Differences in the Distribution of IGF-I Concentrations Between European and US Populations.

Context: Method-specific reference intervals (RIs) determine utility of IGF-I as a biomarker in GH-related diseases. Differences between populations might affect applicability of RIs. Objective: To compare population-specific RIs derived from IGF-I routine testing in laboratories in the United States and Europe using the same assay. Design and setting: Uncensored routine IGF-I testing results generated over 5 years in 4 accredited laboratories (US, n = 778 173 males/710 752 females; Europe, n = 23 220 males/40 183 females). Main outcome measures: Construction of RIs by indirect statistical methods designed to use routine testing data (modified Hoffmann approach). Comparison to published RIs, between the US and Europe, and between regions in the United States with lower and higher mean body mass indexes (BMIs). Results: Lower limits (LLs) of RIs calculated from all routine data sets do not differ from the published LLs. The same is true for upper limits (ULs) calculated from European routine data. ULs derived from US routine data are significantly higher (children, 10-18 years [mean, %]: boys + 149.3 ng/mL [+34.6%]; girls + 94.9 ng/mL [+19.8%]); adults (19-95 years: males + 45 ng/mL [+20.3%]; and females + 29.7 ng/mL [+13.8%]). Average IGF-I is higher in samples from Colorado (lower mean BMI) compared with Alabama (P < 0.0001), although the difference is smaller than between each of them and Europe. Conclusions: We provide evidence that in large datasets from the same population, direct sampling and the indirect Hoffmann approach provide comparable RIs. Although LLs are comparable between Europe and the United States, the UL is significantly higher in the United States. We suggest use of adapted RIs for the United States.

Epidemiology of the human circadian clock.

Humans show large inter-individual differences in organising their behaviour within the 24-h day-this is most obvious in their preferred timing of sleep and wakefulness. Sleep and wake times show a near-Gaussian distribution in a given population, with extreme early types waking up when extreme late types fall asleep. This distribution is predominantly based on differences in an individuals' circadian clock. The relationship between the circadian system and different "chronotypes" is formally and genetically well established in experimental studies in organisms ranging from unicells to mammals. To investigate the epidemiology of the human circadian clock, we developed a simple questionnaire (Munich ChronoType Questionnaire, MCTQ) to assess chronotype. So far, more than 55,000 people have completed the MCTQ, which has been validated with respect to the Horne-Østberg morningness-eveningness questionnaire (MEQ), objective measures of activity and rest (sleep-logs and actimetry), and physiological parameters. As a result of this large survey, we established an algorithm which optimises chronotype assessment by incorporating the information on timing of sleep and wakefulness for both work and free days. The timing and duration of sleep are generally independent. However, when the two are analysed separately for work and free days, sleep duration strongly depends on chronotype. In addition, chronotype is both age- and sex-dependent.