Introduction

Among common medical complications of solid organ transplant (SOT), infections remain a major cause of morbidity and mortality [1]. Most infections arise in the SOT recipient a variable time after transplant and are due to opportunistic pathogens [2]. However, the transplantation procedure itself may represent an efficient way of nosocomial infection transmission, usually involving pathogens with a multidrug resistance (MDR) pattern [3]. Infections due to MDR Gram-negative rods are specifically recognized as a risk factor for increased mortality in this setting [4–11].

Organ donors are critically ill subjects staying in high dependency units and as such at risk for nosocomial infections. Thus, great interest has been given to donor-derived infections [12]. However, recipients of life-saving organs may also be critically ill patients, staying long term in intensive care units (ICU) waiting for a life-saving SOT, and thus at high risk for nosocomial infections themselves. Notwithstanding, little if any data are available concerning the incidence, management and outcome of recipient SOT with a pre-existing nosocomial infections.

Extensively drug-resistant (XDR) Acinetobacter baumannii (Acb) [13] is a major etiologic agent of nosocomial infections associated with high mortality in critically ill patients. It shows resistance to almost all classes of antimicrobials, including carbapenems [15]. The limited available data on the outcome of XDR Acb infections in organ transplant recipients suggest an increased mortality compared with non-transplant patients [14]. In this report, we describe our experience with the complex management of a recipient-born XDR Acb infection already present shortly after heart transplant. It is important to stress the presence of Acb colonization before heart transplant.

Case report

A 60-year-old male, with a long term history of hypertension and diabetes mellitus on oral treatment, was admitted to our hospital on July 1, 2013, because of acute myocardial infarction. Despite percutaneous coronary interventions and supportive care, the patient rapidly developed refractory heart failure, requiring aortic counterpulsation. After 15 days of ICU stay, due to persistence of pulmonary edema, the patient underwent emergency heart transplantation. At the time of transplantation, microbiological cultures were negative except for the culture of the tip of the cannula through which the aortic counterpulsator had been in place, obtained just before transplantation that turned out positive for XDR Acinetobacter baumannii on the fourth post-transplant day. The patient was started that day on colistin, 2 MU tid. The immediate post-transplant course was uneventful, and there was no fever, leucocytosis or other evidence of systemic infection. A surveillance blood culture was negative. After 14 days, colistin was withdrawn. The patient was discharged home asymptomatic 1 week later on a standard cyclosporine A, mycophenolic acid and prednisone regimen.

Ten days after discharge, the patient experienced low grade fever with back pain. At the scheduled follow up, a single blood culture and a urine culture were obtained and turned out negative. Due to persistence of fever, the patient was admitted to our unit (see Fig. 1). On entry, two blood cultures and one urine culture were obtained and all turned out positive for XDR Acb with meropenem MIC of 64 mg/l. The antimicrobial susceptibility profile was characterized by MICs (in mg/l) of 0.5(S) for colistin, 2(S) for tigecycline, 256(R) for fosfomycin, 32(R) for gentamicin, 8(R) for ciprofloxacin and 320(R) for cotrimoxazole. Patient conditions rapidly deteriorated with sepsis syndrome (leukocytes 14,550 cells/μL; lactates 2.2 mmol/L; arterial blood pressure 110/70 mmHg) and acute kidney failure documented by a glomerular filtration rate (GFR) drop to 32 ml/min. Colistin was restarted with a loading dose of 6 MU followed by 3 MU bid iv. After 4 days of treatment, fever peaked at 40.2 °C, clinical conditions worsened and blood cultures remained persistently positive for XDR Acb. Hence, despite renal dysfunction, we increased colistin dose to 4.5 MU bid and added to the antibiotic treatment teicoplanin, at the dose of 800 mg q12 h for the first 3 doses, then 800 mg q24 h. This decision was based on recent in vitro data suggesting a synergistic effect of this combination against XDR Acb [16]. Moreover, due to evidence on the isolated strains of in vitro synergism between colistin and tigecycline, the latter was added to therapy at the dose of 50 mg bid iv following a 100 mg loading dose. After 4 more days, blood cultures were still positive for XDR Acb, with a very short time to positivity. Due to a progressive drop in platelet count with rise of D-dimers, possibly due to disseminated intravascular coagulation, continuous infusion of fresh frozen plasma at the dose of 10 ml/kg/day was initiated. Subsequently, a severe microangiopathic anemia developed with haptoglobin depletion, LDH rise, presence of schistocytes on blood smear and drop of platelets to 5000/μL. A moderate activity of anti-ADAMTS13 antibodies was detected, raising the suspicion of thrombotic thrombocitopenic purpura possibly related to sepsis or cyclosporine A treatment. For this reason, daily total plasma exchange associated with red blood cell transfusions was performed. Immunosuppressive treatment was changed, replacing cyclosporine A with everolimus in association to mycophenolic acid. On the 10th day of therapy, blood cultures were all still positive for XDR Acb. Blood purification with Toraymixyn® treatment was started, without any measurable effect on plasma endotoxin levels or microbiological eradication. Colistin-induced progressive kidney injury occurred with a GFR nadir down to 21 mL/min, proteinuria, hypo-stenuria and marked polyuria. On the 14th day of treatment, blood cultures were still positive for the same XRD Acb, showing an antimicrobial susceptibility profile identical to the prior isolates and an MIC for colistin always equal to 0.5 mg/l. Treatment was continued without further changes and on the 17th day of antibiotic treatment we obtained the first negative blood culture. Colistin therapy was continued for further 15 days (Fig. 1). With sepsis control, thrombocytopenia slowly abated. After colistin withdrawal, renal function also improved, allowing patient discharge after 35 days of hospitalization in stable clinical conditions. Patient is alive and well at 9-month follow-up.

Fig. 1
figure 1

Graphic illustration of the patient clinical course, including fever pattern, blood culture results, antimicrobial and immune suppressive therapies received, and the kinetics of serum creatinine and platelet count

Full size image

Discussion and review of the literature

The present clinical case, while describing the burden of performing a transplant in a recipient with an ongoing XDR Gram-negative infection, provides several clues for discussion. First, it shows that organ transplant candidates staying in ICU should be very carefully monitored for ensuing nosocomial superinfections, irrespective of the length of ICU stay. Indeed, hospitalization time as short as 2 days was shown to be enough to acquire an MDR nosocomial pathogen [2]. A very low threshold for treatment of microorganisms isolated solely from removed intravascular devices should be used in this setting, as infection by MDR Gram-negative bacilli, including XDR Acb, cause high morbidity and poor outcomes.

The most likely initial source of Acb infection was healthcare related contamination of the aortic counterpulsator cannula. Despite of its removal, Acb likely persisted within the intravascular compartment, in the face of appropriate Colistin coverage started 4 days later. We were surprised to observe recurrence of XDR Acb despite recovery of this microorganism in a single, removed, catheter introducer tip. There was neither evidence of subsequent nosocomial transmission, nor other Acb-positive cultures or concurrently admitted patients colonized or infected by Acb. A major role could have been played by the high intensity immune suppression started shortly after the alleged source of bacteremia had been removed. Immune suppression likely played a role also in the persistence of bacteremia. Because of colistin failure to clear Acb from blood, we added high dose teicoplanin to the treatment regimen. This was based on recent literature data showing a synergistic in vitro effect of glycopeptides and colistin [16–18], and also to provide empirical anti-Gram-positive coverage. Under this combination treatment, blood cultures eventually cleared. The role of teicoplanin is unclear in this case and we believe additional reports would be useful before this combination is further studied in clinical trials.

We were challenged by the possibility to add rifampicin to colistin in this patient. Indeed, our group recently completed a randomized clinical trial that showed the superiority of colistin–rifampicin combination versus colistin alone in terms of XDR Acb microbiological eradication from the primary infection site [19]. In this study, no improvement in 30-day survival was observed with the addition of rifampicin. In cases like the one presented here, microbiological eradication would have been a major goal of therapy to improve the initial transplant outcome and minimize the risk of infection recurrence. However, to maintain adequate plasma levels of cyclosporine A and avoid further biliary dysfunction shortly after transplant, we withheld rifampicin treatment in this case, a further suggestion that the most appropriate and well-targeted use of rifampicin in XDR Acb remains to be defined.

Literature data regarding MDR/XDR Gram-negative infections in the post-transplant setting are limited. Table 1 presents a synopsis of all cases of MDR/XDR Gram-negative infections retrieved from our literature review. As shown, most cases were due to Pseudomonas aeruginosa, followed by Klebsiella pneumoniae and Acb, presented as bloodstream infection or arteritis or pneumonia, and occurred in recipients of kidney or liver transplant. The majority of infections were donor derived with a substantial number of cases of nosocomial acquisition unrelated with the transplantation procedure. Apparently, there were no reports of transplantation procedures in previously infected recipients. Overall, the in-hospital death rate was around 40 %, but most patients who were discharged from hospital were still alive at 1-year follow up.

Table 1 Synopsis of transplant-related, XDR/MDR Gram-negative infection cases reported in the literature, current as of June 2014
Full size table

Conclusions

To date, no data are available concerning recipient-born XDR Gram-negative infections in patients undergoing heart transplant. There are no recommendations regarding the optimal therapeutic management in this setting and prognosis remains elusive. The effect of an XDR infection diagnosis on the decision making of wait-listed transplant candidates deserves further evaluation. Clinical observations like ours are needed to provide some evidence to support physicians in the optimal management of these difficult-to-manage conditions.