Participants
We recruited 90 patients, and 84 (93.3%) participants were randomized into three groups. In the standard group, 1 (3.5%) participant denied a surgical plan owing to personal issues, and 27 (96.5%) participants received TKA. Twenty-five (89.3%) participants finished the 12-week follow-up, and 2 (10.7%) participants dropped out in week 4. In the pregabalin group, all participants received TKA and finished the 12-week follow-up. In the duloxetine group, all 28 participants received TKA, 23 (82.1%) participants finished the 12-week follow-up, and 5 (17.9%) participants dropped out in week 2. No significant difference was found in sleeping medication use among the three groups 1 day before TKA and at weeks 2, 4, and 12. More details of the participants’ characteristics are presented in Table 1. There were no significant differences in these characteristics among the three cohorts. The postoperative joint function and pain score after TKA are presented in Supplementary Table 1.
Table 1
Basic characteristics of participants.
| Standard group (n = 28) | Pregabalin group (n = 28) | Duloxetine group (n = 28) | P value |
Age, median (range), y | 69.61 (53–83) | 68.29 (54–81) | 68.29 (51–81) | 0.758 |
Gender, No. (%) | | | | 0.651 |
| Male | 4 (14.3) | 3 (10.7) | 6 (27.3) | - |
| Female | 24 (85.7) | 25 (89.3) | 22 (72.7) | - |
Height, median (Q1-Q3), cm | 155 (152.25–160.00) | 157 (154.25–161.50) | 155 (150.25–163.50) | 0.671 |
Weight, median (Q1-Q3), kg | 62.39 (10.24) | 63.53 (9.73) | 62.23 (7.91) | 0.851 |
BMI, median (Q1-Q3), kg/m2 | 25.32 (22.75–27.34) | 25.98 (23.05–27.69) | 24.22 (22.87–26.87) | 0.687 |
Diabetes Mellitus, No. (%) | 5 (17.9) | 4 (14.3) | 4 (14.3) | 1.000 |
Hypertension, No. (%) | 12 (42.9) | 17 (60.7) | 8 (28.6) | 0.176 |
| Class I | 4 | 10 | 0 | - |
| Class II | 6 | 2 | 7 | - |
| Class III | 2 | 5 | 1 | - |
PSQI total, mean (SD) | 8.04 (3.88) | 8.71 (3.47) | 8.54 (3.57) | 0.771 |
| Use of sleeping medication, No. (%) | | | | 0.896 |
| | 0/week | 0 (0.0) | 0 (0.0) | 0 (0.0) | - |
| | <1/week | 0 (0.0) | 0 (0.0) | 0 (0.0) | - |
| | 1–2/week | 1 (3.6) | 2 (7.1) | 0 (0.0) | - |
| | ≥ 3/week | 2 (7.1) | 1 (3.6) | 2 (7.1) | - |
Duration of KOA, median (Q1-Q3), y | 10 (7.25-15.00) | 10 (5.25-15.00) | 10 (5.00–12.00) | 0.135 |
Kellagran-Lawrence grade, No. (%) | | | | 0.180 |
| 3 | 8 (28.6) | 8 (28.6) | 3 (10.7) | - |
| 4 | 20 (71.4) | 20 (71.4) | 25 (89.3) | - |
VAS, median (Q1-Q3) | 4.00 (4.00–5.00) | 4.00 (4.00–4.00) | 4.00 (4.00–5.00) | 0.130 |
painDETECT score, mean (SD) | 15.13 (7.30) | 16.39 (6.18) | 15.61 (5.48) | 0.759 |
PPT, mean (SD) | | | | |
| Knee with surgery plain | | | | |
| | Medial | 3.12 (0.66) | 3.28 (0.63) | 2.90 (0.84) | 0.145 |
| | Lateral | 3.42 (0.73) | 3.58 (0.79) | 3.32 (0.70) | 0.432 |
| Knee of the opposite side | | | | |
| | Medial | 4.01 (0.77) | 4.05 (0.98) | 3.98 (0.75) | 0.956 |
| | Lateral | 4.40 (0.74) | 4.60 (1.05) | 4.49 (0.82) | 0.681 |
HSS score, mean (SD) | | | | |
| Pain | 16.79 (3.39) | 17.32 (4.19) | 17.86 (3.95) | 0.585 |
| Function | 8.75 (2.91) | 8.89 (2.36) | 8.64 (2.73) | 0.940 |
| Mobility | 12 (2.33) | 12.18 (2.14) | 11.96 (1.82) | 0.920 |
| Muscle strength | 8.21 (0.63) | 8 (0) | 8.14 (0.53) | 0.232 |
| Deformity | 6.29 (3.81) | 6.29 (3.56) | 5.54 (3.64) | 0.678 |
| Stability | 8 (1.05) | 8.14 (0.525) | 7.75 (1.04) | 0.266 |
| Minus | 4 (2.75) | 4.93 (2.65) | 4.11 (1.87) | 0.307 |
| Total | 56.04 (9.77) | 55.89 (10.01) | 55.79 (9.29) | 0.995 |
y, year; No., number; Q1-Q3, first quartile to third quartile; cm, centimeter; kg, kilogram; BMI, body mass index; kg/m2, kilogram in metres squared; PSQI, Pittsburgh sleep quality index; SD, standard deviation; KOA, knee osteoarthritis; VAS, visual pain score; PPT, pressure pain threshold; HSS, hospital for special surgery. |
Primary End Points
There was a significant difference in VAS score at week 4 (standard group 1.00, Q1-Q3[1.00–2.00] vs. pregabalin group 1.00, Q1-Q3[1.00–1.00] vs. duloxetine group 1.00, Q1-Q3[1.00–1.00], P < 0.001). No significant difference was found in VAS score in week 2 (P = 0.147) and week 12 (P = 0.124), and absolute VAS score change in week 2 (P = 0.718), 4 (P = 0.283) and 12 (P = 0.389).
For operated knee PPTs, there were significant differences in PPTs in week 2 (medial knee space: standard group 4.90, 95% CI [4.54–5.25] vs. pregabalin group 10.30, 95% CI [9.45–11.16] vs. duloxetine group 9.47, 95% CI [8.83–10.11], P < 0.001; lateral knee space: standard group 6.18, 95% CI [5.77–6.60] vs. pregabalin group 11.45, 95% CI [10.45–12.45] vs. duloxetine group 10.19, 95% CI [9.72–10.66], P < 0.001), 4 (medial knee space: standard group 5.73, 95% CI [5.43–6.03] vs. pregabalin group 11.04, 95% CI [10.26–11.83] vs. duloxetine group 10.49, 95% CI [9.87–11.12], P < 0.001; lateral knee space: standard group 6.81, 95% CI [6.39–7.23] vs. pregabalin group 11.99, 95% CI [11.05–12.93] vs. duloxetine group 11.10, 95% CI [10.58–11.62], P < 0.001), and 12 (medial knee space: standard group 5.93, 95% CI [5.60–6.26] vs. pregabalin group 11.40, 95% CI [10.65–12.16] vs. duloxetine group 10.79, 95% CI [10.13–11.46], P < 0.001; lateral knee space: standard group 6.94, 95% CI [6.52–7.36] vs. pregabalin group 12.14, 95% CI [11.30-12.98] vs. duloxetine group 11.31, 95% CI [10.66–11.97], P < 0.001).
For operated knee PPTs absolute change, there were significant differences in PPT absolute change in week 2 (medial knee space: standard group 1.88, 95% CI [1.50–2.27] vs. pregabalin group 7.08, 95% CI [6.14–8.02] vs. duloxetine group 6.40, 95% CI [5.74–7.06], P < 0.001; lateral knee space: standard group 2.84, 95% CI [2.35–3.33] vs. pregabalin group 7.88, 95% CI [6.80–8.96] vs. duloxetine group 6.58, 95% CI [6.02–7.14], P < 0.001). However, no significant difference was found in the PPT absolute change in the medial space at weeks 4 (P = 0.336) and 12 (P = 0.687) or in the lateral space at weeks 4 (P = 0.430) and 12 (P = 0.918). More details are presented in Supplementary Table 2 and Fig. 2.
Secondary End Points
There were significant differences in PDQ scores at weeks 2 (standard group 16.19, 95% CI [13.80-18.57] vs. pregabalin group 9.04, 95% CI [7.36–10.71] vs. duloxetine group 9.57, 95% CI [6.82–12.32], P < 0.001), 4 (standard group 12.36, 95% CI [10.61–14.11] vs. pregabalin group 4.46, 95% CI [3.07–5.86] vs. duloxetine group 5.74, 95% CI [4.13–7.35], P < 0.001) and 12 (standard group 7.88, 95% CI [6.47–9.29] vs. pregabalin group 1.29, 95% CI [0.45–2.12] vs. duloxetine group 2.57, 95% CI [1.31–3.82], P < 0.001). A significant difference was also found in the absolute PDQ score change in week 2 (standard group − 1.59, 95% CI [-3.89-0.71] vs. pregabalin group 5.29, 95% CI [3.55–7.02] vs. duloxetine group 5.70, 95% CI [2.70–8.69], P < 0.001). However, there was no significant difference in the absolute PDQ score change at weeks 4 (P = 0.558) and 12 (P = 0.437). For sleep quality, no significant difference was found in PSQI and its change in weeks 2 (P = 0.765), 4 (P = 0.658) and 12 (P = 0.818).
For the opposite knee PPTs, significant differences were found in PPTs for medial space in weeks 2 (P < 0.001), 4 (P < 0.001) and 12 (P < 0.001) and lateral space in weeks 2 (P < 0.001), 4 (P < 0.001) and 12 (P < 0.001). Meanwhile, there were significant differences in PPT absolute change for the medial space (P < 0.001) and lateral space (P < 0.001) in week 2. However, no significant difference was found in PPTs absolute change for medial space in weeks 4 (P = 0.950) and 12 (P = 0.390) or lateral space in weeks 4 (P = 0.085) and 12 (P = 0.437).
For HSS score, there were significant differences in HSS deformity in week 2 (P = 0.026), HSS pain score in weeks 2 (P = 0.041), 4 (P = 0.002) and 12 (P < 0.001), HSS function score in week 4 (P = 0.027), and HSS total score in week 4 (standard group 77.32, 95% CI [74.77–79.87] vs. pregabalin group 82.29, 95% CI [80.40-84.17] vs. duloxetine group 78.83, 95% CI [75.88–81.77], P = 0.010), and 12 (standard group 86.28, 95% CI [3.51–89.05] vs. pregabalin group 90.68, 95% CI [89.15–92.21] vs. duloxetine group 89.87, 95% CI [87.75–91.99], P = 0.009). Meanwhile, significant differences were found in the absolute change in HSS pain (P = 0.018), HSS mobility (P = 0.021), HSS deformity (P = 0.011), and HSS total score (standard group 4.93, 95% CI [1.57–8.28] vs. pregabalin group 10.71, 95% CI [8.02–13.41] vs. duloxetine group 7.13, 95% CI [3.84–10.42], P = 0.023) in week 2. No significant difference was found in other HSS scores and absolute HSS score changes at weeks 2, 4, and 12. More data details are presented in Supplementary Table 2 and Figs. 3 to 5.
Adverse Events
In total, 7 (8.4%) participants reported treatment-related adverse events. Two (7.4%) participants in the standard group changed their treatment plan owing to aggravated pain in week 2. Five (17.9%) participants in the duloxetine group changed their treatment plan owing to neuropsychiatric symptoms in week 4. More details of AEs are presented in Supplementary Table 3.