|
medpundit |
||
 |
 |
|
Friday, August 30, 2002"Surprisingly, this hypothesis doesn't seem to have gone through (rigorous scrutiny) very much. It seems to have been adopted very widely, very rapidly without too much in the way of critical appraisal," said Oxford University scientist Rory Collins, one of the lead investigators in the new analysis Not so surprising. The studies that proport to show an association between fetal weight and adult blood pressure were based on meta-analysis which means they took a bunch of findings and studies and mixed the results in a statistical stew. Give us numbers and we’ll think them true, regardless of how they were obtained. Combining studies in this way has its risks: One problem is that smaller studies are more likely to find more extreme results and positive results are more likely to be published, which means published evidence from the small studies might overestimate the true effect, Collins said. Therefore, more importance was given to the larger studies by the new analysis. The new analysis, as compared with the old analysis finds less of a blood pressure difference with weight. However, the surprising thing is that both analyses found only very small differences in blood pressure among the adults studied: differences of only one to four points. That difference is well within the margin of error for measuring a blood pressure: Proponents of the fetal hypothesis have estimated that a 1 kilogram (2.2 pound) higher birth weight is typically associated with a 2-4 point lower blood pressure. After reweighting the studies to eliminate any bias introduced by only publishing only the small studies with strong results, the new analysis concluded that 1 kilogram (2.2 pound) extra birth weight was linked to about a half point lower blood pressure. The studies also adjusted their statistics to take into account the current weight of the patients when they had their blood pressure measured as adults. That obscures the real relationship between birth weight and blood pressure, Collins said. "When we removed the adjustment for current weight, the relationship appeared to disappear," Collins said. A two pound difference in weight is a lot of weight for a baby, but a four point difference in blood pressure between adults isn’t a lot of blood pressure. It’s not surprising that the re-analysis found the evidence wanting: "It is a serious concern (for the theory) when something that is considered by the proponents to be their best evidence appears to have such serious problems," he said. "That's certainly not to say there's nothing in it, but it may mean that the strength of any association is much less than had been claimed, if at all." posted by Sydney on 8/30/2002 06:09:00 AM 0 comments 0 Comments: |
|
