Skip to main content

Table 16 Meta-analyses for age started to smoke a

From: Systematic review with meta-analysis of the epidemiological evidence in the 1900s relating smoking to lung cancer

Age started

Statisticb

All lung cancerc

Squamousd

Adenoe

Number of setsf

 

69

15

14

About age 26 yearsg

n

60

14

13

 

F

2.70 (2.62-2.79)

10.42 (8.25-13.15)

3.79 (3.12-4.60)

 

R

3.89 (3.33-4.56)

11.06 (6.87-17.81)

3.21 (2.12-4.87)

 

H, PH

8.73, p < 0.001

3.36, p < 0.001

4.09, p < 0.001

About age 18 yearsh

n

29

6

5

 

F

7.75 (7.18-8.37)

20.15 (15.26-26.61)

9.74 (7.63-12.43)

 

R

7.48 (5.94-9.42)

20.28 (13.92-29.53)

8.84 (6.14-12.73)

 

H, PH

7.44, p < 0.001

1.60, NS

1.80, NS

About age 14 yearsi

n

35

6

5

 

F

11.11 (10.23-12.06)

29.91 (22.62-39.54)

13.62 (10.78-17.22)

 

R

10.32 (8.04-13.26)

31.07 (17.93-53.85)

12.34 (7.23-21.08)

 

H, PH

7.42, p < 0.001

3.28, p < 0.01

3.76, p < 0.01

Earliest vs. latestj

n

73

18

17

 

F

1.69 (1.64-1.74)

1.99 (1.72-2.29)

1.94 (1.64-2.29)

 

R

2.35 (2.08-2.65)

2.23 (1.66-2.98)

1.99 (1.48-2.67)

 

H, PH

4.52, p < 0.001

3.19, p < 0.001

2.62, p < 0.001

  1. a Within each study, results are selected in the following preference order, within each sex, for:
  2. smoking status – ever, current;
  3. smoking product – any, cigarettes (ignoring other products), cigarettes only;
  4. cigarette type – any, manufactured (with or without handrolled), manufactured only;
  5. unexposed group – never any product, never cigarettes, near equivalent (see Methods), but see also footnote j;
  6. follow-up period – longest available;
  7. lung cancer type – see notes c to e;
  8. race – all or nearest available, otherwise by race;
  9. overlapping studies – principal, subsidiary;
  10. age – whole study, widest available age group;
  11. Results are then selected for:
  12. sex – single sex results, combined sex results;
  13. adjustment for potential confounders – most available.
  14. b n = number of estimates combined, F = fixed-effect meta-analysis RR (95% CI), R = random-effects meta-analysis RR (95% CI), H = heterogeneity chisquared per degree of freedom, PH = probability value for heterogeneity expressed as p < 0.001, p < 0.05, p < 0.1 or NS (p ≥ 0.1).
  15. c All or nearest available, must include at least squamous cell carcinoma and adenocarcinoma.
  16. d Squamous cell carcinoma or nearest available, but not including adenocarcinoma.
  17. e Adenocarcinoma or nearest available, but not including squamous cell carcinoma.
  18. f Number of sets of RRs available for the key value analysis, where base for comparison is never smoked.
  19. g Category for which results are provided includes 26 years but does not include 18 years.
  20. h Category for which results are provided includes 18 years but does not include 14 or 26 years.
  21. i Category for which results are provided includes 14 years but does not include 18 years.
  22. j For this analysis only, the exposed and unexposed group have the same smoking status, product and cigarette type. There is an additional preference to select the results with least adjustment for other aspects of smoking, followed by a preference to select the results for the earliest (=exposed) and latest (=unexposed) starters. Alternatively preferring results with most adjustment for other aspects of smoking gives n = 73, F = 1.67 (1.62-1.72), R = 2.20 (1.96-2.47), H, PH = 4.01, p < 0.001 for all lung cancer.