Table 6

Effect estimates for never smokersa

Study

First author, year of publication

Adjustment factors

Exposure

Sex

Exposed cases

OR/RR (95%CI)


Studies providing result specific for oral cancer

8

Keller

1970 [75]

None

Smokeless

M

4

3.04 (0.62–14.99)

16

Blot

1988 [100]

Age, race, location, respondent

Smokeless

F

6

6.20 (1.90–19.80)

19

Blomqvist

1991 [103]

None

Snuff

M+F

2

0.67 (0.08–5.75)

25

Kabat

1994 [59]

None

Smokeless

M

4

1.59 (0.51–4.96)

F

4

38.7 (2.1–723.6)b

29

Schildt

1998 [73]

Age, sex, residence

Snuff

M+F

19

0.70 (0.40–1.20)

32

Henley

2005 [66]

Age, race, education, alcohol, exercise, aspirin, body mass index, diet, occupationc

Smokeless

M (CPS-I)

4

2.02 (0.53–7.74)

M (CPS-II)

1

0.90 (0.12–6.71)


Fixed-effect meta-analysis estimate for six studies

1.30 (0.87–1.93)

Random-effects meta-analysis estimate for six studies

1.94 (0.88–4.28)

Heterogeneity

χ2 (df)

18.99 (7)

p

0.0082

I2

63.13


Additional study providing result for oral, larynx and oesophagus cancer combined


27

Lewin

1998 [64]

Age, region

Snuff

M

9

4.70 (1.60–13.80)


Fixed-effect meta-analysis estimate for all seven studies

1.51 (1.04–2.19)

Random-effect meta-analysis estimate for all seven studies

2.20 (1.04–4.67)

Heterogeneity

χ2 (df)

23.82 (8)

p

0.0025

I2

66.42


aSee the discussion section for the reasons why data from study 15 [65] were not included in Table 6

b There were no exposed controls and the OR was estimated by adding 0.5 to each cell in the 2 × 2 table

cOccupation only adjusted for in CPS-II

Weitkunat et al. BMC Public Health 2007 7:334   doi:10.1186/1471-2458-7-334

Open Data