Multiple imputation for missing continuous and dichotomous data: The effects of patterns of missingness and variable correlations on type I error with small samples

The purpose of this study is to investigate the effects of predictor variable correlations and patterns of missingness with dichotomous and/or continuous data in small samples when missing data is multiply imputed. Missing data of predictor variables is multiply imputed under three different multivariate models: the multivariate normal model for continuous data, the multinomial model for dichotomous data and the general location model for mixed dichotomous and continuous data. Subsequent to the multiple imputation process, Type I error rates of the regression coefficients obtained with logistic regression analysis are estimated under various conditions of correlation structure, sample size, type of data and patterns of missing data. The distributional properties of average mean, variance and correlations among the predictor variables are assessed after the multiple imputation process. ^ For continuous predictor data under the multivariate normal model, Type I error rates are generally within the nominal values with samples of size n = 100. Smaller samples of size n = 50 resulted in more conservative estimates (i.e., lower than the nominal value). Correlation and variance estimates of the original data are retained after multiple imputation with less than 50% missing continuous predictor data. For dichotomous predictor data under the multinomial model, Type I error rates are generally conservative, which in part is due to the sparseness of the data. The correlation structure for the predictor variables is not well retained on multiply-imputed data from small samples with more than 50% missing data with this model. For mixed continuous and dichotomous predictor data, the results are similar to those found under the multivariate normal model for continuous data and under the multinomial model for dichotomous data. With all data types, a fully-observed variable included with variables subject to missingness in the multiple imputation process and subsequent statistical analysis provided liberal (larger than nominal values) Type I error rates under a specific pattern of missing data. It is suggested that future studies focus on the effects of multiple imputation in multivariate settings with more realistic data characteristics and a variety of multivariate analyses, assessing both Type I error and power. ^

NF-kappaB and AP-1 connection: mechanism of NF-kappaB-dependent regulation of AP-1 activity.

Nuclear factor kappaB (NF-kappaB) and activator protein 1 (AP-1) transcription factors regulate many important biological and pathological processes. Activation of NF-kappaB is regulated by the inducible phosphorylation of NF-kappaB inhibitor IkappaB by IkappaB kinase. In contrast, Fos, a key component of AP-1, is primarily transcriptionally regulated by serum responsive factors (SRFs) and ternary complex factors (TCFs). Despite these different regulatory mechanisms, there is an intriguing possibility that NF-kappaB and AP-1 may modulate each other, thus expanding the scope of these two rapidly inducible transcription factors. To determine whether NF-kappaB activity is involved in the regulation of fos expression in response to various stimuli, we analyzed activity of AP-1 and expression of fos, fosB, fra-1, fra-2, jun, junB, and junD, as well as AP-1 downstream target gene VEGF, using MDAPanc-28 and MDAPanc-28/IkappaBalphaM pancreatic tumor cells and wild-type, IKK1-/-, and IKK2-/- murine embryonic fibroblast cells. Our results show that elk-1, a member of TCFs, is one of the NF-kappaB downstream target genes. Inhibition of NF-kappaB activity greatly decreased expression of elk-1. Consequently, the reduced level of activated Elk-1 protein by extracellular signal-regulated kinase impeded constitutive, serum-, and superoxide-inducible c-fos expression. Thus, our study revealed a distinct and essential role of NF-kappaB in participating in the regulation of elk-1, c-fos, and VEGF expression.

Changes in serum lipids with change and fluctuation in body mass index and body fat distribution in Mexican Americans with type 2 diabetes

This dissertation was written in the format of three journal articles. Paper 1 examined the influence of change and fluctuation in body mass index (BMI) over an eleven-year period, on changes in serum lipid levels (total, HDL, and LDL cholesterol, triglyceride) in a population of Mexican Americans with type 2 diabetes. Linear regression models containing initial lipid value, BMI and age, BMI change (slope of BMI), and BMI fluctuation (root mean square error) were used to investigate associations of these variables with change in lipids over time. Increasing BMI over time was associated with gains in total and LDL cholesterol and triglyceride levels in women. Fluctuation of BMI was not associated with detrimental lipid profiles. These effects were independent of age and were not statistically significant in men. In Mexican-American women with type 2 diabetes, weight reduction is likely to result in more favorable levels of total and LDL cholesterol and triglyceride, without concern for possible detrimental effects of weight fluctuation. Weight reduction may not be as effective in men, but does not appear to be harmful either. ^ Paper 2 examined the associations of upper and total body fat with total cholesterol, HDL and LDL cholesterol, and triglyceride levels in the same population. Multilevel analysis was used to predict serum lipid levels from total body fat (BMI and triceps skinfold) and upper body fat (subscapular skinfold), while controlling for the effects of sex, age and self-correlations across time. Body fat was not strikingly associated with trends in serum lipid levels. However, upper body fat was strongly associated with triglyceride levels. This suggests that loss of upper body fat may be more important than weight loss in management of the hypertriglyceridemia commonly seen in type 2 diabetes. ^ Paper 3 was a review of the literature reporting associations between weight fluctuation and lipid levels. Few studies have reported associations between weight fluctuation and total, LDL, and HDL cholesterol and triglyceride levels. The body of evidence to date suggests that weight fluctuation does not strongly influence levels of total, LDL and HDL cholesterol and triglyceride. ^

Temporal variations of dyspnea, fatigue, and lung function in chronic lung disease

Background. Research investigating symptom management in patients with chronic obstructive pulmonary disease (COPD) largely has been undertaken assuming the homeostatic construct, without regard to potential roles of circadian rhythms. Temporal relations among dyspnea, fatigue, peak expiratory flow rate (PEFR) and objective measures of activity/rest have not been reported in COPD. ^ Objectives. The specific aims of this study were to (1) explore the 24-hour patterns of dyspnea, fatigue, and PEFR in subjects with COPD; (2) examine the relations among dyspnea, fatigue, and PEFR in COPD; and (3) examine the relations among objective measures of activity/rest and dyspnea, fatigue, and PEFR in COPD. ^ Methods. The repeated-measures design involved 10 subjects with COPD who self-assessed dyspnea and fatigue by 100 mm visual analog scales, and PEFR by peak flow meter in their home 5 times a day for 8 days. Activity/rest was measured by wrist actigraphy. Single and population mean cosinor analyses and correlations were computed for dyspnea, fatigue, and PEFR; correlations were done among these variables and activity/rest. ^ Results. Circadian rhythms were documented by single cosinor analysis in 40% of the subjects for dyspnea, 60% for fatigue, and 60% for PEFR. The population cosinor analysis of PEFR yielded a significant rhythm (p < .05). The 8-day 24-hour means of dyspnea and fatigue was moderately correlated (r = .48, p < .01). Dyspnea and PEFR, and fatigue and PEFR, were weakly correlated in a negative way (r = −.11, p < .05 and r = −.15, p < .01 respectively). Weak to moderate correlations (r = .12–.34, p < .05) were demonstrated between PEFR and mean activity level measured up to 4 hours before PEFR measurement. ^ Conclusions. The findings suggest that (1) the dyspnea and fatigue experienced by COPD patients are moderately related, (2) there is a weak to modest positive relation between PEFR and activity levels, and (3) temporal variation in lung function may not affect the dyspnea and fatigue experienced by patients with COPD. Further research, examining the relations among dyspnea, fatigue, PEFR, and activity/rest is needed. Replication of this study is suggested with a larger sample size. ^

Cardiovascular diseases risk factors during the first year of life

The pattern of change in cardiovascular risk factors, blood pressure (SBP and DBP) and plasma total cholesterol (TC), over time, their tracking and their relation to anthropometric measurements during the first year of life were investigated. Also, the effect of breast feeding on TC and the relationship of blood pressure measurements and family history of CVD risk factors were examined. One hundred five newborn term, healthy infants who were seen at a pediatric clinic in The Woodlands, Texas were followed longitudinally from 2 weeks to 1 year of age. TC, blood pressure, weight and length of the infants were measured at age 2 weeks, and again at 2, 4, 6, 9 and 12 months. In addition, family history, maternal and paternal, of CVD risk factors was obtained. Data analyses included only 40 infants who completed one year of follow up.^ At 2 weeks of age, the median value for TC was 23 mg/dl higher for females than for males. This difference disappeared as infants got older. For males, most of the increase in TC median levels, from 114 to 137 mg/dl, occurred between the ages of 2 weeks and 2 months, whereas for the female group, TC levels increased moderately, about 10 mg/dl, between 9 and 12 months of age. Tracking of TC was examined by using Spearman's correlation analysis. There were strong correlations between measurements taken as early as 2 weeks of age with later measurements. These correlations were stronger and more significant for males than for females (for males, r varied between 0.51 to 0.70, whereas for females, r varied between 0.11 to 0.70). The association of body measurements with TC is no more than modest and is closer for female infants than for male infants. Analysis, also, showed that infants who received breast milk had a TC mean value 47 mg/dl higher than that for infants who received formula milk only during the period of breast feeding and this difference disappeared by age 12 months.^ In both genders, most of the increase in blood pressure (about 10-15 mmHg in both SBP and DBP) occurred during the first 4 months of life. Most of the increase for male infants occurred during the first 2 months of life, while for females, the increase in SBP and DBP was between the age of 2 and 4 months. Neither SBP nor DBP track well during the first year of life and most of the correlations between measurements at different ages were not significant for either gender. The cross-sectional relationship of blood pressure measurements and selected body measurements was assessed. For females, only at age of 12 months did DBP have positive and significant correlations with weight, length and Quetelet index (r = 0.57, 0.60 and 0.57, respectively). There were no significant correlations between blood pressure and body measurements for males. Finally, analysis showed that maternal history of CV risk factors was significantly related to SBP in the female infant group, but not for males. For DBP, neither maternal nor paternal history was related. ^