Table 2 Opium use in ESCC cases and matched controls

From: Opium, tobacco, and alcohol use in relation to oesophageal squamous cell carcinoma in a high-risk area of Iran

 

ESCC cases

Matched controls

Unadjusted

Adjusted

 
 

N (%)

N (%)

OR (95% CI)

OR (95% CI) a

P for trend b

Opium use

 Never

210 (70)

465 (81)

Referent

Referent

—

 Ever

90 (30)

106 (18)

1.95 (1.36–2.78)

2.00 (1.39–2.88)

 

Average intensity

 Never used

210 (70)

465 (81)

Referent

Referent

<0.001

 ⩽ median (1.5 units per day)c,d

42 (14)

53 (9)

1.75 (1.11–2.75)

1.72 (1.08–2.72)

 

 > median

48 (16)

53 (9)

2.19 (1.37–3.51)

2.38 (1.47–3.85)

 

Duration

 Never used

210 (70)

465 (81)

Referent

Referent

<0.001

 ⩽ median (8.5 years)d

34 (11)

53 (9)

1.44 (0.88–2.39)

1.47 (0.89–2.44)

 

 > median

56 (19)

53 (9)

2.37 (1.54–3.65)

2.47 (1.59–3.84)

 

Cumulative use e

Never used

210 (70)

465 (81)

Referent

Referent

<0.001

 ⩽ median (15.5 unit-years)c,d

41 (14)

53 (9)

1.74 (1.10–2.75)

1.73 (1.09–2.74)

 

 > median

49 (16)

53 (9)

2.19 (1.38–3.47)

2.34 (1.45–3.78)

 

Age started

 Never used

210 (70)

465 (81)

Referent

Referent

<0.001

 > median

29 (10)

50 (9)

1.26 (0.74–2.16)

1.26 (0.74–2.16)

 

 ⩽ median (50 years)d

61 (20)

56 (10)

2.50 (1.63–3.84)

2.64 (1.71–4.09)

 

Type of opium used

 Never used

210 (70)

465 (81)

Referent

Referent

—

 Crude opium (teriak) only

66 (22)

95 (17)

1.56 (1.06–2.31)

1.62 (1.09–2.40)

 

 Refined opium (shireh) only

14 (5)

8 (1)

3.84 (1.57–9.42)

3.41 (1.35–8.60)

 

 Both

10 (3)

3 (1)

7.36 (1.99–27.2)

8.80 (2.28–33.9)

 

Route of administration

 Never used

210 (70)

465 (81)

Referent

Referent

—

 Smoked only

45 (15)

62 (11)

1.67 (1.06–2.59)

1.67 (1.06–2.63)

 

 Ingested only

30 (10)

38 (7)

1.86 (1.07–3.21)

1.90 (1.09–3.32)

 

 Both

15 (5)

6 (1)

5.41 (2.05–14.2)

6.61 (2.42–18.0)

 
  1. 95% CI=95% confidence interval; ESCC=oesophageal squamous cell carcinoma; OR=odds ratio. ORs were obtained from conditional logistic regression models.
  2. aAdjusted for education and ethnicity.
  3. bP for trend was obtained from conditional logistic regression models by assigning values of 0, 1, and 2 to no use, below median, and above median use, respectively.
  4. cEach unit is one ‘nokhod’, which is approximately equal to 0.2 g.
  5. dWe used the median in control subjects as the dividing cut point.
  6. eCumulative use was calculated by multiplying intensity of use (per day) by duration of use (in years).
Back to article page