APPENDIX 1: Cut-off points used for doses of NSAIDs and analgesics.

NSAID / High dose (mg)
Aceclofenac / >150
Meclofenamic acid / >300
Mefenamic acid / >1000
Niflumic acid / >500
Celecoxib / > 200
Desketoprofen / > 50
Dexibuprofen / > 800
Diclofenac / >100
Flurbiprofen / >150
Ibuprofen / > 1200
Indometacine / > 75
Ketoprofen / > 150
Ketorolac / >10
Lornoxicam / >8
Meloxicam / > 7,5
Nabumetone / >1000
Naproxen / > 1000
Nimesulide / >100
Piroxicam / > 10
Rofecoxib / >25
Sulindac / >200
Tenoxicam / >10
Paracetamol / Low (<=1,000); Medium (1,000-<=2,000); High (>2000)
Aspirin / Low (<=300); Medium (300-1000); High (>=1000)

Appendix 2. Current use of acid-suppressing agents at index date among controlsby age and exposure to different drugs and drug combinations.

Patients / PPI (%) / H2-blockers (%)
All
40-54
55-64
65-74
75+ / 10,000
2,781
2,093
2,464
2,662 / 957 (9.6)
105 (3.8)
148 (7.1)
290 (11.8)
414 (15.6) / 232 (2.3)
24 (0.9)
37 (1.8)
66 (2.7)
105 (3.9)
NSAIDs
40-54
55-64
65-74
75+ / 841
148
163
261
269 / 244 (29.0)
29 (19.6)
48 (29.5)
77 (29.5)
90 (33.5) / 44 (5.2)
4 (2.7)
4 (2.5)
18 (6.9)
18 (6.7)
Low-dose aspirin
40-54
55-64
65-74
75+ / 822
54
106
272
390 / 171 (20.8)
10 (18.5)
20 (18.9)
52 (19.1)
89 (22.8) / 44 (5.4)
3 (5.6)
6 (5.7)
10 (3.7)
25 (6,4)
Paracetamol
40-54
55-64
65-74
75+ / 1,029
94
146
319
470 / 238 (23.1)
16 (17.0)
31 (21.3)
64 (20.1)
127 (27.0) / 63 (6.1)
4 (4.3)
10 (6.9)
19 (6.0)
30 (6.4)
Metamizole
40-54
55-64
65-74
75+ / 224
25
34
78
87 / 62 (27.7)
3 (12.0)
8 (23.5)
27 (34.6)
24 (27.6) / 10 (4.5)
0
2 (5.9)
4 (5.1)
4 (4.6)
Oral corticosteroids
40-54
55-64
65-74
75+ / 110
13
19
37
41 / 51 (46.4)
4 (30.8)
11 (57.9)
18 (48.6)
18 (43.9) / 5 (4.6)
0
1 (5.3)
0
4 (9.8)
Oral anticoagulants
40-54
55-64
65-74
75+ / 233
11
25
97
100 / 45 (19.3)
2 (18.2)
5 (20.0)
20 (20.6)
18 (18.0) / 19 (8.2)
0
2 (12.0)
5 (5.2)
11 (11.0)
Antiplatelet drugs
40-54
55-64
65-74
75+ / 267
3
32
88
144 / 78 (29.2)
1(66.7)
9 (28.1)
23 (26.1)
44 (30.6) / 24 (9.0)
0
3 (9.4)
7 (8.0)
14 (9.7)
SSRIs
40-54
55-64
65-74
75+ / 360
78
46
85
151 / 88 (24.4)
9 (11,5)
10 (21.7)
25 (29.4)
44 (29.1) / 16 (4.4)
0
3 (9.4)
7 (8.0)
14 (9.7)
NSAIDs + aspirin / 84 / 31 (36.9) / 4 (4.8)
NSAIDs + oral anticoagulants / 14 / 5 (35.7) / 1 (7.1)
NSAIDs + oral corticosteroids / 31 / 18 (58.1) / 0
NSAIDs + antiplatelet drugs / 33 / 12 (36.4) / 2 (6.1)
NSAIDs + SSRIs / 56 / 26 (46.4) / 1 (1.8)
Aspirin + antiplatelet drugs / 37 / 13 (34.1) / 1 (1.9)
Aspirin + SSRI / 60 / 22 (36.7) / 1 (1.8)
NSAIDs + metamizole / 55 / 22 (40.0) / 2 (3.6)
NSAIDs + paracetamol / 189 / 65 (34.4) / 13 (6.9)

APPENDIX 3. Linear regression analysis of the correlation between calendar year and ORs of UGIB associated with NSAIDs

. regress OR Year

Source | SS df MS Number of obs = 5

------+------F( 1, 3) = 66.25

Model | .320409944 1 .320409944 Prob > F = 0.0039

Residual | .014510002 3 .004836667 R-squared = 0.9567

------+------Adj R-squared = 0.9422

Total | .334919945 4 .083729986 Root MSE = .06955

------

OR | Coef. Std. Err. t P>|t| [95% Conf. Interval]

------+------

Year | -.179 .0219924 -8.14 0.004 -.2489897 -.1090103

_cons | 360.281 44.05084 8.18 0.004 220.0915 500.4704

------

APPENDIX 4. Sensitivity analysis for different scenarios of misclassification of the exposure for OTC drugs.

RESULTS OBTAINED IN THE STUDY

Current Users / Cases / Controls
176 / 841
Non-users / 852 / 7,319
Total (excluding past users) / 1,028 / 8,160

Crude OR = (176 * 7,319) / (852 * 841) = 1.80

a)First scenario: Results corrected for a non-differential misclassification in which 50% of users are classified as non-users in both cases and controls

True users detected / Cases / Controls
176
/ 841 (50%)
True users misclassified as non-users / 176 / 841 (50%)
True non-users / 676 / 6,478
Total (excluding past users) / 1,028 / 8,160

Crude OR = (352 * 6,478) / (676 * 1682) = 2.01

b) Second scenario: Differential misclassification with 50% misclassification among the controls, of whom 25% are medium-high dose aspirin users (1682 * 0.25 = 421) and 25% (1682 * 0.25= 421 --420 to round numbers-- ) are non-aspirin NSAIDs (NA-NSAIDs) users. A differential ratio of use prevalence case:control of 3.5 and 2.0 are applied for medium-high dose aspirin and NA-NSAIDs, respectively

Cases / Controls
True users detected
True users misclassified as non-users
/ 176 / 841 (50%)


185 aspirin users
/ 421 aspirin users
106 NA-NSAIDs users / 420 NA-NSAIDs users
True non-users / 561 / 6,478
Total (excluding
past users) / 1,028 / 8,160

Crude OR = (467 * 6,478) / (561 * 1,682) = 3.21

Table 7web. Effect of acid-suppressing drugs (ASD) on UGIB associated with different drugs.

Drugs combined with ASD / Cases (%)
N=1,193 / Controls (%)
N=10,000 / Adjusted RR* (95% CI) as compared to non-users of any / Adjusted RR* (95%CI) as compared to drug-only users
Low-dose aspirin
Aspirin only
Aspirin + ASD / 66 (5.5)
30 (2.5) / 374 (3.7)
169 (1.7) / 1.58 (1.18-2.11)
1.26 (0.82-1.94) / 1 (ref.)
0.77 (0.47-1.25)
Corticosteroids
Cort only
Cortico + ASD / 8 (0.7)
9 (0.8) / 36 (0.4)
54 (0.5) / 1.49 (0.66-3.34)
0.67 (0.31-1.46) / 1 (ref.)
0.45 (0.15-1.37)
Oral anticoagulants
OAC only
OAC + ASD / 29 (2.4)
16 (1.3) / 145 (1.5)
64 (0.6) / 1.78 (1.17-2.71)
1.69 (0.94-3.04) / 1 (ref.)
0.95 (0.47-1.93)
Other antiplatelet drugs
OAD only
OAD + ASD / 28 (2.4)
23 (1.9) / 128 (1.3)
101 (1.0) / 1.62 (1.04-2.51)
1.30 (0.79-2.14) / 1 (ref)
0.80 (0.42-1.53)
Metamizole
Meta only
Meta + ASD / 31 (2.6)
12 (1.0) / 119 (1.2)
72 (0.7) / 1.88 (1.24-2.87)
0.82 (0.43-1.57) / 1 (ref.)
0.44 (0.20-0.93)
SSRIs
SSRIs only
SSRIs + ASD / 35 (2.9)
13 (1.1) / 210 (2.1)
103 (1.0) / 1.15 (0.78-1.68)
0.60 (0.32-1.10) / 1 (ref.)
0.52 (0.26-1.06)

1