Long-term survival after surgical aortic valve replacement in patients aged 80 years and over

Abstract

 

Open in new tabDownload slide

OBJECTIVES

Surgical aortic valve replacement can be safely performed in people aged 80 years and older with early benefits on both symptomatic and prognostic ground. While new approaches are advocated to treat this elderly and frail population, data on long-term outcomes are not available.

METHODS

We conducted a retrospective analysis of 1870 patients aged 80 years and over who underwent first time surgical aortic valve replacement during the period 2000–2019. The Kaplan–Meier method was used to calculate survival and comparisons among groups were performed by log-rank test. Cox analysis was used to determine the independent risk factors for late mortality.

RESULTS

The patients’ mean age was 84 years and 53% were male. Isolated aortic valve replacement was performed in 42% of the patients, and coronary artery bypass grafting (n = 956), mitral valve (n = 94) or aortic surgery (n = 69) were associated in the remaining cases. One hundred eighty-one patients (8%) sustained at least 1 postoperative complication (reopening for bleeding or tamponade 3%, renal replacement therapy 3%, new cerebral stroke 1.5%). In-hospital mortality was 3.2% in the overall population (60/1870) and 2.2% after isolated aortic valve replacement (18/790). Survival was 90%, 66%, 31% and 14% at 1, 5, 10 and 15 years, respectively, and was similar to the expected survival of a sex- and age-matched population (log-rank P = 0.96). A complicated postoperative course was an independent risk factor for mortality during the follow-up [hazard ratio 1.32 (1.03, 1.68), P = 0.026].

CONCLUSIONS

Surgical aortic valve replacement can be performed with an acceptable early mortality rate and provides excellent long-term survival in people aged 80 years and older.

INTRODUCTION

Cardiac surgery has been increasingly performed in octogenarian and older patients during the last decades [1, 2]. The initial experiences reported a postoperative mortality up to 17% [3–6]; however, more recent studies reported a progressive reduction of early mortality for these patients [7, 8]. Particularly, surgical aortic valve replacement can be nowadays safely performed in patients aged 80 years and over with an acceptable in-hospital mortality rate (∼2%) [9–11] and the provision of improved survival [12] and a satisfactory quality of life at mid-term follow-up [5, 7]. Transcatheter techniques have shown similar procedural safety while providing a shorter hospital stay and a quicker recovery [7, 11]. These advantages have not translated so far into a better mid-term survival [13]; nevertheless, long-term results and the potential impact on the long run of postoperative complications such as acute kidney injury and bleeding, especially in surgical patients [8, 11], or a high incidence of pacemaker implantation and paravalvular regurgitation in patients treated with transcatheter procedures [14, 15], remain undetermined.

No long-term data are available for octogenarian and older patients who underwent surgical aortic valve replacement; for this reason, we have reviewed our surgical experience of the last 20 years to provide a picture of the survival of this population and to study the risk factors associated with mortality during the follow-up.

MATERIALS AND METHODS

Ethical statement

Approval was obtained for the use of data (Safeguard System approval number SEV/0029, date 24 October 2018). Considering the type of the study involving anonymized and previously collected data, patients’ consent was waived.

Population

The internal database of Wessex Cardiothoracic Centre at University Hospital Southampton was searched to identify patients who underwent aortic valve replacement during the period January 2000 to December 2019 using the following criteria.

Inclusion criteria:

• Age ≥80 years

• First time sternotomy

• Associated procedures including coronary artery bypass grafting (CABG), mitral valve surgery, tricuspid valve surgery and aortic surgery

Exclusion criteria:

• Redo procedure (any previous cardiac operation)

A total of 1870 patients fulfilled the criteria.

Study design, data collection and outcomes

This study is a retrospective outcome evaluation from institutional records with prospective data entry collected and used in compliance with institutional data protection and confidentiality policies. The data were collected from the hospital database system, patients’ records and records of the general practitioners.

The following data were collected:

Preoperative details:

• Age

• Gender

• History/diagnosis of systemic hypertension

• Diabetes Mellitus

• Chronic obstructive pulmonary disease (COPD)

• Smoking history (ex-smoker or active smoker)

• Creatinine >200 µmol/l

• Previous cerebral stroke

• Extracardiac arteriopathy

• Previous acute myocardial infarction

• New York Heart Association (NYHA) class

• LVEF

• Haemodynamic of valve pathology (aortic stenosis, regurgitation)

• Operation status (elective/urgent)

• Logistic EuroSCORE

Operative data:

• Associated procedures (CABG, MV surgery, TV surgery, aortic surgery)

• Cross-clamp time

• Cardiopulmonary bypass time

Postoperative outcomes:

• In-hospital mortality (death before hospital discharge)

• Reopening for bleeding/tamponade

• Renal replacement therapy/continuous veno-venous haemodyalisis

• New postoperative neurological deficit [permanent stroke, transient ischaemic attack (TIA)]

• Deep sternal wound infection

• Permanent pacemaker implantation (predischarge and post-discharge)

• Postoperative hospital length of stay

• Discharge destination

Long-term outcomes:

• Survival

Definitions

The preoperative data collected were as previously defined for EuroSCORE [16]. All-cause mortality was considered for survival during the follow-up. Postoperative outcomes (occurred before hospital discharge) were recorded according to the VARC-2 criteria [17]. Particularly, new neurological deficit was coded according to the criteria for ‘stroke and TIA’ [17].

Surgical techniques

All the operations were performed through midline sternotomy. Cardioplegic arrest was achieved with cold blood cardioplegia. Conventional aortic valve prostheses were used in all the cases; a tissue valve was implanted in 99.7% of the patients (1865/1870). Intraoperative TOE has been used routinely in the period 2010–2019. All the patients underwent a predischarge transthoracic echocardiogram and a clinic follow-up during the first 3 months since the operation.

Statistical analysis

Continuous variables were presented as mean [standard deviation (SD)] or median (1 interquartile range, 3 interquartile range). Categorical variables were presented as number (%). Univariable comparisons of preoperative, operative and postoperative variables were performed among the groups using the Student’s t-test, Kruskall–Wallis test or the χ² or Fisher’s exact test as appropriate.

Survival probabilities, with time starting on the date of surgery, were calculated using the Kaplan–Meier method. Survival probabilities of a sex- and age-matched English population were calculated using the Kaplan–Meier method and were based on the data available on https://www.ons.gov.uk/peoplepopulationandcommunity/birthsdeathsandmarriages/lifeexpectancies/datasets/singleyearlifetablesuk1980to2018.

Log-rank test was used to compare the survival curves.

A Cox regression analysis, with time starting on the date of hospital discharge, was performed to determine the hazard ratios (HRs) of long-term survival; a backward stepwise model with a significance of P < 0.15 was used. The variables included were gender, age, NYHA III–IV, previous myocardial infarction, diabetes mellitus, hypertension, COPD, history of cerebrovascular accident, extracardiac arteriopathy, LVEF < 30%, type of haemodynamic dysfunction (aortic valve stenosis or regurgitation), associated procedures, cardiopulmonary bypass (CPB) time, cross-clamp time, period when operation was performed (2000–2009 or 2010–2019), occurrence or at least 1 postoperative complication or single complication, pacemaker implantation.

A P-value of ≤0.05 was considered statistically significant.

Statistical analyses were performed using the Stat-View Statistical Software Package 5.0 (SAS Institute, Inc., Cary, NC, USA), NCSS 2001 (Number Cruncher Statistical System, Kaysville, Utah) and Stata/MP version 13 (StataCorp, College Station, TX, USA).

RESULTS

Preoperative characteristics and operative data

During the period 2000–2019, 1870 patients aged 80 years and over underwent first time surgical aortic valve replacement.

There was a progressive increase in the number of surgical aortic valve replacement procedures with octogenarian and older patients accounting for almost the 30% of the overall population undergoing aortic valve surgery. Table 1 reports the preoperative characteristics of our population; Fig. 1 and Supplementary Material, Table S1 detail about patients’ characteristics and number of procedures for every 5-year interval time.

The mean patients’ age was 84 (SD: 3) years, and among them, 53% (987/1870) were male. Logistic EuroSCORE was 13.3% (SD: 7.2%). Forty percentage of the patients presented with an NYHA class III–IV; admission with advanced heart failure symptoms was more common at the beginning of our study period (2000–2004) with 63% of the patients presenting a marked functional limitation compared with the 31% of the cases in the period 2015–2019 (P < 0.001). Similarly, preoperative creatinine >200 µmol/l was more common in the early period (7%), while its prevalence declined to 3% in 717 patients operated during 2010–2014 and to 1% in 600 patients operated between 2015 and 2019 (Supplementary Material, Table S1).

Isolated aortic valve replacement was performed in 42% of the patients (790/1870), and details of the associated procedures and operative times are reported in Table 2.

Early outcomes

The postoperative course before hospital discharge was complicated in 9.6% of the cases. Reopening for bleeding or tamponade was necessary in 57 patients (3%); a new neurological deficit was reported in 47 patients (2.5%) and among them 28 suffered a new permanent cerebral stroke (1.5%); 57 patients (3.0%) required renal replacement therapy; a permanent pacemaker was implanted in 78 cases (4.1%) before hospital discharge.

Overall in-hospital mortality was 3.2% (60/1870) at a median time of 11 days (3, 30). In-hospital mortality after isolated aortic valve replacement was 2.2% (18/790) at a median time of 14 days (4, 33). The mortality rate decreased significantly over time from 6% (9/150) in the period 2000–2004 to 1.5% (9/600) in the period 2015–2019 (P = 0.006) (Table 3 and Supplementary Material, Table S1).

In 54% of the cases (977/1810), patients were discharged at home while in the remaining cases a further period of rehabilitation, nursing or medical assistance was required. The overall median postoperative length of stay was 10 (8, 15) days (Table 3 and Supplementary Material, Table S1).

During the first 6 months, there were further 83 deaths among patients who were discharged from hospital for an overall 6-month mortality of 7.6%.

Long-term outcomes

During a median follow-up of 4.6 (3.5, 11.0) years, there were 907 deaths including in-hospital mortality. The survival probabilities at 1, 5, 10 and 15 years from the date of surgery were 90% [standard error (SE): 0.7%), 66% (SE: 1.2%), 31% (SE: 1.6%) and 14% (SE: 2.0%], respectively (Fig. 2). There was no significant difference in survival between the study population and an age- and sex-matched general English population (log-rank test P = 0.96) (Fig. 3).

The survival for patients who experienced at least 1 postoperative complication (reopening, new neurological deficit, renal replacement therapy, deep sternal wound infection) was 71% (SE: 3.7%), 46% (4.2%) and 18% (SE: 4.2%) at 1, 5 and 10 years from the date of surgery, respectively. It was significantly lower (log-rank test P = 0.002) when compared with the survival of patients who had an uncomplicated postoperative course, 92% (SE: 0.7%) at 1 year, 68% (SE: 1.3%) at 5 years and 33% (SE: 1.7%) at 10 years. Similarly, patients who were operated during the period 2000–2009 showed a significantly lower survival compared with patients who underwent surgery during the period 2010–2019: 10-year survival 27% (SE: 1.9%) vs 41% (SE: 2.5%), respectively (log-rank test, P = 0.002).

Supplementary Material, Fig. S1 provides a graphical view of the impact of the occurrence of postoperative complication(s) and the time of surgery on 10-year survival when compared with the expected survival of an age- and sex-matched general English population.

A Cox regression, with follow-up starting at the discharge date, was performed to study independent predictors for late mortality in patients who were successfully discharged from hospital. Alongside co-existent preoperative comorbidities (previous myocardial infarction, extracardiac arteriopathy), a complicated postoperative course (P = 0.026; HR 1.32, confidence interval 1.03–1.68) and surgery in the first 10 years of the study period (P = 0.003; HR 1.25, confidence interval 1.08–1.46) were independent risk factors for mortality at the long term (Tables 4 and 5].

DISCUSSION

Nowadays, cardiac surgery can be performed safely in octogenarians and older patients. Several studies reported favourable early- and mid-term outcomes across all the spectrum of cardiac diseases and surgical procedures including CABG, aortic valve and aortic surgery, mitral valve surgery and emergency operation [9–11, 18, 19].

Aortic valve surgery represents one of the most common operations undertaken in people aged over 80 years as, in our experience, they accounted for more than one-fourth of the global cases of surgical aortic valve replacement. In these patients, we have found an in-hospital mortality rate of 3.2% over a 20-year period with a progressive and significant reduction in the mortality rate across the years (1.5% in the last period). Our data confirmed recent findings reporting an early mortality rate of 2% in octogenarian and older patients undergoing surgical aortic valve replacement [9–11]. Alongside this evidence of procedural safety, other studies have highlighted that the mid-term survival of elderly patients with aortic valve stenosis is dramatically improved by aortic valve replacement (68% surgery versus 22% medical therapy at 5 years) [12] and that this longer life expectancy is characterized by an acceptable quality of life and independent functional status [6].

We have further explored the long-term outcome of these patients and reported a survival rate of 31% at 10-year and 14% at 15-year follow-up. No long-term data from similar surgical populations were available for a comparison; however, we found that the 10-year survival probability of our patients was not different from the survival rate of an English age- and sex-matched population.

Preoperative status and comorbidities play a major role in determining long-term survival. In more recent years, people aged 80 years and over underwent surgery in a better functional status and less frequently presented with severe chronic kidney dysfunction and history of previous myocardial infarction; nevertheless, the population operated in the last period of observation had a significantly lower mean logistic EuroSCORE value. Reasons for this shift in practice lie probably in an earlier referral driven by a new awareness of the benefits of surgery also in elderly people, the confidence in surgical results and the availability of alternative interventional procedures associated with an established activity of the Heart Valve Team in evaluating frail, high-risk and elderly patients. These changes in preoperative management and the possibility of undertaking a cardiac operation before the development of advanced heart failure symptoms translated in a lower early mortality and a better long-term outcome.

Despite a low perioperative death rate, the first postoperative months were burdened by a significant mortality. This represents a common finding [5, 6, 18] and underlines the vital importance of preserving the already reduced functional reserves of frail and comorbid patients. Prolonged operative times have been extensively associated with perioperative mortality and morbidity in octogenarians undergoing cardiac surgery [3, 8, 20]. A complicated postoperative course may also expose these elderly patients to a difficult and prolonged recovery and has been associated with a worse outcome during the first year since surgery with a mortality at 1 year of 67% in patients requiring postoperative dialysis and 35% in patients suffering perioperative cerebral stroke [8]. In our experience, any deviation from an uncomplicated postoperative course was significantly associated with a worse survival at long-term follow-up. Particularly, new permanent cerebral injury after surgery was an independent predictor of long-term mortality, as it can affect the postoperative recovery and patients’ quality of life and independence.

We recorded a lower in-hospital mortality and a better long-term survival in patients operated during the period 2010–2019. There were no significant technical changes in the intraoperative management of these patients throughout the study period; full sternotomy was the common surgical access and conventional stented prostheses were used. Noteworthy, the last 10 years have been characterized by the establishment and the progressive growing of a transcatheter aortic valve implantation (TAVI) programme.

The implementation of a TAVI programme has been associated with an increase in overall aortic valve intervention and surgical aortic valve replacement activities [21, 22]. Similarly to our findings, a French administrative hospital-discharge database showed that the introduction of TAVI procedures translated in an earlier referral with an improved patients’ preoperative clinical profile [22].

In French and German nationwide experiences, the number of TAVI procedures is now exceeding the surgical activity [22, 23] and this is particularly evident in patients aged over 80 years [22]. A less invasive procedure can undoubtedly be associated with a lower degree of postoperative morbidity especially in high-risk elderly patients [11]. Transcatheter procedures have been associated with a lower risk of bleeding and postoperative renal failure compared with surgical aortic valve replacement [11] and, according to our findings, this could translate in a potential benefit on survival. However, several observational studies have failed to demonstrate, so far, a difference in early mortality after TAVI versus surgical aortic valve replacement in elderly patients [11, 22, 23]. No data are available of mid- and long-term outcomes after transcatheter procedures in low-risk octogenarians and, generally, the impact of prosthesis degeneration and paravalvular regurgitation has not been investigated on the long term due to the inclusion of cohort of patients with reduced life expectancy (survival <30% at 5 years) [14].

On these bases, in our practice, advanced age alone does not represent a sufficient factor to favour a transcatheter procedure over conventional surgery. In our experience, a better preoperative functional status and the careful selection of surgical candidates, while offering transcatheter procedures to higher-risk patients, were the 2 most important factors that ultimately led to a reduced incidence of postoperative complications, lower early mortality, better late outcomes and the opportunity to offer an interventional treatment (conventional surgery or transcatheter procedure) to a wider population of elderly patients.

Limitations

Within the limitations of a single-centre retrospective evaluation, we have reported in a large population an historical view of long-term outcomes of patients aged 80 years and older who underwent surgical aortic valve replacement. This evidence represents a reliable picture of the contemporary surgical outcomes of aortic valve surgery in elderly patients and could represent a benchmark for the forthcoming long-term results of non-surgical interventional treatment of aortic valve disease.

CONCLUSIONS

Conventional surgical aortic valve replacement during the last 20 years provided a safe and successful treatment for aortic valve disease in elderly patients with an in-hospital mortality that in more recent years was for both isolated and combined procedures below 2%. An early surgical referral before the development of advanced signs of heart failure and the availability of alternative interventional procedures and a careful patients’ selection contributed to a progressive and significant reduction in postoperative complications and periprocedural mortality and led to an improved long-term survival.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

Conflict of interest: none declared.

Author contributions

Pietro Giorgio Malvindi: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Validation; Writing—original draft. Suvitesh Luthra: Conceptualization; Data curation; Investigation; Methodology; Writing—original draft. Suresh Giritharan: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Validation; Writing—review & editing. Mariusz Kowalewski: Data curation; Formal analysis; Methodology; Software; Validation; Writing—review & editing. Sunil Ohri: Conceptualization; Investigation; Methodology; Project administration; Supervision; Validation; Writing—original draft.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Walter J. Gomes, Yoshiro Matsui and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

 
• HRs

  Hazard ratios

 
• SD

  Standard deviation