Intrathecal morphine administration reduces postoperative pain and peripheral endocannabinoid levels in total knee arthroplasty patients: a randomized clinical trial

The endocannabinoid system regulates nociception in rodents and humans both through central and peripheral mechanisms [1‐3]. Endocannabinoids are implicated in stress-induced analgesia such as that triggered by mild electric foot shock [4], from strenuous exercise [5], and in modulating the intertwined pain and emotional responses [6]. The endocannabinoids anandamide (AEA) and 2-arachidonoyl glycerol (2-AG) are endogenous lipids that serve as agonists for cannabinoid receptors (CB1 and CB2) [7, 8]. Blockade of endocannabinoid inactivating enzymes (monoacylglycerol lipase cleaves 2-AG; and fatty acid amide hydrolase cleaves AEA) elevates tissue endocannabinoid levels and produces antinociceptive effects [2, 9]. The N-acylethanolamines (NAEs) palmitoylethanolamide (PEA) and oleoylethanolamide (OEA) are lipids structurally related to AEA but do not activate cannabinoid receptors; instead they serve as agonists at the nuclear peroxisome proliferator-activated receptor alpha [10, 11]. Activation of peroxisome proliferator-activated receptor alpha by PEA produces antinociceptive effects in preclinical models of pain [11, 12]. Recent reports indicate that tissue PEA levels are suppressed in animal models of pain [13, 14] and circulating PEA levels are decreased in irritable bowel syndrome patients experiencing abdominal pain compared to pain-free patients [15], suggesting an antinociceptive role of PEA in certain active pain states.

Interactions between the endocannabinoid and opioid systems are well documented [16‐18]. Previous preclinical studies demonstrated synergistic analgesic effects between opioids and cannabinoids and furthermore that analgesic effects of opioids can be reversed by blockade of cannabinoid receptors [19‐23]. Clinical data also suggest that the opioid and cannabinoid systems regulate placebo-induced analgesia in humans [24, 25]. Endocannabinoid levels are dynamically regulated under chronic pain states [26, 27] and augmentation of endocannabinoid levels has been suggested as a novel therapeutic strategy for analgesic development [28, 29].

Total knee arthroplasty (TKA) is a procedure associated with significant acute and chronic postoperative pain [30‐32]. Although the regulation of nociception between the endocannabinoid and opioid systems has previously been described, the influence of opioid receptor activation upon endocannabinoid levels in humans has never been examined. In a rat model of morphine sensitization, acute morphine administration elevated brain AEA levels while concomitantly decreasing 2-AG [33]. Herein, we examined whether acute intrathecal morphine administration, which suppresses postoperative pain in TKA patients [34], alters circulating levels of endocannabinoids prior to and after TKA surgery. Because the endocannabinoid tone may be recruited to dampen pain [35‐37], and because the analgesic effects between opioids and cannabinoids are synergistic, we hypothesized that administration of intrathecal morphine would reduce postoperative pain in patients undergoing TKA by at least 30% [38] and consequently suppress circulating endocannabinoid levels.

Patients undergoing TKA were subdivided into two groups receiving either intrathecal morphine or placebo (Fig. 1). Table 1 shows key demographic features of the two patient groups. As expected and consistent with previous data [34], patients receiving intrathecal morphine reported reduced acute postoperative pain after surgery compared to placebo (Fig. 2a, p = 0.0049). Intrathecal morphine also reduced postoperative systemic opioid consumption measured using patient controlled analgesia (Fig. 2b, p = 0.001). Serum endocannabinoid and NAE levels were examined in patients receiving intrathecal morphine or placebo; and pre-surgical baseline (0 h) blood was drawn approximately 10–15 min after morphine administration. There were no differences in serum PEA levels between the placebo and morphine groups (Fig. 3, p = 0.532). In contrast, AEA (p < 0.0001), 2-AG (p = 0.0008), and OEA (p = 0.0025) levels were significantly lower in the morphine group at baseline compared to placebo (Fig. 3).

We subsequently examined whether intrathecal morphine altered endocannabinoid levels 4 h after surgery. In the placebo group, 2-AG levels were significantly elevated (p = 0.008) 4 h after surgery while the change in levels of AEA (p = 0.424), PEA (p = 0.143), and OEA (p = 0.474) were not significant (Fig. 3 and Table 2). In contrast, compared to baseline, levels of AEA (p = 0.009), PEA (p = 0.0001), and OEA (p = 0.0008) were significantly reduced in the morphine group while the change in 2-AG levels was not significant (p = 0.0698). Trajectories of changes in serum endocannabinoid levels between the 0 and 4 h time points are shown in Fig. 4. Compared to placebo, the levels of AEA, PEA, and OEA were significantly lower in patients treated with intrathecal morphine at 4 h (Fig. 3). A significant group by time effect was observed for PEA (p = 0.0072) for the placebo and morphine groups (Table 2).

Surgical procedures produce a stress response that manifests as an elevation in circulating cortisol levels [43]; and corticosteroids enhance the biosynthesis of endocannabinoids [44‐46]. Therefore, it is possible that the lower endocannabinoid levels observed in the morphine group may have resulted from a suppressed cortisol tone secondary to less acute pain. At baseline, there was no difference in cortisol levels (p = 0.7457) between the placebo and morphine groups (Fig. 5). However, cortisol levels were elevated 4 h after surgery in the placebo group (p = 0.043) while they were reduced in the morphine group (Fig. 5, p = 0.044). A significant group by time effect (p = 0.0103) was observed for cortisol for the placebo and morphine groups (Table 2); suggesting a reduced stress response in the patients receiving intrathecal morphine.

Our data demonstrate that intrathecal morphine reduced postoperative pain and the need for systemic opioids and altered baseline and postoperative levels of endocannabinoids in TKA patients. We demonstrated that intrathecal morphine lowered post-surgical levels of AEA, PEA, and OEA when compared to placebo. Interestingly, 2-AG as well as cortisol levels increased in TKA patients treated with placebo whereas 2-AG levels were not significantly altered and cortisol levels decreased in patients treated with morphine. These results are noteworthy given our previous data demonstrating that 2-AG is higher in CSF and synovial fluid in patients who report more severe postoperative pain after TKA surgery [41], possibly due to its conversion into downstream proalgesic eicosanoids [47].

Our novel data from human subjects agree with preclinical and clinical evidence consistently documenting crosstalk between the endocannabinoid and opioid systems. However, to date only a single preclinical study has examined the effects of opioids on endocannabinoid levels in rats [33]. To our knowledge, this is the first study to examine the effects of spinal opioid receptor activation upon circulating endocannabinoid/NAE levels. We hypothesized that morphine administration would reduce postoperative pain and bypass the need to recruit analgesic endocannabinoids.

The majority of patients scheduled for TKA surgery experience chronic pain due to end-stage osteoarthritis and endocannabinoid levels are known to be elevated in the blood and tissues of patients with chronic pain [35‐37, 48]. In our study, patients that received intrathecal morphine had significantly lower levels of circulating endocannabinoids, which were measured approximately 10 to 15 min after morphine administration. Therefore, it is tempting to speculate that activation of central opioid receptors by morphine results in a rapid onset of analgesia that supplants the necessity to recruit endogenously produced analgesic endocannabinoids. Indeed, compared to patients receiving intrathecal placebo, the morphine group reported reduced postoperative pain 4 h after surgery that was accompanied by a suppressed endocannabinoid/NAE tone.

In our study, the lower preoperative levels of AEA, 2-AG, and OEA observed in subjects receiving intrathecal morphine are unlikely to stem from downregulation of endocannabinoid biosynthetic enzymes or upregulation of catabolic enzymes. Instead, the most plausible explanation is that morphine interferes with the activity of cells that produce endocannabinoids. A recent report found a similarly rapid reduction in AEA levels in the amygdala of mice after administration of corticotropin-releasing hormone [49]. The reduction in AEA levels was likewise rapid (within 10 min) and these effects were likewise independent of de novo synthesis of the endocannabinoid catabolizing enzyme fatty acid amide hydrolase.

Currently, the source(s) of systemically circulating endocannabinoids is not well defined. Previous work has demonstrated that leukocytes produce and secrete endocannabinoids [50, 51]. Intrathecal morphine has been shown to reduce leukocyte function in humans [52], which could theoretically account for the reduced endocannabinoid levels in our cohort. However, a high dose of morphine (0.5 mg) was required to establish this immunosuppressive effect, which likewise only manifested twenty four hours but not two hours after surgery [52]. Therefore, it is unlikely that leukocyte suppression accounts for the rapid reduction in circulating endocannabinoids after morphine administration.

Endocannabinoids are also produced in primary sensory neurons in a calcium-dependent manner [1, 53, 54]. Therefore, it is possible that activation of spinal opioid receptors reduced the excitability of primary sensory neurons [55] and consequently suppresses endocannabinoid/NAE production. Because endocannabinoids can regulate nociception at the level of the primary sensory neuron [1], the analgesic effects of morphine may dampen local biosynthesis of endocannabinoids.

General anesthesia reduces circulating AEA and augments 2-AG levels during the postoperative period [56‐58]. In our patient cohort receiving intrathecal placebo, there were no changes in AEA, PEA, or OEA levels between the preoperative baseline and the 4 h postoperative time point while 2-AG levels as well as cortisol were elevated at 4 h. In contrast, the morphine group showed a significant suppression of AEA, PEA, and OEA levels, no 2-AG upregulation, and a decrease in cortisol postoperatively. These data suggest that the alterations in endocannabinoid levels are unlikely to stem from a general suppressive effect of anesthesia.

Surgical procedures induce a rapid increase in circulating cortisol levels and more effective analgesia blunts this postoperative cortisol response [43, 52, 59]. Interactions between stress and the endocannabinoid system are well established in rodents [60, 61]. Furthermore, acute stress in pain-free individuals stimulates cortisol release that is accompanied by a rapid increase in circulating AEA, OEA, and PEA levels [44, 46, 62]. Interestingly, a heightened cortisol response has been associated with lower pain in human subjects [63]. Therefore, in the morphine treated group, the endogenous postsurgical analgesic response that is normally characterized by release of cortisol and endocannabinoids, may have been supplanted by the potent analgesia attained after morphine administration.

One major limitation of the current study is the lack of baseline circulating endocannabinoid levels measured prior to intrathecal morphine or placebo administration. Another limitation stems from our inability to demonstrate that the suppression of the endocannabinoid tone after intrathecal morphine administration arises from its analgesic effects. Although we speculate that the reduction in endocannabinoids after morphine administration may reflect morphine analgesia and reduced necessity to recruit analgesic endocannabinoids, our study lacked patients without underlying chronic pain to validate this claim. We cannot rule out the possibility that the postsurgical increase in 2-AG in the placebo group might reflect pain-induced stress, which is known to increase 2-AG production [64] and which might have been blocked by intrathecal morphine as evidenced by the decrease in postsurgical cortisol levels in the patients who underwent the active treatment. Another limitation stems from the gender imbalance in favor of female subjects in the placebo group; therefore, the outcomes of this study may not completely extend to the male population and this must be controlled better in future more rigorous trials. Another limitation stems from the premature decision to stop the trial due to slower than expected recruitment; and incomplete dataset for the secondary outcome analyses of endocannabinoid and cortisol levels between the placebo and morphine groups. Lastly, the patients in the placebo group utilized more postoperative opioids to control their pain, which may have altered the levels of endocannabinoids. However, it is noteworthy that the baseline endocannabinoid levels in the placebo group are comparable to those observed in our previous study of TKA patients who were administered oral opioids prior to TKA surgery [41].

In conclusion, our study demonstrates that intrathecal morphine significantly reduces acute postoperative pain after TKA surgery. In addition, this study is the first to document a previously undescribed regulation of peripheral endocannabinoid levels by spinal opioid receptor activation and extends the previously documented cannabinoid-opioid interactions observed in rodents and non-human primates to humans [16‐18, 20, 21]. Our results indicate that activation of central opioid receptors negatively modulates the endocannabinoid tone, which may suggest that activation of strongly analgesic opioid receptors may obviate the need to recruit weakly analgesic endocannabinoids.