Chronic Pain Treatment With Opioid Analgesics

Benefits versus Harms of Long-term Therapy

Nalini Sehgal; James Colson; Howard S Smith

Expert Rev Neurother. 2013;13(11):1201-1220.

Adverse Effects of Opioids

Common Side Effects

Opioid use, both short term and long term, is associated with a high rate of multiple adverse effects.^[49,53] The pooled RR is estimated to be 1.55 (95% CI: 1.41–1.70) for any adverse event, and 4.05 (95% CI: 3.06–5.38) for dropouts due to adverse events.^[48] Adverse effects occur at all dose ranges,^[92] although the frequency increases with daily opioid use (as compared to intermittent use),^[75] higher doses,^[93] long-term therapy, polypharmacy and decreased renal or hepatic function.^[94] Adverse effects are often cited reasons for discontinuing treatment in clinical trials; 33% subjects dropped out of an open-label trial during titration or maintenance phase and 60% withdrew from another trial because of side effects (Table 1, Table 2 & Table 3).^[39,62] As many as 80% of patients experience at least one adverse event,^[61] the most common are gastrointestinal and CNS side effects. Across studies, a high percentage of patients experience dry mouth (42%), constipation (20–41%), sweating (34%), weight gain (29%), somnolence (14–29%), problems with sleep (25%), memory deficits (24%), loss of appetite (23%), nausea (17–33%), concentration deficits (19%), fatigue (19%), sexual dysfunction (18%), dizziness (12–22%), vomiting (11–15%), pruritus/dry skin (10%) and urinary retention.^{[52,61,63,75,95–97]} While tolerance develops to most side effects, constipation does not improve with time and prophylactic measures to improve laxation are required.

Opioid Endocrinopathy – Effects of Suppression of Hypothalamopituitary–Adrenal & –Gonadal Axes

In the past decade opioid induced changes in endocrine function, specifically sexual dysfunction and hypogonadism, has been reported in men and women on intrathecal, transdermal or sustained release oral opiates.^[98–103] Opioids inhibit release of gonadotropin-releasing hormone (GnRH) and corticotrophin releasing hormones from the hypothalamus and lower the circulating levels of luteinizing hormone (LH), follicle stimulating hormone (FSH), adrenocorticotropic hormone (ACTH) and β-endorphin from the pituitary. Decreased pituitary trophic hormones suppress cortisol, testosterone and estrogen release. A strong inhibition of androgen production occurs rapidly after opioid administration, and there is some evidence for a dose-related decrease in androgens, affecting most patients on sustained release opioids.^[99,100,104] Hypogonadotropic hypogonadism manifests clinically in males as decreased libido, erectile dysfunction, decreased sexual arousal and satisfaction, fatigue, depression, hot flashes and night sweats.^[103,104] Women become amenorrheic or develop anovulatory menstrual cycles and reported reduced libido soon after initiating treatment with sustained-action opioid analgesics.^[103] Hormone supplementation with testosterone in males and estrogen and progestin in females improves libido.^[47,98,105] Infertility is a major negative health consequence of low levels of GnRH, FSH and LH in women. Without the pulsatile release of GnRH, FSH and LH secretion by anterior pituitary stops and there is failure of follicular development and ovulation. Other less appreciated but serious effects of low testosterone are osteoporosis,^[106,107] increased fracture risk,^[108] diminished muscle mass, increased fat mass and anemia.^[109] It is recommended that screening for opioid endocrinopathy should be routine in CNCP patients treated with opioids to allow for early identification and treatment.^[110]

Respiratory Depression & Opioid-related Deaths

Unintentional drug overdose death rates in the USA increased by 124% between 1999 and 2007.^[22] Opioids were largely responsible for this upward trend,^[111] and in 2008 accounted for 73.8% of all prescription drug overdose deaths, surpassing deaths due to heroin and cocaine combined.^[25] Opioid overdose rates in adolescents aged 15–19 years also increased between 2000 and 2009 and partly explained the doubling of death rates from 1.7 to 3.3 per 101,000 during this time.^[112] The increase in drug poisoning coincided with a dramatic increase in prescription of strong opioids for CP treatments in the 1990s. Opioid sales surged between 1999 and 2010 and by 2010 'enough OPR were sold to medicate every American adult with 5 mg of hydrocodone every 4 h for 1 month'.^[25] Most of the opioid related deaths are attributed to increased use of hydrocodone, oxycodone and methadone.^[112] CP patients on long-term opioid therapy and those with nonmedical use are the two main groups at-risk for opioid overdose.^[112] In the CP population, risk for opioid related deaths increases in presence of high dose opioid therapy,^[112–114] concomitant use of CNS depressants, sleep apnea, obesity, prescription opioid misuse or substance abuse. It is estimated that 20% patients with pain receive opioids at >100 mg/day MED (10% from a single prescriber and 10% from multiple providers) and account for 80% of opioid related overdose deaths.^[22,111,115] The remaining 80% patients are prescribed lower doses of opioids (<100 mg/day MED) and account for 20% of all prescription drug overdoses.^[22] A maximum prescribed daily dose of ≥100 mg MED is associated with an adjusted hazard ratio (HR) of 7.18 (95% CI: 4.85–10.65),^[113] and an 8.9-fold increase in the overdose rate (1.8% annual overdose rate) as compared to <20 mg/day of opioids (0.2% annual overdose rate).^[114] A strong association is reported between daily opioid dose and mortality.^[26,92] A significant increase in opioid related deaths is seen even at intermediate doses (50–99 mg/day MED) and the mortality rate increased threefold at average daily dose of 200 mg or more of morphine equivalent (OR: 2.88; 95% CI: 1.79–4.63) relative to daily doses of <20 mg of morphine, or equivalent.^[116] Respiratory depression, a potentially lethal complication of opioid therapy^[117] occurs when initial doses are too high, opioids are titrated too rapidly or opioids are combined with other drugs that may potentiate opioid-induced respiratory depression such as benzodiazepines.^[118] Most deaths occur at night, and instead of improving sleep, long-acting opioids produce more adverse effects during the night. Concomitant medication use, especially nighttime sedatives, obesity and sleep apnea increase the risk. A dose-dependent relationship is seen between chronic opioid use, central sleep apneas and ataxic breathing and this is more frequent at a morphine dose equivalent of 200 mg or higher (OR: 15.4; p = .017).^[119] Due to a greater risk for respiratory depression in patients with comorbid cardiovascular, cerebrovascular or respiratory disorder, opioids should be initiated at as low a dose as possible and any further dose increase monitored closely.

Sudden Deaths & Opioid Induced Cardic Toxicity

Sudden death is a potential problem with all opioids, especially at high doses and with drugs that prolong QTc interval, for example, methadone. In the case of methadone, where mortality rates are disproportionately higher than the prescription rates,^[111,120] the increase in deaths is partly a reflection on the pro-arrhythmic properties of methadone. Methadone is a potent inhibitor of the delayed rectifier potassium ion channel (IKr), and causes prolonged QTc interval and torsades de pointes (TdP) in susceptible individuals. The risk of TdP and sudden death increases at a QTc >500 ms. Buprenorphine and extended release morphine also prolong QTc, but have significantly fewer consequences and less morbidity in terms of TdP and sudden cardiac death. Current recommendations for methadone prescribing include screening EKG to identify those with QTc ≥450, repeat EKG at 1 month and at yearly intervals or whenever the dose exceeds 100 mg/day, discontinuing or reducing methadone dose if QTc interval is ≥500 ms and limiting use of drugs that possess QTc prolonging properties or slow the elimination of methadone.^[121]

Opioid Misuse, Abuse & Harmful Use

In 1994, the WHO developed a lexicon to provide definitions for terms concerning alcohol, tobacco and other drugs.^[201] It describes "abuse (drug, alcohol, chemical, substance or psychoactive substance) is a group of terms in wide use but of varying meaning. In DSM-IIIR, psychoactive substance abuse is defined as a maladaptive pattern of use indicated by …continued use despite knowledge of having a persistent or recurrent social, occupational, psychological or physical problem that is caused or exacerbated by the use (or by) recurrent use in situations in which it is physically hazardous".^[201] The term 'abuse' is sometimes used disapprovingly to refer to any use at all, particularly of illicit drugs. Because of its ambiguity, the term is not used in ICD-10; harmful use and hazardous use are the equivalent terms.^[202] In other contexts, abuse has referred to non-medical or unsanctioned patterns of use, irrespective of consequences. Thus the definition published in 1969 by the WHO Expert Committee on Drug Dependence was "persistent or sporadic excessive drug use inconsistent with or unrelated to acceptable medical practice".^[201,203] Harmful use according to ICD-10 is "a pattern of psychoactive substance use that is causing damage to health. The damage may be physical (e.g., hepatitis following injection of drugs) or mental (e.g., depressive episodes secondary to heavy alcohol intake). Harmful use commonly, but not invariably, has adverse social consequences; social consequences in themselves, however, are not sufficient to justify a diagnosis of harmful use".^[202] The term replaces 'non-dependent use' as a diagnostic term. Hazardous use is 'a pattern of substance use that increases the risk of harmful consequences for the user. Some would limit the consequences to physical and mental health; some would also include social consequences. In contrast to harmful use, hazardous refers to patterns of use that are of public health significance despite the absence of any current disorder in the individual user.' The term is used currently by WHO but is not a diagnostic term in ICD-10. The American Society of Addiction Medicine, American Academy of Pain Medicine and American Pain Society define prescription drug misuse as "Use of a medication (for a medical purpose) other than as directed or indicated, whether willful or unintentional, and whether harm results or not" and prescription drug abuse as "Any use of an illegal drug, and the intentional self-administration of a medication for a nonmedical purpose, such as altering one's state of consciousness, for example getting high".^[122,123] Opioid abuse rates in CP patients, who are on long-term opioid therapy, have been variously reported as 5–26%.^{[21,124–127]} Opioid use disorder in a primary care sample receiving daily opioid therapy (n: 801) is four-times higher than in the general population (3.8% vs 0.9%).^[128] Prescription drug abuse is the fastest growing drug problem in the USA. Data from the 2002–2005 National Survey on Drug Use and Health (NSDUH), showed that an annual average of 4.8% individuals (11.4 million persons) in the USA, aged 12 years or more, used a prescription pain reliever non-medically in the preceding 12 months.^[129] Approximately 5.2 million individuals in the USA reported using prescription analgesics non-medically in the prior month, up from 4.7 million in 2005.^[130] The mean annual direct health care costs for patients who abuse opioids are 8.7-times higher than non-abusers.^[131] Hydrocodone, oxycodone and methadone are the most commonly abused opioids.^[132] Risk factors for prescription drug abuse include a family history of substance use disorder (SUD), serving time in prison, living in poor rural communities, younger age, male gender, White race, pain related functional limitations, post-traumatic stress disorder or mental illness.^{[124,128,133–135]} A history of mental illness disorder (OR: 1.46; p = 0.005) and substance abuse (OR: 2.34; p < 0.001),^[136] are moderate to strong predictors of opioid abuse/dependence. Mental health disorders account for a greater number of 'at risk' population because of a higher prevalence of mental health d/o (45.3%) versus SUD (7.6%) in CNCP.^[136] Patients with CP often mislead physicians about illicit drug use.^[128] There is a strong association of four aberrant drug related behaviors (ADRB) with substance abuse and opioid addiction (OR: 48.27; 13.63–171.04). The four ADRBs are: sedating oneself, using opioids for non-pain reasons, increasing dose without authorization and having felt intoxicated when using opioids.^[128] Treatment with high daily doses (>120 mg/day MED), greater day supply of prescription opioids and use of short-acting schedule II drugs increases risk of opioid misuse.^[125] None of these factors by themselves pose an increased risk, but when several of the risk factors coexist in the same individual, drug abuse risk increases significantly. Thus a pain patient on stable doses of opioids, in controlled settings and absence of genetic or psychosocial factors is at no greater risk. On the other hand, a patient with a personal or family history of substance abuse, and coexisting psychosocial risk factors is at increased risk, especially if treatment with opioids is unstructured and not carefully monitored. Accurately diagnosing addiction in CP patients receiving opioids is complex. Managing the patient with pain and co-occurring opioid abuse is equally challenging. Even in older patients, addiction, abuse and misdirection of prescribed opioids are of concern.

Clearly, there is a need to change prescribing practices to curb the upward trend in opioid related morbidity and mortality. Providers should carefully consider the risks and benefits of opioid therapy in each patient, avoid frequent dose escalations and high dose therapy and terminate opioid therapy if pain and function do not improve despite an adequate trial. All patients should be carefully screened and monitored for misuse and abuse before and during treatment with opioids.^[137] Data from prescription drug monitoring programs and insurance claims, where available, should be used to identify and address misuse and abuse.^[22,137] Recent evidence suggests that provider education and changes in prescribing practices is associated with decline in opioid related deaths. In the state of Utah, providers were educated on six safe opioid prescribing practices: start low and go slow, obtain sleep studies for all patients on moderate or high doses of any long-acting opioid, obtain EKGs for methadone dose increases to and above 50 mg/day, avoid sleep aids and benzodiazepines with opioids, avoid long-acting opioids in acute pain and educate patients and their families about risks.^[138] This basic provider education was associated with decrease in opioid prescription related death rates and improved provider self-reported prescribing behaviors. A 14% drop in number of unintentional overdose deaths from prescription opioids was reported in just 1 year.^[138] In 2007, the Washington (WA) workers compensation system introduced a 'yellow flag' warning dose and required pain medicine consultation in CNCP patients receiving >120 mg/day MED of long-acting opioids and demonstrating no substantial improvement in pain and function. After adoption of these guidelines, the number of WA workers claimants treated with opioids, the mean daily dose for long-acting opioids and the percent of time loss claimants on opioid doses, ≥120 mg/day MED declined for the first time since 1999, the year when opioid regulations were liberalized in WA. A substantial decline in the mean daily dose of long-acting DEA schedule II opioids (by 27%), and the proportion of workers on doses ≥120 mg/day MED (by 35%) was followed by a 50% decrease in number of prescription-opioid related unintentional deaths (from 2009–2010) among injured workers.^[139]

1 2 3 4 5

Next Section

Table 1. Opioid efficacy randomized controlled trials of ≥12 weeks' duration.
Study (year)	Study duration (weeks)	Pain condition and sample size (n)	Drug, dose/day (mg)	Pain relief	Function	Treatment discontinued	Ref.
Jamison et al. (1998 )	16 (RCT) & 16 (open label)	Chronic LBP n = 36	Oxycodone 20 mg (fixed dose) versus ER morphine + oxycodone (titrated dose, maximum allowed 200 mg) versus naproxen	+	+ No validated function tools	Due to AE: 3/36	[33]
Babul et al.(2004)	12	OA knee n = 246	Tramadol ER (100–400 mg/day) versus placebo	+	+	27%	[54]
Markenson et al. (2005)	12.8 (90 days)	OA n = 107	Oxycodone CR versus placebo	+	+	Not reported	[66]
Gana et al. (2006)	12	OA knee or hip n = 1020	Tramadol ER daily 100, 200, 300, 400 versus placebo	+	+ (WOMAC)	Not reported	[141]
Katzet al. (2007)	12 Enriched enrollment	Chronic LBP n = 205	Oxymorphone versus placebo	+	Not reported	D/C due to AE: 18% (during titration) & 8% (later in each gp) LOE: 1% completed: 68 versus 47% (oxymorphone vs placebo)	[56]
Burch et al.(2007)	12	OA knee n = 646	Tramadol contramid (200–300 mg/day) versus placebo	+	Not reported	32% in tramadol gp	[57]
Hale et al. (2007)	12 Enriched enrollment	Ch LBP n = 143	Oxymorphone ER (20–260 mg, median 60 mg) versus placebo	+	Not reported	Total: 30% Due to LOE: 11%	[58]
Hanna et al.(2008)	12	DPN n = 338	Gabapentin (1200–1800 mg/day) + oxycodone (10–80 mg/day) versus gabapentin (1200–1800 mg/day) + placebo	+ Superior pain relief in G + O gp versus G alone	Not reported	64% in oxycodone gp D/C rates lower in G + O	[142]
Vorsanger et al. (2008)	15 (3 open label; & 12 RCT w/o washout)	Ch LBP n = 619 (open) 386 (randomized)	Tramadol ER 200 mg versus 300 mg versus placebo	+	Not reported	38% in tramadol gp	[34]
Vondrackova et al. (2008)	12	Ch LBP n = 463	Oxycodone PR-naloxone (10/5 or 20/10 mg) versus oxycodone PR (10 or 20 mg) versus placebo	+	Not reported	11.9% Oxy PR 11.69% oxycodone PR-naloxone	[59]
Afilalo et al. (2010)	15	OA n = 1023	Tapentadol ER 100–250 mg b.i.d. versus oxycodone CR 20–50 mg b.i.d. versus placebo	+	Not reported	Tapentadol 19.2% [66/34] Oxycodone 43.0% [147/342] higher d/c rate due to LOE (placebo), AE (opioid)	[35]
Buynak et al. (2010)	15	Chronic LBP n = 981	Tapentadol ER 100–250 mg b.i.d. versus oxycodone CR 20–50 mg b.i.d. versus placebo	+	Not reported	52% tapentadol, 47% placebo, 40% oxycodone gp completed study	[143]
DeLemos et al. (2011)	12	OA hip or knee n = 1001	Tramadol ER (100, 200, 300 mg) versus celecoxib (200 mg) versus placebo	+/-	WOMAC	Not reported	[37]
Schwartz et al. (2011)	15 Enriched enrollment	DPN n = 395	Tapentadol ER 100–250 mg b.i.d. versus placebo	+ 30% pain relief at 12 weeks	Not reported	Not reported	[36]

AE: Adverse effect; D/C: Discontinue; DPN: Diabetic polyneuropathy; LOE: Loss of effect; LBP: Low back pain; OA: Osteoarthritis; RCT: Randomized controlled trial; WOMAC: Western Ontario & McMaster Universities Osteoarthritis Index.

Table 2. Opioid efficacy prospective uncontrolled trials of ≥12 weeks' duration.
Study (year)	Study duration	Pain condition & sample size	Drug, dose/day (mg)	Pain relief	Function	Treatment discontinued	Ref.
Roth et al.(2000)	Extension trial 18 months (2 weeks RCT)	OA pain >1 month n = 106	Oxycodone Cr 20 mg versus oxycodone CR 40 mg versus placebo	+ 20% or > decrease from baseline pain	Not significant Activity & lifestyle questionnaire	In first 2 weeks (RCT phase): 52.6% Long term: 56.6%	[40]
Allan et al. (2005)	Open-label comparison study 13 months	Chronic LBP n = 680	Transdermal fentanyl (25–50 μg/h) versus SR morphine (60–140 mg/day)	+	Not reported	Fentanyl 37% versus morphine 31%	[55]
Agarwal et al.(2007)	Open label 16 weeks	Neuropathic pain (neuropathy, DPN, post-amputation) n = 53	Transdermal fentanyl (25–150 μg/h) Average dose: 105.6 μg/h	+	+ activity measured on actigraph	13/53 withdrew at varying time	[45]
Portenoy et al. (2007)	Open-label 3-year registry; terminated before all completed 3 years of treatment	Chronic pain (OA, DPN, LBP) n = 233 39 completed 3 years	Oxycodone CR (10–293.5 mg/day) Mean dose 52.5 mg/day	+	Not reported	D/C:133/233 due to AE: 40 (17%) LOE: 18 (8%) Lost f/u: 27(12%) Other: 48 (21%)	[31]
Rauck et al. (2008)	Open label 6 months	Chronic pain (LBP, OA) n = 126	Oxymorphone (5–100, median 20 mg)	+	+ BPI	16% in titration & 17% in maintenance d/c due to AE	[39]
Wallace et al. (2009)	Open label 12 months	Chronic pain: (cancer & non-cancer) n = 388	OROS hydromorphone (32–48 mg/day)	+	Not reported	Not reported	[42]
Wild et al. (2010)	Open-label comparison study Up to 1 year	Chronic pain (LBP, OA) n = 1117	Tapentadol ER 100–250 mg b.i.d. versus oxycodone CR 20–50 mg b.i.d.	+ around 40% pain relief with both opioids	Not reported	Tapentadol gp : 22.1% Oxycodone gp: 36.8% due to AE	[41]
Friedmann et al. (2011)	Open-label, Phase III trial 12 months	Chronic pain (OA, LBP) n = 823	Remoxy (ER oxycodone) 5–80 mg b.i.d.; mean (SD) dose 32.1 (24.7) mg b.i.d.	+	Not reported	D/C: 443 (50%), due to AE 173 (21%)	[60]
Richarz et al. (2013)	Open-label comparison study 52 weeks	Chronic non-cancer pain n = 112	Hydromorphone ER versus oxycodone CR	+	Not reported	Hydromorphone 1.6% versus oxycodone 0.4%	[43]

AE: Adverse effect; b.i.d.: Twice daily; D/C: Discontinue; DPN: Diabetic polyneuropathy; LBP: Lower back pain; LOE: Loss of effect; OA: Osteoarthritis; RCT: Randomized controlled trial; SD: Standard deviation.

Table 3. Systematic reviews of opioid efficacy trials.
Study (year)	Pain condition/study type & #/n (# subjects)	Drug & route	Pain relief	Function, QOL	AE	D/C due to inadequate relief (A) or AE (B)	Opioid addiction	Conclusions	Ref.
Kalso et al. (2004)	CNCP (NeuP, NoP, mixed) RCT (15): 11/15 (on oral opioids) Open label (6/11) n = 1025 subjects (11/15 RCT)	Oral opioids (oxycodone, morphine, methadone) versus placebo Daily dose: oxycodone (20–45 mg), morphine (30–120 mg), methadone (15 mg) Route: oral & IV Study duration: 4 days to 8 weeks (RCT); 6–24 months (open label)	At least 30% mean decrease in pain in NoP & NeuP	3 out of 8 studies reported improvement. Mood & sleep improved with pain relief. No significant improvement in depression	At least 1 AE in 80% patients: constipation (41%), nausea (32%), somnolence (29%)	54% D/C 44% of 388 (open label) still on opioids at 7–24 months	Assessment limited by short f/u and small number of selected patients	Long-term efficacy indeterminate due to lack of sufficient data Concealment of blinding was questionable	[61]
Sandoval et al. (2005)	CNCP (NeuP, LBP, failed opioid Rx; no diagnosis in 50%) RCT (1), case reports (13), Case series (7) n = 545 subjects	Methadone (previous opioid therapy was not tolerated or not effective) Starting dose 0.2–80 mg/day Max dose: 20 –930 mg/day Oral	Meaningful pain relief in 59% in uncontrolled studies; Statistically significant pain relief in RCT at 20 mg/day	Not reported	Minor side effects	Not reported	Not reported	Lack of high quality studies. Response rates based on poor quality uncontrolled studies Great variability in dose	[144]
Eisenberg et al. (2006)	NeuP (peripheral or central) RCT (23): ST =14 IT = 9	Any opioid (ST: morphine, fentanyl, alfentanyl, meperidine, codeine; LT: morphine, oxycodone, methadone, Levorphanol) Any route Study duration: ST: <24 h IT: median 28 (8–70) days	ST: mixed results IT: opioids versus placebo mean decrease 13 points (0–100 VAS) in spontaneous pain equal to 30% relief	Not reported	Nausea (33%) Constipation (33%) Drowsiness (29%) Dizziness (21%) Vomiting (15%) Calculated NNH	B: 11% (opioids) versus 4% (placebo)	Not reported	No evidence on long-term efficacy	[50]
Furlan et al. (2006)	CNCP NoP 80% NeuP 12% FMS 7% Mixed 1% RCT (41) n = 6019 subjects	Any opioid: Weak (codeine, tramadol, propoxyphene) Strong (morphine, oxycodone) Oral route Study duration average 5 (1–16) weeks	Meta-analysis Opioid versus 1. placebo (n = 28) SMD -0.60, 95% CI: -0.69 to -0.50 2. Other drugs (n = 8) SMD -0.05, 95% CI: -0.32 to -0.21	Meta-analysis Opioid versus 1. Placebo (n = 20) SMD -0.31, 95% CI: -0.41 to -0.22 2. Other drugs (n = 8) SMD 0.16, 95% CI: 0.03–0.30	Compared with placebo: Constipation, nausea, dizziness, somnolence, itching Compared with other drugs: nausea, constipation, somnolence	A: 15% in opioid gp versus 30% in placebo gp. B: 21% in opioid gp versus 10% in placebo gp Overall average dropout rates 33% (opioids) versus 38% (placebo)	Drug craving in 8.7% (morphine) versus 4.3% (placebo) in 1 study	Only strong opioids were statistically superior to naproxen & nortriptyline for pain relief No differences in functional outcomes for 3 drug classes	[63]
Martell et al. (2007)	Ch back pain Total studies (35): Prevalence (11) Efficacy (15; RCT 10, open label 5) ADRB/SUD (9) n = 1008 subjects in efficacy studies	Any opioid: codeine tramadol, dextropropoxyphene, oxymorphone ER, oxycodone CR, SR morphine, morphine (topical) TD fentanyl Compared with placebo (3) Acetaminophen (2) NSAID (1) Other opioid (6) Opioid dose: 73 mg/day (SD 89; 30–232 mg/day) Oral, topical, TD Study duration 64 days (1–16 weeks)	Studies report opioids superior to placebo or active control Meta-analysis (4 studies) no reduced pain with opioids (g, -0.199 composite SMD [95% CI:, -0.49 to 0.11]; p = 0.136)	Functional status not necessarily improved with decrease in pain	ADRB: 5–24%	Retention rates 67–99%	Current SUD 3–43%; life time SUD prevalence 36–56%	Long-term efficacy (≥16 weeks) is unclear SUD are common in LBP patients on opioids Result may be biased, 73% trials were industry sponsored	[32]
Deshpande et al. (2007)	Ch LBP RCT (n = 4)	Oral tramadol versus placebo (3 trials) or oxycodone versus naproxen (1 trial) Study duration 90 days (2 trials) 16 weeks (1 trial)	Tramadol more effective than placebo: SMD 0.71 (0.39–1.02) On reanalysis no statistical significance SMD = -0.58 (-1.42 to 0.26)	Tramadol more effective than placebo: SMD 0.17 (0.04–0.30) On reanalysis no statistical significance SMD= -0.06(-0.88 to 0.76)	Headache (RD 9%, 95% CI: 6–12%) Nausea (RD 3%; 95% CI: 0–6%)	Not reported	Not reported	Question role in LBP. On repeat analysis statistically non-significant improvements in pain and function	[47]
Cepeda et al. (2007)	OA RCT (11) n = 1939 subjects	Tramadol or tramadol/paracetamol versus placebo or active control	12% relative decrease in pain intensity in tramadol or tramadol/paracetamol group	8.5% relative improvement in joint stiffness and function on WOMAC index	1 in 5 with minor AE in tramadol or tramadol/paracetamol group	B: 1 in 8 (95% CI: 7–12) in tramadol or tramadol/paracetamol group	Not reported	Small benefits in pain and function with tramadol or tramadol/paracetamol in patients with OA	[69]
Nuesch et al. (2009)	OA ( hip, knee or both) Total studies (20): RCT (10): parallel group (9) , crossover (1) n = 2268 subjects	Weak and strong non-tramadol opioids: codeine (3 trials), TD fentanyl (1 trial), morphine (1 trial), oxycodone (4 trials), oxymorphone (2 trials) versus placebo Oral or TD routes Median daily dose 51 mg morphine equivalent (13–160 mg) Study duration 4 weeks (3 days to 3 months)	Pain relief more effective with opioids than control: SMD -0.36 (95% CI: -0.47 to -0.26) or 15% (95% CI: 11–20%) difference between opioids & placebo NNT 8 (95% CI: 7–11)	Larger improvements in opioid gp SMD -0.33, (95% CI: -0.45 to -0.12); 0.7 units difference on WOMAC or 13% (95% CI: 9–18%) improvement over placebo NNT 10 (8–15)	55% more likely to have AE. RR 1.55 (95% CI: 1.41–1.70) for any AE and 3.35 (95% CI: 0.83–13.56) for serious AE NNH 12 (95% CI: 10–16)	B: 4-times more likely to dropout or withdraw in opioid group versus placebo RR 4.05, (95% CI: 3.06–5.38) NNH for dropout 19 (95% CI: 13–29)	Not reported	Non tramadol opioid benefits (small to moderate) outweighed by AE (large increases)	[48]
Noble et al. (2010)	CNCP (OA, back pain, NeuP) Total studies (27): (RCT (1); uncontrolled trial continuations or case series (26) n = 4893 subjects	Opioids: weak and strong RCT compared two opioids Route (oral (12 trials), TD (5 trials), IT (10 trials) Study duration: at least 6 months	Clinically significant decrease Variable amount of pain relief among studies	Inconclusive	Many minor AE, for example, nausea, HA	A: Oral: 10.3% TD: 5.8% IT: 7.6% B: Oral: 22.9% (weak 11.4%; strong 34.1%) TD: 12.1% IT: 8.9%	Rare at 0.27% reported in some studies	Weak evidence for clinically significant pain relief long term; Inconclusive evidence for improved function or QOL	[46]
Papaleontiou et al. (2010)	CNCP in >60 years OA (70%), NeuP (13%), Other (17%) RCT (43) Safety & efficacy (40) Misuse or abuse (3)	Any opioid (codeine, propoxyphene, tramadol, morphine, oxycodone, hydromorphone, oxymorphone, fentanyl, methadone, buprenorphine) versus placebo/NSAID/acetaminophen/TCA/gabapentin Study duration: 4 (1.5–156) weeks. Only 5 studies (12%) were >12 weeks	ES -0.557 (p < 0.001) on meta-analysis	ES -0.432 (p < 0.001) on meta-analysis	Constipation (30%), Nausea (28%), dizziness (22%)	A: 8% (0–47%) in opioid versus 16% (0–67%) in control group B: 25% (3–52%) in opioid group versus 8% (-24%) in comparator group	Abuse and misuse behaviors negatively associated with older age (in 1 study 1% >60 and 4% <60 years)	Long-term safety, efficacy, and abuse potential not known. Short-term improvement in pain intensity and physical function but poorer mental health function	[52]
Furlan et al. (2011)	CNCP: NoP (87.1%; OA, RA, non-radicular back pain) NeuP (9.2%; DPN, PHN, phantom limb pain, regional cervico-brachial) FMS (3.2%) Mixed (0.41%) RCT (62): EERW & non-EERW trials n = 11,927 participants (7807 completed, 4120 dropped out)	9 opioids: tramadol, propoxyphene, morphine oxycodone, oxymorphone, methadone, TD fentanyl, TD buprenorphine 47 trials with placebo control 8 trials compared with NSAID Study duration: <6 weeks (46 trials) 11–16 weeks (15 trials) 24 weeks (1 trial on propoxyphene)	Medium ES in favor of opioids (ES = 0.58, 95% CI: 0.48–0.67) Only strong opioids more effective than Naproxen 1000 mg/day Weak or strong opioids not more effective than TCA or AED	Small ES (ES = 0.34, 95% CI: 0.25–0.43) NSAIDs significantly better than opioids but with very small ES (0.18) Opioids not more effective than TCA or AED	Statistically & clinically significant AE: Nausea (17%), constipation (20%), somnolence/drowsiness (14%), dizziness/vertigo (12%), dry skin/itching/pruritus (10%), vomiting (11%)	Average dropout is 35% (opioid) versus 38% (control) A: 15% (opioid group) versus 30% (control) B: 21% (opioid group) versus 10% (control) AE Underestimated in EERW trial design	Not reported	Opioids more effective than placebo: NoP (ES = 0.60, 95% CI: 0.49–0.72) and NeuP (ES = 0.56, 95% CI: 0.38–0.73) Industry funding in 47 trials & 87% with some kind of relationship	[51]
White et al. (2011)	Chronic LBP A summary of previous Cochrane reports RCT	Three drug classes versus alternative Rx for CLBP: opioids, NSAIDs, antidepressants	Opioids more effective than placebo Opioids versus NSAIDs, no greater benefit in terms of pain & disability	Opioids more effective than placebo with much greater effect size for pain than disability Opioids are no better than NSAIDs	Compared with placebo, significantly greater AE (differences of 2 and 9%)	Not reported	Not reported	Favorable effectiveness compared with placebo, but similar effectiveness compared with NSAIDs & significant side effects & decreased effect with long-term use	[53]
Whittle et al. (2011)	RA RCT: 11 n = 672 participants	Weak & strong opioids: codeine, dextropropoxyphene, tramadol, tilidine, pentazocine, morphine versus placebo or (NSAID, paracetamol/acetaminophen) Study duration: <1 week: 4 trials 1–6 weeks: 7 trials	Superior to placebo RR 1.44 (95% CI: 1.03–2.03) NNT 6 (95% CI: 3–84) NEAR RR (1.20, 95% CI: 0.89–1.61) no difference between opioids & placebo	3 trials assessed function, 1 trial reported improvement	Risk of AE higher with opioids OR 3.90 (95% CI: 2.31–6.56)	A: no different than placebo based on pooled data (4 trials). RR 0.82 (95% CI: 0.34–2.01)		No long-term evidence >6 weeks	[49]

ADRB: Aberrant drug-related behaviors; AE: Adverse effect; AED: Antiepileptic drugs; CNCP: Chronic non-cancer pain; D/C: Discontinue; DPN: Diabetic polyneuropathy; EERW: Enriched enrollment; FMS: Fibromyalgia syndrome; IT: Intermediate term; LBP: Low back pain; LOE: Loss of effect; NeuP: Neuropathic pain; NoP: Nociceptive pain; OA: Osteoarthritis; PHN: Post-herpetic neuralgia; RA: Rheumatoid arthritis; RCT: Randomized controlled trial; RR: Relative risk; SMD: Standardized mean difference; ST: Short term; SUD: Substance use disorder; TCA: Tricyclics.