Heterogeneous associations of polygenic indices of 35 traits with mortality: a register-linked population-based follow-up study

  1. Hannu Lahtinen  Is a corresponding author
  2. Jaakko Kaprio
  3. Andrea Ganna
  4. Kaarina Korhonen
  5. Stefano Lombardi
  6. Karri Silventoinen
  7. Pekka Martikainen
  1. Helsinki Institute for Demography and Population Health, University of Helsinki, Finland
  2. Max Planck – University of Helsinki Center for Social Inequalities in Population Health, Finland
  3. Institute for Molecular Medicine FIMM, University of Helsinki, Finland
  4. FIMM, University of Helsinki, Finland
  5. VATT Institute for Economic Research, Finland
  6. Institute of Labor Economics (IZA), Germany
2 figures, 2 tables and 2 additional files

Figures

Hazard ratios of polygenic indices for all-cause mortality.

Population (N=40,097 individuals) and within-sibship (N=10,174 individuals in 5071 sibships) estimates. Estimates from Cox proportional hazards models adjusted for indicators for the baseline year, sex and 10 first principal components of the genome. Capped bars are 95% confidence intervals. For a table of corresponding estimates, see Supplementary file 1E. Abbreviations: HR = Hazard ratio; ADHD = Attention deficit hyperactivity disorder.

Hazard ratios of polygenic indices for all-cause mortality by sex (Panel A), educational group (Panel B), age-specific mortality follow-up period (Panel C), and cause of death (Panel D).

Estimates from Cox proportional hazards models adjusted for indicators for the baseline year, sex and 10 first principal components of the genome. Capped bars are 95% confidence intervals. N=40,097 individuals. For a table of corresponding estimates, see Supplementary file 1F. Abbreviations: HR = Hazard ratio; ADHD = Attention deficit hyperactivity disorder.

Tables

Table 1
Hazard ratios of selected polygenic indices and corresponding phenotypes for all-cause mortality (N: 37,548 individuals; 5407 deaths).

All presented variables were standardised to standard-deviation units. For corresponding models with categorical phenotypes, see Supplementary file 1G.

Model 1 a–1 lModel 2 a–2 fModel 3 a & 3bModel 4
HR95% CIHR95% CIHR95% CIHR95% CI
lowerupperlowerupperlowerupperlowerupper
Smoking1.411.371.451.391.351.441.351.311.401.341.291.38
PGI-ever smoking1.121.091.151.071.041.101.071.041.101.041.011.08
BMI1.071.041.101.051.011.081.020.991.051.000.971.04
PGI-BMI1.101.071.131.081.051.111.051.021.091.031.001.06
Depression indicators1.161.141.191.161.131.191.061.031.081.051.031.08
PGI-depressive symptoms1.071.041.101.061.031.091.020.991.051.000.971.03
Alcohol intake1.161.131.191.151.121.181.121.091.141.111.081.14
PGI drinks per week1.061.031.091.051.021.081.051.021.081.031.001.06
Education years0.860.830.880.870.840.900.900.880.930.910.880.94
PGI educational attainment0.910.890.940.940.920.970.960.930.991.020.991.05
Self-rated health0.740.720.760.750.730.770.770.750.800.780.750.80
PGI self-rated health0.900.880.930.940.910.960.950.920.990.980.941.01
  1. All models adjusted for baseline covariates: indicators of the baseline year, sex, and ten first principal components of the genome.

  2. Models 1: Baseline covariates + each phenotype/PGI separately.

  3. Models 2: PGI and corresponding phenotype mutually adjusted.

  4. Model 3 a: Phenotypes mutually adjusted, Model 3b: PGIs mutually adjusted.

  5. Model 4: Full model.

  6. HR=Hazard ratio.

  7. 95% CI=95% confidence interval.

  8. PGI=Polygenic index.

  9. BMI=Body mass index.

Table 2
Hazard ratios of selected polygenic indices for all-cause mortality regarding belonging to a risk category in the related phenotypes.

‘Without phenotype’ categories consist of: Never smokers (N: 21,189), BMI 18.5–24.9 (N:15,092), no depression indicators (N: 31,089), non-alcohol drinkers (N: 5498), higher-tertiary-degree holders (N: 4765), the best self-rated health (5-point scale, N: 7566). ‘With phenotype’ categories are those in the population sample with valid information on the phenotype in question excluding the ‘without risk factor’ category.

Without phenotypeWith phenotype
HR95% CIHR95% CI
lowerupperlowerupper
PGI-ever smoking1.071.031.111.091.051.12
PGI-BMI1.131.081.191.091.061.13
PGI-depressive symptoms1.051.021.081.061.011.12
PGI-drinks per week1.091.041.151.071.031.10
PGI-educational attainment0.880.761.020.930.900.96
PGI-self-rated health0.880.800.960.910.890.94
  1. All models adjusted for indicators of the baseline year, sex, and ten first principal components of the genome.

  2. HR=Hazard ratio.

  3. 95% CI=95% confidence interval.

  4. PGI=Polygenic index.

  5. BMI=Body mass index.

Additional files

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Hannu Lahtinen
  2. Jaakko Kaprio
  3. Andrea Ganna
  4. Kaarina Korhonen
  5. Stefano Lombardi
  6. Karri Silventoinen
  7. Pekka Martikainen
(2026)
Heterogeneous associations of polygenic indices of 35 traits with mortality: a register-linked population-based follow-up study
eLife 14:RP107496.
https://doi.org/10.7554/eLife.107496.3