Combined therapy with nebulized and intravenous colistin for the treatment of multidrug resistance pneumonia in burn patients: experience from a third referral burn care center in Mexico City

Abstract

Purpose

Nebulized colistin has shown promise in treating ventilator-associated pneumonia (VAP) due to carbapenem-resistant strains; evidence in burn patients is limited. This study evaluated the efficacy of nebulized colistin as an adjunct to intravenous colistin in burn patients with pneumonia caused by multidrug-resistant (MDR) Gram-negative bacteria.

Materials and methods

A retrospective cohort was conducted at a burn center in Mexico City (August 2011–May 2017). Patients diagnosed with VAP or hospital-acquired pneumonia (HAP) due to MDR rods, treated during this period with nebulized colistin plus intravenous colistin or intravenous colistin alone, were included. Clinical response and complications were evaluated on days 2, 5, and 10. Thirty-day survival was analyzed using the Kaplan–Meier method.

Results

Fifty patients were included; 41 developed VAP and 9 HAP; combined therapy group included 31 patients. Thermal burns predominated, with airway injury in 26%. Among 65 isolates, Acinetobacter baumannii (n = 28) and Pseudomonas aeruginosa (n = 27) were frequent. Combined therapy was associated with higher frequency of acute respiratory distress syndrome events (61% vs. 32%, p = 0.041), vasopressor use (27% vs. 11%, p = 0.010), longer mechanical ventilation (median 27 vs. 10 days, p = 0.005), elevated SOFA scores (median 3 vs. 1, p < 0.001), and prolonged ICU stay (median 42 vs. 26 days, p = 0.025). Bronchospasm occurred in 13% and alveolar hemorrhage in one patient. No difference in 30-day mortality was observed (log-rank p = 0.63).

Conclusion

Adding nebulized colistin to intravenous therapy did not improve outcomes in burn patients with multidrug-resistant Gram-negative pneumonia. Prospective studies are needed.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s10096-026-05550-0.

Background

Burn injuries disrupt the skin, the body’s natural physical barrier, rendering it susceptible to microbial invasion. This vulnerability may lead to infection and sepsis, irrespective of the burn mechanism. Consequently, burn treatment strategies are designed to mitigate damage and restore the protective barrier, aiming to enhance quality of life [1].

Burn injuries induce a state of immune system dysregulation, predisposing patients to infections [2]. Studies indicate that patients with extensive burns who subsequently develop infections exhibit mortality rates ranging from 42% to 65% [3–5]. Skin and soft tissue infections typically emerge within the first week following the burn injury, with pneumonia developing within 30 days [4]. Inhalation injuries to the respiratory tract increase the risk of respiratory infections [1–3], often requiring ventilatory support in severe cases. In addition to inhalation injury, prolonged hospital stays, and respiratory support are associated with pneumonia development. Approximately 22.9% of patients in burn care units develop pneumonia, with 10.9% of these cases resulting in mortality [1, 3, 6, 7].

Several risk factors for pneumonia have been identified, primarily associated with antibiotic use, prolonged ventilator dependence, complications, and extended hospitalization [8]. The predominant bacteria linked to pneumonia include Pseudomonas aeruginosa, Acinetobacter baumannii, and Enterobacterales, mainly Klebsiella pneumoniae and Escherichia coli. However, antibiotic use in burn patients exerts selection pressure on bacteria, increasing the risk of pneumonia caused by multidrug-resistant (MDR) bacteria, which are challenging to treat due to limited therapeutic options capable of reaching the lungs. The prevalence of MDR microorganisms can reach up to 41% [3]. The scarcity of effective antimicrobial options has prompted the use of older antibiotics, such as colistin, for treating MDR infections. Colistin demonstrates bactericidal activity against most Gram-negative bacilli and is particularly valuable in regions with high antimicrobial resistance and limited access to newer agents. However, colistin is a complex molecule subject to multiple derivatization processes, which may compromise its plasma concentrations, hindering the achievement of bactericidal activity. Several reports have highlighted challenges in achieving adequate plasma concentrations with colistin monotherapy in critically ill patients [9–11].

The aim of this work is to compare clinical outcomes and 30-day mortality in burn patients with hospital-acquired pneumonia (HAP) or ventilator-associated pneumonia (VAP) caused by multidrug-resistant (MDR) Gram-negative bacilli, who were treated with either intravenous colistin alone or combined therapy with intravenous and nebulized colistin.

Materials and methods

Study design and population

We conducted a retrospective, observational, single-center study of burn patients who developed HAP or VAP caused by MDR bacteria and received either intravenous colistin or combined intravenous plus nebulized colistin from August 2011 to May 2017 at the Centro Nacional de Atención a Quemados (CENIAQ) of the Instituto Nacional de Rehabilitación Luis Guillermo Ibarra Ibarra (INR LGII) in Mexico City. This national reference center for burn patients has a 222-bed capacity and admits over 120 burn patients annually.

A comprehensive review of medical records from the infection control surveillance database of the Infectious Diseases Division was performed to identify patients meeting the eligibility criteria: (a) patients aged 18 years or older diagnosed with HAP or VAP, defined as new or progressive pulmonary infiltrates on chest radiograph 48 h after admission or initiation of invasive mechanical ventilation (IMV), accompanied by fever (> 38 °C) or hypothermia (< 35 °C), blood leukocyte count ≥ 12,000 × 10³ cells/µL or < 4,000 × 10³ cells/µL, purulent tracheal secretions, or decreased oxygen saturation [12]; (b) positive cultures from endotracheal aspirates (ETA) and/or bronchoalveolar lavage (BAL) for MDR bacteria, including resistance to carbapenems; and (c) treatment with intravenous colistin or combined intravenous plus nebulized colistin administered for at least 3 days and initiated within 5 days of a positive culture report.

In order to facilitate the analysis of treatment comparisons, patients were grouped according to whether they received intravenous colistin alone or combined therapy (intravenous colistin + nebulized colistin) .

Treatment regimens, dosage and administration

The dose used for intravenous colistin administration was weight-based, with a loading dose of 5 mg/kg followed by 2.5 mg/kg divided every 12 h of sodium colistimethate (Colmesdant, Stendhal), according to the recommendations of the International Consensus Guidelines for the Optimal Use of Polymyxins [13]. The dose for patients with acute kidney injury was calculated according to the recommendations of Garonzik et al. 2011 [14]. If AKI developed during colistin treatment, the daily dose was reduced to the appropriate renally adjusted dose to achieve an average colistin concentration of 2 mg/L. For the combined therapy group, nebulized colistin was administered by diluting 75 mg of colistimethate sodium in 6 ml of sterile 0.9% saline solution with gentle agitation, for immediate administration and an approximate duration of 30 min on average. The type of inhaler used in all patients was an ultrasonic nebulizer, and was positioned proximal to the Y-adapter, delivering more than twice the pulmonary dose compared to the position adjacent to the ventilator. Humidification was not provided during nebulization. The cycling regimen varied according to the individual patient’s needs and was determined by the attending physician. Patients with incomplete medical records or colistin treatment lasting less than 72 h were excluded.

This study was approved by the Institutional Review Board of INR LGII (16/23). Due to its retrospective nature and mandatory surveillance reporting for HAP and VAP, informed consent was not required.

Data collection and definitions

Data were retrieved from the infection control surveillance database, electronic medical records, and microbiological lists of MDR isolates from the Clinical Microbiology Laboratory. Baseline clinical data, comorbidities, burn injury mechanism, total body surface area burned (TBSA), admission to the intensive care unit (ICU), microbiological cultures, treatment duration, use of synergistic antimicrobials against MDR microorganisms (e.g., rifampicin or carbapenems), hospital stay duration, and 30-day mortality were collected. Sequential Organ Failure Assessment (SOFA) scores and ventilatory parameters, including oxygen saturation, positive end-expiratory pressure (PEEP), and PaO2/FiO2 ratio, were evaluated before colistin treatment and on days 2, 5, and 10 thereafter.

Clinical response was defined as favorable if fever resolved for at least 48 h after treatment initiation and ventilatory parameters improved or leukocyte counts normalized by day 10. Stationary clinical conditions or death were considered treatment failures.

Microbiological testing

Samples were processed according to international guidelines [15–17], which assess the presence of leukocytes and epithelial cells via Gram staining. To ensure an adequate quality of the samples, ETA samples were assessed to have higher than 25 leukocytes and less than 10 epithelial cells per 100x field. In BAL samples the presence of < 1% of epithelial was considered as adequate. The quantitative culture thresholds used to define a positive culture were > 10⁶ CFU/mL for ETA samples and > 10⁴ CFU/mL for BAL samples [18]. Samples were inoculated onto 5% sheep blood agar, MacConkey agar, and chocolate agar, and incubated under aerobic and microaerophilic conditions at 37 °C with 7.5% CO2. Antimicrobial susceptibility was determined by the broth microdilution method following Clinical and Laboratory Standards Institute recommendations [19]. Due to the lack of a categorical definition for colistin susceptibility, a minimum inhibitory concentration (MIC) of ≤ 2 µg/mL was considered intermediate, and ≥ 4 µg/mL was considered resistant [19].

Microbiological isolates exhibiting resistance to more than two antimicrobial groups were classified into resistance categories as follows: MDR bacteria were defined as those with acquired non-susceptibility to at least one agent in three or more antimicrobial categories. Extensively drug-resistant (XDR) bacteria were characterized by non-susceptibility to at least one agent in all but two or fewer antimicrobial categories (i.e., susceptible to only one or two categories). Pan-drug-resistant (PDR) bacteria demonstrated resistance to all reported antimicrobial agents for that species [20].

Outcome definitions

The primary outcome of the study was to assess the clinical response at days 2, 5 and 10 between the treatment groups. The following parameters were considered as primary endpoints: the evolution of ventilatory parameters (SatO2, PEEP, FiO2, PAFi); the development of ARDS; mechanical ventilation; the evolution of the SOFA score; the requirement of vasopressors; and laboratory parameters such as the leucocyte count and the PCR levels. Secondary outcomes encompassed the prevalence of adverse effects and survival probability at day 30.

Propensity score analysis

A propensity score analysis was conducted in order to address imbalances in basal characteristics between the groups. Patients were matched using nearest-neighbor matching in a 1:1 ratio without replacement, applying a caliper of 0.2 of the propensity score distance. A logistic regression model was used to estimate the propensity scores, adjusting by age, sex and MDR/XDR infection, and basal clinical and laboratory parameters including ABSI score at admission, temperature, SOFA score, leucocyte count PaO2/FiO2 ratio, and vasopressor requirement at day 0 of treatment.

Statistical analysis

For clinical and sociodemographic data, categorical variables were summarized using frequency measures and percentages, while continuous variables were reported as medians and interquartile ranges (IQR). Comparisons between treatment groups were conducted using chi-square tests, Fisher’s exact tests, and Wilcoxon rank-sum tests.

The longitudinal evolution of clinical and laboratory parameters at days 0, 2, 5, and 10 of treatment was analyzed using mixed-effects models to account for repeated measurements within patients. Continuous variables were evaluated using aligned rank transform mixed-effects models, whereas categorical variables were analyzed using generalized linear mixed-effects models. Between-group comparisons included treatment day, treatment group, and their interaction as fixed effects, while within-group temporal changes were assessed using stratified models including treatment day as a fixed effect. In all models, patients were treated as a random intercept to account for repeated measurements over time.

Survival analysis

To evaluate 30-day survival probability between treatment groups, Kaplan-Meier curves were calculated. Follow-up duration was defined as the time from initiation of antimicrobial therapy for pneumonia until death or completion of 30 days. Comparisons between groups were performed using the log-rank test, with a p-value of < 0.05 considered significant.

Results

Clinical and sociodemographic characteristics

Fifty burn patients were included, with VAP diagnosed in 41 (82%) cases, representing the most frequent type of pneumonia during the study period. Intravenous colistin alone was administered to 19 (38%) patients, while 31 (62%) received combined therapy. Thermal burns caused by flames were the most common injury mechanism (74%), with airway compromise observed in 13 patients (26%). Similar severity was observed between both groups, as shown in Table 1.

Table 1.

Patient’s sociodemographic characteristics

	IV Colistin group (n = 19) n (%); median (IQR)	Combined therapy group (n = 31) n (%); median (IQR)	p-valuea
Sex (male)	14 (74)	22 (71)	0.8
Age (years)	40 (26–52)	31 (22–51)	0.4
Comorbidities
- Diabetes mellitus - Hypertension - Obesity - Other comorbidities	1 (5.3) 2 (11) 1 (5.3) 2 (10.6)	4 (13) 3 (9.7) 1 (3.2) 3 (9.7)	0.6 > 0.9 > 0.9 0.7
Burn location
- Soft tissue - Airway	15 (79) 4 (21)	22 (71) 9 (29)	0.7
Burn injury mechanism
- Thermal: fire - Thermal: scalded - Thermal: contact - Chemical - Electrical	13 (68) 1 (5.3) 2 (11) 1 (5.3) 2 (11)	24 (77) 2 (6.5) 2 (6.5) 0 (0) 3 (9.7)	0.8
Burn Total Surface Body Area (%)	40 (26–56)	50 (40–60)	0.08
ABSI score at admission	8 (7–10)	9 (8–10)	0.13
Pneumonia classification
- Ventilator-associated (VAP) - Hospital-acquired (HAP)	17 (89) 2 (11)	24 (77) 7 (23)	0.5
Time to diagnosis (days)
- Ventilator-associated (VAP) - Hospital-acquired (HAP)	8 (3–19) 5 (5–5)	9 (7–16) 14 (6–16)	0.20.5
Resistant Gram-negative bacilli from BAL or ETA requiring colistin treatment (n = 56) b
- Multidrug-resistant bacteria (MDR) - Extensively resistant bacteria (XDR) - Pandrug-resistant bacteria (PDR) - Other resistant Gram-negative bacilli Other non-Gram-negative bacilli isolates (n = 9) b	12 (55) 6 (27) - 5 (23) 6 (67)	14 (42) 16 (49) 1 (3) 2 (6) 3 (33)	0.13 -
Polymicrobial pneumonia	4 (21)	9 (29)	0.7
Extrapulmonary infections - Soft Tissue Infection - Urinary Tract Infection	16 (84) 13 (68) 3 (16)	25 (81) 24 (77) 6 (19)	0.15 0.5 > 0.9

a) Pearson’s chi square test; Wilcoxon rank sum test; Fisher’s exact test

b) Percentages were calculated considering the total number of Gram-negative isolates from BAL or ETA by treatment group

Abbreviations: Intravenous (IV); Abbreviated Burn Severity Index (ABSI); bronchoalveolar lavage (BAL); endotracheal aspiration (ETA); multidrug-resistant bacteria (MDR); extensively resistant bacteria (XDR); pandrug resistant bacteria (PDR)

Microbiological characteristics of isolates

A total of 65 microorganisms were isolated from endotracheal aspirates (ETA) or bronchoalveolar lavage (BAL). A. baumannii (n = 28) and P. aeruginosa (n = 27) were the most frequent pathogens identified as causative agents of pneumonia. Approximately 26% (13/50) of cases were polymicrobial (see Table 1), with Staphylococcus aureus (n = 3), K. pneumoniae (n = 2); Enterobacter ludwigii, Achromobacter xylosoxidans, and Enterococcus faecalis (n = 1 each) being the most prevalent co-infecting pathogens.

Regarding resistance profiles among clinical isolates, 26 were classified as MDR, 22 as XDR, and one as PDR (Fig. 1). Seven isolates exhibited various resistance mechanisms but did not meet the criteria for MDR classification. No statistically significant differences were observed between treatment groups in the distribution of resistant isolates. However, a higher prevalence of MDR and XDR bacteria was noted in the combined therapy group (Table 1).

Fig. 1.

Frequency of multidrug resistant (MDR), extensively-drug resistant (XDR) and pandrug resistant (PDR) strains of the cohort

Alt text

The bar chart compares the frequency of multidrug resistant bacteria differentiating by specie, isolated from burn patients who developed hospital acquired pneumonia or ventilator-associated pneumonia. A higher frequency of MDR A. baumannii and XDR P. aeruginosa was observed in this cohort.

ETA and BAL isolates from all patients who received combined therapy were resistant to most antimicrobial groups, including carbapenems (Supplementary Fig. 1). On the other hand, patients in the intravenous colistin group exhibited a high frequency of carbapenem resistance isolates, but not all, with MIC50 and MIC90 ≥ 64 µg/mL (Table 2). Within the same group, while some ETA/BAL isolates demonstrated susceptibility to carbapenems, aminoglycosides, quinolones, and/or cephalosporines, concomitant MDR or XDR strains were isolated from soft tissue infections. Consequently, intravenous colistin was selected for the treatment of both infections.

Table 2.

Minimum inhibitory concentration (MIC) of Gram-negative isolates obtained from the VAP/HAP patients. MIC₅₀ and MIC₉₀ are shown for each group of isolates. The clinical breakpoints recommended by the CLSI M100 are also shown

Antimicrobial	Acinetobacter baumannii					Pseudomonas aeruginosa
	MIC (µg/mL)		Clinical breakpointsa			MIC (µg/mL)		Clinical breakpointsa
	MIC50	MIC90	S	I	R	MIC50	MIC90	S	I	R
Amikacin	32	≥ 64	≤ 8	16	≥ 32	≥ 64	≥ 64	≤ 16	32	≥ 64
Gentamicin	16	≥ 64	≤ 2	4	≥ 8	-	-	-	-	-
Ceftazidime	≥ 64	≥ 64	≤ 8	16	≥ 32	≥ 64	≥ 64	≤ 8	16	≥ 32
Cefepime	≥ 64	≥ 64	≤ 8	16	≥ 32	≥ 64	≥ 64	≤ 8	16	≥ 32
Ciprofloxacin	32	≥ 64	≤ 1	2	≥ 4	8	≥ 64	≤ 0.5	1	≥ 2
Levofloxacin	16	≥ 64	≤ 2	4	≥ 8	8	≥ 64	≤ 1	2	≥ 4
Doripenem	≥ 64	≥ 64	≤ 2	4	≥ 8	≥ 64	≥ 64	≤ 2	4	≥ 8
Imipenem	≥ 64	≥ 64	≤ 2	4	≥ 8	32	≥ 64	≤ 2	4	≥ 8
Meropenem	≥ 64	≥ 64	≤ 2	4	≥ 8	≥ 64	≥ 64	≤ 2	4	≥ 8
Colistinb	0.5	1	-	≤ 2	≥ 4	0.5	2	-	≤ 2	≥ 4
Piperacillin/Tazobactam	≥ 128/4	≥ 128/4	≤ 16/4	32/4–64/4	≥ 128/4	≥ 128/4	≥ 128/4	≤ 16/4	32/4	≥ 64/4
Minocycline	2	125	≤ 1	2	≥ 4	-	-	-	-	-
Ampicillin/Sulbactam	32/16	64/32	≤ 8/4	16/8	≥ 32/16	-	-	-	-	-

^a Clinical breakpoint according to the Clinical & Laboratory Standards Institute M100 [19].

^b There is no categorical definition of susceptible for colistin. CLSI guidelines just refer to MIC ≤ 2 mg/L as intermediate.

Clinical outcomes

Concerning treatment characteristics, no differences were observed between groups in the prescription of complementary antimicrobials. However, the combined therapy group exhibited longer treatment durations for intravenous colistin (22 days; IQR 10–35 days vs. 11 days; IQR 7–21 days, p = 0.02) and adjuvant carbapenems (28 days; IQR 14–40 days vs. 11 days; IQR 9–24 days, p = 0.011). Additionally, inhaled corticosteroids were more frequently prescribed in patients receiving combined therapy (Table 3). Adjustment by baseline characteristics between treatment groups, replicate the same results for the combined therapy group.

Table 3.

Treatments characteristics and clinical outcomes between treatment groups

	All cases			Propensity score analysis
	IV Colistin (n = 19)	Combined therapy (n = 31)	p-valuea	IV Colistin (n = 14)	Combined therapy (n = 14)	p-valuea
Treatments characteristics
Colistin duration - Nebulized colistin (days) - Intravenous colistin (days)	- 11 (7–21)	12 (8–20) 22 (10–35)	- 0.02	- 12 (7–21)	13 (9–19) 20 (11–45)	- 0.032
Rifampicin complementary therapy - Duration (days)	5 (26) 11 (8–12)	3 (9.7) 28 (11–32)	0.2 > 0.9	3 (21) 12 (7–17)	1 (7.1) 28 (28–28)	0.6 > 0.9
Carbapenem complementary therapy - Duration (days)	18 (95) 11 (9–24)	31 (100) 28 (14–40)	0.4 0.011	14 (100) 16 (9–24)	14 (100) 28 (20–45)	> 0.9 0.017
Inhaled complementary steroids	4 (21)	17 (55)	0.019	3 (21)	6 (43)	0.4
Clinical outcomes related to colistin therapy
Ventilatory outcomes - ARDS development - IMV length (days)	6 (32)10 (5–24)	19 (61) 27 (16–38)	0.041 0.005	5 (36) 9 (6–18)	9 (64) 29 (14–33)	0.13 0.084
AE associated to combined therapy - Bronchospasm - Hemorrhage	- -	4 (13) 1 (3.2)	- -	- -	1 (7.1) -	-
Renal outcomes - Acute kidney injury development - Renal replacement therapy requirement (RRT)	7 (37) 5 (26)	13 (43) 12 (40)	0.7 0.3	6 (43) 4 (29)	7 (50) 6 (43)	0.7 0.7
Infectious complications - Septic shock - C. difficile infection - Bacteriemia - Candidemia	10 (53) 1 (5.3) 10 (53) 1 (5.3)	26 (84) - 16 (52) 6 (19)	0.017 0.4 > 0.9 0.2	8 (57) 1 (7.1) 5 (36) 1 (7.1)	11 (79) - 9 (64) 1 (7.1)	0.4 > 0.9 0.13 > 0.9
Bacterial reinfection	1 (5.3)	4 (13)	0.6	1 (7.1)	1 (7.1)	> 0.9
Hospitalization - ICU (days) - Total in-hospital stay (days)	26 (15–40) 34 (26–65)	42 (25–62) 56 (30–84)	0.025 0.2	25 (16–40) 32 (25–65)	34 (29–62) 55 (29–84)	0.11 0.3
30-days mortality	5 (26)	7 (23)	> 0.9	5 (36)	4 (29)	> 0.9

a) Wilcoxon rank sum test; Pearson’s chi square test; Fisher’s exact test

Abbreviation: Intravenous (IV); Nebulized (Neb); Acute Respiratory Distress Syndrome (ARDS); Invasive Mechanical Ventilation (IMV), adverse events (AE), Intensive Care Unit (ICU)

At the end of the treatment, patients receiving combined therapy were more likely to develop ARDS (6 [32%] vs. 19 [61%], p = 0.041), required longer time of invasive mechanical ventilation (10 days, IQR 5–24 vs. 27 days, IQR 16–38, p = 0.005), developed a higher frequency of bacteriemia events (53% vs. 84%, p = 0.017) and had longer Intensive Care Unit (ICU) stay (26 days; IQR 15–40 vs. 42 days; IQR 25–62, p = 0.025) (Table 3). However, no differences in these parameters, including renal adverse events, infectious complications and bacterial reinfection, were observed in the matched-propensity score subanalysis.

In terms of toxicity, neither the intravenous nor combined therapy groups required an increased frequency of renal replacement therapy; however, a trend toward higher renal replacement therapy was noted in the combined therapy group compared to the intravenous colistin group. In this group of patients, the daily dose was reduced to the appropriate renally adjusted dose to achieve a Css, avg colistin concentration of 2 mg/L. Respiratory adverse effects secondary to nebulized colistin administration included bronchospasm in four patients (13%) and alveolar hemorrhage in one patient (3.2%) (Table 3).

On the other hand, clinical and laboratory parameters evaluated on treatment days 2, 5, and 10 showed favorable longitudinal changes in both treatment groups (Table 4), with statistically significant improvements in severity-related parameters, including SOFA score and PaO₂/FiO₂ ratio. Notably, the combined therapy group showed significant improvements in a broader range of ventilatory parameters and vasopressor requirement, compared with the IV colistin group when changes over time were assessed within each group. However, when comparisons are made between treatment groups during the first 10 days of treatment, no statistically significant differences in any clinical outcome were observed, suggesting no additional measurable clinical benefit of combined therapy over IV colistin alone. Adjustment for baseline severity characteristics did not modify these findings when compared between treatment groups (Supplementary Table 1).

Table 4.

Clinical and laboratory parameters evolution during day 2, 5 and 10th between treatment groups

	IV Colistin group (n = 19) n (%); median (IQR)					Combined therapy group (n = 31) n (%); median (IQR)					Comparing clinical parameter by time and type of treatment
	Day 0	Day 2	Day 5	Day 10	p-valuea	Day 0	Day 2	Day 5	Day 10	p-valuea	p-valueb
Clinical outcomes
- Temperature (°C)	37.6 (36.8–38.5)	37.5 (36.3–38.2)	37.0 (36.5–38.0)	37.0 (36.5–37.0)	0.16	38.1 (36.6–38.5)	37.0 (36.8–38.0)	37.3 (37.0–38.0)	37.0 (36.5–37.4)	0.004	0.58
- SOFA score	4 (2–5)	3 (2–5)	2 (1–4)	1 (1–3)	< 0.001	6 (3–8)	5 (3–6)	4 (2–5)	3 (2–5)	< 0.001	0.50
Laboratory parameters
- Leucocytes (1 × 103 cel/µL)	10.1 (8.1–11.8)	10 (7.0-13.1)	10.2 (8.1–13.5)	9.4 (7.5–13.6)	0.98	10.1 (7.7–14.0)	9.9 (6.8–13.6)	10.2 (8.2–12.4)	9.6 (7.1–11.8)	0.51	0.86
- CRP (mg/L)	107 (85–183)	124 (95–267)	124 (93–217)	90 (65–153)	0.34	166 (137–245)	165 (130–245)	143 (116–268)	132 (101–193)	0.009	0.18
Ventilatory support
- O2 saturation (%)	96 (95–98)	97 (96–98)	96 (96–98)	97 (96–98)	0.65	97 (95–98)	96 (95–99)	96 (95–99)	98 (96–98)	0.29	0.95
- PEEP (cm H2O)	8 (6.5–11)	7 (6–10)	6 (5–9)	6 (5–10)	0.06	8 (7.5–10)	7 (6–8)	6 (6–8)	8 (6–8)	0.023	0.87
- FiO2 (%)	40 (40–60)	40 (40–50)	40 (30–40)	30 (30–40)	< 0.001	45 (40–60)	40 (40–50)	40 (40–40)	40 (40–40)	0.03	0.13
- PaO2/FiO2	201 (180–267)	278 (206–300)	288 (233–348)	220 (140–358)	0.014	188 (109–240)	223 (160–277)	225 (185–291)	238 (161–293)	0.02	0.39
Cardiovascular support
- Vasopressor requirement	8 (42)	7 (37)	5 (26)	2 (11)	0.17	18 (58)	19 (61)	15 (48)	8 (27)	0.03	0.95

a) Aligned Rank Transform mixed-effects model and Chi-square test were performed, as appropriate, considering the changes in the clinical/laboratory/ventilatory parameter by day of follow-up in the same group of treatment. Comparison of main effects of continuous variables were assessed by ANOVA

b) Aligned Rank Transform mixed-effects model and generalized linear mixed-effects model with binomial distribution test were performed considering the changes in the clinical/laboratory/ventilatory parameter by day of follow-up and group of treatment. Comparison of main effects were assessed by ANOVA.

Abbreviations: Intravenous (IV); Sequential Organ Failure Assessment (SOFA); C-reactive protein (CRP); Positive End-Expiratory Pressure (PEEP); inspired oxygen fraction (FiO₂); Partial pressure of carbon dioxide (P_aO₂).

Survival analysis

Kaplan-Meier analysis of all-cause mortality revealed no statistically significant differences in 30-day mortality between treatment groups. During the first two weeks of follow-up, patients receiving combined therapy exhibited higher survival probabilities; however, after this period, mortality rates equaled those of the intravenous colistin group, and this trend persisted until day 30 (Fig. 2).

Fig. 2.

Survival probability at 30 days of follow-up between treatment groups

Alt text

Kaplan-Meier curve shows in the first panel (up-to-down) the survival probability of patients during 30 days of follow-up, depending on if they received combined therapy (blue) or intravenous colistin alone (pink). No differences were observed regarding 30 day-mortality between treatment groups. The second panel represents the number of patients that are at risk of developing the outcome (death) during the follow-up. The third panel shows the frequency of censored patients in each point time.

Discussion

Infections caused by Gram-negative microorganisms represent a growing global threat because of limited therapeutic options [21]. The latest report on the global burden of bacterial antimicrobial resistance (AMR) estimates that 1.14 million deaths annually may be associated with AMR [22]. Consequently, there is an urgent need for effective treatments against MDR bacteria; however, in countries such as Mexico, newer antibiotics, including cefiderocol, meropenem/vaborbactam, and sulbactam/durlobactam, remain unavailable or difficult to access, leading to the continued use of older agents such as colistin. Burn patients are particularly vulnerable to MDR infections with diverse clinical manifestations [23]. In our burn unit, we previously identified P. aeruginosa and A. baumannii as the main infecting microorganisms [24, 25], both considered priority pathogens by the World Health Organization.

Although colistin remains a potential option for severe infections, its intravenous use in respiratory tract infections is controversial because of poor alveolar penetration [26] and increased risks of nephrotoxicity and neurotoxicity [27, 28]. Antibiotic nebulization enables direct drug delivery to the infection site, achieving concentrations above the MIC; however, its use in critically ill patients remains debated [29]. Despite limited evidence, nebulized antibiotics are increasingly used in mechanically ventilated patients with pneumonia [30, 31]. Evidence on nebulized colistin in specific populations, such as burn patients, remains scarce.

In this study, the most frequent respiratory complication was ventilator-associated pneumonia. In burn patients with inhalation injuries, pulmonary complications are common; however, all burn patients experience more respiratory complications, even without direct lung injury [32]. The median time to VAP development was at the beginning of the second week of hospitalization, earlier than previously reported [3]. Regarding bacterial etiology, an alarming number of Gram-negative MDR microorganisms were observed in our unit, with A. baumannii and P. aeruginosa being the most frequent bacteria associated with infection development in burn units [3, 33, 34], harboring various enzymatic and non-enzymatic antimicrobial resistance genes. As previously reported [34], our country is facing an increasing number of carbapenem-resistant microorganisms, with non-fermenters exhibiting the highest rates. It is important to highlight that both groups had extrapulmonary infections concurrent with respiratory infections, a situation commonly observed in this patient population. The 2016 National Burn Repository Report indicated that 70% of all complications during the acute phase are related to infections, with respiratory, urinary, and skin and soft tissue infections being the most prevalent [35].

Treatment efficacy of combined therapy has been proven to be non-inferior to intravenous colistin as monotherapy [28]. Furthermore, it has been associated with shorter durations of mechanical ventilation support after pneumonia resolution and higher rates of microbiological eradication [28, 36]. These findings support a beneficial effect of combined treatment among patients with pneumonia, despite no differences having been proven in overall mortality at 30 days or ICU length of stay. In our study, clinical improvement on day 10 was not different between treatment groups, and approximately 70% of patients exhibited clinical and ventilatory improvement during this period, which is consistent with findings reported in previous studies. Of particular interest, a similar effect on the PaO2/FiO2 ratio was observed by Abdellatif et al. [28], where patients who received nebulized colistin achieved higher oxygenation diffusion capacity on day 14. In our study, the PaO2/FiO2 ratio was higher in the combined therapy group on day 10; paradoxically, ARDS was more frequently observed in the nebulized colistin group, alongside higher SOFA scores and vasopressor requirements during treatment. The observed differences from other studies may be related to factors beyond pulmonary infection and the colistin administration modality, where septic shock and airway burn injuries may influence ARDS development [37]. In this regard, approximately 80% of patients who received combined therapy presented with septic shock in our cohort (84% vs. 53%, p = 0.017), and approximately 26% had airway burn injuries (21% vs. 29%, p = 0.7), which partially explains the high proportion of ARDS development.

In addition, it’s important to note that, in our study, patients of the combined therapy group had larger courses of intravenous colistin treatment. This can be explained, partially, by the fact that some of our patients developed concomitant soft tissue infection at the site of the burn, which was influenced by the same MDR/XDR microorganism found a causative of the pneumonia, which could have prolonged their treatment. With this in mind, because soft tissue infection in burn patients represents a common complication, the treatment of pneumonia caused by the same microorganisms could influence the treatment time course, and so on influence treatment efficacy.

In the same line of our findings, a recent retrospective cohort study by Katip et al. [38] valuated adjunctive nebulized colistin plus high-loading-dose intravenous colistin compared with high-loading-dose intravenous colistin alone in critically ill patients with CRAB pneumonia. After propensity score adjustment, the authors found no significant additional benefit in 30-day survival, clinical response, microbiological response, or nephrotoxicity. Although their population was not limited to burn patients, these findings are consistent with our adjusted analysis, in which adjunctive nebulized colistin was not associated with significant improvements in ventilatory parameters, mortality, or adverse events compared with intravenous colistin alone.

Acute kidney injury is a common finding reported in most randomized clinical trials associated with intravenous colistin therapy [39]. Most studies evaluating nephrotoxicity associated with nebulized colistin have not observed an increase in renal adverse events compared to intravenous colistin monotherapy [36, 40]. This aligns with our findings, where renal function appeared unaffected in individuals receiving combined treatment. This outcome may be attributed to nebulization promoting a localized effect in the lungs, reducing systemic toxicity despite higher colistin dosages. Additionally, the incidence of adverse events such as bronchospasm, alveolar hemorrhage, or neurotoxicity was low, as confirmed in our study and consistent with prior reports [28, 36].

Mortality in individuals treated with nebulized colistin for carbapenem-resistant pneumonia has been evaluated in several studies. A multicenter observational study conducted in Taiwan by Feng Jia-Yi et al. [40], reported higher survival rates at two weeks among those receiving adjunctive nebulized colistin compared to intravenous therapy alone. However, no differences were observed at 28 days. These results are consistent with other studies [28, 36] and our findings, where individuals receiving combined treatment exhibited survival probabilities below 75% at 30 days, similar to those treated with intravenous colistin alone. Factors contributing to the lack of differences between treatment groups include burn severity and multidrug resistance profiles. Specifically, mortality is closely associated with the extent and depth of burns and airway involvement [41]. Infection with MDR microorganisms is also linked to reduced survival rates [42]. In our cohort, a higher proportion of individuals treated with nebulized colistin had severe burns, airway involvement, and a greater prevalence of MDR and XDR microorganisms. Although most studies exclude burn patients, the presence of severe illness and increased risk of MDR or XDR bacterial infections may diminish the efficacy of nebulized colistin therapy.

This study has several limitations. First, its retrospective design and non-randomized treatment allocation preclude establishing a causal relationship between treatment strategy and clinical outcomes. In our case, treatment decisions were based on the best medical recommendation in accordance with the clinical characteristics of the patients, whichcould result in an imbalance of baseline severity clinical characteristics between treatment groups. Although a propensity score matching was performed to reduce this imbalance, this approach reduced the analytical sample size, and in consequence the statistical power to detect differences. Therefore, these findings should be interpreted as exploratory and hypothesis-generating and larger studies are needed to assess the real efficacy of combined nebulized plus intravenous colistin treatment. Second, the presence of concurrent soft-tissue infections and the development of VAP or HAP in patients with extensive burn injuries complicate the evaluation of clinical improvement and microbiological eradication highlighting the complexity of treating pneumonia in patients with severe burns. Third, direct measurement of nebulized colistimethate bioavailability was not performed; such assessment would be useful to confirm adequate penetration into the lower respiratory tract, and it should be addressed in future studies. Finally, prospective follow-up including control cultures to confirm microbiological eradication and assessment of radiologic improvement would enhance the comprehensive evaluation of treatment outcomes. Prospective studies with randomized treatment assignment and larger sample sizes are needed to more robustly evaluate the potential clinical benefit of adjunctive nebulized colistin.

Conclusions

In conclusion, the use of combined nebulized therapy for the treatment of MDR/XDR pneumonia in burn patients may not offer greater benefits compared to intravenous colistin alone. Given the non-randomized treatment allocation, potential confounding by indication, and limited sample size, these findings should be interpreted cautiously. The assessment of treatment response in this population is particularly challenging due to the high prevalence of XDR or PDR organisms, polymicrobial infections, and concurrent extrapulmonary infections. Larger prospective studies with standardized microbiological, radiologic, and pharmacokinetic follow-up are needed to better define the potential clinical role of adjunctive nebulized colistin.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: R.F.C, L.E.L.J.

Methodology: B.P.A.R, A.F.C.O, M.A.D.M.

Formal analysis and investigation: M.A.D.M, R.F.C, C.A.C.C, M.H.D, T.C.H, M.C.M.W, M.A.G.L.

Writing - original draft preparation: M.A.D.M, R.F.C, L.E.L.J.

Writing - review and editing: B.P.A.R, A.F.C.O, M.A.D.M, R.F.C, L.E.L.J

Supervision: R.F.C, L.E.L.J.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.

Declarations

Ethics approval

This study was approved by the Institutional Review Board of the Instituto Nacional de Rehabilitación Luis Guillermo Ibarra Ibarra (16/23) and is in line with the principles of the Declaration of Helsinki.

Consent to participate

Due to the retrospective nature of the study and the mandatory surveillance reporting for HAP and VAP, informed consent was waived.

Competing interests

The authors declare no competing interests.