Setting | Study period | Mortality outcome | Daily temperature exposure | Lag structure of temperature | Heat threshold °C | % change in mortality per 1°C temperature increase (95% CI) | Notes: | |
Baccini et al32 | Athens, Greece Barcelona, Spain Budapest, Hungary Dublin, Ireland Helsinki, Finland Ljubljana, Slovenia London, UK Milan, Italy Paris, France Prague, Czech Rep Rome, Italy Stockholm, Sweden Turin, Italy Valencia, Spain Zurich, Switzerland | 1992–96 1992–2000 1992–2001 1990–2000 1990–2000 1992–99 1992–2000 1991–98 1992–2000 1992–2000 1992–2000 1990–2000 1991–99 1995–2000 1990–96 | All-cause (exc. ext.) + + + + + + + + + + + + + + | Max app. Mean app. Max app. + + + + + + + + + + + + | Lags 0–3 days + + + + + + + + + + + + + + | 32.7 22.4 22.8 23.9 23.6 21.5 23.9 31.8 24.1 22.0 30.3 21.7 27.0 28.2 21.8 | 5.54 (4.30 to 6.80) 1.56 (1.04 to 2.08) 1.74 (1.47 to 2.02) −0.02 (−5.38 to 5.65) 3.72 (1.68 to 5.81) 1.34 (0.32 to 2.37) 1.54 (1.01 to 2.08) 4.29 (3.35 to 5.24) 2.44 (2.08 to 2.80) 1.91 (1.39 to 2.44) 5.25 (4.57 to 5.93) 1.17 (0.41 to 1.94) 3.32 (2.53 to 4.13) 0.56 (−0.35 to 1.47) 1.37 (0.49 to 2.25) | |
Keatinge et al26 | Netherlands Baden-Wurttemberg, Germany | 1988–92 + | All-cause (65–74 years) + | Mean + | Lag 0 + | 17.3 to 20.3 19.0 to 22.0 | 1.3 (0.4, 2.3) 1.4 (0.6, 2.2) | - Minimum mortality temperature given in 3° bands. Mid-point value assumed for threshold in meta-regression - No indication of seasonal adjustment |
Curriero et al27 28 | Boston, USA Chicago, USA New York, USA Philadelphia, USA Baltimore, USA Washington, USA Charlotte, USA Atlanta, USA Jacksonville, USA Tampa, USA Miami, USA | 1973–94 + + + + + + + + + + | All-cause (exc. ext.) + + + + + + + + + + | Mean + + + + + + + + + + | Lag 0 + + + + + + + + + + | 20.9 18.4 19.1 21.4 21.4 21.4 32.4 24.6 24.9 27.1 27.2 | 2.93 1.28 2.61 2.81 2.57 1.37 N/A 2.98 1.91 1.55 2.31 | |
Zanobetti et al31 | Birmingham, USA Dallas, USA Detroit, USA Houston, USA Minneapolis, USA Phoenix, USA | 1999–2000 1999–2002 + + + 1999–2000 | All-cause (exc. ext.) + + + + + | Mean app. + + + + + | Lag 0 + + + + + | 0.36 (−0.48 to 1.16) −0.08 (−0.54 to 0.36) 0.60 (0.34 to 0.90) 0.32 (−0.12 to 0.74) 0.58 (0.16 to 1.00) 0.64 (−0.04 to 1.32) | - Summer months only - Heat thresholds for meta-regression taken as minimum values within summer months - Heat slopes approximated from graphical presentation | |
Basu et al30 | Contra Costa, USA Fresno, USA Kern, USA Los Angeles, USA Orange, USA Riverside, USA Sacramento, USA San Diego, USA Santa Clara, USA | May 1999–Sep 2003 + + + + + + + | All-cause (exc. ext.) + + + + + + + + | Mean app. + + + + + + + + | Lag 0 + + + + + + + + | 0.40 (−0.16 to 0.96) 0.16 (−0.60 to 0.92) 0.44 (−0.16 to 1.04) 0.88 (0.60 to 1.16) −0.08 (−0.76 to 0.60) 0.36 (−0.20 to 0.92) 0.56 (0.04 to 1.08) 0.24 (−0.20 to 0.68) 1.28 (0.60 to 1.96) | - Summer months only - Heat thresholds for meta-regression taken as minimum values within summer months - Heat slopes approximated from graphical presentation | |
Kim et al29 | Seoul, South Korea Daegu, South Korea Incheon, South Korea Gwangju, South Korea Daejeon, South Korea Busan, South Korea | 1994–2003 + + + + + | All-cause (exc. ext.) + + + + + | Mean + + + + + | Lag 0 + + + + + | 28.1 28.1 26.6 26.6 28.1 N/A | 9.60 (8.04 to 11.17) 4.62 (2.74 to 6.54) 5.31 (3.10 to 7.56) 2.64 (0.34 to 5.00) 3.91 (−0.63 to 8.66) N/A | |
McMichael et al33 | Bucharest, Romania Sofia, Bulgaria New Delhi, India Monterrey, Mexico Mexico City, Mexico Chiang Mai, Thailand Bangkok, Thailand Salvador, Brazil Sao Paulo, Brazil Santiago, Chile Cape Town, South Africa | 1994–97 1996–99 1991–94 1996–99 1994–98 1995–97 1991–92 1996–99 1991–94 1988–91 1996–99 | All-cause (exc. ext.) + + + + + + + + + + | Mean + + + + + + + + + + | Lags 0–1 days + + + + + + + + + + | 22 16 29 31 18 28 29 23 23 16 17 | 3.30 (2.35 to 4.26) 2.88 (2.11 to 3.65) 3.94 (2.80 to 5.08) 18.8 (13.0 to 25.0) 0.77 (0.14 to 1.39) 2.39 (−0.49 to 5.35) 5.78 (3.52 to 8.09) 2.48 (0.93 to 4.05) 3.46 (2.62 to 4.31) 1.04 (0.28 to 1.81) 0.47 (−0.31 to 1.24) | |
Hales et al35 | Christchurch, New Zealand | June 1988 –Dec 1993 | All-cause | Max | Lag 0 | 20.5 | 1.0 (0.4 to 2.1) | |
Hu et al34 | Sydney, Australia | 1994–2004 | All-cause | Max | Lag 0 | 0.9 (0.6 to 1.3) | - Summer months only - Heat threshold for meta-regression taken as minimum value within summer months | |
Kan et al36 | Shanghai, China | June 2000–Dec 2001 | All-cause (exc. ext.) | Mean | Lags 0–2 days | 26.7 | 0.73 | - Slope is 0.73% in abstract but 1.21% in full paper - Slope of 0.73% assumed based on figure |
El-zein et al41 | Beirut, Lebanon | 1997–99 | All-cause | Mean | Lag 0 | 27.5 | 12.3 (5.7 to 19.4) |
‘+’, As above; app, apparent temperature; exc. ext, excluding deaths from external causes; Max, maximum.