Rapid Responses to:

PAPERS: David Gaist, Michael Væth, Ioannis Tsiropoulos, Kaare Christensen, Elisabeth Corder, Jørn Olsen, and Henrik Toft Sørensen Risk of subarachnoid haemorrhage in first degree relatives of patients with subarachnoid haemorrhage: follow up study based on national registries in Denmark BMJ 2000; 320: 141-145 [Abstract] [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

incidence of hospital admission does not equal incidence of disease

Louis H Pobereskin   (26 January 2000)

Re: incidence of hospital admission does not equal incidence of disease

David Gaist   (28 January 2000)

incidence of hospital admission does not equal incidence of disease 26 January 2000
Louis H Pobereskin, Consultant Neurosurgeon Derriford Hospital, Plymouth Send response to journal: Re: incidence of hospital admission does not equal incidence of disease	Editor - We read with interest the report by Gaist and Colleagues on the incidence of subarachnoid haemorrhage in first degree relatives using data from the National Registries in Denmark1. We are, however, concerned about their methods and feel the conclusions which they draw from their data are incorrect. Validation of the diagnosis was only performed in a sample from one county. Is Funen County representative of Denmark and how was it selected? Hospitals with ten or less registered patients in the study period were excluded. Why is it appropriate to exclude the smaller hospitals when they may be a source of patients with particularly low predictive value for a registered diagnosis of subarachnoid haemorrhage? The cohort of first degree relatives was overwhelmingly made up of children and they were the only group in Table 2 for whom the incidence rate ratios were significantly different from 1. The main problem with this study is one that plagues much of the literature on subarachnoid haemorrhage, it was hospital based. The strongest predictor of survival in subarachnoid haemorrhage is age. This is just as true when one considers who will survive to reach medical attention. In a recent population based review of 824 cases of subarachnoid haemorrhage in Devon and Cornwall, the proportion of patients dying outside hospital was 7% and 17% for those under and over 45 years of age respectively. It is therefore not surprising that, as has been shown in this study, a group aged 33 is more likely to be admitted to hospital following a subarachnoid haemorrhage than a group aged 53. In addition, these first degree relatives were in the main patients for whom one of their parents had experienced subarachnoid haemorrhage within the past few years. They and their own families would therefore have a high level of awareness of this condition and its serious consequences and be more likely than the general population to refer themselves for investigation and to press for referral to a specialist centre. This study suggests that first degree relatives are more likely than the general population to be admitted to hospital following a subarachnoid haemorrhage. Considering the age and awareness differences between index cases and relatives, however, it would seem to go beyond the data to conclude from this that they are more likely to suffer a subarachnoid haemorrhage. Louis H Pobereskin Consultant Neurosurgeon J Robert Sneyd Consultant Anaesthetist Derriford Hospital, Plymouth, PL6 8DH 1. Gaist D, Vaeth M, Tsiropoulos I, Christensen K, Corder E, Olsen J, Sorensen H T Risk of subarachnoid haemorrhage in first degree relatives of patients with subarachnoid haemorrhage: Follow-up study based on National Registries in Denmark BMJ 2000; 320: 141-5.
Re: incidence of hospital admission does not equal incidence of disease 28 January 2000
David Gaist, postdoctoral fellow Epidemiology, Institute of Public Health, Odense University, University of Southern Denmark Send response to journal: Re: Re: incidence of hospital admission does not equal incidence of disease	Editor, Pobereskin and Sneyd [1] have raised three issues regarding our paper on the risk of subarachnoidal haemorrhage (SAH) in first degree relatives of patients with this condition [2]. They are concerned about the choice of a single county for validation of the register diagnosis and query our exclusion of cases admitted to hospitals with less than 10 cases from the validation study. Secondly, they point out that young age per se can have a bearing on the chances of being admitted to hospital for SAH. Finally, the authors are concerned with the issue of awareness, i.e. that the threshold for contacting a physician might be lower in patients that have a close relative who suffered from SAH and are familiar with the symptoms. We validated the register diagnosis of SAH in two groups of patients. A random sample of patients discharged from hospitals in the county of Funen (N=210), and all identified familial cases (N=37) from hospitals all over Denmark. The random sample was restricted to the county of Funen for purely logistical reasons, i.e. ease of access to the medical records. The county is geographically well delimited and the inhabitants (450,000) have been found to be a representative 10% sample of the Danish population [3,4]. We found that the incidence rate of SAH in the county of Funen was highly comparable with that of the entire country. The county of Funen does not differ in any major way from other counties in Denmark with regard to access to health care and referral of SAH patients. The county is served by a number of smaller hospitals and a single University hospital (1400 beds), the only one in the county with a neurology and a neurosurgery unit. When sampling the cases for validation we found it practical to exclude patients from two smaller hospitals in Funen, who had discharged fewer than 10 patients each during the entire 18-year follow- up. This resulted in the exclusion of 15 patients or 0.2% of all SAH patients identified in Funen county during the period. We do not believe that this exclusion had any methodological consequences whatsoever. Pobereskin and Sneyd have found that younger patients with SAH in Devon and Cornwall are more likely to be admitted to hospital [1], and are concerned whether this is a problem in our cohort of first degree relatives comprising mainly young adults. Even if such an age effect were present in our material, however, it would not have influenced our results since we controlled for age: The incidence of SAH in the cohort of relatives was compared with the incidence of SAH in the general population in Danes from the same age-group. The final issue raised by Pobereskin and Sneyd is that of awareness. First degree relatives of patients with SAH may have a lower threshold for seeking help, since they are familiar with the symptoms, a potential limitation, which we also pointed out in our paper. However, according to an evaluation of the medical records undertaken by an experienced neurologist who was blinded with regard to family relationships, familial cases did not differ in any significant way from non-familial cases concerning severity or presenting symptoms. In conclusion, we do not find that the issues raised by Pobereskin and Sneyd invalidate our findings. David Gaist, MD, postdoctoral fellow Kaare Christensen, MD, PhD, research professor Epidemiology, Institute of Public Health, Odense University, University of Southern Denmark Michael Væth, professor Department of Biostatistics, University of Aarhus Ioannis Tsiropoulos, Consultant Neurologist Department of Neurology, Odense University Hospital Jørn Olsen, professor Danish Epidemiology Science Centre, University of Aarhus Henrik Toft Sørensen, associate professor Department of Clinical Epidemiology of Aarhus and Aalborg University Hospitals at the Institute of Epidemiology and Social Medicine, University of Aarhus 1. Pobereskin LH, Sneyd JR. Incidence of hospital admission does not equal incidence of disease. BMJ 2000; 320 electronic response. 2. Gaist D, Væth M, Tsiropoulos I, Corder B, Christensen K, Olsen J, Sørensen HT. Risk of subarachnoidal haemorrhage in first degree relatives of patients with subarachnoid haemorrhage: follow-up study based on national registries in Denmark. BMJ 2000; 320: 141-5. 3. Green A. The county of Funen 1972-74. A comparative demographic study of the county of Funen and Denmark as a whole with regard to use of data from Funen county for epidemiological studies [In Danish]. Report. Department of Genetics, Odense University, 1978. 4. Gaist D. Use and overuse of sumatriptan. Pharmacoepidemiological studies based on prescription register and interview data. Cephalalgia 1999; 19: 735-61.