Article Text
Abstract
Background Recent data have suggested that stroke incidence in young people may be rising. In this population-based study, we aimed to determine whether the incidence of stroke in people aged <55 years old had changed over the last three decades.
Methods All cases of first-ever stroke (ischaemic stroke, spontaneous intracerebral haemorrhage, and undetermined stroke) occurring in Dijon, France, from 1985 to 2011 were prospectively collected from a population-based registry. Incidence rates were calculated and temporal trends were analysed by age groups and stroke subtypes using a Poisson regression to estimate incidence rate ratios (IRR). Risk factors and premorbid treatments were analysed.
Results Over the 27-year study period, 4506 patients were recorded (53% women, mean age 74.6±14.4, 10.1% aged <55 years). An increase in overall stroke incidence was noted, as was a rise in ischaemic stroke in individuals aged <55 years (IRR 1.308; 95% CI 0.982 to 1.741, p=0.066 for period 1994–2002 vs period 1985–1993, and IRR 1.697; 95% CI 1.340 to 2.150, p<0.001 for period 2003–2011 vs period 1994–2002), which was consistent for men and women. In these young patients, smoking was the most frequent risk factor (43%).
Conclusions Multiple factors may account for the increased incidence of ischaemic stroke in people aged <55 years including changes in vascular risk factors, better awareness of the disease and treatment options in the population and among practitioners leading to more frequent referrals for specialised care, and improvements in stroke diagnosis. Stroke prevention must be encouraged even in young adults.
- CEREBROVASCULAR DISEASE
- EPIDEMIOLOGY
- STROKE
Statistics from Altmetric.com
Introduction
The incidence of stroke dramatically rises with age. The incidence increases by a factor of approximately 100 between the age of 40 and 80 years,1 ,2 which has led healthcare professionals to consider stroke a disease of the elderly, and an uncommon disorder in the young. Most recent population-based studies conducted in Western countries have reported that the overall incidence of stroke over time is either stable3 ,4 or falling,1 ,5–7 and also a rise in the mean age at stroke onset.8 A meta-analysis of available data concluded that there was a 42% decrease in stroke incidence in high-income countries from 1970 to 2008, whereas stroke incidence in low to middle-income countries more than doubled over the same time.9 Improvements in preventive treatment and reductions in risk factors at the population level accounted for the reduction in stroke incidence in high-income countries, whereas the rise in stroke incidence observed in low to middle-income countries could reflect transitions in the general health status and demographics in these countries.9 A recent study conducted in the Greater Cincinnati/Northern Kentucky region pointed out a significant increase in the incidence of ischaemic stroke in individuals aged 20–44 years old between 1993 and 2005.10 Among the reasons that could explain their findings, the authors pointed out an increase in the prevalence of vascular risk factors, including diabetes mellitus, obesity and illicit drug use in younger people.
In this study, we aimed to analyse temporal trends in the incidence and associated risk factors of stroke in people aged <55 years old, from 1985 to 2011, using data from a prospective population-based registry.
Methods
Case-ascertainment procedures and stroke definition
The Dijon Stroke Registry is a population-based study conducted among the residents of the city of Dijon, France (2007 census: 151 543 inhabitants), since 1985. The methodology used to achieve exhaustive collection of stroke cases has been extensively described elsewhere.2 ,11–13 This registry complies with the defined criteria for conducting ‘high quality’ stroke incidence studies.14 ,15 Indeed, the use of multiple overlapping sources of information allows to identify fatal and non-fatal stroke in hospitalised and non-hospitalised patients in the geographically defined area.2 ,11–13
Stroke is defined according to WHO diagnostic criteria as ‘rapidly developing clinical signs of focal (at time global) disturbance of cerebral function, lasting more than 24 h or leading to death with no apparent cause other than that of vascular origin’.16 Patients who received intravenous thrombolytic therapy or mechanical revascularisation are considered as having stroke even though total recovery was observed within 24 h. The diagnosis of stroke subtype is made on clinical signs with the help of cerebral imaging, and complementary examinations. Most patients (86%) are managed at the University Teaching Hospital, where the research team is located. These patients are ascertained, thanks to a ‘hot pursuit’ procedure, since they are systematically referred to a trained study neurologist, who is responsible for stroke care and data collection. A detailed history is obtained from each patient with a standardised questionnaire, and a neurological examination is systematically performed. Data are recorded in the patients’ medical files that are then reviewed by investigators. A ‘cold pursuit’ procedure is used to collect data from patients not referred to the University hospital. Hence, patients hospitalised in the clinical departments of the three private hospitals of the city and its suburbs (10%) are seen by a neurologist working in these establishments. Medical records are then transmitted to and reviewed by the investigators of the Dijon Stroke Registry. For out-of-hospital patients (4%), study investigators are responsible for contacting practitioners in charge of patient care in order to obtain clinical information. When disagreements between adjudicators need to be resolved, stroke adjudication is based on meetings of study investigators. For this study, only first-ever symptomatic stroke in a lifetime was considered and was classified as ischaemic stroke, spontaneous intracerebral haemorrhage (ICH) or undetermined stroke. The present study included patients registered between 1 January 1985 and 31 December 2011.
Vascular risk factors and prestroke treatments
Over the whole study period, the same methodology was used to identify vascular risk factors among stroke patients2 ,11–13: hypertension, diabetes mellitus, hypercholesterolaemia, atrial fibrillation, coronary heart disease, smoking and previous transient ischaemic attack (TIA). Premorbid therapy was also recorded, including oral anticoagulants (warfarin, acenocoumarol, or fluindione), antiplatelet agents (aspirin, clopidogrel, ticlopidine, or dipyridamole), antithrombotics (either anticoagulant and/or antiplatelet agent) and antihypertensive treatment. Statins were not included because they were only recorded from 2006 onwards.
Statistical analysis
So as to measure the incidence rates, census data for 1982, 1990, 1999 and 2007 concerning the population in Dijon in 1-year age groups and by sex were provided by the National Institute of Statistics. The population was estimated from these censuses by linear interpolation. We calculated the crude incidence rate per age group and the total standardised rate for every year and according to sex using the direct method with the European standard population.17 We assumed Poisson distribution for the annual number of events to estimate 95% CIs for the rates. The results were presented according to three study periods: 1985–1993, 1994–2002, and 2003–2011. To evaluate the impact of the time periods on stroke incidence, incidence rate ratios (IRRs) were calculated using a Poisson regression. To investigate the reasons for the observed trends in the incidence in patients <55 years old with ischaemic stroke, we evaluated changes in the prevalence of vascular risk factors and prestroke treatments. For comparisons between periods, we used a logistic regression. The ORs and their related 95% CIs were calculated. We used a dummy indicator for smoking status so as to prevent the deletion of data for patients with missing values. The proportion of missing values for other variables was less than 1%. p Values <0.05 were considered statistically significant. Statistical analysis was performed with STATA@ V.9.0 software (StataCorp LP, College Station, Texas, USA).
Ethics
The Dijon Stroke Registry was approved by the Comité National des Registres (French National Committee of Registers) and the InVS (French Institute for Public Health Surveillance). Authorisation of the CNIL (‘Commission Nationale de l’Informatique et des Libertés’; National Commission for the Protection of the Privacy of Electronic Data) was obtained.
Results
Over the 27-year study period, 4506 patients with a first-ever stroke were recorded (53% women, mean age±SD: 74.6±14.4). Among these patients, 453 (10.1%) were aged <55 years old. The proportion of strokes among individuals aged <55 years old was 9% in 1985–1993, 9% in 1994–2002 and 11.8% in 2003–2011.
The overall incidence of stroke slightly increased over time, especially in the last study period (IRR 1.031; 95% CI 0.957 to 1.111, p=0.43 for period 1994–2002 vs period 1985–1993, and IRR 1.097; 95% CI 1.023 to 1.176, p=0.009 for period 2003–2011 vs period 1994–2002). Tables 1 and 2 show age-specific incidence trends. A great increase in the incidence of stroke was observed for patients aged <55 years between period 1994–2002 and period 2003–2011 (IRR 1.597; 95% CI 1.282 to 1.988, p<0.001), whereas only a non-significant trend was noted for period 1994–2002 versus period 1985–1993. Of note, a slight but significant increase in stroke incidence was also observed for individuals aged 55–64 years old between the first two periods. Incidence rates were stable over time in the other age groups.
The observed rise in incidence in younger people between 1994–2002 and 2003–2011 was consistent in men (IRR 1.672; 95% CI 1.234 to 2.263, p=0.001) and in women (IRR 1.514; 95% CI 1.102 to 2.080, p=0.010). Stratified analyses by stroke subtypes revealed that this increased incidence was explained by a rise in the incidence of ischaemic stroke, which was particularly pronounced between the last two study periods (IRR 1.308; 95% CI 0.982 to 1.741, p=0.066 for period 1994–2002 vs period 1985–1993, and IRR 1.697; 95% CI 1.340 to 2.150, p<0.001 for period 2003–2011 vs period 1994–2002) (tables 3 and 4), in men and women (figure 1). Conversely, there was no significant change in the incidence of ICH and undetermined stroke, although the number of patients in the latter group was small (table 2). The observed trends were consistent in people aged <45 and those aged 45–54 years old, though the magnitude of the increase in the incidence of ischaemic stroke between period 1994–2002 and 2003–2011 was greater in people <45 years old (see online supplementary tables S1 and S2).
The vascular risk factor profile and premorbid treatments in ischaemic stroke patients aged <55 years old are shown in table 5. Whatever the time period, smoking was the most frequent risk factor. The prevalence of hypercholesterolaemia increased between 1985–1993 and 1994–2002 (OR 2.09; 95% CI 0.99 to 4.40, p=0.053), as did the prestroke use of antiplatelet agents (OR 3.84; 95% CI 1.07 to 13.8, p=0.040) (see online supplementary table S3). By contrast, no change was observed for other risk factors and prestroke treatments.
Discussion
This population-based study revealed an apparent rise in the incidence of ischaemic stroke in people aged <55 years over the last past three decades. This rise was particularly pronounced in recent years and contrasts with stable overall incidence of stroke in older age groups. In this group of younger patients, smoking was the most frequent vascular risk factor.
Our results are consistent with the increase in the incidence of ischaemic stroke in individuals aged 20–44 years old between 1993 and 2005 that was reported in the Greater Cincinnati/Northern Kentucky region.10 Our findings are also in accordance with the observed trend toward increasing hospitalisation rates for stroke and TIA between 2000 and 2007 in people aged <65 years old at a nationwide level in France.18 This latter study was based on administrative data, as was that from the Nationwide Inpatient Sample of the Healthcare Cost and Utilisation Project, which also noted a rise in the prevalence of hospitalisation for ischaemic stroke in people <45 years old between 1995 and 2008 in the USA.19
Taken together, these findings raise important questions about the reasons for this trend. It has been assumed that the rising prevalence of vascular risk factors in young adults may have accounted for the observed increase in the incidence of ischaemic stroke at this age. Hence, the National Health and Nutrition Examination Surveys (NHANES) described a rise in the prevalence of diabetes, hypercholesterolaemia, and obesity in the USA from 1988 to 2006.20 A high prevalence of cigarette smoking and alcohol abuse may also be of great importance to explain these trends.21 ,22 Additionally, in the Cincinnati study, the authors reported a rise in the use of illicit drugs among young stroke patients over time.23 Although we have no specific data for the Dijon population, nationwide estimates showed an increase in the prevalence of diabetes in young adults in France over time.24 Additionally, a large study conducted in southwestern France showed that smoking did not decrease in men aged 35–44 years old, and even increased in women of the same age group between 1985 and 1997.25 Similarly, the prevalence of obesity rose in young men but remained stable in women between 1986 and 2006 in the MONICA-France survey.26 In the present study, we found that smoking was the most frequent vascular risk factor in ischaemic stroke patients aged <55 years old. This high prevalence is consistent with results from previous studies,10 and contrasts with what is observed in older patients in whom hypertension and atrial fibrillation are more common.27 Concerning temporal trends, only the prevalence of prestroke hypercholesterolaemia increased. One reason could be the fact that the definition of hypercholesterolaemia included the use of cholesterol lowering drugs, which may have increased in a similar manner to that observed for antiplatelet agents. Given that our database did not include statin use before 2006, we cannot confirm this hypothesis. However, the contribution of these results to the interpretation of the observed trends in incidence is limited since we included patients with stroke, that is, those in whom prevention had failed. Nevertheless, these data indicate that the vascular risk factor profile of young stroke patients did not dramatically change with time. In particular, we did not observe any increase in prestroke coronary heart disease, which contrasts with what was previously reported in other studies.10 ,19
Beyond these hypotheses, other explanations not explored in this study could have accounted for the observed increase in stroke incidence in young people. Minor changes in socioeconomic factors in young people living in Dijon might have occurred over the study period, but we do not think that they really influenced our results about stroke incidence. More importantly, the attitude of patients towards seeking medical attention or of practitioners towards referring their patients with stroke symptoms to specialists may have changed with time, thus leading to a better identification of stroke patients. Additionally, even though we employed constant methodology over 27 years, we cannot exclude the possibility that case-ascertainment improved with time. Thanks to the use of brain MRI, stroke diagnosis has been made easier, especially in patients presenting with minor or rapidly regressive symptoms. Hence, it could be argued that the apparent increase in stroke incidence in young people may reflect a shift from the diagnosis of TIA towards that of ischaemic stroke. However, in order to avoid the issue, we used the epidemiological definition of stroke, that is, neurological signs lasting more than 24 h, throughout the study period. Therefore, the differential diagnosis between ischaemic stroke and TIA was not based on imaging results, especially MRI, but exclusively on the duration of symptoms, which avoided any excess in the diagnosis of stroke rather than TIA related to the increased use of MRI over time. In support of this remark, there was also an increase in the incidence of TIA between 1994–2002 and 2003–2011 in people aged <55 years old (data not shown). Finally, if the observed change in incidence was exclusively related to better capture and diagnosis of stroke, there should also have been an increase in incidence in older people, which was not the case in our study.
The major strength of this study is the continuous prospective ascertainment of all cases of stroke in a geographically defined population based on overlapping sources of information. To the best of our knowledge, this is the longest ongoing population-based registry. However, our study was limited by the fact that 14% of patients were identified using a ‘cold pursuit’ procedure, which is associated with a lower quality of data collection than with ‘hot pursuit’.
To conclude, many factors may account for the increase in the incidence of ischaemic stroke in people aged <55 years old, and these need to be elucidated. Stroke prevention must be encouraged even in young adults.
Acknowledgments
We thank Mr Philip Bastable for reviewing the English.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Footnotes
-
Contributors All authors have contributed to the article by participating in the conception and design (YB, MG), acquisition of data (YB, BD, JD, AJ, G-VO, OR, MG) or analysis and interpretation of data (YB, JD, MG), drafting the article (YB, MG) or revising it critically for important intellectual content (YB, BD, JD, AJ, G-VO, OR, MG), and approving the final manuscript (YB, BD, JD, AJ, G-VO, OR, MG).
-
Funding The Dijon Stroke Registry is supported by the French Institute for Public Health Surveillance (InVS) and Inserm. The study sponsors had no role in either the study design, or the collection, analysis and interpretation data, the writing of the report, or in the decision to submit the paper for publication.
-
Competing interests None.
-
Ethics approval The Dijon Stroke Registry was approved by the Comité National des Registres (French National Committee of Registers) and the InVS (French Institute for Public Health Surveillance). Authorisation of the CNIL (Commission Nationale de l'Informatique et des Libertés; National Commission for the Protection of the Privacy of Electronic Data) was obtained.
-
Provenance and peer review Not commissioned; externally peer reviewed.