Systematic approach to diagnosis and management of infected prosthetic grafts in the proximal aorta

Abstract Objectives Management of infected prosthetic aortic grafts in the ascending and or root is complex and multifaceted. We report our diagnostic pathway, management and outcomes, identifying successful strategies. Methods This was a retrospective, single center, observational study. Consecutive patients who underwent management of infected aortic grafts in the ascending and/or root at our institution between October 1998 and December 2019 were included. The main outcome measures were: discharge from hospital alive with at least 1 year survival, operative mortality and success of primary treatment strategy. Results Twenty‐six patients presented with infection of proximal aortic grafts and were managed through a number of strategies with an overall hospital‐survival of 81% and 1 year survival of 69%. Twenty of them ultimately underwent redo surgery with 25% operative mortality (within 24 h of surgery). Five patients underwent washout and irrigation of which two were successfully treated and cured with adjunctive antibiotics and two went on to have staged explant and definitive surgery. Interval between surgery and infection was 42.5 ± 35.8 months. All patients had at least one major criterion and three minor criterions with no diagnostic uncertainty. The commonest primary strategy was 3a (definitive surgery), (13/26, 50%). Conclusions Adopting a systematic and flexible patient specific approach to the diagnosis and management of patients with proximal aortic graft infections results in reasonable overall 1 year survival. In the majority of patients surgery is ultimately required in an attempt to achieve a curative treatment; however this comes with high operative mortality risk.


| INTRODUCTION
Patients with infected aortic root present a challenge to aortic surgeons as they are not easily eradicated and this gets more complicated when there is an infected prosthetic graft, such as those with previous Bentall root replacement with aortic valve and coronary buttons reimplantation. They can be of a diagnostic challenge and present a complex scenario where careful planning and intervention should be approached through multi-disciplinary team approach. 1 Initial treatment with antibiotics can help at early stages and some of these patients may not require surgical intervention; however the ultimate and definitive treatment remains through high-risk surgical re-exploration. 2,3 Several hospitals have published small series of such high-risk patients including management and outcomes; however, even in major Aortic centers, mortality outcomes are poor and current reported rates are varying between 14% and 55% depending on strategy, timing and approach. [2][3][4][5] There remains a diagnostic challenge as well as unclear management strategies with little evidence base. We describe a contemporary and pragmatic approach to diagnosis and the management approach developed at Liverpool Heart and Chest Hospital.

| METHODS
We performed a retrospective observational review of our entire practice between October 1998 and December 2019 during which records exist. This study was registered in our institution as a service review, therefore ethical approval and informed consent were not deemed necessary. All patients presenting with infected prosthetic proximal aortic grafts were included. We excluded patients with potential early prosthetic graft infection during the index admission and those sent home on antibiotics following primary surgery. A separate prospective database was collected of patients treated medically. Data for all interventions performed in our center is systematically reported to the National Outcomes for Cardiovascular Outcomes Research and our electronic records have been designed to collect relevant data on all patients undergoing aortic surgical procedures. Data was extracted from this database and included demographics, comorbidity, anatomical and pathological features of aneurysms, morbidity and mortality.
The primary outcome measure of the study was in-hospital mortality. Secondary outcome measures were success of 1 year survival and success of primary treatment strategy.

| Diagnostic approach
In diagnosing proximal aortic graft infections we have used two approaches dependent on whether the patient has an isolated proximal graft or proximal graft with prosthetic aortic valve, either aortic valve replacement (AVR) and ascending graft or AVR as part of an aortic root replacement with coronary buttons. Our diagnostic approach has evolved over the duration of the study into a contemporary approach using a combination of Dukes criteria 6 for diagnosing valve endocarditis and an in-house modification of the management of aortic graft infection collaboration (MAGIC) guidelines for diagnosing vascular graft infections. 7

| Isolated graft infections
The management of aortic graft infections collaboration team has set out criteria for diagnosing vascular graft infections. 7 The approach is

| Proximal graft and prosthetic valve infections
Dukes modified criteria 6 was used to diagnose endocarditis of the prosthetic valve, with the features described within the modified MAGIC criteria describe above, as additional major criteria, in diagnosing infection of a valved conduit ( Figure 1).

| Management approach
Our approach to the management of patients with infected proximal aortic grafts is influenced by several aspects of the patient (

| Generality of surgical methods
Patients undergoing redo surgery were treated in a variety of ways in terms of redo-sternotomy, bypass and operative procedures which commonly involved complex root reconstructions with a variety of prosthetic roots including homografts and Cabrol reconstructions.

| RESULTS
We performed a total of 2079 aortic procedures via sternotomy between the period of 1998-2019. Of these, 173 were redo aortic procedures and 20 were for infected proximal aortic grafts. A total of 26 patients presented with proximal graft infections and were treated accordingly.

| Demographics
The mean age of the cohort was 57 ± 11 years. For patients with a primary non-surgical strategy median age was 62.4 years and for those with a primary operative strategy was 56.4 years.

| Diagnosis
All patients presented with signs and symptoms of sepsis including general malaise and at least one major and three minor criteria

| Choosing a strategy: what works?
There are key features in the presentation (Figure 2) that help guide us in choosing a strategy ( Figure 3). As stated above, antibiotics are the main stay. The four principle drivers of strategy are patient age, frailty, structural pathology and sternal wounds. We attempted T A B L E 1 Summary of our cohort demographics, microbiology, results and outcomes antibiotics as a single modality approach in six patients. Treatment with antibiotics in isolation was chosen as a strategy with two groups of patients. First, in fit and healthy individuals with no hard indications for surgery and with the intentional outcome of successful treatment with antibiotics stopped. The second group was characterized by patients who were elderly, frail, co-morbid, with and without structural defects. The intention in this group was discharge home, either for palliation or long term management on antibiotics.
Interestingly even some patients with a root abscess or pseuodoaneurysm were successfully managed into a sterile chronic state.
Three of these six patients went on to have surgery (3a) and three continued on antibiotics treatment and were alive a year later. This success is in broad agreement with other published series.
A number of surgical based strategies were adopted dependent on presentation, either re-sternotomy and drainage of collections as a definitive strategy, as a staged bridge to definitive explant of infected material, or re sternotomy and explant as a primary strategy. In summary the key message in our data is that starting with strategy 1 (antibiotics), for any patient other than palliation, will be unsuccessful as a definitive treatment in 75% of the cohort. Irrigation as an adjunct to antibiotics and definitive treatment (3a) was successful in just two patients out of five, eventually coming off antibiotics all together however as an adjunct to further surgery (3b), the strategy worked in 2. Primary graft explant (3a) was the commonest approach and the commonest default after failed other strategies.
Our experience suggests we can attempt to treat sepsis, drain collections and manage wounds but in the end most patients will end up with explant of the infected prosthesis. These strategies are not binary decisions but often evolve over time.

| Postoperative management
Patients have routine predischarge CT aorta and echocardiography to act as baseline imaging. Antibiotic administration regime and duration are based on the organism, intraoperative findings and clinical progress. As a routine patients get between 6 and 12 weeks postoperative parenteral therapy with weekly c-reactive levels (C-reactive protein [CRP]) and further imaging at the end point with a decision on further treatment. This is in disagreement to recent data suggesting that in patients with endocarditis, patients only require two weeks antibiotics following their last positive blood culture. 13 Weekly CRPs are maintained for 6 weeks following cessation of antibiotics. Finally, such patients remain under lifelong surveillance. 14