Long-term bleeding events after mechanical aortic valve replacement in patients under the age of 60

Background Although younger patients are supposed to be less susceptible to bleeding complications of mechanical aortic valve replacement (mAVR) than older patients, there is a relative paucity of data on this subject. Therefore, it remains uncertain whether younger patients are really at a lower risk of these complications than older patients. Methods Incidence rates of bleeding events during 15 years of follow-up after mAVR were compared between 163 patients under 60 (group I), 122 patients between 60 and 65 (group II), and 145 patients over 65 (group III) years of age at operation. The target international normalised ratio (INR) was 3.0–4.0. Results During 15 years of follow-up, the annual incidence rate of major bleeding events (excluding haemorrhagic stroke) was lower in the youngest as compared with the oldest group (3.0 versus 4.7 %, respectively; p = 0.030). However, the annual incidence rate of haemorrhagic stroke was as high in the youngest as in the two older groups (0.6 versus 0.7 % and 0.7 %, respectively; p = 0.928). Conclusions With a target INR of 3.0–4.0, patients under 60 years of age are at equally high risk of haemorrhagic stroke after mAVR as older patients. This finding confirms the relevance of a lower target INR as used in international guidelines.


Introduction
Mechanical aortic valve replacement (mAVR) is generally reserved for patients under the age of 60 years because of the durability of a mechanical prosthesis and a supposed lower susceptibility of younger patients to bleeding complications of oral anticoagulation therapy [1][2][3][4][5]. However, there is a relative paucity of data on long-term bleeding events after mAVR in patients under 60 years [6][7][8]. Therefore, it remains uncertain whether younger patients are really at lower risk of these complications than older patients. We aimed to compare incidence rates of bleeding events between patients under 60 and those over 60 years of age during 15 years of follow-up after mAVR.

Study design
In this retrospective longitudinal cohort study, 430 patients were followed for 15 years after mAVR, which was performed in the St. Antonius Hospital in Nieuwegein, the Netherlands, between 1990 and 1994. Incidence rates of bleeding events, occurring after discharge from hospital, were compared between three groups of patients: 163 patients under 60 (group I), 122 patients between 60 and 65 (group II), and 145 patients over 65 (group III) years of age at operation. Target international normalised ratio (INR) of oral anticoagulation therapy was 3.0-4.0, which was the standard at that time regarding mechanical aortic prostheses [9]. During follow-up, target INR did not change [10]. Data were obtained from our own or the referring cardiology departments, general practitioners, and telephone calls to patients and relatives. INR values within 48 h of the bleeding events, except the minor ones, were retrieved from the regional thrombosis services. The study object was agreed upon by the Hospital Committee on Ethics and Medical Experiments.

Definitions
Bleeding events were divided into minor and major bleeding and haemorrhagic stroke events. Definitions were based on the official guidelines for reporting mortality and morbidity after cardiac valve interventions [11] and defined as follows. Minor bleeding: bleeding not requiring admission or blood transfusion. Major bleeding: fatal or nonfatal bleeding requiring admission or blood transfusion, excluding haemorrhagic stroke. Haemorrhagic stroke: focal neurological deficit of sudden onset as diagnosed by a neurologist, lasting more than 24 h and caused by cerebral bleeding.

Data analysis
Calculation of late overall mortality was performed by Kaplan-Meier analysis. To calculate incidence rates of first bleeding events, Kaplan-Meier cumulative incidence rates were computed, whereas formal hypothesis testing was done by means of the log-rank test. To calculate incidence rates of multiple events (up to three per patient for minor or major bleeding events, and up to two for haemorrhagic stroke), linearised annual incidence rates (% per year, with exact 95 % confidence intervals [CI]) were computed, whereas formal hypothesis testing was done by means of an exact method.

Baseline characteristics
Baseline characteristics are depicted in Table 1. The youngest patient was 21 and the oldest 80 years of age at operation. In all three groups more male than female patients were operated upon. In patients under 60 years at operation, mAVR was more often performed because of aortic regurgitation, as compared with aortic stenosis in the older groups. None of the patients had a history of haemorrhagic stroke.

Mechanical protheses
The different types of implanted mechanical aortic valve prostheses are depicted in Table 2. They were similarly distributed among the three groups, half of them being bileaflet and the other half tilting disc prostheses.

Follow-up
Mean follow-up after mAVR was 18.1±1.2 years. All patients were followed for at least 15 years after operation or until death. Follow-up was complete in all patients.

Bleeding events
Incidence rates of bleeding events are depicted in Table 3. Total numbers of patient-years of follow-up were 2479, 1541, and 1481 years in group I, II, and III, respectively. During 15 years of follow-up, the annual incidence rate of major bleeding events (excluding haemorrhagic stroke) was lower in the youngest as compared with the oldest group (3.0 versus 4.7 %, respectively; p=0.030). However, the annual incidence rate of haemorrhagic stroke was as high in the youngest as in the two older groups (0.6 versus 0.7 % and 0.7 %, respectively; p=0.928).  Values are presented as N (%)

Discussion
In this study, incidence rates of bleeding events during 15 years of follow-up after mAVR were determined in a group of patients under 60 years of age, and compared with two groups of patients aged between 60 and 65 and over 65 years at operation, respectively. The patients under 60 years of age were not at low risk of long-term bleeding complications as compared with the older patients. Although the annual incidence rate of major bleeding events (excluding haemorrhagic stroke) was lower in the youngest as compared with the oldest group (3.0 versus 4.7 %, respectively; p=0.030), the annual incidence rate of haemorrhagic stroke was as high in the youngest as in the two older groups (0.6 versus 0.7 % and 0.7 %, respectively; p=0.928). We do not know why the patients under 60 years, despite their younger age, were at equally high risk of haemorrhagic stroke as the older patients. There were no suggestions of a selection of younger patients more prone to bleeding, because risk factors for bleeding (female gender, renal failure, hypertension, history of ischaemic stroke or gastrointestinal bleeding) were not more common in the youngest than in the older groups (  [18,19]. Current [20,21] and past [22] international guidelines recommend a target INR of 2.0-3.0 to 2.5-3.5 (depending on prosthesis thrombogenicity and patient-related risk factors for thromboembolism) for most mechanical aortic valves of the last decades, including the prostheses used in the present study. These target INR values are based on studies, including three major randomised trials [23][24][25], weighing thromboembolic risks against haemorrhagic risks in patients on oral anticoagulation therapy after mechanical heart valve replacement. Although we do not know what the bleeding figures in our patients would have been if the target INR had been as low as in the international guidelines, it seems plausible that with a lower target INR the incidence rates of both major bleeding and haemorrhagic stroke events would have been lower in all three age groups. An important shortcoming of the present study is the high percentage (30.6 %) of unavailable INR values within 48 h of the major bleeding and haemorrhagic stroke events. It is therefore not known how many patients might have suffered these events due to an excessively high INR. However, the INR values which were available within 48 h of the major bleeding or haemorrhagic stroke events were not excessively high.

Conclusions
With a target INR of 3.0-4.0, patients under 60 years of age are at equally high risk of haemorrhagic stroke after mAVR as older patients. This finding confirms the relevance of a lower target INR as used in international guidelines.