Robert C. Elston

Continuous mechanical turning of intensive care unit patients shortens length of stay in some diagnostic-related groups

Periodic changing of body position is a commonly used, generally accepted modality of intensive care unit (ICU) patient care, but the optimum frequency for turning has not been established. In order to evaluate the effect of continuous turning, 86 patients were randomized to either continuous turning on a Kinetic Treatment Table (KTT) or routine turning on conventional beds every two hours by the nursing staff. Although overall ICU mortality, length of stay, APACHE-II score, duration of mechanical ventilatory management, and incidence of nosocomial pneumonia were similar in the two groups, significant differences occurred in certain subsets of patients: continously turned patients with sepsis and pneumonia had a 3.48-day shorter adjusted length of ICU stay compared with controls (P < .001). Chronic obstructive pulmonary disease patients on the KTT also required 6.84 fewer days (P < .001) in the ICU and 4.6 days less of mechanical ventilation (P < .001) than controls and had a lower final APACHE-II score. These data suggest that continuous turning on a mechanical platform shortens the requirements for length of stay and assisted ventilation in certain critically ill patients, thereby providing a major reduction in the financial costs of medical care.

Power and robustness of sib-pair linkage tests and extension to larger sibships

We investigate the power and robustness of Haseman and Elstonfs sib-pair test for genetic linkage between a marker locus and a locus affecting a quantitative trait, and compare the test to that of Penrose. The Haseman-Elston test is more powerful than Penrroses test; its power is acceptable for cases of tight linkage and high heritability due to the hypothesized quantitative trait locus, but is quite low in other situations. Computer simulations indicate that both tests are valid for normally distributed trait values, and that the Haseman-Elston test is robust for a variety of continuous distributions of the trait values. Several linkage tests are developed for sib trios that are much more powerful , for the same total number of sibs, than the test on independent sib pairs. The Haseman-Elston test on all possible sib pairs is suggested for sibships of size larger than three and for samples including sibships of various sizes.

The Elston-Stewart Algorithm for Continuous Genotypes and Environmental Factors

The Elston-Stewart algorithm for a normally distributed trait under a polygenic model is explained in detail and extended to allow for other continuous environmental variables. This formulation is especially useful for large pedigrees, as it avoids the need to invert matrices. Whereas it may not be feasible by this method to estimate all the various components of previously suggested models for polygenic inheritance, it can allow for a reasonably flexible pedigree correlational structure under which valid tests can be performed for fixed effects that may affect the phenotype.

Generalized modulus power transformations

Several power transformations proposed in the past are examined to find out the type of distributions that they can normalize, and a general family of transformations the Generalized Modulus Power Transformation- (GEMPT), is proposed, The GEMPT will remove skewness and kurtosis and induce normality from a broad class of distributions, which we investigate, implying certain limitations for all power transformations. The use of GEMPT is illustrated and shown to iead to a better approximation to a normal distribution in an example in which the response is expected to follow a rectangular hyperbola.

Prolonged survival as a component of hereditary breast and nonpolyposis colon cancer

Identification of prolonged survival in hereditary breast and colon cancer (exclusive of familial multiple adenomatous polyposis coli) have been shown to be statistically significant (P less than .05) when compared to the American College of Surgeons Tumor Registry for these respective neoplasms. Our hypothesis is that the genotype in hereditary cancer determines both susceptibility and natural history, such that increased survival is a manifestation of the natural history.

Relationship of hassles, uplifts, and life events to psychological well-being of freshman medical students

Hassles, uplifts, and life events were related to psychological well-being with a representative sample of 55 (of 179) freshman medical students. Students were sent measures of hassles, uplifts, and affect-moods measures for 9 consecutive months and a measure of life stress at the beginning and middle of the school year. The hassles measure was found to be a better predictor of concurrent and subsequent negative mood than was the life stress measure, whereas life stress was found to be a better predictor of subsequent positive mood than hassles. On the whole, uplifts were unrelated to mood. The relationship between life stress and hassles was also investigated. The implications of the findings for future stress and health outcome research are discussed.

Search for biological/genetic markers in a long-term epidemiological and morbid risk study of affective disorders

A long-term epidemiological genetic study was conducted in which all new patients were evaluated prospectively at the Foundation for Depression and Manic Depression and two Lithium/Affective Disorders clinics at the Columbia-Presbyterian Medical Center between the years of 1972 and 1978. All patients met Feighner, RDC and DSM III criteria for Major Depressive Disorder after initial clinical screening interviews and were further subtyped using the Fieve-Dunner 7-point criteria. All 604 probands and 90% of 2711 first-degree relatives were interviewed blindly by diagnosticians trained in the use of the SADS structured interview. Cumulative morbid risk in parents, siblings and children of 490 bipolar probands was 15.6 +/- 3% and 14.0 +/- 1.7% in the first-degree relatives of 114 unipolar probands. A number of biological and genetic marker studies were simultaneously performed on samples of the overall population. The enzymes catechol O-methyltransferase and dopamine beta-hydroxylase, and the dexamethasone suppression test (SDT) did not show any biological marker value for outpatients even though both enzymes were determined to have hereditability. The HLA system, monoamine oxidase and acetylcholinesterase segregated differently from normal controls in samples of the patient population. The positive association findings with monoamine oxidase and the HLA system conflicted with the positive findings of other investigators, leaving doubtful their biological marker value. Red cell acetylcholinesterase was found to be significantly lower in affective disorder patients than in controls. This positive association finding was recently replicated by Mathews et al. (1982) but needs further confirmation. Using 28 blood group markers, a prior association study between the trait defining susceptibility to affective disorder and the genetic marker was positive for haptoglobin GC, and properdinfactor B, confirming earlier findings. Using the sib-pair method on the remaining 25 blood groups revealed that none other than peptidase A showed significant linkage with affective disorder since one significant finding is expected by chance. We conclude from the overall morbid risk data and segregation analyses that bipolar manic-depressive illness is a spectrum disease inherited through a multifactorial mode of genetic transmission (which is not synonymous with polygenetic inheritance) with possible genetic heterogeneity and find no evidence for X-linkage. Additional studies with acetylcholinesterase, haptoglobin, GC, and properdin-factor B are needed to confirm their positive biological/genetic marker value suggested by our long-term epidemiological study.

Methodological Issues in Linkage Analyses for Psychiatric Disorders: Secular Trends, Assortative Mating, Bilineal Pedigrees

A Task Force was assembled to address three data problems in genetic linkage analyses: (1) a secular trend, e.g. cohort effect; (2) positive assortative mating, and (3) bilineal pedigrees. All are cited as reasons for failure to replicate genetic linkage reports. However, we knew of no work demonstrating that these factors could invalidate or bias linkage analyses, nor that they were complications (e.g., variable age of onset). The Task Force concluded that these factors can reduce the power of linkage analyses and result in bias in the estimate of the recombination frequency due to the fact that they represent noise in the system. There was little evidence that the three factors would invalidate a linkage analysis or be directly responsible for negating a linkage finding.

Linkage and Association to Genetic Markers

Genetic markers that are sufficiently polymorphic (as measured by their heterozygosities) can be used in linkage and association analyses to detect Mendelian segregation underlying disease phenotypes. Each type of analysis can either be based on a specific genetic model or not make any assumptions about the mode of inheritance of the disease. Principles underlying these methods are reviewed, and the assumptions underlying them stressed. Association analyses are more powerful, provided there is linkage disequilibrium between the marker and disease loci; however, only linkage analyses have power in the absence of such disequilibrium. For this reason, models that allow for both kinds of tests are preferred, and such models must adequately approximate the complexity of the disease being studied.

ABO Associations with Blood Pressure, Serum Lipids and Lipoproteins, and Anthropometric Measures

Discriminant analysis was used to explore multivariate associations with ABO blood types in a biracial sample of 898 Bogalusa youths. Dependent variables included blood pressure (systolic and diastolic), serum lipid and lipoprotein levels (total cholesterol, alpha-, beta-, and pre-beta-lipoprotein cholesterol, and triglycerides), and anthropometric variables (height, weight, right arm length, triceps skinfold thickness, and a computed ponderal index). Analyses performed within race showed that several variables including beta-lipoprotein cholesterol, systolic blood pressure, and the ponderal index were sufficient to discriminate between individuals possessing the B antigen (B and AB) and those not possessing the B antigen (A and O) in the White subsample. However, height in itself can account for the detected difference, B individuals being taller than non-B individuals by a mean value of 2.4 cm. A concordant, but not significant effect was found in the Black subsample. Further tests support the conclusion that the strongest association is between ABO blood type and height.