Antimicrobial resistance is becoming a global health challenge that threatens to disengage many medical achievements of the past century. Since 1940s when antibiotics were introduced into medicine, the health and well-being of people worldwide has significantly improved. However, after many decades of success with antibiotics, the world is facing a serious threat of bacterial infections and antibiotic resistance which is present in all countries in the world, contributing to the global specter of a “post-antimicrobial era”. Apart from resistance in malaria, other resistance in Klebsiella pneumoniae, E. coli is also highly prevalence in developing countries. Extensive drug-resistant tuberculosis (i.e., resistant to at least 4 of the core anti-tuberculosis drugs) has been found in 105 countries [1]. In the US alone, the annual cost associated with antimicrobial resistance was estimated to be $55 billion [2]. In the face of such problems, in September 2016, the United Nations General Assembly convened a high-level conference to solve the problem of antimicrobial resistance [3].

Among healthcare settings of high-risk infection, intensive care unit (ICU) is considered an “epicenter of infections”. Patients in ICU are vulnerable to infections because they are exposed to a variety of invasive procedures, including intubation, mechanical ventilation, and vascular access. Moreover, some drugs (sedatives, muscle relaxants) commonly used for ICU patients also increase the risk for infection. In a multicenter study in Europe, between 20 and 30% of ICU admissions had nosocomial infections [4, 5]. Another prospective study in 1265 ICUs from 75 countries found that hospital-acquired infection was present in ~50% of ICU patients [6]. Taken together, these studies clearly indicate that ICU patients are at high-risk for hospital acquired infections, and ICU is an ideal setting for studying the pattern of antibiotic resistance.

Although the vast majority of antibiotic resistance burden is in developing and poor countries, the prevalence and patterns of resistance have not been well documented for these countries. Vietnam is a developing country with a population of 90 million with a heavy demand for healthcare. As far as we know, there have been no studies on the prevalence of and risk factors for antibiotic resistance among ICU patients in Vietnam. In this study we sought to identify patterns of antibiotic resistance in patients admitted to ICU of a major hospital in Vietnam. The finding from this study can contribute to the development of strategies for a better use of prophylactic and empiric antibiotic therapy in ICU patients in developing countries.

At the individual patient level, the prevalence of antimicrobial resistance was 93% (n = 153/164), with the majority (87%) being resistant to at least 2 drugs. Drug resistance was commonly found in ceftriaxone (88%), ceftazidime (80%), ciprofloxacin (77%), cefepime (75%), levofloxacin (72%). Resistance to colistin was found in ~2% (n = 2) patients. Approximately 85% (n = 139) were found to be multiple-drug resistance (MDR).

The patterns of antimicrobial resistance in Acinetobacter, Klebsiella, and Pseudomonas formed a network of multiple drugs as depicted in Fig. 1. In this figure, the degree of thickness of the connecting lines represents the strength of association between any two drugs. For instance, Acinetobacter was resistant to ceftazidime, ceftriaxone, cefepime, ertapenem, and ciprofloxacin, and these resistant pathways formed a complex network. Similar complex networks of drug resistance were also observed for Klebsiella and Pseudomonas.

During the follow-up period, 149 patients had died (incidence rate 68%). The risk of death was associated with advancing age (P = 0.01) and shorter length of ICU stay (P = 0.01). On average, the deceased group had a greater APACHE II score than survivors (23.9 vs 22.1), but the difference was not statistically significant (P = 0.08). The risk of mortality among patients with drug resistance was 67% (103/153) compared with the risk among patients without drug resistance (45%; 5/11); however, the difference did not reach a statistical significance (P = 0.19, Fisher’s exact test). Approximately 91% of those who died in the drug resistance group were MDR. The risk of mortality among patients infected with Acinetebacter, P. aeruginosa, and Klebsiella spp. was 67%, 60%, and 55%, respectively, but there was no statistically significant difference between the three organisms.

Drug resistance, including antimicrobial resistance, is a global issue, because it affects people all over the world, particularly people in developing countries. Based on a sample of ICU patients in a major hospital in Vietnam, we have shown that approximately 95% of ventilator-associated pneumonia patients were resistant to antibiotic drugs. More disturbingly, ~85% (n = 139) were multiple-drug resistant. These results deserve some further elaboration.

In this study, we found that the distribution of pathogens was different from previous studies in developing and developed countries. In most Asian countries, common pathogens isolated from ICU patients were P. aeruginosa, Klebsiella spp., E. coli, Enterococcus, and Staphylococcus aureus [7]. Moreover, the Canadian National Intensive Care Unit study found that P. aeruginosa, Staphylococcus aureus, Haemophilus influenzae, Enterococcus spp., Staphylococcus pneumoniae, and K. pneumoniae are the most common isolates [8]. However, in our study, the most common pathogen was Acinetobacter, followed by Klebsiella and P. aeruginosa. Both Acinetobacter and P. aeruginosa demonstrated multidrug resistance to several antibiotics.

A striking finding from this study was the degree of drug resistance among key pathogens. We found a very high rate of resistance (>70%) among Acinetobacter isolates to most antibiotics, except cefoperazone-sulbactam (4.3%) and colistin (1.5%). Among Klebsiella isolates, low resistance was found for amikacin (5.1%), but high for cefoperazone, ceftazidime, ceftriaxone, cefepime, and levofloxacin. Resistance to meropenem, imipenem, gentamycin, ciprofloxacin and levofloxacin among P. aeruginosa was consistently over 80%. This is probably due to the extensive use of third generation of cephalosporins and quinolone antibiotics in ICU patients.

Previous studies found that the opportunistic pathogens Acinetobacter spp., mainly A. baumannii, are commonly found in ICUs. Our study also found that Acinetobacter was the most common pathogen in ICU patients, and this finding has significant clinical implications. This Acinetobacter spp. has shown a remarkable resistance to many antibiotic classes. Acinetobacter infection is also associated with a high risk of mortality, as there are limited treatment options for this infection. In our study, apart from colistin, Acinetobacter was resistant to virtually all antibiotics, including imipenem (93%), meropenem (90%), ertapenem (100%), ceftriaxone (95%), ceftazidime (93%), cefepime (94%), piperacillin (95%), and ciprofloxacin (95%). Therefore, Acinetobacter is increasingly recognized as an important cause of hospital-acquired infection [9, 10], and our finding confirms the significance of this species as a leading cause of MDR infection in critically ill patients with ventilator-associated pneumonia.

Klebsiella is considered an opportunistic pathogen, and Klebsiella infection is commonly found in hospitalized patients with conditions such as diabetes and chronic pulmonary obstruction [11]. In human, K. pneumoniae, the most important species of the genus, was identified as a pulmonary pathogen more than 100 years ago. While K. pneumoniae has been declining in the United States [12], it is still common in China [13], particularly in ventilator-associated pneumonia patients [14]. In this study, we found that the pathogen accounted for ~22% of all infected patients, and the pathogen is also common in patients with diabetes and pulmonary disorders. In Southeast Asia, among K. pneumoniae isolates, the resistance rate to ciprofloxacin was 62% in Philippines, 29% in Thailand, and 22% in Singapore [15]. In our study, the rate of resistance to ciprofloxacin was 53%. It is not clear why there exists geographic differences in the rates of resistance between countries, but the interaction between environmental reservoir and host variables may be a contributory factor.

In this study P. aeruginosa was the third most common pathogen in ventilator-associated pneumonia. Previous studies have also observed that P. aeruginosa is a prominent cause of pneumonia in hospital setting in Southeast Asia. In previous studies, ciprofloxacin resistance rates among Pseudomonas spp. were over 10% in Malaysia and Singapore [15], which was much lower than our study’s rate (80%). Although the infection may be due to reservoirs within the hospital or ICU environment, selective antimicrobial pressure and compromise of the respiratory tract can also be a risk factor.

Our findings have important clinical implications in the treatment and management of ICU patients, particularly those with ventilator-associated pneumonia. First, clinicians should realize that there is a high possibility that ventilator-associated pneumonia patients are infected with the three common pathogens, and that multiple drug resistance is a reality. Second, the high rate of multidrug resistance observed in this study is a serious concern in the management of ICU patients. It calls for a more systematic approach to reduce antibiotic resistance rates, and minimizing the use of broad-spectrum antibiotics. Third, in the presence of multidrug resistance, the development of rapid diagnostic test for prompt targeted therapy is an important priority. There is also a need for implementing a drug monitoring system that optimizes drug administration and enable a more personalized approach to treatment.

The present study’s findings nevertheless should be interpreted within context of strengths and weaknesses. The study was conducted in a well characterized cohort of patients, who had been thoroughly followed. We have been able to analyze a whole spectrum of pathogens and antibiotic treatments, which allow a relatively comprehensive documentation of antimicrobial resistance in ICU patients. However, we did not have the capacity to conduct a 16S–rRNA gene sequence analysis to identify bacterial species, and this is a significant weakness of the study. Moreover, we did not undertake a phenotypic test that would provide an insight into the mechanism of drug resistance. The study was based on a tertiary hospital and the data were limited to ICU patients; thus the findings may not represent the community acquired infection in Vietnam. The sample size for the study was modest, and did not have adequate power to detect smaller effect sizes or rare incidence. We did not ascertain the specific causes of infection and comorbidities. We also did not ascertain all aspects of care that may have resulted in the prescription of inappropriate antibiotics in our patients.

In this first hospital based study in Vietnam, we observed that ICU-acquired Acinetobacter, Klebsiella and P. aeruginosa predominate ventilator-associated pneumonia patients. More disturbingly, ~85% were multiple-drug resistant. The finding here reinforces the view that multidrug resistance is a global public health issue, and emphasizes the need to study combined therapies and rational treatment strategies.