Skip to main content
. Author manuscript; available in PMC: 2018 Nov 28.
Published in final edited form as: Circulation. 2017 Oct 23;136(22):2100–2116. doi: 10.1161/CIRCULATIONAHA.117.028753

Table 1.

Study Characteristics

Study Description of Study
Sample
No. * Age,
Median
(Range),
y
Women,
No. (%)
Cardio-
vascular
Disease,
No. (%)
Endogenous
Subclinical
Hypo-
thyroidism,
No. (%) §
Medication Users during
Follow-up, No. (%)
Follow-up #

Thyroxine Anti-
thyroid
medication
Start, y Duration,
Median
(IQR), y
Person-
Years
AF
events,
No. (%)
United States
Cardiovascular Health Study13 Community-dwelling adults with Medicare eligibility in 4 US communities 3,328 73 (64–98) 1917 (57.6) 845 (25.4) 436 (13.1) 308 (9.5) NA 1994–1995 11.7 (7.0–18.1) 32,632 886 (26.6)
Health ABC Study25 Community-dwelling adults with Medicare eligibility in 2 US communities 2,346 74 (69–81) 1143 (48.7) 625 (27.1) 270 (11.5) 114 (4.9) 2 (0.1) 1997 8.1 (7.4–8.3) 16,155 201 (8.6)
Osteoporotic Fractures in Men (MrOS) Study26 Community-dwelling men aged 65 years or older in 6 US clinical centers 678 72 (65–91) 0 135 (19.9) 45 (6.6) 18 (2.7) 0 2000–2002 12.6 (11.2–13.1) 7,668 62 (9.1)
Europe
Bari Study27 Outpatients with heart failure followed up by Cardiology Department in Bari, Italy 268 65 (21–92) 55 (20.5) 103 (38.4) 23 (8.6) 22 (8.2) 8 (3.0) 2006–2008 1.3 (0.6–1.9) 339 14 (5.2)
Leiden 85-plus Study28 All adults aged 85 years living in Leiden, the Netherlands 432 85 (85–85) 281 (65.1%) 166 (38.4) 27 (6.3) 5 (1.2) 3 (0.7) 1997–1999 5.5 (2.7–9.0) 1,575 44 (10.2)
SHIP29** Adults living in Western Pomerania, Germany 2,339 45 (20–85) 1191 (50.9) 100 (4.3) 12 (0.5) 172 (7.4) 8 (0.3) 1997–2001 11.5 (11.1–12.1) 22,006 40 (1.7)
InChianti Study30 Community-dwelling adults aged 65 years or older living in Tuscany, Italy 1,051 71 (21–103) 581 (55.3) 123 (11.7) 26 (2.5) 17 (1.6) 2 (0.2) 1998 9.0 (8.3–9.2) 8,453 14 (1.3)
Rotterdam Study16 Inhabitants of Ommoord (The Netherlands) aged ≥ 55 years 1,607 68 (55–93) 975 (60.7) 412 (25.6) 91 (5.7) NA NA 1990–1993 15.5 (11.4–16.9) 20,892 226 (14.1)
PROSPER Study31 Community-dwelling elderly with high cardiovascular risk in The Netherlands, Scotland and Ireland 5,334 74 (69–83) 2,645 (49.6) 2,356 (44.2) 384 (7.2) 57 (1.1) 11 (0.2) 1997–1999 3.3 (3.0–3.5) 16,529 496 (9.3)
EPIC-Norfolk Study32 Adults aged 40 to 79 years living in Norfolk, England 11,642 58 (39–78) 6,181 (53.1) 442 (3.8) 607 (5.2) NA NA 1995–1998 17.0 (16.1–18.0) 137,861 575 (4.9)
Australia
Busselton Health Study33 Adults living in Busselton, Western Australia 1,060 46 (18–81) 539 (50.9) 54 (5.1) 37 (3.5) 20 (1.9) 1 (0.1) 1981 14.0 (14.0–14.0) 14,840 16 (1.5)

Overall 11 Cohorts 30,085 69 (18–103) 15,508 (51.6) 5,371 (17.9) 1,958 (6.5) 733 (2.4) 35 (0.1) 1981–2008 16.6 (10.7–18.7) 278,955 2,574 (8.6)

Studies where IPD were not available
Framingham Heart Study9 Adults aged ≥ 60 years from Framingham, USA 1,759†† ≥ 60 (NA) 1037 (59.0) NA 183 (10.4) NA NA 1978–1980 10.0 (NA) NA 156 (8.9)
Rotterdam Study Cohorts I, II and III24‡‡ Adults aged ≥ 55 years for Cohort II and ≥ 45 years for Cohort III from Ommoord, The Netherlands 8,740 63 (45–105) 5010 (57.3) NA 722 (8.3) NA NA 2000–2001 Cohort II, 2006–2008 Cohort III 6.8 (3.9–10.9) 61,935 403 (4.6)

Abbreviations: AF, atrial fibrillation; EPIC, European Prospective Investigation of Cancer; Health ABC Study, Health Aging and Body Composition Study; InChianti, Invecchiare in Chianti; IPD, individual participant data; IQR, Interquartile Range (25th–75th percentiles); NA, Data not Available; PROSPER, Prospective Study of Pravastatin in the Elderly at Risk; SHIP, Study of Health in Pomerania; TSH, thyroid stimulating hormone; y, Years

*

We excluded from our analyses participants with prevalent atrial fibrillation at baseline, missing outcomes for atrial fibrillation, subclinical hyperthyroidism and overt thyroid dysfunction, and intake of thyroxine or antithyroid medication at baseline.

We excluded participants younger than 18 years.

Cardiovascular disease at baseline was defined as known history of stroke, transient ischemic attack, myocardial infarction, angina pectoris, coronary angioplasty, or bypass surgery.

§

We used a common definition of subclinical hypothyroidism, TSH 4.5 mIU/L to 19.9mIU/L and normal free thyroxine level, but TSH cutoff values varied among the previous reports from each cohort, so numbers are different than in the original articles. To analyze only endogenous subclinical hypothyroidism, we excluded 253 participants in the Cardiovascular Health Study, 207 in the Health ABC Study, 43 in the Osteoporotic Fractures in Men Study, 15 in the Bari study, 12 in the Leiden 85+ study, 107 in the Study of Health in Pomerania, 21 in the Invecchiare in Chianti Study, 26 in the Rotterdam study, 188 in the PROSPER study, 301 in the EPIC-Norfolk study, and 4 in the Busselton Health study because they used thyroid medication at baseline.

We had no data on thyroid medication use during follow-up for 481 participants in the Study of Health in Pomerania, and all participants in the EPIC-Norfolk and the Rotterdam study. 91 participants in the Cardiovascular Health Study did not have information on thyroxin during follow-up, and information on antithyroid medication during follow-up was missing for all patients of the Cardiovascular Health Study. 5 persons took both thyroxine and anti-thyroid medication during the course of follow-up.

#

For all cohorts, we used the maximal follow-up data that were available, which may differ from previous reports of some cohorts. For the Cardiovascular Health Study, we set the baseline for our analysis to the year 5 visit of the original cohort because free thyroxine was measured at the year 5 visit.

**

SHIP includes participants from Pomerania, where an iodine supplementation program began in the mid-1990s. This shifted the distribution of TSH values towards the left in its first years, which lowered TSH values in the population of the SHIP Study during baseline examinations in 1997–2001.

††

Number of participants with euthyroidism and subclinical hypothyroidism. Participants with subclinical hyperthyroidism are not listed here, since they were not included in our sensitivity analysis or in the aggregate data from this cohort.

‡‡

Data on characteristics of 8740 participants included in the longitudinal analysis by Chaker et al.24 was obtained through contact with the authors. Individual participant data of 1,602 participants was available for Rotterdam Study Cohort I (see above).

HHS Vulnerability Disclosure