|
|
| |||||||||||||||||||||||||||||||||||||||
Background Reports of an increase in asthma-related mortality have been based on studies of death certificates from the general U.S. population on which asthma was listed as an underlying cause of death. We addressed the issue in a different way by analyzing long-term survival in a defined, population-based cohort of patients with asthma.
Methods We identified all residents of Rochester, Minnesota, in whom asthma was diagnosed from January 1, 1964, through December 31, 1983, by reviewing the medical records of all patients with asthma and associated diagnoses, using explicit predefined criteria. The patients' vital status at last follow-up was ascertained. Medical records, death certificates, and autopsy reports were reviewed to classify deaths as due either to asthma or to other conditions.
Results We identified 2499 patients with definite or probable asthma. The mean duration of follow-up was 14 years (range, 0 to 29). There were 140 deaths during 32,605 person-years of follow-up. Overall survival was not significantly different from the survival that was expected for residents of Rochester. Survival was less than expected in patients who were 35 years of age or older when their asthma was diagnosed and who also had another lung disease (predominantly chronic obstructive pulmonary disease). Four percent of all deaths in the study cohort were due to asthma, and all were among adults. Survival was not related to the year of onset of asthma.
Conclusions Survival among patients with asthma but no other lung disease was not significantly different from expected survival. However, patients 35 or older who had asthma associated with chronic obstructive pulmonary disease did have worse than expected survival. Asthma was classified as the cause of death in only 4 percent of the patients, and there was no evidence of an increased risk of death among patients with a more recent diagnosis of asthma. These results provide assurance that community-based patients with asthma usually have a good prognosis.
Methods
Study Cohort
Population-based epidemiologic research in Rochester, Olmsted County, Minnesota, is possible because the complete inpatient and outpatient medical records of all residents are retrievable for review and all medical diagnoses and surgical procedures have been indexed in an automated form since 1935. The methods of identifying the incidence cohort have been described previously12. Briefly, the medical diagnostic index was used to retrieve the medical records of all residents with asthma or bronchial asthma; asthmatic bronchitis; allergic bronchitis; allergic, exogenous, or extrinsic asthma; intrinsic, endogenous, or infectious asthma; chronic obstructive pulmonary disease with asthma or bronchospasm; chronic obstructive pulmonary disease in persons under 30 years of age; bronchiolitis or recurrent bronchiolitis; chronic bronchitis; bronchospasm; or wheezing. The medical records of a random sample of patients with chronic obstructive pulmonary disease who were over 30 years old and a random sample of patients with a diagnosis of cough were also reviewed, but the yield from these lists was less than 3 percent and the records were not reviewed in detail. Overall, 18,000 medical records were reviewed by trained nurse abstractors to determine whether patients' illnesses met our predetermined criteria for asthma.
The date of onset of asthma (incidence date) was defined as the date of the earliest diagnosis of asthma by a physician as recorded in the medical record or the earliest description of the constellation of symptoms that met our predetermined criteria for asthma. Residency in Rochester on the date of onset of asthma was verified through birth records, city and county directories, or earlier medical records. A case was included only if the patient (unless an infant) had been a resident of Rochester for at least one year before the incidence date. All residents whose dates of the onset of asthma were from January 1, 1964, through December 31, 1983, were included in the incidence cohort. Questions about eligibility and date of onset were resolved through a review of the medical record with one of the physician investigators.
Diagnosis of Asthma
Patients were considered to have definite asthma if each of three conditions was present on or near the incidence date -- a history of cough, dyspnea, or wheezing or a history of cough or dyspnea with wheezing on examination; substantial variability in symptoms over time or no symptoms for weeks or longer; and two or more of the following: sleep disturbed by nocturnal cough or wheezing; no current smoking (for patients of 14 years or older); nasal polyps; a blood eosinophil index (the leukocyte count per cubic millimeter times the percentage of eosinophils) of more than 300; positive results on allergy skin tests or an elevated serum IgE concentration; a history of hay fever or atopic dermatitis or of cough, dyspnea, and wheezing regularly on exposure to an antigen; and pulmonary-function tests showing at least one value for forced expiratory volume in one second or forced vital capacity that was less than 70 percent of the predicted value and another with at least 20 percent improvement after bronchodilator therapy to more than 70 percent of the predicted value. If a physician had diagnosed asthma but these conditions were not met, the patient was considered to have probable asthma. Single episodes of wheezing were tabulated separately.
Patients were considered not to have asthma and were excluded from the study if any of the following were present on or near the incidence date: dyspnea during exercise (only for patients older than 50); pulmonary-function tests that showed forced expiratory volume in one second to be consistently below 50 percent of the predicted value or diminished diffusing capacity of the lung for carbon monoxide; a tracheal or bronchial foreign body; and hypogammaglobulinemia (serum IgG concentration less than 2.0 mg per milliliter) or another immunodeficiency disorder. Patients were also excluded from the study if they had one of the following diseases before the incidence date: bullous emphysema or pulmonary fibrosis; homozygous 
1-antitrypsin deficiency; cystic fibrosis; or another major chest disease, such as bronchogenic carcinoma, bronchiectasis, or juvenile kyphoscoliosis.
Data Collection
Data were collected from the medical records of all providers of care for each patient, entered into a computer data base, and checked for consistency and accuracy. On four occasions during a six-year period, random samples of records were reviewed by a different nurse abstractor and analyzed for agreement between observers. On each occasion, a high degree of concordance was found13.
The vital status of all the patients was followed through the medical records. Information about vital status was obtained as of the date of the last follow-up visit or the last communication by telephone or mail with a provider of health care to the residents. In addition, vital status was ascertained by searching data bases of ambulatory care or hospitalization for county residents and by searching tapes provided by the state of Minnesota that contained death-certificate data for all deaths in the state through December 31, 1990. A follow-up questionnaire was mailed to a subgroup of patients in whom definite asthma began during childhood (approximately 800 subjects).
The Rochester Epidemiology Project14,15 routinely obtains the death certificates of residents of Rochester and Olmsted County. Information about the underlying cause of death from part I of the death certificate, as coded by state nosologists, was obtained for deaths that occurred in Minnesota.
Statistical Analysis
Survival after the onset of asthma was assessed with Kaplan-Meier survival curves and compared with the expected survival of Olmsted County residents of like age and sex. A Cox proportional-hazards model was used to test whether survival had changed during the 20-year study period after control for age at the onset of asthma and for sex. Whether survival had changed was analyzed by testing whether the year of diagnosis and the interaction between the age at diagnosis and the year of diagnosis were significantly associated with hazard rates. All P values are two-tailed.
Results
We identified 2499 residents of Rochester, Minnesota (1410 male and 1089 female residents), with a new diagnosis of asthma from 1964 through 1983. Of those, 1547 met our criteria for definite asthma and 952 met the criteria for probable asthma (Table 1). The median age at onset was 4 years (3 for male and 8 for female patients), and the mean age was 15 years. The mean duration of follow-up was 14 years (range, 0 to 29), and there were 140 deaths during 32,605 person-years of follow-up. The mean age at death was 71 years (range, 14 to 96). Figure 1 shows the distribution of ages at the onset of asthma in the study cohort, and Figure 2 shows the ages at death for the 140 patients who died.
|
|
|
|
|
|
|
The rate of survival in this large, population-based cohort of residents of Rochester, Minnesota, with onset of asthma from 1964 through 1983 was high and not significantly different from the expected rate of survival calculated on the basis of vital statistics for Olmsted County, of which Rochester is a part. There was no significant reduction in survival in the patients who had asthma alone. The subgroup of patients with asthma and another lung disease (who were older than those with asthma alone) had a significantly lower rate of survival than expected. The associated lung disease was most often chronic obstructive pulmonary disease. Survival was not significantly related to the calendar year of the onset of asthma, after adjustment for age and sex, or to whether the diagnosis of asthma was definite or probable. Survival was somewhat better than expected in persons who had asthma before the age of five, but the difference was small. The trend toward decreased survival in the adults with asthma diagnosed after the age of 35 was explained entirely by survival in the small number of patients with asthma and another lung disease.
Epidemiologic studies of asthma-associated mortality have focused on young patients, because death in childhood is unusual, deaths due to asthma are unexpected, and the identification of asthma as an underlying cause of death on death certificates is believed to be accurate in the group from 5 to 34 years old16,17. However, as our study demonstrates, most deaths among patients with asthma occur at older ages, as do most deaths due to asthma. Indeed, only six of the deaths in our population-based cohort were classified as due to asthma, and all six occurred after the age of 35.
The strengths of our study are that it was population-based, it used predefined epidemiologic criteria to identify patients with asthma, the follow-up period was long, and observed mortality was compared with expected mortality for residents of Olmsted County. Although the mean duration of follow-up was 14 years, the small number of deaths due to asthma limited our ability to test for trends in mortality according to year of incidence. We used the state nosologists' classification of underlying causes of death as stated on the death certificates but supplemented this by a detailed review of the medical records of patients who died in Olmsted County. We recently reported that the classification of underlying causes of death on death certificates underestimates asthma-associated mortality18. We therefore reviewed all medical records and identified two deaths that in our judgment were due to asthma but for which asthma was not listed as the underlying cause. Even with this additional level of ascertainment, asthma was found to be the cause of death in only 4 percent of the cohort.
Extensive efforts were made to determine vital status through a review of medical records, local hospital and outpatient-clinic data bases, and Minnesota death certificates. Death certificates were requested for all deaths, including those outside Minnesota. Patients who had moved to another state were included in the analysis until their last known visit to a health care provider in Olmsted County or their last communication with such a provider by telephone or mail. Of the 140 cohort members who died, 13 were not Minnesota residents at the time of death, and these 13 were included in the survival analysis. The observed survival of cohort members did not differ from the survival expected for residents of Olmsted County. It is unlikely that this finding resulted from underascertainment of deaths among a cohort of patients with asthma, who actually have an increased risk of death.
Previous reports of long-term follow-up of patients with asthma either have not reported the outcome of death or have reported mortality rates and standardized mortality rates rather than survival curves9,10,19,20,21,22,23,24. For 27 patients with asthma (30 percent of them men) whose mean age was 62 years, Burrows et al.25 reported a 10-year survival rate of approximately 85 percent, with significantly fewer deaths than expected. Among 343 18-to-80-year-old patients with asthma who were referred to a chest clinic in Copenhagen, Denmark, from 1976 to 1981, Almind et al.26 found approximately 78 percent survival in men and 88 percent survival in women over an eight-year period; observed survival was less than expected survival only in the men. The mean age at base line of the 69 patients with allergy who were followed up in 1988 was 36 years, and that of the 144 patients without allergy was 52 years26. This cohort was therefore considerably older than ours, and their expected survival accordingly lower.
Four percent of all deaths in our cohort were due to asthma, and none of them were in children. We did not find any evidence to suggest that asthma-associated mortality among the incident cases increased from 1964 through 1983. This observation is especially noteworthy because we previously reported an increase in the incidence of asthma among children in Rochester during this period12. Our study suggests that there is geographic heterogeneity in recent reports of an increase in asthma-associated deaths. Further follow-up studies of our 1964-1983 cohort, as well as studies of the 1984-1993 cohort, will be important in determining whether trends reported elsewhere are occurring later in Rochester. Rochester is a medium-sized midwestern community 96 percent of whose residents are white. There is access to high-quality health care, and there is minimal outdoor air pollution. Continued surveillance of asthma-associated mortality in Rochester may also provide valuable contrasting data that may explain why mortality from asthma has increased in some communities but not in others27.
Supported in part by grants (AI 25187 and AR 30582) from the National Institutes of Health.
We are indebted to Ms. Hazel M. Anderson, Ms. Joan M. LaPlante, Ms. Joanne Mair, and Ms. Marie L. Notermann, nurse abstractors, through whose untiring efforts these data were obtained; to Ms. Mary M. Ho and Mr. Randall E. Stick for assistance with data analysis; and to Ms. Darcy Jacobson for assistance in the preparation of the manuscript.
Source Information
From the Divisions of Area General Internal Medicine (M.D.S.) and Allergic Diseases and Internal Medicine (C.E.R.) and the Departments of Pediatric and Adolescent Medicine (E.J.O., J.W.Y.) and Health Sciences Research (L.J.M., W.M.O.), Mayo Clinic and Mayo Foundation, Rochester, Minn.
Address reprint requests to Dr. Silverstein at the Mayo Clinic, 200 First St. S.W., Rochester, MN 55905.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
Patterns of Mortality from Asthma
Jay S. J., Kane G. C., Petsonk E. L., Castellan R. M., Wagner G. R., Silverstein M. D., Yunginger J. W., Lang D. M., Polansky M., Buist A. S., Vollmer W. M.
Extract |
Full Text
N Engl J Med 1995;
332:1379-1381, May 18, 1995.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2010 Massachusetts Medical Society. All rights reserved. |
