*3.3. Paired t-Tests*

Paired *t*-tests were used to compare dataset means from psychological questionnaires pre- and post-LTOT cessation (Table 2). As indicated in the table, six out of ten paired *t*-tests yielded a significant *p*-value (<0.0001). Because multiple *t*-tests were conducted, the Type 1 error rate might be inflated. The Bonferroni correction rectifies the situation by dividing the alpha level by the number of tests. In this case, the adjusted alpha level is 0.05/10 = 0.005. Nevertheless, because the *p*-value of all significant results is <0.0001, which is far lower than 0.005, the conclusion remains the same.


**Table 2.** Paired *t*-tests of pre- and post-LTOT cessation psychological assessment questionnaires.

\* Significant at the 0.01 Alpha level.

Data imputation was considered for this study for the inconsistent n value, but was ruled out because some participants deliberately chose to skip several questions or entire assessments. When data are missing due to inherent lack of randomness, it is impossible to account for systematic differences between the missing and the observed values based on the existing data, especially when the reference data are sparse. Further, when all or many responses in a group of questions are missing, multiple imputation will yield misleading results [46].

#### **4. Discussion**

#### *4.1. Improved Psychological Assessment Scores Post-LTOT Cessation*

This study found that the improvement between the pre- and post-cessation indicators of quality of life, depression, pain interference, catastrophizing, and fear avoidance (assessed via the SF-36, PHQ, BPI-Pain, BPI-Impairment, PCS, and FAB-PA, respectively) among subjects with CNCP who successfully ceased LTOT are significant at the level of *p* < 0.00l. These findings are consistent with those in previous studies examining mood changes affiliated with problematic opioid use, which have described an intercorrelation between such use and increased depression [8–11,47], fear avoidance [43], catastrophizing [10,12,13,38,48], and decreased quality of life [7] with increased states of pain [12,14–16,47–60]. Interpretation of the present findings must consider not only the impact of LTOT cessation, but also that of buprenorphine introduction and concurrent participation in a group multidisciplinary program, or even a potential synergy between these factors.

Most subjects utilized buprenorphine for LTOT cessation. Buprenorphine acts as a partial mu agonist and as an antagonist at the delta [26,55] and kappa [24,26,55] opioid receptors [26,57–59]. These receptor dynamics have been theorized to confer anxiolytic [56] and anti-depressive effects [50,58–64] in opioid-dependent [56] and opioid-naïve [47,55,58,60] patients and to improve patient-reported scores regarding quality of life [28,50]. Perhaps the addition of buprenorphine can explain, at least in part, the discrepancy between the participants' mood improvements compared with recent studies documenting clinical deterioration with LTOT cessation [19–21]. This dynamic would be an interesting area of future study.

The effects of both partial and full mu agonist opioid exposure on quality of life have been previously studied. One study showed that full mu agonist opioid exposure lowered the SF-36 score initially, but no difference was found between patients with CNCP exposed to opioids vs. not exposed after 4 years duration [7]. Another reported that opioidnaïve patients exhibited improvements in SF-36 scores after 8 weeks of daily treatment with buprenorphine for the indication of treatment-resistant depression [50]. Similarly, a different study found that switching from LTOT to sublingual buprenorphine for at least 60 days resulted in improved quality of life scales, as measured using the QOLS [28]. Longer term follow-up studies would be helpful to better understand the relationship between LTOT, buprenorphine, and quality of life.

In the current dataset, the post-LTOT cessation depression scores (assessed via the PHQ) show a mean improvement from moderately severe to moderate depression, which trends consistently with the results in previous publications. The link between depression and problematic full mu agonist opioid use has been previously documented from a variety of angles. Higher depression scores have been correlated with opioid misuse in CNCP patients with no prior substance use disorder [10]. Depression has also been identified as a risk factor for prolonged post-surgical opioid use [11]. Similarly, researchers have found that opioid use is a risk factor for depression, independent of pain [8,9]. Buprenorphine's effect on depression has also been studied in patients with and without concurrent opioid use disorder and was shown to provide marked improvement in suicidal ideation and treatment-resistant depression after administration [47,58]. Similarly, in a meta-analysis conducted by Serafini et al., 13 studies concluded that buprenorphine alone, or in coadministration with other opioid antagonists, may significantly reduce depression [59]. The current data, supported by previous studies, strongly suggest that depression is an important target for assessment and therapy in the pursuit of LTOT cessation.

Previous reports regarding the PCS and FAB are nuanced. For the PCS, it is interesting to note that a score of 30 or higher was initially validated as clinically relevant [38,42]. However, lower scores have been documented to be associated with the chronicity of prolonged recovery and delayed return to work [45]. Therefore, the substantial improvement in the PCS score with LTOT cessation seen in this study can be argued to be clinically significant, despite being below 30 on the pretest, especially considering the scale of the

improvement. Targeted psychosocial therapy to improve catastrophizing has been shown to significantly aid in returning to work after a period of disability [48]. Current results suggest that addressing catastrophizing may be an avenue worthy of future study for promoting LTOT cessation for patients with CNCP.

The optimal cut off for determining significant FAB scores has been studied in several contexts and varies respectively [36,40,61–64]. Higher FAB scores have been correlated with an increased probability of current and future work loss and disability [36,65,66], as well as social withdrawal [67] and increased LTOT reliance [14]. FAB analysis may also help determine which clinical interventions have an increased probability of a successful outcome to decrease patient-reported disability and pain [36,68,69]. Of note, some of the utility of the FAB, when correlated in these specific ways, has been validated only in industrially insured patients [69], and the FAB-W has been validated for currently—or recently—working patients [36,68,70]. Thus, the present FAB results are difficult to analyze based on previous validations, as the current study did not exclude participants who identified as disabled or retired, nor did it control for insurance payer type.

The BPI scores warrant notice in that their improvement implies a general lack of suffering among the participants, despite LTOT cessation.

#### *4.2. Psychological Assessment Scores Lacking Significant Improvement Post-LTOT Cessation*

Equally interesting is the lack of association seen between LTOT cessation and daytime sleepiness, generalized anxiety, and kinesiophobia (assessed via the ESS, GAD, and TSK, respectively). Nighttime sleep disturbance and daytime impaired cognition are common anecdotal complaints among patients on LTOT. Disturbed sleep architecture and increased obstructive and central sleep apnea have long been recognized as side effects of chronic opioid use [70,71]. Thus, it was surprising that LTOT cessation did not pair with improved daytime sleepiness. However, this outcome is consistent with previous studies documenting that sleep disturbance does not necessarily correlate with daytime sleepiness, as measured by the ESS, in chronic opioid users [70–72].

The improvement in anxiety (assessed via the GAD) between pre- and post-LTOT cessation was not significant. Previous studies have suggested that a self-medication model has been implicated in inciting, or exacerbating, opioid use disorder for patients with clinically significant anxiety [12,16,51], and higher anxiety scores have also been implicated in the increased length of post-operative opioid analgesic use [10]. Moreover, studies have found buprenorphine administration to be correlated with improvement in anxiety [47,56]. The relationship between anxiety and LTOT reliance warrants increased study to help reconcile present and previous findings.

Kinesiophobia (TSK) scores did not follow the significant improvement trajectory of the PCS and FAB- PA. These tools are often grouped together in the category of anxiety, or fear-based, belief and behavior assessments. However, previous investigation into their interchangeability has failed to show cross-over reliability [73]. Of note is the fact that many participants chose not to complete the TSK, for unknown reasons, and it was by far the most neglected of the tests by the participants, with an "n" of 34. Perhaps the smaller number of datasets affected the significance of this outcome.

#### *4.3. Contribution to the Field*

The present findings help to identify psychological improvements associated with successful LTOT cessation. While this study does not prove whether the intervention of the medication changes improved the mood outcomes, or whether the multidisciplinary program promoted a foundation of mood change that allowed for LTOT cessation, it is clear that specific mood improvements were associated with LTOT cessation. With further study, this dynamic may eventually suggest inroads to safer and more optimal treatments in the field of CNCP. Just as previous studies have shown that addressing high fear avoidance and catastrophizing beliefs improved disability measures [48,74], the current findings

could provide the foundation to target specific psychological states for education and multidisciplinary support to aid in successful LTOT cessation in future trials.
