Analysis plan for “The Association of eGFR With Carotid Plaque Risk Is Largely Explained by Age in a Cardiovascular High-Risk Population: A Reappraisal With a Survey of Estimated Pulse Wave Velocity Mediation Analyses” — retrospective registration of the analysis plan
Year: 2026
Type: study-registration
Abstract: IMPORTANT — DISCLOSURE OF RETROACTIVE STATUS Data collection (China PEACE Million Persons Project, Wenzhou pilot screening) and all statistical analyses were completed BEFORE this registration, and the manuscript reporting these analyses has been prepared. This record is made for transparency: it timestamps the hypotheses and the analysis plan exactly as they were conducted, and it is explicitly labeled as a RETROACTIVE registration. It is not a prospective preregistration and did not influence the analyses. Description / Abstract Cross-sectional analysis of 1,097 adults at high cardiovascular risk from the Wenzhou pilot of the China PEACE Million Persons Project, examining (i) the association of kidney function (eGFR, CKD-EPI equation) with carotid plaque, (ii) its apparent mediation by estimated pulse wave velocity (ePWV, a deterministic formula of age and mean blood pressure), and (iii) how every conventional finding changes when age and sex are controlled. A companion structured survey of published ePWV mediation analyses (PubMed, September 2026) documents how the genre handles the formula’s demographic inputs. Hypotheses (as tested) Under the conventional specification used in the literature (age and sex omitted from covariate sets), lower eGFR is associated with higher odds of carotid plaque (logistic Models 1–3; dose–response by restricted cubic splines with 3 knots at the 10th, 50th, and 90th percentiles; two-piecewise threshold model). Conventional causal mediation analysis attributes a substantial proportion of this association to ePWV (Imai et al. framework; 1,000 bootstrap iterations). All conventional findings are null after adjustment for age and sex, within age strata (<60 vs ≥60 years, interaction tested by likelihood ratio), and after excluding established cardiovascular disease; the mediated proportion collapses to approximately zero. A structured survey of published ePWV mediation analyses (2024–2026) shows inconsistent handling of chronological age across the genre. Data Individual-level data from the China PEACE Million Persons Project Wenzhou pilot (6,105 screened; 1,097 analyzed). Not publicly shareable: individual-level health information under project data-use agreements. De-identified data available from the corresponding author on reasonable request. Analysis plan — cohort component Conventional specification. Continuous variables are reported as mean (SD) or median (interquartile range), and categorical variables as numbers (percentages); eGFR-tertile comparisons used ANOVA, the Kruskal–Wallis test, or the chi-square test. eGFR, entered continuously (per 10-unit increment) or in tertiles, was related to carotid plaque by logistic regression in three models: Model 1 unadjusted; Model 2 adjusted for marital status, BMI, waist circumference, family income, and education; and Model 3 adding occupation, smoking, alcohol use, hypertension, diabetes, dyslipidemia, ALT, and AST. Consistent with much of the prior literature—and because the CKD-EPI equation already incorporates age and sex—age and sex were omitted from these conventional models and made the object of explicit sensitivity analysis. The dose–response was modeled with restricted cubic splines (3 knots at the 10th, 50th, and 90th percentiles) and a two-piecewise logistic model anchored at the spline-informed cut point. Mediation used the approach of Imai et al. with 1,000 bootstrap iterations, eGFR coded per 10-unit increment, and Model 3 covariates. Age-sensitive re-specification. All conventional analyses were re-estimated with (a) age and sex added to Models 1–3, plus a minimal age-and-sex-only model; (b) stratification by age (<60 vs ≥60 years) with a likelihood-ratio interaction test; (c) mediation re-estimated with age and sex; (d) the two-piecewise model with age and sex; and (e) exclusion of participants with established cardiovascular disease. Software: R and Free Statistics version 2.2; two-sided P < 0.05 defined statistical significance. Analysis plan — survey component PubMed search on 10 September 2026: ("estimated pulse wave velocity"[tiab] OR ePWV[tiab]) AND mediat*[tiab]. The 37 records returned were screened by title and abstract and classified as main-group studies (ePWV as the proposed mediator; 18), secondary studies (ePWV as exposure; 9), or other (10). For main-group studies: abstraction of data source, sample size, design, exposure, outcome, proportion mediated, and age handling, verified from PubMed Central full texts and supplementary files retrieved through Europe PMC where available. Seven studies were classified from abstracts. One investigator screened and abstracted all records; the record-level screening output is archived. Ethics Approved by the Central Ethics Committee of the National Center for Cardiovascular Diseases and the Ethics Committee of Wenzhou People’s Hospital (approval No. KY-202605-131). All participants gave written informed consent.
Source: Open Science Framework
Author: Zhangke Tang
Language: English
Subfield: Cardiology and Cardiovascular Medicine
Field: Medicine
Domain: Health Sciences
Keywords: estimated glomerular filtration rate, carotid plaque, pulse wave velocity, mediation analysis, age as confounder, sex adjustment, CKD-EPI equation, cardiovascular risk, Wenzhou, arterial stiffness, renal function, cardiovascular screening, sex differences in cardiovascular risk, dose-response relationship
Open Access status: green
OpenAlex ID: W7222831385
