- Email: [email protected]
Seasonal variation of hospital admissions caused by acute stroke in Athens, Greece
Seasonal Variation of Hospital Admissions Caused By Acute Stroke in Athens, Greece Konstantinos Spengos, MD,* Kostas N. Vemmos, MD,† Georgios Tsivgoulis, MD,† Andreas Synetos, MD,† Nikolaos Zakopoulos, Vassilios P. Zis, MD,* and Demitris Vassilopoulos, MD, PhD*
MD,†
Several studies have reported a seasonal variation in stroke incidence with a peak during winter. In a hospital-based study, we investigated the distribution pattern of stroke occurrence in general, as well as its subtypes throughout the year. We studied the monthly and seasonal distribution of acute stroke occurrence among 1299 first ever stroke patients admitted at our hospital and compared it with the assumed equal distribution of stroke over the year (null hypothesis) using 2 techniques. The seasonal distribution of the different stroke subtypes was also examined. Overall stroke incidence shows a seasonal variation with a significant peak during wintertime. Stroke etiology, cardioembolic stroke, and intracerebral hemorrhage are also characterized from a similar seasonal fluctuation with a significant trough in summer and peak in winter. On the contrary, cases of stroke caused by atherosclerotic large vessel disease, small vessel disease, and stroke caused by unusual or undetermined types show no seasonal periodicity. Seasonal variation of stroke seems to depend primarily on the etiology of stroke. Further studies are necessary in order to clarify this issue and provide useful evidence for stroke prevention and therapy. Key Words: Acute stroke—Greece— hospital admission—stroke subtypes. Copyright © 2003 by National Stroke Association
During the last decades, several studies from different countries and climate zones examined the seasonal variation of cerebrovascular diseases and led to discrepant findings. While some authors strongly doubt the existence of a typical circannual pattern,1 a major series of studies suggest a fluctuation with a peak in the colder period of the year.2-7 Recently published data also indicate a similar circannual pattern with a typical winter peak for several factors of etiologic relevance for the
From the *Department of Neurology, University of Athens School of Medicine, Eginition Hospital, and the †Department of Clinical Therapeutics, University of Athens School of Medicine, Alexandra Hospital, Athens, Greece. Received September 9, 2002; accepted January 14, 2003. Address reprint requests to Konstantinos Spengos, MD, Vasilissis Sofias Ave 82, 11528 Athens, Greece. E-mail: [email protected]. Copyright © 2003 by National Stroke Association 1052-3057/03/1202-0001$30.00/0 doi:10.1053/jscd.2003.15
occurrence of stroke.8-12 However, some other studies report a summer peak of stroke incidence,13 while others suggest an inverse seasonal relationship between different stroke types.14 Though the whole issue remains not completely clear, at least in all of its parameters, a seasonal variation of stroke seems to be now widely accepted. Using the recorded data of the Athens Stroke Registry (ASR) we examined the seasonal variation of stroke admission and hospitalization at our hospital for a period of 8 years, in order to clarify whether stroke occurrence in citizens in the capital city of Greece shows a certain seasonal pattern. Because little is reported about the seasonal variation of the different subtypes of stroke, we also investigated this issue throughout the year. Stroke subtypes were classified according to the diagnostic criteria of the ASR.15 Athens is the capital of Greece. It is characterized by its mild climate. Winter in Athens is relatively cold, with a
Journal of Stroke and Cerebrovascular Diseases, Vol. 12, No. 2 (March-April), 2003: pp 93-96
93
K. SPENGOS ET AL.
94
mean maximum and mean minimum ambient temperature of 12.9°C and 6.5°C, respectively, in January, while the summer is quite hot with a mean maximum and mean minimum ambient temperature of 32.6°C and 22.5°C, respectively, in July.
Subjects and Methods Data of 1299 patients with first-ever acute stroke, who were admitted at our hospital and included in the ASR in an 8-year period (from January 1, 1993 to December 31, 2000), were pooled for the analyses, and each year was divided into four seasons. Winter included December, January, and February; spring included March, April, and May; summer included June, July, and August; and autumn included September, October, and November. All patients were examined by an internist and a neurologist within the first 48 hours. In the same period of time a non-contrast CT scan of the brain was also performed. According to the diagnostic criteria of the ASR stroke was classified as stroke caused by large vessel atherothrombotic disease (LVA), cardioembolic stroke (CE), lacunar stroke (LAC), other (undetermined or unusual) type of stroke (OTH) and intracerebral hemorrhage (ICH). LVA stroke included stroke caused by distal insufficiency (vascular imaging showing occlusion or ⬎ 50% linear stenosis in the ipsilateral-to-infarct large vessel supply, border zone infarcts in the computed tomography [CT]/magnetic resonance imaging [MRI] scans, and additional clinical features as previous transient ischemic attacks [TIAs] in the same territory, stepwise or gradual onset, and history of atherosclerotic disease) as well as stroke caused by artery-to-artery embolism (vascular imaging showing occlusion or ⬎ 50% linear stenosis and/or ulceration ⬎ 2 mm in depth in the ipsilateral-to-infarct large vessel supply, ischemic changes in superficial fields in the ipsilateral cerebral artery supply, and additional clinical features, such as a history of atherosclerotic disease, sudden onset, and absence of major source of cardioembolism). CE stroke was diagnosed in presence of one or more cardiac sources of embolism with high-risk criteria (atrial fibrillation, atrial flutter or sick sinus syndrome, prosthetic valves, recent myocardial infarction, etc). In these cases vascular imaging should disclose occlusion or ⬎ 50% linear stenosis and/or ulceration ⬎ 2 mm in depth in the ipsilateral-to-infarct large vessel supply. Infarction in the territory of a single or in combination with multiple infarcts in other divisions of the major cerebral arteries and hemorrhagic infarction were the corresponding CT/ MRI findings. The diagnosis of LAC stroke was made with certain CT/MRI findings (small infarctions ⬍ 2 mm in the distribution of deep penetrating artery appropriate for clinical symptoms) in presence of history of hypertension and/or diabetes mellitus and absence of large vessel atherosclerotic changes. Diagnosis of ICH was based on
the initial CT scan findings. The group of OTH stroke included patients with inadequate evaluation, two possible pathogenic mechanisms, extensive workup with results that do not meet previous diagnostic criteria, or with less common causes of stroke (systemic hypoperfusion, dissection, vasculitis, cerebral venous thrombosis, and others). Only patients with first-ever stroke were included in our registry. Patients with subarachnoid hemorrhage were from the beginning treated by the neurosurgeons and therefore excluded. Patients with TIAs or recurrent stroke were also excluded from our study. Our stroke unit, housed in a university hospital, is a tertiary care center that covers the needs of the innercity population of Athens for stroke care together with other institutions, periodically providing emergency care services every fourth day. In order to verify the existence of a seasonal variation of admissions of stroke patients over the year, we documented the month of hospital admission for all 1299 cases. Assuming that stroke occurrence has no relation to a certain period of the year, the observed stroke cases should be evenly distributed throughout all 12 months and all 4 seasons of the year. We compared the observed global stroke frequency in 12 monthly and four 3-month periods with the expected frequency using the 2 techniques for a single population (for goodness to fit to the null model of equal distribution of strokes), along with 95% confidence intervals. The same statistical techniques were also used in order to evaluate a possible seasonal pattern for the above mentioned etiologic stroke subtypes.
Results In our evaluation we considered the month of hospital admission of 1299 first-ever stroke patients (778 men and 521 women) ranging in age from 22 years to 95 years (median age 71). According to the above-mentioned criteria, LVA as the cause of stroke was diagnosed in 192 patients (15%) whereas CE was the most frequent diagnosis, seen in 412 patients (32%). LAC stroke was identified in 219 cases (17%) while OTH was registered in 279 cases (21%). Finally, the remaining 197 patients (15%) presented stroke symptoms caused by an ICH. The monthly distribution of all 1299 stroke cases showed as expected a decline of stroke admissions, beginning in May and reaching its nadir in August, while an increase is observed in November (Fig 1). Stroke occurrence remains at high levels almost stable over the cold period of the year (November to April). According to the null hypothesis of equal distribution of stroke, we would expect stroke cases to be evenly distributed over all 12 months of the year. However, the observed number of cases differed significantly from the number expected (2 ⫽ 25.8; df ⫽ 11; P ⫽ .007), indicating the existence of
SEASONAL VARIATION OF STROKE ADMISSIONS
95
Figure 1 Monthly distribution of admissions of first ever stroke patients (n ⫽ 1299) recruited in The Athens Stroke Registry in the period between January 1, 1993 and December 31, 2000.
a circannual variation of hospital admissions caused by acute stroke. August and September were the months with fewer admissions, 73 and 90, respectively, while January and March were the months with the most stroke cases, both with 123 hospital admissions. By dividing the year into 4 seasons as above described, we observed a significant decline of stroke incidence during summer while most stroke cases were registered, as expected during winter (2 ⫽ 15.77; df ⫽ 3; P ⫽ .00126) (Table 1). We further examined the distribution of different stroke subtypes over the 4 seasons of the year. The same seasonal pattern with the peak of incidence in winter and its nadir in summer was identified for both groups of CE and ICH (Fig 2; Table 1). On the other hand, the distribution of stroke caused by LVA, LAC, and OTH did not differ significantly from random and showed no periodicity, as seen in Table 1.
Discussion Several studies identified a temporal variation of stroke incidence with a significant peak during winter and a trough during summer.2-7 There is a series of published data concerning the seasonal fluctuation of the different stroke subtypes. Results of other studies describe a similar seasonal distribution of different physiologic parameters that may etiologically correlate with stroke occur-
Table 1. Stroke by subtype and season during the eight years of the Athens Stroke Registry Winter Spring Summer Autumn LVA CE LAC OTH ICH All
47 126 48 69 74 364
45 114 59 79 49 346
47 78 48 65 31 269
53 94 64 66 43 320
2
P
0.750 13.165 3.557 1.76 19.993 15.775
.8614 .0043 .3134 .6237 .0002 .0013
Figure 2 Seasonal distribution of patients in the different etiologic stroke subtype groups. CE, cardioembolic stroke; ICH, intracerebral hemorrhage; LAC, lacunar stroke; LVA, stroke caused by large vessel atherothrombotic disease; OTH, other (undetermined or unusual) type of stroke.
rence. Positive antiphospholipid antibody titers occur less frequently in the summer months.8 Plasma levels of the vasoconstrictor endothelin show a fluctuation with a peak in the winter, particularly in January and February. During the same period of the year, the fluctuation of the vasorelaxant nitric oxide shows its nadir. These combined findings may account in part for the increased incidence of stroke and coronary heart disease, as it is mostly seen during winter.9 It has also been reported that fibrinogen and factor VII plasma values are significantly higher during winter. This seasonal variation could eventually provide possible explanation for the marked seasonal variation in death from ischemic heart disease and stroke in the elderly.10,11 Other authors suggest a correlation between the incidence of human diseases, including stroke, and the higher occurrence of geomagnetic storms during the colder period of the year.1 Analysis of meteorological data identifies cold ambient temperature as a risk factor for stroke.16,17 Other specialists described the central low pressure area with transition of the frontal system as the most unfavourable weather type for stroke occurrence.18 All these findings try to give a plausible explanation for the winter peak of stroke fluctuation. However, taking the different stroke subtypes in concern some other groups published discrepant findings, reporting the highest incidence of ischemic stroke and ICH during spring3 and autumn, respectively.4 Some authors seem to be doubtful about the existence of the reported winter excess of ischemic stroke,1 which could be an artifact caused by a referral bias in hospital-based studies and increased case fatality during winter in mortality studies.19,20 Regarding the incidence of subarachnoid hemorrhage, most studies report a higher case frequency during winter months.21,22 Our results suggest a seasonal variation in general stroke incidence with a significant decrease during summer and are in accordance with most published similar studies. Analysis of the variation in occurrence of the
K. SPENGOS ET AL.
96
different stroke subtypes leads to more differentiated conclusions. The predominance of ICH in winter is highly significant and in accordance with the results of various studies from different countries, which assume a triggering function of low ambient temperature.16,17 We also revealed a higher frequency of CE during the colder months of the year, while this kind of ischemic stroke was significantly less frequent during the summer. The already mentioned varying levels of coagulation parameters may result in a higher risk of formation of embolic material. Further studies about the seasonal variation in occurrence of acute atrial fibrillation and its possible relation with stroke incidence23 are needed. The subgroups of stroke caused by LAV, LAC, and OTH showed a distribution pattern that did not differ from random and were found not to follow a certain periodic cycle. In conclusion, our results are in accordance with the widely accepted model of seasonal variation of stroke incidence in general, with a peak in winter and a trough in summer. Further differentiation of stroke in its etiologic subtypes leads to non-uniform findings. ICH and CE follow the above described periodicity. In contrast, stroke caused by LVA, LAC, and OTH do not show a certain seasonal variation. In our opinion, these findings may partly explain the previously released discrepant data. Another interesting finding of our study is the large number of CEs that are already observed among Greek stroke patients and might be explained by the fact that Greek primary physicians still undertreat patients who should receive anticoagulants, especially those with nonvalvular atrial fibrillation.15 A major limitation of our study is the fact that it is hospital-based, which makes referral bias, especially into a tertiary care center, unavoidable. The summer vacation of the Athenians is another limitating factor, the importance of which is difficult to estimate. However, our hospital is open 365 day a year, which means that there are no possible biases because of a lack of medical personnel and capacities. Further investigations of the seasonal occurrence of the different stroke subtypes in correlation with possible pathogenic parameters are necessary in order to clarify the issue and may lead to important conclusions which could further optimize stroke prevention and therapy.
References 1. Rothwell PM, Wroe SJ, Slattery J, et al. Is stroke incidence related to season or temperature? The Oxfordshire Community Stroke Project. Lancet 1996;347:934-936. 2. Tsementzis SA, Kennet RP, Hitchcock ER, et al. Seasonal variation of cerebrovascular diseases. Acta Neurochir (Wien) 1991;111:80-83.
3. Oberg AL, Ferguson JA, McIntyre LM, et al. Incidence of stroke and season of the year: evidence of an association. Am J Epidemiol 2000;152:558-564. 4. Ricci S, Celani MG, Vitali R, et al. Diurnal and seasonal variations in the occurrence of stroke: a communitybased study. Neuroepidemiology 1992;11:59-64. 5. Jakovljevic D, Salomaa V, Sivenius J, et al. Seasonal variation in the occurrence of stroke in a Finnish adult population. The FINMONICA Stroke Register. Finnish Monitoring Trends and Determinants in Cardiovascular Disease. Stroke 1996;27:1774-1779. 6. Kochanowicz J, Kulakowska A, Drozdowski W. Seasonal variations in stroke incidence in North-Eastern Poland. Neurol Neurochir Pol 1999;33:1005-1013. 7. Shinkawa A, Ueda K, Hasuo Y, et al. Seasonal variation in stroke incidence in Hisayama, Japan. Stroke 1990;21: 1262-1267. 8. Luong TH, Rand JH, Wu XX, et al. Seasonal distribution of antiphospholipid antibodies. Stroke 2001;32:17071711. 9. McLaren M, Kirk G, Bolton-Smith C, et al. Seasonal variation in plasma levels of endothelin-1 and nitric oxide. Int Angiol 2000;19:351-353. 10. Stout RW, Crawford V. Seasonal variations in fibrinogen concentrations among elderly people. Lancet 1991;338:913. 11. Woodhouse PR, Khaw KT, Plummer M, et al. Seasonal variations of plasma fibrinogen and factor VII activity in the elderly: winter infections and death from cardiovascular disease. Lancet 1994;343:435-439. 12. Kuleshova VP, Pulinets SA, Sazanova EA, et al. Biotropic effects of geomagnetic storms and their seasonal variations. Biofizika 2001;46:930-934. 13. Gifford DK. Monthly incidence of stroke in rural Kansas. Kans Nurse 1996;71:3-4. 14. Biller J, Jones MP, Bruno A, et al. Seasonal variation of stroke— does it exist? Neuroepidemiology 1988;7:89-98. 15. Vemmos KN, Takis CE, Georgilis K, et al. The Athens stroke registry: results of a five-year hospital-based study. Cerebrovasc Dis 2000;10:133-141. 16. Azevedo E, Ribeiro JA, Lopes F, et al. Cold: a risk factor for stroke? J Neurol 1995;242:217-221. 17. Feigin VL, Nikitin YP, Bots ML, et al. A populationbased study of the associations of stroke occurrence with weather parameters in Siberia, Russia (1982-92). Eur J Neurol 2000;7:171-178. 18. Cabajova Z, Snopkova Z, Traubner P, et al. Effect of weather on the occurrence of cerebral stroke. Bratisl Lek Listy 1999;100:267-270. 19. Sheth T, Nair C, Muller J, et al. Increased winter mortality from acute myocardial infarction and stroke: the effect of age. J Am Coll Cardiol 1999;33:1916-1919. 20. Lanska DJ, Hoffmann RG. Seasonal variation in stroke mortality rates. Neurology 1999;52:984-990. 21. Nyquist PA, Brown Jr RD, Wiebers DO, et al. Circadian and seasonal occurrence of subarachnoid and intracerebral hemorrhage. Neurology 2001;56:190-193. 22. Passero S, Reale F, Ciacci G, et al. Differing temporal patterns of onset in subgroups of patients with intracerebral hemorrhage. Stroke 2000;31:1538-1544. 23. Kupari M, Koskinen P. Seasonal variation in occurrence of acute atrial fibrillation and relation to air temperature and sale of alcohol. Am J Cardiol 1990;66:1519-1520.
MD,†
Several studies have reported a seasonal variation in stroke incidence with a peak during winter. In a hospital-based study, we investigated the distribution pattern of stroke occurrence in general, as well as its subtypes throughout the year. We studied the monthly and seasonal distribution of acute stroke occurrence among 1299 first ever stroke patients admitted at our hospital and compared it with the assumed equal distribution of stroke over the year (null hypothesis) using 2 techniques. The seasonal distribution of the different stroke subtypes was also examined. Overall stroke incidence shows a seasonal variation with a significant peak during wintertime. Stroke etiology, cardioembolic stroke, and intracerebral hemorrhage are also characterized from a similar seasonal fluctuation with a significant trough in summer and peak in winter. On the contrary, cases of stroke caused by atherosclerotic large vessel disease, small vessel disease, and stroke caused by unusual or undetermined types show no seasonal periodicity. Seasonal variation of stroke seems to depend primarily on the etiology of stroke. Further studies are necessary in order to clarify this issue and provide useful evidence for stroke prevention and therapy. Key Words: Acute stroke—Greece— hospital admission—stroke subtypes. Copyright © 2003 by National Stroke Association
During the last decades, several studies from different countries and climate zones examined the seasonal variation of cerebrovascular diseases and led to discrepant findings. While some authors strongly doubt the existence of a typical circannual pattern,1 a major series of studies suggest a fluctuation with a peak in the colder period of the year.2-7 Recently published data also indicate a similar circannual pattern with a typical winter peak for several factors of etiologic relevance for the
From the *Department of Neurology, University of Athens School of Medicine, Eginition Hospital, and the †Department of Clinical Therapeutics, University of Athens School of Medicine, Alexandra Hospital, Athens, Greece. Received September 9, 2002; accepted January 14, 2003. Address reprint requests to Konstantinos Spengos, MD, Vasilissis Sofias Ave 82, 11528 Athens, Greece. E-mail: [email protected]. Copyright © 2003 by National Stroke Association 1052-3057/03/1202-0001$30.00/0 doi:10.1053/jscd.2003.15
occurrence of stroke.8-12 However, some other studies report a summer peak of stroke incidence,13 while others suggest an inverse seasonal relationship between different stroke types.14 Though the whole issue remains not completely clear, at least in all of its parameters, a seasonal variation of stroke seems to be now widely accepted. Using the recorded data of the Athens Stroke Registry (ASR) we examined the seasonal variation of stroke admission and hospitalization at our hospital for a period of 8 years, in order to clarify whether stroke occurrence in citizens in the capital city of Greece shows a certain seasonal pattern. Because little is reported about the seasonal variation of the different subtypes of stroke, we also investigated this issue throughout the year. Stroke subtypes were classified according to the diagnostic criteria of the ASR.15 Athens is the capital of Greece. It is characterized by its mild climate. Winter in Athens is relatively cold, with a
Journal of Stroke and Cerebrovascular Diseases, Vol. 12, No. 2 (March-April), 2003: pp 93-96
93
K. SPENGOS ET AL.
94
mean maximum and mean minimum ambient temperature of 12.9°C and 6.5°C, respectively, in January, while the summer is quite hot with a mean maximum and mean minimum ambient temperature of 32.6°C and 22.5°C, respectively, in July.
Subjects and Methods Data of 1299 patients with first-ever acute stroke, who were admitted at our hospital and included in the ASR in an 8-year period (from January 1, 1993 to December 31, 2000), were pooled for the analyses, and each year was divided into four seasons. Winter included December, January, and February; spring included March, April, and May; summer included June, July, and August; and autumn included September, October, and November. All patients were examined by an internist and a neurologist within the first 48 hours. In the same period of time a non-contrast CT scan of the brain was also performed. According to the diagnostic criteria of the ASR stroke was classified as stroke caused by large vessel atherothrombotic disease (LVA), cardioembolic stroke (CE), lacunar stroke (LAC), other (undetermined or unusual) type of stroke (OTH) and intracerebral hemorrhage (ICH). LVA stroke included stroke caused by distal insufficiency (vascular imaging showing occlusion or ⬎ 50% linear stenosis in the ipsilateral-to-infarct large vessel supply, border zone infarcts in the computed tomography [CT]/magnetic resonance imaging [MRI] scans, and additional clinical features as previous transient ischemic attacks [TIAs] in the same territory, stepwise or gradual onset, and history of atherosclerotic disease) as well as stroke caused by artery-to-artery embolism (vascular imaging showing occlusion or ⬎ 50% linear stenosis and/or ulceration ⬎ 2 mm in depth in the ipsilateral-to-infarct large vessel supply, ischemic changes in superficial fields in the ipsilateral cerebral artery supply, and additional clinical features, such as a history of atherosclerotic disease, sudden onset, and absence of major source of cardioembolism). CE stroke was diagnosed in presence of one or more cardiac sources of embolism with high-risk criteria (atrial fibrillation, atrial flutter or sick sinus syndrome, prosthetic valves, recent myocardial infarction, etc). In these cases vascular imaging should disclose occlusion or ⬎ 50% linear stenosis and/or ulceration ⬎ 2 mm in depth in the ipsilateral-to-infarct large vessel supply. Infarction in the territory of a single or in combination with multiple infarcts in other divisions of the major cerebral arteries and hemorrhagic infarction were the corresponding CT/ MRI findings. The diagnosis of LAC stroke was made with certain CT/MRI findings (small infarctions ⬍ 2 mm in the distribution of deep penetrating artery appropriate for clinical symptoms) in presence of history of hypertension and/or diabetes mellitus and absence of large vessel atherosclerotic changes. Diagnosis of ICH was based on
the initial CT scan findings. The group of OTH stroke included patients with inadequate evaluation, two possible pathogenic mechanisms, extensive workup with results that do not meet previous diagnostic criteria, or with less common causes of stroke (systemic hypoperfusion, dissection, vasculitis, cerebral venous thrombosis, and others). Only patients with first-ever stroke were included in our registry. Patients with subarachnoid hemorrhage were from the beginning treated by the neurosurgeons and therefore excluded. Patients with TIAs or recurrent stroke were also excluded from our study. Our stroke unit, housed in a university hospital, is a tertiary care center that covers the needs of the innercity population of Athens for stroke care together with other institutions, periodically providing emergency care services every fourth day. In order to verify the existence of a seasonal variation of admissions of stroke patients over the year, we documented the month of hospital admission for all 1299 cases. Assuming that stroke occurrence has no relation to a certain period of the year, the observed stroke cases should be evenly distributed throughout all 12 months and all 4 seasons of the year. We compared the observed global stroke frequency in 12 monthly and four 3-month periods with the expected frequency using the 2 techniques for a single population (for goodness to fit to the null model of equal distribution of strokes), along with 95% confidence intervals. The same statistical techniques were also used in order to evaluate a possible seasonal pattern for the above mentioned etiologic stroke subtypes.
Results In our evaluation we considered the month of hospital admission of 1299 first-ever stroke patients (778 men and 521 women) ranging in age from 22 years to 95 years (median age 71). According to the above-mentioned criteria, LVA as the cause of stroke was diagnosed in 192 patients (15%) whereas CE was the most frequent diagnosis, seen in 412 patients (32%). LAC stroke was identified in 219 cases (17%) while OTH was registered in 279 cases (21%). Finally, the remaining 197 patients (15%) presented stroke symptoms caused by an ICH. The monthly distribution of all 1299 stroke cases showed as expected a decline of stroke admissions, beginning in May and reaching its nadir in August, while an increase is observed in November (Fig 1). Stroke occurrence remains at high levels almost stable over the cold period of the year (November to April). According to the null hypothesis of equal distribution of stroke, we would expect stroke cases to be evenly distributed over all 12 months of the year. However, the observed number of cases differed significantly from the number expected (2 ⫽ 25.8; df ⫽ 11; P ⫽ .007), indicating the existence of
SEASONAL VARIATION OF STROKE ADMISSIONS
95
Figure 1 Monthly distribution of admissions of first ever stroke patients (n ⫽ 1299) recruited in The Athens Stroke Registry in the period between January 1, 1993 and December 31, 2000.
a circannual variation of hospital admissions caused by acute stroke. August and September were the months with fewer admissions, 73 and 90, respectively, while January and March were the months with the most stroke cases, both with 123 hospital admissions. By dividing the year into 4 seasons as above described, we observed a significant decline of stroke incidence during summer while most stroke cases were registered, as expected during winter (2 ⫽ 15.77; df ⫽ 3; P ⫽ .00126) (Table 1). We further examined the distribution of different stroke subtypes over the 4 seasons of the year. The same seasonal pattern with the peak of incidence in winter and its nadir in summer was identified for both groups of CE and ICH (Fig 2; Table 1). On the other hand, the distribution of stroke caused by LVA, LAC, and OTH did not differ significantly from random and showed no periodicity, as seen in Table 1.
Discussion Several studies identified a temporal variation of stroke incidence with a significant peak during winter and a trough during summer.2-7 There is a series of published data concerning the seasonal fluctuation of the different stroke subtypes. Results of other studies describe a similar seasonal distribution of different physiologic parameters that may etiologically correlate with stroke occur-
Table 1. Stroke by subtype and season during the eight years of the Athens Stroke Registry Winter Spring Summer Autumn LVA CE LAC OTH ICH All
47 126 48 69 74 364
45 114 59 79 49 346
47 78 48 65 31 269
53 94 64 66 43 320
2
P
0.750 13.165 3.557 1.76 19.993 15.775
.8614 .0043 .3134 .6237 .0002 .0013
Figure 2 Seasonal distribution of patients in the different etiologic stroke subtype groups. CE, cardioembolic stroke; ICH, intracerebral hemorrhage; LAC, lacunar stroke; LVA, stroke caused by large vessel atherothrombotic disease; OTH, other (undetermined or unusual) type of stroke.
rence. Positive antiphospholipid antibody titers occur less frequently in the summer months.8 Plasma levels of the vasoconstrictor endothelin show a fluctuation with a peak in the winter, particularly in January and February. During the same period of the year, the fluctuation of the vasorelaxant nitric oxide shows its nadir. These combined findings may account in part for the increased incidence of stroke and coronary heart disease, as it is mostly seen during winter.9 It has also been reported that fibrinogen and factor VII plasma values are significantly higher during winter. This seasonal variation could eventually provide possible explanation for the marked seasonal variation in death from ischemic heart disease and stroke in the elderly.10,11 Other authors suggest a correlation between the incidence of human diseases, including stroke, and the higher occurrence of geomagnetic storms during the colder period of the year.1 Analysis of meteorological data identifies cold ambient temperature as a risk factor for stroke.16,17 Other specialists described the central low pressure area with transition of the frontal system as the most unfavourable weather type for stroke occurrence.18 All these findings try to give a plausible explanation for the winter peak of stroke fluctuation. However, taking the different stroke subtypes in concern some other groups published discrepant findings, reporting the highest incidence of ischemic stroke and ICH during spring3 and autumn, respectively.4 Some authors seem to be doubtful about the existence of the reported winter excess of ischemic stroke,1 which could be an artifact caused by a referral bias in hospital-based studies and increased case fatality during winter in mortality studies.19,20 Regarding the incidence of subarachnoid hemorrhage, most studies report a higher case frequency during winter months.21,22 Our results suggest a seasonal variation in general stroke incidence with a significant decrease during summer and are in accordance with most published similar studies. Analysis of the variation in occurrence of the
K. SPENGOS ET AL.
96
different stroke subtypes leads to more differentiated conclusions. The predominance of ICH in winter is highly significant and in accordance with the results of various studies from different countries, which assume a triggering function of low ambient temperature.16,17 We also revealed a higher frequency of CE during the colder months of the year, while this kind of ischemic stroke was significantly less frequent during the summer. The already mentioned varying levels of coagulation parameters may result in a higher risk of formation of embolic material. Further studies about the seasonal variation in occurrence of acute atrial fibrillation and its possible relation with stroke incidence23 are needed. The subgroups of stroke caused by LAV, LAC, and OTH showed a distribution pattern that did not differ from random and were found not to follow a certain periodic cycle. In conclusion, our results are in accordance with the widely accepted model of seasonal variation of stroke incidence in general, with a peak in winter and a trough in summer. Further differentiation of stroke in its etiologic subtypes leads to non-uniform findings. ICH and CE follow the above described periodicity. In contrast, stroke caused by LVA, LAC, and OTH do not show a certain seasonal variation. In our opinion, these findings may partly explain the previously released discrepant data. Another interesting finding of our study is the large number of CEs that are already observed among Greek stroke patients and might be explained by the fact that Greek primary physicians still undertreat patients who should receive anticoagulants, especially those with nonvalvular atrial fibrillation.15 A major limitation of our study is the fact that it is hospital-based, which makes referral bias, especially into a tertiary care center, unavoidable. The summer vacation of the Athenians is another limitating factor, the importance of which is difficult to estimate. However, our hospital is open 365 day a year, which means that there are no possible biases because of a lack of medical personnel and capacities. Further investigations of the seasonal occurrence of the different stroke subtypes in correlation with possible pathogenic parameters are necessary in order to clarify the issue and may lead to important conclusions which could further optimize stroke prevention and therapy.
References 1. Rothwell PM, Wroe SJ, Slattery J, et al. Is stroke incidence related to season or temperature? The Oxfordshire Community Stroke Project. Lancet 1996;347:934-936. 2. Tsementzis SA, Kennet RP, Hitchcock ER, et al. Seasonal variation of cerebrovascular diseases. Acta Neurochir (Wien) 1991;111:80-83.
3. Oberg AL, Ferguson JA, McIntyre LM, et al. Incidence of stroke and season of the year: evidence of an association. Am J Epidemiol 2000;152:558-564. 4. Ricci S, Celani MG, Vitali R, et al. Diurnal and seasonal variations in the occurrence of stroke: a communitybased study. Neuroepidemiology 1992;11:59-64. 5. Jakovljevic D, Salomaa V, Sivenius J, et al. Seasonal variation in the occurrence of stroke in a Finnish adult population. The FINMONICA Stroke Register. Finnish Monitoring Trends and Determinants in Cardiovascular Disease. Stroke 1996;27:1774-1779. 6. Kochanowicz J, Kulakowska A, Drozdowski W. Seasonal variations in stroke incidence in North-Eastern Poland. Neurol Neurochir Pol 1999;33:1005-1013. 7. Shinkawa A, Ueda K, Hasuo Y, et al. Seasonal variation in stroke incidence in Hisayama, Japan. Stroke 1990;21: 1262-1267. 8. Luong TH, Rand JH, Wu XX, et al. Seasonal distribution of antiphospholipid antibodies. Stroke 2001;32:17071711. 9. McLaren M, Kirk G, Bolton-Smith C, et al. Seasonal variation in plasma levels of endothelin-1 and nitric oxide. Int Angiol 2000;19:351-353. 10. Stout RW, Crawford V. Seasonal variations in fibrinogen concentrations among elderly people. Lancet 1991;338:913. 11. Woodhouse PR, Khaw KT, Plummer M, et al. Seasonal variations of plasma fibrinogen and factor VII activity in the elderly: winter infections and death from cardiovascular disease. Lancet 1994;343:435-439. 12. Kuleshova VP, Pulinets SA, Sazanova EA, et al. Biotropic effects of geomagnetic storms and their seasonal variations. Biofizika 2001;46:930-934. 13. Gifford DK. Monthly incidence of stroke in rural Kansas. Kans Nurse 1996;71:3-4. 14. Biller J, Jones MP, Bruno A, et al. Seasonal variation of stroke— does it exist? Neuroepidemiology 1988;7:89-98. 15. Vemmos KN, Takis CE, Georgilis K, et al. The Athens stroke registry: results of a five-year hospital-based study. Cerebrovasc Dis 2000;10:133-141. 16. Azevedo E, Ribeiro JA, Lopes F, et al. Cold: a risk factor for stroke? J Neurol 1995;242:217-221. 17. Feigin VL, Nikitin YP, Bots ML, et al. A populationbased study of the associations of stroke occurrence with weather parameters in Siberia, Russia (1982-92). Eur J Neurol 2000;7:171-178. 18. Cabajova Z, Snopkova Z, Traubner P, et al. Effect of weather on the occurrence of cerebral stroke. Bratisl Lek Listy 1999;100:267-270. 19. Sheth T, Nair C, Muller J, et al. Increased winter mortality from acute myocardial infarction and stroke: the effect of age. J Am Coll Cardiol 1999;33:1916-1919. 20. Lanska DJ, Hoffmann RG. Seasonal variation in stroke mortality rates. Neurology 1999;52:984-990. 21. Nyquist PA, Brown Jr RD, Wiebers DO, et al. Circadian and seasonal occurrence of subarachnoid and intracerebral hemorrhage. Neurology 2001;56:190-193. 22. Passero S, Reale F, Ciacci G, et al. Differing temporal patterns of onset in subgroups of patients with intracerebral hemorrhage. Stroke 2000;31:1538-1544. 23. Kupari M, Koskinen P. Seasonal variation in occurrence of acute atrial fibrillation and relation to air temperature and sale of alcohol. Am J Cardiol 1990;66:1519-1520.