High-dimensional versus conventional propensity scores in a comparative effectiveness study of coxibs and reduced upper gastrointestinal complications.

PURPOSE: High-dimensional propensity score (hd-PS) adjustment has been proposed as a tool to improve control for confounding in pharmacoepidemiological studies using longitudinal claims databases. We investigated whether hd-PS matching improved confounding by indication in a study of Cox-2 inhibitors (coxibs) and traditional nonsteroidal anti-inflammatory drugs (tNSAIDs) and their association with the risk of upper gastrointestinal complications (UGIC). METHODS: In a cohort study of new users of coxibs and tNSAIDs we compared the effectiveness of these drugs to reduce UGIC using hd-PS matching and conventional propensity score (PS) matching in the German Pharmacoepidemiological Research Database. RESULTS: The unadjusted rate ratio (RR) of UGIC for coxib users versus tNSAID users was 1.21 [95 % confidence interval (CI) 0.91-1.61]. The conventional PS matched cohort based on 79 investigator-identified covariates resulted in a RR of 0.84 (0.56-1.26). The use of the hd-PS algorithm based on 900 empirical covariates further decreased the RR to 0.62 (0.43-0.91). CONCLUSIONS: A comparison of hd-PS matching versus conventional PS matching resulted in improved point estimates for studying an intended treatment effect of coxibs versus tNSAIDs when benchmarked against results from randomized controlled trials.

Design and results of the pretest of the IDEFICS study.

BACKGROUND: During the preparatory phase of the baseline survey of the IDEFICS (Identification and prevention of dietary- and lifestyle-induced health effects in children and infants) study, standardised survey procedures including instruments, examinations, methods, biological sampling and software tools were developed and pretested for their feasibility, robustness and acceptability. METHODS: A pretest was conducted of full survey procedures in 119 children aged 2-9 years in nine European survey centres (N(per centre)=4-27, mean 13.22). Novel techniques such as ultrasound measurements to assess subcutaneous fat and bone health, heart rate monitors combined with accelerometers and sensory taste perception tests were used. RESULTS: Biological sampling, physical examinations, sensory taste perception tests, parental questionnaire and medical interview required only minor amendments, whereas physical fitness tests required major adaptations. Callipers for skinfold measurements were favoured over ultrasonography, as the latter showed only a low-to-modest agreement with calliper measurements (correlation coefficients of r=-0.22 and r=0.67 for all children). The combination of accelerometers with heart rate monitors was feasible in school children only. Implementation of the computer-based 24-h dietary recall required a complex and intensive developmental stage. It was combined with the assessment of school meals, which was changed after the pretest from portion weighing to the more feasible observation of the consumed portion size per child. The inclusion of heel ultrasonometry as an indicator of bone stiffness was the most important amendment after the pretest. DISCUSSION: Feasibility and acceptability of all procedures had to be balanced against their scientific value. Extensive pretesting, training and subsequent refinement of the methods were necessary to assess the feasibility of all instruments and procedures in routine fieldwork and to exchange or modify procedures that would otherwise give invalid or misleading results.

Repeatability of maternal report on prenatal, perinatal and early postnatal factors: findings from the IDEFICS parental questionnaire.

OBJECTIVE: To investigate the repeatability of maternal self-reported prenatal, perinatal and early postnatal factors within the IDEFICS (Identification and prevention of dietary- and lifestyle-induced health effects in children and infants) study. DESIGN: Data are from the baseline survey of the longitudinal cohort study IDEFICS in eight European countries. SUBJECTS: A total of 420 parents from eight countries (43-61 per country) were asked to complete the parental questionnaire (PQ) twice at least 1 month apart. MEASUREMENTS: The PQ assesses prenatal (maternal weight gain), perinatal (child's birth weight and length, Caesarean (C)-section, week of delivery) and early postnatal factors (exclusive breastfeeding, breastfeeding, introduction of solid food). Intra-class correlation coefficients (ICCs) were calculated to compare maternal reports on prenatal, perinatal and early postnatal factors between the first and second PQ. RESULTS: In total, 249 data sets were considered for the analyses. Overall, maternal reports for prenatal and perinatal factors showed higher repeatability (ICC=0.81-1.00, P≤0.05 for all) than those for early infant nutrition (ICC=0.33-0.88, P≤0.05 for all). Perfect agreement was found for parental reports on C-section (ICC(all)=1.00, P≤0.05). There was stronger agreement for duration of breastfeeding (ICC=0.71, P≤0.05) compared with exclusive breastfeeding (ICC=0.33, P≤0.05). Maternal reports showed moderate correlation for the introduction of several types of food (cereals ICC=0.64, P≤0.05; fruits ICC=0.70, P≤0.05; meat ICC=0.83, P≤0.05; vegetables ICC=0.75, P≤0.05), and high correlation (ICC=0.88, P≤0.05) for cow's milk. CONCLUSION: Maternal reports on pregnancy and birth were highly reproducible, but parental recall of early infant nutrition was weaker and should be interpreted more cautiously.

Quaternary and subunit structure of Calliphora arylphorin as deduced from electron microscopy, electrophoresis, and sequence similarities with arthropod hemocyanin.

Arylphorin was purified from larvae of the blowfly Calliphora vicina and studied in its oligomeric form and after dissociation at pH 9.6 into native subunits. In accordance with earlier literature, it was electrophoretically shown to be a 500 kDa hexamer (1 x 6) consisting of 78 kDa polypeptides (= subunits). Electron micrographs of negatively stained hexamers show a characteristic curvilinear, equilateral triangle of 12 nm in diameter (top view) and a rectangle measuring 10 x 12 nm (side view). Alternatively, particles in the top view orientation exhibit a roughly circular shape 12 nm in diameter. Crossed immunoelectrophoresis revealed the presence of a major subunit type; the nature of a very minor and a third immunologically separated component remains unclear. A novel 2 x 6 arylphorin particle was detected and isolated. It comprises less than 10% of the total arylphorin material and shows a long, narrow interhexamer bridge in the electron microscope. An arylphorin dissociation intermediate identified as a trimer (1/2 x 6) was isolated; its possible quaternary structure is discussed on the basis of electron micrographs. The epitope of monoclonal antibody Ec-7 directed against tarantula (Eurypelma californicum) hemocyanin subunit d and also reactive to Calliphora arylphorin was traced to a highly conserved peptide of 27 amino acids localized in the center of the protein. The primary structure of Calliphora arylphorin as published in our preceding paper (Naumann and Scheller 1991) is compared in detail to the sequences of spider and spiny lobster hemocyanin. This revealed a basic framework of 103 strictly conserved amino acids. Isofunctional exchanges are proposed for another 76 positions. On the basis of these similarities, and the published three-dimensional model of spiny lobster hemocyanin, a detailed model of the quaternary structure of Calliphora arylphorin is presented. A second larval storage protein previously termed protein II was purified from Calliphora hemolymph. It was demonstrated to be a 500 kDa hexamer of 83 kDa subunits. In the electron microscope it shows a cubic view 9 nm in length with a large central hole and a rectangular view (9 x 10 nm) with a large central cavity. A morphologically very similar hemolymph protein was detected in Drosophila melanogaster larvae. From its structural appearance it is uncertain whether protein II belongs to the hemocyanin superfamily or not.