Table 3

Daily time-series studies with temperature exposures and myocardial infarction (MI) outcomes: study results

First author and yearTemperature variable*, range (if reported) (°C)Relative risk (RR) or rate ratio for temperature (95% CI if reported)(Change in temperature variable to which RR refers)Subgroup to which RR refers (if applicable)Lag for temp effect (days)Comment
Studies of fatal and non-fatal events
Gerber 200614tmax −29 to 390.93 (0.73 to 1.14)>30°C0
    1.00 (reference)18–30°C
0.97 (0.89 to 1.06)0–17°C
1.03 (0.92 to 1.14)<0°C
Wang 200615tmean −0.9 to 32.71.00 (reference)>20°C0
    1.12 (0.99 to 1.27)10–20°C
1.17 (1.01 to 1.35)<10°C
Enquselassie 199316tmax <16 to >301.2 (0.9 to 1.5)*>30°CFatal MIs0
1.0 (reference)23–27°CFatal MIs
1.4 (1.1 to 1.8)*⩽16°CFatal MIs
1.2 (0.9 to 1.4)*>30°CNon-fatal MIs
1.0 (reference)23–27°CNon-fatal MIs
1.0 (0.8 to1.3)*⩽16°CNon-fatal MIs
Ohlson 199117twind <−20 to >01.00 (reference)⩾0°C0
    1.09 (0.82 to 1.44)−10 to −1°C
1.10 (0.79 to 1.52)−20 to −11°C
1.12 (0.67 to 1.85)<−20°C
Barnett 200518tmean 1.5 to 23.01.008 (1.004 to 1.012)Per 1°C decreaseAverage 0–3Only a linear temperature effect appears to have been considered
Morabito 200519tmean, discomfort hours1.03 (p<0.01)Per extra two heat discomfort hoursSummer, men, <65 years0Only linear effects of the exposure variables appear to have been considered. However, discomfort hours analysis was performed separately for winter and summer
    1.06 (p<0.01)Per extra two cold discomfort hoursWinter, women, <65 years
Ebi 200420tmin, tmax −1 to 37 (tmin)1.072 (1.057 to 1.086)Per 3°C decrease (tmin)Men, 55–69 years, LA7Only linear temperature effects appears to have been considered. Effects for age ⩾70 years and for women were all in the same direction as those shown here
    1.008 (1.004 to 1.012)Per 3°C increase (tmin)Men, 55–69 years, SF
1.223 (1.083 to 1.381)Per 3°C increase (tmin)Men, 55–69 years, Sa
1.025 (1.000 to 1.052)Per 3°C increase (tmax)Men, 55–69 years, LA
1.066 (1.023 to 1.111)Per 3°C decrease (tmax)Men, 55–69 years, SF
1.109 (1.051 to 1.169)Per 3°C decrease (tmax)Men, 55–69 years, Sa
Koken 200321tmax 16.7 to 30.81.175 (1.029 to 1.343)Per 5.9°C increase⩾65 years0Analysis was for July/August only
Messner 200222temp, tchange −38 to 30 (temp)1.001 (0.993 to 1.008)Per 1°C decrease (temp)Fatal MIs0Only linear temperature effects appears to have been considered
    1.000 (0.997 to 1.003)Per 1°C increase (temp)Non-fatal MIs|
1.003 (0.979 to 1.028)Per 1°C increase (tchange)Fatal MIs
1.015 (1.003 to 1.027)Per 1°C increase (tchange)Non-fatal MIs
Danet 199923tmean −15 to 281.05 (1.02 to 1.09)Per 5°C decrease0Only linear temperature effects appears to have been considered
Fatal events only
Dilaveris 200624tmean 1 to 391.13†1.00 (reference)30°C23.3°CMean of last 7 daysResults are from the regression line, which shows minimum event rate at 23.3°C, with the event rate increasing smoothly above and below this temperature (levelling off at mean temperatures below 10°C)
    1.40† (p<0.001)10°C
Sharovsky 200425tmean 11 to 271.11 (1.06 to 1.16)†1.00 (reference)23.8–27.3°C21.6–22.6°C2-Day average
1.16 (1.05 to 1.27)†16.4–17.3°C
1.17 (1.07 to 1.28)†15.2–16.4°C
1.31 (1.19 to 1.44)†11.0–15.2°C
Rossi 199926tmean −6 to 321.44 (1.10 to 1.90)>27°C1Effect of colder temperatures is not described
    1.00 (reference)14°C
  • Percentage changes were converted to relative risk by dividing by 100 and adding one.

  • *tmax, daily maximum temperature; tmin, daily minimum temperature; tmean, daily mean temperature; temp, temperature (unspecified); discomfort, number of discomfort hours per day; tchange, temperature change from previous day; twind, windchill-adjusted temperature; †approximate RR, derived from graphical presentation of results.

  • LA, Los Angeles region; Sa, Sacramento region; SF, San Francisco region.