Rapid Responses to:

PAPERS: Pernille Poulsen, Allan Vaag, and Henning Beck-Nielsen Does zygosity influence the metabolic profile of twins? A population based cross sectional study BMJ 1999; 319: 151-154 [Abstract] [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Authors' doubts unsubstantiated by their data

Lesley V Campbell   (5 August 1999)

Why not analysis the difference within zygosity?

Clayman ZK Zhang   (27 April 2002)

Authors' doubts unsubstantiated by their data 5 August 1999
Lesley V Campbell Send response to journal: Re: Authors' doubts unsubstantiated by their data	We have concerns regarding a recent study of the influence of zygosity on the Insulin Resistance Syndrome (IRS) (1). The underlying well-known hypothesis is that foetal undernutrition, usually measured as low birth-weight, results in foetal changes detectable in features of IRS in later life (2). The authors studied the incidence of diabetes and impaired glucose tolerance (IGT) in monozygotic (MZ) and dizygotic (DZ) twins, postulating "undernutrition" in 2/3 of the MZ twins from placental sharing. Diabetes and IGT were NOT more frequent in the MZ twins, not supporting an effect of zygosity on IRS. IRS-related measures did not differ: BMI,waist-to-hip ratio, fasting glucose or insulin (a surrogate for insulin resistance). The only difference was in a higher 30 minute glucose tolerance test (GTT) insulin in MZ twins. On excluding those with glucose intolerance or diabetic family history, the MZ group had a higher 30 minute glucose. Such differences have a questionable significance. Without more accurate measures (hyperinsulinaemic clamp, minimal modeling, homeostasis models),the similar fasting insulins do not support differences in insulin resistance. The lack of gold standard measures allows no conclusions about visceral fat. If intrauterine malnourishment is implicated, available birthweight data should be supplied, not assumed. Although perinatally MZ twins may suffer malnutrition, there are no data to prove that this is also true in a natural selection of MZ twins surviving over 60 years. We question the statistical methodology, including the lack of adjustment for multiple comparisons and for the non-independence of within twin pair observations. The latter reduces cohort variance, resulting in underestimation of standard errors and, consequently, overstatement of P values. Use of maximum likelihood methods or generalised estimating equations, standard in genetic epidemiology, would adjust for this. If intrauterine deprivation of the extent postulated in monochorionic twinning produces IRS, surely,after age 65 yr, there should be greater phenotypic manifestation? Zygosity may influence the 30 minute GTT glucose/insulin level in this study, but it does not, by age 65yr affect the incidence of obesity, diabetes or IGT or influence the measured phenotypic features of IRS. The authors' doubts regarding the validity of twin studies are thus unsubstantiated by their data. As there is evidence for maternal and foetal gene effects causing an association between IRS and low birthweight (3) and a possibility that undernutrition selects "thrifty" genes, more careful studies are essential to tease out the specific influence of intrauterine nutrition on later disease. Lesley V Campbell Katherine Samaras Cornelis B Lambalk Nick Martin
Why not analysis the difference within zygosity? 27 April 2002
Clayman ZK Zhang, Research Staff MRC Environmental Epidemiology Unit,Southampton,SO16 6YD Send response to journal: Re: Why not analysis the difference within zygosity?	Sir- Dr. Poulsen's report(BMJ 1999;319:151-154)tried to explain that dygosity is related to glucose homeostasis.The glucose profiles have been compared in detail between monozygotic and dizygotic twins.This will give a wrong information and waste the data.I think the best way to surport the researcher's points is to compare the glucose profiles within the same type of zygosities.The results from this method can be used to seperate the diffenrent influence from genetic and fetal grow environment.And additonally,a single research can not be used to deny the long accepted belife:in fact,genetic play a major role in type 2 diabetes.