Characteristics of studies
| Total (deep/lobar) | ||||||||
|---|---|---|---|---|---|---|---|---|
| Study* | Year of publication | Study population | Methodological quality criteria fulfilled† | Exclusions due to secondary causes | % of patient excluded | Number of patients in analyses | Mean or median age | Time to scan |
| AVM, arteriovenous malformation; CVT, cerebral venous thrombosis; NR, not reported; SLE, systemic lupus erythematosus. | ||||||||
| *Studies are ordered first according to whether they included mixed-age or younger populations, and then by year of publication. | ||||||||
| †; BP-pre, definition of hypertension based on pre-stroke blood pressure (not post-stroke or combination of pre-stroke and post-stroke blood pressure); FES, first-ever stroke only; P, population based. | ||||||||
| ‡Percentage of all deep and lobar supratentorial haemorrhages excluded because of secondary causes (except for Rome, Cincinnati (b) and USA, where the percentage excluded is calculated with all haemorrhages, including those in the posterior fossa, as the denominator). | ||||||||
| §A total of 3930 patients with haemorrhage were included in the analyses, 2196 of whom had a deep haemorrhage. | ||||||||
| ¶Mean or median (indicated by **) age of total population with intracerebral haemorrhage (including those with posterior fossa haemorrhage). In cases where age was reported as the number of patients within various age bands, mean age was derived by assigning the middle value of each age band to the number of patients included. | ||||||||
| Studies of populations unselected for age | ||||||||
| New Orleans, USA48 | 1979 | Consecutive admissions to the neurology department | — | AVM, blood dyscrasia, coagulopathy or on anticoagulants | 9% (3%/41%) | 242 (222/20) | NR | Within 8 days |
| Heidelberg, Germany34 | 1982 | Non-consecutive admissions to hospital | — | AVM, CVT, or on anticoagulants | 19% (2%/37%) | 71 (44/27) | 58 (NR) | NR |
| Copenhagen, Denmark33 | 1984 | Non-consecutive admissions to neurology and neurosurgery departments | — | AVM, on anticoagulants, alcohol misuse | 25% (0%/41%) | 36 (19/17) | 54** (NR) | Median 2 days |
| Cincinnati (a), USA27 | 1986 | Retrospective review of admissions to 16 general hospitals | BP-pre | None | 0% | 124 (51/73) | NR | NR |
| New York, USA38 | 1987 | Consecutive admissions to one hospital in the Bronx area of the city | BP-pre | Blood dyscrasia, vasculitis or AVM | 18% (11%/29%) | 92 (62/30) | NR | Within 24 h of admission |
| Rome, Italy29 | 1988 | Admissions to one city-centre hospital | BP-pre | AVM or on anticoagulants | 6% (NR) | 87 (56/31) | 62 (NR) | Mean 1.6 days |
| Florence, Italy36 | 1990 | Non-consecutive patients identified from the neuroradiology service | BP-pre | Blood dyscrasia, on anticoagulants, or AVM | 31% (24%/48%) | 70 (54/16) | 63 (NR) | NR |
| Giessen, Germany44 | 1990 | Admissions to neurology department of one hospital | — | AVM, haemorrhagic diathesis, or on warfarin | 13% (8%/15%) | 79 (57/22) | 66** (NR) | NR |
| Linkoping, Sweden41 | 1991 | Consecutive admissions to neurology department | — | None | 0% | 182 (102/80) | 65 (NR) | NR |
| Riyadh, Saudi Arabia50 | 1991 | Admissions to hospital | FES, BP-pre | NR | NR | 13 (10/3) | 49 (51/43) | NR |
| USA39 | 1991 | Admissions to hospitals (multicentre involving 4 different cities) | — | AVM, coagulopathy, ventricular haemorrhage, multiple haemorrhages or on anticoagulants | 17% (NR) | 172 (107/65) | 62** (59/68) | Mean 1 day |
| Cincinnati (b), USA26 | 1993 | Review of medical records from 20 acute-care hospitals and 5 coroner’s offices | P, BP-pre | Haemorrhagic infarction, AVM, anticoagulants, thrombolytic treatment, cocaine use | 11% (NR) | 143 (77/66) | NR | NR |
| Oxford, UK25 | 1993 | Community-based (overlapping sources used to identify cases occurring in a defined area) | P, FES, BP-pre | None | 0% | 42 (18/24) | 71 (67/72) | NR |
| Durham, USA30 | 1994 | Consecutive admissions to one hospital | BP-pre | Thrombocytopenia, inherited coagulopathy, or AVM | NR | 45 (29/16) | 61 (56/67) | NR |
| Essen, Germany24 | 1994 | Admissions to hospital | — | None | 0% | 300 (46/254) | NR | Within 24 h of admission |
| Perth, Australia23 | 1994 | Community-based (overlapping sources used to identify cases occurring in a defined area) | P, BP-pre | None | 0% | 37 (18/19) | 68 (NR) | Median 4 days |
| Massachusetts, USA32 | 1996 | Consecutive patients aged >50 years with lobar haemorrhage, and with non-lobar haemorrhage. Unclear if both groups were recruited from same place and during the same time period | — | AVM, vasculitis or coagulopathy | NR | 63 (18/45) | 73 (69/75) | NR |
| Cologne, Germany42 | 1997 | Retrospective review of admissions to two hospitals | — | Haemorrhagic infarcts, AVM, cavernoma, coagulation disorders, on thrombolytic or anticoagulation therapy | 23% (10%/33%) | 575 (278/297) | 57 (NR) | NR |
| Victoria, Australia46 | 1998 | Consecutive admissions to all hospitals serving a defined population and regular inspection of coroner’s reports | P, FES, BP-pre | AVM, haemorrhagic transformation, bleeding diathesis or drug misuse | NR | 264 (122/142) | 64 (NR) | NR |
| Besançon, France45 | 2000 | Consecutive admissions to neurology, neurosurgery or intensive care units of the only hospital in the county to which patients with neurological diseases are referred | P, FES, BP-pre | Haemorrhagic infarction, AVM, cavernoma or on thrombolytic treatment, | NR | 295 (167/128) | 67 (NR) | Mean 1 day |
| Sweden40 | 2000 | Community-based (12 hospitals and four pathology departments serving a defined population) | P | AVM, haemorrhagic infarction | NR | 297 (121/176) | 74** (72/75) | 1–2 days |
| Cincinnati (c), USA49 | 2002 | Non consecutive patients, identified by surveillance of emergency and radiology departments, and hospital discharge diagnoses | BP-pre | Haemorrhagic infarction, AVM or cavernoma | NR | 188 (121/67) | 65 (65/65) | NR |
| Izumo, Japan35 | 2003 | Admissions to the four hospitals in the city, and review of general practitioner death certificates | P | AVM, moyamoya disease haemorrhagic infarction or coagulation disorder | NR | 274 (229/45) | 68 (68/71) | NR |
| Studies in younger patient populations | ||||||||
| Iowa, USA47 | 1987 | Patients aged 15–45 years admitted to hospital | — | AVM, haemorrhage as a result of drug or alcohol misuse, SLE, moyamoya, cryoglobulinaemia, or preeclampsia | 58% (41%/67%) | 22 (10/12) | 31 (NR) | NR |
| Dijon, France31 | 1991 | Patients aged <45 years admitted to neurosurgery, neurology and rehabilitation departments of the city’s university hospital | — | AVM, cerebral vein thrombosis, SLE, endocarditis, leukaemia or on anticoagulants | 59% (27%/76%) | 12 (8/5) | 33 (NR) | NR |
| Mexico City, Mexico43 | 1991 | Consecutive admission of patients aged <40 years to stroke unit | — | AVM, cavernous angioma, CVT, drug use, toxaemia or other known causes | 75% (49%/85%) | 38 (22/16) | 27 (NR) | NR |
| Tainan Taiwan (a)37 | 1997 | Patients aged 14–40 years admitted to hospital | — | AVM, drug misuse, blood dyscrasia, alcohol misuse, SLE, moyamoya or infective endocarditis | 35% (24%/50%) | 40 (26/14) | 34 (NR) | NR |
| Kaohsiung, Taiwan (b)28 | 1999 | Patients aged 15–44 years admitted to hospital | — | AVM, blood dyscrasia and “other rare causes” (including alcohol and drug misuse, uraemia, etc). We could not exclude 4 patients with tumours. | 33% (12%/53%) | 126 (102/24) | 36 (NR) | NR |