Tobacco Documents Online

Rt:u e-,as 2- # 3 MS# 96-274 Title: Increased serum mitogenic activity for arterial smooth muscle cells associated with relaxation and low educational level in human subjects with high but not low hostility traits: Implications for atherogenesis Authors: Gutstein, Teresi. Wu, Salimian, Cui, Paul, & Jabr This study examined whether platelet-derived growth factor (PDGF) or PDGF- independent mitogenic activity can be altered by the interventions of "frustration," "harassment," or "relaxation" in a group of 225 adults. In addition the subjects had completed the Cook-Medley Hostility Inventory to determine if there was an interaction between the intervention and hostility scores. Though this is a potentially very interesting study, the statistical analyses and the r =lts section are unclear and confusing. I was at such a loss with the statistical analyses that I asked two of my colleagues for their optnton (I include the written comments I received from one of my colleagues). Both concurred with my opinion that the results and statistical sections are unnecessarily complex and may be a fishing expedition because no significant effects were found with a more straig t orwazd analysis. The preferred analyses would be a repeated measures ANOVA of the pre- and post- PDGF and thymidine values using the independent values of intervention group (frustration, anger and control) and high/low hostility. Furthermore, there may be interactions with gender that were not examined. fiX The results on p. 13 are confusing. Was there a significant main effect for the intervention group? The text did not specify what main effect was found. The values for the pre- and post measures should be provided in a table. Furthermore, F values need to be given for the main effect of ? and the nonsignificant trend for hostility. Using 17 as a median split for hostility is somewhat low and results may be found if quartiles are used. What was the mean and standard deviation for the Cooke-Medley hostility inventory? My colleague questions the ecological validity of the experimental manipulation used to induce anger and frustration. This would not be questioned if results were provided for blood pressure and heart rate changes found during the intervention in comparison to the relaxed state and similarly, if mood analyses of the anger and irritation found during the frustration and harassment conditions were provided. I am not clear about the procedures. After inserting the IV catheter, subjects rested for 15 minutes and towards the end of this rest, blood pressure, heart rate and blood samples were obtained. Subjects then underwent the intervention which lasted 30 minutes? If so, subjects will quickly habituate to the intervention and the intervention was too long. Why was 30 minutes selected - is 30 minutes needed to find changes in PDGF and thymidine measures? Were blood pressures taken at the end of the intervention or throughout the intervention?

I am afraid this study has all the hallmarks of a negative result in regard to the original ~hypothesis, leading to the authors launching a statistical fishing expedition to find some positive '~ti esult that would allow the study to be published. Unfortunately, their final conclusion is simply not convincing. The original hypothesis (which can be inferred from the design of the experiment) was that humans may show increased mitogenic activity in the face of increased anger, and possibly frustration. To test this idea, three groups of subjects were subjected to situations which the authors claimed to be frustration inducing, anger inducing and relaxing (control), respectively. A number of dependent measures were taken, some presumably to check whether the experimental manipulation worked (blood pressure and pulse rate) and others as indices of mitogenic activity. The approach to testing the hypothesis is simple: two way ANOVAs on the mitogenic activity measures (or better still a MANOVA) and the hypothesis would be supported if an interaction emerged between group and pre/post, with one or both of the experiments groups showing an increase in mitogenic activity relative to the control group. I am of the opinion that this did not occur, and the reader cannot discover this easily because the means are not provided for the pre/post measures, rather difference scores are given in Table 1, where the data is broken down by hostility level (although this variable was not part of the original design and appears to have been introduced as an afterthought, since it is not mentioned in the Introduction). My reason for saying that the original hypothesis was not supported is that numerical results are not provided for the ANOVA performed on the data in Table 1. Indeed, the description of the analyses on p.13 is very confusing. Unfortunately, things go downhill from there. It appears the authors tried all kinds of multivariate approaches (which in my Qpinion aze described unintel~ibly) until a relationship emerged. The relationship is that high hostility relaxed individuals show increased mitogenic activity. They'sta e that this finding appears counterintuitive (p.17, 2nd para), but then roceed to uch of their discussion trying to jusu it. n m opinion, the results of this study are best ascribed to a Type 1 error ansm e-ef-statistics. My recommendation is that the authors treat the present study as a pilot, and perform a second study specifically designed to test their surprising hypothesis before attempting to publish. I am not at all familiar with this area, but I have to wonder about the ecological validity of the experimental manipulations used to induce anger and frustration. Perhaps this is the reason that the original hypothesis was not supported.