GoKawiilStudy population
The genetic discovery analysis of TG:HDL ratio encompassed multiple cohorts with a cumulative sample size of 1,032,116 participants. Study participants were recruited from various population-based and hospital-based health systems across three continents (Fig. 1 and Supplementary Table 1). These cohorts included: the UK Biobank population-based study (UKB, n = 409,602)60; the Geisinger Health System MyCode cohort (GHS, n = 143,129)61; the Mexico City Prospective Study (MCPS, n = 136,725), a study of residents in two districts of Mexico City (Coyoacán and Iztapalapa)62 with a genetically admixed population with components of Indigenous American, European and African ancestry as described in detail previously63; the Mayo Clinic Project Generation health system-based cohort (MAYO-RGC, n = 80,941)64; the population-based BangladEsh Longitudinal Investigation of Emerging Vascular and nonvascular Events (BELIEVE, n = 69,663)65; the Colorado Center for Personalized Medicine biobank, a health system-based cohort (CCPM, n = 56,173)66; the UCLA ATLAS Precision Health Biobank, a health system-based cohort (ATLAS, n = 45,533)67; the Mount Sinai BioMe BioBank, a health system-based cohort (BioMe, n = 45,440)68; the University of Pennsylvania Penn Medicine BioBank, a health system-based cohort (PMBB, n = 25,785)69; the population-based Dallas Biobank Study at UT Southwestern Medical Center (DBS, n = 13,979)70; and the population-based Malmö Diet and Cancer Study (MDCS, n = 5,146)71. Analyses of other cardiometabolic phenotypes included participants from the aforementioned cohorts and also 4,755 participants from the Indiana University School of Medicine (Indiana-CLDB) study and 53,325 participants from the South Asia Biobank72. All studies received approval from the relevant ethics committees as listed below, and the participants gave their informed consent to take part in the research. UKB cohort: ethical approval for the UKB was obtained from the North West Centre for Research Ethics Committee (11/NW/0382). The work described here was approved by the UKB under application number 26041. GHS cohort: The MyCode Community Health Initiative was approved by the Geisinger institutional review board (IRB; study 2006-0258). MCPS cohort: the study was approved by scientific and ethics committees within the Mexican National Council of Science and Technology (0595 P-M), the Mexican Ministry of Health and the Central Oxford Research Ethics Committee (C99.260). MAYO-RGC cohort: the study was approved by the Mayo Clinic IRB under protocol number 09-007763. BELIEVE cohort: BELIEVE has received approvals from the relevant institutional review boards of the Bangladesh Medical Research Council, the National Heart Foundation Hospital and Research Institute, icddr,b and Bangabandhu Sheikh Mujib Medical University (BMRC/NREC/2013- 2016/390; BMRC/NREC/2016-2019/243; BSMMU/2019/1184; BSMMU/2019/1185; PR-18051; HBREC.2019.09). CCPM cohort: biospecimens and associated data used in this study were obtained from the biobank at the Colorado Center for Personalized Medicine (CCPM) at the University of Colorado Anschutz Medical Campus (CU AMC). All samples and data were collected under IRB approved protocol (15-0461). ATLAS cohort: ATLAS is an approved study by the University of California Los Angeles (UCLA) IRB (UCLA IRB 17-001013). BioMe/Mount Sinai Million Health Discoveries Program: The Icahn School of Medicine at Mount Sinai’s IRB, Program for the Protection of Human Subjects (PPHS), approved the BioMe Biobank and Mount Sinai Million Health Discoveries Program (PPHS IRB 11-01139 and 21-01743). PMBB cohort: the PMBB is approved by the University of Pennsylvania IRB under protocol number 813913. DBS cohort: The University of Texas Southwestern (UTSW) Biobank was reviewed and approved by the UTSW IRB under the study STU 022011-116: “Genetic Variation in a Multi-Ethnic Population”. MDCS cohort: IRB approval for MDCS was obtained from the Ethics committee of Lund University (LU 51-90) and the Regional Board of Ethics in Lund (Dnr 2016/479). Indiana-CLDB: the study has been reviewed by Indiana University IRB and approved under the protocol number 1105005445. South Asia Biobank: the study received ethical approval from the institutional Research Ethics Committee (18IC4698).
Clinical phenotypes
Clinical biomarkers
Measurements of blood lipids, glycaemic markers, transaminases, CRP and blood pressure were obtained either through direct clinical assessments or electronic health record (EHR) extraction, depending on the cohort type. For participants in population-based cohort studies, these measurements were conducted during standardized visits to designated assessment centres. By contrast, for participants from health-system-based cohorts, data were extracted retrospectively from EHR datasets and participant-specific median values were used. Blood lipid levels were primarily assessed using routine clinical chemistry methods across most cohorts. In the MCPS and BELIEVE studies, lipid measurements were derived from nuclear magnetic resonance metabolomics profiling, performed using the Nightingale Health platform73. Blood lipids were corrected for medication use by dividing by a class-specific correction factor obtained from previously reported clinical trials19,74,75,76,77,78. The TG:HDL ratio was calculated using triglyceride and HDL cholesterol values obtained from the same blood draw at baseline in epidemiological cohort studies (5 cohorts and 635,115 participants). In EHR-based studies (6 cohorts and 397,001 participants), where multiple measurements were available, median values across clinical encounters were calculated for TG and HDL separately and then the ratio of the median values was calculated. The use of median values across encounters is a standard approach in EHR-based genetic research to define a patient’s typical lipid profile79,80,81. Importantly, this avoids the exclusion of participants with valid but non-concurrent lipid measurements, thereby preserving sample size and discovery power. A sensitivity analysis was conducted restricting the TG:HDL ratio to values derived from TG and HDL measured concurrently from the same blood draw. A stratified analysis by fasting status was also conducted in the UKB, comparing TG:HDL ratios derived from fasting samples (defined as at least 6 h since the last meal) with those from non-fasting samples (defined as at most 2 h since the last meal). Blood pressure levels were corrected by adding 15 mmHg to systolic and 10 mmHg to diastolic blood pressure in those individuals receiving anti-hypertensive treatment as previously reported82.
Liver biopsy
The bariatric surgery cohort at Geisinger Health System (GHS) comprises 3,599 participants of European ancestry who underwent bariatric surgery procedures61. Intraoperative wedge liver biopsies were systematically obtained 10 cm left of the falciform ligament before any manipulation of the liver or stomach. Biopsies were sectioned, with the principal portion allocated for clinical histopathology—fixed in 10% neutral buffered formalin and stained with haematoxylin and eosin for general assessment and Masson’s trichrome for fibrosis evaluation. The remaining tissue was preserved in a research biobank, either stabilized in the RNAlater tissue collection system (Thermo Fisher Scientific) or snap-frozen in liquid nitrogen. Histopathological evaluation was conducted by an experienced pathologist and independently reviewed by a second pathologist. Scoring followed the Nonalcoholic Steatohepatitis Clinical Research Network (NASH CRN) system83, with steatosis graded as 0 (<5% parenchymal involvement), 1 (5 to <34%), 2 (34 to <67%) or 3 (>67%); lobular inflammation as 0 (none), 1 (mild, <2 foci per ×200 field), 2 (moderate, 2–4 foci per ×200 field) or 3 (severe, >4 foci per ×200 field); hepatocyte ballooning as 0 (none), 1 (few ballooned cells), or 2 (many/prominent ballooned cells); and fibrosis staged from 0 (none) to 4 (cirrhosis) according to established criteria. The MASLD activity score was generated by summing of the steatosis, lobular inflammation and ballooning scores.
Imaging phenotypes
MRI
MRI images were obtained from the UKB imaging cohort. For quantification of liver fat, we used data from a two-dimensional abdominal MRI scan that included the liver; a subset of individuals was scanned using a Dixon gradient echo protocol, while those imaged from 2016 onward were scanned using the IDEAL (iterative decomposition of water and fat with echo asymmetry and least-squares estimation) protocol. The resulting data have an in-plane pixel size of 2.5 × 2.5 mm and a slice thickness of 6 mm.
... continue reading