Deep sternal wound infections are not uncommon after cardiac surgery, and their occurrence is reported at between 0.5% and 6.8%.1
Such infections are even more common after ventricular assist device (VAD) implantation and represent the Achilles heel of long-term VAD therapy.2,3 Infectious complications associated with left ventricular assist devices (LVADs) are extremely heterogeneous3 and carry an increased risk for morbidity. Left ventricular assist device pocket infections alone are associated with a 40–50% mortality rate and may also represent a relative contraindication to heart transplantation.3,4 Infection strongly increases the cost of care in these patients,5 as they often require prolonged hospitalization and admission to the intensive care unit (ICU).
The risk of postoperative mediastinitis is specifically exacerbated in VAD patients by the presence of the artificial surfaces of the device, which represent a suitable milieu for bacterial colonization and growth. For the same reason, once infection is established, it is more difficult to eradicate than in other cardiac surgery patients.
Aggressive surgical treatment is required in patients failing on medical therapy, but device exchange is extremely invasive and the outcome uncertain. A combined medical and surgical strategy, including aggressive debridement and use of vascularized and immune active omentum majus for coverage, may represent a valuable option to achieve full sterilization of the thorax.6 To date, a surgical strategy including transposition of the omentum majus has been reported anecdotally.3,7–9
At our institution, we have implemented a consistent surgical approach to treat mediastinitis in VAD patients in recent years. The aim of this study is to analyze the results of this surgical strategy including surgical debridement, jet lavage, vacuum assisted closure (VAC) treatment, and, finally, omentoplasty in VAD patients with refractory mediastinitis.
We retrospectively collected data on all VAD patients who received surgical treatment and plasty with the omentum majus because of mediastinitis at Deutsches Herzzentrum Berlin up to October 30, 2015. Under German law, retrospective analysis of an institution’s own data does not require ethics committee approval.
For the diagnosis of mediastinitis, we used the classification given in the guidelines provided by the International Society of Heart and Lung Transplantation.10 Key elements for the diagnosis were preoperative computed tomography, the examination of the wound itself, and intraoperative examination of the surgical field, the presence of secretions, bacteremia, and, of course, positive microbiological mediastinal samples. All baseline and preoperative data, the pathogens involved, and postoperative outcomes (including complications such as bleeding, renal failure requiring dialysis, sepsis requiring vasopressor therapy, and the length of ICU and hospital stay) were recorded from patients’ record charts.
Infection Management and Surgical Technique
All patients were consistently managed with high-standard medical care including antibiotics titrated to the antibiogram. The surgical approach consisted of debridement of mediastinum, jet lavage (Figure 1) with 3 L NaCl 0.9%, and VAC. All the patients underwent jet lavages and VAC therapy until three documented consecutive negative microbiological swaps were obtained. Normalization of infection parameters (white blood cells, C-reactive protein, procalcitonin) was concomitantly checked. Once sterilization of the mediastinum was finally achieved, omentoplasty was performed.
The surgical technique included transposition of the omentum majus (Figure 2) into the mediastinum with coverage of the outflow graft and intrapericardially placed pump with or without irrigation with saline for up to 4 weeks (Figure 3). The flap was created using one of the major vessels, most of the time half of the omentum, to cover the heart to the ascending aorta and apex if possible. A pectoralis flap was mobilized from the left and right side to cover the closed sternum in most cases.
During the perioperative period, oral anticoagulation was discontinued and intravenous unfractionated heparin was administered in all patients, with an activated partial thromboplastin time target of 50–60 seconds.
We did not analyze the incidence of mediastinitis in our population.
Data were anonymized before analysis and managed using Microsoft Excel 2007 (Microsoft Office 2007, Redmond, WA). Continuous data are reported as mean value ± standard deviation or as median (minimum value − maximum value), as appropriate.
A further analysis was performed to compare data of surviving patients with those of patients who died, to identify the baseline and perioperative variables possibly associated with hospital mortality. To this aim, Fisher’s exact test was performed for the categorical variables, and the Mann–Whitney U test for continuous variables.
Since 2008 and up to October 30, 2015, 845 patients have received an implantable continuous flow VAD at Deutsches Herzzentrum Berlin. HeartWare HVAD (HeartWare International, Framingham, MA) is the most frequently implanted device (587 patients: 552 as LVAD, 29 as biventricular assist device (BIVAD), and 6 as right ventricular assist device [RVAD]). The second most frequent device was HeartMate II LVAD (183 patients).
Omentoplasty after refractory infective mediastinitis was performed in 17 cases.
Sixteen of the patients (94%) were male, and mean age was 59 ± 13 years. The majority of patients had type 2 diabetes mellitus (65%), chronic renal failure (53%), and arterial hypertension (59%). Among the 17 patients with mediastinitis, nine patients (1.4% of total implants) had a HeartWare HVAD as LVAD, three patients (10% of total implants) had two HeartWare HVAD as BIVAD, four (2% of total implants) had a HeartMate II LVAD device, and one patient (0.5% of total implants) had a Berlin Heart Incor LVAD (INCOR, Berlin Heart GmbH, Berlin, Germany). The cause of heart failure was ischemic heart disease in 11 patients (65%) and nonischemic dilatative cardiomyopathy in six patients (35%). Intention to treat at the time of VAD implantation was destination therapy in eight patients (47%) and bridge to transplant in the other nine patients (53%).
Sixteen patients underwent sternotomy at time of VAD implantation, and one patient received implantation with a left lateral thoracotomy approach because of previous coronary artery bypass grafting operation. Two patients (12%) had a temporary RVAD with Levitronix Centrimag (Levitronix, Framingham, MA), and two (12%) had a temporarily open chest after surgery. Eight patients (47%) underwent rethoracotomy after VAD implantation because of bleeding or the need for pump exchange after an episode of pump thrombosis: two patients had one rethoracotomy, five patients had two re-thoracotomies, and one patient underwent two rethoracotomies and two mini rethoracotomies. Nine patients (53%) experienced early mediastinitis, i.e., postoperative mediastinitis after VAD implantation, and underwent omentoplasty during the same hospitalization. In the other eight patients (47%) mediastinitis occurred later, after hospital discharge. Among the nine patients who had early mediastinitis, two had a temporary RVAD and three BIVAD, two had a temporarily open chest, and three had postoperative bleeding.
Microorganisms isolated from the sternal wound are shown in Table 1: overall, Gram-positive Staphylococcus spp. were the predominant cause of infection. Eight patients had concomitant positive blood cultures.
Outcomes are presented in Table 2. Four cases of postoperative bleeding (24%) requiring surgical revision were recorded: three of intraabdominal and one of wound bleeding. Eight patients (47%) survived. The cause of death in the nine patients (53%) who died before hospital discharge were septic shock and multiorgan failure (MOF) in six cases, intracerebral bleeding in one case, pump thrombosis in one case, and sudden decrease of LVAD flow refractory to resuscitation maneuver in one case.
Among the eight patients who were discharged from hospital, four died within the first postoperative year because of intracranial bleeding in two cases and pulmonary embolism in one case, whereas for the last patient the cause of death was not available. One patient survived 6 years after omentum plasty and experienced a new driveline infection during the follow-up. The last three patients are now in the first year of follow-up after the operation with no signs of recurrence of infection. These patients are routinely screened for infection during the routine follow-up at our outpatient department.
Among patients who did not survive to hospital discharge, two had a BIVAD, two had support with a temporary RVAD, two had a temporarily open chest, and five had undergone rethoracotomy for bleeding.
Sixteen patients (94%) required ICU admission during hospitalization, and median ICU stay was 15 (182 max) days. Overall hospital stay was 21 (1–182) days.
The infection by Gram-positive pathogens or by Staphylococci was not associated with poor outcome. Development of renal failure requiring continuous venovenous hemofiltration and septic shock requiring inotropic support after surgery showed a statistically significant association with mortality (p = 0.009 and p = 0.05, respectively).
Our experience showed that the staged approach with final omentum plasty may be a useful tool for the management of deep sternal infections in VAD patients. Mediastinitis remains a severe complication, and despite aggressive treatment, the mortality rate of 53% is still high, but in accordance with that reported in the literature.4
The critical illness of these patients involves many aspects, beyond mediastinitis itself, which are interrelated and may influence each other in a negative manner and ultimately be responsible for the poor prognosis registered in this population. Indeed, the relationship between systemic inflammation, infection, bleeding, thromboembolic complication, and MOF is well known, and its effects are frequent in these patients. Namely, septic shock, circulatory failure, thromboembolic, and bleeding events were the cause of death in all the nine patients who died before hospital discharge.
Therefore, treatment of infective mediastinitis in this patient population should be started before MOF dominates the clinical situation. On the one hand, medical treatment is the first option and may succeed in containing the infection in some cases. On the other hand, early aggressive treatment is often needed to achieve full sterilization of the mediastinum, as the presence of the foreign surfaces of the device itself facilitates and promotes the proliferation of microorganisms. Severe pump infection may be controlled by device explantation if the cardiac function has fully recovered, or by pump exchange employing a different connection to the heart (e.g., left atrium instead of left ventricle) and descending instead of ascending aorta.11 Surgical device exchange is a major intervention including employment of cardiopulmonary bypass, and implanting the new pump in a nonsterile field. This approach may be too invasive in such a fragile population, but may still be useful in selected cases after the operating field has become sterile. For example, we recently reported a case of complete pump exchange in a patient with mediastinitis and apex involvement with loosening of the fixation ring.12 However, resolution of the mediastinal infection is also a key step in the case of pump exchange; the operative risk is nevertheless high, and the outcome uncertain.
Our strategy is aggressive and aimed at early and full resolution of the mediastinal infection, as it involves debridement of the infective site, jet lavages, and VAC therapy until sterilization of the mediastinum is achieved, and, finally, transposition of the immune active omentum majus, which should act as a biological barrier to further microbial proliferation. At the same time, it is much less invasive than pump exchange. The main advantage of our treatment is that it follows the gold standard rules for treatment of infected wounds: debridement, cleaning, and secondary closure with well-vascularized tissue with continuous irrigation of the infected wound,3,7–9,13–16 in a rational multistep approach. Our experience proved to be useful and resulted in a good rate of hospital discharge, with an acceptable complication rate; indeed, we reported a bleeding rate of 24% in an extremely fragile population that requires continuous anticoagulation therapy. Furthermore, the rate of recurrence of infection is low.
Although this therapy is effective, clinical results are jeopardized by concurrent severe elements, which dominate the clinical scenario, as mentioned above, independently of the operation itself: septic shock, MOF, thromboembolic complications. For the same reason, open treatment for mediastinitis was not performed for patients deemed too sick at baseline (i.e., with sepsis and multiorgan damage), who would not have benefited from this approach. Although no preoperative or baseline data were found to be associated with mortality in our population, it should be noted that 47% of our patients indeed had previous rethoracotomy because of bleeding.
On the basis of the provided data, it is clear that strategies aimed at reducing the risk of bleeding, with the related hemodynamic instability, and the risk of renal failure are needed. A possible approach in this sense could be a reduction of the anticoagulation degree in the perioperative period (i.e., during surgery and on the first postoperative day).
Another possible surgical strategy to treat this complication in selected cases involves the rerouting of the outflow vascular prosthesis to the descending aorta through the left thoracic cavity. Indeed, this strategy proved to be successful in one patient, as reported in the literature,17 but we have no experience with that.
The role of such alternative techniques should not be underestimated, especially in those patients who are on the waiting list for heart transplantation. Indeed, no transplantation can be scheduled until the device is infected and until mediastinal cultures are positive.
At the same time, clinical experience suggests that omentoplasty for previous mediastinitis should not be considered a relevant contraindication for cardiac transplantation,18 although data on VAD patients who have undergone omentoplasty and then transplantation are lacking.
This novel treatment was needed only in a small minority of patients treated with VAD at our institution. Nevertheless, we observed that patients with BVAD are probably more prone to mediastinitis, probably as they undergo more implants. No conclusion can be extrapolated from our limited experience, but both prevention and treatment of mediastinitis in patients with BIVAD may warrant specific consideration. In addition, the surgical treatment of mediastinitis with pump exchange in these patients is further complicated by the presence of two devices, and this may encourage the development of alternative approaches.
This study presents some limitations. First, it is observational and the sample size is small; furthermore, a historical control group is not available. We did not present the incidence of mediastinitis in our study, as we could not retrieve these data. Patients who died before or during surgical treatment (e.g., before omentoplasty) were not included in the analysis, and the incidence of omentoplasty reflects neither the incidence or severity of the mediastinitis nor its relationship to the different devices. We also cannot show data suggesting that our approach has led to an improved outcome in such patients compared with that of the patients who experienced mediastinitis before the development of our strategy. Furthermore, even though we were able to identify predictors of mortality, further analysis, e.g., the identification of preoperative predictors of postoperative need for hemofiltration, was not possible because of the limited sample size.
In conclusion, omentoplasty represents a valuable treatment option for mediastinitis in patients with VAD in otherwise dead-end situations. The outcome of these patients is often poor because of the intrinsic characteristics of the population itself. However, approximately half of the patients survive the acute phase, and in these cases the procedural success is good, with acceptable complication rate and low recurrence of infection.
The authors thank Julia Stein for statistical support and Anne Gale for editorial assistance.
1. Cotogni P, Barbero C, Rinaldi MDeep sternal wound infection after cardiac surgery: Evidences and controversies.World J Crit Care Med20154265273
2. Topkara VK, Kondareddy S, Malik F, et al.Infectious complications in patients with left ventricular assist device: Etiology and outcomes in the continuous-flow era.Ann Thorac Surg20109012701277
3. Tong MZ, Smedira NG, Soltesz EG, et al.Outcomes of heart transplant after left ventricular assist device specific and related infection.Ann Thorac Surg201510012921297
4. Kretlow J, Allred BD, Marks JD, et al.Use of antibiotic beads in the treatment of left ventricular assist device infections.J Heart Lung Transplant201332S276
5. Slaughter MS, Bostic R, Tong K, Russo M, Rogers JGTemporal changes in hospital costs for left ventricular assist device implantation.J Card Surg201126535541
6. Nelson JA, Shaked O, Fischer JP, et al.Complex wound management in ventricular assist device (VAD) patients: The role of aggressive debridement and vascularized soft tissue coverage.Ann Plast Surg201473suppl 2S165S170
7. Kimura M, Nishimura T, Kinoshita O, et al.Successful treatment of pump pocket infection after left ventricular assist device implantation by negative pressure wound therapy and omental transposition.Ann Thorac Cardiovasc Surg201420suppl842845
8. Hui C, Ooi A, Tan TE, Tan BKRectus sparing approach to left ventricular assist device exchange and use of the omental flap for coverage.J Plast Reconstr Aesthet Surg201568278280
9. Rodriguez LE, Bruckner BA, Loebe MOmental flap for treatment of mediastinitis post-left ventricular assist device implantation.Artif Organs20133710811082
10. Hannan MM, Husain S, Mattner F, et al.International Society for Heart and Lung TransplantationWorking formulation for the standardization of definitions of infections in patients using ventricular assist devices.J Heart Lung Transplant201130375384
11. Levy DT, Guo Y, Simkins J, et al.Left ventricular assist device exchange for persistent infection: A case series and review of the literature.Transpl Infect Dis201416453460
12. Scandroglio AM, Potapov E, Pieri M, Arlt G, Falk V, Krabatsch TThree-stage treatment of late mediastinitis after implantation of left ventricular assist device.Med Intensiva2016pii: S0210-5691(16)00043-7
13. Kaye AE, Kaye AJ, Pahk B, McKenna ML, Low DWSternal wound reconstruction: Management in different cardiac populations.Ann Plast Surg201064658666
14. Kawata M, Nishimura T, Hoshino Y, et al.Negative pressure wound therapy for left ventricular assist device-related mediastinitis: Two case reports.J Artif Organs201114159162
15. Turek D, Netuka I, Malý J, et al.[Management of deep device related infection in patient implanted with mechanical cardiac support device using V.A.C. system].Rozhl Chir200988693696
16. Kinoshita O, Nishimura T, Kawata M, Ando M, Kyo S, Ono MVacuum-assisted closure with Safetac technology for mediastinitis in patients with a ventricular assist device.J Artif Organs201013126128
17. Kurihara C, Nishimura T, Kinoshita O, et al.Successful treatment of mediastinitis after ventricular assist device implantation with rerouting of the outflow vascular prosthesis.J Artif Organs201114155158
18. Castedo E, Cañas A, Varela A, Ugarte JDoes omentoplasty
preclude cardiac retransplantation?Chest200112014251426