Skip to main content
. 2011 Jun 17;119(10):1409–1414. doi: 10.12989/ehp.1003026

Table 3.

IR% and 95% CI associated with IQR increases in different PM fractions by cause of death and the presence or absence of Saharan dust.

PM2.5 (IQR = 12.8 μg/m3) PM2.5–10 (IQR = 10.8 μg/m3) PM10 (IQR = 19.8 μg/m3)
Cause of death IR% (95% CI) p-Value IR% (95% CI) p-Value IR% (95% CI) p-Value
Natural causes (lag 0–2) (ICD-9 codes 1–799) 1.24 (–0.65 to 3.17) 2.96 (1.23 to 4.72) 3.04 (1.53 to 4.56)
Dust free 0.91 (–1.09 to 2.94) 3.33 (1.29 to 5.40) 3.01 (1.29 to 4.75)
Dust affected 3.22 (–1.13 to 7.75) 0.316 2.07 (–1.07 to 5.31) 0.502 3.20 (–0.04 to 6.55) 0.914
Cardiac diseases (lag 0–2) (ICD-9 codes 390–429) 1.38 (–1.68 to 4.55) 3.72 (0.78 to 6.73) 4.04 (1.49 to 6.65)
Dust free 0.97 (–2.24 to 4.29) 0.86 (–2.47 to 4.31) 2.09 (–0.76 to 5.02)
Dust affected 1.37 (–5.92 to 9.21) 0.920 9.73 (4.25 to 15.49) 0.005 9.55 (3.81 to 15.61) 0.019
Cerebrovascular diseases (lag 0) (ICD-9 codes 430–438) –0.32 (–4.26 to 3.78) 5.41 (1.74 to 9.22) 2.64 (–1.26 to 6.68)
Dust free –0.07 (–4.29 to 4.33) 4.58 (–0.18 to 9.58) 3.24 (–1.78 to 8.51)
Dust affected –3.22 (–11.72 to 6.10) 0.514 7.03 (1.22 to 13.18) 0.526 1.71 (–5.05 to 8.95) 0.720
Diseases of the circulatory system (lag 0–2) (ICD-9 codes 390–459) 1.23 (–1.42 to 3.95) 4.06 (1.50 to 6.69) 2.99 (0.82 to 5.19)
Dust free 1.19 (–1.60 to 4.05) 2.21 (–0.74 to 5.25) 1.82 (–0.61 to 4.32)
Dust affected –0.91 (–7.04 to 5.62) 0.531 7.93 (3.20 to 12.88) 0.039 5.91 (1.02 to 11.03) 0.134
Diseases of the respiratory system (lag 0–5) (ICD-9 codes 460–519) 0.25 (–9.90 to 11.54) 12.65 (1.18 to 25.42) 4.97 (–2.18 to 12.63)
Dust free –2.17 (–12.72 to 9.66) 8.67 (–4.14 to 23.19) 5.00 (–2.80 to 13.43)
Dust affected 8.38 (–12.66 to 34.49) 0.368 19.43 (0.34 to 42.15) 0.349 2.66 (–12.78 to 20.83) 0.798
The p-value is for the interaction between PM and the Saharan dust indicator. PM2.5 and PM2.5–10 data are from the Italian NIH monitoring station, and PM10 data are from regional EPA monitoring stations. Results from multivariate Poisson regression models are adjusted for time trend, seasonality, day of the week, summer population decrease, holidays, influenza epidemics, high apparent temperatures (penalized spline, lag 0–1), low air temperatures (penalized spline, lag 1–6), and barometric pressure (penalized spline, lag 0).