Ionescu D A, Marcu I, Bicescu E
Lancet. 1976 Feb 7;1(7954):278-80. doi: 10.1016/s0140-6736(76)91405-7.
The ABO blood-group distributions of 1460 patients who had died from a stroke were compared with those of a control group of 20 705 controls selected at random from the healthy population at risk (i.e., over thirty-five years of age and matched for age and sex ratio). The cause of death was certified as cerebral thrombosis in 329 cases and as cerebral haemorrhage in 482 cases, these diagnoses being established in neurological hospitals; the remaining 649 cases had an unspecified type of stroke, the diagnosis being made by general practitioners. In the group with unspecified type of stroke the blood-group distribution was practically the same as the distribution in the controls. In the thrombosis cases there was an excess of blood-groups A and AB and a deficiency of O and B; in cerebral haemorrhage this situation was reversed. However, these were only trends; the differences were not significant at the 5% level. A statistically significant difference did emerge when the A+AB excess in thrombosis was contrasted with the O+B excess in haemorrhage, suggesting that this difference might be accounted for the major A subgroup (A1) and, consequently, A1B.
将1460例死于中风的患者的ABO血型分布与从有风险的健康人群（即35岁以上且年龄和性别比例匹配）中随机选取的20705例对照组的血型分布进行了比较。在神经科医院确诊329例死因是脑血栓形成，482例是脑出血；其余649例中风类型未明确，诊断由全科医生做出。在中风类型未明确的组中，血型分布与对照组的分布基本相同。在脑血栓形成病例中，A和AB血型过多，O和B血型不足；在脑出血病例中，情况则相反。然而，这些只是趋势；差异在5%水平上不显著。当将脑血栓形成中A+AB过多与脑出血中O+B过多进行对比时，确实出现了具有统计学意义的差异，这表明这种差异可能是由主要的A亚组（A1）以及因此的A1B造成的。
