brainscienceblogs.com

 Pharmacotherapy.  2007;27(11):1571-1587.

 Antidepressant Agents for the Treatment of Chronic Pain and Depression

Michael W. Jann, Pharm.D.; Julian H. Slade, Pharm.D.

Abstract

Depression and painful somatic symptoms commonly occur together. Depression and chronic pain can have devastating effects on a patient’s health, productivity, and overall quality of life. When moderate-to-severe pain exists, it can impair patient function while making treatment more difficult or resistant, with increased severity in depressive symptoms and worse outcomes. A variety of chronic pain syndromes exist, including diabetic neuropathy. A high prevalence of patients with chronic pain display depressive symptoms. Treatment for these conditions relies on pharmacologic therapy coupled with diligent, periodic assessments of changes in symptom severity. The link between pain and depression lies in the central and peripheral nervous systems. The brain stem serves as an important connection between the higher brain centers and the spinal cord. In the brain stem, the neurotransmitters serotonin and norepinephrine modulate pain transmission through ascending and descending neural pathways. Both serotonin and norepinephrine are also key neurotransmitters involved with the pathophysiology of depression. Tricyclic antidepressants are effective treatments for pain and depression; selective serotonin reuptake inhibitors provide less benefit. Duloxetine and venlafaxine, which are serotonin and norepinephrine reuptake inhibitors, were shown in clinical trials to alleviate pain and depressive symptoms. Diabetic neuropathy and other chronic pain syndromes were also shown to benefit from duloxetine and venlafaxine. Antidepressants remain fundamental therapeutic agents for depression and anxiety disorders. Their extended use into chronic pain, depression with physical pain, physical pain with or without depression, and other potential medical conditions should be recognized.

Introduction

Chronic pain syndromes and depression are major medical problems facing our society. Approximately 80% of depressed outpatients who completed self-rating questionnaires reported painful somatic symptoms that included stomach pain, neck and back pain, headache, and nonspecific generalized pain.[1] Among hospitalized patients with depression, 92% reported at least one painful symptom, and 76% reported the presence of multiple painful symptoms.[1] Chronic pain originates from a variety of medical illnesses. Although the term “chronic” can be imprecise, we define it as a time period of at least 3–6 months’ duration. Pain can be categorized into three groups: nociceptive (somatic and visceral), neuropathic (central [e.g., stroke], peripheral [e.g., nerve compression by cancer, diabetic neuropathy], or mixed [e.g., postherpetic neuralgia]), and psychogenic. Nociceptive pain occurs when a tissue or organ is damaged by injury or disease. Neuropathic pain is a result of direct damage to the nervous system or spinal cord. Psychogenic pain has no discernible physical source.

Pain is an unpleasant experience, and it is reasonable that its symptoms are closely linked to depression. In a population of patients with various sources of chronic pain, 28% reported at least one depressive symptom and 43% fulfilled the diagnosis for major depression.[1] The frequency of clinical depression in patients with other diseases in which chronic pain is a significant component is staggering and has been reported to be 30–54%.[2]

Other symptoms that overlap both chronic pain and depression are found in anxiety disorders and would be more commonly associated with generalized anxiety disorder. In patients with generalized anxiety disorder, the psychological symptoms of excessive anxiety, constant worries that are difficult to control, feeling on edge, and poor concentration combined with the physical symptoms of fatigue, muscle tension, restlessness, and sleep disturbance can easily lead to depression. Whereas patients with chronic pain can have many symptoms also found in generalized anxiety, sleep disturbance may be one of the most common indistinguishable symptoms among the three categories of chronic pain, depression, and anxiety.[3] It can be difficult for clinicians to discern the origin of these overlapping symptoms that lead to subsequent problems. We examine the relationship between chronic pain and depression and provide a pharmacologic rationale for the efficacy of antidepressants for the treatment of both debilitating conditions. We also discuss the implications of antidepressants that possess both norepinephrine and serotonergic properties that have been shown to be effective in treating chronic pain and depression in clinical studies.

Chronic Pain Model

As previously stated, pain can be grouped into three basic categories. It is beyond the scope of this article to review extensively its complex peripheral and central mechanisms. However, the basic pathophysiology of pain is provided to establish a foundation for its treatment with antidepressant agents.

For pain to occur, an organic or environmental stimulus must be converted into an electrochemical signal, and then transmitted to higher brain centers for interpretation. At that point, it is determined whether the signal is innocuous or noxious in nature. Pain has been described as a complex emotional experience involving not only the transduction of noxious stimuli, but also cognitive and emotional processing by the brain.[4] Pain is not homogeneous and involves multiple genetic and biochemical mechanisms, nervous system pathways, and neuronal plasticity.[4–7] We briefly review only nociceptive and neuropathic pain mechanisms.

For nociceptive pain that originates from a noxious stimulus, the process is initiated at the nociceptor. The two main nociceptor classes include the lightly myelinated, medium-diameter, rapidly conducting Ad fibers, and the unmyelinated, small-diameter, more slowly conducting C fibers.[4] Thereby, Aδ fibers mediate rapid, acute, sharp pain, and C fibers mediate delayed, more diffuse, dull pain. A wide range of stimuli triggers the pain sequence, which could involve a rapid and/or delayed response. Each stimulus has a corresponding receptor that triggers the pain process ( Table 1 ). For example, heat or thermal exposure elicits a rapid response through activation of vanilloid receptor subtype 1 and vanilloid-like receptor subtype 1 on Aδ fibers, launching the pain process. Tissue damage from other mechanisms (e.g., medical disease) can release various biochemical stimuli (e.g., glutamate), which act through their corresponding receptors to commence the delayed pain process by way of the C fibers.[4]

Neuropathic pain is associated with disease or injury of the peripheral or central nervous system and presents difficult therapeutic paradigms for clinicians. Diabetes mellitus, immune disorders, cancer, and ischemic disorders are examples of disease processes that can lead to neuropathic pain. Classification of neuropathic pain can be based on its location in the periphery, spinal cord, or brain ( Table 2 ). Some disorders could have multiple locations (e.g., multiple sclerosis).[8] An essential aspect of neuropathic pain is a partial or complete loss of afferent sensory function and paradoxic hypersensitivities (i.e., hyperalgesia and allodynia). Hyperalgesia is the lowering of the pain threshold and an increased response to noxious stimuli, whereas allodynia is the evocation of pain by nonnoxious stimuli. Mechanical hyperalgesia can be divided into three groups: static, punctuate, and dynamic. These groups are easily distinguishable, as static hyperalgesia begins from gentle pressure, punctate starts with a pinprick, and dynamic comes from a light brush that evokes a painful sensation. Allodynia is characterized by sensations or stimuli that are not considered painful, such as touch, warmth, cold, or simple movements eliciting a painful response. Like nociceptive pain, similar biochemical and molecular mechanisms occur in neuropathic pain with the involvement of Aδ and C fibers.[4,8]

Both nociceptive and neuropathic pain stem from the primary sensory neurons and terminate in the dorsal horn of the spinal cord. The dorsal horn is the first site of synaptic transfer to the brain and can be influenced by neuronal plasticity or modulation integral to pain generation and pain hypersensitivity.[5] Neural pathways from the spinal cord dorsal horn activate many brain structures through an ascending pathway that involves the autonomic, perceptual, and cognitive systems, which elicit the pain response displayed clinically.[4]

Depression Model

Clinical symptoms of depression can be grouped into three basic categories: emotional (depressed mood, lack of motivation, disinterest in social activity, anxiety), cognitive (inability to concentrate, poor memory), and physical (insomnia, headache, fatigue, and stomach, back, and neck pain). The physical pain aspects of depression are well recognized among clinicians. For example, the Hamilton Rating Scale for Depression (HAM-D), a clinical rating scale developed in 1960 to assess depression, contains 21 items, eight of which are questions that ask patients about their physical symptoms.

Theories about the biologic basis for depression have evolved over more than 25 years. The principal biochemical basis for depression has focused on two neurotransmitters: serotonin and norepinephrine.[9,10] These two neurotransmitters have also been implicated in the underlying pathophysiology of chronic pain.[4–7] Serotonergic and norepinephrine neurons overlap in the brain, and these two systems interact biochemically and neuroanatomically. In patients with depression, alterations or reductions of these two neurotransmitters and their respective receptors become dysfunctional over time, leading to a dysregulated system. The following six criteria for dysregulation have been proposed: the system is impaired in one or more regulatory or homeostatic mechanisms; basal output of the system is erratic; normal periods of functioning are disrupted; the system is less responsive to environmental stimuli; a slow return to basal activity occurs after the disturbance; and effective agents restore or reregulate the system.[11] Basically, norepinephrine and serotonin concentrations and output become erratic in patients with depression, and antidepressants attempt to restore a “normal” firing rate in neuronal areas and neurotransmitter activity at the synaptic cleft.

Both the serotonin and norepinephrine pathways in the brain and their associated symptoms have been determined. Both pathways originate in the brain stem and project to various brain regions (Figure 1). Serotonin pathways originate at the raphe nucleus and project to the frontal cortex, basal ganglia, hypothalamus, and limbic areas. Norepinephrine pathways originate in the locus ceruleus and project to the frontal cortex, limbic areas, hypothalamus, and cerebellum. The clinical symptoms for mood disturbance can be associated with the frontal cortex and limbic regions. Loss of appetite, weight loss or gain, and loss of pleasure can be connected to the hypothalamus. Therefore, depressive symptoms originate from various brain areas that result in a complex set of clinical presentations to the health care professional. Each symptom can vary over time in intensity and duration, challenging the role of pharmacotherapeutic interventions.
    
Interrelationship Between Chronic Pain and Depression

The link between the higher brain centers involved with depression and pain and the peripheral body regions occurs in the brain stem, with neurotransmission relayed through the spinal cord. Abnormal body activity and functions (e.g., musculoskeletal) are suppressed from the consciousness by the serotonin and norepinephrine descending pathways in the spinal cord that originate in the brain stem.[12,13] This suppression is not always constant and functions as a homeostatic regulator. For example, these descending pathways suppress the body’s input from minor discomforts such as aching muscles and joints.

As the descending serotonin and norepinephrine neurons arise from the brain stem, two areas within the brain stem have been identified as the source of these neurons (Figure 1). As previously mentioned, the dorsal raphe nucleus serves as a basis for serotonin neurons, and the locus ceruleus serves as a foundation for norepinephrine neurons.[14] In fact, specific norepinephrine cell groups A5 and A7 in the locus ceruleus have been identified and provide anatomic evidence of neuronal projections from the brain stem to the spinal cord.[15–17]

A dysfunctional serotonin and norepinephrine system that promotes emotional and vegetative depressive symptoms is likely to also have dysfunctional descending serotonin and norepinephrine pathways, which provides the explanation for depressed patients who also complain of headache, abdominal pain, and musculoskeletal pain in the lower back, joints, and neck, as well as fatigue and energy loss. The serotonin and norepinephrine descending neurons project from the brain stem into the spinal cord’s dorsal horn. Within this area, a complex set of biochemical actions takes place involving other types of neurotransmitters (e.g., γ-aminobutyric acid) and peptides (substance P) modulating the pain stimuli that originates in the peripheral neurons (Figure 2).[18]

Peripheral pathways and descending projections from the brain stem to the spinal cord. The serotonin and norepinephrine descending neurons project from the brain stem into the spinal cord’s dorsal horn, where a complex set of biochemical actions takes place involving other types of neurotransmitters (e.g., γ-aminobutyric acid [GABA]) and peptides (substance P), modulating the pain stimuli that originate in the peripheral neurons. A host of peptides that includes substance P characterize one primary afferent nociceptor pathway, and the other pathway is identified as IB4 for its binding action to the peptide IB4 lectin. Another peripheral pathway that involves norepinephrine neurons contains sympathetic postganglionic neurons (SPGN) and also includes neuropeptide Y.

    
There are two main primary afferent nociceptor pathways that lead into the spinal cord from the periphery (Figure 2).[19] One pathway is characterized by a host of peptides that includes substance P, and the other pathway has binding action to the peptide IB4 lectin. Both pathways are composed of Aδ and C fibers and terminate in the superficial region of the dorsal horn. Another peripheral pathway that involves norepinephrine neurons contains sympathetic postganglionic neurons, as well as neuropeptide Y. A painful stimulus from the periphery can affect multiple pathways, eliciting the complex set of mechanisms leading to and found within the central nervous system.

As stated earlier, numerous receptor systems located on these two nociceptor pathways exist that form the starting point of pain and its interactions with depression. Glutamate, a major excitatory neurotransmitter, affects both pathways and could serve as one common pharmacologic model for the rapid action of the Ad fibers and the slower actions of C fibers.[5] Many other receptor systems and their corresponding stimuli complete this discussion of the circuitry of pain modulation that incorporates depressive symptomatology. This discussion focuses only on serotonin and norepinephrine pathways. Other models are certainly involved, most notably, enkephalin and opioid peptide links.[18]

Many questions remain, such as, what are the functional consequences of coexistent neurotransmitters and other neuromodulator receptor systems in a single neuron? Are there presynaptic influences on descending pathways? What types of interactions occur between serotonin and norepinephrine terminals in the spinal cord? Regardless of these many unanswered questions, depression and pain possess a physiologically linked basis with a discrete central nervous system network involving opioid-like peptides, biogenic amines, glutamate, and other transmitters.

Role of Antidepressant Drugs

For many years, antidepressants have been used in different chronic pain syndromes with or without the presence of clinical depression. Numerous case reports and noncontrolled studies have appeared in the literature. More than 40 double-blind, placebo-controlled trials in which antidepressants had been used to treat headache, facial pain, rheumatic pain, peripheral neuropathic pain, central pain, chronic pain of various origins, and cancer pain reported that these agents benefit these patients.[20] The studies also indicated that the antidepressant’s beneficial effect on pain does not seem to be related to its effect on mood. By incorporating these concepts with the newer models of pain and depression discussed in the previous sections, we unify these concepts together with the animal studies that have evaluated the potential of antidepressant agents to improve pain symptoms.

Animal Pain Models and In Vitro Data

Based on the interrelationship between pain and depression, the serotonin and norepinephrine reuptake inhibitors (SNRIs) may be preferred over other antidepressant pharmacologic classes (e.g., selective serotonin reuptake inhibitors [SSRIs]) because of their dual action on noradrenergic (norepinephrine) and serotonergic (serotonin) activities in the central nervous system. Animal models have shown a dose-dependent response with SNRIs’ ability to decrease pain sensitivity, whereas SSRIs were ineffective.[21] In a comparison of tertiary tricyclic antidepressants (TCAs), SSRIs, and SNRIs in animal torture subjects, TCAs and SNRIs had positive results of at least 80% in acute and chronic pain tests; the response rates to SSRIs were considerably lower for acute pain (44%) and chronic pain (33%).[22] Mice deficient in serotonin and norepinephrine transporters showed altered increased pain sensitivity to allodynia after nerve injury, and TCAs decreased pain sensitivity, which could provide a basic pharmacologic model for antidepressant efficacy in pain.[22]

An in vitro comparison was made between the newer SNRIs duloxetine and venlafaxine in assessing the inhibition of monoamine uptake and transporter binding.[23] Duloxetine was shown to have more potent tritiated hydrogen binding affinity than that of paroxetine (serotonin) and nisoxetine (norepinephrine) on both human cell lines and rat brain area and synaptosomes. Venlafaxine also displayed these properties, but at a lower potency. In another in vitro study, duloxetine, venlafaxine, desipramine, and other antidepressants were evaluated for their inhibition of monoamine transporter binding.[24] Duloxetine showed potent inhibition (mean concentration of which the inhibitor elicits 50% of maximal inhibition [Ki]) of both norepinephrine and serotonin transporters compared with the other agents (norepinephrine: duloxetine Ki 7.5 nmol/L vs venlafaxine Ki 2483 nmol/L and desipramine Ki 3.8 nmol/L; serotonin: duloxetine Ki 0.8 nmol/L vs venlafaxine Ki 82 nmol/L and desipramine Ki 179 nmol/L). Duloxetine was equally potent as sertraline for serotonin inhibition (0.9 nmol/L) and much more potent than sertraline in norepinephrine inhibition (715 nmol/L), demonstrating that duloxetine is a potent norepinephrine and serotonin inhibitor that differs from other antidepressants.

Duloxetine was compared with other antidepressants in a variety of animal pain models.[25] Duloxetine 3–15 mg/kg in a dose-dependent manner was reported to significantly attenuate (p<0.05) late-phase paw-licking behavior, whereas paroxetine lacked any effect. In a formalin model, duloxetine was reported to be more potent than venlafaxine and amitriptyline and also did not produce any motor coordination problems with use of the rotorod (a rotating-rod device used in rodent studies) test. Duloxetine was also shown to be effective in reversing mechanical allodynia behavior in the L5–6 spinal nerve ligation model of neuropathic pain, but was only minimally effective for the tail-flick model of acute nociceptive pain.

In another study, duloxetine was compared with gabapentin, morphine, and ibuprofen in acute nociceptive pain models and inflammatory or persistent pain models.[26] Duloxetine did not produce a significant effect on response latency in the mouse tail-flick model, but had a modest increase in the mouse hot-plate test. Morphine produced dose-dependent analgesic effects in both of these tests. In models of inflammatory or persistent pain, duloxetine, gabapentin, and ibuprofen reversed thermal hyperalgesia and mechanically induced allodynia in a dose-dependent manner. Duloxetine and gabapentin did not have a substantial effect on the rotorod test compared with morphine. Ibuprofen only at doses of 1000 mg/kg produced modest effects on the rotorod test, and doses less than 300 mg/kg had no effect.

Venlafaxine’s in vitro profile somewhat resembles that of duloxetine for both norepinephrine and serotonin activities, and venlafaxine was reported to be effective in several animal models of pain.[27] For example, venlafaxine showed a dose-dependent improvement in reducing pain in rats with the vincristine model of neuropathy.[28] Venlafaxine 10-, 20-, 40-, and 80-mg/kg doses were evaluated. At 40 and 80 mg/kg, venlafaxine produced maximal effects on C-reflex inhibition with a median effective dose calculated at 27.2 mg/kg. Although some benefit was observed with the venlafaxine 10- and 20-mg/kg doses, it was not significant over a 2-hour time course during the study. These results may indicate that higher venlafaxine doses could be needed for treating chronic pain symptoms. The need for higher venlafaxine doses could be due to its weaker potency for norepinephrine activity.

In summary, both duloxetine and venlafaxine were shown to possess significant dose-dependent effects in alleviating pain in a variety of animal models. Both of these agents were compared with other antidepressants including TCAs and SSRIs. The SSRIs did not produce any consistent improvement in the pain models. The tertiary TCAs had improvements in the animal pain models, but also had predictable adverse effects (e.g., motor coordination problems). The animal models support the long-established clinical observations of TCAs being effective and well recognized in various chronic pain syndromes, but also show that these agents have predictable adverse effects that can limit their use. These findings in animal models provide a pharmacologic basis to evaluate the efficacy of duloxetine and venlafaxine in clinical studies of various pain syndromes in patients with or without depression.

Clinical Trials

Depression With Physical Pain Symptoms. Although many double-blind, placebo-controlled trials of antidepressants in patients with depression and pain have been conducted, these studies were usually completed in a small number of patients (< 100).[20] We present newer information with these agents from large multicenter trials. Table 3 presents a summary of the clinical trials that analyzed the efficacy of antidepressants in patients with depression and symptoms of physical pain. Duloxetine was the most studied agent in clinical trials that focused on these two paradigms of depression and physical pain. These clinical studies were conducted in adult populations with a minimum age of 18 years.

A clinical study compared duloxetine with the SSRI paroxetine.[29] Subjects were randomly assigned to receive duloxetine 40, 60, or 80 mg/day, paroxetine 20 mg/day, or placebo for 9 weeks. The HAM-D (17 items) was used as the primary indicator of efficacy. Secondary measures included the 100-mm Visual Analog Scale (VAS), Somatic Symptom Inventory (SSI), Clinical Global Impression of Severity scale (CGIS), and the Patient Global Impression of Improvement scale (PGI-I); these were completed at every visit. The Quality of Life in Depression scale (QoL-D) was conducted at baseline and after 9 weeks of treatment. Significant improvement in VAS overall pain severity was found only with duloxetine 80 mg/day (p<0.01) at weeks 6 and 8. Although improvement in VAS scores were observed with duloxetine 80 mg/day at week 1, it was not statistically significant. Duloxetine 40 mg/day, paroxetine 20 mg/day, and placebo showed minimal improvement, which also lacked statistical significance. On analysis of the HAMD subscale item 13, which specifically assessed painful physical symptoms (backaches, headaches, and muscle aches), significant improvement was reported with duloxetine 80 mg/day (p<0.05) but not with the other treatments. The SSI scores did not show improvement in any treatment group. This preliminary study does indicate that duloxetine 80 mg/day could be effective in the treatment of physical painful symptoms, whereas a lower 40-mg dose was ineffective. Lack of efficacy was also found with paroxetine, which supports the theory that an antidepressant with a dual pharmacologic mechanism of action is needed to achieve therapeutic benefit for both depression and pain.

Another double-blind, placebo-controlled clinical study evaluated duloxetine 60 mg/day versus placebo and was conducted with an assigned randomization of 1:1.[30] Unlike other studies, this was the only study, to our knowledge, to screen for depressed patients who also had levels of self-reported pain before the start of treatment. (In the other studies, patients were not required to meet a minimum threshold for pain, and the studies were not powered for pain outcomes.) Study drug consisted of three duloxetine 20-mg capsules or three identical placebo capsules taken once/day for 9 weeks. The dose could be decreased to two capsules for tolerability reasons (only once during the study) but had to be increased back to three capsules after 3 weeks and remain at that level for the remainder of the study. Prescription pain drugs were not allowed. Antihypertensive drugs were allowed only if the subject had been taking a stable dose for at least 3 months. The primary efficacy evaluation was by the HAM-D, and the secondary measures included the CGIS, VAS for pain assessment, and PGI-I, which were completed at every visit. The QoL-D was conducted at baseline and after 9 weeks of treatment.

Study results indicated that duloxetine showed significant improvement in total HAM-D scores at 2 weeks (p<0.001) with sustained improvement at 9 weeks based on the consistent reduction in total HAM-D scores versus placebo. The mean change between baseline and study end-point scores for duloxetine and placebo at 9 weeks was -10.91 and -6.05 points, respectively (p<0.001). The analysis of the HAM-D subscale item 13 reported significant improvement with duloxetine compared with placebo, when comparing baseline with study end-point scores (duloxetine -0.78 vs placebo -0.49, p<0.013). All other secondary measures also showed significant improvement with duloxetine versus placebo (p<0.001). The most commonly reported adverse events were nausea, dry mouth, and somnolence. This study indicated that duloxetine could be an effective agent for both depression and painful physical symptoms and established a minimum dose-response threshold.

Results of two subsequent multicenter clinical trials with duloxetine substantiated the efficacy of the 60-mg/day dosage versus placebo.[31,32] The study methodology was similar to the previous study[30] with duloxetine 60 mg/day in which randomization was in a 1:1 duloxetine:placebo ratio. The clinical assessments were also identical and included the HAM-D 17 items, CGIS, VAS, PGI-I, QoL-D, and SSI, which focused on symptomatic pain. The VAS was expanded to include subscales of overall pain, headaches, back pain, interference with daily activities, and time in pain while awake. One of the studies[32] also used the Brief Pain Inventory (BPI) as its primary efficacy instrument for pain assessment. The other study[31] was designed to have 80% power to detect a 2.73-point difference in the total HAM-D score but was not powered for pain outcomes. The former study[32] was powered at 80% to detect a treatment effect size of 0.36 with an α of 0.05 and a 15% increase in sample size to factor for early subject terminations. These parameters yielded a treatment group of 141 patients/group with an estimated group difference of VAS scores of 20 or greater (scale 0–100 mm). Both studies used the likelihood mixed-effects model repeated measures to assess the mean changes from baseline for all efficacy outcomes. The primary analysis addressed the association between painful physical symptoms and depression remission (defined as HAM-D score ≤ 7) and response (defined as ≥ 50% reduction in total HAM-D score).

In the first study,[31] the VAS scores significantly improved for all subscales except for headache. Duloxetine was superior to placebo in interference with daily activities and pain while awake (p<0.05) and with shoulder pain, back pain, and overall pain (p<0.05) when comparing baseline with study end point. The VAS overall pain severity showed significant improvement in the duloxetine group at 2 weeks versus the placebo group (p<0.005) and continued to be sustained at week 9 (p<0.05). Weak correlations between VAS overall pain scores and HAM-D total scores were found throughout the study (e.g., week 3, r = 0.226). Of interest, improvement was found with total HAM-D scores. Therefore, although improvement in depression and pain scores were found, the time of improvement and the magnitude of improvement differed between these two paradigms. In the duloxetine group, remission rates were 38.8% versus 24.8% (p=0.027) for pain responders versus nonresponders. In the placebo group, the remission rates were 32.6% versus 12.3% (p<0.001) in the pain responders versus nonresponders. These findings indicated that pain response is needed for remission to occur in a significant number of subjects. The likelihood of pain nonresponse could contribute toward diminished efficacy with duloxetine. Patients with early pain response noted at 2 weeks had a significantly higher probability of depression response compared with patients not showing early pain response (35.4% vs 20.9%, p=0.009). All other secondary scales used in the trial demonstrated significant improvements from baseline to study end point. Significant correlations between QoL-D and VAS pain scores were found (p<0.001). Treatment-emergent adverse events were not reported in this study.

Results from the second study[32] showed that the mean change in BPI score in the duloxetine group had significant improvement by week 1 (p<0.005) with noted improvements at weeks 2 and 5 (p<0.05). Improved BPI scores were observed at other time periods, but these scores were not statistically significant compared with placebo. Mean changes from baseline to end point in VAS pain measures, CGIS, and PGI-I did not differ significantly between groups. When pooled BPI assessments from all the visits were analyzed together, duloxetine was shown to be significantly greater than placebo for mean improvements in pain severity, worst pain, least pain, average pain, and pain right now (p<0.05). Improvement ranged from 33.5–49.6% for the duloxetine group compared with 19.0–39.3% for the placebo group. Mean change in total HAM-D scores did not differ significantly between the two groups (duloxetine -10.85 vs placebo -10.27, p=0.544). The data were then analyzed by comparing patients with no previous depressive episode with patients with one or more previous episodes. Significant improvements were found with duloxetine in the mean change in average BPI scores (p=0.012) and BPI walking interference scores (p<0.001) in patients with one or more previous episodes. These results showed that patients with previous depressive episodes and physical pain symptoms responded to duloxetine compared with placebo. Nausea, dry mouth, and fatigue were the most commonly reported adverse events with duloxetine.

Physical Pain With or Without Depression. A multicenter, double-blind, placebo-controlled trial was conducted to investigate the potential benefits of duloxetine in patients with primary fibromyalgia with or without depression.[33] The pathophysiology of fibromyalgia is unknown, but evidence for dysfunctional serotonin and norepinephrine neurotransmission might play a role in its pain modulation. Only patients with primary fibromyalgia were allowed into the study and were required to score 4 or higher on the pain intensity item (scale 0–10) of the Fibromyalgia Impact Questionnaire (FIQ). After screening, patients had a 1-week, single-blind, placebo lead-in phase and then were randomly assigned in a 1:1 ratio to receive duloxetine or placebo. Duloxetine dosage was a forced titration method from 20 mg/day to 60 mg twice/day for the first 2 weeks in the following manner: 20 mg/day for 5 days, 20 mg twice/day for 5 days, 40 mg twice/day for at least 2 days, and then 60 mg twice/day by week 2. The dosage of 60 mg twice/day remained fixed throughout the remainder of the study.

The coprimary efficacy measures were the FIQ pain severity score and the total FIQ score. Secondary measures included other FIQ subscales such as fatigue, morning tiredness, and stiffness. Other secondary assessments included the BPI, PGI-I, Short Form–36 General Health Survey (SF-36), QoL-D, CGIS, and Sheehan Disability Scale. The severity of depression and anxiety was evaluated by the Beck Depression Inventory-II (BDI-II) and Beck Anxiety Inventory (BAI). The FIQ and BPI were completed at each weekly visit, whereas the remaining assessments were conducted at weeks 4, 8, and 12.

The study results reported that duloxetine showed significant improvement in total FIQ scores versus placebo at week 4 (p=0.005) and consistent improvement to week 12 (-5.53 points, p=0.027). Early improvement was noted at weeks 1 and 2, but the change in total FIQ score was not significant. However, improvement in the pain severity FIQ score was found to be significant at weeks 1 and 2 (p=0.004 and p<0.001, respectively). Maximum pain relief was noted to occur at week 4 (p<0.001). A strong trend was found in the response rate (defined as at least 50% reduction in total FIQ score) in the duloxetine group compared with the placebo group (27.7% vs 16.7%, p=0.06). Significant differences (p<0.05) between duloxetine and placebo were found for the BPI, PGI-I, and CGIS scores, but not with the BDI-II and BAI scores. These findings demonstrate that duloxetine can be effective for the physical pain symptoms regardless of the presence of depression or anxiety in patients with fibromyalgia. Efficacy was reported to be greater in female patients compared with male patients; however, more than 88% of the study patients were women. This study’s results could be biased toward the female population. Another possibility is that male patients may not recognize this disease and, therefore, would not seek treatment.

A 1-year, open-label, single-center study evaluated the efficacy of venlafaxine extended release (XR) for the treatment of chronic pain associated with depression.[34] Patients with a variety of chronic pain syndromes for at least 3 months entered the study. Patients were recruited by referral from their primary care physicians who were treating the patients for their depression. These pain syndromes included chronic back pain, migraine, and chronic regional pain syndromes. Venlafaxine-XR was started after at least a 10-day washout from previous antidepressants except for fluoxetine, which had a 30-day washout. Venlafaxine immediate-release was started at 12.5 mg twice/day for 3 days, and then increased to 37.5 mg twice/day for an additional 3 days. Venlafaxine-XR was then started at 37.5 mg/day and the dose increased every 3 days to 75 mg/day and then to 150 mg/day or higher. The mean study dose was 225 mg/day. Only nonsteroidal antiinflammatory drugs as needed were permitted for intermittent pain. The HAM-D and VAS were used to evaluate depression and pain; QoL-12 was used to assess quality of life. Patients were evaluated in two groups: depression only and depression with pain. After 1 year, HAM-D scores were significantly lower in both groups compared with baseline scores (depression-only group 8.9 vs 17.6, p<0.001; depression and pain group 8.9 vs 16.8, p<0.001). The VAS scores for the depression and pain group also showed significant improvement from baseline to the study end point (8.4 vs 3.6, p<0.001). Baseline and study end-point QoL-12 scores were similar for both patient groups, which showed significant improvement (p<0.001). Only 11 patients discontinued treatment because of adverse events, which were nausea, anxiousness, agitation, and sexual side effects.

In another open-label study, depressed patients were evaluated on the outcome of physical symptoms during 9 months of antidepressant treatment.[35] Patients were included in this study if the primary care physician thought that antidepressant therapy was warranted based on their clinical evaluation. Patients were randomly assigned to receive one of three SSRIs— paroxetine 20 mg/day, fluoxetine 20 mg/day, or sertraline 50 mg/day. Depression outcomes were assessed by the Hopkins Symptoms Checklist–20 items (HSCL-20) and the 9-item Patient Health Questionnaire (PHQ). The HSCL includes a 13- item depression subscale and 7 items that allow for the assessment of the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV). The PHQ-9 is a self-administered test that includes nine DSM-IV depressive symptoms and is a validated measure of depression severity. The PHQ-15, which also was used, is a different assessment that evaluates physical symptoms. Quality-of-life evaluations included the SF-36 and three scales from the Work Limitations Questionnaire (output demand, time management, interpersonal relationships). Medical comorbidity was calculated for each patient with the Chronic Disease Score. Evaluations were conducted at baseline and at 1, 3, 6, and 9 months after enrollment. Major depression was noted in 74% of the patients, dysthymia in 18%, and minor depression in 8%. At least 33–50% of the patients had physical complaints, with 10–20% having severe symptoms (e.g., fatigue, sleep disturbance, pain, and headaches).

The SSRI treatments resulted in a substantial improvement in physical pain symptoms as shown by PHQ-15 scores for only the first month, then the scores plateaued with minimal improvement during the remainder of the study. Depressed patients without physical pain symptoms had the most benefit when PHQ-9 scores were compared with those of depressed patients with physical pain symptoms. Similar to the PHQ-15 scores, improvement in PHQ-9 scores was found to occur in the first month of treatment with consistent benefits. Remitters and partial responders to antidepressant therapy (defined as ≥ 50% improvement in HSCL-20 scores after 3 mo) had significant improvement (p<0.001) in both pain and nonpain physical symptoms compared with nonresponders. Significant differences were not found between remitters and partial responders. A logistic regression analysis revealed that significant predictors of nonresponse included advanced age, presence of depression, poor physical functioning, and low energy level (p<0.10).

This study showed that physical symptoms in depressed patients only initially improved with SSRI treatment, and unlike depression, minimal resolution occurred afterward. The fixed-dose design with the SSRIs could have been a limiting factor. Whether further improvements could take place with increased drug doses, especially with sertraline, was not investigated. However, this was an open-label study, which lends support that SSRIs may not be the best drug choice for depressed patients with physical painful symptoms; a double-blind, placebo-controlled trial would strengthen these findings.

Summary. Most of the above-mentioned clinical trials were conducted mainly with duloxetine and other agents in a double-blind, placebo-controlled environment. These studies demonstrate that antidepressants are safe and effective treatments of physical painful symptoms whether or not comorbid depression and/or anxiety are present for a relatively short time period of up to 12 weeks. An early response was noted during the first few weeks, with continued benefits throughout the study. The trials with a doubleblind, placebo-controlled design typically were 9–12 weeks in duration compared with the much longer duration of 9–12 months for the open-label studies. The shorter study duration for the double-blind, placebo-controlled trials is directly related to their costs, subject compliance, and retention. Nevertheless, whether the clinical trials used double-blind, placebo-controlled or open-label methodologies, they showed a consistent pattern of clinical efficacy for drug response.

Antidepressant Agents in the Treatment of Diabetic Neuropathy

A focus on the treatment of diabetic neuropathy with antidepressants was selected because of the recent approval by the United States Food and Drug Administration (FDA) of duloxetine for this specific disease state. Although numerous studies have been published with TCAs in different pain syndromes including diabetic neuropathy,[20] only a relatively small number of studies that have compared TCAs with SSRIs have explored the requirement of a noradrenergic effect in an agent for analgesia to occur in patients with diabetic neuropathy. Studies with SNRIs in this population of patients have also been conducted. In addition, combination therapy for diabetic neuropathy (e.g., pregabalin) can be prescribed where different pharmacologic approaches are used to maximize therapeutic efficacy in reducing these painful symptoms.

Tricyclic Antidepressants and Selective Serotonin Reuptake Inhibitors

Imipramine was compared with paroxetine in 20 patients with diabetic neuropathy.[36] After 1 week of taking placebo, subjects were randomly assigned into a three-way crossover study to receive imipramine, paroxetine, or placebo for 2 weeks and then cross over to another treatment group. The treatment group duration was 2 weeks, and each subject participated in all three treatment groups. There was no washout period between treatments except for in three patients who were identified as poor metabolizers of cytochrome P450 (CYP) 2D6 and in whom a 2–3-week washout was allowed. Subjects were given a fixed imipramine daily dose of either 50 or 75 mg to achieve a plasma concentration (imipramine + desipramine) of 400–600 nmol/L, and paroxetine was administered as 40 mg/day. Clinical assessments were conducted by patient self-ratings with use of the VAS and by a single physician rating the neuropathy with use of a 6- item observation scale.

Both imipramine and paroxetine were reported to be significantly better than placebo in reducing median VAS scores. The improvement in median VAS scores from baseline was placebo 141.5, paroxetine 81.5, and imipramine 37.0 (p<0.001). Of interest, imipramine was shown to have slightly greater effect than paroxetine. In the neuropathy findings, both agents showed significant improvement compared with placebo in five of six items (not hypesthesia). Improvement with imipramine was significant compared with paroxetine in four of six items: pain, dysesthesia, nightly aggravation, and sleep disturbance (p<0.05). Although imipramine’s adverse-effect profile could partially explain the improvements in nightly aggravation and sleep disturbance, benefits in pain efficacy were clearly noted. Significantly more adverse effects were reported with imipramine than paroxetine and included dry mouth, sweating, fatigue, and palpitations that resulted in five patients dropping out of the study. Four patients reported withdrawal symptoms after imipramine discontinuation. This study showed that imipramine and paroxetine were effective agents in reducing diabetic neuropathy pain in this short time span, with imipramine being slightly more effective than paroxetine, but with more adverse effects.

Amitriptyline, desipramine, and fluoxetine were evaluated in a double-blind, crossover study.[37] Fifty-seven subjects were randomly assigned to one of two treatment groups: one group compared amitriptyline with desipramine and the other compared fluoxetine with placebo. After a 1-week baseline period, subjects were treated with antidepressants for 6 weeks, had a washout period for the next 2 weeks, and then crossed over to the other agent for the next 6 weeks. The placebo selected for the fluoxetine comparison was benztropine to mimic the dry mouth adverse effect of desipramine and amitriptyline. Patients rated their pain daily by choosing from a scale of 13 words describing different magnitudes of pain intensity that showed internal consistency, reliability, and objectivity. At the end of each study period, each patient made a global rating of pain relief in six increments that ranged from complete relief to pain worsening. The mean daily doses were amitriptyline 105 mg, desipramine 111 mg, fluoxetine 40 mg, and placebo (benztropine 1.3 mg).

For the desipramine versus amitriptyline arm (29 patients), both drugs were reported to be effective in decreasing weekly pain from baseline to study point without significant differences between each agent. Seventy-four percent and 61% of patients receiving amitriptyline and desipramine, respectively, reported with the global descriptors of moderate or greater pain relief. In the fluoxetine versus placebo arm, fluoxetine did not differ significantly from placebo in the weekly pain relief scores (p=0.34). Only 48% of the fluoxetine group reported moderate or greater pain relief in the global assessment compared with 41% in the placebo group (p=NS). A few (16) patients had depression in the fluoxetine versus placebo arm, and in those patients fluoxetine was reported to decrease their pain to a similar extent as that in the amitriptyline versus desipramine arm. The pain improvement in the amitriptyline versus desipramine arm occurred with the same magnitude in depressed and nondepressed patients. The most commonly reported adverse effects were dry mouth (amitriptyline, desipramine, and placebo) and headache (fluoxetine). This study showed that modest doses of amitriptyline and desipramine were effective in pain relief, and despite moderate fluoxetine doses, pain relief occurred only in depressed patients. These studies indicate that TCAs should be the first-line agents to treat patients with diabetic neuropathy, and SSRIs should be reserved for those patients who have coexistent depression. Again, efficacy from TCAs support the hypothesis that a noradrenergic pharmacologic profile is needed, and since these agents also possess a serotonergic property, a dual-action antidepressant can be more efficacious compared with a compound with only a single pharmacologic specificity (e.g., SSRIs).

Serotonin and Norepinephrine Reuptake Inhibitors

Several double-blind, placebo-controlled trials with the SNRIs duloxetine and venlafaxine were conducted in patients with diabetic mellitus type 1 or 2 who experienced pain due to bilateral peripheral neuropathy per the Michigan Neuropathy Screening Instrument ( Table 4 ).[38–40] These studies enrolled an adult population with a minimum age of 18 years.

Duloxetine. In one study with duloxetine, subjects were randomly assigned to receive duloxetine 20 mg/day, 60 mg/day, 60 mg twice/day, or placebo for 12 weeks.[38] The primary efficacy instrument used was the weekly assessment from the 24- hour Average Pain Score (APS) on an 11-point Likert Scale. Secondary evaluations included pain severity (from the weekly scores of 24-hr worse pain severity and night pain severity), BPI, CGIS, McGill Pain Questionnaire (MPQ), and SF-36. Clinical response was defined as a 30% reduction from the baseline APS. Sustained response was defined as at least a 30% reduction from baseline in the APS at a visit at least 2 weeks before the last visit with at least 20% reduction consistently from baseline to every visit and in between visits.

Significant mean APS reductions occurred as early as week 1 for the duloxetine 60-mg/day and 120-mg/day groups (p<0.001) versus placebo with continued improvement until week 6. After week 5, both groups had consistent improvement until week 12. A significant difference between the duloxetine 60-mg/day group and the duloxetine 120-mg/day group was not found. The duloxetine 20-mg/day group noted minimal improvement compared with placebo, but at each weekly time point, significance was not achieved.

Response rates were significantly higher in the duloxetine 60-mg/day and 120-mg/day groups (64% and 65%, respectively) versus placebo (47%, p<0.01 for both comparisons), but not in the duloxetine 20- mg/day group (51%). Similarly, the sustained response rate was significantly greater with duloxetine 60 mg/day and 120 mg/day than with placebo (56% both groups vs 33%, p<0.001) but not with the duloxetine 20-mg/day group (46%).

Concomitant acetaminophen was allowed for pain during the study. The median average daily dose of acetaminophen was significantly lower in the duloxetine 60-mg/day and 120-mg/day groups versus the placebo group (74.1 and 80.1 mg, respectively, vs 335.3 mg, p<0.01). Use of acetaminophen was lower in the duloxetine 20- mg/day group (178.3 mg), but it was not statistically significantly different from that in the placebo group.

Secondary measures were reported to significantly improve (p<0.05) with both duloxetine groups (60 and 120 mg/day). Significant differences between duloxetine 20 mg/day and placebo were noted only with the 24- hour worse pain score and the MPQ (p<0.05). Occurrence of adverse effects was higher in the duloxetine 60- and 120-mg/day groups. The most commonly reported adverse effects were nausea, somnolence, dizziness, and constipation. Hemoglobin A1c levels and lipid profiles did not significantly differ among the duloxetine groups and placebo group. These findings support that duloxetine 60 and 120 mg/day is effective in alleviating pain from diabetic neuropathy and that duloxetine 20 mg/day is an ineffective dose for this condition.

Two subsequent, double-blind, placebo-controlled trials used the duloxetine doses of 60 and 120 mg/day.[39] Both studies used identical study methodology for duration and primary and secondary efficacy assessments as that of the above-mentioned study.[38] Clinical response rates and sustained response rates were not reported. Each study showed that both duloxetine doses had significant improvement in 24-hour APS scores by week 1 (p<0.001) with continued improvement at week 2. In one study (334 patients), improvement plateaued after week 2 and remained consistent until week 12. After week 5, the duloxetine 120-mg/day group had a slight improvement, greater than that in the duloxetine 60-mg/day group, but the difference was not significant. For the second study (348 patients), improvement gradually continued until week 5, then plateaued, and improvement remained consistent throughout the time points. No significant differences between both duloxetine groups were found throughout the study in mean 24-hour APS scores. All secondary measures of improvement with both duloxetine groups were found to be significant compared with placebo (p<0.05).

Pooled safety data from all three of these studies were presented, and nausea, somnolence, dizziness, and fatigue were the most frequently reported.[38–40] A slightly higher rate of these adverse effects was found with the duloxetine 120-mg/day group, except for nausea for which the frequency was identical to that in the 60- mg/day group. A slightly greater change in fasting blood glucose level from baseline was found to be significant for the duloxetine 120- mg/day group versus placebo (+9.9 vs -2.47 mg/dl, p<0.01) but not for the duloxetine 60- mg/day group (+8.1 mg/dl). No significant change in hemoglobin A1c level was found between duloxetine 60 mg/day and 120 mg/day compared with placebo.

In summary, these three pivotal studies[38–40] demonstrate that duloxetine 60 mg/day is the best effective dose for treatment of diabetic neuropathy. If the patient does not adequately respond to that dose, a further dose increase may be warranted depending on tolerability. Duloxetine 20 mg/day was not shown to benefit patients, but a 40-mg/day dose was not evaluated. As duloxetine is partially metabolized by CYP2D6, polymorphism status may play a role in the dose-response relationship for this agent, and future studies should be conducted to elucidate this potential correlation.

Venlafaxine Extended-Release. Venlafaxine XR was evaluated in patients with painful diabetic neuropathy.[41] Patients were randomly assigned to three groups: placebo, venlafaxine XR 75 mg/day, or venlafaxine XR 150–225 mg/day for 6 weeks. During the first 3 weeks, patients were in a dose-titration phase; they started at 37.5 mg/day for 1 week and then increased to 75 mg/day at week 2. In the higher dose group, the dose was increased to 150 mg/day at week 3. By week 4, the number of capsules could be individually adjusted to the maximum dose of 225 mg/day within this group. Primary efficacy was assessed with the 100-mm VAS for pain intensity (VAS-PI) and pain relief (VAS-PR). Secondary measurements included the CGIS and percentage of patients achieving a 50% reduction in pain intensity.

The venlafaxine XR 150–225-mg/day group showed a statistically significant difference versus the placebo group only at week 6 (p<0.001) for VAS-PI and VAS-PR scores. Although improvement was found during the other study weeks, the difference was not statistically significant. Minimal improvement was found in the venlafaxine XR 75-mg/day group, but it did not achieve statistical significance at any time point. Similar findings with the CGIS scores were also reported. The response rate was significantly higher for the venlafaxine XR 150–225-mg/day group versus placebo (56% vs 34%, p<0.01) and also for the venlafaxine XR 75-mg/day group (40%, p<0.05). Nausea and somnolence were the most commonly reported adverse effects for both venlafaxine XR groups. A higher number of patients in the venlafaxine XR 75-mg/day group (20%) had a postural decrease in systolic blood pressure exceeding 25 mm Hg compared with venlafaxine XR 150–225 mg/day (12%) and placebo (13%). Seven venlafaxine XR–treated patients were reported to have clinically significant electrocardiographic changes during treatment, but only four of these patients withdrew from the study. Overall, venlafaxine XR was suggested to be efficacious and safe in this short-term study.

This study was of shorter duration compared with the duloxetine clinical trials. Although different clinical assessments were used, it appears that venlafaxine XR does provide some benefit in patients with pain due to diabetic neuropathy.

 Treatment of Other Pain Syndromes

Antidepressants, especially TCAs, have been used for a variety of chronic pain syndromes.[20] Updated information is presented only for the SNRIs. Venlafaxine has been evaluated in a small series of pain syndromes associated with different diseases.[27] A few double-blind, placebo-controlled studies and useful case reports are presented in this section.

Venlafaxine was compared with imipramine in 32 patients with painful polyneuropathy for at least 6 months in whom the diagnosis was confirmed by nerve conduction studies.[42] Patients aged 20–70 years were eligible, and most of the patients had pain due to diabetic neuropathy (15 patients) or nondiabetic pain (17 patients). Patients were randomly assigned in a three-way crossover study to receive venlafaxine, imipramine, or placebo for 4 weeks with a 1- week washout period between treatments. Venlafaxine was dosed as 37.5 mg twice/day for week 1, then 75 mg twice/day for week 2, and then 112.5 mg twice/day for the remaining 2 weeks. Imipramine was dosed 25 mg twice/day for week 1, 50 mg twice/day for week 2, and then 75 mg twice/day for the final 2 weeks. Pain ratings were collected by using a 0–10-point scale for constant pain (burning or pressing), pain paroxysms, touch-evoked pain, and weekly pain pressure, but only baseline and week-4 scores were evaluated. Patients self-rated their global pain relief as complete, good, moderate, slight, or none, or worse pain. Serum venlafaxine and imipramine concentrations were obtained at week 4 of each study period.

Both venlafaxine and imipramine were reported to be effective in improving all of the pain scores (p<0.05). No significant difference between venlafaxine and imipramine was found. In the patient’s global self-ratings, only 7 patients receiving placebo reported benefit (21 patients had no relief or worse pain) compared with venlafaxine (13 had benefit, 17 had no relief or worse pain) and imipramine (17 had benefit, 12 had no relief or worse pain). From the global evaluation, venlafaxine had only a strong trend toward efficacy compared with placebo (p=0.073), whereas imipramine was reported to be significant versus placebo (p=0.003). Again, no significant differences between venlafaxine and imipramine were found (p=0.16).

The most commonly reported adverse events for imipramine were dry mouth and sweating, whereas only tiredness was reported for venlafaxine. Significantly higher mean serum venlafaxine and total drug (venlafaxine + metabolite) concentrations were found for the responders versus nonresponders (p=0.02 and p=0.006, respectively) in patients assigned to receive venlafaxine. No correlations were found between imipramine and imipramine plus desipramine serum concentrations and response. Venlafaxine was reported to be a useful alternative to TCAs for painful polyneuropathy. The duration of this study was much shorter than that of previous studies, and full benefit of the drugs may not have been realized. The adverse-effect profile was not unexpected for these two drugs. Since higher drug concentrations and improvement was found with venlafaxine, this finding should be further explored.

In a case report, venlafaxine was noted to benefit a patient with cancer who had neuropathy from oxaliplatin.[43] After the 10th cycle with oxaliplatin, therapy was stopped because of development of hypesthesias of the fingers (patient was unable to grip utensils). Venlafaxine was started at 37.5 mg/day, and the patient reported a quick recovery of the sensation in the fingers. Three days later, the venlafaxine dosage was increased to 37.5 mg twice/day, which was maintained for the next 6 months without any disruption in functional ability.

In another case report, oxaliplatin was stopped after nine cycles because of painful neuropathy in both legs. Venlafaxine 37.5 mg twice/day was started, and after 2 weeks of treatment, no improvement was noted; venlafaxine was discontinued. Topiramate 50 mg/day was started and increased to twice/day for 1 week. The patient noticed pain relief and was able to walk; this effect was sustained during the month of follow-up.

Only one of the case reports showed that venlafaxine could benefit some difficult cases of patients with neuropathy from other conditions. Although the second case report did not report any improvement with venlafaxine, the dosage may have been too low and the time period too short to fully evaluate its potential use. This patient’s condition was very serious, however, and topiramate was beneficial.

Nonpharmacologic Intervention With Vagal Nerve Stimulation

Nonpharmacologic interventions could also be useful in treating patients with chronic pain and depression. For example, vagal nerve stimulation (VNS) was recently approved by the FDA for treatment of refractory depression.[44,45] Continued antidepressant treatment is recommended along with VNS. The vagal nerve afferents have been implicated to influence nociception and pain.[45] Several case reports have shown that VNS reduces pain in patients with migraine and cluster headaches and those with epilepsy.[46,47] Perhaps, VNS combined with antidepressants could be useful in patients with chronic pain syndromes and depression. Other nonpharmacologic interventions such as transcranial magnetic therapy also used for depression may be useful in patients with chronic pain.

 Conclusion

Chronic pain and depression pose significant clinical issues to the health care system, with long-term consequences for patients. Coexistence of both diseases often negatively affects the patient’s performance at work. Some patients can become disabled and unable to continue in their employment, thereby increasing health care utilization and the economic impact to society. The biologic process for chronic pain and depression appears to share similar mechanisms where serotonin and norepinephrine are involved in the dorsal raphe nucleus and locus ceruleus. Antidepressants with serotonin and norepinephrine activity were found to be effective in the treatment of depression associated with physical pain symptoms and chronic pain syndromes. The SSRIs were clearly shown to have marginal benefit compared with the SNRIs. Tricyclic antidepressants have been used for chronic pain syndromes but possess adverse effects that could limit their long-term use. The SNRIs like venlafaxine and duloxetine have been studied extensively in double-blind, placebo-controlled trials and in open-label studies, which have demonstrated the efficacy of these drugs for depression with physical pain symptoms, physical pain with or without depression, diabetic neuropathy, and other associated pain syndromes.

To date, only one antidepressant agent, duloxetine, has been FDA approved for diabetic neuropathy. The SNRIs, which have a dual pharmacologic profile of serotonin and norepinephrine action, have been consistently shown to be efficacious for pain symptoms. Combination therapy that uses different pharmacologic approaches also could be prescribed.

Antidepressants remain the standard of care for treating depression as well as generalized and other anxiety disorders. Their use extends beyond these areas, however, and it is well accepted that the antidepressants are efficacious in treating chronic pain syndromes. Numerous clinical studies supported these findings. Clinicians should be knowledgeable about the various uses of antidepressants in the treatment of these debilitating medical conditions. Future development of antidepressants may involve multiple neurotransmitters beyond serotonin and norepinephrine that could include dopaminergic pathways, neuropeptides, corticotropin-releasing factor, and other neuropharmacologic systems that account for these wide therapeutic applications.