Decision-Making and Outcomes in Severe Symptomatic Aortic Stenosis Erik Charlson 1, Anna T. R. Legedza 2, Mary Beth Hamel 2

Transcription

1 Decision-Making and Outcomes in Severe Symptomatic Aortic Stenosis Erik Charlson 1, Anna T. R. Legedza 2, Mary Beth Hamel 2 1 Mayo Clinic College of Medicine, 2 Division of General Medicine and Primary Care, Harvard Medical School, Boston, USA Background and aim of the study: Aortic stenosis (AS) remains the most common valvular disease of the elderly in the United States. Though valve replacement has proven effective among older adults, decision-making regarding surgery is difficult for these patients and their physicians. Herein, the clinical outcomes and decision-making process for elderly patients with severe symptomatic AS was assessed. Methods: A retrospective cohort study of patients aged 60 years with severe AS was conducted at two large urban teaching hospitals. Severe AS was defined by a mean valve gradient 50 mmhg or valve area <0.8 cm 2 by echocardiogram, and associated symptoms (angina, congestive heart failure, dyspnea, fatigue, or exercise intolerance). Demographic and clinical data and information about decision-making were obtained from inpatient and outpatient medical records. Results: Of the 124 patients studied, 49 (39.5%) had aortic valve replacement (AVR) surgery. In a logistic regression analysis adjusting for gender, comorbidity and baseline functional status, those patients aged <80 years were significantly more likely to have surgery than older patients. Surgery was associated with a large reduction in mortality in all age groups. At one-year follow up, 87.8% of all patients (87.5% of those aged 80 years) who had undergone surgery were alive, while only 54.7% (49.1% of those aged 80 years) who did not receive surgery were alive. Postoperative complications were similar among older and younger elderly patients. Comorbidity and age were the most common reasons for not offering elderly patients valve replacement. Conclusion: The results of the present study showed that AVR surgery improves the survival of elderly patients with severe AS, and patients aged >80 years experience benefits similar to younger patients. Nevertheless, these findings suggest that surgery may not always be offered to elderly patients who might benefit from it. The Journal of Heart Valve Disease 2006;15: Aortic stenosis is a disease of the elderly. Indeed, an estimated 3% of adults aged over 65 years are affected by this condition, making it the most common valvular lesion in this age group (1,2). The well-documented natural history of aortic stenosis has long shown that, in its severe form, the disease has a poor prognosis (3,4). With the onset of the cardinal symptoms of angina, dyspnea or syncope, numerous autopsy studies have confirmed a survival time for patients with aortic stenosis of less than five years (5,6). In a more recent study of 125 patients with symptomatic aortic stenosis who refused aortic valve replacement (AVR) surgery, only 12% of the patients survived for five years (7). Address for correspondence: Mary Beth Hamel MD, MPH, Division of General Medicine and Primary Care, Harvard Medical School, 330 Brookline Avenue, LY- 330, Boston, MA 02215, USA mhamel@bidmc.harvard.edu Other recent studies have consistently shown worse outcomes with medical management than with surgery (8). In fact, when outcomes of surgery patients were compared with age-matched controls, those patients who had surgery were seen to have a similar life expectancy as those who did not have aortic stenosis (9,10). These established benefits of surgery, moreover, have been extended even to the oldest patients with severe aortic stenosis (11-13). Decision-making with regard to AVR surgery remains difficult for both older patients and their physicians, however, as age continues to be cited as a factor in selecting patients for surgical treatment (14,15). In the present study, the clinical outcomes were assessed for a cohort of 124 elderly patients with severe aortic stenosis, which included 73 patients aged 80 years. Outcomes were compared of patients treated surgically to those of patients treated medically, and an examination was made as to how older age was Copyright by ICR Publishers 2006

2 J Heart Valve Dis Aortic stenosis decision-making 313 associated with outcomes. In addition, the study attempted to describe in detail the decision-making processes of both physicians and patients with regard to AVR surgery, and to categorize the various reasons why patients may not be offered surgery. Clinical material and methods Study overview A cohort study was conducted of patients with severe aortic stenosis who were cared for at two urban teaching hospitals. The study involved medical record review and telephone interviews with patients. The study protocol was approved by the hospital s institutional review board (IRB). Data acquisition Study patients were identified from a database containing the results of inpatient and outpatient echocardiograms performed between 1995 and The study population included patients aged 60 years who, at the time of the qualifying echocardiogram, had been diagnosed with severe aortic stenosis. For the purposes of the study, severe aortic stenosis was defined as a mean valve gradient 50 mmhg or a valve area <0.8 cm 2 by echocardiogram, with associated symptoms (angina, congestive heart failure (CHF), dyspnea, syncope, fatigue, or exercise intolerance). In order to collect decision-making and outcomes data, two nurses reviewed electronic and paper hospital records, as well as office notes from the patients primary care physicians, cardiologists, and cardiothoracic surgeons. Inpatient and outpatient data collection was designed to capture information within 60 days before or after the echocardiogram. Relevant data collected during this period included the following: patient demographics; data on echocardiogram and cardiac catheterization results; patients clinical characteristics, including baseline functional status and comorbid illnesses; patient care preferences; free text entries taken from medical records that captured any relevant decision-making passages; final decisions of patients with respect to AVR surgery; and the patient outcomes, including survival, NYHA functional status, and complications from AVR surgery. If a final decision had not been made within 60 days after echocardiography, nurses reviewed subsequent medical records to ascertain the final decision regarding AVR. Information was collected on patient care preferences, free text descriptions of relevant decisionmaking passages, and the final decision of patients with regard to surgery. Finally, in order to record subsequent information not found in patient medical records, a follow up interview with all patients still alive in the spring of 2001 was undertaken. After notifying the study patients physicians of the interview, an attempt was made to contact patients by telephone. A professional interviewer conducted a survey to collect information on the patients surgical history, present functional status, and current living situation. Clinical and demographic characteristics Patients clinical and demographic characteristics were collected from medical records. The patient s age at baseline was based on age at the date of study enrollment, defined as the date of the qualifying echocardiogram. For baseline functional status, the Activities of Daily Living (ADL) and Independent Activities of Daily Living (IADL) scales were used (16,17). The ADL scale measures independence in the areas of bathing, dressing, toileting, transferring, walking, and eating, with a score of 0 assigned to independence and 1 assigned to needing assistance or dependence. The sum of these is the ADL score, which ranges from 0 to 6. The IADL scale measured independence in the areas of cleaning, shopping, and doing laundry, and scores of 0 for independence and 1 for needing assistance or dependence were similarly assigned. The sum of these items is the IADL score, which ranges from 0 to 3. Higher ADL and IADL scores are associated with poorer function. In order to measure the burden of comorbid illness, the Charlson Comorbidity Index was used (18). The scale assigns a value of 1 each to the presence of myocardial infarction, CHF, peripheral vascular disease, cerebrovascular disease, dementia, chronic pulmonary disease, connective tissue disease, peptic ulcer disease, mild liver disease, and mild to moderate diabetes; a value of 2 each to diabetes with end-organ damage, hemiplegia, moderate to severe renal damage, any tumor, leukemia, and lymphoma; a value of 3 each to moderate or severe liver disease; and a value of 6 each to metastatic solid tumor and AIDS. The scale is additive, and possible scores range from 0 to 32. Higher scores are associated with a poorer prognosis. In classifying the specialties of the ordering physician for the qualifying echocardiogram, physicians were classified as cardiologists versus general internists/other specialties. Decision-making Patients were first separated into three groups: (i) those who were offered and accepted AVR surgery; (ii) those who were offered but declined surgery; and (iii) those who were never offered surgery. The group which accepted AVR was further divided into those who had AVR and those who died before surgery could be performed.

3 314 Aortic stenosis decision-making Two methods were used to identify reasons why patients declined AVR surgery, and reasons why physicians did not offer surgery to their patients. First, key variables were identified that indicated patient or physician decision rationale regarding AVR surgery. Second, the principal investigator and a research assistant reviewed full text passages both for evidence of final decision-making and for specific reasons why patients were not offered AVR and why patients declined the surgery. Follow up and outcome events All patients were monitored for survival, surgery complications, and follow up NYHA functional status data. Information about survival was obtained from medical records and office notes at the time of record review. Final survival data were obtained from an online Interactive Social Security Death Index (ssdi.genealogy.rootsweb.com) in January For patients already known to have died, the dates of death were verified for accuracy and consistency with the Death Index. As the index is current to three months, patients who had not died were considered alive as of 1st September A censoring variable was created for each patient representing the date of death or the last known date alive (1st September 2004). The primary outcome of interest was the time from the initial echocardiogram to the date of death, or the last known date alive. Information about AVR surgery complications was obtained at the time of the medical records review. Complications included major events (intraoperative death, postoperative death within 60 days after surgery, myocardial infarction, stroke, CHF, arrhythmia, renal failure, or dialysis) and minor events (pneumonia, wound infection, or postoperative delirium). Follow up functional status was analyzed in two ways. First, the number of patients alive at one year was determined. To estimate functional outcomes for those who survived at least one year, mean ADL scores were calculated from assessments documented in medical records dated six to 18 months after the initial echocardiogram. Those patients who were alive at one year, but were lacking ADL assessments within the appropriate window of time (6-18 months postechocardiography), were assigned the median value of patients with ADL available. The mean follow up ADL score was then compared to the baseline ADL score for the same group. Comparison yielded the number of patients whose functional status improved, declined, or remained the same at one year. Second, follow up functional status data were analyzed for all patients who completed a follow up telephone interview in March and April of 2001 (four to six years after study enrollment). Among those patients interviewed, the number who had surgery and the number of ADL dependencies for survivors was examined. Statistical analysis Comparisons were made between patients aged years and those aged 80 years with regard to baseline demographic and clinical characteristics such as gender, ethnicity, baseline function, comorbidity score, site of qualifying echocardiogram (inpatient versus outpatient), and specialty of the physician ordering the echocardiogram. A further comparison was then made between patients who had surgery and those who did not have surgery, with respect to age, gender, baseline function, and comorbidity score. The probability of having AVR surgery was then modeled in a logistic regression with respect to the patients age group, adjusting for potential cofounders of gender, comorbidity, and baseline function. With regard to patients decision-making, the patients were divided into three groups: (i) those who decided to have AVR surgery; (ii) those who declined AVR surgery; and (iii) those who were not offered AVR surgery. (Two patients who accepted AVR surgery died before it was performed.) Through chart review of medical records, those patients declining surgery were grouped according to the reasons for declining, and patients not offered surgery according to the reasons why they were not offered AVR surgery. Specific quotes from the medical records accompanied each group. A comparison was also made between the survival times of patients treated surgically and those treated medically. In a Cox proportional hazard model, survival probability was modeled with respect to having AVR surgery or not, adjusting for the potential cofounders of age, gender, comorbidity, and baseline functional status. The analysis was conducted using SAS statistical software version 8.1 (SAS Institute, Cary, NC, USA), with Kaplan-Meier curves estimating survival time for each group. In addition, the rates of complications after AVR were described, and in a multivariable logistic regression analysis, an assessment made as to how older age was associated with rates of surgical complications, adjusting for gender, comorbidity, and baseline functional status. All analyses were carried out using SAS software, and a p-value <0.05 was considered to be statistically significant. Results J Heart Valve Dis Baseline characteristics of study population The study included 124 patients with severe aortic stenosis. The mean (±SD) age was 81.5 ± 8.3 years (range: 60 to 97 years), with 10.5% of patients aged years, 30.7% aged years, 43.6% aged 80-89

4 J Heart Valve Dis Aortic stenosis decision-making 315 years, and 15.3% aged years. Thus, a majority of the study patients (58.9%) was aged >80 years. Most patients were female, and for those whose race was recorded in the medical records, over 99% were white. For baseline function, patients averaged 1.2 ADL dependencies and 1.9 IADL dependencies, though patients aged years had significantly fewer ADL and IADL dependencies on average than did patients aged 80 years, on average (p <0.0001). Study patients had a mean of 1.8 for the Charlson Comorbidity Index score, with 60- to 79-year-old patients having a significantly lower mean score than those aged >80 years. Most study patients had a normal ejection fraction ( 50%); a decreased ejection fraction was not associated significantly with any one age group (Table I). Characteristics of surgically and medically treated patients In total, 49 (39.5%) patients underwent AVR surgery to treat their aortic stenosis. When compared with nonsurgery patients, AVR patients tended to be younger and healthier, and had significantly less baseline functional impairment (Table II). The mean age of patients treated surgically was 76.6 years, while the mean age of those treated medically was 84.7 years (p = 0.01). Moreover, a majority (76.9%) of patients aged years had AVR surgery, and 60.5% of those aged years had surgery compared to 21.9% of those aged >80 years (p <0.001). The mean comorbidity score was 1.3 for surgery patients, and 2.2 for non-surgery patients. For patients with comorbidity scores 2, almost half (46.7%) had surgery, while a large majority (79.4%) with comorbidity scores 3 did not. With respect to baseline functional status, as the number of dependencies increased, the percentage of patients receiving AVR decreased in both ADL and IADL scales. In particular, 52.2% of patients with zero ADL dependencies had surgery, compared to only 9.4% with one or more ADL dependencies. Similarly, 56.3% of patients with zero IADL dependencies had surgery, while 27.0% of patients with one to three IADL dependencies had AVR. Though male patients tended to have surgery more frequently than female patients (48.8% and 34.6%, respectively), the effect of gender was not statistically significant (p = 0.129). Patients whose echocardiogram was ordered by a cardiologist were significantly more Table I: Demographic and clinical characteristics of patients aged 60 years with severe aortic stenosis. Parameter All patients Age Age 80+ p-value (n = 124) (n = 51) (n = 73) Gender (n) Female 81 (65.3) 23 (45.1) 58 (79.5) < Male 43 (34.7) 28 (54.9) 15 (20.5) Race/ethnicity (n) White 102 (82.3) 39 (76.5) 63 (86.3) 0.21 Black 1 (0.8) 1 (2.0) 0 (0) Not stated 21 (16.9) 11 (21.5) 10 (13.7) Baseline functional scores * ADL dependencies (1-6) (0,2) 0.36 (0,0) 1.8 (0,3) < IADL dependencies (1-3) 1.9 (0,3) 0.97 (0,3) 2.3 (1.5,3) < Comorbidity score * 1.8 (0,3) 1.7 (0,3) 1.9 (1,3) 0.05 Ejection fraction (n) 50% 86 (69.9) 40 (78.4) 46 (63.9) % 20 (16.3) 5 (9.8) 15 (20.8) <30% 17 (13.8) 6 (11.8) 11 (15.3) Location of qualifying echocardiogram (n) Inpatient 80 (64.5) 26 (51.0) 54 (74.0) Outpatient 44 (35.5) 25 (49.0) 19 (26.0) Specialty of physician ordering echocardiography (n) Cardiologist 58 (46.8) 26 (51.0) 32 (43.8) 0.56 General internist/other 61 (49.2) 22 (43.1) 39 (53.5) Not stated 5 (4.0) 3 (5.9) 2 (2.7) * Values are mean (CI 25%,75%). Values in parentheses are percentages missing ADL scores at baseline. 29 missing IADL scores at baseline. One missing baseline ejection fraction; echo reports describing ejection fraction as normal were classified as >50%.

5 316 Aortic stenosis decision-making likely (p <0.0001) to have AVR surgery than those whose echocardiograms were ordered by general internists or other physicians. In particular, 56.9% of the former had surgery as opposed to 24.6% of the latter. In a logistic regression analysis adjusting for gender, comorbidity and baseline functional status, patients aged 80 years were less likely to have AVR surgery than those aged years (adjusted odds ratio 0.19 [95% CI ]). Decision-making Final decisions for all study patients fell into three categories: 51 patients (41.1%) decided to have AVR surgery; 30 patients (24.2%) were offered but declined surgery; and the remaining 43 patients (34.7%) were never offered AVR (Tables III and IV). (Among the 51 patients with chart records indicating they were offered and decided to have surgery, two died before the procedure could be performed.) Among patients aged years, according to the medical record review, 64.7% discussed AVR with their provider and agreed to have surgery, while only 24.7% of patients aged >80 years agreed. The medical records were J Heart Valve Dis examined to identify the reasons why patients declined AVR surgery, as well as reasons why they were not offered AVR surgery. Among the subgroup of patients who declined AVR surgery, 30.0% preferred medical rather than surgical management, 30.0% had no explicit reason recorded, 23.2% chose valvuloplasty, and the remaining 16.8% preferred comfort care only. Among the subgroup of patients never offered AVR, the most frequent reason given by physicians for not offering surgery was the presence of comorbid illnesses (32.6%) (Table IV). No discussion of AVR surgery occurred in 32.6% of cases and, although AVR was discussed in an additional 14.0% of study patients, no rationale was provided in the medical records. In 9.3% of the population, age was cited as the reason for not offering AVR surgery, and in another 7.0% the combination of comorbidity and age was cited. In the remaining 4.6%, the patient s aortic stenosis was considered clinically stable and not warranting surgical intervention at the time. Table II: Comparison of patients aged 60 years with severe aortic stenosis, with and without AVR surgery. Parameter AVR No AVR p-value surgery surgery No. of patients 49 (39.5) 75 (60.5) - Age (years) (76.9) 3 (23.1) (60.5) 15 (39.5) < (21.9) 57 (78.1) Gender Female 28 (34.6) 53 (65.4) Male 21 (48.8) 22 (51.2) Baseline ADL functional impairment No dependencies 36 (52.2) 33 (47.8) 1 dependency 3 (9.4) 29 (90.6) < Not stated 10 (43.5) 13 (56.5) Baseline IADL functional impairment No dependencies 18 (56.3) 14 (43.8) 1-3 dependencies 17 (27.0) 46 (73.0) Not stated 14 (48.3) 15 (51.7) Comorbidity score (46.7) 48 (53.3) (20.6) 27 (79.4) Specialty of physician ordering echocardiography Cardiologist 33 (56.9) 25 (43.1) General internist/other 15 (24.6) 46 (75.4) Not stated 1 (20.0) 4 (80.0) Values in parentheses are percentages.

6 J Heart Valve Dis Clinical outcomes Survival At the time of the final follow up (January 2005), 25.8% of the study patients were still alive. Survival was significantly better for patients who had AVR surgery (Fig. 1). At the one-year follow up, 87.8% of patients who had AVR were alive compared to 54.7% of those who did not have AVR (p <0.001). In the subgroup of patients aged 80 years, 87.5% of those who had AVR were alive compared with 49.1% of those who did not have AVR (p = 0.006). In a Cox proportional hazard model adjusting for age, baseline functional status, and comorbid illness, AVR surgery was associated with a large mortality reduction (adjusted odds ratio 0.39 [95% CI ]). Aortic stenosis decision-making 317 Complications For those patients who underwent AVR surgery (n = 49), a majority (32; 65.3%) experienced one or more complication (Table V). Although there were no intraoperative deaths or myocardial infarctions, four patients (8.2%) died within 60 days (two of the four died within 30 days). Postoperative complications were not uncommon: one patient (2.0%) had a stroke, six patients (12.2%) experienced CHF, 24 (49.0%) had an arrhythmia, five (10.2%) had renal failure, two (4.1%) were placed on dialysis, four (8.2%) had pneumonia, eight (16.3%) had a wound infection, and eight (16.3%) experienced postoperative delirium. Among the four postoperative deaths, one death occurred on the day of surgery, while the other deaths occurred at five, 45 and 48 days after surgery, respectively. Although no statistically significant difference in complication rates for younger and older patients was found, there was a trend toward a significant difference for one of the 12 complications studied (congestive heart failure, with 25% of patients aged 80 years experiences this complication and only 6% of patients age years; p = 0.08). When the probability of having any type of complication with AVR surgery was modeled with respect to the patients age group in an unadjusted logistic regression analysis, complications were not more common among older patients (comparing aged 80 years to years; odds ratio 0.83 [95% CI ]). The results Figure 1: Survival curves for patients with (dashed line), and without (solid line), aortic valve replacement surgery. were similar after adjustment for gender, comorbidity, and baseline functional status (adjusted odds ratio 0.65 [95% CI ]). When age was measured in terms of years instead of age groups, again there was no statistically significant relationship between patient age and the rate of complications after AVR. Functional outcomes At one year after enrollment, 84 patients (67.7%) were still alive. Follow up functional status data from six to 18 months after enrollment were available in the medical records for 47 (60.0%) patients who survived one year. (Of the 37 survivors without follow up function data at one year, no functional status data were available for six of these patients, recorded earlier than six months for 18, and recorded after 18 months for the remaining 13 patients.) The mean ADL score for survivors at one-year follow up was 0.67 ± 1.87 dependencies, compared to 0.64 ± 1.59 dependencies at baseline. Among all surviving patients, there were 36 survivors (42.9%) who had both follow up data at one year and baseline data, and the vast majority (80.6%) maintained or improved their baseline ADL score, while 19.4% declined in ADL independence. Changes Table III: Reasons for absence of AVR surgery in patients aged 60 years with severe aortic stenosis (n = 75). Decision All patients Age Age 80+ p-value category (n = 124) (n = 51) (n = 73) Accepted AVR surgery * 51 (41.1) 33 (64.7) 18 (24.7) < Declined AVR surgery 30 (24.2) 4 (7.8) 26 (35.6) Not offered AVR surgery 43 (34.7) 14 (27.5) 29 (39.7) * Two patients decided to have AVR but died before the procedure.

7 318 Aortic stenosis decision-making J Heart Valve Dis Table IV: Reasons (taken from medical records) why patients declined AVR, or why surgery was not offered. Reason No. of patients (%) Example quotes from medical records Reasons for declining AVR (n = 30) Patient preferred medical 9 (30.0) As in the past, the patient adamantly refuses to consider valve management replacement and prefers to follow a conservative approach with medical therapy. (Age 84 years) Patient does not want surgical intervention. On 8/14, she was concerned that she should have aortic valve surgery... Risks were explained to patient, and patient is reassured with the status quo. (Age 87 years) No explicit reason listed 9 (30.0) Patient is totally uninterested in pursuing an AVR and we agree that this is prudent. We will see him back in three months. (Age 79 years) The patient has severe AS with mild to moderate AR. The patient is not interested in AVR at this time. (Age 92 years) Patient chose valvuloplasty 7 (23.2) Patient initially presented with critical AS. At that time, she refused instead of AVR AVR and underwent valvuloplasty which increased valve area. (Age 88 years) We discussed at great length again balloon valvuloplasty, surgery or medical therapy It would be appropriate to proceed with balloon valvuloplasty with hopes of symptomatic improvement. The patient and daughter are going to think about it. (Age 87 years) Patient or family asked 5 (16.8) Discussed with daughter the issue of DNR/DNI [Do-not-resuscitate/do for comfort care only not hospitalize]. She does not want extraordinary things done. Wants comfort measures. This is in accord with son s wishes. (Age 92 years) Patient stated clearly that she did not wish to be hospitalized should she become ill. Does not wish to be transferred for any medical care. (Age 93 years) Reasons why physicians did not offer AVR (n = 43) Comorbidity cited 14 (32.6) Cardiac echo yesterday revealed severe AS. I m in favor of a conservative approach in this elderly diabetic with CAD and severe AS. (Age 84 years) Given high comorbidity, patient is not a candidate for valve replacement. Need to try to maximize her medical regime. (Age 66 years) No discussion of AVR 14 (32.6) N/A in medical records AVR discussed, but no 6 (14.0) AS seems to be the reason for congestive heart failure, and a noninterverationale for not offering tional medical approach was determined to be the most appropriate. (Age 91 years) The patient was felt not to be a candidate for valvular replacement, so no further workup was pursued. (Age 96 years) Age of the patient cited 4 (9.3) As you know, [the patient] has always refused AVR, and I don t think she is any longer a candidate, even if she would agree. (Age 89 years) The echo shows significant AS and MR. In view of his age and the fact that he will need a two valve replacements with possible CABG, I think he is best managed medically. (Age 70 years) Both age and 3 (7.0) The patient is not an ideal for cardiac catheterization, AVR, or CABG, comorbidity cited secondary to her age, increased creatinine, and decreased hematocrit. (Age 89 years) We cannot render a definitive opinion in the absence of her old note and without a current cath. Nevertheless, the simple answer is that this unfortunate woman poses far too great a risk for AVR given her condition, age, and status. (Age 85 years) AS considered clinically 2 (4.6) AS not severe enough to warrant intervention at this time. It is unlikely stable N/A: Not applicable. that the patient will need surgery. (Age 78 years) The patient appears clinically stable and I would see no reason to alter her medical therapy at this time, especially in the absence of any clinical symptomatology. (Age 74 years)

8 J Heart Valve Dis Aortic stenosis decision-making 319 in ADL dependencies for the 36 patients were as follows: -2 (improved by gaining independence in two ADLs) for two patients; -1 for one patient, 0 (no change in ADL score) for 26 patients, +3 (worsened by losing independence in three ADLs) for two patients, and +6 for five patients. Among patients having AVR surgery, 43 were alive at one year, and follow up functional status data during the six- to 18-month period following surgery were available for 21 patients (48.8%). (Of the remaining 22 patients (51.2%), follow up data were either recorded outside the six- to 18-month window, or were missing.) Within this group, ADL scores improved in all but one patient, and were as follows: -2 (improved by gaining independence in two ADLs) for one patient, 0 (no change in ADL score) for 19 patients, and +6 (worsened by losing independence in six ADLs) for one patient. When long-term follow up data were collected in the spring of 2001 (four to six years after study enrollment), only 43 patients (34.7%) were alive to be contacted for a telephone interview. Among these patients, follow up data were obtained from 37 (86.0%). Of all patients still alive in 2001, most reported good functional status: 29 (67.4%) had zero ADL dependencies, two (4.7%) had one dependency, one (2.3%) had three dependencies, one (2.3%) had four dependencies, four (9.3%) had six dependencies, and six (14.0%) had unknown functional status. At long-term follow up, 33 of the 43 surviving patients had received AVR surgery. Long-term functional status follow up data for this subgroup of surviving surgery patients were as follows: 24 (72.7%) had zero ADL dependencies, one (3.0%) had one ADL dependency, one (3.0%) had three dependencies, one (3.0%) had four dependencies, three (9.1%) had six dependencies, and three (9.1%) had unknown functional status. Among the surgery patients who had survived to long-term follow up, nine were aged 80 years, and their functional outcomes were similar: five (55.5%) had no dependencies, one (11.1%) had three dependencies, and three (33.3%) had six dependencies. Discussion The study findings confirm that AVR surgery improves survival for elderly patients with severe aortic stenosis. Survival and functional outcomes were good for patients aged 80 years who had AVR surgery, and older patients in the present cohort did not have higher postoperative complication rates than younger patients. These findings suggest, however, that older age is at times associated with a reluctance on the part of patients to agree to, and physicians to recommend, AVR surgery. Older patients were more likely to decline surgery than younger patients (35.6% of those aged >80 years declined AVR surgery compared to 7.8% of younger patients), and among the 30 patients who declined surgery, eight were aged 90 years. As patients reached very old age, some may Table V: Reported complications from AVR surgery (n = 49). Complication All patients Age Age 80+ p-value (n = 49) (n = 33) (n = 16) Mortality Intraoperative death (n) 0 (0.0) 0 (0.0) 0 (0.0) - Postoperative death (n) * 4 (8.2) 3 (9.1) 1 (6.3) Postoperative morbidity Myocardial infarction (n) 0 (0.0) 0 (0.0) 0 (0.0) - Stroke (n) 1 (2.0) 1 (3.0) 0 (0.0) Congestive heart failure (n) 6 (12.2) 2 (6.1) 4 (25.0) Arrhythmia (n) 24 (49.0) 18 (54.6) 6 (37.5) Renal failure (n) 5 (10.2) 4 (12.1) 1 (6.3) Dialysis (n) 2 (4.1) 1 (3.0) 1 (6.3) Pneumonia (n) 4 (8.2) 3 (9.1) 1 (6.3) Wound infection (n) 8 (16.3) 5 (15.2) 3 (18.8) Postoperative delirium (n) 8 (16.3) 5 (15.2) 3 (18.8) Any of the above complications One or more complication 32 (65.3) 22 (66.7) 10 (62.5) from AVR (n) Values in parentheses are percentages. * Death occurring within 60 days of surgery.

9 320 Aortic stenosis decision-making have chosen to focus their care on comfort instead of life extension. The results of the present study also found older age to be associated in some cases with a reluctance on the part of physicians to offer AVR surgery. While the presence of significant comorbidity was the most common reason explicitly offered by physicians in this chart-analysis review, patient age proved to be the second most common reason. Interestingly, not all physicians were equally likely to recommend AVR surgery to older patients. The study findings suggested that surgery was more likely when cardiologists were involved in the care of patients with aortic stenosis, which may reflect cardiologists greater appreciation of the benefits of AVR surgery in elderly patients with severe symptomatic AS. These results agree with a growing body of research into the safety and efficacy of surgery in managing older patients with aortic stenosis. While the benefit of valve replacement for symptomatic aortic stenosis in the general population has been well documented, numerous studies during the past decade have confirmed that older patients benefit from this surgery with an acceptable level of risk. Indeed, one recent study reported perioperative mortality rates of 7% for patients aged years, with no significant difference in mortality between patients in their eighties compared to those in their seventies (11). Many other studies have also shown low mortality and morbidity rates (9,19-21) and, importantly, excellent functional outcomes following AVR, even in patients aged >80 years (22,23). In fact, studies have shown that older patients demonstrate an improvement in symptoms, physical abilities, and general well-being equivalent to that of younger patients. The results of the present study also showed age to be a significant predictor of whether patients had valve replacement surgery. The idea that older patients are being denied surgery on the basis of age has been acknowledged in the literature on valve replacement surgery (14,22), though few studies have actually attempted to analyze the specific reasons why surgery has been denied in this patient population. Those that have done so agree that age has functioned as an important factor in the decision-making process. In one such study of Dutch cardiologists (15), the authors reported that 41% of those surveyed by mail appeared to base their decisions regarding AVR surgery in older adults on age alone. Another study employing chart review at a British hospital, while not addressing age specifically, found little discussion of valve replacement surgery in patients aged over 75 years who did not undergo surgery, presumably because the option of surgery was never discussed among the oldest aortic stenosis patients (24). J Heart Valve Dis Study limitations One major limitation of the present study was the use of chart-review data for determining decisionmaking. While information was reviewed from both inpatient charts and outpatient office visits, including general internists, cardiologists, and cardiothoracic surgeons, it is likely that some amount of decisionmaking surrounding valve replacement surgery was not captured by the methods employed. Whenever feasible, however, patient records from other institutions were sought and included in the data analysis. For some patients, no rationale was evident in the medical records for not offering AVR surgery. It is possible that some discussions were not documented, and that some of the patients declined surgery. In addition, information about functional status was not available for all patients. In conclusion, as the elderly population of the USA continues to grow, an increasing number of older adults will be faced with decisions regarding severe aortic stenosis. The results of the present study provide additional evidence that older adults benefit from valve replacement surgery. The study findings corroborate and extend prior research in this area by attempting to categorize precisely how patients and physicians make decisions about AVR surgery. In this cohort of elderly patients with symptomatic, severe aortic stenosis, elderly patients (even those aged 80 years) were seen to experience excellent outcomes after AVR surgery. It is important to note that patients in the present cohort who had surgery were selected for such treatment; thus, it can be concluded that appropriately selected elderly patients, including those aged >80 years, can achieve good outcomes. References 1. Freeman RV, Crittenden G, Otto C. Acquired aortic stenosis. Exp Rev Cardiovasc Ther 2004;2: Sawhney N, Hassankhani A, Greenberg BH. Calcific aortic stenosis in the elderly: A brief overview. Am J Geriatr Cardiol 2003;12: Ross J, Braunwald E. Aortic stenosis. Circulation 1968;38(Suppl.V):V61-V67 4. Lester SJ, Heilbron B, Gin K, Dodek A, Jue J. The natural history and rate of progression of aortic stenosis. Chest 1998;113: Takeda J, Warren R, Holzman D. Prognosis of aortic stenosis. Arch Surg 1963;87: Bergon J, Abelmann WH, Vazquez-Milan H, Ellis LB. Aortic stenosis: Clinical manifestations and course of the disease. Arch Intern Med 1954;94: Turina J, Hess O, Sepulcri F, Krayenbuehl HP. Spontaneous course of aortic valve disease. Eur

10 J Heart Valve Dis Aortic stenosis decision-making 321 Heart J 1987;8: Horstkotte D, Loogen F. The natural history of aortic valve stenosis. Eur Heart J 1987;9(Suppl.): Culliford AT, Galloway AC, Colvin SB, et al. Aortic valve replacement for aortic stenosis in persons aged 80 years and over. Am J Cardiol 1991;67: Wong JB, Salem DN, Pauker SG. You re never too old. N Engl J Med 1993;328: Chiappini B, Bergonzini M, Gallieri S, et al. Clinical outcome of aortic valve replacement in the elderly. Cardiovasc Surg 2003;11: Bouma BJ, van den Brink RB, van der Meulen JH, et al. To operate or not on elderly patients with aortic stenosis: The decision and its consequences. Heart 1999;82: Tseng EE, Lee CA, Cameron DE, et al. Aortic valve replacement in the elderly. Risk factors and longterm results. Ann Surg 1997;225: ;discussion Bramstedt KA. Aortic valve replacement in the elderly: Frequently indicated yet frequently denied. Gerontology 2003;49: Bouma BJ, van der Meulen JH, van den Brink, et al. Variability in treatment advice for elderly patients with aortic stenosis: A nationwide survey in The Netherlands. Heart 2001;85: Katz S, Ford AB, Moskowitz RW, Jackson BA, Jaffe MW. Studies of illness in the aged. The index of ADL: A standardized measure of biological and psychosocial function. JAMA 1963;185: Lawton MP, Brody EM. Assessment of older people: Self-maintaining and instrumental activities of daily living. Gerontologist 1969;9: Charlson ME, Pompei P, Ales KL, McKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 1987;40: Bergus BO, Feng WC, Bert AA, Singh AK. Aortic valve replacement (AVR): Influence of age on operative morbidity and mortality. Eur J Cardiothorac Surg 1992;6: Chiappini B, Camurri N, Loforte A, Di Marco L, Di Bartolomeo R, Marinelli G. Outcome after aortic valve replacement in octogenarians. Ann Thorac Surg 2004;78: Craver JM, Puskas JD, Weintraub WW, Shen Y, Guyton RA, Gott JP, Jones EL. 601 octogenarians undergoing cardiac surgery: Outcome and comparison with younger age groups. Ann Thorac Surg 1999;67: Sundt TM, Bailey MS, Moon MR, et al. Quality of life after aortic valve replacement at the age of >80 years. Circulation 2000;102(19 Suppl.3):III70-III Olsson M, Janfjall H, Orth-Gomer K, Unden A, Rosenqvist M. Quality of life in octogenarians after valve replacement due to aortic stenosis. A prospective comparison with younger patients. Eur Heart J 1996;17: Abdul-Hamid AR, Mulley GP. Why do so few older people with aortic stenosis have valve replacement surgery? Age Ageing 1999;28: