Table 1.
Characteristics of the studies included in the meta-analysis
| Authors | Study setting | Study period | Study design | Exposure assessment level | Data source | No. of participants | No. of cases | Exposure period | Exposure range (mean (IQR) μg/m3) | Quality scorea |
|---|---|---|---|---|---|---|---|---|---|---|
| Wilhelm et al. [39] | California, USA | 1999–2000 | Retrospective | Semi-individual level | Monitoring network data | 106,483 | 92,68 | TS | 21.0 (NA) | 8 |
| Huynh et al. [7] | California, USA | 1999–2000 | Retrospective | Semi-individual level | Monitoring network data | 42,692 | 10,673 | WP and TS | 18.0 (8.7) | 8 |
| Jalaludin et al. [33] | Sydney, Australia | 1998–2000 | Retrospective | Regional level and semi-individual level | Monitoring network data | 123,840 | 6011 | TS | 9.0 (4.5) | 8 |
| Ritz et al. [37] | California, USA | 2003 | Prospective | Semi-individual level | Monitoring network data | 58,316 | 5924 | TS | 20.0 (NA) | 7 |
| Brauer et al. [8] | Vancouver, Canada | 1999–2002 | Prospective | Semi-individual level | Monitoring network data | 70,249 | 3748 | WP | 5.1 (1.1) | 7 |
| Wu et al. [16] | California, USA | 1997–2006 | Retrospective | Individual level | Monitoring network data | 81,186 | 6712 | WP | 1.8 (1.4) | 9 |
| Gehring et al. [17] | North, west, and center of the Netherlands | 1996–1997 | Prospective | Individual level | Monitoring network data and land use regression model | 3853 | 165 | WP and TS | 20.1 (4.6) | 7 |
| Rudra et al. [38] | Washington, USA | 1996–2006 | Retrospective | Semi-individual level | Monitoring network data | 3509 | 369 | TS | 10.1 (NA) | 9 |
| Kloog et al. [34] | Massachusetts, USA | 2000–2008 | Retrospective | Semi-individual level | Remote sensing data | 634,244 | 61,972 | WP | 9.6 (5.3) | 9 |
| Lee et al. [35] | Pittsburgh, USA | 1997–2002 | Prospective | Semi-individual level | Monitoring network data | 34,705 | 1940 | TS | 15.6 (4.0) | 7 |
| Chang et al. 2015 | Atlanta, USA | 1999–2005 | Retrospective | Semi-individual level | Monitoring network data | 175,891 | 18,648 | WP and TS | 17.3 (3.1) | 8 |
| Fleischer et al. [10] | 22 countries | 2004–2008 | Retrospective | Regional level | Remote sensing data | 192,900 | 13,379 | WP | 1.4–98.1 (NA) | 7 |
| Nannam et al. 2014 [41] | Northwest England | 2004–2008 | Retrospective | Semi-individual level and individual level | Monitoring network data | 265,613 | 38,608 | WP and TS | 22.1 (4.6) | 9 |
| Ha et al. [42] | Florida, USA | 2004–2005 | Retrospective | Regional level and semi-individual level | Monitoring network data | 423,719 | 39,082 | WP and TS | 9.9 (2.0) | 8 |
| Hyder et al. [18] | Connecticut and Massachusetts, USA | 2000–2006 | Retrospective | Regional level and semi-individual level | Monitoring network data and remote sensing data | 647,942 | 41,868 | WP and TS | 11.9 (2.4) | 8 |
| Gray et al. [32] | North Carolina, USA | 2002–2006 | Retrospective | Regional level | Monitoring network data | 457,642 | 40,746 | WP | 13.6 (2.0) | 8 |
| Pereira et al. [9] | Connecticut, USA | 2000–2006 | Prospective | Semi-individual level | Monitoring network data | 61,688 | - | WP and TS | 12.4 (2.3) | 9 |
| Pereira et al. [56] | Perth, Australia | 1997-2007 | Prospective | Semi-individual level | Monitoring network data | 31,567 | - | WP and TS | 8.6 (2.2) | 9 |
NA: Data not available
a: Newcastle-Ottawa quality score
-: The number of cases was not available because these studies were longitudinal studies that assessed the effects of PM2.5 on preterm birth across successive pregnancies. NA: Data not available