Table 2 Association between lung cancer mortality and categorical cumulative silica in the Vermont granite workers study, 1950–94.

	Cumulative silica*	Cases	Mortality rate ratio† (95% CI)	p for trend
Original cohort‡		n = 213
	0 to 0.4	53 (24.9%)	Referent
	>0.4 to 1.3	53 (24.9%)	1.84 (1.22 to 2.77)	0.001
	>1.3 to 2.9	53 (24.9%)	2.54 (1.58 to 4.06)
	>2.9 to 48.2	54 (25.4%)	2.35 (1.36 to 4.07)
Stratified by hire status: cutpoints same as original cohort
Prevalent hires§		n = 157
		15 (9.6%)	Referent
		37 (23.6%)	1.63 (0.89 to 2.98)	0.02
		51 (32.5%)	2.14 (1.18 to 3.90)
		54 (34.4%)	2.08 (1.09 to 3.98)
Incident hires¶		n = 56
		38 (67.9%)	Referent
		16 (28.6%)	2.01 (1.05 to 3.83)	0.04
		2 (3.6%)	2.22 (0.49 to 10.02)
		0	NA
Stratified by hire status: cutpoints determined within hire status
Prevalent hires§		n = 157
	0 to 1.0	39 (24.8%)	Referent
	>1.0 to 2.0	39 (24.8%)	1.29 (0.81 to 2.04)	0.15
	>2.0 to 3.6	39 (24.8%)	1.90 (1.18 to 3.08)
	>3.6 to 48.2	40 (25.5%)	1.33 (0.78, 2.26)
Incident hires¶		n = 56
	0 to 0.04	19 (33.9%)	Referent
	>0.04 to 0.4	18 (32.1%)	1.39 (0.67 to 2.85)	0.02
	>0.4 to 2.0	19 (33.9%)	2.51 (1.14 to 5.55)

*15‐year lag, mg‐year/m³.

†Models used age as the timeline and controlled for calendar year and age at hire.

‡Subjects who worked in granite industry anytime between 1 January 1950 (start of follow‐up) and 31 December 1982 (includes incident and prevalent hires).

§Subjects who were already at work prior to the start of follow‐up on 1 January 1950 and still at work on that date.

¶Subjects who started working in the granite industry between 1 January 1950 (the start of follow‐up) and 31 December 1982.