitivity pattern. Indian J Cancer. 2010;47(2):184.
6. Thacker N, Pereira N, Banavali S et al. Epidemiology of blood stream infections in pediatric patients at a Tertiary Care Cancer Centre. Indian J Cancer. 2014;51(4):438.
7. Wattal C, Raveendran R, Goel N, Oberoi JK, Rao BK. Ecology of blood stream infection and antibiotic resistance in intensive care unit at a tertiary care hospital in North India. Braz J Infect Dis. 2014;18(3):245–51.
8. Santoro A, Franceschini E, Meschiari M et al. Epidemiology and Risk Factors Associated With Mortality in Consecutive Patients With Bacterial Bloodstream Infection: Impact of MDR and XDR Bacteria. Open Forum Infect Dis. 2020;7(11): ofaa461.
9. CLSI. Performance Standards for Antimicrobial Susceptibility Testing. 33rd ed.
10. Segala FV, Pafundi PC, Masciocchi C et al. Incidence of bloodstream infections due to multidrug-resistant pathogens in ordinary wards and intensive care units before and during the COVID-19 pandemic: a real-life, retrospective observational study. Infection. 2023;51(4):1061–9.
11. Naveenraj P, Kumar D, Meena DS et al. Difficult-to-treat resistant gram-negative blood stream infections - the beginning of a superbug era - a prospective observational study. Indian J Med Microbiol. 2023;44:100364.
12. Unterberg M, Rahmel T, Rump K et al. The impact of the COVID-19 pandemic on non-COVID induced sepsis survival. BMC Anesthesiol. 2022;22(1):12.
13. Bass SN, Bauer SR, Neuner EA, Lam SW. Impact of Combination Antimicrobial Therapy on Mortality Risk for Critically Ill Patients with Carbapenem-Resistant Bacteremia. Antimicrob Agents Chemother. 2015;59(7):3748–53.
14. Wang J, Zhang J, Wu Z-H et al. Clinical Characteristics and Prognosis Analysis of Acinetobacter baumannii Bloodstream Infection Based on Propensity Matching. Infect Drug Resist. 2022;15:6963–74.
15. Mantzarlis K, Deskata K, Papaspyrou D et al. Incidence and Risk Factors for Blood Stream Infection in Mechanically Ventilated COVID-19 Patients. Antibiotics. 2022;11(8):1053.
16. Vidaur L, Eguibar I, Olazabal A et al. Impact of antimicrobial stewardship in organisms causing nosocomial infection among COVID-19 critically ill adults. Eur J Intern Med. 2024;119:93–8.
17. Palanisamy N, Vihari N, Meena DS, Kumar D, Midha N, Tak V, et al. Clinical profile of bloodstream infections in COVID-19 patients: a retrospective cohort study. BMC Infect Dis. 2021;21(1):933.
18. Lin XC, Li CL, Zhang SY, Yang XF, Jiang M. The Global and Regional Prevalence of Hospital-Acquired Carbapenem-Resistant Klebsiella pneumoniae Infection: A Systematic Review and Meta-analysis. Open Forum Infect Dis. 2023;11(2):ofad649.
19. Rumbus Z, Matics R, Hegyi P et al. Fever is Associated with Reduced, Hypothermia with Increased Mortality in Septic Patients: A Meta-Analysis of Clinical Trials. PLoS One. 2017 Jan 12;12(1):e0170152.
20. Shimazui T, Nakada T-A, Walley KR et al. Significance of body temperature in elderly patients with sepsis. Crit Care. 2020;24(1):387. DOI:10.1186/s13054-020-02976-6
21. Kaur J, Singh H, Sethi T. Emerging trends in antimicrobial resistance in bloodstream infections: multicentric longitudinal study in India (2017–2022). Lancet Reg Health - Southeast Asia. 2024 Jul;26:100412.
22. Cataldo MA, Tetaj N, Selleri M et al. Incidence of bacterial and fungal bloodstream infections in COVID-19 patients in intensive care: An alarming “collateral effect.” J Glob Antimicrob Resist. 2020; 23:290–1.
23. Amarsy R, Trystram D, Cambau E et al. Surging bloodstream infections and antimicrobial resistance during the first wave of COVID–19: a study in a large multihospital institution in the Paris region. Int J Infect Dis. 2022; 114:90–6.
24. Giacobbe DR, Battaglini D, Ball L, Brunetti I, Bruzzone B, Codda G, et al. Bloodstream infections in critically ill patients with COVID-19. Eur J Clin Invest. 2020;50(10):e13319.
25. Polly M, De Almeida BL, Lennon RP, Cortês MF, Costa SF, Guimarães T. Impact of the COVID-19 pandemic on the incidence of multidrug-resistant bacterial infections in an acute care hospital in Brazil. Am J Infect Control. 2022;50(1):32–8.
26. Gajic I, Jovicevic M, Popadic V et al. The emergence of multi-drug-resistant bacteria causing healthcare-associated infections in COVID-19 patients: a retrospective multi-centre study. J Hosp Infect. 2023; 137:1–7.
27. Budhiraja S, Tarai B, Jain D et al. Secondary infections modify the overall course of hospitalized patients with COVID-19: a retrospective study from a network of hospitals across North India. IJID Reg. 2022; 3:44–53.
28. Chowdhary A, Tarai B, Singh A, Sharma A. Multidrug-Resistant Candida auris Infections in Critically Ill Coronavirus Disease Patients, India, April–July 2020. Emerg Infect Dis. 2020;26(11):2694–6.
29. Vijay S, Bansal N, Rao BK et al. Secondary Infections in Hospitalized COVID-19 Patients: Indian Experience. Infect Drug Resist. 2021;14:1893–903.
30. Lim C, Hantrakun V, Teerawattanasook N et al. Impact of low blood culture usage on rates of antimicrobial resistance. J Infect. 2021;82(3):355–62.