Table 3. All-Cause Mortality Within One Year Among Patients Initiating Tramadol Prescription Compared With Other Propensity Score–Matched Analgesics .
| Nonselective NSAIDs | COX−2 Inhibitors | Weak Opioid | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Tramadol vs Naproxen | Tramadol vs Diclofenac | Tramadol vs Celecoxib | Tramadol vs Etoricoxib | |||||||||
| Tramadol (n = 12 397) | Naproxen (n = 12 397) | Tramadol (n = 6512) | Diclofenac (n = 6512) | Tramadol (n = 5674) | Celecoxib (n = 5674) | Tramadol (n = 2946) | Etoricoxib (n = 2946) | Tramadol (n = 16 922) | Codeine (n = 16 922) | |||
| Deaths, No. | 278 | 164 | 226 | 121 | 171 | 102 | 73 | 37 | 519 | 552 | ||
| Rate of death, per 1000 person-years | 23.5 | 13.8 | 36.2 | 19.2 | 31.2 | 18.4 | 25.7 | 12.8 | 32.2 | 34.6 | ||
| RD (95% CI), per 1000 person-years | 9.7 (6.3 to 13.2) | 17.0 (11.2 to 22.8) | 12.8 (6.9 to 18.7) | 12.8 (5.7 to 20.0) | −2.3 (−6.3 to 1.7) | |||||||
| HR (95% CI) | 1.71 (1.41 to 2.07) |
1 [Ref] | 1.88 (1.51 to 2.35) |
1 [Ref] | 1.70 (1.33 to 2.17) |
1 [Ref] | 2.04 (1.37 to 3.03) |
1 [Ref] | 0.94 (0.83 to 1.05) |
1 [Ref] | ||
| PS trimminga,b | 1.74 (1.42 to 2.13) |
1.87 (1.49 to 2.34) |
1.74 (1.35 to 2.24) |
2.00 (1.33 to 3.01) |
0.94 (0.83 to 1.06) |
|||||||
| History of other opioidsb | ||||||||||||
| Yes | 1.58 (1.17 to 2.14) | 1.83 (1.23 to 2.71) | 1.42 (0.98 to 2.06) | 2.39 (1.14 to 5.04) | 0.96 (0.83 to 1.09) | |||||||
| No | 1.80 (1.40 to 2.31) | 1.88 (1.44 to 2.46) | 1.93 (1.38 to 2.69) | 1.95 (1.21 to 3.14) | 0.88 (0.68 to 1.13) | |||||||
| As-treated approachb,c | 2.75 (1.86 to 4.06) | 2.04 (1.36 to 3.06) | 2.38 (1.54 to 3.69) | 2.55 (1.19 to 5.48) | 0.83 (0.68 to 1.02) | |||||||
| Incident OA patientsb,d | 1.50 (1.18 to 1.90) | 1.61 (1.19 to 2.16) | 1.91 (1.39 to 2.62) | 2.40 (1.44 to 4.01) | 0.97 (0.83 to 1.14) | |||||||
| Missing data imputationb,e | 1.63 (1.37 to 1.96) | 1.62 (1.33 to 1.96) | 1.92 (1.54 to 2.40) | 1.98 (1.28 to 3.06) | 0.92 (0.83 to 1.02) | |||||||
| Lagb,f | ||||||||||||
| Six months | 1.51 (1.22 to 1.88) | 1.87 (1.45 to 2.41) | 1.50 (1.15 to 1.96) | 1.96 (1.26 to 3.06) | 0.91 (0.80 to 1.05) | |||||||
| One year | 1.63 (1.3 to 2.05) | 1.95 (1.49 to 2.54) | 1.54 (1.16 to 2.05) | 1.92 (1.21 to 3.05) | 0.87 (0.76 to 1.01) | |||||||
Abbreviations: COX-2, cyclooxygenase 2; HR, hazard ratio; NSAIDs, nonsteroidal anti-inflammatory drugs; OA, osteoarthritis; PS, propensity score; RD, rate difference.
Asymmetric trimming was used to exclude participants whose propensity score was below the 2.5th percentile of the propensity score of the tramadol cohort and above the 97.5th percentile of the propensity-score of the comparator cohort.
The reference group is the second pair in each comparison.
This analysis censored the follow-up at the time when participants either changed (eg, switching from tramadol to naproxen or vice versa, when comparing tramadol with naproxen) or discontinued (ie, no prescription refill for the respective class of medication for over 60 d) their initiated medication.
This analysis was performed among participants whose osteoarthritis was diagnosed during the study period (ie, incident osteoarthritis) to minimize potential misclassification of duration of osteoarthritis.
Imputation analysis was performed to deal with missing data. Specifically, missing values of the variables (ie, body mass index, smoking, drinking status, or Townsend Deprivation Index) were imputed by a sequential regression method based on a set of covariates as predictors.
This analysis introduced a 6-month or 1-year exposure lag period to account for a potential latency time window (eg, excluding cancer cases that occurred within 6 months or 1 year.