Nowadays, although multimodal analgesia strategies have been widely implemented to control postoperative pain in TKA patients, it was reported that over 20% of patients still suffered from severe postoperative pain . To deal with this problem, surgeons never stopped to explore new methods. Recently, some researchers started to turn their eyes out of regular analgesia drugs, onto the application of antidepressant drugs like tricyclic antidepressants, selective serotonin reuptake inhibitors, and SNRIs, especially duloxetine. The latest review in 2014 which discussed the application of antidepressant drugs for postoperative pain concluded that there was insufficient evidence to support clinical use and called for more high-quality trials to confirm the exact effect of antidepressant drugs on postoperative pain . Duloxetine (Cymbalta), a potent selective SNRI, primarily found to treat depression, has already been demonstrated the analgesic effect in patients with chronic musculoskeletal pain . However, its effect on postoperative pain relief after TKA remains controversial.
So, we conducted this study to detect if duloxetine could be used as an adjuvants or supplement of multimodal analgesia strategies after TKA. In this trial of oral duloxetine 60 mg administered daily for 17 days initiated on preoperative day 2 and continued for 14 days postoperatively, postoperative pain, including both rVAS and aVAS, was significantly reduced relative to placebo through the whole postoperative 7 days. Besides, the placebo group consumed markedly more morphine equivalent during the postoperative period. In terms of secondary outcomes, postoperative recovery of function of the knee (range of motion at different time points) was better in the duloxetine group on the early days after surgery. When it comes to the side effects, patients took duloxetine had better sleep quality and less cases of nausea, vomit or constipation than placebo group, and the difference was marked.
SNRIs, like duloxetine, are targeted presynaptically and postsynaptically on serotonergic and noradrenergic nerve terminals and are the key regulator of synaptic serotonin and noradrenaline levels in central nervous system . The basic function of SNRIs is to block and then down-regulate serotonin and noradrenaline transporters, which therefore increases the concentrations of these two neurotransmitters . The increase of unbound serotonin and noradrenaline would enhance the descending inhibitory pain pathways in the central nervous system [5, 6]. Although the antidepression effect of SNRIs is relatively independent with its analgesia effect, there are still some studies demonstrated that serotonergic signaling in brain regions relevant to affective cognition have been demonstrated to be coupled to tonic pain ratings in healthy volunteers [17, 18]. It was reported that 7–14 days were necessary for SSRIs to take effect in treating anxiety or depression because the brain need time to establish the response [19, 20], so we hypothesized that it might likewise take days for SNRIs to take effect in relieving pain.
Lunn et al.  randomized 120 patients scheduled for TKA in a double-blind manner to either 10 mg escitalopram (a kind of SSRI) or placebo daily from pre-anesthesia to postoperative day 6 in addition to a standardized analgesic regime and found that no between-group differences were observed in overall pain upon ambulation from 4 to 48 h or overall pain at rest from 2 to 48 h after surgery. But conversely, overall pain upon ambulation and overall pain at rest from day 2 to 6 after surgery were lower in the escitalopram versus placebo group, which suggested that SSRI could relief postoperative pain but with a delayed attribute. Similarly, Koh et al.  randomized 80 patients scheduled for TKA to either duloxetine group (30 mg from the day before surgery and for 6 weeks after surgery) or control group and finally found that patients received duloxetine experienced better pain control during postoperative weeks 2 to 12, while there was no apparent benefit in terms of reduced pain during the immediate postoperative period (weeks 1 and 2). Koh et al. attributed the delayed analgesia to the study design and to the already potent analgesic efficacy of their current multimodal regimen. According to the abovementioned trials, we hypothesized that it might take days as buffer time for duloxetine to take the effect of analgesia. So, we conducted this trial, initiating the intervention of duloxetine much earlier than previous studies (60 mg daily for 17 days initiated on preoperative day 2 and continued for 14 days postoperatively), so that the duloxetine might take effect quickly after surgery, when the pain was supposed to be severe. Consequently, we found that since postoperative 6 h, both the rVAS and aVAS were significantly lower in duloxetine group till postoperative 7 days. It demonstrated that initiating duloxetine early preoperatively could effectively relief the pain of patients after TKA, which was consistent with prior studies and our initial hypothesis.
In terms of the opiate consumption, it has the different function with VAS. Opiate consumption is also a parameter that measures the postoperative pain of patients. We give patients opioids only when the immediate VAS score exceeded 6, which indicates the patient is suffering from severe pain, and the opioids act as a remedial treatment. If we can say the VAS score reflects the patients’ regular condition of pain, we can also say the opiate consumption reflects patients’ fulminant pain. In 2010, Ho et al.  first conducted a randomized placebo-controlled trial in 50 patients who received either two doses of oral duloxetine 60 mg (2 h before TKA and on the first day postoperatively) or placebo to specially detect the effect of duloxetine in morphine requirements after TKA and found that the morphine consumption during the 24 h and 48 h after TKA were significantly lower in the duloxetine group, compared with the placebo group. However, the sample size was small in Ho’s study, and only two doses of duloxetine were administrated which could not totally reflect the effect of duloxetine during the postoperative period. Later in 2016, based on the study of Ho et al., YaDeau et al.  implemented a longer duration of therapy starting from the day of surgery to the postoperative 14 days (1 dose of 60 mg oral duloxetine daily, 15 doses in all) and found the opioids-reduced effect of duloxetine started quickly after administration. The total daily opioid consumption was significantly less in the duloxetine group from postoperative day 1 to 14, compared with the placebo group, which was consistent with Ho’s study. Similarly, in our study, we prolonged the application of duloxetine before surgery and found that the daily morphine equivalent was markedly less in the duloxetine group from postoperative day 1 to 7, through the early postoperative period. To sum up, the duloxetine could effectively relief the fulminant pain after TKA and this function took effect quickly after application.
Range of motion, which acted as a measurement of recovery of knee function, was significantly better in the duloxetine group during the early postoperative period. Because, both VAS (rVAS and aVAS) and opioid consumption were significantly less in the duloxetine group, and the relatively milder pain would encourage patients in the duloxetine group to be more willing to exercise more frequently and hard, which would therefore accelerate the recovery of the knee function after surgery.
Tolerability and side effects are important issues in analgesic trials. It was reported by Lunn MP et al.  that minor adverse effects are common with duloxetine and the regimen of duloxetine 60 mg daily has lower rate of side effects than the dose of 120 mg daily. In our study, although some mild side effects occurred, all the patients were tolerated to the duloxetine regimen of 60 mg per night. Some patients in duloxetine group said they had more drowsiness at night. It is an interesting finding because the drowsiness at night isn’t strictly a side effect, it was reported that a deep sleep at night could relief the nervous mind and unnecessary anxiety. Besides, it was reported by Blagestad et al.  that the pain after surgery may obviously contribute to impaired sleep. Oppositely, an improved sleep has also been demonstrated antinociceptive effects . Of note, we noticed that the rates of nausea and vomit as well as constipation were markedly lower in the duloxetine group than in the placebo group. We inferred that the minimum in opioid use by duloxetine might therefore reduce the opioid-related adverse effects especially nausea, vomit and constipation. Although no serious side effects were observed, other studies have suggested an increased risk of perioperative bleeding from the cut and other postoperative morbidities . However, in our study, no such events were observed.
This study had several limitations, firstly, the sample size of this study was calculated according to the primary outcome, so, it might be underpowered to detect other outcomes like rates of various side effects, which calls for future studies to investigate. What’s more, although we initiated the duloxetine intervention earlier than prior studies [9, 22], it still couldn’t be determined whether the duloxetine exerted an effect exactly after surgery, where pain is more pronounced. So, future studies could focus on the optimal duration of perioperative use of duloxetine.