SEARCH · PubMed Health

Results for “multiple linear regression”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 775 records · Page 43Linked to original sources

PubMed2005

Atherosclerosis and the Glu298Asp polymorphism of the eNOS gene in white patients with end-stage renal disease.

BACKGROUND: We investigated whether the eNOS G/T polymorphism (Glu298Asp variant) is linked to the severity of carotid atherosclerosis and whether it is independent of asymmetric dimethylarginine (ADMA) in determining vascular damage in patients with end-stage renal disease (ESRD). METHODS: The eNOS polymorphism, ADMA, carotid intima-media thickness (IMT), and carotid artery (CCA) wall-to-lumen ratio (an indicator of arterial remodeling) were determined/measured in 131 patients with ESRD. RESULTS: Both in the co-dominant and dominant model approach, IMT as well as CCA wall-to-lumen ratio were directly related to the T allele (P < or = .009) and these relationships held true in multiple linear regression analyses including ADMA and traditional and emerging risk factors. The relationship between eNOS genotypes and CCA wall-to-lumen ratio was further analyzed by a categorical approach and in a multiple logistic regression analysis, the odds ratio (OR) of increased CCA wall-to-lumen ratio was strongly associated to the T allele (codominant model: GG, OR = 1; GT, OR = 2.1; TT, OR = 8.2; P for trend = .01; dominant model: GG, OR = 1; GT and TT, OR = 2.7; P = .02). CONCLUSIONS: The T allele of eNOS gene is an independent predictor of intimal lesions and vascular remodeling and it is associated with the severity of atherosclerosis independently of ADMA.

Aged↗

PubMed1990

Quantitative structure-retention relationship studies of odor-active aliphatic compounds with oxygen-containing functional groups.

High-quality regression equations (R greater than 0.996) modeling the gas chromatographic retention indices of 115 odor-active compounds on stationary phases of different polarity are generated by using the ADAPT software system. Multiple linear regression techniques are used to describe the statistical relationship between the Kováts retention indices and structure-based molecular descriptors. These descriptors encode topological, geometrical, and electronic features of the molecules. The utility of several new descriptors encoding functions of partial atomic charge and solvent accessible surface area is demonstrated. Quantitative predictions of odor threshold values for a subset of these compounds containing the alcohol functional group are also calculated by using a similar methodology.

Hydrocarbons↗

PubMed1996

Determinants of Economic Cost Related to Low Back Pain among Nurses at a University Hospital.

The objectives of this study were to determine the economic cost related to low back pain (LBP) among nurses, and to identify factors associated with this cost. All 998 nurses at the National Taiwan University Hospital (NTUH) were asked to complete a questionnaire designed to examine whether the nurses had experienced LBP and whether their LBP had been related to the incurring of economic cost between May 1, 1990, and April 30, 1991. Of the 863 respondents, 417 had had LBP and 102 of these had incurred economic costs related to the LBP. Personal interviews of the 102 nurses, as well as of an additional four nurses who had resigned due to LBP within the 12 months, were performed to determine the related economic costs. The total monthly costs over the 12 months ranged from U.S. $105,405 to $149,083. Twenty percent of the 106 cases were responsible for 70% of the total overall cost. There was no difference in demographics between the two groups of non-LBP and LBP-without-cost. However, statistical analysis using logistic regression showed that the occurrence of economic cost was positively associated with subjective pain, lifting and carrying babies and small children at home, and age or work years or parity. In addition, multiple linear regression analysis showed that medical cost was positively associated with parity and sick leave; and productivity loss was associated with sick leave, lifting and carrying babies at home, unspecified housework, patient transport, and performing cardiopulmonary resuscitation.

Journal Article↗

PubMed1992

A controlled study of the prevalence of cognitive dysfunction in randomly selected patients with systemic lupus erythematosus.

OBJECTIVE: To investigate the prevalence of cognitive dysfunction in randomly selected patients with systemic lupus erythematosus (SLE). METHODS: Randomly selected, ambulatory patients with SLE (n = 49) or with rheumatoid arthritis (RA) (n = 40) completed neuropsychological tests. These included Associate Learning, Switching Attention, Continuous Performance, Associate Recall, Hand-Eye Coordination, Pattern Comparison, Pattern Memory, the Stroop Color and Word Test, and the Symptoms Checklist-90R. Results were evaluated by multiple linear regression analysis. RESULTS: SLE patients had poorer performance than RA patients on the test of attention (P = 0.002) and tests of visuospatial ability (P = 0.03) and P = 0.04), independent of age, education, or steroid use. The conservative level of statistical significance, adjusting for multiple comparisons, was 0.005. SLE patients reported more symptoms of cognitive difficulty. CONCLUSION: Cognitive dysfunction is common in ambulatory SLE patients as measured by standardized tests and is a cause of distress and impaired functioning. Self-reported cognitive difficulty appears to correlate with objective performance.

Adult↗

PubMed2000

Sperm factors related to in vitro penetration of porcine oocytes.

This study was designed to evaluate the relationship between sperm factors and penetration capacity in an in vitro system with immature porcine oocytes. The sperm parameters evaluated in 145 ejaculates were volume, sperm concentration, total cells in the ejaculate, ATP content, morpho-anomalies, percentage of motile sperm cells, forward progressive motility (FPM), acrosome status (NAR), hypo-osmotic swelling test (HOS), osmotic resistance test (ORT), eosin-nigrosin viability stain and sperm membrane integrity (DCF). Porcine oocytes (a total of 8,736) were used to evaluate the capacity of the different sperm assays to predict penetration. Many parameters were found to be related to in vitro penetration ability; all conventional semen parameters, except sperm concentration and eosinnigrosin staining, were significantly better in high (>75%) than in low penetration rates (<75%). When the ejaculates were preselected the number of significantly related parameters was lower. When studying all conventional semen parameters through a stepwise multiple linear regression analysis of seminal measurements, up to 72.3% of total variance of the penetration rate could be predicted. However, as many as 4 parameters were needed (FPM in fresh semen, folded tail, NAR in post-treatment semen and DCF) for accurate prediction. On the other hand, the multiple logistic regression needed 7 parameters to discriminate 83.96% of the cases correctly. In summary, the results from the present study showed that almost all studied parameters were significantly different for predicting penetration process attained or failed, but most of them were correlated together. These findings emphasize the complexity of sperm functions and the difficulty of assessing the fertilizing ability.

Acrosome↗

PubMed2005

Racial disparities in the association of foetal growth retardation to childhood blood pressure.

BACKGROUND: Foetal growth retardation (FGR), defined as less than the 10th percentile of birth weight for gestational age, is reported to be an important contributor to hypertension and cardiovascular disease in children and adults, but findings are not consistent. For this reason we re-examined the role of FGR in childhood blood pressure. METHODS: We performed univariate and multivariate analyses on data gathered from 262 children, age 5 years, born to mothers at risk for pre-term delivery or FGR infant. The characteristics of the mothers and the children were evaluated using Student's t-test. Rates and proportions were compared using either chi-square or Fisher's exact test. Linear regression models evaluated the effect of birth weight and body mass index on systolic and diastolic blood pressure. Multivariate linear regression was used to model the effects of FGR, gestational age, body mass index, race, gender, maternal smoking, maternal gestational diabetes on blood pressure while adjusting for possible confounders. RESULTS: Systolic blood pressure was inversely associated with birth weight in white children while a small direct association was noted in African Americans. Body mass index was positively associated with systolic blood pressure in both groups. Multiple linear regression analyses showed FGR and early gestational age were associated with higher blood pressure in white but not African American children, accounting for a 13.2 mmHg difference between FGR and appropriate for gestational age groups. Blood pressure in African Americans was strongly affected by maternal gestational diabetes and smoking. CONCLUSIONS: Birth weight influences childhood blood pressure but the effects may vary depending on ethnic group. The relative importance of birth weight on blood pressure may depend on other prenatal and post-partum risks.

Adult↗

PubMed2005

Quality of life of caregivers in Parkinson's disease.

OBJECTIVE: To assess the impact of PD on informal caregivers of patients and identify the main factors related to caregiver strain. PATIENTS AND METHODS: Pairs of PD patients and their caregivers. Evaluation by neurologists included the Hoehn and Yahr, Schwab and England, UPDRS (parts 1-3), ISAPD, and Pfeiffer's SPMSQ rating scales. Patients completed the Euro-QoL 5D, PDQ-8, and Hospital Anxiety and Depression Scale. The SQLC was used to assess caregivers' quality of life (QoL), with caregivers, in turn, applying the Euro-QoL and PDQ-8 to assess patients' health-related quality of life (HRQoL). Multiple linear regression models were fitted to ascertain factors linked to the SQLC. RESULTS: Significant correlations were in evidence between the following scores: SQLC and clinical rating scales and SQLC and patients' HRQoL. Based on multiple regression analysis, patients' functional state (ADL) proved to be the main predictor of caregivers' QoL. Self- and caregiver-assessed patients' HRQoL also proved to be a relevant factor. CONCLUSIONS: (1) Patients' functional state was significantly related to caregivers' psychosocial burden; (2) patients' HRQoL proved to be an additional factor linked to caregiver QoL; (3) improvement of patient disability and HRQoL might alleviate caregiver strain.

Aged↗

PubMed2005

Lower hypertension risk in Mexico City than in San Antonio.

BACKGROUND: We examined the effects of blood pressure (BP), weight, and weight gain on hypertension risk in two similar ethnic origin populations, subjects in Mexico City and Mexican Americans in San Antonio. METHODS: The Mexico City Diabetes Study and San Antonio Heart Study are population-based, epidemiologic studies with identical survey protocols. Incident hypertension (BP > or = 140/90 mm Hg or current antihypertensive treatment) was analyzed in subjects aged 35 to 64 years of Mexican ethnicity living in low-income neighborhoods (n = 1467 in Mexico City, n = 628 in San Antonio). RESULTS: In Mexico City, 10.6% of men and 13.1% of women developed hypertension in a 6.5-year period; in San Antonio, 28.6% and 28.7% in a 7.5-year period, respectively. Poisson regression analysis demonstrated a greater hypertension risk in San Antonio for both men (risk ratio [RR] = 1.75, 95% confidence interval [CI]: 1.19-2.56) and women (RR = 1.40, 95% CI: 1.05-1.86). In a multiple linear regression analysis, systolic BP change was associated with weight gain in Mexico City (P < .001 in men and women) and San Antonio (P = .045 in men, and P = .027 in women) independently of age, BP, obesity, alcohol consumption, cigarette smoking, diabetes, and antihypertensive treatment. These covariates did not fully explain greater increments of systolic BP in San Antonio than in Mexico City (P < .001 in men and women). CONCLUSIONS: Hypertension risk is lower in Mexico City than in San Antonio. Systolic BP increases with weight gain, independently of other determinants of hypertension.

Adult↗

PubMed1996

Sciatic function index, nerve conduction tests, muscle contraction, and axon morphometry as indicators of regeneration.

The purpose of this study was to determine which parameters were the best measure of nerve regeneration, assuming that the sciatic functional index (SFI) represented the "gold standard." Three different sciatic functional indexes and 11 commonly used electrophysiologic and morphologic indicators of regeneration were all determined in 24 rats 12 weeks after one of three lesions was created in the sciatic nerve. With linear regression analysis, only fiber/axon diameter ratio (D/d) and myelin thickness/axon diameter ratio showed statistically significant correlations with sciatic functional index (r = 0.55 and 0.53, respectively). The other 11 parameters had poorer correlation. Therefore, if sciatic functional index is the best measure of comprehensive nerve function, then other parameters are not. It is probable that each parameter measures some different component of the regeneration process. A stepwise multiple linear regression analysis produced a model that included D/d, nerve conduction velocity, and nerve action potential amplitude that gave a slightly better correlation (r = 0.67). The relatively poor correlation between sciatic functional index and the other parameters of nerve function indicates that all nerve regeneration studies must be interpreted carefully before comparisons are made. Furthermore, the best measure of nerve function remains unproved or undiscovered in the experimental animal.

Animals↗

PubMed2004

Obesity, emerging risk in China: trend of increasing prevalence of obesity and its association with hypertension and hypercholesterolaemia among the Chinese.

1. The aim of the present study was to investigate the trend of the prevalence of obesity in China and its association with hypertension and hypercholesterolemia. 2. A multicentre cross-sectional study was conducted in Chinese men and women aged 48-56 years between 1985 and 2000. In the report, three study periods were classified as survey 1 (1985), survey 2 (1988-1989) and survey 3 (1997-2000) in order to describe the long-term trend. 3. The results show that: (i) mean body mass index (BMI), prevalence of obesity (BMI > or = 28 kg/m2) and overweight (BMI > or = 25 and < 28 kg/m2) increased significantly from 1985 to 2000 in both sexes (P < 0.001); (ii) similar to the trend for BMI, the prevalence of hypertension and hypercholesterolaemia (total cholesterol (TC) > or = 220 mg/dL) also increased significantly from 1985 to 2000 (P < 0.001); (iii) partial correlation and multiple linear regression analyses indicated that increased BMI was significantly positively correlated with blood pressure and TC and negatively correlated with serum high-density lipoprotein-cholesterol (P < 0.01 or P < 0.001); and (iv) multiple logistics regression analysis models indicated that obese subjects had a 3.9-fold higher risk of being hypertensive (relative risk (RR) 4.9; 95% confidence interval (CI) 3.4-7.3) compared with those subjects who had a BMI less than 25 kg/m2. The corresponding RR (95% CI) of obesity for hypercholesterolaemia was 2.63 (1.57-4.40). 4. In conclusion, the results highlight the epidemic of obesity, an emerging risk in China. Great efforts must be made to alter this unwelcome trend.

Asian People↗

PubMed2026

Sex-specific differences in liver DNA methylation patterns and epigenetic aging in mice.

Biological sex has been shown to influence aging outcomes, contributing to distinct trajectories in disease susceptibility and lifespan. DNA methylation patterns provide a quantitative measure of biological aging. This study investigated whether aged male and female mice display distinct liver DNA methylation patterns and differences in epigenetic aging. Liver samples were collected from 17 aged c57BL/6 mice (6 males, 11 females). Genomic DNA was extracted and bisulfite-converted before targeted enrichment of 2,045 murine age-associated CpG loci. Biological age (DNAge) was estimated using a previously developed DNA methylation-based predictor generated through elastic net regression. The difference (ΔDNAge) between DNAge and chronological age was computed. Sex-specific differences were assessed by comparing site-specific methylation ratios, ΔDNAge values, and through principal component analysis (PCA) and multiple linear regression. Twelve CpG sites across six genes (Fam84b, Zswim6, Hsf4, Mn1, Qprt, and Rapgefl1) showed significant sex-associated differences in methylation. Fam84b demonstrated the largest and most consistent sex-associated effect, with all three associated CpG sites showing higher methylation in males (regression coefficients: -0.204, -0.281, and -0.294). Zswim6 exhibited consistent lower methylation ratios in females, whereas the other genes showed higher methylation in females. There were no sex differences in biological age or ΔDNAge (P = 0.596). Although the epigenetic clock did not reveal differences between sexes in aging, aged mice did exhibit sex-specific liver methylation patterns different from those reported in younger mice, suggesting that sex-dependent epigenetic changes may emerge later in life and may reflect sexual dimorphism in liver function with age.NEW & NOTEWORTHY Males and females are known to age differently and develop certain diseases at different rates. Here, we examined the livers of aged male and female mice to see if they show different DNA methylation patterns. We found that aged male and female mice had distinct DNA methylation patterns at specific genes. Interestingly, most of these methylation differences were not present in younger mice, suggesting that sex differences in the genome may change with age.

Animals↗

PubMed1998

Power and sample size calculations for studies involving linear regression.

This article presents methods for sample size and power calculations for studies involving linear regression. These approaches are applicable to clinical trials designed to detect a regression slope of a given magnitude or to studies that test whether the slopes or intercepts of two independent regression lines differ by a given amount. The investigator may either specify the values of the independent (x) variable(s) of the regression line(s) or determine them observationally when the study is performed. In the latter case, the investigator must estimate the standard deviation(s) of the independent variable(s). This study gives examples using this method for both experimental and observational study designs. Cohen's method of power calculations for multiple linear regression models is also discussed and contrasted with the methods of this study. We have posted a computer program to perform these and other sample size calculations on the Internet (see http://www.mc.vanderbilt.edu/prevmed/psintro+ ++.htm). This program can determine the sample size needed to detect a specified alternative hypothesis with the required power, the power with which a specific alternative hypothesis can be detected with a given sample size, or the specific alternative hypotheses that can be detected with a given power and sample size. Context-specific help messages available on request make the use of this software largely self-explanatory.

Bacterial Vaccines↗

PubMed2003

Dioxins released from incineration plants and mortality from major diseases: an analysis of statistical data by municipalities.

The aim of this study is to examine the relation between mortality and the existence of incineration plants and dioxin released from the plants in 590 municipalities across Japan. The concentration of dioxins in emissions from incinerators, the amount of dioxins per population, the cumulative amount of dioxins, and the cumulative amount per land area were used as dioxin-related municipal indices. Age-adjusted mortality rates from all causes and five major disease categories by municipality in 1995 were used as health indices. The relation was examined using t-test, analysis of covariance (ANCOVA), correlation coefficients and multiple linear regression analysis, considering the effects of cities' socioeconomic conditions. Although municipalities with plants had significantly higher mortality from female stroke and lower mortality from male cancer at all site and lung than municipalities without plants, these differences were not significant in ANCOVA with socioeconomic indicators. The significant relation between mortality and dioxin indices in correlation coefficient was ruled out when the socioeconomic conditions were adjusted in multiple regression analysis. This study did not show the statistical relation between increased mortality from major causes and the existence of incineration plants and dioxins from the plants at the municipal level.

Age Factors↗

PubMed2004

Long-term follow-up and factors influencing success of silastic ring vertical gastroplasty.

BACKGROUND: Careful selection of patients for the operation plays a major role in long-term results of silastic ring vertical gastroplasty. The objective of the current study is to identify predictive factors for the objective (excess weight loss) and subjective (satisfaction) success of silastic ring vertical gastroplasty (SRVG) for morbid obesity. METHODS DESIGN: Retrospective cohort study. SETTING: University hospital. SUBJECTS: 300 patients of 450 who were operated on between 1984 and 1997. Follow-up time was 4.4 +/- 2.3 years. INTERVENTION: SRVG. STATISTICAL METHODS: Correlations, multi-linear regression model. RESULTS: Average BMI (body mass index) loss: 13.6 +/- 7.4 kg/m(2), average excess weight loss was 67.4 +/- 33.0%. Satisfaction rate: 81.3%. Correlation was found between objective parameters of successes and the satisfaction of the patient (p < 0.001). Excess weight loss was correlated to younger age (p < 0.005), pre-operative weight and BMI (p < 0.005, p < 0.01, respectively), and shorter follow-up (p < 0.001). Multiple linear regression model revealed that age and preoperative weight were independent variables and correlated to the excess weight loss after SRVG (R(2) = 0.303, p < 0.01; R(2) = 0.026, p < 0.05). Social support was correlated to satisfaction (p < 0.05). CONCLUSIONS: SRVG is an operation with high rates of objective and subjective success rate. Younger and heavier subjects will mostly enjoy SRVG in terms of excess weight loss. Patients who have social support have the most satisfactory emotional outcome.

Adult↗

PubMed1994

Human HDL cholesterol levels are determined by apoA-I fractional catabolic rate, which correlates inversely with estimates of HDL particle size. Effects of gender, hepatic and lipoprotein lipases, triglyceride and insulin levels, and body fat distribution.

High-density lipoprotein (HDL) cholesterol (HDL-C) levels are a strong inverse predictor of atherosclerosis risk, but the physiological determinants of HDL-C levels are poorly understood. We selected 57 human subjects (30 women and 27 men) with a broad range of HDL-C levels and performed turnover studies of apolipoprotein (apo)A-I and apoA-II, the two major apolipoproteins of HDL, to measure the fractional catabolic rate (FCR) and production or transport rate (TR) of these proteins. We also measured several other parameters known to correlate with HDL-C levels to test for their interrelations and to postulate mechanisms of regulation of HDL-C levels. As expected, the women had higher levels of HDL-C (56.7 +/- 21.4 versus 45.1 +/- 16.3 mg/dL, mean +/- SD; P = .03) and apoA-I (147 +/- 32 versus 126 +/- 29 mg/dL, P = .01) than men and did not differ in apoA-II levels (34.5 +/- 7.4 versus 33.3 +/- 7.5 mg/dL, P > .2). The FCR of apoA-I tended to be lower in the women (0.248 +/- 0.077 versus 0.277 +/- 0.069 pools/d, P = .1), although the difference was not statistically significant. The FCR of apoA-II was also lower (0.184 +/- 0.043 versus 0.216 +/- 0.056 pools/d, P = .02). In contrast, the apoA-I TR was equal in women and men (12.0 +/- 1.6 versus 12.1 +/- 2.8 mg/kg per day, P > .2), and there was a trend toward lower apoA-II TR in women (2.19 +/- .62 versus 2.61 +/- 1.06 mg/kg per day, P = .07). Linear regression analysis revealed a strong inverse correlation between HDL-C levels and the FCRs of apoA-I and apoA-II (r = -.81 and -.76, respectively; P < .0001 for both). In contrast, there was little or no association between HDL-C and the TRs of apoA-I and apoA-II (r = .06 and -.35, P = not significant and .01, respectively). In stepwise multiple linear regression analysis, apoA-I FCR alone accounted for 66% of the variability in HDL-C; two other variables accounted for an additional 7%. Due to the importance of apoA-I FCR, its determinants were sought among the remaining variables.(ABSTRACT TRUNCATED AT 400 WORDS)

Adipose Tissue↗

PubMed2005

The use of quantile regression in health care research: a case study examining gender differences in the timeliness of thrombolytic therapy.

Investigators are frequently interested in determining patient and system characteristics associated with delays in the provision of essential medical treatment. Investigators have typically used either multiple linear regression or Cox proportional hazards models to assess the impact of patient and system characteristics on the timeliness of medical treatment. A drawback to the use of these two methods is that they allow, at best, a partial exploration of how a distribution of delays in treatment or of waiting times changes with patient characteristics. In contrast, quantile regression models allow one to assess how any quantile of a conditional distribution changes with patient characteristics. We illustrate the utility of quantile regression by examining gender differences in the delivery of thrombolysis in patients with an acute myocardial infarction. We demonstrate that richer inferences can be drawn through the use of quantile regression. Females were more likely to experience delays in treatment compared to males. Furthermore, gender had a greater impact upon those patients who had the greatest delays in treatment. Investigators who want to determine how a distribution of delays in treatment or of waiting times changes with patient or system characteristics should consider complementing their analyses with the use of quantile regression.

Aged↗

PubMed2004

Admission lactate level and the APACHE II score are the most useful predictors of prognosis following torso trauma.

BACKGROUND: Markers of dysoxic metabolism and scoring systems for triage have been widely used in critically injured patients. However, so far, no model is sufficiently reliable to predict the outcome in trauma victims. The purposes of the present study, therefore, were to determine whether a correlation exits between the main trauma scoring systems and the markers of dysoxic metabolism. Moreover, to assess if any of the admission parameters can be used to indicate outcome. METHODS: Sixty-four patients were included in this study. Admission data, including arterial lactate level, base deficit (BD), pH, revised trauma score (RTS), injury severity score (ISS), shock index (SI), and Acute Physiology and Chronic Health Evaluation (APACHE II), were collected and analysed by logistic regression analysis. Degree of association between continuous variables were calculated by either Pearson's or Spearman's correlation coefficient, where applicable. The dependence of lactate on two or more other variables was evaluated by multiple linear regression analysis. RESULTS: Logistic regression analysis showed that the fatal outcome following major torso trauma was principally associated with the APACHE II score and lactate. The specificity and the sensitivity of this logistic regression model was 94.6 and 79.2%, respectively. According to standardised linear regression coefficients, BD was the best single predictor of lactate, and APACHE II added a small amount of predictive power. The proportion of total variation in lactate level explained by base deficit, APACHE II and age is R2=85.2%. CONCLUSION: APACHE II score and the arterial lactate level are the most important determinants of clinical outcome in critically injured patients. A correlation exits between lactate and APACHE II and between lactate and base deficit.

APACHE↗

PubMed2006

The association of lead exposure and motor performance mediated by cerebral white matter change.

The goals were to determine if lead exposure is associated with cerebral white matter changes (WMC) and if so, does WMC mediate the relation between lead and psychomotor slowing as measured by Grooved Pegboard (GP). In the literature, age is the strongest predictor of WMC and therefore 61 lead smelter workers age 50 and under were included in the study population. Mean (range) age was 40 (23-50) years, years of education was 9 (0-13), duration of employment was 19 (1-26), current blood lead (PbB) was 29 (16-42) microg/dl, working lifetime weighted integrated blood lead (IBL) was 826 (65-1451) microg year/dl, working lifetime weighted average blood lead (TWA) was 42 (17-59) microg/dl, and bone lead (PbBn) was 39 (-12-90) microg Pb/g bone mineral. WMC, recorded as hyperintensities on T2-weighted MRI of the brain were graded. Lead variables were entered in a logistic regression attempting to differentiate normal versus abnormal MRI, after controlling for age and cerebrovascular risk factors. Direct effects of lead on GP and indirect effects of lead on GP through WMC was modeled using multiple linear regression analyses after controlling for the covariates. WMC were present in 23% of MRIs. Logistic regression of WMC on lead exposure metrics demonstrated significantly elevated odds ratios for IBL, TWA, and PbBn after the covariates. Of the lead exposure variables, IBL (beta=0.339, p<0.10) had a larger direct effect on GP after adjusting for the covariates than PbBn (beta=0.265, p<0.10). After adjusting for the lead term and covariates WMC accounted for an additional effect on GP performance after PbBn (beta=0.261, p<0.10) and after IBL (beta=0.278, p<0.05). Path analysis demonstrated that some of the relationship of both PbBn and IBL with GP is mediated by WMC.

Adult↗