Research Article

Complications of pregnancy and delivery in relation to psychosis in adult life: data from the British perinatal mortality survey sample.

BMJ 1991; 302 doi: (Published 29 June 1991) Cite this as: BMJ 1991;302:1576
  1. D J Done,
  2. E C Johnstone,
  3. C D Frith,
  4. J Golding,
  5. P M Shepherd,
  6. T J Crow
  1. Division of Psychiatry, Clinical Research Centre, Harrow, Middlesex.


    OBJECTIVE--To evaluate whether events occurring at or around the time of birth contribute to the onset of psychotic illness in adult life. DESIGN-Pregnancy and birth complications as possible causes of adult mental illness were studied in the population sample of the British perinatal mortality survey. Subsequent psychiatric admissions were independently identified through the Mental Health Enquiry and records of regional and special health authorities. Logistic regression was used to compare data on perinatal deaths with those on survivors to determine factors independently associated with perinatal death, and this equation was then used to calculate the risk of perinatal death for each survivor. SUBJECTS--16,980 people born in a single week in 1958 (the British perinatal mortality survey sample), including 252 patients admitted to psychiatric care; case notes of 235 patients were supplied. MAIN OUTCOME MEASURES AND RESULTS--Patients with a schizophrenic illness (whether defined by "broad" (n = 57) or "narrow" (n = 35) diagnostic criteria) did not have a greater mean risk of perinatal death than the population in general, but there was some evidence of increased liability (relative risk 2.43; 95% confidence interval 1.17 to 5.05) for those with affective psychosis (n = 32). Specific high risk variables for affective psychosis were decreased gestation time (273.9 v 281.2 days; mean difference 7.3 days, 95% confidence interval 3.1 to 11.5; p less than 0.002) and prescription of vitamin K to the child in the first week of life (19% of patients v 5% of controls, p = 0.016). CONCLUSIONS--The findings give no support to theories that factors predicting perinatal mortality contribute significantly to causation of schizophrenic illness. Further investigation of decreased gestation length in relation to affective disorder is required.