**3. Results**

Between the end of 1986 and 1991, 80 patients received a Carpentier-Edwards valve in aortic position. This number was 216 between 1992 - 1996, 345 between 1997 - 2001 and 365 between 2002 and the end of 2006. In our series, the changes in the four 5-year periods showed a significant increase in time for


All increases were highly significant (p<0.001). Only for the need of digitalis, a significant decrease was observed: from 39.7% to 9.4% (p<0.001). These results show that patients referred for AVR became older and have more co-morbidity. The reduction of use of digitalis could be due to the introduction of ACE inhibitors, angiotensin receptor, renin and beta blocking agents.

The hospital results showed a completely different pattern. Only for non-cardiac complications, a significant increase (p=0.001) has been found over time, from about 20% in the first two five-year periods to about 30% in the last two five-year periods. Mortality and major cardiovascular complications rates such as ventricular arrhythmias, bleeding, congestive heart failure and thromboembolism all remained well below 5% throughout these observation periods, without a noteworthy increase. Atrial fibrillation remained between 30 and 40%, without an increase over time.

Although patients referred for AVR became older and sicker, 30-day postoperative survival seemed not affected. Therefore, age and the presence of co-morbid conditions do not necessarily represent formal contra-indications for AVR. Nevertheless, an increase of noncardiac postoperative complications with concomitant increase in length of stay and use of economic resources could be expected if older and sicker patients are operated upon.

### **Hospital or 30 day mortality**

Significant preoperative factors for mortality in an unvariate analysis were:


The other risk factors had no significant effect. NYHA functional class II had a protective effect (p<0.001) for mortality.


Multivariate analysis showed following results

110 Special Topics in Cardiac Surgery

documented on ECG (n=381)

with 0.3 mg% (n=53)

showed a significant increase in time for - patients over 80 (from 6,3% to 25,5%),



between 30 and 40%, without an increase over time.


beta blocking agents.

**Hospital or 30 day mortality** 

effect (p<0.001) for mortality.


**3. Results** 

atrial fibrillation, new or recurrence of a previously paroxysmal atrial fibrillation,

acute renal function impairment, documented by an increase of plasma creatinine

pulmonary complications: clinical and radiological signs of atelectasis or

Between the end of 1986 and 1991, 80 patients received a Carpentier-Edwards valve in aortic position. This number was 216 between 1992 - 1996, 345 between 1997 - 2001 and 365 between 2002 and the end of 2006. In our series, the changes in the four 5-year periods

All increases were highly significant (p<0.001). Only for the need of digitalis, a significant decrease was observed: from 39.7% to 9.4% (p<0.001). These results show that patients referred for AVR became older and have more co-morbidity. The reduction of use of digitalis could be due to the introduction of ACE inhibitors, angiotensin receptor, renin and

The hospital results showed a completely different pattern. Only for non-cardiac complications, a significant increase (p=0.001) has been found over time, from about 20% in the first two five-year periods to about 30% in the last two five-year periods. Mortality and major cardiovascular complications rates such as ventricular arrhythmias, bleeding, congestive heart failure and thromboembolism all remained well below 5% throughout these observation periods, without a noteworthy increase. Atrial fibrillation remained

Although patients referred for AVR became older and sicker, 30-day postoperative survival seemed not affected. Therefore, age and the presence of co-morbid conditions do not necessarily represent formal contra-indications for AVR. Nevertheless, an increase of noncardiac postoperative complications with concomitant increase in length of stay and use of

P<0.001: need for urgent surgery, age over 80, decreased left ventricular function, P<0.01: renal function impairment, LVEF below 50%, previous AMI, previous

The other risk factors had no significant effect. NYHA functional class II had a protective

economic resources could be expected if older and sicker patients are operated upon.

Significant preoperative factors for mortality in an unvariate analysis were:

heart failure, need for digitalis, P<0.05: diabetes, chronic or paroxysmal atrial fibrillation,

respiratory infection or prolonged intubation (n=58).


Mortality in the hospital phase was the most important outcome. It varied between 1.5% (8) and 24% (9). These differences were due to large differences in patient characteristics such as age and co-morbid conditions. With a time span of almost 20 year between the first and the last publication, improvement in surgical techniques and peri-operative care could also be responsible for these differences in mortality.

In most series, the hospital mortality was below 10%. In series where the mortality was over 10%, several risk factors were usually present. Independent predictors for hospital mortality were identified and confirmed that specific co-morbid conditions increased the early postoperative risk.

Most of these factors could be related to the left ventricle and hence to the patient. An emergency need for AVR (i.e. to operate on the same day as the admission in order for the patient to survive) has been identified as the most important predictor, with a increase of mortality of 10 times (10). This has been confirmed in other series (11,12). This indicated to an exhaustion of all compensatory mechanisms to maintain an adequate circulation. A need for urgent AVR has also been identified as a predictor for early postoperative congestive heart failure, which is a highly lethal condition (13). A high preoperative functional class NYHA IV (14,15) and a low-flow low-gradient problem also could be related to a protracted burden, and hence a decreased left ventricular function. Coronary artery disease, previous and the need for concomitant CABG (9,11,12,15) as well as a previous myocardial infarction (10,11) and previous CABG (12) could add to a decrease in left ventricular function.

Valvular factors such as severity of valvular disease and the type and size of valve prosthesis implanted also had an effect (15,16). The effect of non-cardiac factors such as diabetes (15) and renal disease (11,12,17) was also observed. Remarkably, the effect of age over 80 (9,12,18,19), although important, was less compared to the effect of need for urgent surgery on mortality(10).

### **Thromboembolic events**

TE events were one of the most important and devastating events after AVR, especially if permanent neurological deficit was present. One also has to keep in mind that many TE events go unnoticed. In one small series, MRI after cardiac surgery could document silent events in 6 of 34 patients, which is rather high (20). We found this clinically evident thromboembolism in 27 of 1000 patients. In 25 cases, this was neurological. An univariate analysis identified a decreased left ventricular function (p<0.01) as sole risk factor.

This risk factor was confirmed in a multivariate analysis, which showed an ejection fraction below 50% as an independent predictor (11/247 patients), with p=0.027, an odds ratio of 2.5 with a 95% confidence interval between 1.1-5.7.

Some studies reported on the predictors for thromboembolism after AVR on long-term and none reported on such events on short-term. The short-term thromboembolic events, however, have their importance since these are a predictor for future events (21). A preoperative CVA seemed to have a comparable significance (21-24) for long-term events.

The 30 Day Complication Rate After Aortic Valve

predictors or risk factors could be identified (30,31).

Literature data only described long-term events.

endocarditis has been the main indication for re-operation (45).

Univariate analysis identified following risk factors:

NYHA class II had a protective effect (p<0.001) Multivariate analysis showed following results

**Prosthetic valve endocarditis** 

**Congestive heart failure** 

Replacement with a Pericardial Valve in a Mainly Geriatric Population 113

Bleeding events could be considered as the other side of the coin of anticoagulation with vitamin K antagonists. Effective protection against thromboembolism holds the risk for bleeding. A continuous balance should be made. The INR does not only depend on the dose of vitamin K antagonists, but also on the adsorption of vitamin K by the mucosa of the colon and on its processing by the liver. No other reports concerning predictors for early bleeding after valve replacement appeared. Current results indicated that early bleeding seems related to the procedure and is different compared to bleeding events at long term, which is probably more related to anticoagulation. Nevertheless, other factors such as age and increased cardiothoracic index were identified as predictors for bleeding during long-term follow-up (26). Fragility in elderly might be an important reason. In other series, no

The number of patients with early prosthetic valve endocarditis was low, hence no risk factors or predictors (such as preoperative endocarditis) could be identified in our series.

The linearized rate for long-term prosthetic valve endocarditis was low and mostly under 1% per patient year (26,27,32-39). The most event free rates after 5 to 25 year were well over 90%. This was true for mechanical heart valves (35,40,41) as well as for the different stented (30,42) and stentless tissue valves (27,36,43) as for homografts (8). No risk factors could be identified, although a previous endocarditis could arouse some suspicion. Prosthetic valve endocarditis carried a high risk for mortality, especially if Staphylococcus aureus has been detected (44). These micro-organisms often lead to ring abscesses and prosthetic paravalvular leak. The occurrence of congestive heart failure with prosthetic valve

P<0.001: Need for urgent surgery, chronic or paroxysmal atrial fibrillation P<0.01: Age over 80, decreased left ventricular function, previous infarction P<0.05: COPD, LVEF below 50%, preoperative heart failure, need for digitalis

**Factor n/N p OR 95%CI**  Urgent AVR 7/25 <0.001 10.5 3.6-30.8 Atrial fibr. 16/197 0.001 3.5 1.7-7.4 EF<50% 10/155 0.055 2.1 1.0-4.4

In-hospital congestive heart failure after AVR is a highly lethal complication (13,46). Hospital heart failure occurred in 2.6% of the patients in an earlier series (13), with need for urgent AVR as the sole independent predictor. This pointed to an exhaustion of all compensatory mechanisms to maintain an adequate circulation and this observation should reason not to postpone AVR, once it has become symptomatic. In another study, the level of B-type natri-uretic peptide was also identified as a predictor (46). Apoptosis or programmed cell death of cardiomyocytes could be an important event in these patients (47). Atrial fibrillation could lead to a decreased ventricular filling and cardiac output since the atrial contraction is lost. In an earlier series, this was also the case in long-term heart failure (13).

A low ejection fraction also has been identified as predictor (24). Congestive heart failure and a dilated ventricle are known risk factors for thromboembolism. Damage to the ventricular wall as well as an abnormal flow of blood (and possible stasis) can promote intra-ventricular thrombosis and hence embolization. If the LVEF is below 30%, life-long anticoagulation is required (25).

Age also has been identified as an independent predictor (23-27). The age distribution of any given patient population should be made known to appreciate fully the effect of this factor. Age might be related, however to co-morbid conditions such as diabetes and atheromatosis of the aorta and the cerebral vessels (23,24).

Carotid artery disease could be a matter for debate. In some series (23,24), it was identified as a predictor for thromboembolism, in other series, this was not (21). A Doppler-duplex investigation is a reliable tool for the detection of lesions and is routinely performed in some institutions. It cannot, however, detect atheromatosis of the intracranial vessels.

A long ECC time (24) might indicate to coronary artery disease (28) and hence, to the need for an associated procedure, which mostly is a CABG. This could confirm atheromatous vessels as a possible source for thromboembolic events. These cannot be considered entirely as valve related. A "smoking gun", however, is often not found. And CABG itself has never been identified as a predictor for thromboembolism.

AF could also be a matter for debate: it is certainly a risk factor for thromboembolic events. However, it is also an indication for anticoagulation, which might cloud the effect of AF. In some series, AF was identified as a predictor (28), in others not (21,24).

It is by no means certain that risk factors for long-term thromboembolic events could be used to predict the occurrence of such events on short term. The latter, however, have their prognostic significance. Moreover, all these predictors indicate that the sources of thromboembolic events are multiple. Hence thromboembolism cannot be considered as entirely valve related (21,29), but also as patient related.

To prevent thromboembolism after AVR with a biological prosthesis, anticoagulant medication is given for three months after implantation, after which it can be replaced by acetyl salicylic acid, with good results (21), unless anticoagulants are indicated for another reason. An RCT to support this strategy does not exist (25), but the general idea is that vitamin K antagonists protect against early thromboembolism while re-endothelialization of the stent and the sutures is not yet complete. For mechanical heart valves, a life-long anticoagulant treatment is necessary. The target level for INR is determined by the valve position (for aortic this is lower than for mitral) and by the type of mechanical heart valves. Older types such as cage ball devices are more thrombogenic and require deeper anticoagulation. A strict classification for thrombogenicity is not available. There are too many patient related factors involved (25).

### **Bleeding events**

Univariate analysis revealed three factors with a significant effect. This was confirmed by a multivariate analysis. The use of vitamin K antagonists could not be included since almost all patients received this medication for three months after operation. The three predictors were:


Bleeding events could be considered as the other side of the coin of anticoagulation with vitamin K antagonists. Effective protection against thromboembolism holds the risk for bleeding. A continuous balance should be made. The INR does not only depend on the dose of vitamin K antagonists, but also on the adsorption of vitamin K by the mucosa of the colon and on its processing by the liver. No other reports concerning predictors for early bleeding after valve replacement appeared. Current results indicated that early bleeding seems related to the procedure and is different compared to bleeding events at long term, which is probably more related to anticoagulation. Nevertheless, other factors such as age and increased cardiothoracic index were identified as predictors for bleeding during long-term follow-up (26). Fragility in elderly might be an important reason. In other series, no predictors or risk factors could be identified (30,31).

### **Prosthetic valve endocarditis**

112 Special Topics in Cardiac Surgery

A low ejection fraction also has been identified as predictor (24). Congestive heart failure and a dilated ventricle are known risk factors for thromboembolism. Damage to the ventricular wall as well as an abnormal flow of blood (and possible stasis) can promote intra-ventricular thrombosis and hence embolization. If the LVEF is below 30%, life-long

Age also has been identified as an independent predictor (23-27). The age distribution of any given patient population should be made known to appreciate fully the effect of this factor. Age might be related, however to co-morbid conditions such as diabetes and atheromatosis

Carotid artery disease could be a matter for debate. In some series (23,24), it was identified as a predictor for thromboembolism, in other series, this was not (21). A Doppler-duplex investigation is a reliable tool for the detection of lesions and is routinely performed in some

A long ECC time (24) might indicate to coronary artery disease (28) and hence, to the need for an associated procedure, which mostly is a CABG. This could confirm atheromatous vessels as a possible source for thromboembolic events. These cannot be considered entirely as valve related. A "smoking gun", however, is often not found. And CABG itself has never

AF could also be a matter for debate: it is certainly a risk factor for thromboembolic events. However, it is also an indication for anticoagulation, which might cloud the effect of AF. In

It is by no means certain that risk factors for long-term thromboembolic events could be used to predict the occurrence of such events on short term. The latter, however, have their prognostic significance. Moreover, all these predictors indicate that the sources of thromboembolic events are multiple. Hence thromboembolism cannot be considered as

To prevent thromboembolism after AVR with a biological prosthesis, anticoagulant medication is given for three months after implantation, after which it can be replaced by acetyl salicylic acid, with good results (21), unless anticoagulants are indicated for another reason. An RCT to support this strategy does not exist (25), but the general idea is that vitamin K antagonists protect against early thromboembolism while re-endothelialization of the stent and the sutures is not yet complete. For mechanical heart valves, a life-long anticoagulant treatment is necessary. The target level for INR is determined by the valve position (for aortic this is lower than for mitral) and by the type of mechanical heart valves. Older types such as cage ball devices are more thrombogenic and require deeper anticoagulation. A strict classification for thrombogenicity is not available. There are too

Univariate analysis revealed three factors with a significant effect. This was confirmed by a multivariate analysis. The use of vitamin K antagonists could not be included since almost all patients received this medication for three months after operation. The three predictors

**Factor n/N p OR 95%CI**  Concomitant CABG 17/610 0.046 3.6 1.0-12.5 COPD 8/235 0.051 2.6 1.0-6.6 Aortoplasty 4/61 0.058 3.5 1.0-12.6

institutions. It cannot, however, detect atheromatosis of the intracranial vessels.

some series, AF was identified as a predictor (28), in others not (21,24).

anticoagulation is required (25).

of the aorta and the cerebral vessels (23,24).

been identified as a predictor for thromboembolism.

entirely valve related (21,29), but also as patient related.

many patient related factors involved (25).

**Bleeding events** 

were:

The number of patients with early prosthetic valve endocarditis was low, hence no risk factors or predictors (such as preoperative endocarditis) could be identified in our series. Literature data only described long-term events.

The linearized rate for long-term prosthetic valve endocarditis was low and mostly under 1% per patient year (26,27,32-39). The most event free rates after 5 to 25 year were well over 90%. This was true for mechanical heart valves (35,40,41) as well as for the different stented (30,42) and stentless tissue valves (27,36,43) as for homografts (8). No risk factors could be identified, although a previous endocarditis could arouse some suspicion. Prosthetic valve endocarditis carried a high risk for mortality, especially if Staphylococcus aureus has been detected (44). These micro-organisms often lead to ring abscesses and prosthetic paravalvular leak. The occurrence of congestive heart failure with prosthetic valve endocarditis has been the main indication for re-operation (45).

### **Congestive heart failure**

Univariate analysis identified following risk factors:


NYHA class II had a protective effect (p<0.001)

Multivariate analysis showed following results


In-hospital congestive heart failure after AVR is a highly lethal complication (13,46). Hospital heart failure occurred in 2.6% of the patients in an earlier series (13), with need for urgent AVR as the sole independent predictor. This pointed to an exhaustion of all compensatory mechanisms to maintain an adequate circulation and this observation should reason not to postpone AVR, once it has become symptomatic. In another study, the level of B-type natri-uretic peptide was also identified as a predictor (46). Apoptosis or programmed cell death of cardiomyocytes could be an important event in these patients (47). Atrial fibrillation could lead to a decreased ventricular filling and cardiac output since the atrial contraction is lost. In an earlier series, this was also the case in long-term heart failure (13).

The 30 Day Complication Rate After Aortic Valve

hence a longer hospital stay (60)

by suturing (58).

A multivariate analysis showed following results:

which illustrates the importance of these defects.

Univariate analysis showed following risk factors P=0.001: postoperative heart failure P<0.01: preoperative PM implantation

Multivariate analysis identified following predictors:

prolonged ventilation, respiratory failure, pulmonary infection).

Pulmonary complications occurred in 58/1000 patients after AVR (66).

**Non cardiac hospital complications** 

P<0.05: COPD

Replacement with a Pericardial Valve in a Mainly Geriatric Population 115

**Factor n/N p OR 95%CI**  Preoperative heart failure 32/215 0.001 3.0 1.7-5.4 Concomitant CABG 69/610 0.007 2.0 1.2-4.2 NYHA class IV 23/249 0.039 1.9 1.0-3.4 AV Block grade 1 7/37 0.078 2.4 1.0-5.7

Conduction defects such as complete atrio-ventricular block, and hence permanent pacemaker implantation, occurred in almost 10% of the patients who underwent AVR (58). The occurrence of all new conduction defects is reported to be 15% (59). Other reports state a lower incidence for permanent pacemaker implantation, since not all conduction defects need this treatment. Sometimes, a new conduction defects could be reversible (58). Development of a new conduction defects resulted to an increased need for monitoring and

The main predictor for permanent pace maker implantation in other series were preoperative conduction defects (58,61-63). Other predictors were female gender, annular calcification, bicuspid aortic valve, hypertension, myocardial infarction, electrolyte imbalance and prolonged cross-clamping time (60,64,65). Aortic regurgitation might also

Preoperative changes in the conduction system might be degenerative and due to older age,

Some factors such as a calcified annulus could lead to trauma of the conduction system during surgery. The link with a congenitally bicuspid aortic valve is less clear. Hypertension could lead to calcification but also to increased septal left ventricular hypertrophy, which makes the conduction system more difficult to protect during cross-clamping (60). Other traumas of the conduction system could be due to impingement by the prosthetic valve or

Development of a new conduction defects such as left bundle branch block in the postoperative period was a marker for future adverse events, such as sudden death (59),

**Factor n/N p OR 95%CI**  Pacemaker implant 7/33 0.002 4.4 1.8 – 11.2 COPD 21/235 0.073 1.7 0.95 – 3.1 Heart failure 7/34 0.001 4.7 1.8 – 11.9

Respiratory complications after AVR are common (9,66,67) but few reports appeared concerning their predictors. Their rate depends largely on the criteria used (need for

play a role since this condition usually required implantation of larger valves (58).

to ischemia and mechanical factors, due to an increased left ventricular pressure (58).

Early congestive heart failure should be distinguished from long-term heart failure: the latter was diastolic in nature and has some other risk factors such as conduction defects and coronary artery disease. An impaired relaxation occurred since after regression of the muscle mass, this fibrosis persists for a longer time. This stiffened the LV wall (48). The LVEF remained often normal. Hence, this could be labeled as a "diastolic" form of CHF (13). Concurrent conduction defects thereby could lead to perfusion defects and wall motion abnormalities, even in absence of coronary artery disease (49). Biventricular pacing often could correct this condition (50,51). Coronary artery disease could lead to ischemic loss of myocardial contractility, adding to the left ventricular function. Concomitant CABG, however was not identified as a predictor for postoperative CHF (13).

A small valve size has not been identified as a predictor for long-term heart failure in a previous series (13), neither in the current series for short-term heart failure. Hence, in patients with a small aortic root, it seemed not necessary to enlarge the root or to implant a stentless valve, which requires a more demanding technique, provided the LVF is normal.

However, it becomes important in patients with a low ejection fraction. When, in such patients, a small valve size leads to an even moderate prosthesis-patient mismatch, this might result in heart failure and an increased mortality (52).

### **Atrial fibrillation**

Atrial fibrillation and other non-sinusal rhythms occurs in 20% of the postoperative patients (32). In other series, this was 40% or more (53,54). The definition and the diagnostic method for AF have also an effect on its incidence (54,55). Postoperative AF has not been considered as innocent: it could lead to other complications and an increased stay in the hospital. With atrial fibrillation, there was a higher incidence of postoperative mortality (54) heart failure, renal function impairment, infection and neurologic events (53).

Univariate analysis showed following risk factors


In a multivariate analysis, previous PTCA was the only independent predictor (p=0.006; odds ratio 2.7; 95% confidence interval 1.3-5.4)

The risk factors for postoperative AF in other series were a history of paroxysmal atrial fibrillation, a large left atrium, a prolonged P wave, heart failure, high age, low ejection fraction and left ventricular hypertrophy (55-57). The effect of age, however remained a matter for debate: atrial fibrillation did not occur more frequently in octogenarians after AVR (27).

It seems that postoperative atrial fibrillation was preventable, at least in part. Pacing could half the frequency of this event (53). Timely and adequate treatment could increase the speed of postoperative rehabilitation and reduce the stay within the hospital (55).

### **Conduction defects**

An univariate analysis showed as risk factors:


An AV block grade 1 showed a trend. A large valve size (27 mm) had no effect.

A multivariate analysis showed following results:

114 Special Topics in Cardiac Surgery

Early congestive heart failure should be distinguished from long-term heart failure: the latter was diastolic in nature and has some other risk factors such as conduction defects and coronary artery disease. An impaired relaxation occurred since after regression of the muscle mass, this fibrosis persists for a longer time. This stiffened the LV wall (48). The LVEF remained often normal. Hence, this could be labeled as a "diastolic" form of CHF (13). Concurrent conduction defects thereby could lead to perfusion defects and wall motion abnormalities, even in absence of coronary artery disease (49). Biventricular pacing often could correct this condition (50,51). Coronary artery disease could lead to ischemic loss of myocardial contractility, adding to the left ventricular function. Concomitant CABG,

A small valve size has not been identified as a predictor for long-term heart failure in a previous series (13), neither in the current series for short-term heart failure. Hence, in patients with a small aortic root, it seemed not necessary to enlarge the root or to implant a stentless valve, which requires a more demanding technique, provided the LVF is normal. However, it becomes important in patients with a low ejection fraction. When, in such patients, a small valve size leads to an even moderate prosthesis-patient mismatch, this

Atrial fibrillation and other non-sinusal rhythms occurs in 20% of the postoperative patients (32). In other series, this was 40% or more (53,54). The definition and the diagnostic method for AF have also an effect on its incidence (54,55). Postoperative AF has not been considered as innocent: it could lead to other complications and an increased stay in the hospital. With atrial fibrillation, there was a higher incidence of postoperative mortality (54) heart failure,

In a multivariate analysis, previous PTCA was the only independent predictor (p=0.006;

The risk factors for postoperative AF in other series were a history of paroxysmal atrial fibrillation, a large left atrium, a prolonged P wave, heart failure, high age, low ejection fraction and left ventricular hypertrophy (55-57). The effect of age, however remained a matter for debate: atrial fibrillation did not occur more frequently in octogenarians after

It seems that postoperative atrial fibrillation was preventable, at least in part. Pacing could half the frequency of this event (53). Timely and adequate treatment could increase the

speed of postoperative rehabilitation and reduce the stay within the hospital (55).

An AV block grade 1 showed a trend. A large valve size (27 mm) had no effect.

however was not identified as a predictor for postoperative CHF (13).

might result in heart failure and an increased mortality (52).

renal function impairment, infection and neurologic events (53).

Univariate analysis showed following risk factors

odds ratio 2.7; 95% confidence interval 1.3-5.4)

An univariate analysis showed as risk factors: P<0.001: preoperative heart failure

P<0.05: NYHA functional class IV

P<0.01: carotid artery disease, concomitant CABG

P<0.01: Previous PTCA P<0.05: LVEF below 50%

**Atrial fibrillation** 

AVR (27).

**Conduction defects** 


Conduction defects such as complete atrio-ventricular block, and hence permanent pacemaker implantation, occurred in almost 10% of the patients who underwent AVR (58). The occurrence of all new conduction defects is reported to be 15% (59). Other reports state a lower incidence for permanent pacemaker implantation, since not all conduction defects need this treatment. Sometimes, a new conduction defects could be reversible (58). Development of a new conduction defects resulted to an increased need for monitoring and hence a longer hospital stay (60)

The main predictor for permanent pace maker implantation in other series were preoperative conduction defects (58,61-63). Other predictors were female gender, annular calcification, bicuspid aortic valve, hypertension, myocardial infarction, electrolyte imbalance and prolonged cross-clamping time (60,64,65). Aortic regurgitation might also play a role since this condition usually required implantation of larger valves (58).

Preoperative changes in the conduction system might be degenerative and due to older age, to ischemia and mechanical factors, due to an increased left ventricular pressure (58).

Some factors such as a calcified annulus could lead to trauma of the conduction system during surgery. The link with a congenitally bicuspid aortic valve is less clear. Hypertension could lead to calcification but also to increased septal left ventricular hypertrophy, which makes the conduction system more difficult to protect during cross-clamping (60). Other traumas of the conduction system could be due to impingement by the prosthetic valve or by suturing (58).

Development of a new conduction defects such as left bundle branch block in the postoperative period was a marker for future adverse events, such as sudden death (59), which illustrates the importance of these defects.

### **Non cardiac hospital complications**

Pulmonary complications occurred in 58/1000 patients after AVR (66).

Univariate analysis showed following risk factors


Multivariate analysis identified following predictors:


Respiratory complications after AVR are common (9,66,67) but few reports appeared concerning their predictors. Their rate depends largely on the criteria used (need for prolonged ventilation, respiratory failure, pulmonary infection).

The 30 Day Complication Rate After Aortic Valve

minutes, severe angina an non-elective surgery (67,78).

compared to the area with low initial plasma creatinine.

**4. Discussion and conclusions** 

factors are liable for alteration.

(76,77).

Replacement with a Pericardial Valve in a Mainly Geriatric Population 117

Endothelial dysfunction of patients with CAVS could be a link with renal function impairment. This could help explaining the observed association between preoperative AF and postoperative renal complications. There are, however, several confounding factors

Other risk factors were hypertension, peripheral arterial disease, bypass time over 70

Three difficulties could arise by comparing different series. First, the definition of previous renal function impairment as well as of postoperative decrease in renal function varied between series. In one series, a level of 2.0 mg/dl was used (78), in another 1.4 mg/dl (17). This could account for differences in results. The second difficulty has been the estimation of renal function: plasma creatinine , as a routine clinical procedure is not sufficient as estimate for glomerular filtration rate. Hence, its results should be interpreted cautiously. The estimation of glomerular filtration rate requires a 24-h urine collection, which is liable to errors. Third, defining a worsening of renal function by an increase of plasma creatinine (79) also had its difficulties: in patients with a plasma creatinine between 1 and 2 mg and hence, a moderate degree in renal function impairment, an additional increase with 0.3% could mean a serious additional renal damage. In patients with higher initial plasma creatinine , such an increase does not mean necessarily a major change in renal function: the slope of the relation between glomerular filtration rate and plasma creatinine is much less

The extracorporeal circulation could have a damaging effect on the glomeruli, especially if the kidney already has been injured. Mechanisms inflicting renal damage are non-pulsatile perfusion, renal hypoperfusion, hypothermia, and increased levels of circulating catecholamines, cytokines, enzymes, free radicals and free hemoglobin (80). Keeping the cross-clamp time as short as possible or installing a minimal ECC could be helpful (81,82). In spite of increasing age and co-morbid conditions in patients referred for AVR, the increase in hospital complications seemed to be limited to non-cardiac complications, which have a lower fatality rate than cardiovascular complications (83). The occurrence of renal and pulmonary postoperative events should be taken into account, however, if one chooses

Postoperative mortality and valve related complications received much attention in most patients series. Recently, other cardiac complications such as heart failure, conduction defects and atrial fibrillation were also scrutinized. Univariate analysis as screening and subsequent multivariate analysis as identification of risk factors for each event could be helpful patient selection and, more importantly, improve postoperative results if these risk

Mortality and congestive heart failure, which was identified as the most lethal cardiac complication, were clearly patient related. One might expect a reduction in these events if patients are referred early, once aortic valve degeneration has become symptomatic. This could avoid the appearance of a major risk factor, i.e. the need for urgent valve replacement. This is the clear consequence of a protracted pressure overload on the left ventricle by the

Thromboembolism and bleeding are typically considered as valve related events. The former, however, could also be related to patient factors, while the latter could also be

diseased aortic valve and the ultimate marker for advanced heart valve disease.

to operate older and sicker patients with symptomatic aortic valve disease.

The basic mechanisms of these complications involved a lack of deep inspiration due to postoperative pain, with a shallow breathing pattern, a prolonged recumbent positioning, a temporary diaphragmatic dysfunction and an impaired mucociliary clearance with a decrease in cough effectiveness, increases the risks associated with retained pulmonary secretions and bronchial obstruction (68). Atelectasis, infection and prolonged stay within the intensive care unit could be the result of such obstruction (69,70).

The use of an extracorporeal circulation during cardiac operations certainly has side effects on the respiratory system, due to inflammation (71,72). With greater hemodilution this effect could increase resulting pulmonary edema and pneumonia, especially when endothelial cells are injured (73). Pulmonary edema was also a hallmark of postoperative heart failure, which has been identified as a clear risk factor for postoperative pulmonary complications.

Chronic obstructive pulmonary disease could worsen the production of mucus, thereby increasing the risk for atelectasis and infection. Early postoperative heart failure, and hence pulmonary edema increase the need for ventilator support and prolong the stay on ICU. This also could make the patient more vulnerable for postoperative pulmonary problems (66).

In 58 of 1000 patients, we observed renal function impairment after AVR (17). Renal and cardiovascular disease could be linked in two ways. On the one hand, renal function impairment, even if this modest, has been an established element in the risk profile for atheromatosis (74,75). On the other hand, renal function impairment could also be the result of atheromatosis and has certainly been observed after major cardiovascular surgery such as AVR. Postoperative decrease of the renal function has been considered a serious complication with an increased mortality rate.

The risk factors in an univariate analysis for this complication were


A multivariate analysis revealed following predictors


Age over 80 has been associated with an increase in atheromatosis and myocardial infarction is certainly a marker for it. If atheromatosis also affects the renal arteries, a postoperative decrease in renal function could be expected (17).

The basic mechanisms of these complications involved a lack of deep inspiration due to postoperative pain, with a shallow breathing pattern, a prolonged recumbent positioning, a temporary diaphragmatic dysfunction and an impaired mucociliary clearance with a decrease in cough effectiveness, increases the risks associated with retained pulmonary secretions and bronchial obstruction (68). Atelectasis, infection and prolonged stay within

The use of an extracorporeal circulation during cardiac operations certainly has side effects on the respiratory system, due to inflammation (71,72). With greater hemodilution this effect could increase resulting pulmonary edema and pneumonia, especially when endothelial cells are injured (73). Pulmonary edema was also a hallmark of postoperative heart failure, which has been identified as a clear risk factor for postoperative pulmonary complications. Chronic obstructive pulmonary disease could worsen the production of mucus, thereby increasing the risk for atelectasis and infection. Early postoperative heart failure, and hence pulmonary edema increase the need for ventilator support and prolong the stay on ICU. This also could make the patient more vulnerable for postoperative pulmonary problems (66). In 58 of 1000 patients, we observed renal function impairment after AVR (17). Renal and cardiovascular disease could be linked in two ways. On the one hand, renal function impairment, even if this modest, has been an established element in the risk profile for atheromatosis (74,75). On the other hand, renal function impairment could also be the result of atheromatosis and has certainly been observed after major cardiovascular surgery such as AVR. Postoperative decrease of the renal function has been considered a serious

Preop. renal impairment 22/108 29/890 <0.001 Age > 80 22/186 31/814 <0.001 Preop. atrial fibrillation 22/197 29/803 <0.001 Preop. pulmonary oedema 22/216 31/781 0.001 Preop. conduction defect 24/270 28/721 0.002 Diabetes 15/149 37/851 0.006 Preop. myocardial infarction 15/151 36/849 0.006 Postop. heart failure 6/34 47/966 0.007 CCT >75 min. (complete procedure) 29/460 7/275 0.015 Previously performed CABG 9/81 42/916 0.018 concomitant CABG 39/610 14/390 0.031 LV ejection fraction<0.50 13/155 31/723 0.033 Previous TIA/CVA 10/108 41/892 0.035

**Factor p OR 95%CI**  Preop. renal impairment <0.001 5.5 2.9 – 10.4 Preop. atrial fibrillation 0.010 2.3 1.2 – 4.2 Age > 80 0.014 2.2 1.2 – 4.1 Myocardial infarction 0.022 2.2 1.1 – 4.4

Age over 80 has been associated with an increase in atheromatosis and myocardial infarction is certainly a marker for it. If atheromatosis also affects the renal arteries, a

the intensive care unit could be the result of such obstruction (69,70).

The risk factors in an univariate analysis for this complication were

complication with an increased mortality rate.

A multivariate analysis revealed following predictors

postoperative decrease in renal function could be expected (17).

Endothelial dysfunction of patients with CAVS could be a link with renal function impairment. This could help explaining the observed association between preoperative AF and postoperative renal complications. There are, however, several confounding factors (76,77).

Other risk factors were hypertension, peripheral arterial disease, bypass time over 70 minutes, severe angina an non-elective surgery (67,78).

Three difficulties could arise by comparing different series. First, the definition of previous renal function impairment as well as of postoperative decrease in renal function varied between series. In one series, a level of 2.0 mg/dl was used (78), in another 1.4 mg/dl (17). This could account for differences in results. The second difficulty has been the estimation of renal function: plasma creatinine , as a routine clinical procedure is not sufficient as estimate for glomerular filtration rate. Hence, its results should be interpreted cautiously. The estimation of glomerular filtration rate requires a 24-h urine collection, which is liable to errors. Third, defining a worsening of renal function by an increase of plasma creatinine (79) also had its difficulties: in patients with a plasma creatinine between 1 and 2 mg and hence, a moderate degree in renal function impairment, an additional increase with 0.3% could mean a serious additional renal damage. In patients with higher initial plasma creatinine , such an increase does not mean necessarily a major change in renal function: the slope of the relation between glomerular filtration rate and plasma creatinine is much less compared to the area with low initial plasma creatinine.

The extracorporeal circulation could have a damaging effect on the glomeruli, especially if the kidney already has been injured. Mechanisms inflicting renal damage are non-pulsatile perfusion, renal hypoperfusion, hypothermia, and increased levels of circulating catecholamines, cytokines, enzymes, free radicals and free hemoglobin (80). Keeping the cross-clamp time as short as possible or installing a minimal ECC could be helpful (81,82).

In spite of increasing age and co-morbid conditions in patients referred for AVR, the increase in hospital complications seemed to be limited to non-cardiac complications, which have a lower fatality rate than cardiovascular complications (83). The occurrence of renal and pulmonary postoperative events should be taken into account, however, if one chooses to operate older and sicker patients with symptomatic aortic valve disease.
