Day-to-day Hedonic and Calming Effects of Opioids, Opioid Craving, and Opioid Misuse Among Patients With Chronic Pain Prescribed Long-term Opioid Therapy

Leah Frimerman; Maria Verner; Amanda Sirois; Katherine Scott; Alice Bruneau; Jordi Perez; Yoram Shir; Marc O. Martel

Disclosures

Pain. 2021;162(8):2214-2224. 

In This Article

Results

Descriptive Statistics

Descriptive statistics for the study sample are presented in Table 1. The sample included 49.5% of women and primarily Caucasians (80.6%). The average pain duration of the sample was 12.4 years (SD = 11.9) and the majority of patients (72.8%) experienced back pain. Opioid regimen characteristics included the prescription of long-acting opioids for 39.8% of patients and short-acting opioids for 86.4% of patients. Up to 26.2% of patients were taking both types of opioid formulations. The average oral morphine equivalent daily dose was 67.67 mg/d (SD = 101.1). The frequency of different types of opioid misuse behaviors reported by patients is shown in Figure 1.

Figure 1.

Frequency of opioid misuse behaviors reported by patients over the 14-day period. Figure 1 corresponds to the percentage (%) of patients having reported each of the opioid misuse behaviors, at different frequencies, over the past 14 days.

Before conducting main study analyses, the potential influence of patient demographics, opioid regimen characteristics, pain characteristics, and psychological or psychiatric variables on primary study outcomes was examined. Results indicated that men reported significantly greater opioid craving than women (B = 9.15, P = 0.041). Results also indicated that a more frequent daily opioid dosage was associated with greater opioid craving (B = 1.93, P = 0.042), and higher DAST-10 scores were associated with heightened reports of opioid craving (B = 4.11, P = 0.001) as well as with more frequent opioid misuse behaviors (rs = 0.203, P = 0.045). None of the other variables were significantly associated with main study outcomes.

Contribution of Hedonic and Calming Effects of Opioids to Prescription Opioid Misuse

Spearman correlation analyses were conducted to examine whether the hedonic and calming effects of opioids contributed to opioid misuse (COMM-Total). As shown in Table 2, feelings of pleasure (r s = 0.08, P = 0.42) and calmness (r s = −0.10, P = 0.30) were not significantly associated with COMM-Total. Correlational analyses also indicated that hedonic and calming effects were not significantly associated with any specific type of opioid misuse behaviors, such as overusing opioids, using opioids for non-pain symptoms, or borrowing/shopping opioids (Table 2).

For exploratory purposes, analyses were also conducted to examine whether opioid misuse behaviors were associated with patients' average daily levels of pain intensity, catastrophizing, NA, or PA. Results revealed that more frequent use of opioids for non-pain symptoms was associated with higher daily levels of catastrophizing (rs = 0.20, P = 0.042) and NA (rs = 0.35, P = 0.000) as well as with lower levels of PA (rs = −0.25, P = 0.012). Results from these correlation analyses can be seen in Table 2 as well as Supplementary Figure 1 (available at http://links.lww.com/PAIN/B282).

Contribution of Hedonic and Calming Effects of Opioids to Opioid Craving

Multilevel linear regression analyses were conducted to examine the degree to which the hedonic and calming effects of opioids contributed to patients' daily levels of opioid craving. Results indicated that higher reports of pleasurable feelings were associated with heightened reports of opioid craving (B = 0.10, SE = 0.028, P = 0.000). Higher reports of calmness were also associated with heightened reports of opioid craving (B = 0.09, SE = 0.026, P = 0.001). Sensitivity analyses were then conducted to examine if these associations were influenced by person-level (ie, level 2) variables found to be associated with opioid craving (ie, age, sex, dosing frequency, and DAST-10). Given that all covariates of interest (ie, sex, dosing frequency, and DAST-10) are level 2 variables, their influence was tested by creating cross-level interaction terms to determine if level 2 variables influenced the magnitude of the association between opioid effects (ie, pleasure and calmness) and opioid craving. Results revealed a significant interaction involving sex, indicating that the associations between opioid effects (ie, pleasure and calmness) and opioid craving were stronger among men than women (both P's < 0.01). Interaction effects involving dose frequency and DAST-10 were not significant (both P's > 0.05), indicating that these variables did not influence the strength of the association between reports of pleasure and opioid craving, or the association between calmness and opioid craving.

For exploratory purposes, multilevel cross-lagged analyses were conducted to explore the potential directionality of associations between opioid effects (ie, hedonic and calming effects) and opioid craving. As can be seen from Figure 2, we first examined whether feelings of pleasure and calmness predicted next day opioid craving. Inverse associations were then explored. Results from these analyses indicated that feelings of pleasure and calmness significantly predicted next-day opioid craving. This was observed for feelings of pleasure (B = 0.08, SE = 0.02; P = 0.002) and calmness (B = 0.06, SE = 0.02, P = 0.011). Results indicated that opioid craving did not significantly predict next day reports of pleasure (B = 0.05, SE = 0.02, P = 0.054) or calmness (B = 0.02, SE = 0.02, P = 0.38.

Figure 2.

Within-person associations between pain, psychological variables, opioid effects, and opioid craving. (A) Multilevel cross-lagged analysis examining the association between feelings of pleasure and opioid craving. b = unstandardized beta coefficient. (B) Multilevel cross-lagged analysis examining the association between feelings of calmness and opioid craving.

Moderators of the Associations Between Opioid Effects (ie, Hedonic and Calming Effects) and Opioid Craving

Before conducting analyses examining the moderators of the association between opioid effects (ie, hedonic and calming effects) and opioid craving, a series of multilevel analyses were conducted to examine the associations between patients' reports of pleasure and calmness and each of the potential Level 1 moderators (ie, daily pain intensity, catastrophizing, NA, and PA). Results of these analyses indicated that patients' reports of pleasure were not significantly associated with pain intensity (B = 0.03, P = 0.27), catastrophizing (B = −0.08, P = 0.69), or NA (B = −0.16, P = 0.31), but were significantly associated with PA (B = 0.36, P = 0.015). Patients reports of calmness were not significantly associated with pain intensity (B = 0.02, P = 0.44), catastrophizing (B = −0.18, P = 0.42), NA (B = −0.04, P = 0.79), or PA (B = 0.29, P = 0.07). Multilevel analyses were then conducted between each of the potential moderators and opioid craving. As can be seen from Supplementary Figure 2 (available at http://links.lww.com/PAIN/B282), results indicated that day-to-day elevations in pain intensity (B = 0.60, SE = 0.023), catastrophizing (B = 4.6, SE = 0.177), and NA (B = 2.4, SE = 0.151) were associated with heightened reports of opioid craving (all P's < 0.001). Inversely, day-to-day elevations in PA were associated with decreased levels of opioid craving (B = - 2.3, SE = 0.142, P = 0.000). The relative contribution of opioid effects (ie, hedonic and calming effects), pain intensity, and psychological factors to daily opioid craving can be seen in Table 3.

A series of multilevel models were built to examine whether patients' daily levels of pain intensity (level 1), catastrophizing (level 1), NA (level 1), or PA (level 1) moderated the association between patients' reports of opioid-related hedonic effects and opioid craving. The same analyses were conducted to examine the moderators of the association between opioid-related calming effects and opioid craving. As can be seen from Figures 3 and 4, models were built using opioid craving as the dependent variable. In all these models, two-way interaction terms between the IVs (ie, reports of pleasure or calmness) and each of the potential moderators (ie, pain intensity, catastrophizing, NA, or PA) were specified and preceded by appropriate main effects. Results revealed significant two-way interactions between patients' reports of pleasure and daily pain intensity (B = 0.003, SE = 0.0005, P = 0.000), catastrophic thinking (B = 0.026, SE = 0.004, P = 0.000), and NA (B = 0.018, SE = 0.003, P = 0.000). Results also revealed significant 2-way interactions between patients' reports of calmness and daily pain intensity (B = 0.003, SE = 0.0005, P = 0.000), catastrophic thinking (B = 0.024, SE = 0.004, P = 0.000), and NA (B = 0.019, SE = 0.003, P = 0.000). As shown in Figures 3 and 4, results revealed that associations between opioid effects (ie, pleasure and calmness) and craving were stronger on days when patients experienced elevations in pain intensity, catastrophizing, and NA. The 2-way interaction effects involving PA were not significant (both P's > 0.05), but visual inspection of data indicated that associations between opioid effects (ie, pleasure and calmness) and craving were stronger on days when patients experienced decreases in PA (Figures 3 and 4).

Figure 3.

Within-person associations between opioid-related pleasure and opioid craving as a function of patients' daily levels of pain intensity (A), pain catastrophizing (B), negative affect (C), and positive affect (D). All variables included in the figure are based on Level 1 units. The independent variable (ie, feelings of pleasure) is within-person centered and thus represent daily deviations over (or below) each patient's average levels of pleasure across the 14-day period. The moderators (ie, daily pain intensity, catastrophizing, negative affect and positive affect) are also within-person centered and represent daily deviations (ie, −2 SD, −1SD, +1 SD, and + 2 SD) from each patient's average levels across the 14-day period.

Figure 4.

Within-person associations between opioid-related calmness and opioid craving as a function of patients' daily levels of pain intensity (A), pain catastrophizing (B), negative affect (C), and positive affect (D). All variables included in the figure are based on level 1 units. The independent variable (ie, feelings of calmness) is within-person centered and thus represent daily deviations over (or below) each patient's average levels of calmness across the 14-day period. The moderators (ie, daily pain intensity, catastrophizing, negative affect and positive affect) are also within-person centered and represents daily deviations (ie, −2 SD, −1SD, +1 SD, and + 2 SD) from each patient's average levels across the 14-day period.

processing....