High Stroke Incidence in the Prospective Community-Based L’Aquila Registry (1994–1998)
First Year’s Results
Background and Purpose Changes in stroke incidence are likely to occur as a consequence of aging of the population, but evidence for this hypothesis is lacking.
Methods A prospective community-based registry of first-ever strokes (1994 to 1998) classified according to the International Classification of Diseases, 9th Revision (ICD-9) was established in the L’Aquila district, central Italy, with a total population of 297 838 (1991 census). Patients were identified by active monitoring of multiple sources, including general practitioners.
Results In 1994, 819 patients (398 men and 421 women; mean±SD age, 74.8±11.3 years) suffered from a first-ever stroke. Eighty-nine percent of the patients had neuroimaging studies of the brain and were reclassified with the recent Application of the International Classification of Diseases to Neurology (ICD-10 NA). The occurrence of subarachnoid hemorrhage, intracerebral hemorrhage, cerebral infarction, and ill-defined events was 2.9%, 14.9%, 80.2%, and 2.0%, respectively. Crude annual incidence of first-ever stroke was 2.75/1000 (95% confidence interval [CI], 2.57 to 2.94) and 24.23/1000 (95% CI, 21.65 to 27.10) in patients older than 80 years. Incidence rates were higher in men and steeply increased with age. The standardized rate was 2.37/1000 for the Italian and 2.28/1000 for the European population. The 30-day case-fatality rate was 25.6% (95% CI, 22.8% to 28.7%). The occurrence of death, disability, and full recovery at 1 year was 36.9%, 38.9%, and 24.2%, respectively. No differences were found in stroke incidence and case-fatality according to income and urban or rural residences.
Conclusions In our population-based study, we found a high stroke incidence notably in the older age subgroups, suggesting that rather than declining, stroke is only being postponed until later in life.
Stroke registries are useful to determine the incidence and case-fatality rate of stroke in defined populations.1 2 Although incidence and death rates from stroke have been declining in Western countries for the past decades,3 4 the number of elderly people at higher risk for stroke is progressively increasing. Therefore, further changes in temporal trends of stroke incidence and mortality are likely to occur.5
Accurate data are needed to organize stroke units and services6 7 and to plan clinical trials on acute stroke.8 Because of the growth in the number of people older than 80 years, the need of long-term care services will increase.9 Differences in life expectancy at the age of 80 years have been reported among various countries, possibly because of income inequalities.10 11
In Italy, the L’Aquila district is an ideal setting to determine stroke incidence and case-fatality rates and their trends, since it represents a well-defined geographic area with a stable population, wide availability of health services, and easy access to hospitals. This report describes data collected during the first year of activity of the L’Aquila Stroke Registry from January 1 to December 31, 1994.
Subjects and Methods
The prospective community-based registry of stroke patients residing in the L’Aquila district started on January 1, 1994, to continue for at least 5 years until December 31, 1998. The Neurology Department of the University of L’Aquila is the coordinating center for the study.
Stroke was defined as rapidly developing signs of focal or global disturbance of cerebral function, lasting longer than 24 hours or leading to death, with no apparent cause other than that of vascular origin (codes 430 to 434 and 436 to 437, International Classification of Diseases, 9th Revision [ICD-9]).12 13 The recently developed codes of the Application of the International Classification of Diseases to Neurology (ICD-10 NA) were also applied to patients who had neuroimaging studies of the brain.14
The L’Aquila district is a delimited mountainous area of 5034.46 km2 of the Abruzzo region, in central Italy, with continental weather patterns.15 Globally, there are 108 towns, 97 rural and 11 urban. The total study population, 48% rural, is 297 838 (1991 census).16 Mean per capita disposable income is 5.5% higher than in the rest of the nation (17 708 000 versus 16 780 000 lira).17
The population is served by 8 public hospitals, 8 private hospitals, 233 general practitioners, and 89 on-call physicians. Medical care is completely free of charge, allowing easy access to medical services within the acute phase of strokes.
To be included in the study, patients had to reside in the L’Aquila district at the time of the stroke occurrence and had to have suffered a first-ever stroke during the study period. All events were identified by active monitoring of all inpatient and outpatient health services. In each clinical ward, all patients admitted for a cerebrovascular event were identified and examined by a senior physician. Thereafter, all patients were seen by the consulting neurologist to validate the event. To verify all admitted stroke patients, eight consulting neurologists screened the admission and discharge lists on a daily basis. Nearby hospitals were regularly monitored to identify those residents who had cross-boundary medical care.8 The records of patients with dizziness, vertigo, confusion, seizures, headache, and transient global amnesia were also reviewed. Neuroradiology, neurophysiology, and neurosonology services were systematically checked. Regular contacts were also maintained with rehabilitation and long-term care services.
The study purpose was explained in advance to all general practitioners and on-call physicians who were asked to refer all stroke cases or give information about patients evaluated at home. Death certificates were checked monthly, and clinical details of all deceased patients with a diagnosis of stroke, not otherwise included in the registry, were reviewed. The use of a case-finding method including multiple overlapping sources allowed an assessment of the completeness of case ascertainment by means of a capture-recapture technique.18
Data Collection and Follow-up
Clinical and laboratory data were classified and recorded on standardized forms and stored in a computerized database.8 19 Basic information included medical history, cardiovascular and neurological evaluations, assessment of disability with the use of the Barthel Index,20 routine laboratory blood tests, 12-lead ECG, Doppler ultrasonography of neck vessels, transthoracic echocardiography, brain CT, and/or MRI. Transesophageal echocardiography, transcranial Doppler sonography of intracerebral arteries, three-lead continuous ECG, MR angiography, and/or digital subtraction angiography of cerebral arteries were performed in selected cases.
Risk factors such as arterial hypertension, diabetes mellitus, hypercholesterolemia, hypertriglyceridemia, smoking, alcohol abuse, history of migraine, and oral contraceptive use were screened, together with cardiac arrhythmias, coronary heart disease, left ventricular hypertrophy, previous transient ischemic attacks, and peripheral arterial disease.
Every effort was made to maintain diagnostic standards uniform throughout the study period. Consulting neurologists met weekly to discuss uncertain cases. Interrater agreement resulted in the almost perfect range (κ index, 0.82; 95% CI, 0.74 to 0.89). The reliability of laboratory data was tested among centers. All cases were followed up with quarterly planned visits or with a structured telephone interview. Outcome events to be considered during the follow-up were transient ischemic attacks, nonfatal stroke, nonfatal myocardial infarction, and death from either cardiovascular or noncardiovascular causes. Definitions are reported in Appendix 2.
Crude incidence rates together with 95% CIs for single binomial proportions were calculated by the exact approach with data from the 1991 census.16 Standardized rates were obtained by the direct method with 10-year age grouping of the Italian16 and European21 populations as standards. The expected number of patients missed by all the case-finding sources was estimated by a log-linear model including inpatient, outpatient, and death certificates sources, together with their second-order interaction terms. Pearson and Mantel-Haenszel χ2 tests or Student’s t test were used to compare groups as appropriate. χ2 for trend was performed to evaluate the linear relation between categorical variables. The linear association between the mean per capita disposable income referring to each of the 108 towns of the L’Aquila district17 and stroke incidence or case-fatality rate was evaluated by the Pearson’s correlation coefficient. Survival curves were estimated by the Kaplan-Meier method. Comparisons among the survival curves for the different subgroups were performed by the log-rank test. Two-sided values of P<.05 were considered to indicate statistical significance. All the data were analyzed with SPSS software.
In the population of the L’Aquila district, the proportion of subjects older than 65 years at the 1991 census showed a 18.3% increase compared with the 1981 census and was higher (18.5%) than in Italy (15.3%) and Europe (14.4%) (Fig 1⇓).
Between January 1 and December 31, 1994, 1047 patients with clinical signs attributable to cerebrovascular events were identified. After comprehensive evaluations, 819 patients were diagnosed as suffering from a first-ever stroke. Two hundred twenty-eight patients were excluded because of nonfatal (n=23) or fatal (n=24) recurrent stroke, residence out of the district (n=53), transient ischemic attacks (n=86), stroke due to head trauma (n=41), and perinatal intraventricular hemorrhage (n=1). Among the 819 patients there were 398 men and 421 women. The mean±SD age was 74.8±11.3 years, with a range of 24 to 106 years. Women were older than men (76.0±10.5 versus 73.4±11.9 years; P=.0008 by Student’s t test). Seven hundred fifty-four patients (92%) were hospitalized, either within (n=683) or out of the district (n=71). Sixty-five first-ever strokes (8.0%) occurred in nonhospitalized patients: 49 were reported by general practitioners, and 16 were traced by screening death certificates. The mean±SD duration of the hospital stay was 13.4±12.6 days (median, 11) without any significant correlation with age, sex, and stroke types (P>.05 for all comparisons). Repeated reporting of the same event from different and independent case-finding sources occurred in 309 patients (37.7%). It is estimated that 10 patients (1.2%) were missed by the capture-recapture technique.
Brain CT (n=599), MRI (n=112), or both (n=18) were performed at least once in 729 patients (89%) within a median time of 7 days from onset. They were not performed in 90 patients because of very early death (n=27), refusal (n=15), exclusive home care of very old patients (n=11), and equipment breakdown (n=37). Coexisting asymptomatic lacunar lesions were observed in 173 patients (24%) and clinically silent cerebral infarctions in 71 (10%). Table 1⇓ presents the distribution of stroke types according to the recently proposed ICD-10 NA in comparison with the ICD-9. Twenty-four patients (2.9%; 95% CI, 2.0 to 4.3) had subarachnoid hemorrhage, 122 (14.9%; 95% CI, 12.6 to 17.5) had intracerebral hemorrhage, 657 (80.2%; 95% CI, 73.4 to 82.8) had cerebral infarction, and 16 (2.0%; 95% CI, 1.2 to 3.1) suffered from an ill-defined cerebrovascular event.
The crude annual incidence rate of first-ever stroke was 2.75/1000 (95% CI, 2.57 to 2.94). Incidence rates showed a steep rise in those older than 65 years and were higher in men than in women (Fig 2⇓, Table 2⇓). The incidence rate, standardized by age and sex to the 1991 Italian population, was 2.37/1000; the corresponding rate, standardized to the 1991 European population, was 2.28/1000. In patients older than 80 years the crude incidence rate was 24.23/1000 (95% CI, 21.65 to 27.10) because of the higher incidence (χ2=4.80; P=.028) in men (28.25/1000; 95% CI, 23.85 to 33.45) than in women (21.79/1000; 95% CI, 18.75 to 25.32). Incidence rates were higher in rural than in urban centers (2.99 versus 2.48/1000; P=.007); after adjustment for age and sex, this difference lost its significance (P=.91). The mean per capita annual disposable income ranged from 12 138 000 to 22 139 000 lira across the district without any correlation with stroke incidence (r=−.0528; P=.294).
Crude annual incidence rates for the considered stroke types are reported in Table 2⇑. Four hundred thirty-four patients (53%) were older than 75 years, while 296 (36%) were older than 80 years, with a higher proportion of women (55%). Regardless of sex, the proportion of patients with a positive history of arterial hypertension (χ2 for trend=14.64; P<.0001), diabetes mellitus (4.78; P<.03), coronary heart disease (23.07; P<.0001), atrial fibrillation (17.15; P<.0001), peripheral arterial disease (13.79; P=.0002), and left ventricular hypertrophy (25.00; P<.0001) increased with age. On the other hand, a decreasing prevalence with age was found for smoking (28.87; P<.0001) and hypercholesterolemia (3.97; P<.05).
Two hundred ten patients died within 30 days of stroke onset (Table 3⇓). Ninety-five were men and 115 were women; 118 (56.2%) were older than 80 years; 161 died because of the qualifying stroke or a new fatal stroke (n=8), cardiovascular death (n=29), and nonvascular death (n=12). The 30-day case-fatality rate was 25.6% (95% CI, 22.8 to 28.7). Ninety-two additional patients, 38 men and 54 women, died within 1 year (Table 3⇓) because of the qualifying stroke (n=6) or a new fatal stroke (n=15), cardiovascular death (n=33), nonvascular death (n=34), or unknown death (n=4). At the end of the first year of follow-up, 198 patients (24.2%) were functionally independent, 319 (38.9%) were functionally dependent, and 302 died. The 1-year case-fatality rate was 36.9% (95% CI, 33.6 to 40.2).
As shown by the Kaplan-Meier estimate, most of the fatal events occurred early after stroke onset in all stroke types. Patients with intracerebral hemorrhage had a lower survival than patients suffering subarachnoid hemorrhage or cerebral infarction (P<.0001, log-rank test; χ2 for trend=44.57; P<.0001). Differences in survival among stroke types persisted until the end of the first year of follow-up.
Stroke incidence depends on the age distribution of the study population and accuracy of case-finding methods to identify fatal and nonfatal events occurring both in and out of the hospital.1 2 Comparisons between the most recent Italian censuses show a clear-cut drop in the birth rate and aging of the population that is particularly evident in our district.16
Compared with previous national registries,22 23 24 our study has shown the highest crude (2.75/1000) and age- and sex-adjusted (2.37/1000) annual incidence rates. The corresponding rate, standardized to the European population (2.28/1000), was also higher with respect to other international registries.5 8 25 26 Although the observed high stroke incidence was largely due to the older age of our population, other reasons for the increased stroke occurrence, culminating in highest rates in the oldest old, should be considered. Diligent case ascertainment might have been favored by the wide availability of free medical care and the tendency toward hospitalization, even in the very elderly.10 Cross-boundary medical care was provided in a minority of cases (8.7%). Complete enumeration of cases in the population might be a possible explanation for our rates. The estimate of possible missed cases by the capture-recapture technique (1.2%)18 shows that we met our recruitment goal. Even if a biased identification of the population denominator might have occurred since the study was performed 3 years after the last census, incidence estimates based on an extrapolation of the 1994 population were only marginally different. Misdiagnosis of stroke leading to the inclusion of patients with strokelike episodes might also have occurred. However, the rates of hospitalization (92%) and of diagnoses supported by neuroimaging techniques (89%) were remarkable8 22 23 24 25 26 27 and might have contributed to avoid this bias. Fifty-nine percent of the missing examinations were due to very early death and refusal or exclusive home care in very old patients. Thus, we should also consider the occurrence of a truly high incidence depending on the greatest increase of stroke risk in the older groups of the population. In our opinion, the observed declining incidence of stroke, thus far interpreted as a consequence of preventive measures,3 might also be attributed to a postponement in the age of onset of the first-ever stroke. Therefore, any further aging of the population will increase stroke occurrence, not only as a consequence of the increased proportion of elderly subjects but also because of their increased risk.5 28
Fifty-three percent of our patients were older than 75 years, confirming that stroke is a condition that affects elderly people.5 8 22 23 24 25 The exclusion of stroke patients older than 75 years may explain the lower incidence rates reported by some other studies.2 29 Moreover, 36% of the first-ever strokes occurred in patients older than 80 years, a subgroup with a higher proportion of women (55%).3 8 However, as a result of the lower life expectancy of men with respect to women, the corresponding annual incidence rates in patients older than 80 years were higher in men (28.25/1000).3 22 23 24 Duration of hospital stay was similar to that reported in other studies and did not show any difference according to age, sex, and stroke type.25 Proportions of arterial hypertension, together with many atherogenic risk factors, and atrial fibrillation increased with age, while smoking and hypercholesterolemia showed an opposite trend that might be explained as a consequence of lifestyle changes or of selective survival. Lack of similar data from other registries precluded comparisons, which would be particularly useful now that a longer survival to the age of 80 years and beyond has been achieved.2 10 29
Despite misclassification that might have occurred in a minority of our patients with very early death, the distribution of stroke types was similar to that shown by other studies in the presence of a particularly low rate of ill-defined events (2.0%) as a consequence of the high rate of clinical diagnoses confirmed by neuroimaging techniques.22 23 24 27 Compared with the ICD-9, the ICD-10 NA helped to better define stroke types and their pathogenic mechanisms.13 14 Unrelated asymptomatic lacunar lesions (24%) and silent cerebral infarctions (10%) were disclosed in proportions similar to those found in other studies.30 31
The overall 30-day case-fatality rate (25.64%) as well as rates referring to patients with cerebral infarction (20.6%) and intracerebral (51.6%) or subarachnoid hemorrhage (29.2%) were in the range of those reported by other studies.8 22 23 24 27 32 33 In addition, we confirmed the direct relationship between age and death,23 32 with the highest proportion of fatal events (56.2%) occurring in patients older than 80 years. The same pattern of mortality extended to 1-year results, showing that stroke still represents a major cause of death.
Cross-national studies have suggested a relation between income and vascular diseases, although the mechanisms underlying this possible association are poorly understood.11 In our study the lack of any association between per capita disposable income and stroke incidence makes this bias unlikely. In fact, if the proportion of the rural population (48%) might have somehow reduced the impact of preventive measures, the wide availability of free medical care might have counteracted this bias.10 The higher crude incidence rate that we found in rural with respect to urban centers (2.99/1000 versus 2.48/1000) might be explained by the age and sex distribution of the population, thus rendering the impact of any socioeconomic bias improbable.11 17
To limit the influence of comorbidity and to enhance the chance of showing any positive effect of potentially harmful treatments,34 acute stroke trials have excluded patients older than 80 years. However, the relevance of any positive findings might be somehow limited if not applicable to patients older than 80 years, in whom the majority of deaths are likely to occur.10 23 32 Nevertheless, as an effect of aging the absolute number of strokes in the population will increase. In addition, the survival bias in the oldest old might contribute toward modifying the proportion and relevance of risk factors and of comorbidity itself. For this reason, monitoring stroke, particularly in the older age groups, should be a priority to identify age-specific preventive measures.2 The corroboration of our findings will be of the greatest importance from a public health perspective.
Together with the general practitioners, the following investigators participated in the study: L’Aquila: Ospedale Civile (G. Azzarone, M. Bonamini, F. Caione, G. Cerone, N. Cimini, E. D’Alessandro, P. D’Amore, R. De Amicis, F. De Paulis, M. Euforbio, E. Farina, V. Festuccia, M. Gallucci, A. Iannessi, P. Marsili, G. Masciocchi, A. Santucci, G. Spacca, C. Spartera, G. Taglieri, G. Tonietti, L. Triggiani); Clinica Sanatrix (L. Onori). Avezzano: Ospedale Civile (G. De Blasis, L. Federici, G. Rabitti, L. Tiburzi, A. Venditti); Clinica Di Lorenzo (A. Petroni). Canistro: Clinica INI (L. Onori). Carsoli: Pronto Soccorso (G. Costantini). Castel di Sangro: Ospedale Civile (P. Caracciolo). Celano: Clinica Immacolata (G. Polino). Pescina: Ospedale Civile (A. Biocca). Sulmona: Ospedale Civile (E. Conti, C. Di Tommaso, P. Guerrini). Tagliacozzo: Ospedale Civile (A. Trotta). Pescara: Ospedale Civile (R. Buonaguidi); Villa Serena (A. Serio). Popoli: Ospedale Civile (U. Colangelo, A. Falco, A. Mobilij, V. Quatraro). Roma: Policlinico Umberto I (M. Rasura); Ospedale S. Giovanni Battista (M.R. Zylberman); Clinica S. Lucia (S. Paolucci). Sora: Ospedale Civile (F. Pericoli Ridolfini, A. Quadrini). Subiaco: Ospedale Civile (M. Marci, F. Russo). Teramo: Ospedale Civile (C. Bernardini, R. Galzio). Project Coordinators: A. Carolei, F. di Orio; Coordinating and Statistics Center: C. Marini, M. Di Napoli; Data Monitoring Committee: A. Carolei, S. Mearelli, R. Totaro.
Cerebral infarction was defined as a neurological deficit of sudden onset, documented by a brain CT or MRI indicating the presence of infarction or the absence of hemorrhage. In the absence of neuroimaging techniques or necropsy examination, cases of stroke with sufficient clinical details were diagnosed as probable cerebral infarction. Intracerebral hemorrhage was defined as a neurological deficit, documented by a brain CT or MRI indicating the presence of an intracerebral hemorrhage. In the absence of neuroimaging techniques or necropsy examination and of occlusive peripheral vascular disease, a diagnosis of probable intracerebral hemorrhage was made in the presence of clinical manifestations reflecting increased intracranial pressure such as headache and vomiting, decreased alertness or coma, and gradual progression to death within 24 hours of onset. Subarachnoid hemorrhage was diagnosed in a patient who presented with a typical clinical syndrome of neurological deficit that may progress to coma, usually with headache, nuchal rigidity, and evidence of hemorrhage in the subarachnoid space by lumbar puncture, brain CT, MRI, or necropsy. Ill-defined or unclassified cerebrovascular events were considered all strokes that could not be categorized as ischemic, hemorrhagic, or due to subarachnoid hemorrhage because of the absence of adequate clinical signs or symptoms and confirmatory investigations.
Transient ischemic attack was defined as an episode of focal cerebral dysfunction, presumably ischemic in origin, lasting less than 24 hours and followed by a return to normality. Recurrent stroke was defined as any new fatal and nonfatal event subsequent to the initial one, with an increased handicap at the time of the event, persisting beyond 24 hours. Cerebral death was defined as death occurring within or after 30 days of the onset of signs or symptoms of the qualifying stroke or of a new stroke, with clinically proven rostrocaudal deterioration, in the absence of other intervening causes. Cardiovascular deaths included sudden death (in which the death was seen by an eye witness, with a reliable observation of the time between the onset of symptoms and death, or the patient was found dead) or death from myocardial infarction, congestive heart failure, systemic embolism, or other cardiovascular causes (including pulmonary embolism and peripheral arterial disease). Noncardiovascular deaths included cancer, pneumonia, sepsis, neoplasia, and deep vein thrombosis. Unknown death was diagnosed in the presence of underlying pathology not otherwise specified.
Arterial hypertension was defined as known hypertension treated with antihypertensive therapy or systolic blood pressure >160 mm Hg and/or diastolic blood pressure >90 mm Hg on two different occasions. Diabetes mellitus was diagnosed if patients gave a history of diabetes that was confirmed in their medical records or were taking insulin or an oral hypoglycemic agent. Alternatively, patients with diabetes were included if there was a random nonfasting blood glucose concentration ≥11 mmol/L. Hypercholesterolemia was defined as fasting cholesterol serum level >220 mg/dL at recruitment. Hypertriglyceridemia was defined as fasting triglycerides serum level >150 mg/dL at recruitment. Smoking status was defined as never, current, or ex-smoker of any kind of tobacco. Alcohol abuse was diagnosed in the presence of a daily consumption >120 g. Migraine was diagnosed in the presence of a history of migraine with or without aura. Oral contraceptive use was recorded as current use during the last 6 months. Cardiac arrhythmias and conduction disturbances were diagnosed in the presence of ventricular ectopic beats, atrial fibrillation, paroxysmal supraventricular tachycardia, Wolff-Parkinson-White syndrome, bundle branch block, or atrioventricular block on a standard 12-lead ECG. Coronary heart disease was defined as a history of acute myocardial infarction or angina pectoris. Left ventricular hypertrophy was coded as present when documented in a standard 12-lead ECG. Peripheral arterial disease was diagnosed in the presence of a history of intermittent claudication or previous arterial intervention or Doppler ultrasonography documentation.
This study was supported by a grant (CNR 96.03027.CT04) from the Consiglio Nazionale delle Ricerche. We are indebted to Frank M. Yatsu, MD, for thoughtful review of the manuscript and to Jenny van Drimmelen, MD, and Matilde Leonardi, MD, for helpful assistance in the use of ICD-10 NA.
Reprint requests to Professor Antonio Carolei, MD, Clinica Neurologica, Dipartimento di Medicina Interna e Sanità Pubblica, Università degli Studi di L’Aquila, 67100 L’Aquila-Collemaggio, Italy.
The persons and institutions contributing to the L’Aquila Stroke Registry are listed in Appendix 1.
- Received July 8, 1997.
- Revision received August 13, 1997.
- Accepted August 26, 1997.
- Copyright © 1997 by American Heart Association
Sudlow CLM, Warlow CP. Comparing stroke incidence worldwide: what makes studies comparable? Stroke.. 1996;27:550–558.
Bonita R, Beaglehole R. Monitoring stroke: an international challenge. Stroke.. 1995;26:541–542.
Bonita R, Broad JB, Anderson NE, Beaglehole R. Approaches to the problems of measuring the incidence of stroke: the Auckland Stroke Study, 1991–1992. Int J Epidemiol.. 1995;24:535–542.
Bamford J, Sandercock P, Dennis M, Burn J, Warlow C. A prospective study of acute cerebrovascular disease in the community: the Oxfordshire Community Stroke Project—1981–86, II: incidence, case fatality rates and overall outcome at one year of cerebral infarction, primary intracerebral and subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry.. 1990;53:16–22.
Smith GD. Income inequality and mortality: why are they related? BMJ. 1996;312:987–988.
World Health Organization. Manual of the International Statistical Classification of Diseases, Injuries, and Causes of Death, 9th Revision. Geneva, Switzerland; WHO; 1977:vol 1.
World Health Organization. Application of the International Classification of Diseases to Neurology (ICD-10 NA). 2nd ed. Geneva, Switzerland: WHO. In press.
Istituto Centrale di Statistica. Popolazione e abitazioni. 13° Censimento generale della popolazione e delle abitazioni, 20 ottobre 1991, Fascicolo provinciale 66, L’Aquila, Roma, 1993:1–300.
Servizio Informatico Statistico Territoriale. Serie statistiche comunali. Provincia di L’Aquila, Roma, 1991. Database.
McCarty DJ, Tull ES, Moy CS, Kwoh CK, LaPorte RE. Ascertainment corrected rates: applications of capture-recapture methods. Int J Epidemiol.. 1993;22:559–565.
Bogousslavsky J, Van Melle G, Regli F, for the Lausanne Stroke Registry Group. The Lausanne Stroke Registry: analysis of 1,000 consecutive patients with first stroke. Stroke.. 1988;9:1083–1092.
EUROSTAT. Demographic Statistics, 1994: Population and Social Conditions. Luxembourg, Belgium: Office for Official Publications of the European Communities; Series Yearbooks 3A; 1994:1–258.
Ricci S, Celani MG, Guercini G, Rucireta P, Vitali R, La Rosa F, Duca E, Ferraguzzi R, Paolotti M, Seppoloni D, Caputo N, Chiurulla C, Scaroni R, Signorini E. First-year results of a community-based study of stroke incidence in Umbria, Italy. Stroke.. 1989;20:853–857.
D’Alessandro G, Di Giovanni M, Roveyaz L, Iannizzi L, Pesenti Compagnoni M, Blanc S, Bottacchi E. Incidence and prognosis of stroke in the Valle d’Aosta, Italy. Stroke.. 1992;23:1712–1715.
Lauria G, Gentile M, Fassetta G, Casetta I, Agnoli F, Andreotta G, Barp C, Caneve G, Cavallaro A, Cielo R, Mongillo D, Mosca M, Olivieri PG. Incidence and prognosis of stroke in Belluno province, Italy: first-year results of a community-based study. Stroke.. 1995;26:1787–1793.
Friday G, Lai SM, Alter M, Sobel E, LaRue L, Gil-Peralta A, McCoy RL, Levitt LP, Isack T. Stroke in the Lehigh Valley: racial/ethnic differences. Neurology.. 1989;39:1165–1168.
Lindenstrøm E, Boysen G, Nyboe J, Appleyard M. Stroke incidence in Copenhagen, 1976–1988. Stroke.. 1992;23:28–32.
Yatsu FM, Becker C, McLeroy KR, Coull B, Feibel J, Howard G, Toole JF, Walker MD. Community hospital-based stroke programs: NC, Oregon, and New York, I: goals, objectives, and data collection procedures. Stroke.. 1986;17:276–284.
Broderick JP, Phillips SJ, Whisnant JP, O’Fallon WM, Bergstralh EJ. Incidence rates of stroke in the eighties: the end of decline of stroke? Stroke.. 1989;20:577–582.
Thorvaldsen P, Asplund K, Kuulasmaa K, Rajakangas A-M, Schroll M, for the WHO MONICA Project: Stroke incidence, case-fatality, and mortality in the WHO MONICA Project. Stroke. 1995;26:361–367.
Ricci S, Celani MG, La Rosa F, Righetti E, Duca E, Caputo N. Silent brain infarction in patients with first-ever stroke: a community-based study in Umbria, Italy. Stroke.. 1993;24:647–651.
Jørgensen HS, Nakayama H, Raaschou HO, Gam J, Olsen TS. Silent infarction in acute stroke patients: prevalence, localization, risk factors, and clinical significance: the Copenhagen Stroke Study. Stroke.. 1994;25:97–104.
Howard G, Walker MD, Becker C, Coull B, Feibel J, McLeroy K, Toole JF, Yatsu F. Community hospital-based stroke programs: NC, Oregon, and New York, III: factors influencing survival after stroke: proportional hazards analysis of 4219 patients. Stroke.. 1986;17:294–299.
Kunitz SC, Gross CR, Heyman A, Kase CS, Mohr JP, Price TR, Wolf PA. The Pilot Stroke Data Bank: definition, design, and data. Stroke.. 1984;15:740–746.