Aerosolized Plus Intravenous Polymyxin B Versus Colistin in the Treatment of Pandrug-Resistant Klebsiella Pneumonia-mediated Ventilator-Associated Pneumonia: A Retrospective Cohort Study in Bangladesh

An infection in the lung parenchyma preceded by invasive mechanical ventilation for at least a duration of 48 to 72 h at intensive care unit (ICU) setup is known as ventilator-associated pneumonia (VAP). VAP is one of the most common mechanical ventilator-associated infections in intubated critically ill patients and a difficult-to-treat infection among the nosocomial infections with high rate of mortality worldwide. VAP accounts for up to 29.1% of nosocomial pneumonia [1]. VAP causes hospitalization time longer, amplified treatment cost, and wastage of medical resources [2]. Despite advancements of microbiological techniques, standard diagnosis criteria and treatment of VAP is still a dueling issue which impedes the overall treatment process and abates clinical outcome in critically ill patients [1,2].

According to the clinical guideline of the Infectious Diseases Society of America (IDSA) and the American Thoracic Society (ATS) in 2016, up to 13% of patients die from VAP in each year [3], while another multi-center study showed that the 30-day mortality due to VAP, early VAP and late VAP was 29.9%, 19.2% and 31.4%, respectively in Europe [4]. A meta-analysis of 195 studies found that from 2006 to 2014 in China’s mainland, the cumulative VAP incidence rate was 23.8% with 22.83 patients per 1000 ventilator days [5]. A recent single-center study in Bangladesh found that 36.5% of mechanically ventilated patients in ICU developed VAP of which 78.9% and 21.2% was early and late VAP, respectively [6].

The gram-negative Klebsiella pneumoniae, a member of Enterobacteriaceae family, is one of the most common organisms for nosocomial pneumonia, including VAP around the world. The emergence of K. pneumoniae with pandrug (PDR)-resistant strains, resistant to all currently available antibiotics [7], is a big threat to global human health [8,9]. Polymyxin B and colistin (polymyxin E) are last-line polypeptide antibiotics intended to use for the treatment of multidrug-resistant K. pneumoniae infections, including VAP. Pandrug-resistant K. pneumoniae (PDRKP) strains frequently show resistance to polymyxins at usual doses and at that time, treatment of serious respiratory tract infections, including VAP becomes challenging [8–11]. To date, in-vivo data are limited regarding the pharmacokinetics of polymyxins in lung tissue after its intravenous or aerosolized administration and the susceptibility of PDR KP strains to polymyxins at differential drug-concentration in the lung tissue. Moreover, the paucity of data on polymyxins’ dose-mediated toxicity in patients with VAP develops conflicts regarding therapeutic modification [12,13]. The optimized therapeutic armamentarium with the last-hope polymyxins may be effective against the PDRKP strains causing VAP. A recent study showed that dual-routed (intravenous and inhalation) polymyxin B regimen exhibited better microbial eradication and increased survival rate compared to intravenous polymyxin B therapy in patients with VAP.

The objective of this study is to compare the ability of eradication of PDR K. pneumoniae and clinical outcome of polymyxin B with colistin while administering concomitantly in intravenous and aerosolized form in the treatment of pandrug-resistant K. pneumoniae VAP in critically ill patients in a tertiary care hospital in South Asia.

Of 222 patients, 106 patients (N) and 116 patients (N) were included in PMB and CLN group, respectively. The Number of male patients was higher than that of female patients in both groups (74%/26%: PMB group and 53%/47%: CLN group), and the median age of patients in PMB and CLN group was 56 years (IQR: 44.5-68) and 55 years (IQR: 46-65), respectively. The body mass index value of all patients of the study was from 20 to 24. Comorbidities in all patients were recorded at the time of hospital admission and mentioned in Table 1. Before starting PMB and CLN therapy, the median peripheral capillary oxygen saturation in blood (SpO₂), the median fraction of inspired oxygen (FiO₂), and the median ratio of arterial oxygen partial pressure to fractional inspired oxygen (PaO₂/FiO₂) was 96% (IQR: 93-98)/96% (IQR: 94-98), 50 (IQR: 40-65)/45 (IQR: 40-65), and 240 mmHg (IQR: 193-255)/243 mmHg (IQR: 206.72-256) in patients of PMB and CLN group, respectively (P>0.05). Median values of blood components, infection markers, and liver enzymes in patients of both groups were mentioned and compared in Table 1. The median Modified Early Warning Score (MEWS) score and Acute Physiology and Chronic Health Evaluation (APACHE II) score was 4 (IQR: 3-4)/ 4 (IQR: 3-4) (P=0.032) and 20 (IQR: 20-24)/ 20 (IQR: 18-24) (P>0.05) in patients of PMB (N=106) and CLN group (N=116), respectively.

Patients treated with intravenous plus aerosolized polymyxin B in PMB group showed better microbial eradication (83%, N=106) than the patients treated with intravenous plus aerosolized colistin in CLN group (68.1%, N=116) (P<0.05). Secondary gram-negative bacterial infection within 21-day of detection of first PDR K. pneumoniae VAP infection was privileged more in CLN group than in PMB group, significantly (27.58% vs 22.64%, respectively). The median day of intubation and ICU stay was less in patients of PMB group than in CLN group’s patients [10 days (IQR: 9-12.25)/14 days (IQR: 11-19), respectively, P=0.001; 12 days (IQR: 10-14)/15 days (IQR: 9-18.5), respectively, P=0.072] (Table 2). Drug-induced adverse events were monitored continuously in every patient of both groups till discharged from the hospital. Patients who received aerosolized polymyxin B (PMB group) developed more bronchospasm (36.79%, N=106) compared to patients who received aerosolized colistin (CLN group) (6.89%, N=116) (P=0.001) but the incidence of nephrotoxicity was higher in CLN group compared to PMB group (28.44%/11.32%, respectively; P =0.002) (Table 3). Antibiotic therapy was continued after developing drug-induced bronchospasm in patients of the study and well managed with inhaled levosalbutamol as bronchodilator. Intravenous dose of polymyxin B and colistin was adjusted preceded by antibiotic-induced nephrotoxicity but therapy was not discontinued. Neurotoxicity associated with polymyxins was not found in any patient of the study.

Sixteen deaths were observed in patients of PMB group (15%, N=106) while 25 deaths occurred in the CLN group (21.55%, N=116). The Cox proportional hazard model (adjusted for age, PaO₂/FiO₂ ratio, APACHE II, microbial non-eradication, length of intubation; and length of ICU stay) showed a significantly greater survival rate in PMB group as compared to CLN group (multivariate HR for death: 0.035; 95% CI=0.007-0.168, P<0.001) (Fig. 1). The survival rate did not differ with age, PaO2/FiO2 ratio, and length-of-ICU stay (multivariate HR: 1.028; 95% CI=1.001-1.055, P =0.045, 1.005; 95% CI=0.994-1.015, P>0.05, and 1.043; 95% CI=0.991-1.099, P>0.05, respectively). Non-eradication of PDR K. pneumoniae caused VAP significantly increased the risk of death (multivariate HR: 1.148; 95% CI: 0.484-2.721) in CLN group compared to PMB group. APACHE II and Length of intubation on mechanical ventilation was similar between the two groups, as shown in the Table 4.

Fig. 1.

In this study, compared to colistin, concomitant intravenous and aerosolized polymyxin B exhibited better eradication of PDR K. pneumoniae with a significant reduction in intubation period and minimized the length-of-ICU stay in patients with PDR K. pneumonia VAP. A recent study on Bangladeshi population found that intravenous plus aerosolized polymyxin B eradicated 92.1% of multidrug-resistant (MDR) K. pneumoniae (only susceptible to polymyxin B) causing VAP [14] whereas in this study, the eradication rate of polymyxin B (intravenous plus aerosolization) was 83% while the K. pneumoniae strains were resistant to all available antibiotics including polymyxins.

The emergence of PDR gram-negative bacteria, such as K. pneumoniae, is currently a serious global public health concern [7–9]. Polymyxins are the reserve antibiotics for treating severe infections with MDR pathogens [8,9]. However, therapeutic outcome becomes very worse while K. pneumoniae isolates represent resistance to polymyxins [9–11]. VAP caused by PDR isolates of K. pneumoniae is most likely untreatable with the standard dose of polymyxin B or colistin given intravenously or aerosolized form [9,11]. Variable pharmacokinetic data of PMB and CLN have been demonstrated in different in-vivo studies to date [12,13,18,19]. The use of polymyxin B and colistin as last-line therapy in intravenous and aerosolized form in serious infections with MDR pathogens has been attributed in clinical practice over the last couple of decades [12]. In practice, in severe upper or lower respiratory tract infections, such as VAP, aerosolized polymyxins are currently more popular among the clinicians due to its outweighed therapeutic benefits over intravenous use [13,19,20]. However, clinical data related to the use of polymyxins in the treatment of PDR K. pneumoniae VAP mostly in South Asian population is limited [21].

Colistin is the bioactive form of CMS (administration form) and following an intravenous single dose of CMS, only 20% to 25% of the total CMS is rapidly converted to colistin-base activity (CBA) in-vivo, and the major portion of the drug is excreted through kidney as unchanged from of drug [20,22]. Following an intravenous dose of 150 mg CBA, the maximum plasma concentration (C_{max, Plasma}) of colistin ranges from 0.4 to 0.77 mg/Liter of blood [22]. Imberti et al. [23] found that in critically ill patients, following intravenous 60 mg CBA every 8 h, at the steady-state phase, the formed colistin concentration was undetectable in bronchoalveolar lavage (BAL) fluid, and they also assumed that an increased intravenous dose may attribute a higher concentration of formed colistin in lung epithelial lining fluid (ELF) but nephrotoxicity would be the major concern. According to the findings of Athanassa et al. [24], an aerosolized 80 mg dose of CMS in mechanically ventilated patients resulted in high median CBA concentration (6.7 μg/mL) in ELF (locally converted) for up to 4 h after inhalation which met the MIC breakpoint for K. pneumoniae (<2 μg/mL). However, after 4 h, colistin concentration was declined below the MIC valueand a variety of sub-therapeutic colistin concentrations was found in BAL and ELFof the patients leading to therapeutic failure. A higher dose of inhaled colistin was suggested by the authors of the above study to achieve effective CLN concentration in the lung tissue for the eradication of MDR/PDR K. pneumoniae isolates causing serious RTIs, including VAP [24]. In this study, intravenous plus aerosolized dosage-regimen of colistin (total 400 mg CBA/day) successfully eradicated 68.1% of colistin resistant K. pneumoniae caused VAP and this may be due to high colistin concentration exposure to lung tissue and fluids.

In contrast, polymyxin B is administered as pharmacologically active sulfate salt form [19]. The in-vivo pharmacokinetic/pharmacodynamic data of intravenous or aerosolized polymyxin B is limited and the exact bioavailability of PMB in respiratory tract is difficult to predict. Yu-Wei et al. [25] in their Mouse Lung Infection Model found that after 5 min of an intravenous dose of 4.12 mg base (one milligram of pure polymyxin B base is equivalent to 10,000 units of polymyxin B sulfate)/kg of body weight of PMB, the average free plasma concentration of 1.07 ± 0.14 mg/liter was obtained and polymyxin B concentration in ELF following intravenous administration was below the ELF limit of quantification (LOQ) (~2.00 mg/liter) due to low rate of free-drug transfer from plasma to ELF (CL_{Plasma, ELF} of 1.14 × 10⁻⁴ liters/h/kg) which does not meet the required MIC level (2 μg/mL as per CLSI) in ELF essential for killing PDR K. pneumoniae in lung. On the other hand, following pulmonary administration of 4.12 mg base/kg of body weight of PMB, the maximum achieved ELF concentration (C_{max, ELF}) was 107.0 ± 24.6 mg/liter and the concentration was significant over 12 h meeting the MIC breakpoint of 2.00 mg/liter. Thus, 12-hourly aerosolized plus intravenous dosages of PMB may ensure a significant PMB level in lung tissue and fluids over 24 h above the MIC level, and with this dual dosage-regimen in the study, we found significant eradication of PDR K. pneumoniae with polymyxin B (83%, N=106). In a recent study on South Asian population, intravenous plus aerosolized PMB resulted in higher eradication of MDR K. pneumoniae compared to intravenous PMB (92.1% vs 70.1%; P<0.05) [14]. Likewise, in our study, PMB exhibited superior microbial eradication rather than colistin in similar dual-route administration model (83% vs 68.1%, P<0.05).

K. pneumoniae exhibits resistance to polymyxins through developing chromosomal mutation or horizontal gene-transfer of plasmid which reduce the binding affinity of polymyxins to bacterial outer cell membrane required for increased membrane permeability leading to bacterial death [26]. The exact mechanism behind the ability of polymyxin B and colistin at high exposed concentration to kill polymyxins resistant K. pneumoniae is still not known. However, in our study, through concomitant intravenous plus aerosolized administration, a relatively higher concentration of polymyxin B and colistin might be attained in the lung but, compared to colistin, PMB showed better microbial eradication, decreased length-of-intubation and ICU stay, and increased survival rate. Following intravenous or inhaled dose, colistin concentration in lung tissue and ELF was found insignificant (below MIC breakpoint) in time-response while polymyxin B showed sustainable drug level and maintained MIC level in the infected lung [23,24]. Considering these findings in previous studies, our study found better recovery and survival with concomitant intravenous and aerosolized polymyxin B compared to colistin in patients with PDR K. pneumoniae VAP. In addition, the multivariate HR for death was significantly more in CLN group compared to PMB group (0.035; 95% CI=0.007-0.168). May be due to high PMB concentration in lung in PMB group’s patients, incidence of bronchospasm was higher in PMB group than the CLN group but, high exposure of colistin in kidney in patients of CLN group developed more nephrotoxicity.

To determine the actual pharmacokinetics and pharmacodynamics of polymyxin B and colistin in post intravenous plus aerosolized administration and the definite mechanism of high polymyxins concentration-mediated killing of pandrug-resistant K. pneumoniae in patients with VAP, a large randomized control trial is highly required during the emergence of highly resistant K. pneumoniae infections worldwide. No genetic identification of PDR K. pneumoniae strains, no pharmacokinetic assessment of intravenous or aerosolized polymyxins in plasma or lung, and lack of pharmacokinetic evidences of inhaled dosages of polymyxins were the major limitations of this study.

Pandrug-resistant K. pneumoniae VAP is a serious life-threatening infection with limited effective antibiotic option. This study found better microbial eradication, reduced length-of-intubation and ICU stay, and improved 60-day survival rate in patients with pandrug-resistant K. pneumoniae VAP treated with intravenous and aerosolized polymyxin B concomitantly compared to colistin.