Adjustment for Regression Dilution in Epidemiological Regression Analyses

Knuiman, Matthew; Divitini, Mark; Buzas, Jeffrey S.; Fitzgerald, Patrick

doi:10.1016/s1047-2797(97)00107-5

Cited by 83 publications

(67 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Even the baseline CRP value measured when the participant was free from acute infections might still not represent the usual CRP level during the long-term follow-up period due to within-person variability. This bias, referred to as regression dilution bias [27] and [28], could have also biased our effect estimates toward the null. Third, due to a relative small sample size, especially a small number of diabetes cases, we combined the two bottom tertile groups as the reference for calculating hazard ratios.…”

Section: Discussionmentioning

confidence: 99%

C-reactive protein and the risk of developing type 2 diabetes in Aboriginal Australians

Wang

Hoy

2007

Diabetes Research and Clinical Practice

View full text Add to dashboard Cite

ObjectiveTo investigate the association between C-reactive protein (CRP) and the risk of developing diabetes in Aboriginal Australians. Research design and methodsHigh sensitivity CRP levels were measured in 620 Aboriginal participants aged 20-74 years free from diabetes at baseline in a remote community in the Northern Territory of Australia. Participants were followed for a median of 11 years to identify newly diagnosed cases of diabetes. Cox proportional hazards models were used to assess the relationship of CRP levels with the risk of developing diabetes over the follow-up period. ResultsA total of 109 participants were newly diagnosed with diabetes. Incident rates were 10.8, 16.6 and 28.8 per 1000 person-years for people in the lower, middle and upper tertile groups of baseline CRP levels, respectively. After adjusting for age, sex, BMI, baseline glucose regulation status, total cholesterol, urine albumin to creatinine ratio, systolic blood pressure, smoking and alcohol drinking, the association between diabetes and CRP remained significant, with a hazard ratio of 1.23 (95% confidence interval (CI) 1.05, 1.45) corresponding to a doubling in CRP values. Similarly, the adjusted hazard ratio for development of diabetes in people in the upper tertile versus the bottom two tertiles of CRP was 1.75 (95% CI 1.19, 2.56). ConclusionsCRP is independently associated with the development of diabetes in Aboriginal people. Our findings support a role of inflammation in the etiology of diabetes in the high risk population of Aboriginal Australians.

show abstract

Section: Discussionmentioning

confidence: 99%

C-reactive protein and the risk of developing type 2 diabetes in Aboriginal Australians

Wang

Hoy

2007

Diabetes Research and Clinical Practice

View full text Add to dashboard Cite

show abstract

“…Average estimated NEAP was then tested for association with the serum bicarbonate value obtained at the 2-year (Ϯ6 months) study visit using linear regression. Given the possibility of measurement error because of day to day variability in diet, for which we could not account with averaging, we performed an additional sensitivity analysis in this population using errors in variables regression (23). These analyses simulate the measures of association that would be obtained if habitual NEAP could be precisely measured.…”

Section: Statistical Analysesmentioning

confidence: 99%

Estimated Net Endogenous Acid Production and Serum Bicarbonate in African Americans with Chronic Kidney Disease

Scialla

Appel

Astor

et al. 2011

Clinical Journal of the American Society of Nephrology

View full text Add to dashboard Cite

SummaryBackground and objectives Metabolic acidosis may contribute to morbidity and disease progression in patients with chronic kidney disease (CKD). The ratio of dietary protein, the major source of nonvolatile acid, to dietary potassium, which is naturally bound to alkali precursors, can be used to estimate net endogenous acid production (NEAP). We tested the association between estimated NEAP and serum bicarbonate in patients with CKD.Design, setting, participants, & measurements NEAP was estimated among 462 African American adults with hypertensive CKD using published equations:. Dietary protein and potassium intake were estimated from 24-hour urinary excretion of urea nitrogen and potassium, respectively. All of the eligible measurements during follow-up were modeled using generalized linear regression clustered by participant and adjusted for demographics, 24-hour urinary sodium, kidney function, and selected medications. ResultsHigher NEAP was associated with lower serum bicarbonate in a graded fashion (P trend Ͻ 0.001). Serum bicarbonate was 1.27 mEq/L lower among those in the highest compared with the lowest quartile of NEAP (P Ͻ 0.001). There was a greater difference in serum bicarbonate between the highest and lowest quartiles of NEAP among patients with stage 4/5 CKD (Ϫ2.43 mEq/L, P Ͻ 0.001) compared with those with stage 2/3 disease (Ϫ0.77 mEq/L, P ϭ 0.01; P-interaction ϭ 0.02). ConclusionsReducing NEAP, through reduction of dietary protein and increased intake of fruits and vegetables, may prevent metabolic acidosis in patients with CKD.

show abstract

“…The attenuation of the regression association between risk factor and risk was first referred to as regression dilution by MacMahon et al, 14 because it attenuates or 'dilutes' the strength of the association. Within-person variation is caused by random intra-individual measurement error, 15 and by other short-or long-term variations. 15,16 For body weight data, short-or long-term variations can be caused by factors such as smoking, illness and changes in diet or exercise.…”

Section: Survival Analysesmentioning

confidence: 99%

“…Within-person variation is caused by random intra-individual measurement error, 15 and by other short-or long-term variations. 15,16 For body weight data, short-or long-term variations can be caused by factors such as smoking, illness and changes in diet or exercise. Within-person variation that causes regression dilution also causes regression to the mean patterns.…”

Section: Survival Analysesmentioning

confidence: 99%

“…The average weight during the months or years before and after the survey is referred to as the usual or true weight. 15 In a regression model survival analysis mortality risk associated with a high usual BMI will be attributed to a higher baseline BMI and vice-versa for low usual BMI. This may artificially decrease the HR at high levels of baseline BMI, 6 and suggest that the mortality risk for obesity is lower than its true value.…”

Section: Survival Analysesmentioning

confidence: 99%

See 1 more Smart Citation

Putative biases in estimating mortality attributable to obesity in the US population

2007

View full text Add to dashboard Cite

Background: A recent analysis concluded that there were fewer excess deaths attributable to obesity in the US population than previously believed. This analysis may not have fully corrected for two putative biases, the regression-dilution and the reversecausation biases. It is not presently known whether correcting for these biases would increase estimates of excess deaths attributable to obesity. Methods: All-cause mortality hazard ratios (HRs) for different body mass index (BMI) categories were calculated and adjusted for confounding factors, using data from the prospective Atherosclerosis Risk in Communities Study. The analysis was based on 12 457 individuals aged 51-70 years and 606 all-cause deaths during a 5.3-year follow-up. The HRs were corrected for the regression-dilution and reverse-causation biases, and compared with data from a previously published study to evaluate the effects of correcting for these putative biases on estimates of excess deaths attributable to obesity in the US population.Results: The uncorrected all-cause mortality HR for obesity (BMIX30) was 1.26 (95% confidence interval (95% CI) ¼ 1.01-1.58), using the 21-25 kg/m 2 as ideal-weight category. Correcting for regression dilution increased the HR to 1.46 (95% CI ¼ 1.17-1.83). Correcting for both putative biases increased it further to 2.70 (95% CI ¼ 1.31-5.57). Such increases in HRs are consonant with increases of several hundred percent in estimates of deaths attributable to obesity in the US. Conclusions: Correcting for putative biases yielded increases in all-cause mortality HRs for obesity that correspond to substantial increases in estimates of excess deaths attributable to obesity in the US population.

show abstract

Adjustment for Regression Dilution in Epidemiological Regression Analyses

Cited by 83 publications

References 28 publications

C-reactive protein and the risk of developing type 2 diabetes in Aboriginal Australians

C-reactive protein and the risk of developing type 2 diabetes in Aboriginal Australians

Estimated Net Endogenous Acid Production and Serum Bicarbonate in African Americans with Chronic Kidney Disease

Putative biases in estimating mortality attributable to obesity in the US population

Contact Info

Product

Resources

About