Understanding the Pain Paradox in Borderline Personality Disorder
School
St. John's University**We aren't endorsed by this school
Course
PSY 621
Subject
Psychology
Date
Dec 10, 2024
Pages
11
Uploaded by SargentElementMosquito33
The Pain Paradox: Borderline Personality Disorder Features, Self-HarmHistory, and the Experience of PainRyan W. Carpenter and Timothy J. TrullUniversity of MissouriIndividuals with borderline personality disorder (BPD), compared to controls, report a relative absenceof acute pain. In contrast, BPD is overrepresented among chronic pain patients, suggesting theyexperience a relative excess of chronic pain. To date, this “pain paradox” has been only partiallyexplored; no study has examined both acute and chronic pain in the same sample. In addition, previousresearch has not fully examined the effect of nonsuicidal self-injury (NSSI) on either acute or chronicpain experience in BPD. Undergraduates (N206), oversampled for those high in BPD features,completed a Cold Pressor Task (CPT), rating their pain every 15 s over a maximum of 4 min. Followingthe CPT, participants completed measures of BPD features, NSSI history, past-year pain, and perceivedpain tolerance. Results did not support the expected negative association between BPD features and acutepain. Multilevel modeling revealed an interaction of BPD features and NSSI history on CPT pain ratings:Among individuals in the no-NSSI group, BPD features were associated with greater acute pain. Amongindividuals in the NSSI group, BPD features were not significantly associated with acute pain. Resultsfor past-year pain indicated that BPD features were associated with greater past-year pain regardless ofNSSI history. This finding, coupled with the difference in the association of BPD features and acute painbetween the NSSI and no-NSSI groups provides tentative evidence that the combination of BPD featuresand NSSI history, among nonclinical samples, is linked to a pain paradox.Keywords:acute pain, borderline personality disorder, chronic pain, nonsuicidal self-injury, pain paradoxBorderline personality disorder (BPD) can be described as adisorder of extremes that involves stark shifts in affect, behavior,and experience. BPD individuals may, at one moment, feel intenseemptiness, dissociation, or social isolation and, at another, beoverwhelmed with intense emotions, urges to engage in impulsivebehavior, or idealization of a loved one. Another domain in whichBPD individuals report widely disparate experiences is in theexperience of pain. When exposed to acute painful stimuli (i.e.,pain that is transient or short-term, often as the result of an injury),BPD individuals report experiencing less pain and show less painsensitivity than individuals without BPD. In contrast, BPD is alsoassociated with increased chronic pain (i.e., pain that persistsbeyond the typical healing period).Sansone and Sansone (2007)labeled these seemingly inconsistent findings the “pain paradox”of BPD.The bulk of the evidence for a relative absence of pain underacute conditions comes from pain induction studies, which haveused a diverse range of painful stimuli, including cold (coldpressor task; CPT), heat (e.g., radiant heat), chemical (capsaicin),and mechanical (e.g., tourniquet pain task). These studies haveconsistently found that BPD patients report less pain compared tocontrols (Bohus et al., 2000;Ludäscher et al., 2007;Ludäscher etal., 2009;Magerl, Burkart, Fernandez, Schmidt, & Treede, 2012;McCown, Galina, Johnson, DeSimone, & Posa, 1993;Niedtfeld etal., 2010;Pavony & Lenzenweger, 2014;Russ et al., 1992;Russ,Campbell, Kakuma, Harrison, & Zanine, 1999;Schmahl et al.,2004;Schmahl et al., 2006;Schmahl et al., 2010). In addition,studies have shown that BPD patients report less pain than indi-viduals with posttraumatic stress disorder, bulimia nervosa, majordepressive disorder, and other personality disorders (McCown etal., 1993;Pavony & Lenzenweger, 2014;Schmahl et al., 2010). Asnoted, all of these studies included samples of BPD patients. Toour knowledge, no previous work has examined the relationship ofBPD features and acute pain in a nontreatment seeking sample.Fewer studies have examined the case for a relative excess ofchronic pain in BPD. However, here, too, the results are consistent:BPD and chronic pain are highly associated. In a review,Sansoneand Sansone (2012)found that, across studies, BPD was present in30% of chronic pain patients (compared to prevalence in thegeneral population of 1–3%;Lenzenweger, Lane, Loranger, &Kessler, 2007;Trull, Jahng, Tomko, Wood, & Sher, 2010). BPDwas also prevalent among patients seeking treatment for chronicheadaches (for a review, seeSaper & Lake, 2002), and BPDfeatures were correlated with pain report among patients seekingtreatment for pain (Sansone, Mueller, Mercer, & Wiederman,2010;Tragesser, Bruns, & Disorbio, 2010). All of these studiesexamined the prevalence of BPD and BPD features in chronic painsamples. The only study to examine chronic pain in a BPD sampleThis article was published Online First February 23, 2015.Ryan W. Carpenter and Timothy J. Trull, Department of PsychologicalSciences, University of Missouri.We thank Sean P. Lane, Phillip K. Wood, Simon L. McCabe, and KristinD. McLaughlin for their feedback on earlier drafts of this paper, andRichard P. Zeuner for his assistance with study design.Correspondence concerning this article should be addressed to RyanW. Carpenter, Department of Psychological Sciences, University ofMissouri, 210 McAlester Hall, Columbia, MO 65211. E-mail:rwc449@mail.missouri.eduThis document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.Personality Disorders: Theory, Research, and Treatment© 2015 American Psychological Association2015, Vol. 6, No. 2, 141–1511949-2715/15/$12.00http://dx.doi.org/10.1037/per0000112141
wasFrankenburg and Zanarini (2004), who found that pain-relatedsyndromes were more prevalent in BPD patients with active symp-toms than those in remission. Indeed, while the prevalence of BPDamong chronic pain patients has attracted notice in the medicalfield (Kalira, Treisman, & Clark, 2013;Sansone & Sansone,2007), it has received less attention in psychology.Thus, under acute conditions (e.g., pain induction tasks), BPDfeatures appear to be related to greater resiliency toward pain,while under chronic conditions, the effect is reversed. These find-ings support the idea that physical pain is a largely overlookeddomain in which BPD features are associated with diverging andopposing experiences. Better understanding this pain paradox ofBPD may provide new perspective on dysregulation in otherdomains associated with BPD (e.g., identity, behavioral, interper-sonal, and emotional). However, no study of BPD to date hasexamined acute and chronic pain in unison, leaving the exactnature of the paradox unclear.It is also worth noting that chronic pain is a significant publichealth problem, affecting over 100 million U.S. adults (Committeeon Advancing Pain Research, Care, and Education, Institute ofMedicin, 2011) and is associated with decreases in quality of lifeand functioning, as well as an increased risk of suicide (Breivik,Collettt, Ventafridda, Cohen, & Gallacher, 2006;Tang & Crane,2006). In addition, reduced experience of acute pain may beassociated with increased risk of future nonsuicidal self-injury(NSSI) and suicidal behavior (Joiner, 2009), both of which areprevalent in BPD. Thus, there is a need to increase our understand-ing of the relationship of pain and BPD features.NSSINSSI is defined as behavior that intentionally causes tissuedamage without suicidal intent (e.g., cutting). Most BPD individ-uals report at least one occasion of NSSI over their lifetime (e.g.,Dulit, Fyer, Leon, Brodsky, & Frances, 1994;Zanarini et al.,2008). Evidence from pain induction studies suggests that manyBPD individuals report little or no pain during NSSI (e.g.,Magerlet al., 2012;Russ et al., 1992;Russ et al., 1999). This correspondswith findings of reduced acute pain in BPD. That they correspond,however, is perhaps not surprising, as pain induction studies haveeither only included BPD participants with a history of NSSI(Bohus et al., 2000;Ludäscher et al., 2007;Ludäscher et al., 2009;Magerl et al., 2012;Niedtfeld et al., 2010;Russ et al., 1992;Russet al., 1999;Schmahl et al., 2004;Schmahl et al., 2006) or have notreported whether participants had a history of NSSI (McCown etal., 1993;Pavony & Lenzenweger, 2014;Schmahl et al., 2010).Thus, the association of acute pain tolerance and BPD features inthe absence of NSSI history is unknown. In the case of chronicpain, its association with NSSI has not been previously exploredand no study of BPD and chronic pain has assessed NSSI.It is therefore possible that reduced pain sensitivity to acutestimuli and, by extension, the pain paradox, may be restricted toBPD individuals with a history of NSSI. In support of this, evi-dence suggests that recency of NSSI is related to acute pain report,such that BPD patients who reported less recent NSSI reportedmore acute pain in the laboratory than BPD patients who reportedmore recent NSSI (Magerl et al., 2012;Ludäscher et al., 2009). Inaddition, BPD patients who reported pain during NSSI reportedmore acute pain in the laboratory than BPD patients, though stillless than controls (Russ et al., 1992;Russ et al., 1999). Thesestudies suggest that NSSI history may be relevant for acute painsensitivity.Complicating this picture, several studies have found reducedacute pain report among individuals with a history of NSSI, as wellas indirect self-harm (e.g., dysregulated eating, substance use),compared to those who had not (Franklin et al., 2012;Gratz et al.,2011;Hooley, Ho, Slater, & Lockshin, 2010;Hooley & St. Ger-main, 2014;St. Germain & Hooley, 2013). Although these studiesdid not assess for BPD, they raise the possibility that previousfindings of reduced acute pain experience in BPD individuals arenot a result of BPD, but of NSSI history. To examine the relativeeffects of NSSI and BPD on pain experience, there is a need tocompare individuals with and without NSSI history who also varyin terms of BPD features.The Current StudyDetermining the nature of the pain paradox in BPD is necessaryto better understand the relationship of physical pain and BPDfeatures. Although previous research has demonstrated a link be-tween BPD and acute and chronic pain, at least two gaps in theliterature remain to be investigated. First, it is unknown whetherthe same BPD individuals experience both a relative absence ofacute pain and a relative excess of chronic pain. Second, theexperience of pain in BPD individuals without a NSSI history, andhow it compares to the experience of BPD individuals with a NSSIhistory, has not been examined. Whether the pain paradox isspecific to BPD features, to NSSI history, or, perhaps, the combi-nation of both may shed light on how the experience of pain inBPD becomes dysregulated.The present study addressed these gaps in a large sample ofnonclinical young adults that were oversampled for significantBPD features. Specifically, we tested the association of BPDfeatures with indices of both acute and past-year pain (a proxymeasure for chronic pain) in individuals with and without a historyof NSSI. Among individuals with NSSI history, we examined theeffects of recency of NSSI and pain during NSSI. Although wedid not know whether past work with BPD patients wouldgeneralize to a sample of nonclinical individuals with BPDfeatures, we hypothesized that BPD features would be nega-tively associated with acute pain in the NSSI group. We had nohypothesis for the association of BPD features and acute pain inthe no-NSSI group, as there was no previous work upon whichto base a prediction for this group. For past-year pain, followingwork with chronic pain patients, we hypothesized that BPDfeatures would be positively associated with past-year painregardless of NSSI history.We additionally examined the relationship of BPD features andperceived pain tolerance, as it is unknown how BPD individualsperceive their ability to handle pain. It is possible that there existsa discrepancy between actual and perceived pain tolerance as BPDfeatures increase. This is potentially important, as perceived paintolerance could serve as an indicator of risk for NSSI and suicidalbehaviors (Joiner, 2009). If BPD features are related to a pain-paradox, then BPD features may be associated with either reducedor increased perceived pain tolerance.This document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.142CARPENTER AND TRULL
MethodParticipantsParticipants were 210 undergraduates recruited from introduc-tory psychology courses at a large Midwestern university. Oneparticipant was excluded because she was a non-native Englishspeaker and had difficulty understanding the experimenter. Asecond participant was excluded because she reported current useof prescription opioids, which have analgesic effects. Finally, twoparticipants were excluded because the temperature of the CPTwater exceeded 10.5°C. Thus, the number of participants includedin the final analyses was 206. The sample was primarily Caucasian(N169; 82%) and female (N164; 80%) and had a mean ageof 18.90 (SD1.27). Thirty-eight participants reported usingpsychoactive medications, usually selective reuptake inhibitors forserotonin or noradrenaline (N17).1MaterialsThe CPT has been established as a low-risk and ethical means ofproducing moderate amounts of pain. A modified version of theCPT protocol standardized byEfran et al. (1989)was used. Tocreate circulation, water moved by tubing from a 10-gallon storagetank into a testing container where participants placed their handduring the CPT. The water then flowed into a second storage tankbefore being pumped back into the main tank. Ice packs were keptin the storage tanks to maintain water temperature. Temperaturewas kept at 10°C (/.5) and was assessed by thermometer in themain storage tank and the testing container.The Personality Assessment Inventory—Borderline Featuresscale (PAI-Bor;Morey, 1991) consists of 24 items rated on a4-point scale ranging from 0 (false) to 3 (very true),.77, andis made up of four subscales representing core features of BPD:affective instability, self-harm/impulsivity, interpersonal relation-ships, and identity problems. Studies have demonstrated the va-lidity and reliability of the PAI-Bor for use in nonclinical samplesfor assessing BPD features (Trull, 1995;Trull, Useda, Conforti, &Doan, 1997). Thus, this scale served as an indicator of BPDfeatures. Participants were recruited with the goal of representingthe entire distribution of possible PAI-Bor scores, with an effort toensure adequate representation at the upper end of the scale. Thiswas accomplished by screening participants for BPD features andrecruiting them over the phone, with a relative emphasis on indi-viduals with scores above 37, which suggest the presence ofclinically significant BPD features (Morey, 1991;Trull, 1995;Trull et al., 1997). The final sample included 48 participants withPAI-Bor scores above 37 (23.30% of the sample). Note that thePAI-Bor was treated continuously in all analyses. Findings did notsignificantly differ when the cutoff for clinically significant fea-tures was used instead.The Pain Module of the Standard Evaluation Questionnaire(Pain-SEQ;Müller et al., 2008) was designed to measure muscu-loskeletal pain in the general population. For the current study,questions referred to the past year. Two sum score indicators,considered to serve as approximate measures of chronic pain, wereused. The first, body location-based past-year pain, consisted of 7items that asked about the intensity of pain in different areas of thebody (e.g., head, left shoulder, back). The second, activity-basedpast-year pain, consisted of 12 items that asked about pain duringdifferent activities (e.g., throwing an object, lying still). Twoparticipants did not answer two of the questions for this indicator.Mean imputation was used in these cases.2All 19 items were ratedfrom 1 (no pain) to 7 (intolerable pain).The Deliberate Self-Harm Inventory (Gratz, 2001) is a 17-itembehaviorally based questionnaire that assesses deliberate damageof body tissue without conscious suicidal intent (i.e., NSSI). Be-haviors included cutting, burning, skin carving, severe scratching,biting, rubbing sandpaper, dripping acid, sticking with sharp ob-jects, breaking bones, banging head, and preventing wounds fromhealing. If cutting was endorsed, follow-up questions were asked(e.g., “When was the last time you did this?”). Participants whoreported one or more NSSI incidents (of any method) were codedas 1 in a single, binary variable (NSSI history) whereas those thatdid not were coded as 0. Participants who reported NSSI wereasked to report the degree of pain they experienced during NSSIfrom 1 (very little or none at all) to 5 (extreme amount).The Pain Anxiety Symptoms Scale (PASS;McCracken, Zayfert,& Gross, 1992) measures fear and anxiety related to pain over foursubscales: cognitive anxiety, escape/avoidance behaviors, fearfulappraisal, and physiological anxiety. Twenty items were ratedfrom 0 (never) to 5 (always),.94. The Pain CatastrophizingScale (PCS;Sullivan, Bishop, & Pivik, 1995) was designed toidentify individuals with negative cognitions regarding pain. Thir-teen items were rated from 0 (not at all) to 4 (all the time),.94.Two participants failed to answer one question each on the PCSand mean imputation was used. Sum scores of the PCS and PASSwere considered indicators of participants’ perception of their paintolerance.The Positive and Negative Affect Schedule (PANAS;Watson,Clark, & Tellegen, 1988) includes two 10-item scales, one forpositive affect (PA;1.90;2.87) and one for negativeaffect (NA;1.75;2.86). Respondents were instructed toread each mood descriptor and then rate the extent to which theyfelt that way at that moment, ranging from 1 (very slightly or notat all) to 5 (extremely).ProcedureParticipants received an overview of the experiment and gaveinformed consent. They first placed their dominant hand into acontainer of lukewarm water and were instructed to sit quietly for5 min. This ensured that initial skin temperature was consistentacross the sample. Participants then filled out the first of twoPANAS forms, reporting their current affective state.Participants next completed the CPT. They were instructed toplace their hand up to the wrist in the insulated test container ofcold water with their hand flat against the bottom. The experi-menter asked participants to rate the intensity and unpleasantnessof their current pain on a scale from 0 (not present) to 9 (maximumpain) every 15 s. The task lasted for 4 min, although participants1Psychoactive medication use was initially included in analyses as acovariate. However, its presence in the models did not significantly alterresults and it was, therefore, dropped.2Dropping participants who did not complete all items of a question-naire from analyses, instead of using mean imputation, did not significantlyalter results.This document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.143PAIN PARADOX AND BPD FEATURES
could remove their hand at any time if they felt that the pain wastoo severe to continue. Following the task, participants filled outthe second PANAS form and placed their hand in the lukewarmbath for 5 min, to ameliorate residual pain. Lastly, participantscompleted the self-report questionnaire battery in a separate room.ResultsPreliminary AnalysesThe person-level means for intensity and unpleasantness re-sponses were highly correlated,r(204).918,p.001, and,therefore, these values were averaged for each trial to create asingle repeated-measures indicator, referred to asCPT pain. Thistrial-level indicator served as the primary dependent variable forCPT analyses. The person-level mean of CPT pain (person-levelCPT pain; the average of CPT pain across all trials within eachparticipant) was also computed to examine patterns of associationwith other person-level variables.Correlations for dependent variables are presented inTable 1.Table 2presents means and standard deviations for the continuouspredictor variables and dependent variables, both overall and bygender. Similar to previous work (for a review, seeGiles &Walker, 1999), men reported significantly lower person-level CPTpain than women, as well as lower PASS scores. There was nogender effect on the remaining measures. Given its relationshipwith several dependent variables, but no independent variables,gender was included as a covariate in analyses.3Fifty-eight individuals reported a positive NSSI history, whichdid not differ significantly by gender: 26% of females (N43)and 35% of males (N15) reported self-harm,2(1,N206)1.49,p.222. Individuals in the NSSI group had higher PAI-Borscores (M36.38,SD15.55) than individuals in the no-NSSIgroup (M22.17,SD11.20),t(81.22)6.34,p.001.However, NSSI was present across the distribution of PAI-Borscores, with 17 individuals below the PAI-Bor mean and 41 individ-uals above reporting past NSSI. The most common form of NSSIreported was cutting (N28) and of those reporting NSSI, mostparticipants reported engaging in more than one NSSI method(M2.83,SD2.15). Time since last reported NSSI rangedfrom 1 month or less to 8 years ago (M2.96 years,SD2.52years,N26).4Most participants reported at least some NSSIpain (M2.67,SD1.08,N52).5NSSI recency wasnegatively correlated with PAI-Bor score,r(24).480,p.013, indicating that scoring higher on the PAI-Bor was associatedwith a shorter period of time since last NSSI. There was nosignificant relationship between PAI-Bor score and NSSI pain(p.746).Cold Pressor Task/Acute PainTo test hypotheses regarding trial-level CPT pain, we usedmultilevel modeling (MLM). MLM was used in part because 15participants6withdrew their hand from the cold water before thefull 4 min had elapsed. MLM is able to accommodate the resultingdifferent number of observations for some participants. In ad-dition, the specified model included a random intercept and arandom slope for trial, which allowed us to adjust for individualdifferences in level of CPT pain and change in CPT pain overtime. To reduce collinearity between the random intercept andslope, trial was first centered at the midpoint of the CPT, takento be Trial 8.Trial and the quadratic effect of trial (centered) were enteredinto the model as Level 1 fixed effects. The quadratic effect of trialwas included as it fit the temporal process of CPT pain moreclosely than the linear effect. PAI-Bor, as a centered, continuouspredictor, NSSI history, the interaction of PAI-Bor and NSSIhistory, and gender were entered as level 2 fixed effects.7Resultsare reported inTable 3. There were significant effects for trial andthe quadratic effect of trial, indicating that pain initially increasedslightly and then decreased over time. There was a main effect forgender, such that females (M6.22,SE0.17) reported moreCPT pain than males (M4.86,SE0.27),b1.36,SE0.30,t(198)4.52,p.001, 95% confidence interval (CI): 0.77, 1.95.8The main effects of PAI-Bor and NSSI history were not signifi-cant, but there was a significant interaction of PAI-Bor and NSSIhistory on CPT pain (seeFigure 1),b0.04,SE0.02,t(201)2.23,p.027, 95% CI: 0.01, 0.08. To examine this interaction,analyses were conducted separately by NSSI history. As PAI-Borscores increased in the no-NSSI group, CPT pain also increased(b0.05,SE0.01,t(148)3.98,p.001, 95% CI: 0.02,0.07). However, in the NSSI group, although the slope was posi-tive, there was no significant change in pain report as PAI-Borscores increased (b0.01,SE0.02,t(54.10)0.47,p.643,95% CI:0.02, 0.04).3Magerl et al. (2012)found an interaction of gender and BPD on acutepain. However, we found no interaction of gender with BPD features orNSSI history on CPT pain in our sample and, thus, these interaction termswere not included in analyses.4NSSI recency was only assessed in a subset of participants whoreported cutting (n28). Two participants provided a response that wasinsufficiently clear to be counted (e.g., “several years ago”).5Six participants who reported past NSSI did not complete the NSSIpain questions.6Of the 15 participants, four reported past NSSI and their mean score onthe PAI-Bor was 28.33 (SD13.10). Excluding these individuals did notsignificantly alter the results.7The interaction effects of trial with PAI-Bor, NSSI history, and theirinteraction were also estimated but were not significant. Therefore, thesewere dropped from the final model.8Degrees of freedom for all effects related to CPT pain were calculatedusing the conservative Satterthwaite approximation.Table 1Correlations Between Acute and Past-Year PainDependent Variables123451. Person-level CPT pain—2. Body location-based PYP.103—3. Activity-based PYP.050.585—4. PCS.268.355.375—5. PASS.370.260.230.785—Note.Person-level CPT painthe average of cold pressor task painacross all trials within each participant; PYPpast-year pain (derivedfrom the Pain Module of the Standard Evaluation Questionnaire); PCSPain Catastrophizing Scale; PASSPain Anxiety Symptom Scale.p.01.p.001.This document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.144CARPENTER AND TRULL
In the NSSI group, NSSI recency and NSSI pain were uncor-related with person-level CPT pain (ps.687). PAI-Bor scoreswere positively correlated with NA both before,r(204).231,p.001, and after,r(204).298,p.001, the CPT, but not withPA scores (ps.565). To examine whether NA changed frombefore the CPT to after, a multivariate analysis of variance modelwith repeated measures was specified, with NA as the dependentvariable. Time was a within-subject factor (before the CPT, afterthe CPT), and PAI-Bor, NSSI history, and their interaction werebetween-subjects factors. There was an effect for PAI-Bor on NAover time,F(1, 201)5.10,p.025, such that higher PAI-Borscores were associated with a greater increase in NA from beforethe CPT to after. There was no effect for NSSI history or theinteraction of PAI-Bor and NSSI history (ps.335).Past-Year Pain ExperienceTo examine the relationship between past-year pain and BPDfeatures, two regressions were conducted, one with body location-based and one with activity-based indicators of past-year pain(derived from the Pain-SEQ) as the dependent variable. PAI-Bor,NSSI history, their interaction, and gender were predictor vari-ables. For body location-based past-year pain, there was a maineffect for PAI-Bor,b0.17,t(204)3.06,p.003, such thathigher PAI-Bor scores were related with greater report of location-based past-year pain. There was no significant effect of NSSIhistory, the interaction of NSSI history and the PAI-Bor, or gender(ps.267).Results were similar for activity-based past-year pain. HigherPAI-Bor scores predicted greater amounts of past-year pain (b0.41,t(204)5.00,p.001). There was no significant effect forNSSI history or gender (ps.489). However, the interaction ofPAI-BorandNSSIhistorywassignificant(b0.27,t(204)2.53,p.012; seeFigure 2). Conducting analysesseparately by NSSI history revealed that the effect of PAI-Bor wassignificant in the NSSI group,(b0.38,t(56)4.44,p.001,as well as the no-NSSI group,b0.14,t(146)1.99,p.048,but the effect was significantly greater in the NSSI group.In the NSSI group, NSSI pain did not correlate with past-yearpain (ps.432). NSSI recency correlated with body location-based,r(24).449,p.022,andactivity-basedpain,r(24).526,p.005). A shorter lag since last NSSI wasassociated with more past-year pain.Table 2Means and Standard Deviations for Predictor and Dependent Variables for the Overall Sample and by GenderTotalWomenMenMSDMSDMSDt(df204)PAI-Bor26.1714.0826.9614.3223.0712.811.60Person-level CPT pain5.901.876.211.754.701.864.92Body location-based PYP16.176.6616.336.6215.526.880.70Activity-based PYP22.3010.2322.7910.4220.409.351.35PCS11.8510.6912.4811.119.368.521.70PASS24.8318.3526.4619.0418.4813.792.55Note.Person-level CPT painthe average of cold pressor task pain across all trials within each participant; PYPpast-year pain (derived from thepain module of the standard evaluation questionnaire); PCSPain Catastrophizing Scale; PASSPain Anxiety Symptom Scale.p.05.p.001.Table 3Parameter Estimates for Multilevel Model of Cold Pressor Task Pain95% CIEstimateSEdfaztpLowerUpperFixed effectsIntercept4.870.3320114.55.0014.215.53Trial0.170.0119514.10.0010.190.15Trial20.010.0012,73120.82.0010.010.01PAI-Bor0.010.011990.42.6730.020.04NSSI history0.300.312010.96.3390.320.91PAI-BorNSSI history0.040.022012.23.0270.010.08Gender1.360.301984.52.0010.771.95Random effects ([co-]variances)Level 1 (within-person)Var(Intercept)2.900.309.77.0012.403.59Var(Trial)0.030.0039.37.0010.020.03Cov(Intercept, Trial)0.040.021.78.0760.0040.08Var(Residual)0.350.0136.84.0010.330.36Note.N206 individuals, 16 trials, 3,135 observations used. CIconfidence interval; Trial2quadratic effect of trial; PAI-BorPersonalityAssessment Inventory—Borderline Features; NSSInonsuicidal self-injury.aDegrees of freedom were calculated using the conservative Satterthwaite approximation.This document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.145PAIN PARADOX AND BPD FEATURES
Perceived Pain-ToleranceFinally, we examined the relationship between pain-related cog-nitions and BPD features. Two regressions were conducted, onewith the PCS and the other with the PASS as the dependentvariable. On the PCS, effects for PAI-Bor, NSSI history, andgender were not significant, but the interaction of PAI-Bor andNSSI history was,b0.27,t(204)2.42,p.017 (seeFigure3A). Conducting the analysis separately by NSSI history revealeda significant positive effect for PAI-Bor in the no-NSSI group,b0.43,t(146)6.02,p.001, but, although still positive, theeffect was not significant in the NSSI group,b0.27,t(56)1.70,p.094.On the PASS, PAI-Bor, and NSSI history were not significant,but their interaction was (b0.70,t(204)3.87,p.001).Analyses revealed a significant positive effect for PAI-Bor in theno-NSSI group (b0.89,t(146)7.56,p.001), but, althoughstill positive, this effect was not significant in the NSSI group,b0.20,t(56)1.41,p.164. There was also a significant gendereffect on the PASS,b5.98,t(204)2.12,p.035, such thatfemales (M27.69,SE1.61) scored higher than males (M21.72,SE2.55).DiscussionThe goal of the present study was to explore the nature of thepain paradox in individuals with BPD features, understood as arelative absence of acute pain and a relative excess of past-yearpain. To accomplish this, acute and past-year pain were as-sessed in a nonclinical sample of individuals with and withoutNSSI history who varied in terms of BPD features. CPT painincreased as BPD features increased in the no-NSSI group, butthere was no association in the NSSI group. In contrast, past-year pain increased as BPD features increased regardless ofNSSI history. Lastly, BPD features were associated with lowerperceived pain tolerance in the no-NSSI group, but not the NSSIgroup.Acute Pain PerceptionIn the no-NSSI group, scoring higher on the PAI-Bor wasassociated with increased pain report. Previous research onBPD and pain induction has consistently found a reduction ofacute pain report among BPD patients. However, as notedearlier, there are differences between the current study andprevious work. First, the current work used a nonclinical sampleand assessed for BPD features and, second, previous workeither did not include participants without NSSI history or didnot report on NSSI history. To our knowledge, this was the firststudy to examine acute pain report in individuals with BPDfeatures and no NSSI history. Thus, it provides novel informa-tion about the experience of acute pain in these individuals.Among individuals in the NSSI group, the association be-tween the PAI-Bor and CPT pain report was not significant. Wealso found no significant relationship between NSSI recency orNSSI pain and CPT pain. Thus, contrary to prediction, we didFigure 1.Effect of borderline personality disorder (BPD) features, as measured by the Personality AssessmentInventory–Borderline Features (PAI-Bor), on overall cold pressor task (CPT) pain by presence versus absenceof nonsuicidal self-injury (NSSI) history across all observations. Because of the large number of observations(N3,135), there was considerable overlap and, therefore, a “jitter” function was used to increase the visibilityof the observations. In addition, the figure did not take into account nesting of observations within individuals,which affects standard errors. For this reason, confidence intervals were not estimated.This document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.146CARPENTER AND TRULL
not find that acute pain decreased as BPD features increased inthe NSSI group. This finding differs from the existing literature(Bohus et al., 2000;Ludäscher et al., 2007;Ludäscher et al.,2009;Magerl et al., 2012;McCown et al., 1993;Niedtfeld et al.,2010;Pavony & Lenzenweger, 2014;Russ et al., 1992;Russ etal., 1999;Schmahl et al., 2004;Schmahl et al., 2006;Schmahlet al., 2010). However, the current work also differed fromthese studies in that we recruited a nonclinical sample andassessed for BPD features. Therefore, the findings do not con-tradict previous work, but instead add to it, providing data onthe acute pain experience of individuals with BPD features.In fact, the present findings may converge with previouswork on the effects of NSSI recency and NSSI pain on acutepain.Magerl et al., (2012)found that, although BPD patientswith recent NSSI reported less acute pain than controls, BPDpatients who had not self-harmed in the last year did not.9Ludäscher et al. (2009)also found an effect for NSSI recency:BPD individuals who had not self-harmed in the last 6 monthsreported more acute pain than BPD individuals with currentNSSI, although still less than controls. Others have found thatBPD patients who report pain during NSSI report more pain inthe laboratory than those who do not (Russ et al., 1992;Russ etal., 1999). In contrast, many of the studies that found BPDpatients reported reduced acute pain used samples where all ornearly all participants reported current NSSI and/or only in-cluded participants who reported little or no NSSI pain (e.g.,Bohus et al., 2000;Ludäscher et al., 2007;Niedtfeld et al.,2010;Schmahl et al., 2004;Schmahl et al., 2006).Although the NSSI reported in the present study was clinicallysignificant (e.g., cutting, burning), many NSSI group participantshad not self-harmed within the past year. Many also reported painduring past NSSI. This may be a consequence of recruiting anonclinical sample. As a result, our NSSI group differed fromsamples in previous work that reported a relative absence of acutepain. Instead, in terms of NSSI, the group more closely matchedgroups of BPD patients that showed smaller or no differences fromcontrols (Ludäscher et al., 2009;Magerl et al., 2012;Russ et al.,1992;Russ et al., 1999). This, at least in part, may explain theabsence of a negative association in the current study betweenacute pain and BPD features in the NSSI group. However, if NSSIrecency and NSSI pain have such a significant effect on acute pain,it is unclear why there was no correlation between CPT pain andeither NSSI recency or NSSI pain.Nevertheless, the fact that CPT pain increased with PAI-Borscores in the no-NSSI group, but not the NSSI group, suggests thatNSSI history was associated with reduced acute pain among indi-viduals higher in BPD features. Notably, there was no evidence fora similar effect among individuals lower in BPD features (i.e., nosignificant main effect for NSSI history on CPT pain). Thus,results did not suggest that NSSI history is negatively associatedwith acute pain regardless of BPD features. Previous work hasfound an association of NSSI history and reduced acute pain(Franklin et al., 2012 Gratz et al., 2011;Hooley et al., 2010;Hooley & St. Germain, 2014), but these studies did not assess forBPD features. Again, given that NSSI was less recent in thecurrent sample than in previous work, it may be that a main effect9Magerl et al. (2012)only had five such BPD individuals, meaning thatthis null finding should be interpreted with caution.Figure 2.Effect of borderline personality disorder (BPD) features, as measured by the Personality AssessmentInventory–Borderline Features (PAI-Bor), by presence versus absence of nonsuicidal self-injury (NSSI) historyon activity-based past-year pain (Pain Module of the Standard Evaluation Questionnaire). 95% confidenceintervals were used.This document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.147PAIN PARADOX AND BPD FEATURES
for NSSI history would have emerged in a sample with morerecent NSSI. However, in the current study, only the interaction ofNSSI history and BPD features revealed an effect for acute pain.Past-Year Pain ExperienceBPD features were positively associated with past-year pain, inline with findings that BPD is associated with chronic pain (Fran-kenburg & Zanarini, 2004;Sansone et al., 2010;Sansone &Sansone, 2012;Saper & Lake, 2002;Tragesser et al., 2010). It alsoextends previous work, demonstrating that BPD features are notonlyprevalentamongindividualswithmedicallysignificantchronic pain, but also associated with pain experience that may ormay not be medically significant. In addition, previous work haslargely been restricted to examining rates of BPD and BPD fea-tures in chronic pain samples. The present study shows that past-year pain is associated with BPD features in a nonclinical sampleoversampled for BPD features.The positive association of BPD features and past-year pain waspresent regardless of NSSI history. Interestingly, some of thefindings suggested that past-year pain was more strongly associ-ated with BPD features in the NSSI group than the no-NSSI group.Past-year pain was also negatively associated with NSSI recency.However, this interaction was not found for body location-basedpast-year pain, making it difficult to interpret. In sum, the evidencesupported the idea that BPD features are associated with a relativeexcess of past-year pain.Perceived Pain ToleranceThere was also a significant interaction effect of BPD featuresand NSSI history on the PCS and on the PASS. While PAI-Borscores were associated with lower ratings of perceived pain toler-ance in the no-NSSI group, there was no effect of PAI-Bor scoresin the NSSI group. That BPD features were, in the no-NSSI group,associated with reporting more catastrophizing of and greateranxiety toward pain corresponds with the findings that BPD fea-tures were also positively associated with acute and past-year painin this group. Thus, there was no discrepancy in the no-NSSI groupin terms of perceived and actual pain tolerance. In contrast, therewas a discrepancy in the NSSI group: BPD features, despite beingassociated with greater past-year pain, were not associated withperceived pain tolerance. This finding that the combination of BPDfeatures and NSSI history is linked to both higher past-year painand normal perceived pain tolerance may indicate that these indi-viduals are at risk for future self-destructive acts, as the aversivequality of pain may normally serve as a deterrent for such acts(Joiner, 2009).The State of the ParadoxThe pain paradox described bySansone and Sansone (2007)may be more nuanced than it would appear at first glance. Elevatedpast-year pain appears to be generally associated with BPD fea-tures, regardless of NSSI history. However, in the no-NSSI group,BPD features predicted greater CPT pain, while there was noassociation in the NSSI group. Though not as expected, comparingthe former positive association to the latter null effect providestentative evidence that NSSI history was associated with reducedacute pain report in individuals higher in BPD features. Thefindings also add to the evidence that effects for acute pain aresmaller among individuals with BPD features who have not re-cently self-harmed. In sum, the current study provides some evi-dence that BPD features are linked to a pain paradox, but thecombination of a relative absence of acute pain and a relativeexcess of past-year pain may be specific to individuals with bothBPD features and a history of NSSI. It is unknown, however,whether these findings would extend to individuals with clinicallydiagnosed BPD. Thus, they suggest a need for future work thatexamines the effect of NSSI and BPD in a clinical sample.Figure 3.Effect of borderline personality disorder (BPD) features, as measured by the Personality AssessmentInventory–Borderline Features (PAI-Bor), by presence versus absence of nonsuicidal self-injury (NSSI) historyon perceived pain tolerance, as measured by (A) the Pain Catastrophizing Scale (PCS) and (B) the Pain AnxietySymptoms Scale (PASS). 95% confidence intervals were used.This document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.148CARPENTER AND TRULL
What might be the cause of the pain paradox? Previous work hassuggested that decreased acute pain sensitivity is a result of stress-induced analgesia or dissociation (Bohus et al., 2000;McCown etal., 1993), habituation to physical pain via repeated NSSI (Nock,2010), or alterations in the brain’s regulation of the affective-motivational component of pain (Niedtfeld & Schmahl, 2009;Niedtfeld et al., 2010). Increased chronic pain sensitivity has beenconceptualized as a failure to self-regulate (Sansone & Sansone,2007) or has been linked to depression (Tragesser et al., 2010).Another possible factor involved in the pain paradox is dysregu-lation of the endogenous opioid system, which plays an importantrole in the body’s analgesic response to pain (Akil et al., 1984) andhas been previously implicated in BPD (Bandelow, Schmahl,Falkai, & Wedekind, 2010;Prossin, Love, Koeppe, Zubieta, &Silk, 2010;Stanley & Siever, 2010). Ultimately, however, themechanisms behind the pain paradox remain largely unknown.Limitations, Future Directions, and ConclusionThe current study had notable strengths, most importantly that itwas the first investigation of pain experience in those with BPDfeatures to examine past-year and acute pain and the first toinclude individuals with and without NSSI history. In addition, thesample included both males and females, whereas most previousresearch on acute pain in BPD has only included females (excep-tions areMagerl et al., 2012, andMcCown et al., 1993). Thecurrent findings suggest that there are no significant differences inthe experience of pain between males and females with BPDfeatures. Finally, the large sample size and the use of MLM toanalyze the CPT data both increase confidence in the findings.There were, however, several limitations. First, the current studywas limited in the depth of information obtained about NSSIbehavior. Related to this, relatively few participants had engagedin NSSI in the last year. Second, past-year pain and chronic painare different, though related, constructs, with past-year pain beingbroader and likely less closely tied to impairment. Past-year painwas assessed because chronic pain, strictly defined, was ex-pected to have a low base rate in this population. The fact thatwe found that BPD features were associated with greater past-year pain despite this probable low prevalence of chronic painis noteworthy.Third, participants were not assessed for a BPD diagnosis.Although using a nonclinical sample broadens our knowledgeabout the association of BPD and pain experience and made itpossible to recruit a larger sample than is typical in pain inductionstudies of BPD, there may be differences in pain experiencebetween individuals high in BPD features and individuals with aBPD diagnosis. It cannot be ruled out that the use of a nonclinicalsample was a reason that we did not find a negative associationbetween BPD features and acute pain. Indeed, we suggest that thisis the case, in that recruiting a nonclinical sample is a possiblefactor behind why many individuals in the NSSI group had notrecently self-harmed. Relatedly, self-harm was positively associ-ated with PAI-Bor scores and, as can be seen inFigure 1, most ofthe individuals scoring highest on the PAI-Bor reported past NSSI.Thus, among individuals with the most BPD features, the relativeeffects of PAI-Bor and NSSI history could not be fully disentan-gled in the present study. The current work then, in conjunctionwith previous studies using BPD patients, provides a fullerpicture of pain experience across the spectrum of BPD severity.Ultimately, however, to clarify the pain paradox in BPD, futureresearch is required that examines chronic and acute pain in aclinical sample of BPD individuals with and without past NSSI.The current study examined the relationship of acute and past-year pain with BPD features and NSSI history. Results did notsupport hypotheses regarding acute pain. However, the interactionof BPD features and NSSI history on acute pain suggests that thepain paradox may involve the combination of BPD features andNSSI history. Although the present study cannot evaluate whetherNSSI causes reduced acute pain, it suggests that NSSI history,along with recency of NSSI and pain during NSSI, are importantfactors to take into consideration. In addition, the present studyadds to the existing evidence that BPD features are associated withchronic pain. Besides the negative impacts on quality of life andfunctioning, chronic pain is a risk factor of suicidal behavior(Breivik et al., 2006;Tang & Crane, 2006), as is reduced acutepain experience (Joiner, 2009), and NSSI history (Cooper et al.,2005;Zahl & Hawton, 2004). Increasing our understanding of theinterplay of acute and chronic pain and NSSI history in BPD may,thus,haveimportantimplicationsforunderstandingself-destructive behavior in this disorder. In addition, it may help us tounderstand other domains in which BPD individuals experiencewidely diverging extremes. Future work should examine the rela-tionship of pain with other forms of dysregulation (e.g., identity,behavioral, interpersonal, emotional) in BPD.ReferencesAkil, H., Watson, S. J., Young, E., Lewis, M. E., Khachaturian, H., &Walker, J. M. (1984). Endogenous opioids: Biology and function.An-nual Review of Neuroscience, 7,223–255.http://dx.doi.org/10.1146/annurev.ne.07.030184.001255Bandelow, B., Schmahl, C., Falkai, P., & Wedekind, D. (2010). Borderlinepersonality disorder: A dysregulation of the endogenous opioid system?Psychological Review, 117,623– 636.http://dx.doi.org/10.1037/a0018095Bohus, M., Limberger, M., Ebner, U., Glocker, F. X., Schwarz, B., Wernz,M., & Lieb, K. (2000). Pain perception during self-reported distress andcalmness in patients with borderline personality disorder and self-mutilating behavior.Psychiatry Research, 95,251–260.http://dx.doi.org/10.1016/S0165-1781(00)00179-7Breivik, H., Collett, B., Ventafridda, V., Cohen, R., & Gallacher, D.(2006). Survey of chronic pain in Europe: Prevalence, impact on dailylife, and treatment.European Journal of Pain, 10,287–333.http://dx.doi.org/10.1016/j.ejpain.2005.06.009Committee on Advancing Pain Research, Care, and Education, Institute ofMedicine. (2011).Relieving pain in America: A blueprint for transform-ing prevention, care, education, and research. Washington, DC: TheNational Academies Press.Cooper, J., Kapur, N., Webb, R., Lawlor, M., Guthrie, E., Mackway-Jones,K., & Appleby, L. (2005). Suicide after deliberate self-harm: A 4-yearcohort study.The American Journal of Psychiatry, 162,297–303.http://dx.doi.org/10.1176/appi.ajp.162.2.297Dulit, R. A., Fyer, M. R., Leon, A. C., Brodsky, B. S., & Frances, A. J.(1994). Clinical correlates of self-mutilation in borderline personalitydisorder.The American Journal of Psychiatry, 151,1305–1311.http://dx.doi.org/10.1176/ajp.151.9.1305Efran, J. S., Chorney, R. L., Ascher, L. M., & Lukens, M. D. (1989).Coping styles, paradox, and the cold pressor task.Journal of BehavioralMedicine, 12,91–103.http://dx.doi.org/10.1007/BF00844751This document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.149PAIN PARADOX AND BPD FEATURES
Frankenburg, F. R., & Zanarini, M. C. (2004). The association betweenborderline personality disorder and chronic medical illnesses, poorhealth-related lifestyle choices, and costly forms of health care utiliza-tion.The Journal of Clinical Psychiatry, 65,1660–1665.http://dx.doi.org/10.4088/JCP.v65n1211Franklin, J. C., Aaron, R. V., Arthur, M. S., Shorkey, S. P., & Prinstein,M. J. (2012). Nonsuicidal self-injury and diminished pain perception:The role of emotion dysregulation.Comprehensive Psychiatry, 53,691–700.http://dx.doi.org/10.1016/j.comppsych.2011.11.008Giles, B. E., & Walker, J. S. (1999). Gender differences in pain.CurrentOpinion in Anaesthesiology, 12,591–595.http://dx.doi.org/10.1097/00001503-199910000-00017Gratz, K. (2001). Measurement of deliberate self-harm: Preliminary dataon the Deliberate Self-Harm Inventory.Journal of Psychopathology andBehavioral Assessment, 23,253–263.http://dx.doi.org/10.1023/A:1012779403943Gratz, K. L., Hepworth, C., Tull, M. T., Paulson, A., Clarke, S., Remington,B., & Lejuez, C. W. (2011). An experimental investigation of emotionalwillingness and physical pain tolerance in deliberate self-harm: The mod-erating role of interpersonal distress.Comprehensive Psychiatry, 52,63–74.http://dx.doi.org/10.1016/j.comppsych.2010.04.009Hooley, J. M., & St. Germain, S. A. (2014). Nonsuicidal self-injury, pain,and self-criticism does changing self-worth change pain endurance inpeople who engage in self-injury?Clinical Psychological Science, 2,297–305.http://dx.doi.org/10.1177/2167702613509372Hooley, J. M., Ho, D. T., Slater, J., & Lockshin, A. (2010). Pain perceptionand nonsuicidal self-injury: A laboratory investigation.Personality Dis-orders: Theory, Research, and Treatment, 1,170–179.http://dx.doi.org/10.1037/a0020106Joiner, T. (2009).Why people die by suicide. Cambridge, MA: HarvardUniversity Press.Kalira, V., Treisman, G. J., & Clark, M. R. (2013). Borderline personalitydisorder and chronic pain: A practical approach to evaluation andtreatment.Current Pain and Headache Reports, 17,350.http://dx.doi.org/10.1007/s11916-013-0350-yLenzenweger, M. F., Lane, M. C., Loranger, A. W., & Kessler, R. C.(2007).DSM–IVpersonality disorders in the national comorbidity sur-vey replication.Biological Psychiatry, 62,553–564.http://dx.doi.org/10.1016/j.biopsych.2006.09.019Ludäscher, P., Bohus, M., Lieb, K., Philipsen, A., Jochims, A., & Schmahl,C. (2007). Elevated pain thresholds correlate with dissociation andaversive arousal in patients with borderline personality disorder.Psy-chiatry Research, 149,291–296.http://dx.doi.org/10.1016/j.psychres.2005.04.009Ludäscher, P., Greffrath, W., Schmahl, C., Kleindienst, N., Kraus, A.,Baumgärtner, U., . . . Bohus, M. (2009). A cross-sectional investigationof discontinuation of self-injury and normalizing pain perception inpatients with borderline personality disorder.Acta Psychiatrica Scandi-navica, 120,62–70.http://dx.doi.org/10.1111/j.1600-0447.2008.01335.xMagerl, W., Burkart, D., Fernandez, A., Schmidt, L. G., & Treede, R.-D.(2012). Persistent antinociception through repeated self-injury in pa-tients with borderline personality disorder.Pain, 153,575–584.http://dx.doi.org/10.1016/j.pain.2011.11.021McCown, W., Galina, H., Johnson, J., DeSimone, P. A., & Posa, J. (1993).Borderline personality disorder and laboratory-induced cold pressorpain: Evidence of stress-induced analgesia.Journal of Psychopathologyand Behavioral Assessment, 15,87–95.http://dx.doi.org/10.1007/BF00960610McCracken, L. M., Zayfert, C., & Gross, R. T. (1992). The Pain AnxietySymptoms Scale: Development and validation of a scale to measure fearofpain.Pain,50,67–73.http://dx.doi.org/10.1016/0304-3959(92)90113-PMorey, L. C. (1991).Personality Assessment Inventory: Professional man-ual. Odessa, FL: Psychological Assessment Resources.Müller, U., Tänzler, K., Bürger, A., Staub, L., Tamcan, O., Roeder, C., . . .Trelle, S. (2008). A pain assessment scale for population-based studies:Development and validation of the pain module of the Standard Evalu-ation Questionnaire.Pain, 136,62–74.http://dx.doi.org/10.1016/j.pain.2007.06.014Niedtfeld, I., & Schmahl, C. (2009). Emotion regulation and pain inborderline personality disorder.Current Psychiatry Reviews, 5,48–54.http://dx.doi.org/10.2174/157340009787315262Niedtfeld, I., Schulze, L., Kirsch, P., Herpertz, S. C., Bohus, M., & Schmahl,C. (2010). Affect regulation and pain in borderline personality disorder: Apossible link to the understanding of self-injury.Biological Psychiatry, 68,383–391.http://dx.doi.org/10.1016/j.biopsych.2010.04.015Nock, M. K. (2010). Self-injury.Annual Review of Clinical Psychology, 6,339–363.http://dx.doi.org/10.1146/annurev.clinpsy.121208.131258Pavony, M. T., & Lenzenweger, M. F. (2014). Somatosensory processingand borderline personality disorder: Pain perception and a signal detec-tion analysis of proprioception and exteroceptive sensitivity.PersonalityDisorders: Theory, Research, and Treatment, 5,164–171.http://dx.doi.org/10.1037/per0000017Prossin, A. R., Love, T. M., Koeppe, R. A., Zubieta, J. K., & Silk, K. R.(2010). Dysregulation of regional endogenous opioid function in bor-derline personality disorder.The American Journal of Psychiatry, 167,925–933.http://dx.doi.org/10.1176/appi.ajp.2010.09091348Russ, M. J., Campbell, S. S., Kakuma, T., Harrison, K., & Zanine, E.(1999). EEG theta activity and pain insensitivity in self-injurious bor-derline patients.Psychiatry Research, 89,201–214.http://dx.doi.org/10.1016/S0165-1781(99)00113-4Russ, M. J., Roth, S. D., Lerman, A., Kakuma, T., Harrison, K., Shindle-decker, R. D., . . . Mattis, S. (1992). Pain perception in self-injuriouspatients with borderline personality disorder.Biological Psychiatry, 32,501–511.http://dx.doi.org/10.1016/0006-3223(92)90218-OSansone, R. A., Mueller, M., Mercer, A., & Wiederman, M. W. (2010).The relationship between self-reported pain and borderline personalitysymptomatology among internal medicine outpatients.Primary CareCompanion to the Journal of Clinical Psychiatry, 12,PCC.09100933.http://dx.doi.org/10.4088/PCC.09l00933whiSansone, R. A., & Sansone, L. A. (2007). Borderline personality and thepain paradox.Psychiatry, 4,40–46.Sansone, R. A., & Sansone, L. A. (2012). Chronic pain syndromes andborderline personality.Innovations in Clinical Neuroscience, 9,10–14.Saper, J. R., & Lake, A. E., III. (2002). Borderline personality disorder andthe chronic headache patient: Review and management recommenda-tions.Headache, 42,663–674.http://dx.doi.org/10.1046/j.1526-4610.2002.02156.xSchmahl, C., Bohus, M., Esposito, F., Treede, R. D., Di Salle, F., Greffrath,W., . . . Seifritz, E. (2006). Neural correlates of antinociception inborderline personality disorder.Archives of General Psychiatry, 63,659–667.http://dx.doi.org/10.1001/archpsyc.63.6.659Schmahl, C., Greffrath, W., Baumgärtner, U., Schlereth, T., Magerl, W.,Philipsen, A., . . . Treede, R.-D. (2004). Differential nociceptive deficitsin patients with borderline personality disorder and self-injurious behav-ior: Laser-evoked potentials, spatial discrimination of noxious stimuli,and pain ratings.Pain, 110,470–479.http://dx.doi.org/10.1016/j.pain.2004.04.035Schmahl, C., Meinzer, M., Zeuch, A., Fichter, M., Cebulla, M., Kleindi-enst, N., . . . Bohus, M. (2010). Pain sensitivity is reduced in borderlinepersonality disorder, but not in posttraumatic stress disorder and bulimianervosa.The World Journal of Biological Psychiatry: The OfficialJournal of the World Federation of Societies of Biological Psychiatry,11,364–371.http://dx.doi.org/10.3109/15622970701849952Stanley, B., & Siever, L. J. (2010). The interpersonal dimension of bor-derline personality disorder: Toward a neuropeptide model.The Amer-ican Journal of Psychiatry, 167,24–39.http://dx.doi.org/10.1176/appi.ajp.2009.09050744This document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.150CARPENTER AND TRULL
St. Germain, S. A., & Hooley, J. M. (2013). Aberrant pain perception indirect and indirect non-suicidal self-injury: An empirical test of Joiner’sinterpersonal theory.Comprehensive Psychiatry, 54,694–701.http://dx.doi.org/10.1016/j.comppsych.2012.12.029Sullivan, M. J. L., Bishop, S. R., & Pivik, J. (1995). The Pain Catastro-phizing Scale: Development and validation.Psychological Assessment,7,524–532.http://dx.doi.org/10.1037/1040-3590.7.4.524Tang, N. K. Y., & Crane, C. (2006). Suicidality in chronic pain: A reviewof the prevalence, risk factors and psychological links.PsychologicalMedicine, 36,575–586.http://dx.doi.org/10.1017/S0033291705006859Tragesser, S. L., Bruns, D., & Disorbio, J. M. (2010). Borderline person-ality disorder features and pain: The mediating role of negative affect ina pain patient sample.The Clinical Journal of Pain, 26,348–353.http://dx.doi.org/10.1097/AJP.0b013e3181cd1710Trull, T. J. (1995). Borderline personality disorder features in nonclinicalyoung adults 1. Identification and validation.Psychological Assessment,7,33–41.http://dx.doi.org/10.1037/1040-3590.7.1.33Trull, T. J., Jahng, S., Tomko, R. L., Wood, P. K., & Sher, K. J. (2010).Revised NESARC personality disorder diagnoses: Gender, prevalence, andcomorbidity with substance dependence disorders.Journal of PersonalityDisorders, 24,412–426.http://dx.doi.org/10.1521/pedi.2010.24.4.412Trull, T. J., Useda, J. D., Conforti, K., & Doan, B. T. (1997). Borderlinepersonality disorder features in nonclinical young adults: 2. Two-yearoutcome.Journal of Abnormal Psychology, 106,307–314.http://dx.doi.org/10.1037/0021-843X.106.2.307Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and vali-dation of brief measures of positive and negative affect: The PANASscales.Journal of Personality and Social Psychology, 54,1063–1070.http://dx.doi.org/10.1037/0022-3514.54.6.1063Zahl, D. L., & Hawton, K. (2004). Repetition of deliberate self-harm andsubsequent suicide risk: Long-term follow-up study of 11,583 patients.The British Journal of Psychiatry, 185,70–75.http://dx.doi.org/10.1192/bjp.185.1.70Zanarini, M. C., Frankenburg, F. R., Reich, D. B., Fitzmaurice, G., Wein-berg, I., & Gunderson, J. G. (2008). The 10-year course of physicallyself-destructive acts reported by borderline patients and axis II compar-ison subjects.Acta Psychiatrica Scandinavica, 117,177–184.http://dx.doi.org/10.1111/j.1600-0447.2008.01155.xThis document is copyrighted by the American Psychological Association or one of its allied publishers.This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.151PAIN PARADOX AND BPD FEATURES