Withdrawal catastrophizing scale: initial psychometric properties and implications for the study of opioid use disorder and hyperkatifeia

ABSTRACT

Background: Discovery of modifiable factors influencing subjective withdrawal experience might advance opioid use disorder (OUD) research and precision treatment. This study explores one factor – withdrawal catastrophizing – a negative cognitive and emotional orientation toward withdrawal characterized by excessive fear, worry or inability to divert attention from withdrawal symptoms.

Objectives: We define a novel concept – withdrawal catastrophizing – and present an initial evaluation of the Withdrawal Catastrophizing Scale (WCS).

Methods: Prospective observational study (n = 122, 48.7% women). Factor structure (exploratory factor analysis) and internal consistency (Cronbach’s α) were assessed. Predictive validity was tested via correlation between WCS and next-day subjective opiate withdrawal scale (SOWS) severity. The clinical salience of WCS was evaluated by correlation between WCS and withdrawal-motivated behaviors including risk taking, OUD maintenance, OUD treatment delay, history of leaving the hospital against medical advice and buprenorphine-precipitated withdrawal.

Results: WCS was found to have a two-factor structure (distortion and despair), strong internal consistency (α = .901), and predictive validity – Greater withdrawal catastrophizing was associated with next-day SOWS (r_s (99) = 0.237, p = .017). Withdrawal catastrophizing was also correlated with risk-taking behavior to relieve withdrawal (r_s (119) = 0.357, p < .001); withdrawal-motivated OUD treatment avoidance (r_s (119) = 0.421, p < .001), history of leaving the hospital against medical advice (r_s (119) = 0.373, p < .001) and buprenorphine-precipitated withdrawal (r_s (119) = 0.369, p < .001).

Conclusion: This study provides first evidence of withdrawal catastrophizing as a clinically important phenomenon with implications for the future study and treatment of OUD.

KEYWORDS:

• Opioid-related disorders
• opioid addiction
• substance withdrawal syndromes
• drug withdrawal symptoms
• catastrophizing

Introduction

Withdrawal is known to play a central role in the neuropathogenesis of opioid addiction (1–4). Individuals with opioid use disorder (OUD) experience cycles of withdrawal which come to drive compulsive opioid use (5). This process is characterized by negative reinforcement. Over time, the removal of withdrawal becomes the primary motive for opioid use (4,6–8). Repeated episodes of withdrawal are believed to gradually sensitize brain-stress-systems – an effect known as hyperkatifeia (9–11). Hyperkatifeia involves the progressive worsening of emotional symptoms of withdrawal. Such negative emotions are also thought to worsen the subjective experience of physical withdrawal symptoms, including pain (12). Indeed, hyperkatifeia and pain are linked in the brain (10,13–17). The originators of this concept have called hyperkatifeia the “emotional corollary of hyperalgesia,” because the neural substrates of pain perception and hyperkatifeia are so significantly overlapping (13,18,19).

Despite extensive literature characterizing the role of withdrawal in addiction neurobiology, less is known about individual difference factors that might modulate subjective withdrawal experience or withdrawal motivated behavior. Prior research has observed substantial between-person variability in withdrawal severity, unaccounted for by pharmacologic or physiologic factors (20–23). Studies of modifiable factors influencing the subjective experience of opioid withdrawal are lacking. One such factor might be “catastrophizing” (24–26).

Catastrophizing describes a tendency to magnify or exaggerate the threat value or severity of aversive stimuli and is associated with excessive fear, worry or an inability to divert attention from such stimuli (27). Catastrophizing has been conceptualized as a “maladaptive cognitive style” involving cognitive distortions that contribute to the development and maintenance of negative affectivity (28). This phenomenon has been well studied in multiple psychopathologies and chronic pain (24,27,29–33). Catastrophizing is related to symptom burden in depression and anxiety and contributes to emotional dysregulation in borderline personality disorder (34–41). In chronic pain, catastrophizing has been linked to a variety of negative clinical outcomes and is a target for psychosocial pain management interventions (42,43). Pain catastrophizing has even been offered as a potential therapeutic target to improve substance use treatment outcomes (26). However, to our knowledge, no prior study has described withdrawal-specific catastrophizing – hereafter referred to as “withdrawal catastrophizing” - or sought to measure withdrawal catastrophizing among individuals with OUD.

Given withdrawal catastrophizing might plausibly modulate withdrawal-related negative reinforcement in OUD, the development of an instrument to measure withdrawal catastrophizing could be of considerable importance (44). Such an instrument might inform new OUD treatments or serve as a control variable in future OUD clinical trials (11). Therefore, the present work intends to fill this gap in the literature. Described is a novel instrument, the Withdrawal Catastrophizing Scale (WCS) adapted from the reference standard psychometric tool for measuring pain catastrophizing – the Pain Catastrophizing Scale (PCS) (45,46). The theoretical basis for adapting a withdrawal catastrophizing instrument from a pain catastrophizing instrument is informed by the overlapping neurobiology of hyperkatifeia and chronic pain. For example, the central nucleus of the amygdala (CeA) – a key structure dysregulated in hyperkatifeia – is also known as the “nociceptive amygdala” for its role in the emotional processing of pain (47).

The present study aimed to introduce the concept of withdrawal catastrophizing and provide initial psychometric evidence for WCS including factor structure, internal consistency, and predictive validity. Additionally, the clinical salience of WCS was evaluated by correlation, with questions pertaining to withdrawal-motivated behaviors and buprenorphine-precipitated withdrawal. Specifically assessed were withdrawal-motivated risk-taking behaviors, OUD maintenance, OUD treatment delay, history of leaving the hospital against medical advice and buprenorphine-precipitated withdrawal.

Materials and methods

This was a prospective observational study with a cross-sectional and a longitudinal component. Approval was obtained prior to study initiation from the Ohio State University Wexner Medical Center (OSUWMC) Institutional Review Board. Participants provided verbal consent and were monetarily compensated.

Participants

A total of 124 participants were recruited between May 8, 2022, and December 16, 2022, from the Ohio State University Wexner Medical Center (OSUWMC) Talbot Hall. Talbot Hall provides comprehensive addiction care, including admission for medically supervised withdrawal. At Talbot Hall, admitted patients receive 24-hour nursing care, group counseling, case management and are evaluated daily by an Addiction Medicine specialist physician. Withdrawal management is protocolized. Per protocol, all patients are offered clonidine, acetaminophen, magnesium hydroxide, promethazine, ondansetron, bisacodyl, dicyclomine, hydroxyzine and melatonin to ease symptoms of withdrawal. Patients may choose to be initiated on buprenorphine/naloxone with a standard dose and frequency of 4/1 mg every 6 hours. Patients may also elect to receive tramadol for myalgia and arthralgia with a tapering schedule as follows: 100 mg every six hours for 12 doses, 100 mg every eight hours for nine doses, 100 mg every twelve hours for 4 doses. Recruitment was conducted by trained staff during a routine admission assessment.

Participants were adults presenting for medically supervised withdrawal from opioids. All participants were diagnosed with OUD at the time of recruitment. OUD was defined by the presence of at least 2 of 11 OUD criteria during the past 12 months as described in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) (48). Exclusion criteria included concurrent use of or withdrawal from alcohol, benzodiazepines, or other central nervous system depressants, inability to provide informed consent, read or comprehend survey items. Eight individuals declined participation. Two were excluded for concurrent alcohol withdrawal. The final sample was 122 (n = 122).

Participants completed an initial survey at the time of admission (hospital day 1) which included demographic information, WCS, and seven additional original items. Beginning the following day (hospital day 2) they were asked to rate their withdrawal using the Subjective Opiate Withdrawal Scale (SOWS) once daily for up to four days (hospital days 2–5). Survey responses were collected using REDCap (49). Participants privately accessed the surveys on tablet computers and did not interact with one another or view others’ responses.

Measures

Subjective Opiate Withdrawal Scale (SOWS)

Participants rated the severity of their withdrawal using SOWS (20,50). SOWS is a 16-item self-report instrument (range 0–64) designed to assess the symptoms of opiate withdrawal. Individual items are scored (0) Not at all (1), A little (2), Moderately (3), Quite a bit (4), Extremely. SOWS has previously been shown to have high internal consistency, good construct validity and moderate-to-high test–retest reliability (20,50).

Withdrawal Catastrophizing Scale (WCS)

WCS is a 13-item scale (range 0–52) intended to measure catastrophizing about withdrawal. WCS was adapted from the widely used Pain Catastrophizing Scale (PCS), which has been referred to as the reference standard psychometric tool for measuring pain catastrophizing and is one of the nine core data elements required by the National Institutes of Health (NIH) Helping End Addiction Long-Term (HEAL) Initiative (45,46). The rationale for creating WCS was based on the fact that both pain and withdrawal are aversive conditions, and their temporary relief by opioids may be negatively reinforcing for individuals with OUD (4,8). WCS responses were scaled as not at all (0), to a slight degree (1), to a moderate degree (2), to a great degree (3) all the time (4). WCS is displayed in the results (see Table 2).

Table 1. Sample characteristics.

Characteristic	Participants (n = 122)
Age mean (SD)
Years	36.4 (9.60)
Gender n (%)
Female	58 (48.7)
Male	61 (51.3)
Racial Identity n (%)
Black	37 (31.1)
White	80 (67.2)
American Indian or Alaska Native	1 (0.8)
Asian	2 (1.7)
More than one race	1 (0.8)
Ethnicity n (%)
Hispanic	4 (3.4)
Non-Hispanic	115 (96.6)
OUD Severity mean (SD)
Number of DSM-5 criteria present	10.8 (0.9)
Opioids Used n (%)
Heroin	11 (9.2)
Illicitly Manufactured Fentanyl	110 (92.4)
Prescription Opioid	15 (12.6)
Any Other Opioid	4 (3.4)
Route of Opioid Administration n (%)
Intravenous	34 (28.6)
Intramuscular or subcutaneous	1 (0.8)
Smoking	53 (44.5)
Snorting/sniffing up the nose	67 (56.3)
Swallowing by mouth	11 (9.2)
Another Route	3 (2.5)
Monetary Cost of Daily Opioid Use mean (SD)
US dollars per day	102.4 (88.8)
History Of Prior OUD Remission n (%)
No prior episodes of remission	28 (23.5)
Prior early remission	91 (76.5)
Prior sustained remission	64 (53.8)

Three participants were missing data for age, race, ethnicity, gender, types of opioids used, route of administration and history of prior OUD remission. Three participants did not respond to cost of daily use and ten responded “I do not know.” Prior history of early and sustained OUD remission was operationalized per DSM-5. All participants had active OUD at the time of participation.

Table 2. Withdrawal catastrophizing scale (WCS).

	Rotated Factor Loading
Questionnaire Item	Factor 1	Factor 2	Communalities	Item Mean (SD)
1. I worry all the time whether the opioid withdrawal sickness will end.	.463	.250	.424	2.76 (1.19)
2. I feel I can’t go on.	−.211	.878	.582	1.83 (1.23)
3. It’s terrible and I think it’s never going to get any better.	−.049	.788	.574	2.09 (1.31)
4. It’s awful and I feel that it overwhelms me.	.069	.690	.541	2.81 (1.16)
5. I feel I can’t stand it anymore.	.181	.628	.571	2.76 (1.18)
6. I become afraid that the opioid withdrawal sickness will get worse.	.400	.402	.525	2.81 (1.19)
7. I keep thinking of other times I’ve had opioid withdrawal sickness.	.570	.131	.436	2.76 (1.17)
8. I anxiously want the opioid withdrawal sickness to go away.	.829	−.121	.575	3.48 (0.83)
9. I can’t seem to keep it out of my mind.	.254	.423	.378	2.96 (1.04)
10. I keep thinking about how sick I feel.	.544	.120	.394	2.99 (1.06)
11. I keep thinking about how badly I want the opioid withdrawal sickness to stop.	.958	−.207	.709	3.36 (0.82)
12. There’s nothing I can do to reduce the intensity of my symptoms except to use opioids.	.358	.308	.362	2.59 (1.21)
13. I wonder whether something serious may happen.	.637	.091	.487	2.44 (1.34)

Factor 1 = Distortion, factor 2 = Despair. All questions were prefaced “We are interested in the types of thoughts and feelings you have when you have OPIOID WITHDRAWAL SICKNESS. Please rate how much each of the following describe how you respond when faced with OPIOID WITHDRAWAL SICKNESS. When I have OPIOID WITHDRAWAL SICKNESS … ” Responses were scaled as not at all (0), to a slight degree (1), to a moderate degree (2), to a great degree (3) all the time (4).

Seven original items pertaining to withdrawal-motivated behavior and buprenorphine-precipitated withdrawal

Seven original questions captured withdrawal-related risk-taking behaviors, OUD maintenance, treatment delay, and buprenorphine precipitated withdrawal. These questions were not part of WCS but instead were written in order to evaluate the potential clinical salience of WCS. Items 1–3 asked how frequently participants engaged in risk-taking behaviors to escape opioid withdrawal and were coded on a six-option Likert scale from (0) never, to (5) very often. These items were based on prior research identifying accidental injury, shame, and arrest or incarceration as relevant risks to individuals who use substances (51–56). Item 4 assessed history of leaving the hospital against medical advice due to uncontrolled withdrawal (57). This item was informed by literature suggesting opioid withdrawal might motivate premature patient-directed discharge (58). Items 5 and 6 measured aversion to withdrawal as a motive for continuing opioid use and delaying OUD treatment. These questions were adapted from the Pain-related OUD Exacerbation Scale (PrOUD-ES) (59,60). Item 7 inquired about whether participants had ever experienced buprenorphine precipitated withdrawal. Buprenorphine-precipitated withdrawal is considered a clinically significant barrier to successful initiation of this treatment for OUD (61,62). Items 4–7 were coded strongly disagree (0), disagree (1), neutral (2), agree (3), strongly agree (4). Original items are presented in the results (Table 3).

Table 3. Correlations between the withdrawal catastrophizing scale (WCS) and original items.

Item	WCS total score	WCS factor 1	WCS factor 2
1. In order to get well from OPIOID WITHDRAWAL SICKNESS, I do things that are dangerous, or that could get me hurt.	.357**	.238**	.407**
2. In order to get well from OPIOID WITHDRAWAL SICKNESS, I do things that make me feel ashamed.	.366**	.306**	.373**
3. In order to get well from OPIOID WITHDRAWAL SICKNESS, I do things that could get me into trouble with the law.	.330**	.239**	.350**
4. In the past, I have left the hospital against medical advice, or before finishing treatment I needed, because of OPIOID WITHDRAWAL SICKNESS.	.373**	.311**	.415**
5. The most important reason I have been unable to stop using opioids is OPIOID WITHDRAWAL SICKNESS.	.287**	.307**	.215*
6. I have avoided or put off getting addiction treatment because I was worried about going through OPIOID WITHDRAWAL SICKNESS.	.421**	.374**	.425**
7. In the past, I have gotten OPIOID WITHDRAWAL SICKNESS “aka precipitated withdrawal” from taking Buprenorphine too soon after my last dose of another opioid.	.369**	.276**	.398**

Factor 1 = Distortion, factor 2 = Despair. Items one through three were scaled never (0), very rarely (1), rarely (2), occasionally (3), often (4), very often (5). Items four through seven were scaled strongly disagree (0), disagree (1), neutral (2), agree (3), strongly agree (4). **Correlation is significant at the 0.01 level (2-tailed). *Correlation is significant at the 0.05 level (2-tailed).

Data analysis

An exploratory factor analysis (EFA) was run on WCS (n = 122). Although WCS adapted from PCS, the structure of this new scale was uncertain. Therefore, exploratory, not confirmatory, factor analysis was appropriate (63). Common factor analysis was chosen over principal components analysis as the goal of this study was to assess WCS’s latent structure (64). Extraction was achieved by principal axis factoring due to the relative tolerance of multivariate nonnormality of principal axis factoring, and its superior recovery of weak factors (65). Minimum average partial analysis and parallel analysis, complemented by screen plot inspection, determined factor retention (66–69). Promax rotation was selected (70). To optimize the practical (10% variance explained) and the statistical significance (p < .05), salience threshold, or the minimum factor loading to be considered meaningful, was predetermined as 0.32 (71). Also predetermined was a maximum acceptable factor correlation of 0.8. Kaiser–Meyer–Olkin (KMO) was used to evaluate sampling adequacy. Bartlett’s test of sphericity was conducted to ensure that the correlation matrix differed significantly from an identity matrix.

The predictive validity of WCS was evaluated using Spearman’s rank order correlations between WCS (measured at the time of admission) and SOWS (collected on hospital day 2) (72). Due to the large amounts of missing SOWS data on hospital day 3 and beyond, assessment of relationships between WCS and subjective withdrawal experience on subsequent hospital days was not possible (see section Subjective Withdrawal & Opioid Medication Use below for further detail). Spearman’s correlations between the seven original items and WCS assessed the clinical salience of withdrawal catastrophizing. All correlations were two-tailed and deemed significant at α < 0.05. Statistical analyses were performed using SPSS software (Version 28.0, SPSS. Inc).

Results

Sample characteristics

Demographics were obtained from 119 (97.5%) participants. Table 1 summarizes sample characteristics.

Subjective withdrawal & opioid medication use

Participants were offered SOWS by the attending Addiction Medicine physician once daily during rounds on hospital days two through five (hospital day 2 was the following day after WCS was completed). One hundred and one participants (N = 101) completed SOWS on day 2, 82 on day 3, 44 on day 4, and 11 on day 5. Twenty-one participants completed no SOWS assessments. One participant declined SOWS. The remaining 20 were lost to premature discharge (transfer to acute care hospital for COVID-19, unstable medical comorbidity or left against medical advice). There were no substantive differences between those who did not complete SOWS and those who completed SOWS based on age (completed SOWS Mdn = 35, did not complete SOWS Mdn = 35), U = 1016.0, z = −0.091, p = .928; OUD severity (completed SOWS Mdn = 11, did not complete SOWS Mdn = 11), U = 1025.0, z = −0.435, p = .663 or degree of withdrawal catastrophizing (completed SOWS Mdn = 34.0, did not complete SOWS Mdn = 37.5), U = 917.5, z = −0.971, p = .332. The sample mean, minimum, and maximum day 2 SOWS score was 28.2 (SD = 15.6), 0 and 64, respectively.

Data regarding buprenorphine/naloxone and tramadol use were missing from the 21 participants who completed no SOWS assessments. Of those who completed at least one SOWS (N = 101), 58 (57.4%) participants were recorded to have received buprenorphine/naloxone and 45 (44.6%) were recorded to have received tramadol at any point during admission. Eleven (11, 10.9%) received both tramadol and buprenorphine/naloxone. At the time of the day 2 SOWS assessment, 25 (24.8%) participants were recorded to have received buprenorphine/naloxone and 42 (41.6%) tramadol. The day 2 SOWS scores of participants who received buprenorphine/naloxone did not differ from those who had not received buprenorphine/naloxone at the time SOWS was measured (received buprenorphine/naloxone Mdn = 29, had not received buprenorphine/naloxone SOWS Mdn = 27), U = 907.5, z = −0.335, p = .738. Similarly, day 2 SOWS did not differ by receipt of tramadol (received tramadol Mdn = 24.5, had not received tramadol SOWS Mdn = 30), U = 1142.5, z = −0.665, p = .506. Degree of withdrawal catastrophizing also did not differ between those who did or did not receive buprenorphine/naloxone (received buprenorphine/naloxone WCS Mdn = 34, had not received buprenorphine/naloxone WCS Mdn = 38.5), U = 807.5, z = −1.122, p = .262; or tramadol (received tramadol WCS Mdn = 36.5, had not received tramadol WCS Mdn = 37), U = 1232.5, z = −0.045, p = .964.

Seven original items pertaining to withdrawal-motivated behavior and buprenorphine-precipitated withdrawal

The seven original items had a good internal consistency as determined by a Cronbach’s alpha of 0.78. Responses to items 1–3 revealed variability in the frequency with which participants endorsed engaging in risk-taking behaviors to relieve opioid withdrawal. A small minority of participants reported they never engaged in risk-taking behaviors to relieve withdrawal: 20 (16.5%) never put themselves in danger of being hurt; 16 (13.2%) never did things that made them feel ashamed; and 23 (19.0%) never broke the law. A larger percentage indicated they often or very often risked their safety 45 (36.9%), performed shameful 38 (31.1%), or illegal acts 46 (37.7%) to relieve opioid withdrawal sickness. The majority of participants agreed or strongly agreed with items 4, 5, and 6 indicating that most participants had left the hospital before completing needed medical care due to uncontrolled withdrawal (61, 60.4% of the 101 who had ever been hospitalized); found withdrawal to be the most important reason they had struggled to stop using opioids (105, 86.1%); and had delayed accessing OUD treatment because of opioid withdrawal (93, 76.2%). A majority also reported they had previously experienced buprenorphine-precipitated opioid withdrawal (69, 61.1% of 113 participants who had tried buprenorphine). Agreement with items 4–7 is illustrated by Figure 1.

Figure 1. Bar graph representing the percent of participants who endorsed agreement, or strong agreement with each displayed statement.

Psychometric analysis of WCS

All 122 participants (100%) completed WCS. Responses to WCS were non-normally distributed per Shapiro–Wilk test (p = .006) and skewed toward greater withdrawal catastrophizing −.276 (standard error = .219). The median, minimum and maximum total WCS scores were 35.5, 11.0 and 52.0, respectively. Total WCS scores reflected prevalent catastrophizing about opioid withdrawal in this sample of individuals with OUD. Nearly 70% (85, 69.7%) scored 30 or above on WCS – the threshold for clinically significant catastrophizing on PCS, the measure on which WCS was based (73).

Factor structure and internal consistency

Exploratory factor analysis (EFA) was run on the 13 item WCS (n = 122). EFA appropriateness was assessed prior to extraction. A correlation matrix showed all 13 variables had correlation coefficients greater than 0.3. Kaiser–Meyer–Olkin (KMO) was meritorious (0.87) for WCS as a whole and individual questionnaire item KMO measures were all greater than 0.8. Bartlett’s test of sphericity indicated that the correlation matrix differed significantly from an identity matrix (p < .001). Due to multivariate non-normality, extraction was achieved via principal axis factoring.

EFA revealed two factors with an eigenvalue greater than one, and a third factor with an eigenvalue very near to one (0.95). Potential factors one, two and three explained 46.52%, 10.91% and 7.31% of the total variance, respectively. Parallel analysis showed that the observed eigenvalue of factor 1 (5.59) and factor 2 (0.99) exceeded the 95th percentile mean eigenvalue of factor 1 (0.87) and factor 2 (0.67) obtained from generating 1000 permutations of the original data with preserved distributions (67). However, the observed eigenvalue of potential factor 3 (0.48) was less than the 95th percentile mean eigenvalue (0.52) in this parallel analysis. Therefore, parallel analysis supported retention of two factors. The minimum average partial analysis revealed that the smallest mean squared partial correlation was 0.04 and the smallest mean 4^th power partial correlation was 0.003, which corresponded to a two-factor solution. Finally, visual inspection of the scree plot also indicated that two factors should be retained.

The two-factor solution explained 57.4% of the variance and exhibited a largely simple structure. Cross-loading above the predetermined threshold of salience (.32) was noted on item 6 only. Item 6 loaded slightly higher on factor 2 (hereafter despair). Because of its slightly higher loading and apparent thematic relationship to the other items in factor 2 (despair), item 6 was considered to be a part of factor 2 (despair). As a whole, WCS exhibited a high level of internal consistency with Cronbach’s alpha of .901. Factor 1 (hereafter referred to as distortion) consisted of 7 items primarily reflecting rumination and magnification and had a Cronbach’s alpha of .862. Factor 2 (despair) contained 6 items mostly reflecting helplessness and was found to have a Cronbach’s alpha of .854. Distortion and despair were well correlated, but this relationship did not exceed the predetermined value of 0.8 (r_s (119) = 0.747, p < .001) WCS factor loadings, communalities and item means are shown in Table 2. Figure 2 is a path diagram illustrating the two-factor solution.

Figure 2. Path diagram illustrating the two-factor solution of the withdrawal catastrophizing scale (WCS) obtained from exploratory factor analysis. Rectangles representing the 13 items of WCS are shaded to represent the strength of their rotated loading on their associated factor (darker = stronger). Items are also labeled magnification (M), rumination (R), or helplessness (H) based on the schema of the pain catastrophizing scale (PCS) from which WCS is derived.

Predictive validity

Spearman’s rank-order correlation was used to assess a hypothesized relationship between baseline withdrawal catastrophizing (WCS measured at the time of admission on hospital day 1) and subjective withdrawal severity (SOWS collected during rounds on hospital day 2). Cases with missing SOWS data were excluded from analysis. The assumption of monotonicity was confirmed by scatterplot inspection. WCS was significantly associated with total SOWS (r_s (99) = 0.237, p = .017). In addition to WCS as a whole, WCS factors 1 and 2 were also assessed for predictive validity. WCS factor 1 (distortion) was not significantly correlated with SOWS (r_s (99) = 0.192, p = .055). WCS factor 2 (despair) was correlated with SOWS (r_s (99) = 0.247, p = .013). Finally, SOWS item 16 “I feel like using now” was correlated with WCS (r_s (99) = 0.299, p = .002), factor 1 (distortion) (r_s (99) = 0.225, p = .024) and factor 2 (despair) (r_s (99) = 0.309, p = .002).

Clinical salience of withdrawal catastrophizing

Clinical salience was assessed by Spearman’s rank-order correlations between WCS and a seven-item battery of original questions (Table 3) regarding withdrawal motivated risk-taking behaviors (items 1–4), withdrawal motivated OUD maintenance and treatment delay (items 5–6) and self-report of prior buprenorphine precipitated withdrawal (item 7). WCS and its two factors were significantly correlated with all original items. Figure 3 presents a) a scatterplot of the relationship between risk-taking (shameful acts) and WCS and b) a pie chart of how often participants resorted to performing shameful acts to relieve withdrawal. Table 3 lists the original items and presents the correlation coefficients of this analysis.

Figure 3. (a) Scatterplot of the relationship between risk-taking (shame) and WCS, and (b) pie chart representing how often participants found themselves doing things that made them feel ashamed to relieve withdrawal. Asterisks (*) in the key correspond to the two exploded pie pieces in the pie chart.

Discussion

This study introduces a novel concept to addiction science - withdrawal catastrophizing – which we define as a negative cognitive and emotional orientation toward withdrawal characterized by excessive fear, worry or inability to divert attention from withdrawal symptoms. We present WCS, an instrument we have adapted to measure withdrawal catastrophizing. WCS exhibited promising psychometric properties including a two-factor solution, high internal consistency, and predictive validity. Greater withdrawal catastrophizing was associated with more severe next-day withdrawal by self-report. Withdrawal catastrophizing was also related to a history of buprenorphine-precipitated withdrawal and leaving the hospital against medical advice due to uncontrolled withdrawal. Finally, withdrawal catastrophizing was associated with greater agreement among participants that withdrawal had motivated them to continue using opioids; avoid OUD treatment and perform acts that were dangerous, illegal, or made them feel ashamed. Together, these early exploratory findings suggest withdrawal catastrophizing might be an important, previously unappreciated individual difference factor related to withdrawal experience and withdrawal-motivated behavior among individuals with OUD.

Withdrawal catastrophizing appeared common and consequential. Nearly 70% of the participants scored 30 or above on WCS. While further studies are needed to determine meaningful cut scores for WCS, 30 is considered to be the threshold for clinically significant catastrophizing on WCS’s parent measure, PCS (19). A plurality of participants reported they were often or very often motivated by withdrawal to do things that were physically dangerous, illegal or made them feel ashamed. Accidental injury, incarceration and diminished self-esteem are known detrimental consequences of OUD (74–79). Sixty percent of participants who had previously been hospitalized indicated they had left the hospital against medical advice because of opioid withdrawal. Leaving the hospital against medical advice exposes patients to an increased risk of adverse outcomes, including death (80,81). Withdrawal catastrophizing was robustly correlated with all four of these high-risk withdrawal-motivated behaviors, suggesting withdrawal catastrophizing might possibly have an unrecognized relationship to some of the worst consequences of OUD.

Withdrawal catastrophizing was also correlated with treatment avoidance. Seventy-six percent of participants agreed or strongly agreed they had avoided or delayed OUD treatment for fear of going through opioid withdrawal. In 2020, only 11.2% of the 2.5 million adolescents and adults living with OUD in the United States received medication treatment (buprenorphine, methadone, or naltrexone) (82). Given the increasing pervasiveness of illicitly manufactured fentanyl and associated loss of life, treatment delay may have lethal consequences for individuals with OUD (83–86). Additional studies are needed to determine if interventions targeting withdrawal catastrophizing might improve OUD treatment engagement.

Over 60% of participants who had previously tried buprenorphine reported having experienced buprenorphine-precipitated withdrawal. Buprenorphine-precipitated withdrawal is considered a significant barrier to buprenorphine treatment and initiation protocols intended to circumvent this barrier are an area of active inquiry (61,62). This inquiry has been based on the premise that buprenorphine-precipitated withdrawal is a pharmacological problem accounted for by the partial opioid agonism of buprenorphine and exacerbated by the lipophilicity of fentanyl (87). However, our finding that withdrawal catastrophizing was correlated with buprenorphine-precipitated withdrawal strongly suggests psychological involvement. To our knowledge, no prior research has controlled for withdrawal catastrophizing when investigating withdrawal treatments or medications for OUD like buprenorphine. WCS might one day be considered a useful control variable in clinical trials of emerging treatments for opioid withdrawal and OUD (88). Additional research is needed to determine if WCS might possibly predict treatment response or assist with mechanism identification (i.e. hyperkatifeia).

WCS may also provide insight into the withdrawal/negative affect stage of the brain disease model of addiction. Eighty-six percent of participants agreed or strongly agreed withdrawal was the most important reason they kept using opioids. WCS was correlated with agreement with this statement, suggesting withdrawal catastrophizing might relate to the refractoriness of OUD as a chronic health condition. Our findings are consistent with the observations of Koob in describing hyperkatifeia and the withdrawal/negative affect stage of addiction (4,8,9,11,89). We build upon this literature by demonstrating a relationship between withdrawal catastrophizing and withdrawal as a key motive for continued opioid use in OUD. As a measure of emotional orientation to withdrawal, WCS might aid community and epidemiological studies of hyperkatifeia, a recently identified research area of high interest to the NIH National Institute on Alcohol Abuse and Alcoholism (19).

Study strengths included adequate sample size and low levels of missing data on the day 1 survey (WCS, demographics). There were also notable limitations. Approximately 17% of day 2 SOWS responses were missing, requiring those cases to be removed from the predictive validity analysis. However, participants excluded from predictive validity analysis were not different from included participants on age, OUD severity or degree of withdrawal catastrophizing. Missing data precluded assessment of predictive validity on day-three and beyond. SOWS was collected once daily during physician rounds. We did not control for the timing of WCS in relation to last use of opioids or other substances or medications. Collection of SOWS by ecological momentary assessment might have yielded more complete data. Additional research is needed to further establish the reliability and validity of WCS.

The present study defined an original concept withdrawal catastrophizing – a negative cognitive and emotional orientation toward withdrawal characterized by excessive fear, worry or inability to divert attention from withdrawal symptoms – and proposed a new scale (WCS) to measure withdrawal catastrophizing among individuals with OUD. WCS exhibited promising initial psychometric properties including: a two-factor solution, internal consistency and predictive validity. Withdrawal catastrophizing was associated with withdrawal-motivated risk-taking, including the performance of dangerous, illegal, or shameful acts to relieve withdrawal and leaving the hospital against medical advice. Withdrawal catastrophizing was also related to next-day withdrawal severity, history of buprenorphine-precipitated withdrawal and OUD treatment avoidance. Taken together, these findings provide first evidence of withdrawal catastrophizing as a clinically important phenomenon with implications for the future study of hyperkatifeia and withdrawal-motivated behavior in OUD.

Acknowledgments

The authors would like to thank Kirk Carruthers, MD; Bradley Lander, PhD; Mamie Martin, RN, MSN, APRN-CNP, DNP; Leah Brown, MS, MPH, APRN-CNP; Jessica Belser, MSW, MS, PMHNP-BC, CARN-NP and Megan Ackley APRN-CNP for their support in conducting this study.

Disclosure statement

Dr. Hall provided expert opinion regarding the opioid overdose crisis to Lumanity and Emergent BioSolutions.

Additional information

Funding

Funding was provided by the Care Innovation and Community Improvement Plan (CICIP), a program of the Ohio Department of Medicaid. The views expressed in this publication do not necessarily reflect the official policies of the Ohio Department of Medicaid nor does mention of trade names, commercial practices, or organizations imply endorsement by the government of Ohio.