Pain Medicine 2016; 17: 25–32 doi: 10.1111/pme.12860
Allopregnanolone Levels Are Inversely Associated with Self-Reported Pain Symptoms in U.S. Iraq and Afghanistan-Era Veterans: Implications for Biomarkers and Therapeutics
*Research and Development/Mental Health Services, Durham Veterans Affairs Medical Center, Durham, North Carolina; †Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, Durham, North Carolina; ‡Veterans Integrated Service Networks (VISN)6 Mid-Atlantic Mental Illness, Research Education and Clinical Center (MIRECC), Durham, North Carolina; §Department of Psychiatry, University of North Carolina, Chapel Hill, North Carolina; ¶Department of Biostatistics, University of North Carolina, Chapel Hill, North Carolina, USA Correspondence to: Christine E. Marx, MD, MA, Associate Professor, Durham VA Medical Center, 508 Fulton St., Durham, NC 27705, USA. Tel: 919-2860411-3626; Fax: 919-416-5912; E-mail: [email protected]; [email protected]. The VA Mid-Atlantic MIRECC workgroup includes: Jean C. Beckham, PhD; Mira Brancu, PhD.; Patrick S. Calhoun, PhD; John A. Fairbank, PhD; Harold S. Kudler, MD; Scott D. Moore, MD, PhD; Kristy A. Straits-Troster, PhD; Richard D. Weiner, MD, PhD Funding sources: This work was supported by the VA Mid-Atlantic Mental Illness Research Education and Clinical Center (MIRECC), PI John Fairbank, which is supported by the Office of Mental Health Services, Department of Veterans Affairs. Additional support is from VA Career Development Transition Award (Marx) and VA Merit Review Awards (Marx). Dr. Naylor is supported by a Department of Veterans Affairs Rehabilitation
Research and Development Career Development Award (1lK2RX000908). Disclosure and conflicts of interest: Dr. Marx is an applicant or co-applicant on pending patent applications focusing on neurosteroids and derivatives in CNS disorders (co-owned by VA and Duke); no patents issued; no licensing in place; VA 208 waiver in place. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.
Abstract Background and Objectives. Pain symptoms are common among Iraq/Afghanistan-era veterans, many of whom continue to experience persistent pain symptoms despite multiple pharmacological interventions. Preclinical data suggest that neurosteroids such as allopregnanolone demonstrate pronounced analgesic properties, and thus represent logical biomarker candidates and therapeutic targets for pain. Allopregnanolone is also a positive GABAA receptor modulator with anxiolytic, anticonvulsant, and neuroprotective actions in rodent models. We previously reported inverse associations between serum allopregnanolone levels and self-reported pain symptom severity in a pilot study of 82 male veterans.
Methods. The current study investigates allopregnanolone levels in a larger cohort of 485 male Iraq/ Afghanistan-era veterans to attempt to replicate these initial findings. Pain symptoms were assessed by items from the Symptom Checklist-90-R (SCL-90-R) querying headache, chest pain, muscle soreness, and low back pain over the past 7 days. Allopregnanolone levels were quantified by gas chromatography/mass spectrometry.
C 2015 American Academy of Pain Medicine. All rights reserved. For permissions, please e-mail: [email protected] V
25
Downloaded from http://painmedicine.oxfordjournals.org/ by guest on March 25, 2016
Jennifer C. Naylor, PhD,*,†,‡ Jason D. Kilts, PhD,*†,‡ Steven T. Szabo, MD, PhD,*,†,‡ Charlotte E. Dunn, BS,*,‡ Francis J. Keefe, PhD,† Larry A. Tupler, PhD,*,†,‡ Lawrence J. Shampine, BS,*,†,‡ Rajendra A. Morey, MD,*,†,‡ Jennifer L. Strauss, PhD,*,†,‡ Robert M. Hamer, PhD,§,¶ H. Ryan Wagner, PhD,*,†,‡ MIRECC Workgroup,‡ and Christine E. Marx, MD*,†,‡
Naylor et al. Results. Associations between pain ratings and allopregnanolone levels were examined with Poisson regression analyses, controlling for age and smoking. Bivariate nonparametric Mann– Whitney analyses examining allopregnanolone levels across high and low levels of pain were also conducted. Allopregnanolone levels were inversely associated with muscle soreness [P 5 0.0028], chest pain [P 5 0.032], and aggregate total pain (sum of all four pain items) [P 5 0.0001]. In the bivariate analyses, allopregnanolone levels were lower in the group reporting high levels of muscle soreness [P 5 0.001].
Key Words. Allopregnanolone; Veteran; Neurosteroid
Pain;
GABA;
Introduction Nearly 1.9 million U.S. troops have been deployed to Iraq and Afghanistan since September 11, 2001 [1]. Of these returning veterans, approximately 40–50% may experience chronic pain symptoms [2,3]. Chronic pain adversely affects an individual’s emotional and physical well-being, cognition, quality of life, and functional abilities, and also increases suicide risk [4]. Pain disorders can be particularly challenging to treat in Iraq/ Afghanistan-era veterans, as this cohort often reports complex co-occurring physical and psychological symptoms [5–9]. Treating pain represents a multibillion dollar annual expense for the Veterans Healthcare System [10]. Pharmacological interventions for pain are frequently suboptimal, however, and many veterans experience persistent and unalleviated pain symptoms. Medications for pain symptoms are limited by unfavorable side effect profiles and/or lack of efficacy. Opioid analgesics produce tolerance, demonstrate misuse and abuse potential, and are associated with multiple health risks (especially if combined with other medications or alcohol). Overdose deaths from prescription pain medications in the United States have risen sharply in recent years, reaching the level of a public health epidemic (Centers for Disease Control and Prevention: http:// www.cdc.gov/drugoverdose/data/overdose.html). New pharmacological interventions for pain disorders that are safe, effective, and do not produce tolerance are thus urgently needed. Neurosteroids may be logical candidates in this regard, as they are endogenous molecules that exhibit analgesic actions. 26
Although several compelling preclinical studies suggest an important role for allopregnanolone in analgesia, to date few clinical studies investigating neurosteroids and pain have been conducted. A prior study using various experimental pain paradigms demonstrated that healthy individuals with higher endogenous allopregnanolone levels were less likely to report increased pain tolerance during challenge studies [23]. Using a different approach that examined the association of neurosteroids at baseline with self-reported pain, our laboratory found that serum allopregnanolone levels were inversely correlated with pain symptoms in a pilot investigation in 82 Iraq/ Afghanistan-era veterans [24]. The current study attempts to replicate and extend these prior findings by investigating serum allopregnanolone levels in a larger independent cohort of 485 Iraq/Afghanistan-era veterans. Methods Study Population All participants provided written informed consent, and the protocol was approved by the local institutional review board at the Durham Veterans Affairs Medical Center (DVAMC). The study cohort was the following: 485 male veterans who participated in the VA MidAtlantic Mental Illness, Research, Education and Clinic Center (VISN 6 MIRECC) Study of Post-Deployment Mental Health (PDMH Study) at the DVAMC site who had a blood draw between 10:30 am and 2:30 pm. This cohort did not include any veterans who were in the initial pilot study consisting of 82 male Iraq/ Afghanistan-era veterans who participated in the same study effort at the Durham site, and who also had a blood draw between 10:30 am and 2:30 pm [24]. Our aim was thus to attempt to replicate and extend the findings of the prior smaller pilot study in 82 veterans (a “discovery” cohort) in the current non-overlapping distinct larger cohort of 485 veterans (a “validation” cohort).
Downloaded from http://painmedicine.oxfordjournals.org/ by guest on March 25, 2016
Conclusions. These findings are generally consistent with our prior pilot study and suggest that allopregnanolone may function as an endogenous analgesic. Thus, exogenous supplementation with allopregnanolone could have therapeutic potential. The characterization of neurosteroid profiles may also have biomarker utility.
Multiple preclinical studies provide evidence for the role of allopregnanolone in pain analgesia. For example, allopregnanolone increases response latencies to tailflick [11], protects against noxious mechanical stimuli in rats [12], and increases response latencies to thermal stimuli in both rats [13] and invertebrates [14]. Moreover, allopregnanolone has been implicated in both neuropathic pain analgesia [5,15–18] and the alleviation of chemotherapy-induced neuropathies [19]. Allopregnanolone may also be the active metabolite responsible for the analgesic actions of progesterone [20,21], a precursor to allopregnanolone. As allopregnanolone is an endogenous molecule that can be quantified in blood using mass spectrometry-based approaches, its characterization may have biomarker potential for understanding pain symptoms (particularly since brain and serum neurosteroid levels appear to be correlated [22]). If allopregnanolone functions as an endogenous analgesic, a therapeutic approach that enhances levels of this neurosteroid could hold clinical promise.
Allopregnanolone and Pain in Iraq/Afghanistan Veterans To participate in the VISN 6 MIRECC PDMH study, participants must have served in the United States military since September 11, 2001 (i.e., Iraq/Afghanistan-era). Data are included from those who participated in the study between 4/14/06 and 9/22/10. Self-Reported Pain Measure: Symptom Checklist-90R (SCL-90-R)
Neurosteroid Analyses Serum allopregnanolone levels were determined by a highly sensitive and specific gas chromatography/mass spectrometry (GC/MS) method as previously described, [25,26] with minor modifications: specifically, the electron impact mode was utilized rather than negative ion chemical ionization [27–29]. Statistical Analyses Poisson regression procedures were conducted for each of the four SCL-90-R pain items and serum allopregnanolone levels, controlling for age and smoking (number of cigarettes per day). A model regressing an aggregate total pain measure (the sum of all four fivelevel pain items of the SCL-90-R) on the allopregnanolone measure was similarly estimated. Uncorrected rates of inference were set at P ¼ 0.05. Corrected rates controlling the family-wise error rate were based on the
To further investigate putative associations between the primary outcome measures and allopregnanolone levels, we estimated a series of bivariate analyses comparing allopregnanolone levels between participants reporting no or minimal pain symptoms (scores of 0 or 1, or “low pain”) with participants reporting moderate to extreme pain symptoms (scores of 2, 3, or 4, or “high pain”) on each of the four SCL-90-R pain items. Because tests of normality based on the Kolmogorov–Smirnov statistic indicated that the distribution of allopregnanolone deviated from normality (primarily reflecting rightward skewing: D ¼ 0.21; P < 0.01), comparisons were estimated using non-parametric Mann–Whitney–Wilcoxon rank procedures (SAS 9.4; PROC NPAR1WAY). Results Study Population The average age (SD) of this cohort of veterans was 37.40 (10.22) years; 51.96% were Caucasian, 43.51% were African American, 2.27% were Native American, 1.65% were Asian, and 0.21% were Islander; 6.24% identified as Hispanic; 59.2% were currently married; average education was 13.49 (3.26) years. More than one-quarter reported currently smoking (26.7%), consuming on average 11.73 (9.02) cigarettes per day (Table 1). Poisson Regression Analyses Poisson analyses regressing the individual SCL-90-R pain items on the quantified serum allopregnanolone levels controlling for age and smoking revealed inverse associations between allopregnanolone levels and selfreported levels of muscle soreness (P ¼ 0.0028) and
Table 1 Demographic characteristics for male Iraq/Afghanistan-era veterans Variable
N
Age Caucasian African American Native American Asian Islander Hispanic/Latino Married Education (years) Current smoker Cigs smoked/day
485 252 211 11 8 1 30 287 480 129 128
Mean or % 37.40 51.96 43.51 2.27 1.65 0.21 6.24 59.2 13.49 26.70 11.73
SD 10.22
3.26 9.02
Race data available for 483 participants
27
Downloaded from http://painmedicine.oxfordjournals.org/ by guest on March 25, 2016
Participants completed the SCL-90-R, which includes four ratings of pain symptom severity over the past 7 days. These four SCL-90-R pain items—headache [SCL-1], chest pain [SCL-12], low back pain [SCL-27], and muscle soreness [SCL-42] [16]—constituted our primary outcome measures, and were accordingly used to assess pain in the current study. Each pain condition was rated with five possible response options: none (level 0), a little bit (level 1), moderate (level 2), quite a bit (level 3), and extreme (level 4). Each rating item was thus characterized by a five-level integer scale. In this Iraq/Afghanistan-era cohort, self-reported pain items demonstrated a pattern of monotonically decreasing frequencies of endorsement at higher levels of selfreported pain (i.e., fewer participants reported more severe pain). In tests of the primary pain outcomes, individual pain items were regressed on quantified serum allopregnanolone levels (adjusting for age and smoking, variables previously reported to impact this neurosteroid). An aggregate pain measure was also created by summing the four individual pain items of the SCL-90-R to obtain a more generalized pain assessment. For nonparametric bivariate Mann–Whitney analyses (see below), pain scores were dichotomized as previously described [24]. Thus, participants reporting pain levels of 0 or 1 on an SCL-90-R pain item were designated to the “low pain” condition (i.e., “none” or “a little bit”), while participants endorsing pain levels of 2, 3, or 4 were designated to the “high pain” condition (i.e., “moderate” to “extreme” pain).
Hochberg step-up procedure [30], and applied to tests involving any of the four individual SCL-90-R items and to the aggregate total pain model analysis.
Naylor et al.
Table 2
Serum allopregnanolone levels as a function of SCL-90-R pain items Est
Wald v
Pr > v2 P value
Corrected Prob Corrected P value
0.22 0.00 0.01 0.00
0.22 0.00 0.00 0.00
1.07 0.36 1.81 3.07
0.30 0.55 0.18 0.08
0.16
0.44 0.00 0.00 0.00
0.34 0.01 0.01 0.00
1.61 0.40 0.14 4.62
0.20 0.53 0.71 0.032
0.09
0.42 0.00 0.00 0.00
0.17 0.00 0.00 0.00
5.67 0.40 0.87 1.49
0.02 0.53 0.35 0.22
0.22
0.45 0.02 0.01 0.00
0.23 0.00 0.00 0.00
4.02 14.15 2.01 8.96
0.05 0.00 0.16 0.0028
0.011
1.35 0.00 0.00 0.00
0.11 0.00 0.00 0.00
149.84 3.29 4.51 14.48
Allopregnanolone Levels Are Inversely Associated with Self-Reported Pain Symptoms in U.S. Iraq and Afghanistan-Era Veterans: Implications for Biomarkers and Therapeutics
*Research and Development/Mental Health Services, Durham Veterans Affairs Medical Center, Durham, North Carolina; †Department of Psychiatry and Behavioral Sciences, Duke University Medical Center, Durham, North Carolina; ‡Veterans Integrated Service Networks (VISN)6 Mid-Atlantic Mental Illness, Research Education and Clinical Center (MIRECC), Durham, North Carolina; §Department of Psychiatry, University of North Carolina, Chapel Hill, North Carolina; ¶Department of Biostatistics, University of North Carolina, Chapel Hill, North Carolina, USA Correspondence to: Christine E. Marx, MD, MA, Associate Professor, Durham VA Medical Center, 508 Fulton St., Durham, NC 27705, USA. Tel: 919-2860411-3626; Fax: 919-416-5912; E-mail: [email protected]; [email protected]. The VA Mid-Atlantic MIRECC workgroup includes: Jean C. Beckham, PhD; Mira Brancu, PhD.; Patrick S. Calhoun, PhD; John A. Fairbank, PhD; Harold S. Kudler, MD; Scott D. Moore, MD, PhD; Kristy A. Straits-Troster, PhD; Richard D. Weiner, MD, PhD Funding sources: This work was supported by the VA Mid-Atlantic Mental Illness Research Education and Clinical Center (MIRECC), PI John Fairbank, which is supported by the Office of Mental Health Services, Department of Veterans Affairs. Additional support is from VA Career Development Transition Award (Marx) and VA Merit Review Awards (Marx). Dr. Naylor is supported by a Department of Veterans Affairs Rehabilitation
Research and Development Career Development Award (1lK2RX000908). Disclosure and conflicts of interest: Dr. Marx is an applicant or co-applicant on pending patent applications focusing on neurosteroids and derivatives in CNS disorders (co-owned by VA and Duke); no patents issued; no licensing in place; VA 208 waiver in place. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.
Abstract Background and Objectives. Pain symptoms are common among Iraq/Afghanistan-era veterans, many of whom continue to experience persistent pain symptoms despite multiple pharmacological interventions. Preclinical data suggest that neurosteroids such as allopregnanolone demonstrate pronounced analgesic properties, and thus represent logical biomarker candidates and therapeutic targets for pain. Allopregnanolone is also a positive GABAA receptor modulator with anxiolytic, anticonvulsant, and neuroprotective actions in rodent models. We previously reported inverse associations between serum allopregnanolone levels and self-reported pain symptom severity in a pilot study of 82 male veterans.
Methods. The current study investigates allopregnanolone levels in a larger cohort of 485 male Iraq/ Afghanistan-era veterans to attempt to replicate these initial findings. Pain symptoms were assessed by items from the Symptom Checklist-90-R (SCL-90-R) querying headache, chest pain, muscle soreness, and low back pain over the past 7 days. Allopregnanolone levels were quantified by gas chromatography/mass spectrometry.
C 2015 American Academy of Pain Medicine. All rights reserved. For permissions, please e-mail: [email protected] V
25
Downloaded from http://painmedicine.oxfordjournals.org/ by guest on March 25, 2016
Jennifer C. Naylor, PhD,*,†,‡ Jason D. Kilts, PhD,*†,‡ Steven T. Szabo, MD, PhD,*,†,‡ Charlotte E. Dunn, BS,*,‡ Francis J. Keefe, PhD,† Larry A. Tupler, PhD,*,†,‡ Lawrence J. Shampine, BS,*,†,‡ Rajendra A. Morey, MD,*,†,‡ Jennifer L. Strauss, PhD,*,†,‡ Robert M. Hamer, PhD,§,¶ H. Ryan Wagner, PhD,*,†,‡ MIRECC Workgroup,‡ and Christine E. Marx, MD*,†,‡
Naylor et al. Results. Associations between pain ratings and allopregnanolone levels were examined with Poisson regression analyses, controlling for age and smoking. Bivariate nonparametric Mann– Whitney analyses examining allopregnanolone levels across high and low levels of pain were also conducted. Allopregnanolone levels were inversely associated with muscle soreness [P 5 0.0028], chest pain [P 5 0.032], and aggregate total pain (sum of all four pain items) [P 5 0.0001]. In the bivariate analyses, allopregnanolone levels were lower in the group reporting high levels of muscle soreness [P 5 0.001].
Key Words. Allopregnanolone; Veteran; Neurosteroid
Pain;
GABA;
Introduction Nearly 1.9 million U.S. troops have been deployed to Iraq and Afghanistan since September 11, 2001 [1]. Of these returning veterans, approximately 40–50% may experience chronic pain symptoms [2,3]. Chronic pain adversely affects an individual’s emotional and physical well-being, cognition, quality of life, and functional abilities, and also increases suicide risk [4]. Pain disorders can be particularly challenging to treat in Iraq/ Afghanistan-era veterans, as this cohort often reports complex co-occurring physical and psychological symptoms [5–9]. Treating pain represents a multibillion dollar annual expense for the Veterans Healthcare System [10]. Pharmacological interventions for pain are frequently suboptimal, however, and many veterans experience persistent and unalleviated pain symptoms. Medications for pain symptoms are limited by unfavorable side effect profiles and/or lack of efficacy. Opioid analgesics produce tolerance, demonstrate misuse and abuse potential, and are associated with multiple health risks (especially if combined with other medications or alcohol). Overdose deaths from prescription pain medications in the United States have risen sharply in recent years, reaching the level of a public health epidemic (Centers for Disease Control and Prevention: http:// www.cdc.gov/drugoverdose/data/overdose.html). New pharmacological interventions for pain disorders that are safe, effective, and do not produce tolerance are thus urgently needed. Neurosteroids may be logical candidates in this regard, as they are endogenous molecules that exhibit analgesic actions. 26
Although several compelling preclinical studies suggest an important role for allopregnanolone in analgesia, to date few clinical studies investigating neurosteroids and pain have been conducted. A prior study using various experimental pain paradigms demonstrated that healthy individuals with higher endogenous allopregnanolone levels were less likely to report increased pain tolerance during challenge studies [23]. Using a different approach that examined the association of neurosteroids at baseline with self-reported pain, our laboratory found that serum allopregnanolone levels were inversely correlated with pain symptoms in a pilot investigation in 82 Iraq/ Afghanistan-era veterans [24]. The current study attempts to replicate and extend these prior findings by investigating serum allopregnanolone levels in a larger independent cohort of 485 Iraq/Afghanistan-era veterans. Methods Study Population All participants provided written informed consent, and the protocol was approved by the local institutional review board at the Durham Veterans Affairs Medical Center (DVAMC). The study cohort was the following: 485 male veterans who participated in the VA MidAtlantic Mental Illness, Research, Education and Clinic Center (VISN 6 MIRECC) Study of Post-Deployment Mental Health (PDMH Study) at the DVAMC site who had a blood draw between 10:30 am and 2:30 pm. This cohort did not include any veterans who were in the initial pilot study consisting of 82 male Iraq/ Afghanistan-era veterans who participated in the same study effort at the Durham site, and who also had a blood draw between 10:30 am and 2:30 pm [24]. Our aim was thus to attempt to replicate and extend the findings of the prior smaller pilot study in 82 veterans (a “discovery” cohort) in the current non-overlapping distinct larger cohort of 485 veterans (a “validation” cohort).
Downloaded from http://painmedicine.oxfordjournals.org/ by guest on March 25, 2016
Conclusions. These findings are generally consistent with our prior pilot study and suggest that allopregnanolone may function as an endogenous analgesic. Thus, exogenous supplementation with allopregnanolone could have therapeutic potential. The characterization of neurosteroid profiles may also have biomarker utility.
Multiple preclinical studies provide evidence for the role of allopregnanolone in pain analgesia. For example, allopregnanolone increases response latencies to tailflick [11], protects against noxious mechanical stimuli in rats [12], and increases response latencies to thermal stimuli in both rats [13] and invertebrates [14]. Moreover, allopregnanolone has been implicated in both neuropathic pain analgesia [5,15–18] and the alleviation of chemotherapy-induced neuropathies [19]. Allopregnanolone may also be the active metabolite responsible for the analgesic actions of progesterone [20,21], a precursor to allopregnanolone. As allopregnanolone is an endogenous molecule that can be quantified in blood using mass spectrometry-based approaches, its characterization may have biomarker potential for understanding pain symptoms (particularly since brain and serum neurosteroid levels appear to be correlated [22]). If allopregnanolone functions as an endogenous analgesic, a therapeutic approach that enhances levels of this neurosteroid could hold clinical promise.
Allopregnanolone and Pain in Iraq/Afghanistan Veterans To participate in the VISN 6 MIRECC PDMH study, participants must have served in the United States military since September 11, 2001 (i.e., Iraq/Afghanistan-era). Data are included from those who participated in the study between 4/14/06 and 9/22/10. Self-Reported Pain Measure: Symptom Checklist-90R (SCL-90-R)
Neurosteroid Analyses Serum allopregnanolone levels were determined by a highly sensitive and specific gas chromatography/mass spectrometry (GC/MS) method as previously described, [25,26] with minor modifications: specifically, the electron impact mode was utilized rather than negative ion chemical ionization [27–29]. Statistical Analyses Poisson regression procedures were conducted for each of the four SCL-90-R pain items and serum allopregnanolone levels, controlling for age and smoking (number of cigarettes per day). A model regressing an aggregate total pain measure (the sum of all four fivelevel pain items of the SCL-90-R) on the allopregnanolone measure was similarly estimated. Uncorrected rates of inference were set at P ¼ 0.05. Corrected rates controlling the family-wise error rate were based on the
To further investigate putative associations between the primary outcome measures and allopregnanolone levels, we estimated a series of bivariate analyses comparing allopregnanolone levels between participants reporting no or minimal pain symptoms (scores of 0 or 1, or “low pain”) with participants reporting moderate to extreme pain symptoms (scores of 2, 3, or 4, or “high pain”) on each of the four SCL-90-R pain items. Because tests of normality based on the Kolmogorov–Smirnov statistic indicated that the distribution of allopregnanolone deviated from normality (primarily reflecting rightward skewing: D ¼ 0.21; P < 0.01), comparisons were estimated using non-parametric Mann–Whitney–Wilcoxon rank procedures (SAS 9.4; PROC NPAR1WAY). Results Study Population The average age (SD) of this cohort of veterans was 37.40 (10.22) years; 51.96% were Caucasian, 43.51% were African American, 2.27% were Native American, 1.65% were Asian, and 0.21% were Islander; 6.24% identified as Hispanic; 59.2% were currently married; average education was 13.49 (3.26) years. More than one-quarter reported currently smoking (26.7%), consuming on average 11.73 (9.02) cigarettes per day (Table 1). Poisson Regression Analyses Poisson analyses regressing the individual SCL-90-R pain items on the quantified serum allopregnanolone levels controlling for age and smoking revealed inverse associations between allopregnanolone levels and selfreported levels of muscle soreness (P ¼ 0.0028) and
Table 1 Demographic characteristics for male Iraq/Afghanistan-era veterans Variable
N
Age Caucasian African American Native American Asian Islander Hispanic/Latino Married Education (years) Current smoker Cigs smoked/day
485 252 211 11 8 1 30 287 480 129 128
Mean or % 37.40 51.96 43.51 2.27 1.65 0.21 6.24 59.2 13.49 26.70 11.73
SD 10.22
3.26 9.02
Race data available for 483 participants
27
Downloaded from http://painmedicine.oxfordjournals.org/ by guest on March 25, 2016
Participants completed the SCL-90-R, which includes four ratings of pain symptom severity over the past 7 days. These four SCL-90-R pain items—headache [SCL-1], chest pain [SCL-12], low back pain [SCL-27], and muscle soreness [SCL-42] [16]—constituted our primary outcome measures, and were accordingly used to assess pain in the current study. Each pain condition was rated with five possible response options: none (level 0), a little bit (level 1), moderate (level 2), quite a bit (level 3), and extreme (level 4). Each rating item was thus characterized by a five-level integer scale. In this Iraq/Afghanistan-era cohort, self-reported pain items demonstrated a pattern of monotonically decreasing frequencies of endorsement at higher levels of selfreported pain (i.e., fewer participants reported more severe pain). In tests of the primary pain outcomes, individual pain items were regressed on quantified serum allopregnanolone levels (adjusting for age and smoking, variables previously reported to impact this neurosteroid). An aggregate pain measure was also created by summing the four individual pain items of the SCL-90-R to obtain a more generalized pain assessment. For nonparametric bivariate Mann–Whitney analyses (see below), pain scores were dichotomized as previously described [24]. Thus, participants reporting pain levels of 0 or 1 on an SCL-90-R pain item were designated to the “low pain” condition (i.e., “none” or “a little bit”), while participants endorsing pain levels of 2, 3, or 4 were designated to the “high pain” condition (i.e., “moderate” to “extreme” pain).
Hochberg step-up procedure [30], and applied to tests involving any of the four individual SCL-90-R items and to the aggregate total pain model analysis.
Naylor et al.
Table 2
Serum allopregnanolone levels as a function of SCL-90-R pain items Est
Wald v
Pr > v2 P value
Corrected Prob Corrected P value
0.22 0.00 0.01 0.00
0.22 0.00 0.00 0.00
1.07 0.36 1.81 3.07
0.30 0.55 0.18 0.08
0.16
0.44 0.00 0.00 0.00
0.34 0.01 0.01 0.00
1.61 0.40 0.14 4.62
0.20 0.53 0.71 0.032
0.09
0.42 0.00 0.00 0.00
0.17 0.00 0.00 0.00
5.67 0.40 0.87 1.49
0.02 0.53 0.35 0.22
0.22
0.45 0.02 0.01 0.00
0.23 0.00 0.00 0.00
4.02 14.15 2.01 8.96
0.05 0.00 0.16 0.0028
0.011
1.35 0.00 0.00 0.00
0.11 0.00 0.00 0.00
149.84 3.29 4.51 14.48
