Skip to main content
. 2024 Jun 5;42:100943. doi: 10.1016/j.lanepe.2024.100943

Table 4.

Risk of incident type 2 diabetes predicted by polygenic risk score only, and by polygenic risk score plus light exposure.

Percentile PRS only
PRS + light
HR [95% CI] p-value HR [95% CI] p-value
Model 1
 Polygenic risk 0–25% (ref.)
25–50% 1.69 [1.44–1.98]a <0.0001 1.69 [1.44–1.98]a <0.0001
50–75% 2.55 [2.20–2.97]a <0.0001 2.54 [2.18–2.95]a <0.0001
75–100% 4.68 [4.03–5.43]a <0.0001 4.66 [4.02–5.40]a <0.0001
 Night light 0–50% (ref.)
50–70% 1.35 [1.20–1.53]a <0.0001
70–90% 1.41 [1.25–1.60]a <0.0001
90–100% 1.73 [1.50–2.00]a <0.0001
 Day light 0–50% (ref.)
50–70% 0.98 [0.87–1.11] 0.74
70–90% 0.88 [0.78–1.00]a 0.048
90–100% 0.74 [0.63–0.87]a 0.00033
Model 2
 Polygenic risk 0–25% (ref.)
25–50% 1.66 [1.41–1.95]a <0.0001 1.65 [1.41–1.94]a <0.0001
50–75% 2.48 [2.13–2.89]a <0.0001 2.46 [2.12–2.87]a <0.0001
75–100% 4.42 [3.80–5.13]a <0.0001 4.40 [3.78–5.11]a <0.0001
 Night light 0–50% (ref.)
50–70% 1.32 [1.16–1.49]a <0.0001
70–90% 1.40 [1.24–1.58]a <0.0001
90–100% 1.66 [1.43–1.92]a <0.0001
 Day light 0–50% (ref.)
50–70% 1.00 [0.89–1.13] 0.97
70–90% 0.92 [0.82–1.05] 0.21
90–100% 0.78 [0.66–0.92]a 0.0032
Model 3
 Polygenic risk 0–25% (ref.)
25–50% 1.59 [1.35–1.88]a <0.0001 1.59 [1.35–1.87]a <0.0001
50–75% 2.36 [2.02-2.75]a <0.0001 2.34 [2.01–2.73]a <0.0001
75–100% 4.17 [3.58–4.85]a <0.0001 4.14 [3.55–4.82]a <0.0001
 Night light 0–50% (ref.)
50–70% 1.31 [1.16–1.49]a <0.0001
70–90% 1.39 [1.23–1.58]a <0.0001
90–100% 1.58 [1.36–1.84]a <0.0001
 Day light 0–50% (ref.)
50–70% 1.08 [0.95–1.22] 0.25
70–90% 1.07 [0.94–1.21] 0.31
90–100% 0.98 [0.83–1.16] 0.82

Data are hazard ratios (95% CI). Exposures are percentiles of type 2 diabetes polygenic risk, night light and day light exposure. Model covariates: model 1: age, sex, ethnicity, and the top five principal components of genetic ancestry; model 2: model 1 covariates plus income, material deprivation, education, and employment status; and model 3: model 2 covariates plus smoking status, alcohol consumption, healthy diet, physical activity, and urbanicity. Models were analysed within a sub-sample of participants with European ancestry only, consisting of 94.6% of participants with complete light data and no type 2 diabetes diagnosis prior to light-tracking.

a

p < 0.05.