Outcomes of conservative treatment for thoracic vascular graft infections

Abstract

OBJECTIVES

Thoracic vascular graft infections are devastating complications after aortic surgery, entailing high mortality. The gold standard treatment combines excisional surgery and antimicrobial therapy, but patients deemed inoperable might benefit from a conservative approach. Outcomes of patients treated only with antimicrobial agents without reoperative surgery are scanty. We aim to describe patients’ characteristics and outcomes using an antibiotic-only strategy without thorough debridement.

METHODS

Retrospectively collected data from a prospective cohort in a tertiary centre. Descriptive analysis for baseline characteristics and Kaplan–Meier estimates for survival were performed.

RESULTS

From November 2012 to December 2022, 66 patients were identified with aortic root, ascending aortic and aortic arch graft infections. Of these, 44 received an antibiotic-only strategy or in combination with selective debridement after achieving multidisciplinary consensus. Median follow-up was 4.8 years [interquartile range (IQR) 1.7–6.1], and cumulative survival was 82.9% (CI 95%, 69.7–96.1). Streptococcus spp were the most common isolated microorganisms.

CONCLUSIONS

In selected cases, a conservative approach with antibiotics only or in combination with selective debridement showed acceptable results at follow-up, suggesting a valuable therapy option for this cohort of patients.

Thoracic vascular graft infections (TVGIs) are devastating complications reported in 1–4% after aortic surgery [1–4].

GRAPHICAL ABSTRACT

INTRODUCTION

Thoracic vascular graft infections (TVGIs) are devastating complications reported in 1–4% after aortic surgery [1–4]. These have been observed since the 60s after the introduction of prosthetic grafts in treating aortic diseases in the late 1950s as an alternative to homografts [5]. Ever since it was identified as a complex pathology requiring multidisciplinary management [6, 7]. The gold standard of treatment consists of reoperative surgery on cardiopulmonary bypass with prosthesis removal and at least 6 weeks of antimicrobial therapy, depending on the extent of surgical debridement [6, 8–11]. Even though advances in imaging techniques and isolation of microorganisms have been made over the last decades, leading to faster identification and diagnosis, morbidity and mortality rates remain high. Outcomes following reoperative surgery with complete prosthesis removal and antimicrobial treatment remain to be improved, with mortality rates around 30% [12]. This can be attributed to an older and more comorbid population as well as to the aggressiveness of the disease and radical surgery. Recently, more conservative approaches have emerged to deal with this pathology: in high-risk or inoperable patients it has been suggested, the combination of selective debridement of the infected prosthesis and the use of omental or muscle pedicle flaps followed by antimicrobial therapy [13–16] might be suitable. Since the infected prosthesis is left in situ, patients have to be meticulously monitored to prevent reinfection or relapse. On the other hand, other authors advocate for a conservative treatment approach without reoperative surgery in patients rejected for reoperation with acceptable results [17]. However, this approach and its results have not gained widespread acceptance.

Overall, the literature addressing TVGI and its outcomes is scanty, with studies reporting results on rather small population of patients. This study aimed to investigate the characteristics and outcomes of patients with aortic root, ascending aortic and aortic arch graft infections treated with a conservative approach without reoperative surgery on cardiopulmonary bypass.

MATERIALS AND METHODS

Study design

This is a retrospective study of prospectively collected data from an ongoing cohort, The Vascular Graft Infections Cohort VASGRA (Clinical Trials registration no. NCT01821664), in a tertiary care institution [18].

Inclusion and exclusion criteria

Patients aged 18 years or older with infections of the aortic root, ascending aorta or aortic arch graft were included, while those with peripheral, endovascular or descending aortic graft infections were excluded.

Data retrieval

Patients’ demographics and characteristics, index operation and indication, operative reports, microbiological findings and outcomes were collected from the Institutional electronic medical record system.

Definitions

TVGI was classified into 2 groups depending on the timing of diagnosis: infection within 4 months after prosthesis implantation were categorized as early infections, whereas those occurring after 4 months were defined as late infections [6]. Differentiation between early and late infections is important, since different microbiological agents can cause one or another infection, altering antimicrobial treatment.

Relapse was defined as recurrent infection with the same microorganism and reinfection as infection with a different microorganism usually more than 6 months after the initial episode and after antimicrobial treatment discontinuation [19, 20].

Cure was defined as a healed infection without clinical, radiological and laboratory signs of inflammation or infection on follow-up consultations.

Postoperative infection was defined as any infection requiring antibiotic treatment after the initial or index surgery, and prior to the episode or diagnosis/suspicion of TVGI.

Treatment modalities

Reoperative surgery and thorough debridement were defined as resternotomy and complete removal of the infected prosthesis on cardiopulmonary bypass.

Conservative treatment was defined as either an antibiotics-only treatment or a combination of antibiotics treatment and selective debridement. The latter included resternotomy and antibiotic irrigation with or without negative-pressure wound therapy.

Indication for surgery

Patients were evaluated for surgery using the available Clinical Practice Guidelines [6, 7, 19, 20].

Follow-up and antimicrobial treatment and discontinuation during follow-up

Follow-up data were collected through our outpatient clinic by telephone calls with the patients, their general practitioners or referral doctors. Patients were evaluated with clinical visits and serial blood sampling. All patients underwent at least one ¹⁸F-fluorodeoxyglucose (FDG)-positron emission tomography (PET)/computed tomography (CT) scans (¹⁸F-FDG-PET/CT) during follow-up to evaluate whether antimicrobial treatment could be stopped or not. This was made by comparing baseline (SUVmax and uptake pattern) with follow-up values and by taking into account clinical examination and laboratory/inflammatory findings.

Antimicrobial treatment was administered in accordance with current guidelines and local epidemiological trends [11]. Decision to discontinue treatment during follow-up was guided by clinical evaluations, laboratory findings and ¹⁸F-FDG-PET/CT scans [21].

Diagnostic of infection

The Management of Aortic Graft Infection Collaboration (MAGIC) criteria and the 2023 Duke-ISCVID criteria for the diagnosis of infective endocarditis (IE) were applied, recognizing that TVGI share similarities with IE and are managed by a multidisciplinary team [6, 19, 20, 22]. Diagnostic evaluation included clinical examination, blood cultures, inflammatory markers, transthoracic or transoesophageal echocardiography, computed tomography and ¹⁸F-FDG-PET/CT scans. Infections were either categorized as ‘confirmed’ or ‘suspected’.

Institutional team approach

The institutional Endocarditis and Cardiovascular Infections Team (ET) evaluated and treated all patients after achieving a multidisciplinary consensus [23]. All patients were first treated with antibiotics and were re-evaluated weekly for either following a reoperative or a conservative approach. Patients receiving antibiotics considered unfit for surgery, those who declined surgical treatment or those with suspected infections were managed with either an antibiotic-only approach or with selective debridement and negative-pressure wound therapy [24].

Outcomes of interest

Outcomes of interest were follow-up survival, reintervention, relapse and reinfection rates.

Statistical analysis

SPSS Software Version 29 (SPSS, Inc., Chicago, Illinois, USA) was used for statistical analyses. Qualitative variables were expressed as numbers and percentages. Quantitative variables were expressed as median and interquartile range (IQR). Kaplan–Meier estimates were calculated for survival. Follow-up time was reported using reverse Kaplan–Meier survival curve. Re-intervention and reinfection were analysed in R version 4.3.1 [R Core Team (2023). R: A Language and Environment for Statistical Computing_. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/] using death as a competing event. Cumulative incidence function from the package cmprsk was used to estimate corresponding probabilities.

Ethics

The local Ethics Committee approved the study (KEK-ZH-Nr. 2012–0583, PB PB_2016-01320). Written informed consent was obtained from all participants.

RESULTS

From November 2012 until December 2022, 66 patients had aortic root, ascending aortic or aortic arch graft infections, and 44 were managed conservatively without reoperative surgery on cardiopulmonary bypass. Of these 44 patients, 63.6% (28/44) received antibiotic-only treatment, while 36.3% (16/44) were treated with a combination of antibiotics, resternotomy and mediastinal irrigation with or without negative-pressure wound therapy.

The cohort was predominantly male (75%; 33/44) with a median age of 64.6 years (IQR 59.1–71.2). The median time to diagnosis was 11 months (IQR 1.2–56.2) with most patients classified as late infections (63.6%; 28/44). At presentation, 63.6% (28/44) had a fever, and the median Charlson comorbidity index was 4 (IQR 2–6) (Table 1). The median hospital length of stay was 26.5 days (IQR 20.5–43.8), and the in-hospital mortality was 2.2% (1/44). On autopsy, the cause of death of this patient was identified as septic shock and multiorgan dysfunction.

Table 1:

Baseline characteristics

	Conservative, n = 44	Surgery, n = 22
Age, median year (IQR)	64.6 (59.1–71.2)	62.5 (55.6–72.7)
Sex (male), n (%)	33 (75)	21 (95.5)
BMI, median kg/m2	25.9 (22.8–27.2)	27.4 (24.9–29.9)
Arterial hypertension,a  n (%)	31 (70.5)	18 (81.8)
Diabetes mellitus, n (%)	3 (6.8)	5 (22.7)
Coronary artery disease, n (%)	10 (22.7)	3 (13.6)
Congestive heart failure, n (%)	9 (22.5)	6 (27.3)
Atrial fibrillation, n (%)	22 (50)	7 (31.8)
Chronic obstructive pulmonary disease, n (%)	3 (6.8)	1 (4.5)
Cerebrovascular disease, n (%)	15 (34.1)	6 (27.3)
Peripheral vascular disease, n (%)	1 (2.3)	/
Chronic kidney disease, n (%)	10 (22.7)	3 (13.6)
Malignancy, n (%)	10 (22.7)	3 (13.6)
Metastasis, n (%)	3 (6.8)	1 (4.5)
Connective tissue disorder, n (%)	4 (9.1)	2 (9)
Charlson comorbidity index, median (IQR)	4 (2–6)	1 (0–4)
NYHA III/IV, n (%)	9 (20.5)	7 (31.8)
LVEF,b median % (IQR)	55 (44.5–60)	54.5 (46.7–70)
Creatinine, median mmol/l (IQR)	102 (84.7–135.7)	97.5 (85.5–120.5)
Fever, n (%)	28 (63.6)	15 (68.2)
Temperaturec median grade Celsius (IQR)	37.2 (36.6–38)	37 (36.7–37.4)
Pulsee, median beats/min (IQR)e	83 (72.7–96)	73 (64–81.2)
CRP, median mg/l (IQR)	95.7 (54–185)	115.5 (24.7–252.2)
WBCd median cells × 109 (IQR)	9 (7–11.3)	10.3 (8.2–13.3)
Hospital length of stay, median days (IQR)	26.5 (20.5–43.8)	16 (12.2–19.2)

Qualitative variables are expressed as numbers and percentages. Quantitative variables are expressed as median and interquartile range (IQR). For patients treated conservatively:

^aArterial hypertension, n = 43.

^bLVEF, n = 40.

^cTemperature, n = 42.

^dWBC, n = 43.

^e n = 42.

BMI: body mass index; CRP: C-reactive protein; LVEF: left ventricular ejection fraction; NYHA: New York Heart Association; WBC: white blood cells.

Regarding the prior procedures before the infection, the most common initial surgery was for acute type A aortic dissection (54.5%; 24/44), followed by surgery for aortic aneurysm (22.7%; 10/44) and IE (13.7%; 6/44). In total, 65.6% (29/44) of these index surgeries were performed on an emergency or urgent basis. Aortic root replacement, typically a Bentall-de Bono-procedure with either a biological or a mechanical prosthesis, had been performed in 63.6% (28/44) of cases, while aortic arch surgery in 15.9% (7/44), indicating the complexity of these index operations. Re-thoracotomy for bleeding or other complications prior to TVGI was necessary in 22.7% (12/44). Table 2 summarizes index surgery data prior to the appearance of TVGI.

Table 2:

Index surgery data before thoracic vascular graft infection

	Conservative, n = 44	Surgery, n = 22
Indication
Dissection, n (%)	24 (54.5)	9 (40.9)
Aneurysm, n (%)	10 (22.7)	12 (54.5)
Infective endocarditis, n (%)	6 (13.7)	/
Other, n (%)	4 (9.1)	1 (4.5)
Surgery duration,a median min (IQR)	326 (252.5–426.5)	258 (230–370)
ECCb time, median min (IQR)	174 (145.5–243.5)	142 (116.2–177)
Aortic-cross-clampc time, median min (IQR)	113 (84–136)	98 (75.2–128.2)
Procedure
Ascending aorta replacement, n (%)	15 (34.1)	7 (21.8)
Aortic root replacement, n (%)	28 (63.6)	15 (68.2)
Aortic arch involvement, n (%)	7 (15.9)	2 (9)
Redo-surgery, n (%)	3 (6.8)	1 (4.5)
Urgent/emergency, n (%)	29 (65.6)	13 (59.2)
Rethoracotomyd, n (%)d	10 (22.7)	3 (13.6)
Postoperative infection, n (%)	15 (34.1)	5 (22.7)

Extracorporeal circulation data from the initial surgery before the appearance of thoracic vascular graft infection. For conservative treated patients:

^aSurgery duration, n = 43.

^bECC, n = 40.

^cAortic-cross-clamp, n = 41. For reoperated patients: Aortic-cross-clamp, n = 21. Quantitative variables are expressed as median and interquartile range (IQR).

^d n = 43.

ECC: extracorporeal circulation.

Microbiology

Microbiological findings are summarized in Table 3. The most common presumed infection route was haematogenous in 43.2% (19/44). Streptococcus spp were the most common isolated microorganisms in 27.3% (12/44), followed by staphylococci at 24.9% (11/44), and in 15.9% (7/44), no pathogen could be isolated (culture-negative, PCR negative)

Table 3:

Microbiological findings

	Conservative, n = 44	Surgery, n = 22
Time to diagnosis, median months (IQR)	11 (1.2–56.2)	54.5 (18–99.5)
Late graft infection, n (%)	28 (63.6)	22 (100)
Route of infection
Haematogenous, n (%)	19 (43.2)	15 (68.2)
Perioperative, n (%)	16 (36.4)	2 (9.1)
Contiguous, n (%)	9 (20.4)	5 (22.7)
Gram-positive bacteria
Staphylococcus aureus, n (%)	6 (13.6)	6 (27.3)
Coagulase-negative staphylococci, n (%)	5 (11.3)	2 (9.1)
Enterococcus faecalis, n (%)	3 (6.8)	/
Cutibacterium acnes, n (%)	2 (4.5)	6 (27.3)
Streptococcus spp., n (%)	12 (27.3)	2 (9.1)
Gram-negative bacteria
Pasteurella multocida, n (%)	1 (2.3)	/
Aggregatibacter aphrophilus, n (%)	1 (2.3)	/
Pseudomonas aeruginosa, n (%)	1 (2.3)	/
Culture negative
Culture negative, n (%)	7 (15.9)	4 (18.2)
Coxiella burnetii, n (%)	1 (2.3)	/
Candida spp, n (%)	1 (2.3)	/
Mycobacterium chimaera, n (%)	1 (2.3)	1 (4.5)

Reasons for not undergoing reoperative surgery

Table 4 presents a comprehensive summary of the various factors contributing to the decision against reoperative treatment involving cardiopulmonary bypass.

Table 4:

Reasons for not undergoing reoperative treatment

Nr.	Treatment	Early/Late	Confirmed/suspected	Reason	Other observations
1	AB only	Late	Suspected	2nd redo-surgery and good response to AB treatment	AB treatment 8 months
2	AB only	Early	Confirmed	Good response to AB treatment	AB treatment, AB treatment 42 months
3	AB only	Late	Confirmed	High risk	Oesophageal-fistula, suppressive treatment
4	AB only	Late	Confirmed	Good response to AB treatment and PET activity regression	AB treatment 5 months
5	AB only	Late	Suspected	Advanced age, suspected infection	AB treatment 6 months
6	AB only	Late	Suspected	Suspected infection	AB treatment 14 months
7	AB only	Late	Confirmed	Good response to AB treatment	AB treatment 6 months
8	AB only	Late	Suspected	2nd redo-surgery, polymorbid	AB treatment 72 months
9	AB only	Late	Suspected	Good response to AB treatment	AB treatment 9 months
10	AB only	Late	Suspected	Neurological impairment	AB treatment 14 months
11	AB only	Late	Suspected	Polymorbid	Died during suppressive treatment
12	AB only	Late	Suspected	Patients wish	AB treatment 60 months
13	AB only	Late	Suspected	Intracerebral bleeding	Suppressive treatment
14	AB only	Late	Suspected	Advanced age	AB treatment 3 months
15	AB only	Late	Suspected	Good response to AB treatment	Ab treatment 6 months
16	AB only	Early	Confirmed	Good response to AB treatment	Suppressive treatment
17	AB only	Late	Confirmed	2nd redo-surgery, shock	Died during AB treatment
18	AB only	Late	Confirmed	5th redo-surgery	Suppressive treatment
19	AB only	Early	Suspected	Suspected infection	AB treatment 1.5 months
20	AB only	Late	Suspected	Good response to AB treatment	AB treatment 6 months
21	AB only	Late	Confirmed	Good response to AB treatment	AB treatment 3 months
22	AB only	Late	Suspected	2nd redo-surgery, advanced age	Suppressive treatment
23	AB only	Early	Suspected	Good response to AB treatment	AB treatment 3 months
24	AB only	Late	Confirmed	Good response to AB treatment	AB treatment 6 months
25	AB only	Early	Confirmed	Recent surgery for infective endocarditis	AB treatment 6 months
26	AB only	Early	Confirmed	Recent surgery for infective endocarditis	AB treatment 4 months
27	AB only	Late	Confirmed	5th redo-surgery	Suppressive treatment
28	AB only	Late	Suspected	3rd redo-surgery	AB treatment 6 months
29	Selective debridement	Late	Confirmed	Patients wish	Mediastinal abscess, died during suppressive AB treatment
30	Selective debridement	Early	Confirmed	Recent surgery for infective endocarditis	Mediastinal abscess, AB therapy 10 months
31	Selective debridement	Late	Confirmed	Good response to AB treatment and low virulence bacteria	Mediastinal abscess, AB treatment 20 months
32	Selective debridement	Early	Confirmed	Good response to intervention and AB treatment	Mediastinal abscess, AB treatment 11 months
33	Selective debridement	Early	Confirmed	Good response to intervention and AB treatment	Mediastinal abscess, AB treatment 4 months
34	Selective debridement	Late	Confirmed	Good response to intervention and AB treatment	Mediastinal abscess, AB treatment 4 months
35	Selective debridement	Early	Confirmed	4th redo-surgery, 32 days after last surgery	Cutaneous abscess, AB treatment 32 months
36	Selective debridement	Early	Confirmed	High risk	Mediastinal abscess, died during suppressive treatment
37	Selective debridement	Early	Confirmed	High risk, patients wish	Mediastinal abscess, AB treatment 36 months
38	Selective debridement	Early	Confirmed	High risk and good response to intervention and AB treatment	Mediastinal abscess, AB treatment 15 months
39	Selective debridement	Early	Confirmed	High risk and good response to intervention and AB treatment	Mediastinal abscess, AB treatment 38 months
40	Selective debridement	Early	Confirmed	High risk, advanced age	Cutaneous abscess, AB treatment 26 months
41	Selective debridement	Late	Confirmed	Cancer/Chemotherapy	Cutaneous abscess, died during suppressive treatment
42	Selective debridement	Late	Confirmed	High risk	Cutaneous abscess, AB treatment 24 months
43	Selective debridement	Early	Suspected	High risk	Cutaneous abscess, AB treatment 24 months
44	Selective debridement	Late	Confirmed	High risk	Mediastinal abscess, suppressive treatment

Survival, reintervention, relapse and reinfection

One patient died 17 days after hospital discharge due to unknown causes. The median follow-up duration was 4.8 years (IQR 1.7–6.1), with cumulative survival shown in Fig. 1. At 4.8 years, 82.9% (95% confidence interval [CI] 95%, 69.7–96.1) were alive. Eight patients died during follow-up, with details provided in Table 5.

Figure 1:

Kaplan–Meier survival estimates.

Table 5:

Characteristics of deceased patients at follow-up

	Index surgery	Early/late	Pathogen	Reason no surgery/treatment	Survival	Cause of death
1	Bentall-De Bono (biological prosthesis) + CABGx2, Rescue-TAVI	Late	C. acnes	Neurological impairment	9.3 months	Cardiac
2	Bentall-De Bono	Late	Culture negative	Good response to AB treatment	19.4 months	Perioperatively (TEVAR)
3	Bentall-De Bono + Debranching Innominate artery + CABGx1	Late	P. multocida	2nd redo-surgery	5.6 years	Intraoperatively (pseudoaneurysm)
4	Bentall-De Bono + Frozen elephant trunk	Late	M. chimaera	Patients wish, suppressive treatment	1.1 years	Graft infection
5	Redo-Bentall-De Bono (biological), LSA-Debranching, EAP	Early	Rothia dentocariosa	High-risk, selective debridement, suppressive treatment	42 days	Graft infection
6	Bentall-De Bono (biological)	Early	Corynebacterium durum, C. acnes	High-risk and advanced age, selective debridement	8.1 years	Other
7	Bentall-de Bono (biological)	Late	Streptococcus spp	Malignancy, selective debridement, suppressive treatment	4.8 years	Malignancy
8	Ascending aorta replacement + AVR (biological), PFO-closure, pacemaker	Early	Streptococcus spp	2nd redo-surgery	1.8 years	Other infection
	Reoperated patients
1	Bentall-De Bono + CABGx2, MitraClip x2	Late	S. aureus	Re-Bentall (biological prosthesis), Cabrol-shunt, CABGx1, MVR (biological prosthesis)	1.5 years	Unknown
2	Bentall-biological + ascending aorta replacement	Late	Strep. dysgalactiae	Re-Bentall (biological prosthesis)	4.3 years	Unknown
3	Bentall-De Bono + CABGx1	Late	Culture negative	Re-Bentall + Cabrol fistula	2.7 years	Unknown

In total, 5 patients required interventions during follow-up (Fig. 2). Patient 1 underwent transcatheter aortic valve implantation (TAVI) 7 weeks after treatment initiation due to refractory heart failure. Patients 2 and 3 also received TAVI for severe regurgitation, at 5 and 6.5 months post diagnosis, respectively. Patient 4 underwent aortic arch replacement and thoracic endovascular aortic repair 16 and 18 months after initial diagnosis, respectively, following 7 months of infection-control and cessation of antimicrobial therapy. Lastly, patient 5 had surgery for pseudoaneurysm formation 5.6 years after diagnosis and nearly 5 years after cessation of antimicrobial therapy. There were no relapses and 1 reinfection with Escherichia coli 19 months after diagnosis (Fig. 3).

Figure 2:

Re-intervention analysis during follow-up with death as a competing event.

Figure 3:

Reinfection analysis during follow-up with death as a competing event.

Overall, 70.4% (31/44) of patients achieved cure during follow-up, with discontinuation of antimicrobial treatment following clinical assessment, laboratory monitoring and imaging. The median duration of antimicrobial treatment was 7 months (IQR 6–18). Eleven patients required life-long suppressive treatment due to persistent infection identified in follow-up visits; 4 of these patients died during the follow-up period. Post-mortem investigations were not available. Patient 1 died of unknown causes at home; patient 2 died due to an uncontrolled M. chimaera infection after deliberately ceasing antimicrobial treatment; patient 3 died at home during suppressive treatment; and patient 4 died due to metastatic cancer.

DISCUSSION

With this investigation, we have analysed outcomes from a cohort of patients with TVGI that did not undergo reoperative surgery on cardiopulmonary bypass, with low in-hospital mortality (2.2%) and an acceptable survival rate (82.9%) at 4.8 years of follow-up. The standard of care for patients with TVGI includes reoperative surgery on cardiopulmonary bypass with complete removal of the infected prosthesis, and targeted antimicrobial therapy [6, 8]. Our results show that this approach yields favourable outcomes, with acceptable survival rates (76.9%) and low rates of reinfection and reoperation at 4.4 years which is in synchrony to the available literature [4, 12].

Factors such as age, comorbidities, patients’ condition at presentation and a patient’s preference to avoid reoperative surgery can influence the decision against reoperative surgery on cardiopulmonary bypass and complete removal of the infected prosthesis. The outcomes of patients receiving antimicrobial treatment without surgical excision remain understudied [1, 3, 4, 12, 25]. Ramos et al. [12] reported no fatalities among 4 patients managed non surgically, while in Akowuah et al. [25] on 8 patients, 2 patients died after 27 days and 10 weeks of treatment, respectively. In contrast, Oda et al. [4] reported no in-hospital deaths following an antibiotic treatment-only approach, though no follow-up were included. In contrast, Kouijzer et al. [17] examined outcomes for 24 patients conservatively, noting a 6-months mortality of 8%. Among these, 19 patients were solely with antimicrobial therapy achieving acceptable outcomes without reoperative surgery.

In our study, 28 patients were treated solely with antimicrobial therapy without reoperative surgery on cardiopulmonary bypass, while 16 received resternotomy, mediastinal irrigation with or without negative-pressure wound therapy based on consensus from our multidisciplinary meeting. In-hospital mortality was low (2.2%), with only 1 patient dying prior to discharge.

In contrast, 22 patients were reoperated on cardiopulmonary bypass with complete prosthesis removal. Outcomes of reoperated patients have been reported by other groups: Ramos et al. reported outcomes of 27 patients from a multicentric study (GAMES registry). Twenty-two patients were reoperated on, with an operative mortality of 22% and no follow-up information [12]. Similarly, Preventza et al. reported results on 39 patients (33 with ascending grafts) during a 13-year period, with an operative mortality of 10.6%, and a survival of 65% during a median follow-up of 2.5 years [26]. Oda et al. from Japan analysed outcomes of 68 patients during 13 years. Of these, 18 were reoperated on cardiopulmonary bypass, with an operative mortality of 33% and overall 1- and 3-year survival rates of 58.6% and 56.2%, respectively [4]. In our study, operative mortality was as low as 13.6% and a follow-up survival at 4 years was 76.9%, which is somewhat higher than in the reported literature. Our results suggest that after overcoming the initial phase of the reoperation and infection, outcomes remain good and stable over time.

Decision for not to reoperate vary in the literature. Kouijzer et al. [17] identified 4 main reasons for avoiding reoperative surgery, with the most common being a favourable setting due to haematogenous source of infection and early initiation of targeted antimicrobial treatment. In our analysis, the primary reasons for not undergoing reoperative surgery were a combination of inoperability due to high preoperative risk, extensive surgical history and team expertise. In contrast, this factor accounted for only 17% of cases in the study mentioned above. In our study, 14 patients showed a good response to the initial treatment and were no longer consider candidates for reoperative treatment. This variability highlights the heterogeneity in patients conditions and the different approaches taken by the endocarditis teams.

Several risk scores have been proposed to assess patient condition and surgical risks for IE [27, 28] with conflicting results. To date, none are widely adopted for IE or TVGI. Additionally, it has been suggested that these scores should not be used solely as a basis to reject patients for reoperative surgery [29, 30].

The cumulative survival of our patients not undergoing reoperation on cardiopulmonary bypass was 82.9% (CI 95%, 69.7–96.1), which is rather high given the mortality rates reported for vascular graft infections [4, 12, 13]. However, these results may be explained by the lower virulence of the responsible microorganisms and the low rate of aorta-oesophageal or aortobronchial fistulae (only 2) in the cohort. The most common microorganisms were Streptococcus spp, unlike other studies where staphylococci were more common. Another factor that may have positively influenced these results is that at presentation, 63.6% had bacteraemia and elevated C-reactive protein (CRP) levels [median 95.7 mg/l (IQR 54–185)] as a surrogate for acute onset infection. This might have led to a rapid initiation of a targeted antimicrobial regime that prevented biofilm formation.

While survival outcomes of conservatively treated patients might appear better than those who underwent reoperation on cardiopulmonary bypass, one cannot forget that these groups are not comparable. First of all, the conservatively treated have not the adjunct risks of a reoperation protecting them to some extent of immediate and short-term mortality as a result of the aggressiveness of reoperative surgery. Second, 17 patients out of 44 (38%), and 16/28 from the antibiotics-only group, had suspected infections according to the diagnostic criteria, whereas only 1 patient out of 22 (4.5%) from the reoperated group was categorized as such. This difference in confirmed and suspected infections might be key in comparing both groups. This could also suggest that those patients with suspected infections could be managed without reoperative treatment, whereas those with confirmed infections required surgical intervention to achieve infection control.

While infection could be controlled in 31 patients, allowing for discontinuation of antimicrobial therapy with only 1 reinfection at follow-up, 25% (11/44) of patients required lifelong antimicrobial therapy due to pathological laboratory findings and metabolic activity observed in their ¹⁸F-FDG-PET/CT-scans.

The strengths of our study include its prospective observational design, which facilitated more detailed and tailored data collection compared to previous retrospective or cross-sectional studies on vascular graft infections. Additionally, we benefited from the adjudication of clinical events and a multidisciplinary approach that involved specialists from cardiac surgery, cardiology, nuclear medicine and infectious diseases. Moreover, our follow-up on conservatively managed patients is one of the longest reported to date, with a median of 4.8 years (IQR 1.7–6.1). Our study also has limitations. First, since this study reflects the experience of a Swiss tertiary care referral centre, the data may not be generalizable to other health care systems. Second, there is no formal record linkage with other hospitals, making it impossible to fully rule out the possibility that participants attended other institutions for care and that information on death, reinfection and relapse was not reported by participants or care providers. Third, surgical rejection and patient selection introduce a potential selection bias, limiting our ability to make comparisons with other treatment approaches. Lastly, the different treatment modalities and the initiation of these in both groups make it complex to compare survival rates hence leading to potential survivorship bias. This was avoided by selecting the hospital discharge date as common time zero.

CONCLUSION

TVGIs are complex diseases associated with high mortality, even when managed with reoperative surgery and extended antimicrobial treatment. The decision to reoperate and remove the infected prosthesis rests with the endocarditis team, which must assess the patients’ clinical condition and life expectancy. If the drawbacks of surgical intervention outweigh its benefits, antimicrobial therapy without reoperative surgery on cardiopulmonary bypass may be an acceptable alternative for selected patients. This strategy does not replace gold-standard care and should not be universally applied. However, it demonstrates that a conservative approach can be a viable option for a specific group of patients.

Glossary

ABBREVIATIONS

CT

Computed tomography

FDG

18F-fluorodeoxyglucose

IE

Infective endocarditis

IQR

Interquartile range

PET

Positron emission tomography

TVGI

Thoracic vascular graft infection

FUNDING

This study was financed within the framework of the Vascular Graft Cohort Study (VASGRA), supported by the Swiss National Science Foundation (SNF) grant #32003B_219351/1 (to B.H.). This work was also supported by the Clinical Research Priority Program (CRPP) of the University of Zurich for the CRPP ‘Precision Medicine for Bacterial Infections’.

Conflict of interest: none declared.

DATA AVAILABILITY

The data presented in this study are available on request from the corresponding author.

Author contributions

Mathias Van Hemelrijck: Conceptualization; Data curation; Formal analysis; Writing—original draft. Juri Sromicki: Writing—review & editing. Petar Risteski: Writing—review & editing. Rasha Boulos: Data curation; Formal analysis; Writing—review & editing. Ronny R. Buechel: Writing—review & editing. Michelle Frank: Writing—review & editing. Barbara Hasse: Data curation; Funding acquisition; Writing—review & editing. Héctor R.C. Biefer: Supervision; Writing—review & editing. Omer Dzemali: Supervision; Validation; Writing—review & editing

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Ari Mennander, Carlos A. Mestres, Eduard Quintana and the other anonymous reviewer(s) for their contribution to the peer review process of this article.