Incidence, risk factors and clinical outcome of multidrug-resistant organisms after heart transplantation

Abstract

BACKGROUND

Transplant recipients commonly harbor multidrug-resistant organisms (MDROs), as a result of frequent hospital admissions and increased exposure to antimicrobials and invasive procedures.

AIM

To investigate the impact of patient demographic and clinical characteristics on MDRO acquisition, as well as the impact of MDRO acquisition on intensive care unit (ICU) and hospital length of stay, and on ICU mortality and 1-year mortality post heart transplantation.

METHODS

This retrospective cohort study analyzed 98 consecutive heart transplant patients over a ten-year period (2013-2022) in a single transplantation center. Data was collected regarding MDROs commonly encountered in critical care.

RESULTS

Among the 98 transplanted patients (70% male), about a third (32%) acquired or already harbored MDROs upon transplantation (MDRO group), while two thirds did not (MDRO-free group). The prevalent MDROs were Acinetobacter baumannii (14%), Pseudomonas aeruginosa (12%) and Klebsiella pneumoniae (11%). Compared to MDRO-free patients, the MDRO group was characterized by higher body mass index (P = 0.002), higher rates of renal failure (P = 0.017), primary graft dysfunction (10% vs 4.5%, P = 0.001), surgical re-exploration (34% vs 14%, P = 0.017), mechanical circulatory support (47% vs 26% P = 0.037) and renal replacement therapy (28% vs 9%, P = 0.014), as well as longer extracorporeal circulation time (median 210 vs 161 min, P = 0.003). The median length of stay was longer in the MDRO group, namely ICU stay was 16 vs 9 d in the MDRO-free group (P = 0.001), and hospital stay was 38 vs 28 d (P = 0.006), while 1-year mortality was higher (28% vs 7.6%, log-rank-χ²: 7.34).

CONCLUSION

Following heart transplantation, a predominance of Gram-negative MDROs was noted. MDRO acquisition was associated with higher complication rates, prolonged ICU and total hospital stay, and higher post-transplantation mortality.

Key Words: Heart transplantation; Multi drug resistant organisms; Transplantation complications; Transplantation outcome

Core Tip: We evaluated multidrug-resistant organisms (MDROs) in heart transplantation and their impact on patient outcome. Carbapenem-resistant Gram-negative bacteria predominated, in line with the epidemiologic pattern in south-eastern Europe. Among comorbidities, renal failure and higher body mass index were shown to be important risk factors pre-transplantation. Surgical and medical complications were shown to be predictive of MDRO acquisition, while no association was shown for the type of cardiomyopathy, for the mode of admission [from home, ward or intensive care unit (ICU)] or for previous cardiac surgery. MDRO presence was associated with longer ICU and hospital length of stay, and higher ICU-mortality and 1-year mortality.

INTRODUCTION

In the south-eastern European countries, a higher prevalence of multidrug-resistant organisms (MDROs) is annually being reported in comparison to the northwest of Europe, pertaining mainly to multidrug-resistant (MDR) Gram-negative organisms[1]. High MDRO prevalence is strongly correlated with nosocomial infections[2]. Previous publications suggest an increased mortality in transplanted patients with MDRO colonization or bacteremia[3,4]. In concert with the patient safety guidelines of the World Health Organization, MDRO infections represent the occurrence of avoidable harm: Therefore MDRO prevalence must be recorded and its effects on patients measured[5].

There is a gap of knowledge on the impact of MDRO acquisition on heart transplant recipients in intensive care unit (ICU) environments with a high burden of gram-negative pathogens displaying advanced resistance patterns, for which limited treatment options are available.

We investigated whether MDRO presence negatively affects heart transplantation outcomes and constitutes a valid concern for patient safety, with an emphasis on the Gram-negative MDR pathogens, which are endemic in ICU environments in south-eastern Europe.

MATERIALS AND METHODS

Study design

This is a retrospective 10-year study conducted at the Onassis Cardiac Surgery Center, a referral hospital for cardiomyopathies and heart failure. The hospital hosts the national heart transplantation program in Greece. The post-surgical care of heart transplant patients takes place in the 16-bed cardiac surgery ICU. The study protocol received approval from the hospital scientific and ethics committee (P.E.E 787/16.10.2023).

Pathogen identification

As per ICU microbiological surveillance protocol, colonization cultures were obtained from transplanted patients daily during the first week post transplantation, followed by three times per week during ICU stay. Nasal, pharyngeal and rectal swabs, bronchial secretions and urine samples were taken. Blood cultures were drawn as per clinical indication, i.e. following the occurrence of signs consistent with an inflammatory response or circulatory compromise. A specific pathogen isolated from sequential cultures from the same patient was analyzed as a single occurrence, given that recurrent identification of the same pathogen in the ICU is generally ascribed to pathogen tolerance and persistence mechanisms, rather than reinfection[6,7].

The MDROs reviewed were those frequently encountered in ICU care, which are currently under active monitoring (epidemiological surveillance), as per national and European guidelines[8]. These include carbapenem-resistant organisms, namely Acinetobacter baumannii, Klebsiella pneumoniae and Pseudomonas aeruginosa isolates, as well as methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant Enterococcus sp (VRE).

Study population

The study comprised all 98 consecutive patients who underwent heart transplantation in our unit between 1/1/2013 and 31/12/2022 (10 years). Group comparison was carried out according to MDRO isolation from patient samples, i.e. patients were separated into two groups as per MDRO presence or MDRO absence.

Risk factors and outcomes

The variables reviewed were: (1) Patient demographic characteristics; (2) Clinical pre-transplantation characteristics: Cause of cardiac failure, presence of ventricular assist device (VAD), comorbidities and acute physiologic derangement e.g. continuous renal replacement treatment (CRRT); (3) Route of admission at transplantation: Ambulant patient/admission from home; inotrope-dependent patient/admission from hospital ward; mechanical circulatory support patient/admission from cardiac ICU; (4) Operating procedure data: Duration of general anesthesia, duration of extracorporeal circulation, aorta closure time, transfusion requirements; (5) Post-transplantation clinical data: Days of mechanical ventilation, use of nitric oxide (NO) as a marker for refractory post-transplantation hypoxemia, days of CRRT, major medical and surgical complications; and (6) Microbiology data: Type of MDRO, MDRO presence upon transplantation or acquisition during ICU stay, donor-acquired MDRO. Outcome measures were ICU length of stay and total hospital length of stay post-transplantation, as well as ICU mortality and early (30-d) and late (1-year) mortality.

Statistical analysis

SPSS v.25 software was employed for statistical analysis. Categorical values are given as absolute numbers and percentages (relative frequency). The Kolmogorov-Smirnoff test was used to check whether continuous variables conformed to a normal distribution. Normally distributed continuous variables, namely weight, height and body mass index (BMI), are given as mean values with SD. Continuous variables non-normally distributed are given as median with percentiles (25-75^th). Demographic and comorbidity data prior to transplantation were explored through descriptive statistics. The Chi-square test was used for the analysis of categorical variables. The Mann-Whitney test was used for the comparison of means of continuous variables with a non-normal distribution, while the t-test was used for normally distributed continuous variables. The Kaplan-Meier curve was used for 1-year post-transplantation survival analysis.

RESULTS

Demographics

The population of heart transplant recipients consisted of 98 Caucasian patients, of whom 69 were male (70.4%) (Table 1). The median age was 49.5 years (range: 15-66 years). The mean patient BMI was 25.1 (range: 15.1-33); six patients fell within the cachexia range (BMI < 18.5), and seven patients within the obesity range (BMI ≥ 30). When comparing patients with or without MDRO presence, a statistically significant association between higher body weight/BMI and MRDO presence was found (P = 0.001), while age was not shown to have an impact.

Table 1 Selected characteristics of heart transplant recipients according to multidrug-resistant organism presence.

Heart transplant recipients	Total (n = 98)	MDRO presence (n = 32)	MDRO-free (n = 66)	P value
Demographics
Age (yr)	50 (37-56)	51 (41-56)	45.5 (35.5-56)	0.48
Sex (male)	69 (70.4)	25 (78.1)	44 (66.7)	0.24
Height (cm), SD	172 (9)	174 (8)	172 (9)	0.25
Body weight (kg), SD	75 (15)	81 (11)	72 (16)	0.002
ΒΜΙ (kg/m2), SD	25.1 (3.8)	26.9 (3)	24.3 (3.9)	0.001
Heart failure etiology
Non-ischemic cardiomyopathy	75 (76.5)	23 (71.9)	52 (78.8)	0.45
Ischemic cardiomyopathy	22 (22.4)	8 (25)	14 (21.2)	0.67
Comorbidities
Diabetes mellitus	21 (21.4)	8 (25)	13 (19.7)	0.54
CRF (eGFR ≤ 60 mL/min/m2)	13 (13.3)	8 (25)	5 (7.6)	0.017
Vasculopathy	40 (40.8)	17 (53.1)	23 (34.8)	0.08
COPD	5 (5.1)	2 (6.3)	3 (4.5)	0.72
Previous cardiac surgery	18 (18.4)	24 (75)	38 (57.6)	0.09
Smoking history	55 (56.7)	21 (65.6)	34 (52.3)	0.21
Status at transplantation
VAD	54 (55.1)	22 (68.8)	32 (48.5)	0.06
Admitted from home	65 (66.3)	20 (62.5)	45 (68.2)	0.10
Admitted from ward	15 (15.3)	5 (15.6)	10 (15.2)	0.95
Admitted from ICU	18 (18.4)	7 (21.9)	11 (16.7)	0.53
Operating room
General anesthesia (hours)	7 (6-8)	6.5 (6-8)	7 (6-8)	0.84
Extracorporeal circulation (min)	168 (144-229)	210 (146-271)	161 (141-193)	0.003
Aorta closure (min)	115 (77-198)	115 (85-130)	115 (67-211)	0.94
Transfusion (RBC units)	4 (2-10)	9.5 (4-13)	4 (2-7)	0.001
Post transplantation
Primary graft dysfunction	13 (13.3)	10 (31)	3 (4.5)	0.001
Surgical re-exploration	20 (20.4)	11 (34)	9 (13.6)	0.017
Ventilator days	2 (1-5)	5 (3-3.5)	1.3 (1-3)	0.001
NO for refractory hypoxemia	25 (25.5)	14 (43.8)	11 (16.7)	0.04
CRRT for renal failure	15 (15.3)	9 (28)	6 (9)	0.014
Post Tx mechanical circulatory support	32 (32.7)	15 (46.9)	17 (25.8)	0.037
ATG treatment (days)	5 (4-7)	6 (5-8)	5 (4-6)	0.08
Outcomes
ICU stay, days	10 (7-17)	15.5 (10-26)	9 (7-12)	0.001
Post transplantation total hospital stay, days	30 (24-41)	38 (25-62)	28 (21-39)	0.006
Early mortality (30 d)	9 (1)	6 (18.8)	3 (4.5)	0.02
Late mortality (1-year)	14 (14.3)	9 (28.1)	5 (7.6)	0.006
Died in ICU	12 (12.2)	8 (25)	4 (6)	0.003

Categorical values are given as absolute numbers and percentages. Normally distributed continuous variables (weight, height, body mass index) are given as mean values with SD. Continuous variables non-normally distributed are given as median with percentiles (25-75^th). ATG: Anti-thymocyte globulin; BMI: Body mass index; CRRT: Continuous renal replacement treatment; ICU: Intensive care unit; NO: Nitric oxide; VAD: Ventricular assist device; COPD: Chronic obstructive pulmonary disease; MDRO: Multidrug-resistant organism; CRF: Chronic renal failure.

Clinical characteristics

The reason for transplantation was non-ischemic cardiomyopathy in 76 patients (77.5%) and ischemic cardiomyopathy in 22 patients (22.4%). Of the former, 60 cases (61%) pertained to dilated cardiomyopathy, 10 cases (10.2%) to hypertrophic obstructive cardiomyopathy and 6 (6.1%) to other non-ischemic cardiomyopathies. Previous chemotherapy for hematologic neoplasia, mostly lymphoma, was the underlying cause for dilated cardiomyopathy in 9 patients (9.2%).

Over half of the patients (54/98 or 55%) were supported by a VAD on transplantation, equally divided between left ventricular and biventricular support (left ventricular assist device and bi-VAD respectively). Apart from these patients, an additional number of 8 patients had undergone cardiac surgery before transplantation, namely coronary artery bypass or valvular surgery. Therefore, the total number of patients with a history of cardiac surgery added up to 62 patients (63.3%).

Most patients were ambulant on admission (admitted from home: 65/98, 66.3%), of whom 16 were supported by a VAD (16.3% of total transplantations, 24.6% of home admissions). Upon transplantation, 33 patients were already hospitalized (33.7%). Of these, 15 were admitted from a cardiology ward (15.3% of total patients, 45% of hospitalized recipients), while 18 were admitted from the cardiac ICU (18.4% of total patients, 54.5% of hospitalized recipients). The patients admitted from a hospital ward had a median pre-transplantation hospitalization of 16 d (range: 2-520 d), while 9/15 (60%) were supported by a VAD. The patients admitted from a cardiac ICU had a median pre-transplantation ICU stay of 47 d (26-429), while 2/18 (11%) had a VAD in place, 14/18 (77.8%) were on circulatory support by intra-aortic balloon pump (IABP) and 1/98 was on Extracorporeal Membrane Oxygenation (ECMO) support and mechanically ventilated.

The type of cardiomyopathy, the presence of VAD or the mode of admission (home/ward/ICU) were not shown to differ between patients with MDROs and MDRO-free patients.

Comorbidities

Pertaining to comorbidities, peripheral vasculopathy was the most prevalent condition (40.8%), followed by diabetes mellitus (21.4%) and chronic renal failure (13.3%). An additional 10% of patients had arterial hypertension without vasculopathy. Chronic obstructive pulmonary disease and sleep apnea was present in 5.1%. Nearly a quarter of patients (24.5%) had suffered a cerebrovascular accident prior to transplantation. The prevalence of dyslipidemia was 23.4%, while 56% of patients had a smoking history.

Renal failure was the only comorbidity shown to be related to MDRO presence in the recipient.

Surgical procedure

The median duration of general anesthesia was 7 h (range: 4-18), the median duration of extracorporeal circulation was 168 min (range: 99-413) and median aorta closure time was 115 min (range: 36-283). The median number of packed-red-cell units transfused was 4 (range: 0-41).

Extracorporeal circulation time was shown to differ across comparison groups, namely longer duration was associated with a greater MDRO prevalence in heart recipients. Similarly, larger transfusion requirements, a marker of surgical complications, were associated with a greater MDRO prevalence. No difference was shown for general anesthesia duration or mean aorta closure time.

Post-transplantation events

Primary graft dysfunction occurred in 13 patients (13.3%). Mechanical circulatory support was required in 32 patients (32.7%). Of the latter, 26/32 (81%) were supported by IABP (median duration: 4 d, range: 1-39 d), of whom 7/26 (26.9%) died in ICU. ECMO support was needed for 11/32 patients (34%, median duration: 13.5 d, range: 1-26 d), of whom 8/11 (72.7%) died in ICU.

Major surgical complications occurred in 32 patients (32.7%). Surgical re-exploration was needed in 21/32 cases (65.6%), mainly due to hemorrhage (16/21, 76.2%) or tamponade (5/21, 23.8%). Delayed sternal closure was necessary in 18 patients (median open chest duration: 2 d, range: 1-20). Sternal wound debridement was carried out in 18/32 patients (18.4%), of whom 12 were open chest cases.

Major medical complications included acute renal failure requiring CRRT in 15 patients (median duration: 13 d, range: 1-33), major thromboembolism in 6 patients (6.1%) and multiple organ failure in 11 patients (11.2%). Septic shock occurred in 7 patients, of whom 5 had MDRO bacteremia.

Ventilator support exceeded 48 h post-surgery in 48/98 patients (49%), while re-intubation was necessary in 16/98 patients (16.3%). Post-transplantation refractory hypoxemia requiring treatment with inhaled NO was noted in 25/98 patients (25.5%). Median post-transplantation sedation was 40 h (range: 5-650). Median chest tube days were 7.5 (range: 6-39). Tracheostomy was performed on 8 patients (median duration: 12.5 d, range: 3-83).

Induction anti-thymocyte globulin (ATG) was used in all but 4 patients (96%, median duration: 6 d, range: 1-13).

Ventilator days, sedation hours and NO administration for refractory hypoxemia, right ventricle support were all associated with MDRO presence in the recipient. This was also true for primary graft dysfunction, post-transplantation mechanical circulatory support and any major surgical or medical complication, such as surgical re-exploration and renal replacement (CRRT). No impact was shown for the administration of ATG or its duration.

Microbiology

MDRO presence pertained almost exclusively to Gram-negative carbapenemase producing pathogens (94,8% of MDRO presence and 100% of MDRO bacteremias, Figure 1). Carbapenem-resistant Acinetobacter baumannii was the prevalent MDRO (36% of MDROs isolated), found in 14 patients (14.3% of all transplanted patients). Half of those already harbored Acinetobacter upon transplantation, five more acquired the organism in the ICU (median length of ICU stay before MDRO acquisition: 11 d, range: 8-33 d), while two acquired the pathogen from a donor subsequently found to have been bacteremic (blood cultures had been drawn from the donor upon organ procurement).

Figure 1 Multidrug-resistant organism carriage and bacteremia in heart transplantation patients (n = 98). Transplanted patients with multidrug-resistant organisms (MDROs): Acinetobacter baumannii was the most prevalent MDRO, followed by Pseudomonas aeruginosa and Klebsiella pneumoniae. MDRO acquisition: Half of the patients with Acinetobacter and a single patient with Pseudomonas were already colonized pre-transplantation, while most Pseudomonas strains and all Klebsiella strains were intensive care unit acquired. MDRO bacteremia was noted in about a third of Acinetobacter and Pseudomonas carriers. Four cases of MDRO bacteremia were donor-acquired (the donor was subsequently found to have been bacteremic at the time of organ procurement). All MDRO bacteremias were due to Gram-negative pathogens. MDRO: Multidrug-resistant organism.

Carbapenem-resistant Pseudomonas aeruginosa followed in frequency (30.8% of MDROs isolated), found in 12 recipients (12.2%). Of those, 10 patients acquired the organism in ICU (median length of ICU stay for acquisition: 20 d, range: 2-31 d), one patient was already a carrier upon transplantation, while one acquired the pathogen from the donor.

Carbapenem-resistant Klebsiella pneumoniae (28.2% of MDROs isolated) was found in 11 patients (11.2%). All patients acquired the organism in ICU (median length of ICU stay for acquisition: 8 d, range: 3-40 d).

Overall, regarding the timeline of Gram-negative pathogen acquisition within the ICU, Klebsiella ICU acquisition generally preceded Acinetobacter acquisition, while Pseudomonas acquisition was a delayed event (median acquisition day: 8^th for Klebsiella, 11^th for Acinetobacter and 20^th for Pseudomonas).

Two patients were found to harbor MRSA, acquired on days 2 and 5 of ICU hospitalization respectively. No VRE cases were recorded.

Gram-negative MDRO bacteremia occurred in 11 patients (11.2%): 5 were caused by A. baumannii, 4 by P. aeruginosa and 2 by K. pneumoniae. Four out of eleven bacteremic patients had received a graft from a donor subsequently found to be bacteremic on the day of procurement (3 cases of A. baumannii, 1 case of P. aeruginosa).

However, not all recipients of bacteremic donors developed bacteremia post transplantation: Four additional patients having received a heart from a donor with MDRO bacteremia (2 cases of A. baumannii and 2 cases of K. pneumoniae) did not develop bacteremia during the first month post transplantation. In total, eight patients received a heart from a MDRO bacteremic donor; half of these developed bacteremia by that pathogen. The characteristics and outcomes of these patients are summarized in Table 2. Nο bacteremia was recorded by gram-positive MDROs.

Table 2 Patients who received a heart from a bacteremic donor and developed multidrug-resistant organism bacteremia post-transplantation.

Patient	1	2	3	4
Diagnosis	DCM	DCM	DCM	DCM
Age (yr)	21	53	58	61
Sex	M	F	M	M
Device pre-transplant	BiVAD	IABP	LVAD	LVAD
Comorbidities	-	DM	-	CRF, vasculopathy
Admitted from (days of stay pre-transplant)	Ward (3 d)	Cardiac ICU (123 d)	Home	Ward (36 d)
Recipient MDRO colonization pre-transplant	No	No	No	No
MDRO donor bacteremia	Acinetobacter baumannii	Acinetobacter baumannii	Acinetobacter baumannii	Pseudomonas aeruginosa
Post-transplant complications	Re-exploration for hemorrhage	IABP (3 d)	IABP (9 d), re-exploration for hemorrhage	-
Septic shock	No	No	Yes	No
Post-transplant ICU stay	10 d	8 d	27 d (death)	11 d
Post-transplant total hospital stay	35 d	50 d	27 d (death)	34 d
Outcome (1 yr)	Fully functional at home	Fully functional at home	Death (day 27) due to thromboembolism/MOF	Partially dependent at home, frequent readmissions

BiVAD: Bi-ventricular assist device; CRF: Chronic renal failure; DCM: Dilated cardiomyopathy; IABP: Intra-aortic balloon pump; ICU: Intensive care unit; LVAD: Left ventricular assist device; MOF: Multiple organ failure; MDRO: Multidrug-resistant organism.

Outcomes

Median ICU stay for all patients was 10 d (range: 3-22), while the median total hospital stay post transplantation was 30 d (range: 6-198). In total, 11 patients died in ICU, of whom 9 during the first 30 d (median survival: 26 d, range: 6-60 d). One male patient died in the operating room. Two more patients died within one year, following a long course of frequent readmissions with poor functional status (survival: 190 and 337 d respectively). Early mortality (30 d), ICU mortality and late mortality (1-year) were higher in the MDRO group (1-year mortality: 28% vs 7.6%, log-rank-χ² = 7.34, Figure 2).

Figure 2 Post-transplantation 1-year survival according to multidrug-resistant organism status. 1-year survival was 72% in heart transplant recipients with multidrug-resistant organism (MDRO) presence (MDRO+) vs 92% in MDRO-free (MDRO-) recipients (P < 0.01). MDRO: Multidrug-resistant organism.

DISCUSSION

MDRO presence occurred frequently in heart transplant recipients and was associated with surgical or medical complications post-transplantation. MDRO presence had a negative impact on patient outcomes, with a longer ICU and hospital stay, as well as an increased early and late mortality rate.

In patients with Acinetobacter, the organism was already present upon transplantation in about half of the cases, while Pseudomonas and Klebsiella were ICU acquired. Pre-transplantation hospitalization was not shown to be a predictor of MDRO presence upon transplantation, i.e. hospitalization at the time of transplantation did not seem to impact MDRO prevalence, as compared to patients admitted from home. However, most advanced cardiac failure patients are intermittently admitted in hospital, a fact which may account for this finding[9].

Among comorbidities, peripheral vasculopathy prevailed, but only renal failure and higher BMI was significantly higher in MDRO patients.

Bacteremia occurred in about a third of patients with MDRO presence. Moreover, bacteremia cases (36%) were donor acquired. This is a remarkably high percentage, which could be attributed to a delay in decision making about brain death declaration and organ procurement, resulting in donor increased length of stay in the ICU, a MDRO burdened environment[10].

Particularities of MDRO prevalence

ICUs harbor a distinct pathogen ecosystem arising through ongoing antimicrobial selection pressure[11,12]. As compared to other acute hospital settings, the ICU endemicity pattern comprises MDROs, which carry genes entailing evolving resistance to advanced antimicrobial agents[13]. In referral hospitals and specialty ICUs, patient previous inter-hospital transfers are an additional source of MDR pathogen importation[14].

In Europe, a geographical north-to-south and west-to-east gradient of bacterial resistance exists, with higher rates observed in the southeast: The selected bacteria surveyed in this study are the most prevalent microorganisms in ICU environments in south-eastern Europe, where carbapenem-resistant gram-negative pathogens prevail[15]. These pathogens are accordingly deemed relevant for hospital infection surveillance purposes in Greece, in consonance with the National Action Plan for Antimicrobial Resistance. The Action Plan draws on the antimicrobial resistance reporting protocol issued by the European Center for Disease Control (ECDC)[8].

Acinetobacter baumannii, the prevailing MDRO in our study, is characterized by the capacity to acquire and harbor a battery of determinants of antimicrobial resistance and environmental persistence[16]. In low-prevalence ICU environments, e.g. in the north and west of Europe, Acinetobacter outbreaks are usually monoclonal and can be traced to an index transmission event; in contrast, high prevalence ICU environments, similar to those of south-eastern Europe, are marked by a diversity of polyclonal isolates and entail an increased complexity regarding the pathways of transmission[17]. In this setting, accelerated evolution of resistance via plasmid transfer between different isolates has been documented[18]. In a similar manner, mobile genetic elements, including plasmids or phages, play a critical role in Klebsiella ICU clusters, via horizontal transmission between polyclonal strains[19].

Remarks on MDRO colonization of heart transplant patients

Heart transplant candidates constitute a heterogenous patient group. Considering the most frequent underlying diagnoses, patients with dilated cardiomyopathy in our cohort were younger, while patients with ischemic cardiomyopathy were middle-aged. Both in this cohort and as a general rule, the latter group is relatively more burdened by comorbidities, such as hypertension, diabetes, peripheral vasculopathy and renal disease[20]. Patients re-transplanted for graft vasculopathy represent a small, but challenging subgroup[21].

Candidates for heart transplantation share numerous risk factors for MDRO acquisition pre-transplant. Notably, heart failure is a leading cause of hospitalization, while asymptomatic colonization of patients with MDRO is a recognized consequence of frequent admissions[22]. Further, a subgroup of heart transplantation candidates (about a third of patients in our study) is confined to the inpatient setting in a state of “dependent stability”, i.e. in need of continuous inotrope infusion (15.3%) and/or support with an IABP (14.3%). A small group of patients pertains to a hyper-acute state, such as ECMO support (one patient in our study).

A feature particular to heart transplantation candidates is the presence of VAD (55% in our study). VAD at the time of transplantation confers an increased risk of MDRO colonization, usually at the driveline entry point (5.5% of patients in our study)[23]. VAD entry point infection occasionally spreads across the driveline (mediastinitis)[24,25]. However, active VAD-related bacterial infection in the transplant recipient does not constitute a contraindication to transplantation[26]. Given that heart transplantation is never a scheduled procedure, bacteremia secondary to active device infection cannot always be accurately excluded at the time of transplantation. Further, weight-gain in patients on VAD is not infrequent. Among VAD patients subsequently transplanted, increased BMI was associated with post-transplantation complications[27].

Regarding the timing of MDRO acquisition, MDRO pathogens in our study were mainly acquired post-transplantation in the ICU, rather than already present at transplantation. Patients in the ICU become colonized via cross-contamination between patients, via direct or indirect contact due to environmental MDRO persistence, or via acquisition of resistant strains in the patient’s gut or skin following prolonged antimicrobial exposure[28,29].

Indeed, newly transplanted patients are commonly being treated with long courses of antimicrobials, even in the absence of complications or clinical indications of infection[30]. To address the use of antimicrobials out of proportion to infection prevalence, current guidelines recommend the discontinuation of prophylactic antimicrobials 24-48 h post transplantation[31]. Nevertheless, the familiarity of common practice prevails over best available evidence; it is notably difficult to de-implement customary treatments and practices devoid of an evidence-based foundation[32]. This effect is further enhanced by the fluctuating post-surgical physiological state of transplanted patients, generating uncertainty about whether antimicrobials may be safely discontinued or even de-escalated.

Infection control strategies and antimicrobial stewardship need to be constantly promoted in the ICU, to improve patient outcomes. However, even when relevant policies are in place, high workload may act as a hindrance to guideline implementation[33]. Our cardiac surgery ICU has a 100% occupancy, while the average nurse-to-patient ratio is 0.7, becoming 1 for transplanted patients. The steadily high bed occupancy rate entails an increased work volume per nurse, a possible impediment to optimal infection prevention practices[34].

Post-transplantation MDRO bacteremia

In the ICU, bacteremia arising from vascular catheter colonization is an important cause of patient destabilization[35]. MDRO bacteremia, in particular, is a recognized cause of sepsis in the ICU[36,37]. Sepsis-related mortality remains high in ICUs, although increasing attention is being given to the prompt recognition and control of septic episodes, and despite the availability of incrementally advanced and precise diagnostics[38].

Due to concerns for post-transplantation bacteremia, MDRO presence in the donor or the recipient was formerly listed as a contraindication for transplantation[39,40]. Indeed, MDRO bacteremias are mostly encountered in the early post-transplantation period, when transplanted patients recover from major surgery, while being treated with high dose immunosuppression[41]. In our unit, immunosuppressive treatment comprises ATG, corticosteroids, mycophenolate and tacrolimus.

Further, an increased propensity for bacterial translocation to the blood through the gut mucosa is noted in heart transplant patients, given that hemodynamic instability is common in the early post-transplantation period, even in the absence of graft dysfunction[42].

All post-transplantation bacteremia episodes in our study were caused by Gram-negative MDROs. Globally, the epidemiology of bacteremia in transplanted patients has shifted from Gram-positive to Gram-negative pathogens, matched with a rising emergence of resistant strains[43].

In this study MDRO acquisition preceded bacteremia episodes, a finding in accordance with other studies[44,45], although MDRO bacteremia without previous colonization has been previously reported[35]. Due to the limited number of post transplantation bacteremia episodes, the study was underpowered to support a valid analysis of the impact of MDRO bacteremia on patient outcomes.

Outcomes: Mortality and ICU length of stay

Whether the presence of MDROs in the ICU has a measurable impact on patient ICU mortality remains an unsettled question. Different studies report opposing results, both positive[43-45] and negative[46]. When matched immunocompetent patients serve as a control group, solid organ transplantation patients generally experience more bacteremia episodes during hospitalization, but mortality does not seem to be greater[47]. Therefore, while MDRO colonization was previously listed as a contra-indication for transplantation, presently, donor or recipient MRDO colonization no longer constitutes an exclusion criterion per se for organ procurement and allocation[48,49]. Rather, comprehensive recipient evaluation and effective interinstitutional communication channels are recommended, so that information about MDRO presence can be rapidly communicated and managed at the time of transplantation[50,51].

In our study, immediate post-transplantation mortality (within 10 days) was due to primary graft dysfunction and surgical complications. Indeed, immediate post transplantation mortality is generally not attributed to microbial causes[52]. However, ICU mortality, 30-d mortality and 1-year mortality were higher in patients harboring MDROs. Therefore, while heart transplant recipients with MDROs are not more likely to die when compared with matched non-transplanted ICU patients, they have a higher mortality rate when compared with MDRO-free transplant recipients. It is unclear whether this association represents a direct causal relation, given that MDRO presence may be either a consequence of medical and surgical complications of the transplantation procedure, or a marker of prolonged poor status, rather than a direct cause of death[53].

A longer ICU stay and total hospital stay were clearly correlated with MDRO presence. Given the detrimental impact of prolonged ICU stay on patient functional status and hospital resource allocation, it becomes clear that infection control practices need to be constantly revisited, so that patient MDRO colonization can be prevented. Teamwork is essential for infection prevention effectiveness, i.e. nurses daily inspecting indwelling devices (venous, arterial and urinary catheters, tracheal tubes etc.), physicians reassessing the need for such devices, and cleaning staff implementing surface disinfection. Antimicrobial stewardship support is decisive for discontinuation of antimicrobials as appropriate[54]. Best efforts for adequate ICU staffing and continuing staff education and support must be made by the hospital administration[55].

Strengths and limitations of the study, and future directions: Although this was a retrospective study, which constitutes a limitation per se, all transplanted patients were consecutively included in the study, and data completeness was 100% for all the variables studied. Despite the relatively low sample size, the study demonstrated a statistically significant association between MDRO presence and patient outcomes. However, the study was underpowered to support a valid analysis of infrequent events, such as receiving a heart from a donor subsequently proven to have been bacteremic at the time of organ procurement. In the future, these parameters need to be re-evaluated by pooling data from several transplantation centers.

Although the study was conducted in a single transplantation center, the pattern of resistance observed is representative of the prevailing resistance pattern in south-eastern Europe (high prevalence of carbapenemase producing MDROs/high level of resistance). Given the geographical epidemiological differences, the results may not be generalizable for institutions in the north-west of Europe, where the pathogen resistance pattern is characterized by a predominance of Gram-positive MDROs or by the prevalence of ESBL-producing Gram-negative MDROs with lower level of resistance[56].

Finally, in this study, bacterial isolation and identification was based on classic microbiological detection methods. In the future, novel sequencing techniques may provide more information on MDRO clonal spread and virulence characteristics. Different patient outcomes may be influenced by dissimilar virulence capacity among distinct isolates of the same pathogen[57,58]. Therefore, molecular pathogen data may clarify phenomena, such as the persistence of colonizing MDRO strains and their interaction with the transplanted host, as well as the way these phenomena determine transplantation outcomes.

CONCLUSION

This single center retrospective study showed that heart transplant recipients had a high incidence of MRDO presence with a clear predominance of Gram-negative carbapenemase-producing pathogens, a pattern characteristic of south-eastern Europe. MDROs were mainly ICU acquired during the early post-transplantation period, rather than already present at transplantation. Higher BMI and pre-existing renal failure were shown to be risk factors pre-transplant, while medical and surgical complications upon transplantation were shown to be risk factors of MDRO acquisition post-transplant. MDRO acquisition was associated with prolonged ICU and total hospital stay, as well as early and late post transplantation mortality.