Psychoneuroendocrinology (2015) 56, 157—167
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/psyneuen
Low-dose hydrocortisone replacement improves wellbeing and pain tolerance in chronic pain patients with opioid-induced hypocortisolemic responses. A pilot randomized, placebo-controlled trial Marni A. Nenke a,b,∗, Clare L. Haylock c,d, Wayne Rankin e, Warrick J. Inder f, Lucia Gagliardi a,b, Crystal Eldridge g, Paul Rolan h, David J. Torpy a a
Endocrine and Metabolic Unit, Royal Adelaide Hospital, Adelaide, Australia School of Medicine, University of Adelaide, Adelaide, Australia c Northern Adelaide Geriatrics Service, Modbury Hospital, Modbury, Australia d School of Medical Sciences, University of Adelaide, Adelaide, Australia e Chemical Pathology Directorate, SA Pathology, Adelaide, Australia f Department of Diabetes and Endocrinology, Princess Alexandra Hospital, Woolloongabba, Australia g Pain and Anaesthesia Research Clinic, University of Adelaide, Adelaide, Australia h Discipline of Pharmacology, School of Medical Sciences, University of Adelaide, Adelaide, Australia b
Received 25 November 2014; received in revised form 6 March 2015; accepted 6 March 2015
KEYWORDS Chronic pain; Opioid; Hypocortisolemia; Hydrocortisone replacement
Abstract Long-term opioid therapy has been associated with low cortisol levels due to central suppression of the hypothalamic—pituitary—adrenal axis. The implications of hypocortisolism on wellbeing have not been established. Our aim was to determine whether intervention with physiologic glucocorticoid replacement therapy improves wellbeing and analgesic responses in patients with chronic non-cancer pain on long-term opioid therapy with mild cortisol deficiency. We performed a pilot randomized, double-blind, placebo-controlled crossover study of oral hydrocortisone replacement therapy in 17 patients recruited from a Pain Clinic at a single tertiary center in Adelaide, Australia. Patients were receiving long-term opioid therapy (≥20 mg morphine equivalents per day for ≥4 weeks) for chronic non-cancer pain with mild hypocortisolism, as defined by a plasma cortisol response ≤350 nmol/L at 60 min following a cold pressor test. The crossover intervention included 28-day treatment with either 10 mg/m2 /day of oral
∗ Corresponding author at: Endocrine and Metabolic Unit, Level 7 Emergency Block, Royal Adelaide Hospital, Adelaide 5000, SA, Australia. Tel.: +61 8 8222 5520; fax: +61 8 8222 5908. E-mail address: [email protected] (M.A. Nenke).
http://dx.doi.org/10.1016/j.psyneuen.2015.03.015 0306-4530/© 2015 Elsevier Ltd. All rights reserved.
158
M.A. Nenke et al. hydrocortisone in three divided doses or placebo. Improvement in wellbeing was assessed using Version 2 of the Short Form-36 (SF-36v2), Brief Pain Inventory-Short Form, and Addison’s disease quality of life questionnaires; improvement in analgesic response was assessed using cold pressor threshold and tolerance times. Following treatment with hydrocortisone, the bodily pain (P = 0.042) and vitality (P = 0.013) subscales of the SF-36v2 were significantly better than scores following treatment with placebo. There was also an improvement in pain interference on general activity (P = 0.035), mood (P = 0.03) and work (P = 0.04) following hydrocortisone compared with placebo. This is the first randomized, double-blind placebo-controlled trial of glucocorticoid replacement in opioid users with chronic non-cancer pain and mild hypocortisolism. Our data suggest that physiologic hydrocortisone replacement produces improvements in vitality and pain experiences in this cohort compared with placebo. Trial registration: Therapeutic Goods Administration Clinical Trials Notification Scheme (Drugs), Trial Number 2012/0476. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Chronic pain is experienced by 20% of the population (Reid et al., 2011). The long-term use of opioid analgesia is becoming increasingly common, particularly in sufferers of chronic non-cancer pain (CNCP) (Manchikanti et al., 2012). This increase in opioid consumption is accompanied by considerable risks; common side effects include constipation, nausea, sedation and physical dependence, while more serious complications include hyperalgesia, immunosuppression, respiratory depression, overdosage death, and neuroendocrine dysfunction (Benyamin et al., 2008; Baldini et al., 2012). Chronic pain and opioid use are also associated with a high incidence of anxiety and depression as well as lower quality of life and inferior self-rated health (Becker et al., 1997; Eriksen et al., 2003). Secretion of the glucocorticoid cortisol is tightly regulated to maintain homeostasis (Chrousos, 2009). Hypothalamic secretion of corticotropin releasing hormone (CRH) is a major driver of pituitary adrenocorticotropic hormone (ACTH) production, which then stimulates secretion of cortisol from the adrenal cortex. Opioids have central effects on the hypothalamic—pituitary—adrenal (HPA) axis, acting via low-affinity delta (␦) and kappa () opioid receptors, where binding results in tonic inhibition of the excitatory ␣1 -noradrenergic pathways that stimulate the release of CRH (Grossman and Besser, 1982; Grossman et al., 1986; Torpy et al., 1993). This is confirmed by the increase in plasma ACTH and cortisol seen following blockade of these opioidergic pathways by the opioid receptor antagonist naloxone (Volavka et al., 1979; Torpy et al., 1993). Direct inhibitory effects of opioids on steroidogenesis by the gonads and adrenals are also described (Aloisi et al., 2009). Episodes of acute adrenal insufficiency in patients receiving chronic opiates have been reported, with recovery of the HPA axis when opioid doses are reduced (Abs et al., 2000; Oltmanns et al., 2005; Mussig et al., 2007). Detailed information and clinical awareness related to opioid-induced hypocortisolism is limited (Vuong et al., 2010) despite the knowledge that states of cortisol deficiency are associated with fatigue, hypotension, altered mood and neurocognitive function, working disability and impaired quality of life (Lovas et al., 2002; Hahner et al., 2007). These signs and
symptoms may overlap with symptoms experienced with chronic pain (Eriksen et al., 2003). Activation of the HPA axis, as well as the adrenomedullary hormonal system and sympathoadrenergic systems, occur in a stressor-specific manner (See Goldstein and Kopin, 2007 for review). Stressors including hemorrhage, insulin, cold, pain and immobilization produce individualized neuroendocrine responses (Pacak et al., 1998). With specific regard to the HPA axis, different peak cortisol responses have been found to the insulin tolerance test, the low-dose (1 g) Synacthen test and the high-dose (250 g) Synacthen test, suggesting that individualized cut-off values should apply to each stimulus (Cho et al., 2014). In this study we have utilized the cold pressor test (CPT), a moderate physiologic stimulator of the HPA axis (Al’absi et al., 2002; Smeets et al., 2008). The CPT is a widely used experimental method for inducing systemic stress via pain (Mitchell et al., 2004) and allows simultaneous study of the HPA axis integrity and pain parameters. We have found that the plasma cortisol response to CPT is impaired in patients receiving long-term opioid therapy (LTOT) (Haylock, 2012) and using our methodology, have obtained method-specific reference ranges to detect hypocortisolemia in these patients. Although low cortisol levels in patients using chronic opioids have been described (Palm et al., 1997; Abs et al., 2000; Oltmanns et al., 2005; Mussig et al., 2007), the implications for health and wellbeing and the response to hormone replacement have not been established. To date, no study has investigated the pain and health-related quality of life effects of glucocorticoid replacement in chronic opioid users. Therefore, we undertook a pilot placebo-controlled double-blind crossover study to determine whether intervention with physiologic glucocorticoid replacement therapy can improve wellbeing and analgesic responses in subjects on LTOT with mild cortisol deficiency.
2. Patients and methods 2.1. Study 1 We undertook a preliminary cross-sectional observational study to determine whether the CPT could detect mild
Hydrocortisone RCT in hypocortisolemic opioid users hypocortisolemia in chronic pain patients using LTOT (Haylock, 2012). The study was approved by the Royal Adelaide Hospital (RAH) Human Research Ethics committee. The normal pain and cortisol responses were examined in a group of older community-dwelling males, recruited via physician referral, local media and flyer advertising as well as a hospital chronic pain database. Patients experiencing chronic pain (pain every day or almost every day for >3 months) treated with sustained release full opioid agonists (≥10 MEDD continuously for ≥4 weeks) or without opioid therapy (controls) were enrolled. Following telephone screening, participants underwent assessment over 2 consecutive days including medical history, anthropometric measurements, waking and 30-min post-waking salivary cortisol measurement, 1 g Synacthen test, and serial plasma cortisol testing with CPT at t = pretest, 30, 60, 90 and 120 min. The Likert 10 point scale was used to assess baseline subjective pain scores. CPT was performed as documented in Section 2.2.5.1, with a maximum tolerance time allowed of 180 s.
2.2. Study 2 2.2.1. Participants and design This was a randomized, double-blind, placebo-controlled crossover trial conducted between August 2012 and April 2014 at a single tertiary center in Adelaide, South Australia. Human Research Ethics Committee Approval was obtained for the Royal Adelaide Hospital. Participants were identified through the Pain and Anaesthesia Research Clinic (PARC) database, and through RAH Pain Clinic patient records. Letters of invitation were sent to potential participants by their treating pain specialist. Inclusion criteria: chronic pain treated with sustained-release opioids at morphine equivalent daily doses (MEDD) of ≥20 mg for ≥4 weeks; age ≥18; adequate venous access; ability to provide written consent; and plasma cortisol levels ≤350 nmol/L obtained 60 min after CPT. Exclusion criteria: use of any glucocorticoid within 2 months of baseline visit; known disease of the HPA axis; severe major organ dysfunction including New York Heart Association class III heart disease, dyspnea on minor exertion, Childs—Pugh score B or C, creatinine clearance 160/100 mmHg despite treatment; Raynaud’s phenomenon; history of clinically significant infection; history or evidence of bradyarrhythmia; current pregnancy or lactation; known hypersensitivity to hydrocortisone; inability to fulfill the requirements of the study; or a change in opioid prescription within 4 weeks of baseline assessments. Withdrawal criteria: intolerable side effects from hydrocortisone; non-compliance with protocol or assessment schedule; withdrawal of participant’s consent; and where there is thought to be potential safety risks to the participant, at the discretion of the study physician. As it was not known prior to the study what the effect of glucocorticoid replacement on the various wellbeing outcome measures would be, power calculations to estimate adequate sample size were not possible. Participants who responded to the study invitation underwent basic screening procedures including familiarization with the cold pressor test and wellbeing questionnaires.
159 Patients who fulfilled criteria were enrolled and provided written consent. 2.2.2. Baseline testing At a minimum of 28 days following the screening procedures, participants underwent baseline testing. Participants collected salivary cortisol at 30 min post wakening, midday, and upon retiring to bed the day before their visit. This was performed as a measure of overall cortisolemia rather than as a test of HPA axis reactivity. Wellbeing questionnaires were completed and CPT with plasma cortisol measurements at pre-test, 30, 60, 90 and 120 min was performed. Participants also underwent low-dose (1 g) Synacthen testing >48 h after baseline CPT with plasma cortisol samples collected immediately before IV injection of tetracosactrin (1 g ACTH 1—24, Synacthen) and at 30 and 60 min post injection. 2.2.3. Interventions The study design is shown in Fig. 1. Participants were randomized to undergo 28 days of either oral hydrocortisone (HCT) or placebo, administered upon waking, at midday and at 1800 h. This was followed by a 2-week washout period then crossover to a 28-day period of the alternate treatment. The washout period was sufficiently long at >10 times the elimination half life of the intervention drug to minimize carry over effect (US Department of Human Services, 2006). Participants completed wellbeing questionnaires at the beginning and end of each treatment period. Repeat CPT occurred at day 28 of each period. Participants received follow up phone calls at day 7 and 14 of each period to assist with compliance and screen for adverse events. Randomization codes were generated and held by the Pharmacy Department, RAH, with an allocation ratio of 1:1. Oral HCT, prescribed at 10 mg/m2 /day in three divided doses, and oral placebo tablets in identical encapsulation were then formulated and dispensed to PARC. Participants were instructed on dose administration and provided with their tablets and compliance diary on day 1 of each treatment period. 2.2.4. Safety considerations Hydrocortisone was prescribed at recommended replacement doses for adrenal insufficiency, reflective of normal adrenal cortisol production, to avoid the development of glucocorticoid excess and symptoms following withdrawal (Nieman, 2014). Nonetheless, participants were asked to report symptoms of glucocorticoid withdrawal following each treatment period. Adverse events (AE) were reported to the investigators at each study visit or at follow up phone calls. Suspension of the study drug for participant health was at the discretion of the primary investigator. 2.2.5. Outcome measures 2.2.5.1. Cold pressor test. We performed CPTs between 1000 h and 1400 h using temperature-controlled water baths with a water pump to prevent laminar warming. Participants submerged their non-dominant hand and forearm into a warm-water bath (34.5—35.5 ◦ C). At 1 min 45 s, a sphygmomanometer cuff was inflated to 20 mmHg below the diastolic blood pressure and at 2 min the limb was removed from the warm-water bath and submerged into a cold-water bath (0.5—1.5 ◦ C). The participant was asked to indicate the time
160
M.A. Nenke et al.
Fig. 1 Study design. CPT—–cold pressor test; BPI-SF—–Brief Pain Intervention Short Form; SF-36v2—–Short Form-36 version 2; AddiQoL—–Addison’s quality of life questionnaire; HCT—–hydrocortisone.
they first experience pain (CPT threshold), then asked to continue submersion until the pain became intolerable (CPT tolerance, a maximum of 120 s). The arm was removed from the cold-water bath and participants indicated when the pain ceased. When performed, cortisol response was measured at pre-test, 30, 60, 90 and 120 min post-CPT. 2.2.5.2. Subjective health status (SHS) questionnaires. Wellbeing was assessed using three SHS questionnaires: Version 2 of the Short Form-36 (SF-36v2), Brief Pain Inventory-Short Form (BPI-SF) and Addison’s disease Quality of Life Questionnaire (AddiQoL). Questionnaires were selfadministered and completed at screening, baseline, day 1 and 28 of each treatment period and at follow up. The SF-36v2 (Ware et al., 2000) is a 36-item survey measuring current functional health, divided into 8 dimensions including physical functioning (PF), role physical (role limitations due to physical health; RP), bodily pain (BP), general health perception (GH), vitality (VT), social functioning (SF), role emotional (role limitations due to emotional health; RE) and mental health (MH). These dimensions combine to produce physical health component (PCS) and mental health component (MCS) summary scores. It has been used extensively for assessment of quality of life and response to treatment in patients with chronic pain (Mielenz et al., 2006; Nicholson et al., 2006; Miller et al., 2013), with higher scores (0—100) indicating better functional health. Scores are standardized to population norms and this paper reports the standardized values. The BPI-SF is a widely used multidimensional pain experience assessment tool, validated for malignant and non-malignant pain (Tan et al., 2004; Keller et al., 2004). It assesses pain severity (questions 3—6) and degree of interference due to pain (question 9 a—g) in the preceding 24 h using 0-to-10 numerical rating scales where 10 is the worst pain (Cleeland, 2009). The AddiQoL is a 30-item questionnaire validated to assess wellness in people with primary adrenal insufficiency [Addison’s disease (AD)]. There are 19 negative and 11 positive statements with each item having 6 response categories, scored in Likert fashion with higher scores (30—120) representing better health-related quality of life (Oksnes et al., 2012).
2.3. Plasma and salivary cortisol measurements Salivettes® (Sarstedt, Nümbrecht Germany) were selfadministered for salivary cortisol collection then stored at
−20 ◦ C until analyzed. All samples were assayed using an electrochemiluminescence immunoassay on a Roche e601 analyzer (Roche Diagnostics, Castle Hill, NSW, Australia).
2.4. Statistical analyses Analyses were performed using GraphPad Prism version 6 for Mac OS X (GraphPad Software, Inc., US). One-sample t test or Wilcoxin signed-rank test was used to compare study values to population values. We compared SHS scores and CPT pain times at the end of HCT treatment to those at the end of placebo treatment using one-tailed paired Student’s t tests. Results are reported as mean ± SD unless otherwise stated. Cortisol conversion 1 nmol/L = 0.036 mcg/dL.
3. 3 Results 3.1. Study 1 See Table 1 for baseline characteristics and cortisol responses. Twenty-six participants including 7 active opioid users were enrolled in and completed the study. All participants were Caucasian. There was no difference in age or anthropometric measurements between the groups. Of the 12 participants with back pain, 6 reported disc prolapse, sciatica or spinal stenosis associated with their degenerative arthritis. Waking salivary cortisol levels as well as basal and post-ACTH-stimulation plasma cortisol levels were 20% lower in opioid-treated patients compared with controls, but this difference was not statistically significant. Five of the seven opioid users met the 550 nmol/L cutoff for adequate adrenal function following 1 g Synacthen test while 18 out of 19 controls reached this level. In contrast, a significantly reduced plasma cortisol response was seen in LTOT users following the CPT, with the blunted responses at 60 and 120 min reaching statistical significance (P = 0.003 and P = 0.046, respectively; Fig. 2). Specifically, mean plasma cortisol levels at 60 min post CPT were 275 nmol/L (95%CI 231—319) in the opioid users and 411 nmol/L (95%CI 334—489) in non-opioid-using controls. Baseline pain scores according to Likert Pain scale were higher in the LTOT group but this was not statistically significant (P = 0.35). Similarly, pain responses to CPT, assessed by pain threshold and tolerance times, were not different between groups. Thus the CPT is able to detect subtle impairment in the HPA axis response to stressful stimuli in
Hydrocortisone RCT in hypocortisolemic opioid users Table 1
161
Study 1 preliminary CPT validation; baseline characteristics and cortisol responses.
Characteristics, mean (SD) Gender (male, %) Age (yr.) BMI (kg/m2 ) Morning salivary cortisol (nmol/L) Pain syndrome, n (%) Back/spine pain Extremity pain* Neuropathic pain Daily opioid dose (oral morphine equivalent), n (%) 10—50 mg 50—100 mg >100 mg Other analgesia, n (%) Paracetamol NSAIDs TCA Pregabalin Response to 1 g Synacthen test (nmol/L), mean (SD) Basal cortisol 30 min cortisol 60 min cortisol Pain response to CPT (s), mean (SD) Threshold Tolerance Likert
Opioid subjects (n = 7)
Control subjects (n = 19)
100% 71 (4) 31 (7) 18 (7.6)
100% 71 (6) 27 (3) 21 (8.2)
5 (71%) 2 (29%) 0
7 (37%) 6 (32%) 6 (32%)
4 (57%) 0 (0%) 3 (43%)
0 0 0
2 (29%) 1 (14%) 0 0 P value 287 (149) 630 (91) 521 (111)
2 4 2 2
338 (107) 695 (80) 587 (109)
0.34 0.09 0.18
26.8 (48.7) 65.9 (78.7) 4.6 (2.9)
14.1 (23.9) 89.4 (68.7) 3.6 (1.9)
0.38 0.46 0.35
(11%) (21%) (11%) (11%)
BMI—–body mass index; NSAIDs—–non-steroidal anti-inflammatory drugs; TCA—–tricyclic antidepressant. * Hand, shoulder, hip, or knee pain due to osteoarthritis.
Fig. 2 Plasma cortisol levels following cold pressor test (CPT). Cortisol is significantly lower in subjects using chronic opioids 60 min after the cold pressor test than healthy controls (Haylock, 2012). * P < 0.05, # P < 0.02.
chronic opioid users, which was not detectable by Synacthen testing. We subsequently used 350 nmol/L as a cut off for detection of mild hypocortisolism in opioid-treated patients.
3.2. Study 2 3.2.1. Participants Supplementary figure (CONSORT 2010 Flow Diagram) documents the flow of participants through enrolment and treatment crossover in the study. Seventeen participants
completed both treatment periods in the study including 9 women and 8 men. The target sample size of 60 participants was not reached due to slow recruitment. Baseline characteristics are detailed in Table 2. The age of participants ranged from 35 to 79 years. Chronic pain diagnoses included low back pain, osteoarthritis, sciatica, headache and complex regional pain syndrome. All participants had experienced their pain for >5 years and almost half had multiple sites of pain. Twelve participants reported a diagnosis of depression (70.6%) and five reported hypertension (29.4%). All except one participant were overweight with body mass index (BMI) > 25 and the majority (58.8%) were obese (BMI > 30). Average MEDD for participants was 124 ± 95 mg. No serious adverse events related to the study drug were reported.
3.2.2. Baseline cortisol levels Five of the seventeen participants had low basal cortisol levels
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/psyneuen
Low-dose hydrocortisone replacement improves wellbeing and pain tolerance in chronic pain patients with opioid-induced hypocortisolemic responses. A pilot randomized, placebo-controlled trial Marni A. Nenke a,b,∗, Clare L. Haylock c,d, Wayne Rankin e, Warrick J. Inder f, Lucia Gagliardi a,b, Crystal Eldridge g, Paul Rolan h, David J. Torpy a a
Endocrine and Metabolic Unit, Royal Adelaide Hospital, Adelaide, Australia School of Medicine, University of Adelaide, Adelaide, Australia c Northern Adelaide Geriatrics Service, Modbury Hospital, Modbury, Australia d School of Medical Sciences, University of Adelaide, Adelaide, Australia e Chemical Pathology Directorate, SA Pathology, Adelaide, Australia f Department of Diabetes and Endocrinology, Princess Alexandra Hospital, Woolloongabba, Australia g Pain and Anaesthesia Research Clinic, University of Adelaide, Adelaide, Australia h Discipline of Pharmacology, School of Medical Sciences, University of Adelaide, Adelaide, Australia b
Received 25 November 2014; received in revised form 6 March 2015; accepted 6 March 2015
KEYWORDS Chronic pain; Opioid; Hypocortisolemia; Hydrocortisone replacement
Abstract Long-term opioid therapy has been associated with low cortisol levels due to central suppression of the hypothalamic—pituitary—adrenal axis. The implications of hypocortisolism on wellbeing have not been established. Our aim was to determine whether intervention with physiologic glucocorticoid replacement therapy improves wellbeing and analgesic responses in patients with chronic non-cancer pain on long-term opioid therapy with mild cortisol deficiency. We performed a pilot randomized, double-blind, placebo-controlled crossover study of oral hydrocortisone replacement therapy in 17 patients recruited from a Pain Clinic at a single tertiary center in Adelaide, Australia. Patients were receiving long-term opioid therapy (≥20 mg morphine equivalents per day for ≥4 weeks) for chronic non-cancer pain with mild hypocortisolism, as defined by a plasma cortisol response ≤350 nmol/L at 60 min following a cold pressor test. The crossover intervention included 28-day treatment with either 10 mg/m2 /day of oral
∗ Corresponding author at: Endocrine and Metabolic Unit, Level 7 Emergency Block, Royal Adelaide Hospital, Adelaide 5000, SA, Australia. Tel.: +61 8 8222 5520; fax: +61 8 8222 5908. E-mail address: [email protected] (M.A. Nenke).
http://dx.doi.org/10.1016/j.psyneuen.2015.03.015 0306-4530/© 2015 Elsevier Ltd. All rights reserved.
158
M.A. Nenke et al. hydrocortisone in three divided doses or placebo. Improvement in wellbeing was assessed using Version 2 of the Short Form-36 (SF-36v2), Brief Pain Inventory-Short Form, and Addison’s disease quality of life questionnaires; improvement in analgesic response was assessed using cold pressor threshold and tolerance times. Following treatment with hydrocortisone, the bodily pain (P = 0.042) and vitality (P = 0.013) subscales of the SF-36v2 were significantly better than scores following treatment with placebo. There was also an improvement in pain interference on general activity (P = 0.035), mood (P = 0.03) and work (P = 0.04) following hydrocortisone compared with placebo. This is the first randomized, double-blind placebo-controlled trial of glucocorticoid replacement in opioid users with chronic non-cancer pain and mild hypocortisolism. Our data suggest that physiologic hydrocortisone replacement produces improvements in vitality and pain experiences in this cohort compared with placebo. Trial registration: Therapeutic Goods Administration Clinical Trials Notification Scheme (Drugs), Trial Number 2012/0476. © 2015 Elsevier Ltd. All rights reserved.
1. Introduction Chronic pain is experienced by 20% of the population (Reid et al., 2011). The long-term use of opioid analgesia is becoming increasingly common, particularly in sufferers of chronic non-cancer pain (CNCP) (Manchikanti et al., 2012). This increase in opioid consumption is accompanied by considerable risks; common side effects include constipation, nausea, sedation and physical dependence, while more serious complications include hyperalgesia, immunosuppression, respiratory depression, overdosage death, and neuroendocrine dysfunction (Benyamin et al., 2008; Baldini et al., 2012). Chronic pain and opioid use are also associated with a high incidence of anxiety and depression as well as lower quality of life and inferior self-rated health (Becker et al., 1997; Eriksen et al., 2003). Secretion of the glucocorticoid cortisol is tightly regulated to maintain homeostasis (Chrousos, 2009). Hypothalamic secretion of corticotropin releasing hormone (CRH) is a major driver of pituitary adrenocorticotropic hormone (ACTH) production, which then stimulates secretion of cortisol from the adrenal cortex. Opioids have central effects on the hypothalamic—pituitary—adrenal (HPA) axis, acting via low-affinity delta (␦) and kappa () opioid receptors, where binding results in tonic inhibition of the excitatory ␣1 -noradrenergic pathways that stimulate the release of CRH (Grossman and Besser, 1982; Grossman et al., 1986; Torpy et al., 1993). This is confirmed by the increase in plasma ACTH and cortisol seen following blockade of these opioidergic pathways by the opioid receptor antagonist naloxone (Volavka et al., 1979; Torpy et al., 1993). Direct inhibitory effects of opioids on steroidogenesis by the gonads and adrenals are also described (Aloisi et al., 2009). Episodes of acute adrenal insufficiency in patients receiving chronic opiates have been reported, with recovery of the HPA axis when opioid doses are reduced (Abs et al., 2000; Oltmanns et al., 2005; Mussig et al., 2007). Detailed information and clinical awareness related to opioid-induced hypocortisolism is limited (Vuong et al., 2010) despite the knowledge that states of cortisol deficiency are associated with fatigue, hypotension, altered mood and neurocognitive function, working disability and impaired quality of life (Lovas et al., 2002; Hahner et al., 2007). These signs and
symptoms may overlap with symptoms experienced with chronic pain (Eriksen et al., 2003). Activation of the HPA axis, as well as the adrenomedullary hormonal system and sympathoadrenergic systems, occur in a stressor-specific manner (See Goldstein and Kopin, 2007 for review). Stressors including hemorrhage, insulin, cold, pain and immobilization produce individualized neuroendocrine responses (Pacak et al., 1998). With specific regard to the HPA axis, different peak cortisol responses have been found to the insulin tolerance test, the low-dose (1 g) Synacthen test and the high-dose (250 g) Synacthen test, suggesting that individualized cut-off values should apply to each stimulus (Cho et al., 2014). In this study we have utilized the cold pressor test (CPT), a moderate physiologic stimulator of the HPA axis (Al’absi et al., 2002; Smeets et al., 2008). The CPT is a widely used experimental method for inducing systemic stress via pain (Mitchell et al., 2004) and allows simultaneous study of the HPA axis integrity and pain parameters. We have found that the plasma cortisol response to CPT is impaired in patients receiving long-term opioid therapy (LTOT) (Haylock, 2012) and using our methodology, have obtained method-specific reference ranges to detect hypocortisolemia in these patients. Although low cortisol levels in patients using chronic opioids have been described (Palm et al., 1997; Abs et al., 2000; Oltmanns et al., 2005; Mussig et al., 2007), the implications for health and wellbeing and the response to hormone replacement have not been established. To date, no study has investigated the pain and health-related quality of life effects of glucocorticoid replacement in chronic opioid users. Therefore, we undertook a pilot placebo-controlled double-blind crossover study to determine whether intervention with physiologic glucocorticoid replacement therapy can improve wellbeing and analgesic responses in subjects on LTOT with mild cortisol deficiency.
2. Patients and methods 2.1. Study 1 We undertook a preliminary cross-sectional observational study to determine whether the CPT could detect mild
Hydrocortisone RCT in hypocortisolemic opioid users hypocortisolemia in chronic pain patients using LTOT (Haylock, 2012). The study was approved by the Royal Adelaide Hospital (RAH) Human Research Ethics committee. The normal pain and cortisol responses were examined in a group of older community-dwelling males, recruited via physician referral, local media and flyer advertising as well as a hospital chronic pain database. Patients experiencing chronic pain (pain every day or almost every day for >3 months) treated with sustained release full opioid agonists (≥10 MEDD continuously for ≥4 weeks) or without opioid therapy (controls) were enrolled. Following telephone screening, participants underwent assessment over 2 consecutive days including medical history, anthropometric measurements, waking and 30-min post-waking salivary cortisol measurement, 1 g Synacthen test, and serial plasma cortisol testing with CPT at t = pretest, 30, 60, 90 and 120 min. The Likert 10 point scale was used to assess baseline subjective pain scores. CPT was performed as documented in Section 2.2.5.1, with a maximum tolerance time allowed of 180 s.
2.2. Study 2 2.2.1. Participants and design This was a randomized, double-blind, placebo-controlled crossover trial conducted between August 2012 and April 2014 at a single tertiary center in Adelaide, South Australia. Human Research Ethics Committee Approval was obtained for the Royal Adelaide Hospital. Participants were identified through the Pain and Anaesthesia Research Clinic (PARC) database, and through RAH Pain Clinic patient records. Letters of invitation were sent to potential participants by their treating pain specialist. Inclusion criteria: chronic pain treated with sustained-release opioids at morphine equivalent daily doses (MEDD) of ≥20 mg for ≥4 weeks; age ≥18; adequate venous access; ability to provide written consent; and plasma cortisol levels ≤350 nmol/L obtained 60 min after CPT. Exclusion criteria: use of any glucocorticoid within 2 months of baseline visit; known disease of the HPA axis; severe major organ dysfunction including New York Heart Association class III heart disease, dyspnea on minor exertion, Childs—Pugh score B or C, creatinine clearance 160/100 mmHg despite treatment; Raynaud’s phenomenon; history of clinically significant infection; history or evidence of bradyarrhythmia; current pregnancy or lactation; known hypersensitivity to hydrocortisone; inability to fulfill the requirements of the study; or a change in opioid prescription within 4 weeks of baseline assessments. Withdrawal criteria: intolerable side effects from hydrocortisone; non-compliance with protocol or assessment schedule; withdrawal of participant’s consent; and where there is thought to be potential safety risks to the participant, at the discretion of the study physician. As it was not known prior to the study what the effect of glucocorticoid replacement on the various wellbeing outcome measures would be, power calculations to estimate adequate sample size were not possible. Participants who responded to the study invitation underwent basic screening procedures including familiarization with the cold pressor test and wellbeing questionnaires.
159 Patients who fulfilled criteria were enrolled and provided written consent. 2.2.2. Baseline testing At a minimum of 28 days following the screening procedures, participants underwent baseline testing. Participants collected salivary cortisol at 30 min post wakening, midday, and upon retiring to bed the day before their visit. This was performed as a measure of overall cortisolemia rather than as a test of HPA axis reactivity. Wellbeing questionnaires were completed and CPT with plasma cortisol measurements at pre-test, 30, 60, 90 and 120 min was performed. Participants also underwent low-dose (1 g) Synacthen testing >48 h after baseline CPT with plasma cortisol samples collected immediately before IV injection of tetracosactrin (1 g ACTH 1—24, Synacthen) and at 30 and 60 min post injection. 2.2.3. Interventions The study design is shown in Fig. 1. Participants were randomized to undergo 28 days of either oral hydrocortisone (HCT) or placebo, administered upon waking, at midday and at 1800 h. This was followed by a 2-week washout period then crossover to a 28-day period of the alternate treatment. The washout period was sufficiently long at >10 times the elimination half life of the intervention drug to minimize carry over effect (US Department of Human Services, 2006). Participants completed wellbeing questionnaires at the beginning and end of each treatment period. Repeat CPT occurred at day 28 of each period. Participants received follow up phone calls at day 7 and 14 of each period to assist with compliance and screen for adverse events. Randomization codes were generated and held by the Pharmacy Department, RAH, with an allocation ratio of 1:1. Oral HCT, prescribed at 10 mg/m2 /day in three divided doses, and oral placebo tablets in identical encapsulation were then formulated and dispensed to PARC. Participants were instructed on dose administration and provided with their tablets and compliance diary on day 1 of each treatment period. 2.2.4. Safety considerations Hydrocortisone was prescribed at recommended replacement doses for adrenal insufficiency, reflective of normal adrenal cortisol production, to avoid the development of glucocorticoid excess and symptoms following withdrawal (Nieman, 2014). Nonetheless, participants were asked to report symptoms of glucocorticoid withdrawal following each treatment period. Adverse events (AE) were reported to the investigators at each study visit or at follow up phone calls. Suspension of the study drug for participant health was at the discretion of the primary investigator. 2.2.5. Outcome measures 2.2.5.1. Cold pressor test. We performed CPTs between 1000 h and 1400 h using temperature-controlled water baths with a water pump to prevent laminar warming. Participants submerged their non-dominant hand and forearm into a warm-water bath (34.5—35.5 ◦ C). At 1 min 45 s, a sphygmomanometer cuff was inflated to 20 mmHg below the diastolic blood pressure and at 2 min the limb was removed from the warm-water bath and submerged into a cold-water bath (0.5—1.5 ◦ C). The participant was asked to indicate the time
160
M.A. Nenke et al.
Fig. 1 Study design. CPT—–cold pressor test; BPI-SF—–Brief Pain Intervention Short Form; SF-36v2—–Short Form-36 version 2; AddiQoL—–Addison’s quality of life questionnaire; HCT—–hydrocortisone.
they first experience pain (CPT threshold), then asked to continue submersion until the pain became intolerable (CPT tolerance, a maximum of 120 s). The arm was removed from the cold-water bath and participants indicated when the pain ceased. When performed, cortisol response was measured at pre-test, 30, 60, 90 and 120 min post-CPT. 2.2.5.2. Subjective health status (SHS) questionnaires. Wellbeing was assessed using three SHS questionnaires: Version 2 of the Short Form-36 (SF-36v2), Brief Pain Inventory-Short Form (BPI-SF) and Addison’s disease Quality of Life Questionnaire (AddiQoL). Questionnaires were selfadministered and completed at screening, baseline, day 1 and 28 of each treatment period and at follow up. The SF-36v2 (Ware et al., 2000) is a 36-item survey measuring current functional health, divided into 8 dimensions including physical functioning (PF), role physical (role limitations due to physical health; RP), bodily pain (BP), general health perception (GH), vitality (VT), social functioning (SF), role emotional (role limitations due to emotional health; RE) and mental health (MH). These dimensions combine to produce physical health component (PCS) and mental health component (MCS) summary scores. It has been used extensively for assessment of quality of life and response to treatment in patients with chronic pain (Mielenz et al., 2006; Nicholson et al., 2006; Miller et al., 2013), with higher scores (0—100) indicating better functional health. Scores are standardized to population norms and this paper reports the standardized values. The BPI-SF is a widely used multidimensional pain experience assessment tool, validated for malignant and non-malignant pain (Tan et al., 2004; Keller et al., 2004). It assesses pain severity (questions 3—6) and degree of interference due to pain (question 9 a—g) in the preceding 24 h using 0-to-10 numerical rating scales where 10 is the worst pain (Cleeland, 2009). The AddiQoL is a 30-item questionnaire validated to assess wellness in people with primary adrenal insufficiency [Addison’s disease (AD)]. There are 19 negative and 11 positive statements with each item having 6 response categories, scored in Likert fashion with higher scores (30—120) representing better health-related quality of life (Oksnes et al., 2012).
2.3. Plasma and salivary cortisol measurements Salivettes® (Sarstedt, Nümbrecht Germany) were selfadministered for salivary cortisol collection then stored at
−20 ◦ C until analyzed. All samples were assayed using an electrochemiluminescence immunoassay on a Roche e601 analyzer (Roche Diagnostics, Castle Hill, NSW, Australia).
2.4. Statistical analyses Analyses were performed using GraphPad Prism version 6 for Mac OS X (GraphPad Software, Inc., US). One-sample t test or Wilcoxin signed-rank test was used to compare study values to population values. We compared SHS scores and CPT pain times at the end of HCT treatment to those at the end of placebo treatment using one-tailed paired Student’s t tests. Results are reported as mean ± SD unless otherwise stated. Cortisol conversion 1 nmol/L = 0.036 mcg/dL.
3. 3 Results 3.1. Study 1 See Table 1 for baseline characteristics and cortisol responses. Twenty-six participants including 7 active opioid users were enrolled in and completed the study. All participants were Caucasian. There was no difference in age or anthropometric measurements between the groups. Of the 12 participants with back pain, 6 reported disc prolapse, sciatica or spinal stenosis associated with their degenerative arthritis. Waking salivary cortisol levels as well as basal and post-ACTH-stimulation plasma cortisol levels were 20% lower in opioid-treated patients compared with controls, but this difference was not statistically significant. Five of the seven opioid users met the 550 nmol/L cutoff for adequate adrenal function following 1 g Synacthen test while 18 out of 19 controls reached this level. In contrast, a significantly reduced plasma cortisol response was seen in LTOT users following the CPT, with the blunted responses at 60 and 120 min reaching statistical significance (P = 0.003 and P = 0.046, respectively; Fig. 2). Specifically, mean plasma cortisol levels at 60 min post CPT were 275 nmol/L (95%CI 231—319) in the opioid users and 411 nmol/L (95%CI 334—489) in non-opioid-using controls. Baseline pain scores according to Likert Pain scale were higher in the LTOT group but this was not statistically significant (P = 0.35). Similarly, pain responses to CPT, assessed by pain threshold and tolerance times, were not different between groups. Thus the CPT is able to detect subtle impairment in the HPA axis response to stressful stimuli in
Hydrocortisone RCT in hypocortisolemic opioid users Table 1
161
Study 1 preliminary CPT validation; baseline characteristics and cortisol responses.
Characteristics, mean (SD) Gender (male, %) Age (yr.) BMI (kg/m2 ) Morning salivary cortisol (nmol/L) Pain syndrome, n (%) Back/spine pain Extremity pain* Neuropathic pain Daily opioid dose (oral morphine equivalent), n (%) 10—50 mg 50—100 mg >100 mg Other analgesia, n (%) Paracetamol NSAIDs TCA Pregabalin Response to 1 g Synacthen test (nmol/L), mean (SD) Basal cortisol 30 min cortisol 60 min cortisol Pain response to CPT (s), mean (SD) Threshold Tolerance Likert
Opioid subjects (n = 7)
Control subjects (n = 19)
100% 71 (4) 31 (7) 18 (7.6)
100% 71 (6) 27 (3) 21 (8.2)
5 (71%) 2 (29%) 0
7 (37%) 6 (32%) 6 (32%)
4 (57%) 0 (0%) 3 (43%)
0 0 0
2 (29%) 1 (14%) 0 0 P value 287 (149) 630 (91) 521 (111)
2 4 2 2
338 (107) 695 (80) 587 (109)
0.34 0.09 0.18
26.8 (48.7) 65.9 (78.7) 4.6 (2.9)
14.1 (23.9) 89.4 (68.7) 3.6 (1.9)
0.38 0.46 0.35
(11%) (21%) (11%) (11%)
BMI—–body mass index; NSAIDs—–non-steroidal anti-inflammatory drugs; TCA—–tricyclic antidepressant. * Hand, shoulder, hip, or knee pain due to osteoarthritis.
Fig. 2 Plasma cortisol levels following cold pressor test (CPT). Cortisol is significantly lower in subjects using chronic opioids 60 min after the cold pressor test than healthy controls (Haylock, 2012). * P < 0.05, # P < 0.02.
chronic opioid users, which was not detectable by Synacthen testing. We subsequently used 350 nmol/L as a cut off for detection of mild hypocortisolism in opioid-treated patients.
3.2. Study 2 3.2.1. Participants Supplementary figure (CONSORT 2010 Flow Diagram) documents the flow of participants through enrolment and treatment crossover in the study. Seventeen participants
completed both treatment periods in the study including 9 women and 8 men. The target sample size of 60 participants was not reached due to slow recruitment. Baseline characteristics are detailed in Table 2. The age of participants ranged from 35 to 79 years. Chronic pain diagnoses included low back pain, osteoarthritis, sciatica, headache and complex regional pain syndrome. All participants had experienced their pain for >5 years and almost half had multiple sites of pain. Twelve participants reported a diagnosis of depression (70.6%) and five reported hypertension (29.4%). All except one participant were overweight with body mass index (BMI) > 25 and the majority (58.8%) were obese (BMI > 30). Average MEDD for participants was 124 ± 95 mg. No serious adverse events related to the study drug were reported.
3.2.2. Baseline cortisol levels Five of the seventeen participants had low basal cortisol levels
