This study found no significant correlation between the type of antibiotic administered and the occurrence of resistance. In a study in Saudi Arabia in 2023, cefazolin, imipenem, amikacin, and cotrimoxazole could increase the risk of developing resistance.⁶ In a systematic review in China, carbapenem and cephalosporin were found to increase the risk of antibiotic resistance.¹² The disparity in results may be attributed to variations in antibiotic resistance trends, as cefazolin, imipenem, amikacin, and cotrimoxazole were infrequently prescribed at Dr Soetomo Academic General Hospital during the study period.⁶ Ceftriaxone was the first empirical therapy choice for systemic infection in critical condition with normal renal function in Dr. Soetomo Academic General Hospital with cefotaxime as its alternative and since empirical usage is the dominant antibiotic indication in this study, ceftriaxone was the most used antibiotics.^13,14 Metronidazole was used for suspected anaerobic infection and commonly used in combination with ceftriaxone for suspected infection that originated from intraabdominal source or complicated skin infection such as necrotizing fasciitis.^13,14 Ceftazidime was the first choice for suspected Pseudomonas species infection, while Cefoperazone-sulbactam used as alternative for severe cases since the local data in Dr Soetomo Academic General hospital showed good susceptibility for it.^13,14 For patients with impaired renal function where sulbactam was avoided, cefoperazone was used. Fluoroquinolone such as levofloxacin and moxifloxacin were an alternative for infection which suspected originated from respiratory tract or some condition where ceftriaxone usage did not show any improvement.^13,14,42 Cefazolin was limited to be used only as surgical prophylaxis, with ceftriaxone can be used as alternative if cefazolin wasn’t available.^13,14 Ampicillin sulbactam single drug was the first choice for urosepsis empirical therapy accordance to local data.¹³ Ampicillin sulbactam in combination with gentamicin was commonly used for alternative in intraabdominal infection.¹³ Other antibiotics like Vancomycin, Amikacin, Meropenem, Tigecycline, Linezolid, and Cotrimoxazole injection were kept as reserve antibiotics and only used as definitive therapy with prior antibiotic susceptibility result.^13,14 However, in certain cases, these antibiotics may be used as empirical therapy, for example, Vancomycin is allowed to be used in a suspected MRSA infection.¹³

There was also difference in the species of the bacteria found from the culture where as in China the dominant species were Klebsiella pneumoniae (38.6%) and Pseudomonas aeruginosa (34.1%), in Saudi Arabia Klebsiella pneumoniae (21.5%) and Staphylococcus aureus (14.8%), while in this study are Escherichia coli (24.7%) and Klebsiella pneumoniae (19.6%).^6,12 The dominant type of resistance also might play a role in this difference since the most common resistance type found in this study is ESBL-E which is similar to previous study in Indonesia while carbapenem-resistant more often found in the study in Saudi Arabia.

There was no association between antibiotic dose or frequency with the onset of resistance. In drug administration, particularly antibiotics, alterations in the patient’s metabolic and excretion parameters typically necessitate dose modifications to mitigate toxicity and optimize circulating drug concentration.¹¹ A systematic study in 2021 revealed that patients with renal impairment receiving a modified dose of meropenem exhibited blood levels of the drug ranging from 158% to 286% of the normal dose observed in individuals without renal impairment.⁴³ A separate investigation into time-dependent antibiotics, specifically beta-lactams, revealed that the duration for which the concentration of beta-lactam antibiotics exceeded the MIC (fT > MIC) did not significantly correlate with the emergence of resistance; however, a ƒAUC/MIC greater than 494 could diminish the likelihood of resistance.⁴⁴ This prove that adjusted regiment is no inferior than standard regiment in the correct situation, and during this study all the antibiotic dosing and frequencies was in accordance to the guideline used in Dr Soetomo Academic Hospital.^15-37

Based on the 2022, the number of antibiotic prescriptions >3 may increase the likelihood of resistance through prescribing errors due to human error.⁸ Since most of the antibiotic combination found in this study <3 (DR group: 98.26%, non-DR group: 96.72%), there is no significant correlation between the antibiotic combination and antimicrobial resistance.

The duration of antibiotic exposure in this study did not have a significant relationship. In another study conducted to see the increase in beta-lactam MIC on Gram-negative bacteria, the first 7 days of exposure did not show a significant effect on the increase in MIC.⁴⁴ In another study observing the effect of antibiotic exposure to Staphylococcus aureus, an increase in MIC was observed on day 8.⁴⁵ In this study, most of the antibiotics were prescribed <7 days (DR group: 65.70%; non-DR group: 69.95%), hence the bacteria did not have enough time to develop resistance. As for the total number and duration of therapy, since both DR and non-DR group using the same guideline for treatment, no significant difference between the two group and the bacterial resistance in bloodstream infection.^14,46

The use of antibiotic prophylaxis inherently carries a risk of exacerbating bacterial resistance; but, under some circumstances, prophylactic medication may effectively prevent infections and enhance patient outcomes.^38-40,47 The most well-defined antimicrobial prophylaxis use is the surgical prophylaxis which is given 30-60 minutes before surgery for up to 24 hours (may extend to 48 hours following cardiothoracic surgery) and proven can reduce rate of surgical site infection by 80%.³⁹ Empirical and definitive therapy is given to treat the infection that already happen for patients with the high likelihood of infection, or the outcome is likely to be affected adversely by delayed therapy, it is appropriate and prudent for hospitals to develop systems in which patients are expeditiously recognized and promptly treated with an antimicrobial regimen that is broad enough to cover all plausibly likely pathogens.⁴⁸

This study revealed that non-surgical prophylaxis significantly elevates the likelihood of antibiotic resistance in bloodstream infections (adjusted OR = 9.187; CI 95% = 1.9-44.37; p = 0.006). In total, there were 32 antibiotics prescribed as prophylaxis, and 27 of them were prophylaxis in non-surgical cases. We analysed 27 non-surgical prophylaxis treatment prescribed in this study and found: 14 (51.85%) of them were prophylaxis in high grade burn patients, 8 (29.63%) in immunocompromised patients either due to cancer or blood disorders, 5 (18.52%) in trauma patients with open wound.

There were 18 burn patients found in this study, and all of them fulfil the criteria of burn patient’s referral in accordance to The American Burn Association recommendation by being partial thickness burn >10% total body surface area.⁴⁹ In burn patient, there were changes in pharmacokinetics since there will be an increase of renal blood flow and loss of albumin through the wound resulting in increased elimination and alteration of drug concentration caused by the loss of albumin.⁵⁰ Cefuroxime, a second-generation cephalosporin, was the drug of choice for prophylactic antibiotic for burn patients in Dr. Soetomo Academic Hospital since it has relatively good effectiveness, much cheaper and low protein binding (16-33%).^51,52 However, the alteration of pharmacokinetics in burn patients varies from case to case therefore individualized dose adjustment is needed.⁵⁰ In 2010, a systematic review found that the use of prophylactic antibiotic in burn patients may reduce 50% in all cause of mortality.⁵³ However, in other research in 2020, antibiotic usage is not recommended in all grade burn patients unless they have pneumonia or inhalation injury.⁴¹ The burn patients which categorized in the non-surgical prophylaxis has neither sign of pneumonia or inhalation injury and none of the antibiotics prescribed has their dose adjusted. These could be the reasons why the resistance occur, however only small number of burn patients found in this study to reach that conclusion.

The usage of prophylactic antibiotic in the immune compromised group is not contraindicated, but must be reserved for those who have high risk of infection i.e. absolute neutrophil count ≤100 cells/mm³ for >7 days.⁴⁷ In this non-surgical prophylaxis group, the immune suppressed patients were those who have malignancy since no HIV, diabetes mellitus or other type of immune suppression found in this specific category. Only 2 out of 8 prescription is given in the high-risk situation, and 7 out of 8 developed bacterial resistance in the bloodstream infection. For the open wound trauma, in fact antibiotic prophylaxis is highly recommended, however prolonged administration (7-10 days) did not have much benefit for the infection prevention.⁴⁰ In our study, 3 of the prescription were given >7 days and resistance developed for all the patients, while the other 2 which prescribed for <7 days, only 1 of them developed resistance.

Other factors found significant in this study is the use of endotracheal tube (ETT) and immune suppression medication. ETT in the trachea disrupts the natural defence mechanism of physiological mucus clearance and results in the formation of biofilm and mucus on the intraluminal surface of the ETT.⁵⁴ Both suctioning and the inspiratory flow through the ETT may create enough shear force to detach biofilm fragments and reach the lower respiratory tract.⁵⁴ Infections involving a biofilm component are often chronic and highly recalcitrant to antibiotic therapy because of intrinsic physical factors including extracellular matrix production, low growth rates, altered antibiotic target production and efficient exchange of resistance genes.⁵⁵ In this study, it was found that 68,22% of the patient with drug-resistant bloodstream infection which was intubated also experienced pneumonia and bloodstream infection possibly resulted secondary from the respiratory tract infection.⁵⁶

Previous studies states that the dysfunctional immune system poses challenges to the effectiveness of antibiotic therapy, increasing the likelihood of treatment failure since antibiotics cooperate with host responses to clear infections and antibiotic failure would increase the incidence of resistance.⁵⁷ The state of immunosuppression may vary from primary immunodeficiency that is caused by genetic disorder or secondary which was acquired such as diabetes mellitus, cancer, AIDS, autoimmune disease, malnutrition, geriatric, or intervention such as surgeries and immune suppression medication.⁵⁷ In type 2 diabetes mellitus for example, there were alteration in absorption due to decrease of 60-70% mucosal blood flow.⁵⁸ Even if bypassing the first pass effect in patients with diabetes mellitus, high blood glucose will cause glycosylation of albumin resulting in increase of unbound antibiotic which in return increase the clearance of antibiotic leading to suboptimal concentration in plasma.⁵⁸ As discussed before, in burn patient, there was an increase of renal blood flow and loss of albumin through the wound resulting in increased elimination and alteration of drug concentration caused by the loss of albumin.⁵⁰ Other immunosuppression state could cause chemotactic and phagocytic defects in neutrophils, and lymphocyte dysfunction, significantly increase the host’s susceptibility to certain infections and reduce the effectiveness of antibiotic treatment.⁵⁷ There were no AIDS patient found during the data collection. In this study, all the patients were critically ill but no significant correlation found between comorbidities that could suppress immune system like old age, diabetes mellitus, malignancy, liver disease, kidney disease, surgery, or major burn.

However, the continuous use of immune suppression medication correlates strongly to the development of resistance (p = 0.008; adjusted OR 2.709; CI 95%: 1.30-5.65). Among those who used immune suppression medication, 2 patients from the drug-resistant group and 1 patient from control group used IVIG due to Guillain-Barrי syndrome. One patient from the drug-resistant group were a Systemic Lupus Erythematosus patient that use long term methylprednisolone. No other autoimmune disease was found, and the other patients were given corticosteroid treatment in form of continuous pump of methylprednisolone or prednisone to alleviate their inflammation for more than 5 days. Corticosteroids produce their effect through multiple pathways, the glucocorticoid receptor is located intracellularly within the cytoplasm and, upon binding, trans-locates rapidly into the nucleus, where it affects gene transcription and causes inhibition of gene expression and translation for inflammatory leukocytes leads to a reduction in proinflammatory cytokines, chemokines, cell adhesion molecules, and other enzymes involved in the inflammatory response.⁵⁹ The resistance found in the drug-resistant group were ESBL (80.0%), CRE (13.33%), and XDR (6.67%). The duration between hospital administration and the development of resistance among the immunosuppressant users varied from 3 days to 41 days (mean 12,5 days) so it is unclear if it will hasten bacterial resistance. The length of stay in the immune suppression medication group averaged 24,6 days (varied from 5-53 days) and 15,8 days (varied from 4-69 days) for those who didn’t get the medication while the mortality rate does not show significant difference being 22.72% (5 death in 22 patients) for the those who get the immune suppressant and 25.58% (44 death in 172 patients) for those who are not (p = 0.772) but it might be bias due to unbalanced amount between the group. The effect of phospholipase A2 inhibition, which is critical for producing inflammatory cytokines, impairing release of arachidonic acid, regulation of apoptosis in thymocytes and inhibition of B cells and T cells production caused by steroid therapy would render the host susceptible to infection.⁵⁹ A glucocorticoid such as dexamethasone was reported having the capability to induce CYP3A4 and CYP2C9 which would alter the concentration antibiotics that metabolized by them.⁶⁰ Another report showed that hormonal steroid therapy may induce drug resistant by increasing the expression of multidrug efflux pump operon mtrCDE through inhibition of regulator MtrR.⁶¹ Further research might be needed to conclude whether continuous immunosuppressant medication really promote the development of resistance different from other states of immune suppression especially the expression of ESBL in Enterobacterales.

This study has limitations, notably that the evaluation of blood cultures was conducted irregularly, contingent upon alterations in the patient’s state, making it challenging to determine the onset of infection by drug-resistant bacteria. This study exclusively assessed bloodstream infections, thereby precluding the evaluation of bacterial resistance development at the local wound or organ site. This study was conducted at a single hospital in Indonesia; hence it scarcely represents the broader community situation. The variability in the quantity of resistances present complicates the analysis of each resistance type. This study did not analyse additional characteristics such as visitor count, infection control protocols, and caregiver involvement. Comprehensive research on a wider scale is required to substantiate the danger associated with antibiotic administration and other factors concerning the occurrence of bacterial resistance, particularly in specific populations such as burn victims, immunocompromised individuals, or trauma patients.