Skip to main content
. 2022 Dec 10;13:7652. doi: 10.1038/s41467-022-35157-w

Table 2.

Characteristics of eligible individuals when emulating a target trial of metformin vs. sulfonylurea initiators on the risk of death and dementia in the US RPDR (2007–2017) and the UK CPRD (2001–2017)

US RPDR cohort (N = 13,191) UK CPRD cohort (N = 108,025)
Patient characteristics Metformin initiators (n = 11,229) Sulfonylurea initiators (n = 1962) Metformin initiators (n = 94,208) Sulfonylurea initiators (n = 13,817)
Age at baseline (mean, years) 68.6 72.2 64.9 70.2
Sex (% male) 49.2 52.9 57.4 58.3
Race (percent)
    White 78.2 83.4 86.5 87.0
    Black or African American 7.6 5.4 3.6 3.0
    Asian 4.4 3.7 Data combined with Other
    Other 9.8 7.5 9.9 10.0
Ethnicity (% Hispanic) 4.4 3.3 Data not available
Year of first prescription (median) 2012 2012 2008 2003
SES/IMDa (% low income/most deprived) 3.1 1.8 17.3 17.5
Cancer (%) 30.2 28.7 9.9 12.4
CVDb (%) 42.8 44.4 53.8 50.8
Hypertension (%) 74.2 67.0 94.9 90.7
COPD (%) Data not available 4.7 4.6
Smokera (%) 17.0 18.7
Strokeb (%) 11.9 12.1 Data combined with CVD
Baseline HbA1C levela (percent)
    <7 (%) 51.8 46.2 16.4 11.8
    7–10 (%) 42.0 47.9 69.1 65.3
    >10 (%) 6.2 5.9 14.5 22.9
Baseline BMIa (kg/m2)
    <25 (%) 11.1 17.5 8.1 34.8
    25–30 (%) 29.5 36.0 32.9 43.5
    ≥30 (%) 59.4 46.5 59.0 21.7
All-cause mortality analysis outcomes
    Median follow-up time (years) 5.3 5.3 6.0 7.0
    Total person-years 63,060 11,047 592,948 103,777
    Deaths (n, %) 527 (4.7) 222 (11.3) 12,941 (13.7) 5173 (37.4)
Competing risk analysis outcomes
    Median follow-up time (years) 5.0 5.0 6.0 7.0
    Total person-years 60,683 10,508 592,072 103,209
    Deaths prior to dementia (n, %) 415 (3.7) 154 (7.8) 11,560 (12.3) 4,570 (33.1)
    Incident dementia cases (n, %) 869 (7.7) 241 (12.3) 5561 (5.9) 1699 (12.3)

BMI body mass index, COPD chronic obstructive pulmonary disease, CVD cardio-vascular disease, HbA1C glycosylated hemoglobin. IMD Index of Multiple Deprivation (official measure of relative deprivation by small geographic region in the UK).

aSES (US RPDR) and IMD (UK CPRD) are distinct, country-specific socioeconomic indicators, intended for comparison within—rather than across—the two cohorts. IMD, smoking status, HbA1C, and BMI had 7%, 2%, 21%, and 3% of missing values in the UK CPRD, respectively. The corresponding statistics presented in the table are valid percentages out of patients without missing information.

bStroke and CVD indicators were collapsed.