General characteristics of the study participants are shown in Table 1. Among the 25,355 participants, 9.7% (n = 2467) were categorized as having platelet elevation. There were significant effects on platelet elevation of age, gender, race, education level, status of smoking and alcohol drinking, indoor decoration, activity per week, WBC counts and having hypertension, heart disease or hyperlipemia.
Table 1
General characteristics of the study participants.
Characteristics | Platelet counts | P-value |
≤P90a (n = 22,888) | >P90b (n = 2467) | |
Age, years | 54.2 (54.1,54.4)c | 51 (50.6,51.5) | < 0.0001d |
Gender, ne (%) | Male | 8126 (35.5)f | 534 (21.7) | < 0.0001 g |
Female | 14761(64.5) | 1932(78.3) |
Race, n (%) | Han | 14571(63.7) | 1622(65.7) | 0.006 |
Manchu | 8287(36.2) | 837(33.9) |
Other | 30(0.1) | 8(0.3) |
Income (10,000 Yuan) | 31.7(13.7,49.8) | 5.8(5.6,6) | 0.355 |
BMI, kg/m2 | 31(25.6,36.4) | 28.4(24.6,32.2) | 0.763 |
Education, n (%) | 0–6 years | 5551(24.3) | 497(20.1) | < 0.0001 |
7–12 years | 13602(59.4) | 1458(59.1) |
> 12 years | 3735(16.3) | 512(20.8) |
Smoke, n(%) | Current | 3597(15.7) | 337(13.7) | < 0.0001 |
Ever | 1016(4.4) | 64(2.6) |
Never | 18275(79.8) | 2066(83.7) |
Alcohol, n(%) | Current | 3996(17.5) | 361(14.6) | < 0.0001 |
Ever | 362(1.6) | 31(1.3) |
Never | 18530(81) | 2075(84.1) |
Decoration, n(%) | No | 18958(82.8) | 1934(78.5) | < 0.0001 |
Yes | 3925(17.2) | 531(21.5) |
Hypertension, n(%) | No | 14425(63) | 1640(66.5) | 0.001 |
Yes | 8463(37) | 827(33.5) |
Diabetes, n(%) | No | 19906(87) | 2161(87.6) | 0.38 |
Yes | 2982(13) | 306(12.4) |
Heart disease, n(%) | No | 21390(93.5) | 2329(94.4) | 0.068 |
Yes | 1498(6.5) | 138(5.6) |
Activity, n(%) | < 1 h/week | 10768(47) | 1186(48.1) | < 0.0001 |
1–2 h/week | 2341(10.2) | 316(12.8) |
> 2 h/week | 9779(42.7) | 965(39.1) |
Hyperlipemia, n(%) | No | 18534(81) | 1805(73.2) | < 0.0001 |
Yes | 4354(19) | 662(26.8) |
White blood cell count (109/ml) | 6.2(6.2,6.3) | 7.2(7.1,7.2) | < 0.0001 |
PM2.5 exposure (µg/m3) | 36.8(36.7,36.8) | 37.7(37.5,37.9) | < 0.0001 |
a, ≤ 90th percentile; b, > 90th percentile; c, least square means with 95% CI intervals – applies to all such values; d, P-values of variance analyses – applies to all such values; e, number of participants; f, total counts with percentages – applies to all such values; g, P-values of c2 tests – applies to all such values.
Average PM2.5 exposure of all participants was 37.25 µg/m3 with a significant difference between participants with elevated platelet counts and those without. Figure 1 shows the distributions of participant locations and variation in PM2.5 exposure across the whole study area.
The associations between PM2.5 exposure and 90th percentile platelet elevations remained stable after adjusting for all possible confounders (Table 2). When PM2.5 was treated as a continuous variable, the OR between every 1 µg/m3 increment of PM2.5 exposure and platelet elevation was 1.1 (95%CI: 1.08–1.12). When PM2.5 was treated as a categorical factor, compared with participants in the first quartile, those in the third (OR = 1.28, 95%CI: 1.08–1.52) and fourth (OR = 2.18, 95%CI: 1.83–2.60) quartiles were more likely to have elevated platelet counts.
Table 2
Effects of PM2.5 exposure on 90th percentile platelet elevation.
PM2.5 | OR (95%CI) |
Model 1 | P-value | Model 2 | P-value | Model 3 | P-value |
1 µg/m3 increment | 1.06 (1.05,1.08) | < 0.0001 | 1.11(1.09,1.12) | < 0.0001 | 1.1(1.08,1.12) | < 0.0001 |
1st quartile | Ref | | Ref | | Ref | |
2st quartile | 0.88(0.77,1.0) | 0.05 | 1.07(0.92,1.25) | 0.37 | 1.1(0.94,1.29) | 0.24 |
3rd quartile | 1.02(0.9,1.16) | 0.73 | 1.31(1.11,1.55) | 0.00 | 1.28(1.08,1.52) | 0.01 |
4st quartile | 1.84(1.64,2.05) | < 0.0001 | 2.21(1.88,2.59) | < 0.0001 | 2.18(1.83,2.6) | < 0.0001 |
P-value for trend | < 0.0001 | < 0.0001 | < 0.0001 |
Model 1, crude model; Model 2, adjusted for age, gender, race, education, income and BMI; Model 3, further adjusted for status of smoking and alcohol drinking, decoration in the previous five years, hypertension, diabetes, heart disease, hyperlipemia, activity per week and white blood cell counts.1st quartile, PM2.5 ≤ 25th percentile; 2nd quartile, 25th percentile < PM2.5 ≤50th percentile; 3rd quartile, 50th percentile < PM2.5 ≤75th percentile; 4th quartile, PM2.5 > 75th percentile.
Model 3 showed significant interactions of PM2.5 exposure with gender (P < 0.01), race (P < 0.0001) and diabetes status (P < 0.001). Figure 2 shows the stratified analysis results according to gender (male and female), race (Han and Manchu) and diabetes (yes and no). Males were more likely to have platelet elevation after long-term PM2.5 exposure compared with females. Those of Han ethnicity were more likely to have platelet elevation compared with Manchu ethnicity. Participants without diabetes were more likely to have platelet elevation after long-term PM2.5 exposure.
Sensitivity analysis results are shown in Supplemental Tables 1 and 2. Analysis results for the mixed linear model between long-term PM2.5 exposure and platelet counts showed that every 1 µg/m3 increment of PM2.5 exposure was associated with 0.29% (95%CI: 0.25–0.32%) increase in platelet counts; and effects of PM2.5 exposure were more evident in participants who were male, of Han ethnicity and without diabetes (Supplemental Table 1). Using the 75th percentile as the cut-off to define elevated platelet counts gave similar results (Supplemental Table 2) to those for the 90th percentile.