|
| Parameter | Value | Uncertainty range/95% credible interval | Assumed uncertainty distribution | Source |
|
| Efficacy | | | | |
| Change from baseline BASFI | | | | |
| Etoricoxib (90 mg) | −17.87 | −22.16; −13.64 | No distribution assumed; posterior distributions directly obtained from mixed treatment comparison of extracted data and simultaneously forwarded into Markov model. For MTC, noninformative prior distributions were used. |
Barkhuizen et al. [26];
Van der Heijde et al. [16]
Dougados et al. [27];
Sieper et al. [17] |
| Celecoxib (200 mg) | −10.12 | −12.34; −7.932 |
| Celecoxib (400 mg) | −11.8 | −14.52; −9.10 |
| Diclofenac | −11.51 | −15.68; −7.34 |
| Naproxen | −14.82 | −17.69; −11.98 |
| Change from baseline BASDAI | | | |
| Etoricoxib (90 mg) | −28.53 | −32.06; −25.05 | Van der Heijde et al. [16];
Sieper et al. [17] |
| Celecoxib (200 mg) | −18.47 | −24.12; −12.9 |
| Celecoxib (400 mg) | −21.77 | −26.95; −16.51 |
| Diclofenac/naproxen | −23.46 | −26.96; −19.96 |
| Probability of discontinuation | | | |
| Etoricoxib (90 mg) | 0.063 | 0.027; 0.117 | |
| Celecoxib (200 mg) | 0.225 | 0.165; 0.292 | Barkhuizen et al. [26] |
| Celecoxib (400 mg) | 0.177 | 0.113; 0.255 | Van der Heijde et al. [16] |
| Diclofenac/naproxen | 0.149 | 0.105; 0.202 | Dougados et al. [27] |
| BASFI without treatment | 45 | 40; 50 | Uniform (40, 50) | Based on baseline characteristics of trials included in MTC (see Table 1) |
| BASDAI without treatment | 45 | 40; 50 | Uniform (40, 50) |
| Disease progression measured using annual changes in BASFI | 0.5 | 0; 0.10 | Uniform (0, 0.10) | Kobelt et al. [6];
Ara et al. [20] |
| BASFI with anti-TNFα | 23 | 20; 26 | Uniform (20, 26) |
Ara et al. [20] |
| BASDAI with anti-TNFα | 19 | 18; 20 | Uniform (18, 20) |
| BASFI when stopped with anti-TNFα | 55 | 50; 60 | Uniform (50, 60) |
| BASDAI when stopped with anti-TNFα | 52 | 47; 57 | Uniform (47, 57) |
| Annual probability of discontinuation from anti-TNFα treatment | 0.10 | 0.05; 0.15 | Beta (13.2, 118.8) | Ara et al. [20] |
|
| Safety and treatment (annual probabilities) | | | | |
| PUBs | | | | |
| Etoricoxib | 0.0111 | 0.0074; 0.0159 | No distribution assumed; posterior distribution directly obtained from indirect comparison analysis of extracted data and simultaneously forwarded into Markov model. For indirect comparison of safety, noninformative prior distributions were used. |
Ramey et al. [11];
Silverstein et al. [14] |
| Celecoxib | 0.0134 | 0.0075; 0.0221 |
| Diclofenac/naproxen | 0.0270 | 0.0216; 0.0334 | |
| Suspected PUBs | | | | |
| Etoricoxib | 0.0016 | 0.0000; 0.0061 | No distribution assumed; posterior distribution directly obtained from analysis of extracted data and simultaneously forwarded into Markov model. For indirect comparison of safety, noninformative prior distributions were used. |
Ramey et al. [11] |
| Celecoxib | 0.0016 | 0.0000; 0.0061 |
| Diclofenac/naproxen | 0.0030 | 0.0000; 0.0115 |
| Minor GI events | | |
MEDAL study [15] |
| Etoricoxib/celecoxib | 0.0463 | 0.0420; 0.0506 |
| Diclofenac/naproxen | 0.0704 | 0.0650; 0.0759 |
|
| PUB risk reduction with PPI | 0.40 | — | — | Moore et al. [29] |
| Dying from PUB | 0.036 | — | — |
Ramey et al. [11];
Laine et al. [28] |
| Hospitalization given PUB | 0.21 | 0.056; 0.358 | Uniform (0.056, 0.358) | Bloom et al. [39];
Maetzel et al. [40];
Knill-Jones et al. [41];
Singh and Ramey, [42];
de Pouvourville [43];
Jönsson and Haglund, [44];
Gabriel and Matteson, [45],
Smalley et al. [46];
Edelson et al. [47] |
|
| Surgery given hospitalization | 0.25 | 0.12; 0.39 | Uniform (0.12, 0.39) | Maetzel et al. [40];
Knill-Jones et al. [41];
Jönsson and Haglund, [44];
Gabriel et al. [48];
Johnson et al. [49];
Kong et al. [50] |
| Inpatient tx given suspected PUB | 0.25 | 0.18; 0.32 | Beta (36.5, 109.5) | Maetzel et al. [40] |
| Treatment given minor GI | 1 | — | — | Assumption |
| Thrombotic CV event rate | | | No distribution assumed; posterior distribution directly obtained from indirect comparison analysis of extracted data and simultaneously forwarded into Markov model. For indirect comparison of safety, noninformative prior distributions were used. | |
| Etoricoxib | 0.0124 | 0.0111; 0.01381 |
Cannon et al. [15];
Curtis et al. [30] |
| Celecoxib | 0.0124 | 0.0111; 0.01381 |
| Diclofenac | 0.0131 | 0.0117; 0.01454 |
| Naproxen | 0.0077 | 0.0039; 0.01381 | |
| Death from thrombotic CV event | | | |
Cannon et al. [15] |
| Etoricoxib/celecoxib | 0.13 | — | — |
| Diclofenac/naproxen | 0.128 | — | — |
| Edema | | | |
MEDAL [15] |
| Etoricoxib/celecoxib | 0.0106 | 0.0086; 0.0127 | Beta (101.4, 9459.7) |
| Diclofenac/naproxen | 0.0070 | 0.0054; 0.0088 | Beta (64.9, 9165.4) |
| Hypertension | | | |
| Etoricoxib/celecoxib | 0.0229 | 0.0200; 0.0260 | Beta (218.9, 9342.0) |
| Diclofenac/naproxen | 0.0153 | 0.0129; 0.0179 | Beta (141.2, 9088.8) |
| Coronary heart failure | | | |
| Etoricoxib/celecoxib | 0.0044 | 0.0032; 0.0058 | Beta (42.1, 9518.9) |
| Diclofenac/naproxen | 0.0026 | 0.0017; 0.0037 | Beta (24.7, 9490.5) |
| Hepatic events | | | |
| Etoricoxib/celecoxib | 0.0036 | 0.0025; 0.00489 | Beta (34.4, 9526.6) |
| Diclofenac/naproxen | 0.0218 | 0.0189; 0.0249 | Beta (201.2, 9028.8) |
| Renal events | | | |
| Etoricoxib/Celecoxib | 0.0114 | 0.0094; 0.0136 | Beta (109.0, 9452.0) |
| Diclofenac/Naproxen | 0.0100 | 0.0081; 0.0120 | Beta (92.3, 9137.7) |
|
| Utility | | | | |
| Relation between EQ-5d and BASFI and BASDAI | | | | Ara et al. [20] |
| Constant | 0.924 | 0.890; 0.957 | Normal (0.924, 0.0172) |
| BASFI | −0.004 | −0.0057; −0.0029 | Normal (−0.004, 0.00072) |
| BASDAI | −0.004 | −0.0056; −0.0024 | Normal (−0.004, 0.00082) |
|
| Disutility due to adverse events (adjusted for duration) | | | | |
| Surgery for PUB | 0.080 | 0.069; 0.092 | Beta distributions |
Moore et al. [29] |
| Inpatient treatment for PUB | 0.062 | 0.052; 0.072 |
| Outpatient treatment for PUB | 0.051 | 0.042; 0.060 |
| Inpatient investigation for suspected PUB | 0.062 | 0.052; 0.072 |
| Outpatient investigation for suspected PUB | 0.025 | 0.021; 0.030 |
| Minor GI symptoms requiring treatment | 0.015 | 0.012; 0.019 |
| Minor GI symptoms not requiring treatment | 0.00004 | 0.00000; 0.00032 |
| Thrombotic CV event | 0.294 | 0.256; 0.331 | Moore et al. [29] |
| Edema | 0.020 | 0.016; 0.024 | Revicki [33] |
| Hypertension | 0.001 | 0.000; 0.002 | Stason and Weinstein, [34] |
| Hepatic | 0.055 | 0.040; 0.072 | Nichol et al. [35] |
| CHF | 0.002 | 0.001; 0.002 | Wong et al, [36] |
| Renal | 0.020 | 0.016; 0.024 | Revicki [33] |
|
| Costs of events (NOK) | | | | |
| Surgery for PUB | 22,904 | 18,900; 27,300 | | |
| Inpatient treatment for PUB | 22,904 | 18,900; 27,300 | | |
| Outpatient treatment for PUB | 2,231 | 2,038; 2,437 | Gamma distributions |
Resource use from Jansen et al. [19]; drug acquisition costs from NoMA (September 2007); GP costs from Norwegian Medical Association; DRG prices from ISF 2007 [37]; |
| Inpatient investigation for suspected PUB | 22,295 | 18,240; 26,700 |
| Outpatient investigation for suspected PUB | 1,297 | 1,157; 1,445 | |
| Minor GI symptoms requiring treatment | 568 | 507; 636 | |
| Thrombotic CV event | 95,555 | — | — | NoMA (September 2007); ISF 2007 [37]; |
| CHF | 45,958 | — | — | NoMA (September 2007); ISF 2007 [37]; |
|
| Annual drug costs (NOK) | | | | |
| Etoricoxib (90 mg) | 4,654 | — | — | NoMA (September 2007) |
| Celecoxib (200 mg) | 3,318 | — | — |
| Celecoxib (400 mg) | 6,636 | — | — |
| Diclofenac (150 mg) | 1,588 | — | — |
| Naproxen (1000 mg) | 1,380 | — | — |
| PPI (omeprazole) | 3,050 | — | — |
| Aspirin (75 mg) | 383 | — | — |
| Anti-TNFα tx cost | 143,322 | — | — | NoMA (September 2007) |
|
