Association between life stress and serious perinatal complications

Journal of Family Practice, Nov, 1989 by Harold A. Williamson, Jr., Michael LeFevre, Melvin Hector

Several psychosocial factors have been associated with the outcome of pregnancy. Associations with adverse outcomes have been found for anxiety,(1-5) life stress,(5-10) low social support,(8-12) and family dysfunction." Although there is generel consensus that these factors influence the complication rate in pregnancy in some manner, many specific issues remain unresolved.

Several studies have demonstrated an effect of stress or social support on pregnancy outcome, but have used such broadly inclusive definitions of pregnancy complicationsl, 5,8-10) that outcomes as diverse as threatened abortion and prolonged labor and as unimportant as nuchal cord have been used to define the complicated pregnancy groups. In these studies, rates of complicated pregnancy reach nearly 50%. Such criteria call into question the clinical significance of the findings and defy any attempt to explain the biological basis of the effect of stress or social support.

Other studies have been differently limited by methodologic constraints. Some research has relied on measures of prenatal stress taken at the time of hospitalization for delivery. 4,1 It seems likely that recall bias would affect responses to such retrospective reports. Other studies have systematically excluded patients with high biomedical risk, have enrolled convenience samples of women presenting to large tertiary care centers, or have been hampered by small sample sizes. Still others have measured life changes (stress), a dynamic variable, at only one point in pregnancy.

Independent of methodologic issues, several credible studies have failed to demonstrate a link between poor outcome and psychosocial risk. Studies that have controlled for biomedical risk by excluding complicated patients or by using multivariate analysis have shown no relationship between pregnancy complications and anxiety, personality characteristics, social support, attitudes, or life stress. 11-15

This current study prospectively measured the effect of stressful life change and social support on serious and important perinatal complications in a large group of pregnant, rural women and controlled for the potential confounding effect of biomedical risk. METHODS Subjects and Sites The study population was composed of pregnant women residing in four rural Missouri counties who came for prenatal care to one of four family practices involved in a research consortium." The four practices are located in towns of 3000 to 10,000 persons; two are family medicine residency training sites. During a 2-year period from 1984 to 1986, pregnant women between 18 and 22 weeks'gestation were asked by physicians or office staff to participate in the project. They were told that this study was to help determine factors that might influence the outcome of pregnancy. Presentation at a gestational age greater than 22 weeks was the only prospective exclusion criterion; a decision was made to exclude late registrants because they could not complete the second trimester questionnaire during the 18- to 22-week gestation period. Pregnancies that ended prematurely before the 32- to 36-week gestation visit also were excluded from this study because data would not be available from both the second and third trimester visits. instruments

The questionnaire used to gather data about life stress was a 39-item adaptation of the Social Readjustment Rating Scale of Holmes and Rahe." Slight changes in wording were made to account for pregnancy, and questions that were inappropriate for a pregnant subject were deleted. This scale was chosen because it has been used in numerous studies of stress and health effects. Each of the questionnaire items has a designated weight. For example, 100 points is assigned to the death of a spouse, 45 points for loss of a job, 20 points for a move, etc. Responses are summed to create a total life change score. The patients were asked to answer the questions to account for life changes occurring in the 12-month period preceding the visit.

Social supports were assessed as a summary score of six components of support derived from a 12-item questionnaire: personal networks, community networks, intimacy, appreciation-understanding, satisfaction-loneliness, and tangible assistance. The conceptual background for these measures of social support was articulated by Berkman and Symell and Henderson et al.(19) Social support, as measured by this technique, has been found by McKay et al (2O) to buffer the adverse effect of life stress on health.

The physician performing the delivery completed an outcome questionnaire. Standard perinatal outcome variables were used, and because adverse outcomes were relatively rare, a composite of perinatal complications was developed. A poor outcome was prospectively defined as the presence of at least one of the following: neonatal death, transfer of the infant to a neonatal intensive care unit, Apgar score of <7 at 5 minutes, or low birthweight (less than 2500 g). These outcomes were chosen because they are clinically important, easily measured, and definite, and because they can also be predicted by the biomedical risk index used in this study, thus allowing a procedure for controlling for biomedical risk.