Methods

We searched the MEDLINE database for stroke genetics studies published in English between January 2000 and January 2002, using the medical subject heading term “cerebrovascular disorders, genetics of” plus the text words “ischaemic” and “stroke.”² We included only original clinical trials and observational studies of human genetic risk factors for ischaemic stroke.

Both authors independently reviewed every article. We used standardised forms to record whether an explicit statement described informed consent and institutional review board (IRB) procedures. We also recorded whether the consent was given orally or in writing, whether it was witnessed, and whether family members and caregivers were involved in the process. In addition, we recorded whether the ethics committee or IRB was identified (named or identified as “the local ethics committee”). We also recorded whether articles referred to previous publications of the same study, the country where the study was done, and the journal in which it was published. We settled any disagreements by consensus. Because this study did not involve human participants or medical records, institutional review board approval was neither required nor obtained.

Results

In 24 separate journals, 41 articles met search criteria.² The number of articles per journal was one for 17 journals, two for four journals, three for two journals, and 10 for one journal.

Informed consent was reported more frequently than was protocol approval by an IRB or ethics committee (table 1). About half reported both obtaining informed consent from the patient and receiving protocol approval from an ethics committee, and 17% did not mention any measures taken to protect patients.

Details regarding the mechanism by which informed consent was obtained were rarely reported (table 2). It was also rare for articles to report whether the use of familial or non-familial surrogate consent was permitted if the subject was not competent to give informed consent. The IRB or ethics committee was identified by name in 24 (92%) of the 26 articles that reported receiving approval.

For 40 articles, no distinction was reported between ethical protection measures for subjects who had a history of stroke and stroke free subjects. One article stated that both ethical protection measures were in place for stroke free controls but did not report whether the same protections were in place for the stroke affected study subjects.³ One investigative group studying stroke genetics in children published two articles in the same year and in the same journal. One article reported the use of familial consent,⁴ and one article did not.⁵

Discussion

After the introduction of the Uniform Requirements for Manuscripts Submitted to Biomedical Journals,⁶ the number of articles reporting use of informed consent in clinical trials increased from 64% to 82%, and the number reporting ethics committee approval increased from 59% to 82%.⁷ The percentage of articles reporting these ethical protections for stroke genetics studies published in 2000 or later appears to be similar to that for clinical trials published before 1997. It is possible that in some studies ethical protections were in place but were not reported, and it is not known whether selection factors influence such reporting.

Only rarely did articles report using consent from a legally authorised representative if the prospective study subject was not competent to give informed consent. Knoppers and colleagues⁸ point out that it is difficult to establish a blanket set of ethical guidelines with respect to the participation of incompetent adults in genetics research. It may not be possible to know whether participation in a genetics study would conform to previous personal values or preferences of a person rendered incompetent by acute or chronic stroke.

The debate over competence of stroke patients to participate in clinical research has focused on therapeutic trials, because truly emergent consent is rarely required in stroke genetics research.⁹ Patients who are enrolled in such trials are on average more severely affected by stroke than population based stroke cohorts. The recovery of decisional capacity of patients with stroke may, in some instances, parallel recovery of neurological deficit. Patient genotype may, however, influence survival after stroke¹⁰ and response to acute therapies.¹¹ Thus, to avoid survival bias it may be important for stroke genetics protocols to have as an option the enrolment of subjects using the combination of personal assent and consent by a legally authorised representative.