Cardiac surgery in nonagenarians: not only feasible, but also reasonable?

Abstract

OBJECTIVES

Changes in the age profile of the population in the western world and improvement in surgical techniques and postoperative care have contributed to a growing number of cardiosurgical patients aged over 90. In periods when transapical and transfemoral aortic valve replacement were done, we aimed at evaluating the outcome of nonagenarians after conventional aortic valve replacement and cardiac surgery in general, and determining perioperative parameters to predict a complicated postoperative course.

METHODS

Between 1995 and 2011, 49 nonagenarians (aged 91.2 ± 3.1 years) underwent cardiac surgery. A subgroup of 30 patients received aortic valve replacement alone (63%; n = 19), in combination with coronary artery bypass grafting (27%; n = 8) or other surgical procedures (10%; n = 3). Most of the patients suffered from combined aortic valve disease with a mean valve orifice area of 0.6 ± 0.3 cm² and a mean antegrade pressure gradient of 86 ± 22 mmHg.

RESULTS

Cardiac surgery in nonagenarians resulted in remarkable postoperative morbidity and an overall in-hospital mortality of 10% (n = 5). In the AVR subgroup, biological valve prostheses were implanted in 29 patients. In this subgroup, the length of stay was 2.9 ± 0.9 days in the intensive care unit and 17.0 ± 5.5 days in the hospital. The in-hospital mortality amounted to 13% (n = 4). Although several general preoperative risk factors of postoperative complications such as renal failure, low cardiac output syndrome and New York Heart Association Class IV were remarkably more frequent among the patients who died after the operation, the small cohort of non-surviving nonagenarians did not allow for significant differences.

CONCLUSIONS

Cardiac surgery in the very elderly, particularly with regard to aortic valve replacement, carries a high risk of early morbidity and mortality. However, in selected nonagenarians, surgery can be performed with an acceptable outcome. The risk may even be reduced by an individual approach to the procedure. With regard to potential risk factors, the selection of these patients should be carried out very carefully.

INTRODUCTION

Increasing life expectancy and improvement in operative techniques as well as postoperative care have contributed to an increasing number of elderly and very elderly surgical patients. In 2010, >50% of all German cardiosurgical operations were performed in patients aged over 70 [1], while 15 years previously, this group accounted for only 28%. Likewise, the group of nonagenarians being considered for cardiosurgical procedures has increased and poses a challenge to surgical skills as well as technical facilities.

Recent studies demonstrated an adequate operative outcome and even promising follow-up data for selected nonagenarians undergoing conventional cardiac surgery [2–4] and particularly aortic valve replacement (AVR) [5, 6]. On the other side, transapical and transfemoral approaches for AVR transcatheter aortic valve implantation (TAVI) offer a good alternative in very elderly patients with multiple comorbidities. In 2007, the first TAVI procedures in Germany were conducted and, 3 years later, the number had increased to 24% of all AVR patients [1].

The present study aimed at evaluating the outcome of nonagenarians undergoing cardiac surgery, and in particular conventional AVR in times of TAVI. Furthermore, we determined perioperative parameters to predict a complicated postoperative course, which may influence the preoperative decision between surgical, interventional or conservative treatment.

PATIENTS AND METHODS

Retrospective data analysis included 49 nonagenarians who were operated on in our cardiosurgical centre from 1995 to 2011. A subgroup of 30 patients suffered from calcifying aortic stenosis exhibiting a mean valve orifice area of 0.6 ± 0.3 cm² and a mean antegrade pressure gradient of 86 ± 22 mmHg. All those patients had undergone conventional AVR via median sternotomy, and all except one had received biological aortic valve prostheses. The AVR operation was performed as single AVR (63%; n = 19), in combination with coronary artery bypass grafting (27%; n = 8) or other additional surgical procedures (10%; n = 3).

In the following, the groups are defined as:

• ‘TOTAL’ all patients ≥90 years (n = 49).

• ‘AVR’ AVR patients ≥90 years (n = 30).

In Tables 1 and 2, the pre- and intraoperative characteristics of all nonagenarians (group TOTAL), and nonagenarians who underwent AVR (group AVR), are depicted.

Table 1:

Pre- and intraoperative characteristics of all nonagenarians who underwent cardiac surgery (group TOTAL)

	Group TOTAL (n = 49)
Preoperative
Age (year)	91.2 ± 3.1
Gender male	32
Morbid obesity	9
Chronic obstructive pulmonary disease	7
Diabetes mellitus	14
Arterial hypertension	38
Former neurological problems	6
Left ventricular ejection fraction <0.4	9
New York Heart Association Class IV	11
Renal failure	6
Unstable angina	9
Intraoperative
Emergent operation	4
Redo operation	3
Coronary artery bypass grafting	19
Valve/combined surgery	30
Operation time (min)	279 ± 61
Extracorporeal circulation time (min)	131 ± 41
Aortic cross-clamp time (min)	88 ± 28
Surgical re-exploration	3
Haemorrhage	12
Myocardial infarction	2
Low cardiac output syndrome	3

Table 2:

Pre- and intraoperative characteristics of nonagenarians who underwent AVR (group AVR), including an analysis of possible predictors for in-hospital mortality

	Group AVR (n = 30)	Survivors of group AVR (n = 26)	Non-survivors of group AVR (n = 4)	P-value (non-surviving vs surviving patients)
Preoperative
Age (year)	91.5 ± 3.3	91.3 ± 3.2	92.7 ± 3.8	0.432
Gender male	19	16	3	1.000
Morbid obesity	5	4	1	0.538
Chronic obstructive pulmonary disease	6	5	1	1.000
Diabetes mellitus	12	10	2	1.000
Arterial hypertension	18	16	2	1.000
Former neurological problems	4	3	1	0.455
Left ventricular ejection fraction <0.4	8	6	2	0.284
New York Heart Association Class IV	8	6	2	0.284
Low cardiac output syndrome	3	2	1	0.360
Renal failure	3	2	1	0.360
Unstable angina	4	3	1	0.455
Intraoperative
Emergent operation	2	1	1	0.253
Redo operation	2	1	1	0.253
Aortic valve replacement alone	19	16	3	1.000
Operation time (min)	268 ± 41	265 ± 40	286 ± 47	0.346
Extracorporeal circulation time (min)	114 ± 33	112 ± 31	126 ± 44	0.431
Aortic cross-clamp time (min)	77 ± 23	76 ± 23	83 ± 23	0.576
Prolonged aortic cross clamping >80 min	12	9	3	0.274
Prosthesis diameter <21 mm	3	2	1	0.360
Surgical re-exploration	2	1	1	0.253
Haemorrhage	5	4	1	0.538
Myocardial infarction	1	0	1	0.133
Low cardiac output syndrome	2	1	1	0.253

Data were expressed as mean values and standard deviations for continuous variables or as percentages for categorical data. In order to identify the predictors of postoperative in-hospital mortality in the groups TOTAL and AVR, unpaired two-tailed Fisher's exact tests, allowing for common surgical and cardiovascular risk factors, were utilized. Significance was assumed if P-values were <0.05. Data analysis was performed with the SPSS software (Chicago, IL, USA), version 11.5.

RESULTS

Postoperative course

In the subgroup of nonagenarians after AVR (group AVR), an equivalent pattern of postoperative morbidity (Fig. 1) and ventilation as well as hospital stay length (Fig. 2) was observed, when compared with the group of all nonagenarians who underwent cardiac surgery (group TOTAL). In the subgroup AVR, the in-hospital mortality rate was 13% (n = 4), when compared with the 10% (n = 5) in-group TOTAL (Fig. 3).

Figure 1:

After cardiac surgery (group TOTAL), and AVR in particular (group AVR), nonagenarians showed the displayed complication rates. CV: cardiovascular; PULM: pulmonary; NEUR: neurological; GI: gastrointestinal.

Figure 2:

After cardiac surgery (group TOTAL), and AVR in particular (group AVR), nonagenarians showed the displayed ventilation (VENT), ICU stay (ICU) and hospital stay (HOSP) periods. Referring to the second y-axis, the broken line separates the corresponding bars.

Figure 3:

In-hospital mortality of nonagenarians after cardiac surgery (group TOTAL), and AVR in particular (group AVR). Broken lines display the range of mortality for cardiac surgery in nonagenarians, as reported in the literature [2].

Predictors of a complicated course in nonagenarians

Multiple Fisher's exact tests were conducted in order to reveal the predictors of postoperative in-hospital mortality in nonagenarians. Since the study was focused on patients undergoing AVR and the results in this subgroup were representative for all study patients, only the analysis in the subgroup AVR is displayed in Table 2. Most general preoperative risk factors of postoperative complications (e.g. renal failure, low cardiac output syndrome, New York Heart Association Class IV, emergent or redo operation, or myocardial infarction) were remarkably more frequent in patients who died after the operation. However, the small cohort of these non-surviving nonagenarians did not allow for revealing significant intergroup differences.

DISCUSSION

Nonagenarians are an increasing cohort of cardiovascular patients and, therefore, clinical research on the adequate treatment of cardiovascular diseases in the very elderly is warranted.

Due to multiple comorbidities and the low potential for compensating perioperative complications, cardiac surgery in nonagenarians implies an enhanced risk of morbidity and mortality. Correspondingly, the present study resulted in an overall mortality of 11 and 14% for AVR patients, respectively. In their report on behalf of the German Society for Thoracic and Cardiovascular Surgery, Gummert et al. [1] published procedure-related mortality rates of 8% for transvascular as well as transapical AVR in 3629 German TAVI patients and 3% for conventional AVR. However, only a small percentage of these patients was aged over 90, so that their procedure-related risk was naturally lower than in our study cohort of nonagenarians. According to Speziale et al. [2] , the mortality associated with cardiosurgical operations in patients aged over 90 ranges from 7 to 18%, which is consistent with the mortality of 11% in the present investigation. The main reason for the worse outcome of nonagenarians, when compared with younger patients, may be an enlarged number of comorbidities such as chronic obstructive pulmonary disease, former neurological events, heart failure or renal failure. Moreover, our patient cohort had a large amount of heart valve operations, combined procedures and emergent cases. Another explanation for the impaired outcome of nonagenarians might be that, in case of a known cardiac disease, the referral to a cardiosurgical centre is delayed due to age and the physical state of a patient. That such delay may further impair the operative outcome seems to be obvious and was recently shown by Pierard et al. [7]. The observers prospectively included 163 octogenarians with severe aortic stenosis and an indication for operation according to the guidelines. Those 40% of the patients, who refused or were denied operative AVR, showed a 5-year mortality of 69%, while it was only 34% after surgery (P < 0.001). In the present study, nonagenarians suffered from remarkable postoperative cardiovascular, renal, pulmonary, neurological, gastrointestinal and wound complications. Nevertheless, the observed mean intensive care unit stay time period of 3.2 days was still short, when compared with 10.2 or 12.0 days, recently reported by other investigators [2, 4].

Since cardiac surgery in nonagenarians is associated with increased morbidity and mortality, it is desirable to elucidate the preoperative predictors of an adverse outcome. Defining these variables would enable a reasonable selection of patients undergoing surgery. Statistical analysis of our study cohort revealed that patients who died after the operation, when compared with surviving patients, showed higher incidences of most of the common risk factors of postoperative complications. However, due to the few cases of nonagenarians not surviving the in-hospital period, these findings did not reach significance. A study in octogenarians with isolated coronary artery bypass grafting showed that a decreased left ventricular ejection fraction (LVEF) and the necessity of emergent surgery are independent risk factors of operative mortality [8]. Ullery et al. [3] reported chronic renal insufficiency and a decreased LVEF to be predictors of mortality of cardiosurgical nonagenarians. With regard to AVR in patients aged over 80, age, male gender, postoperative myocardial infarction, urgency status, dialysis, low LVEF and mean aortic gradient were shown to be risk factors of impaired survival [5].

In patients with aortic stenosis, the choice of the heart valve substitute also affects the outcome. The implantation of bioprosthetic valves avoids the need for life-long oral anticoagulation and, therefore, it reduces the related risk of severe bleeding, which is all the more noteworthy in the very elderly with impaired liver function or reduced therapeutical compliance. Moreover, bioprostheses were shown to have a slower rate of structural deterioration in these patients [9, 10].

The major limitations of our study are its retrospective design and the lack of direct control groups who were treated conventionally or interventionally. Furthermore, the study population is heterogeneous with regard to the conducted surgery. However, this aspect is due to the small number of nonagenarians undergoing cardiac surgery.

In summary, our study shows that cardiac surgery and particularly conventional AVR can be performed in nonagenarians with acceptable outcomes. As responsible, evidence-based patient selection is crucial in the very elderly, further studies including larger patient cohorts are necessary to define the statistically significant predictors of postoperative complications and in-hospital mortality.

Conflict of interest: none declared.

APPENDIX. CONFERENCE DISCUSSION

Dr P. Gerometta (Milan, Italy): May I ask you one question? This paper really addresses a population that is constantly increasing. I saw in your slides that in Germany, 13% or more of the population are aged over 80 years, so it is really a population that it is not negligible. Do you have any information on the mid-term survival of the patients? I mean, how many of them do survive the 30-day mortality term? Do you have any additional information about this cohort of patients?

Dr Boeken: I was sure that this question would come up, and we finalized this study shortly before the deadline for this meeting. Subsequently we tried to contact the patients, and found that the first ten we called were no longer alive, so it is somewhat difficult to find out when they died and to assess the mid-term outcome. Up until now, we do not have long-term data, but we are trying to complete this.

Dr C. Kik (Rotterdam, Netherlands): I noted that a lot of the patients were in New York Class IV. Have you any idea why the cardiologists did not refer these patients earlier?

Dr Boeken: It is difficult to analyze retrospectively why the patients were referred to us and not to interventional therapy by the cardiologists. I cannot really answer this question.