Epidemiology

When considered to be a discrete disorder, fibromyalgia (FM) is generally considered to be the second-most common “rheumatic” disease, behind osteoarthritis. Depending on the diagnostic criteria used, the prevalence of FM ranges from 2% to 8% of the population. It can develop at any age, including in childhood. It occurs in relatively equal frequency in different countries, cultures, and ethnic groups; there is no evidence that this condition at increased rates in “industrialized” countries and cultures.

Individuals who develop fibromyalgia nearly always have a lifelong history of chronic pain in various regions of the body. Most individuals who eventually develop fibromyalgia have pain in multiple body regions beginning earlier in life. The prevalence of any (regional or widespread) chronic musculoskeletal pain in the population is about 30%; thus, if a single individual has had chronic pain in multiple bodily regions over the course of their lifetime, he or she is exhibiting a “pain prone phenotype” that is an important part of the history in a chronic pain patient. Often beginning in childhood or adolescence, individuals who eventually go on to develop fibromyalgia are more likely to experience headaches, dysmenorrhea, temporomandibular joint disorder (TMJD), chronic fatigue, irritable bowel syndrome (IBS) and other functional gastrointestinal (GI) disorders, interstitial cystitis/painful bladder syndrome, endometriosis, and other regional pain syndromes (especially back and neck pain). In fact, what often looks to one healthcare provider as a new episode of acute or subacute pain is in fact simply the latest region of the body experiencing pain. Consequently, many experts in the pain field have begun to feel that especially these “centralized” pain states are best thought of as a single lifelong disease that merely tends to manifest in different bodily regions over time.

In addition to fibromyalgia patients having a high personal lifetime history of chronic pain, there is often a strong family history of chronic pain identifiable. The first degree relatives of fibromyalgia patients are 8 times as likely to have this condition as the family members of controls, and they have very high rates of other chronic pain states. The strong familial predisposition to FM is shared by many chronic pain conditions, and there has been an explosion of knowledge recently regarding the strong role of genetic factors in chronic pain states. Many of the genes that have been identified to date in leading to increases or decreases in the frequency of chronic pain states, or of pain sensitivity, are mostly involved in regulating the breakdown or binding of neurotransmitters that generally increase pain sensitivity (e.g., glutamate) or decrease pain sensitivity (serotonin, norepinephrine, gamma-aminobutyric acid [GABA]). The fact that pain sensitivity is polygenic and that different individuals develop increased pain sensitivity because of imbalances or altered activity of many different neurotransmitters likely partly explains the “ U -shaped curve” seen with many analgesics, wherein they either work fairly well or not at all. Twin studies suggest that approximately 50% of the risk of developing FM or related pain conditions such as IBS and headache is genetic and 50% environmental.

The environmental factors that are most likely to trigger the development of FM are various types of “stressors,” typically those that involve acute pain for at least a few weeks. Psychological stress is but one such stressor. FM or similar illnesses can be triggered by certain types of infections (e.g., Epstein-Barr virus, Lyme disease, Q fever, viral hepatitis), trauma (motor vehicle collisions), and deployment to war.

FM also is commonly seen as a comorbidity in other chronic pain conditions such as osteoarthritis, rheumatoid arthritis, and lupus. Approximately 20% to 30% of individuals with these established rheumatic disorders will meet the criteria for FM. This phenomenon had previously been termed “secondary fibromyalgia,” but since this is so common and might occur in a subset of nearly any chronic pain cohort, a more popular term we prefer to use for this phenomenon is that these individuals have “centralized” their pain. Thus, although peripheral nociceptive input might be responsible for some or even much of that individual’s pain, superimposed central nervous system (CNS) factors also likely amplify pain and lead to other comorbid symptoms such as fatigue, memory problems, and disturbances of sleep and mood. The term “central sensitization” is also sometimes used to describe this phenomenon, but many feel that this particular term should be reserved for the original spinal mechanism that was identified and called by this name rather than the more general phenomenon we now believe can result via a multitude of different spinal and supraspinal mechanisms. Regardless of what term an individual uses to describe this phenomenon, it is becoming increasingly important to identify it. Because emerging evidence suggests that therapies that work best for peripheral, nociceptive pain (e.g., NSAIDs, opioids, injections, and surgical procedures) are less likely to be effective in these individuals.

FM, especially the “primary” form, is also very comorbid with early life and current stress, and many if not most individuals will have a lifetime history of a psychiatric disorder such as depression or anxiety. There is typically more psychiatric and psychological comorbidity seen in tertiary care settings or in individuals who are refractory to treatment. This bidirectional relationship between FM and psychiatric conditions is likely due in part to the fact that there are common triggers to both sets of conditions (e.g., early life stress or trauma), as well as shared pathophysiology (i.e., most of the same neurotransmitters that affect pain transmission also affect mood, memory, fatigue, sleep). Other potentially modifiable risk factors for developing FM include poor sleep, obesity, physical inactivity, or poor job or life satisfaction. Similarly, cognitive factors such as catastrophizing (the feeling that pain is very bad and associated with a poor prognosis for recovery) or fear of movement have been shown to be poor prognostic factors in FM and other chronic pain states.

Pathophysiology

The physiological hallmark of centralized pain, central sensitization, or FM is augmented central pain processing. This was originally identified in FM (and still can be clinically) by noting that an individual is diffusely tender to palpation. The scientific terms for this phenomenon are diffuse hyperalgesia (increased pain to normally painful stimuli) and/or allodynia (pain in response to normally nonpainful stimuli). In the absence of an identifiable diffuse “peripheral” inflammatory process involving the body tissues, this strongly suggests that the CNS (i.e., spinal cord and brain) is causing augmented pain processing. In 1990 when the original criteria for FM were first published, this feature of diffuse tenderness was incorporated into the diagnostic criteria by requiring that an individual had a certain number of tender points (11 or greater) in addition to chronic widespread pain in order to qualify for this diagnosis. Subsequent studies using more sophisticated measures of experimental pain testing showed that individuals with FM are more tender everywhere in the body, not just in the 18 regions considered to be “tender points.” Subsequent experimental pain testing studies have identified multiple potential mechanisms that may be responsible for pain amplification in FM, including a decrease in the activity of descending analgesic pathways, as well as a diffuse increase in the processing of all sensory stimuli (not just pain).

These initial observations that individuals with FM were diffusely tender led to subsequent functional, chemical, and structural brain neuroimaging studies that have been amongst the best “objective” evidence that the pain in FM and related pain amplification syndrome is “real.” These methods such as functional magnetic resonance imaging (fMRI) clearly demonstrate that when individuals with FM are given a mild pressure or heat stimuli that most individuals would feel as “touch” rather than “pain,” they experience pain, and similar brain activation patterns in brain areas involved in pain processing. fMRI has also proved useful in determining how comorbid psychological factors influence pain processing in FM. For example, in FM patients with variable degrees of comorbid depression, the authors found that the anterior insula and amygdala activations were correlated with depressive symptoms, consistent with these “medial” and prefrontal brain regions being involved with affective or motivational aspects of pain processing (and being more closely related to unpleasantness rather than the sensory intensity of pain). However, the degree of neuronal activation in more lateral structures generally thought to be associated with the “sensory” processing of pain (i.e., where the pain is localized and how intense it is) were not associated with levels of depressive symptoms, or the presence or absence of major depression, consistent with a plethora of evidence in the pain field that pain and depression are largely independent but overlapping physiological processes.

A more recent advance in the use of fMRI is to look at the extent brain regions are “connected” to each other, that is, simultaneously activated (or deactivated). The advantage of resting state analysis is that it is a window into brain changes associated with chronic, ongoing spontaneous pain. Studies have shown that individuals with FM have increased connectivity between brain regions involved in increasing pain transmission and neural networks not normally involved in pain such as the default mode network, and the degree of this hyperconnectedness is related to the severity of ongoing pain. A different group has shown that during a painful stimulus, connectivity is decreased between key antinociceptive regions (e.g., the brainstem—the origin of descending analgesic pathways) and a region they had previously identified to be a potential source of dysfunctional pain inhibition in FM.

Other imaging techniques have been used to identify the neurotransmitter abnormalities that may be “driving” the pain amplification seen in FM and other chronic pain disorders. Wood and colleagues used positron emission tomography (PET) to show that attenuated dopaminergic activity may be playing a role in pain transmission in FM, and Harris and colleagues showed evidence of decreased mu opioid receptor availability (possibly due to increased release of endogenous mu opioids) in FM. This latter finding as well as previous studies showing increases in endogenous opioids in the cerebrospinal fluid (CSF) of FM patients has been suggested as evidence of why opioid analgesics appear to not be efficacious in FM.

Other groups have used proton spectroscopy (proton magnetic resonance spectroscopy [H-MRS]) to probe other neurotransmitters. Several groups have shown there are increases in brain concentrations of the body’s major excitatory neurotransmitter, glutamate, in pain processing regions such as the insula in FM. This finding has also been noted in the CSF in FM. Drugs such as pregabalin and gabapentin are likely working in part in FM by reducing glutamatergic activity. This has been nicely demonstrated by Harris and colleagues who showed that individuals with FM that had the highest pretreatment levels of glutamate in the posterior insula were those most likely to respond to pregabalin. When pregabalin led to improvement in symptoms in these individuals, there was normalization of fMRI and connectivity findings, all suggesting that this neurotransmitter is playing a critical role in the pathogenesis of FM in some individuals. An even more important finding from this study was the fact that individuals with FM with normal or low baseline levels of glutamate in their posterior insula did not respond to pregabalin, even though this drug further lowered glutamate levels in these individuals as well. This helps us understand why no single class of CNS analgesic is likely to work in every patient with pain of CNS origin.

Conversely, H-MRS has recently been used to demonstrate low levels of one of the body’s major inhibitory neurotransmitters, GABA. This likely accounts for the efficacy of drugs such as gamma-hydroxybutyrate in FM. This finding may also suggest biological plausibility for the finding that FM patients that have low alcohol consumption (compared to none or high) and have less symptoms and better functionality. Alcohol is known to be a GABA agonist and analgesic, likely showing a U -shaped curve for analgesic effects just as for some beneficial cardiovascular effects. The fact that these imbalances between excitatory and inhibitory neurotransmitters are not diffusely noted in brain structures and seem to be confined to brain regions such as the insula that are known to be involved in polysensory processing is concordant with the notion that there is a global problem with sensory hyperresponsiveness that is partly responsible for the pathophysiology of FM and related conditions.

Although most agree that the core symptoms of FM are likely due to changes in the CNS, peripheral factors also play an important role in both the pathogenesis and treatment of FM. For example, some elements of the processes of central sensitization can be worsened or driven by ongoing nociceptive input. Thus, it is possible or likely that the many individuals with FM that also have comorbid conditions causing ongoing peripheral nociceptive input (e.g., myofascial pain, osteoarthritis, obesity) would potentially benefit from therapies aimed at reducing peripheral drive of central sensitization. This was suggested in a short-term study by Affaitati et al., which showed that treating these common comorbid conditions could lead to improvement in the widespread pain and tenderness seen in FM.

There is also a current ongoing controversy regarding the meaning of finding decreased intraepidermal nerve fiber density (i.e., small fiber neuropathy) in FM. There is no question that this has been shown by many groups, but these authors feel that this is a nonspecific finding seen in many other pain and nonpain conditions, so at present it is not clear whether this plays any role in the pathogenesis of FM or whether doing such testing helps at all in treatment.

Diagnosis of Fibromyalgia

The 1990 American College of Rheumatology criteria for FM were never intended to be used as strict diagnostic criteria for use in clinical practice. Many individuals who clearly have FM do not have pain throughout their entire body, or do not have 11 tender points. Moreover, pain and tenderness occur across a continuum in the population, and it is impossible to know how to distinguish between an individual with symptoms and someone with an “illness.”

For this and many other reasons, the alternative 2010 (which need to be administered by a healthcare provider) and 2011 (which are entirely self-report) FM criteria may represent a preferred manner of diagnosing, or thinking about, FM. The survey version of these criteria are entirely patient self-reported and can be administered on a single piece of paper. There is a body map with 19 areas (each counted as one point towards a total possible score of 31), a symptom survey that asks about the presence and severity of fatigue, sleep disturbances, memory difficulties (each scored 0–3 for the presence and severity), and irritable bowel, headaches and mood problems (one point each). When used as a dichotomous measure with a variety of “cut points” that can be used, this measure roughly identifies most of the same individuals as the old criteria (except many more males). However, when FM is considered more as a physiological construct to determine the location of an individual on the continuum of pain sensitivity, then this can be used as a continuous measure (i.e., degree of “fibromyalgia-ness” or the degree to which pain is centralized) that can be useful in the diagnosis and treatment of virtually any patient with a rheumatic disorder who is experiencing pain.

In several recent studies, this concept of fibromyalgianess or subsyndromal FM has been shown to be very clinically important. In these studies by Brummett and colleagues, individuals who were scheduled for either lower extremity joint replacement or hysterectomy completed a broad self-report battery prior to their surgery. The group hypothesized that individuals with higher FM scores on the 2011 FM Survey Criteria would predict decreased responsiveness to opioids given in the perioperative period and to surgery, ultimately relieving pain. Again, this measure is scored from 0 to 31, with a level of 13 being typically used as the diagnostic cut point of FM. These studies demonstrated that for each one-point increase in this measure from 0 to 31, individuals needed an adjusted 7–9 mg more oral morphine equivalents to control their pain in the first 24–48 hours following surgery and were 15%–20% less likely to improve following arthroplasty. These findings were independent of a number of preoperative characteristics, including age, sex, anxiety, depression, catastrophizing, and opioid use. More importantly, these findings were linear and the same incremental increase in opioid and surgery nonresponsiveness was seen in individuals well below the threshold used to diagnose FM and extending well into the range of individuals exceeding this threshold. Fig. 26.1 shows two different individuals with osteoarthritis, but neither of whom meet the criteria for FM and are at different portions of the FM continuum; the figure also shows the marked difference in opioid responsiveness and improvement in pain following arthroplasty that these two individuals experience. These data suggest that this measure of fibromyalgianess might serve as the degree of pain centralization that an individual is experiencing and help identify individuals in perioperative or other settings who are less likely to respond to peripherally directed analgesics such as surgery or opioids.

Distribution of fibromyalgianess. After adjustment, compared to Patient A with localized pain and no somatic symptoms, Patient B would need 80 mg more oral morphine equivalents (OME) during first 48 hours of hospitalization and would be 5 times less likely to have improvements in pain at 6 months.

Brummett CM, Urquhart AG, Hassett AL, et al: Characteristics of fibromyalgia independently predict poorer long-term analgesic outcomes following total knee and hip arthroplasty. Arthritis Rheumatol. 67(5):1386-1394, 2015; Brummett CM, Janda AM, Schueller CM, et al. Survey criteria for fibromyalgia independently predict increased postoperative opioid consumption after lower-extremity joint arthroplasty: a prospective, observational cohort study. Anesthesiology. 119(6):1434-1443, 2013.

In clinical practice, the clinician should suspect FM in individuals with multifocal pain that cannot be fully explained based on damage or inflammation in those regions of the body. In most cases, musculoskeletal pain is the most prominent feature, but because pain pathways throughout the body are amplified, pain can occur anywhere. Thus, chronic headaches, sore throats, and visceral pain and sensory symptoms are very common in individuals with FM. As previously noted, it is especially helpful to take a lifetime history of chronic pain, since a patient new to a given clinician often arrive with their most recent location of pain. If the clinician recognizes that this same individual has had chronic pain most of their lives in different bodily regions, they (hopefully) are less likely to perform surgery or injections that will solve this individual’s problems.

Because pain is a defining feature of FM, it is helpful to focus on the features of the pain that can help distinguish it from other disorders. The pain of FM is typically diffuse or multifocal, often waxes and wanes, and is frequently migratory in nature. These characteristics of “centralized pain” are quite different from “peripheral” pain, where both the location and triggers of pain are typically more constant and predictable. Patients with FM or pain from neural origin may complain of discomfort when they are touched or when wearing tight clothing, and they may experience dysesthesias or paresthesias that accompany the pain.

Another defining characteristic of “centralized” pain is the company it keeps. In addition to pain, individuals typically experience a number of other somatic symptoms. These symptoms can generally be broken into two categories: (1) symptoms of CNS origin that are controlled by the same neurotransmitters that control pain processing and (2) symptoms due to generalized sensory hyperresponsiveness. In the first category, fatigue, memory difficulties, and sleep and mood disturbances are all very common in FM and other centralized pain states. These symptoms are at least partially due to some of the same neurotransmitter abnormalities that contribute to the pathophysiology of as FM, because several of these symptoms typically improve along with pain when an individual is successfully treated with drugs that alter these neurotransmitters. The second type of symptom due to generalized hyperresponsiveness often challenges clinicians to doubt the veracity of an FM patient’s complaints. This is often responsible for the “pan-positive review of symptoms” that has often characterized these individuals as “somatizers.” The biology of somatization is becoming increasingly understood as a biology of sensory hyperresponsiveness. These sensory symptoms throughout the body can be better understood by clinicians if they realize that all sensory experiences are interpreted by brain regions such as the insula, which are known to be hyperactive in FM and other centralized pain states.

Although the physical examination is generally unremarkable in individuals with FM, it is helpful to assess for diffuse tenderness, and there are many ways that this can be done clinically other than performing a tender point count. For example, individuals with FM are more sensitive to the inflation of a blood pressure cuff. Another way to assess overall pain threshold while also getting other valuable diagnostic information is to assess pain thresholds in the hands and arms of all chronic pain patients. A rapid examination is performed by applying firm pressure over several interphalangeal (IP) joints of each hand and over the adjacent phalanges and then more proximally to include firm palpation of the muscles of the forearm including the lateral epicondyle region. This is one way to assess overall pain threshold and obtain additional diagnostic information about the patient. If the individual is tender in many of these areas, or in just the muscles of the forearm, then they are likely diffusely tender (i.e., have a low central pain threshold). However, if the individual is only tender over the IP joints and not the other regions, and especially if there is any swelling over these joints, one should be more concerned about a systemic autoimmune disorder, whereas if tenderness is confined to just the bones, then this is a symptom of metabolic bone disease or condition causing periostitis (e.g., hyperparathyroidism).

Laboratory testing is generally not useful, except for the purpose of differential diagnosis. The intensity of the diagnostic work-up can be partially guided by the length of time the patient has had symptoms. If the patient’s symptoms have persisted for several years, minimal testing is required, whereas a more aggressive strategy should be employed for acute or subacute onset of symptoms. Simple testing should be limited to complete blood count and routine serum chemistries, along with thyroid-stimulating hormone (TSH) and erythrocyte sedimentation rate (ESR) and/or C-reactive protein. Serologic studies such as antinuclear antibody (ANA) and rheumatoid factor assays should generally be avoided unless there are historical features not seen in FM or abnormalities on physical examination.

Once a clinician excludes other potential disorders, an important and at times controversial step in the management of FM is asserting the diagnosis. Despite some assumptions that being “labeled” with FM may adversely affect patients, all existing studies suggest that this is not the case and that the diagnosis of FM is often a tremendous source of relief for the patient and leads to decreased healthcare utilization because of a reduction in referrals and diagnostic testing, or “looking for the cause of the pain.”

General Treatment Approach

All individuals diagnosed with a condition such as FM should first receive some basic education regarding this disorder. Clinicians can do this in the context of their practice (e.g., with nurse educators or other allied health professionals) or electronically using websites and videos (see below). This education should emphasize that it is important that the patient plays an active role in their management and that some of the most effective therapies are nondrug therapies such as exercise, improving sleep, and reducing stress. Most patients need some combination of pharmacological and nonpharmacological therapies to have meaningful improvement in symptoms and function. Table 26.1 outlines all the evidence-based treatments for individuals with FM.

TABLE 26.1

Evidence-Based Treatments for Fibromyalgia

From Clauw DJ: Fibromyalgia: a clinical review. J Am Med Assoc . 311(15):1547-1555, 2014.

Treatment	Cost	Specifics	Evidence Level	Side Effects	Clinical Pearls
General Recommendations
Patient education	Low	Incorporate principles of self-management including a multimodal approach	1, A		Following initial diagnosis spend several visits (or use separate educational sessions) to explain the condition and set treatment expectations.
Nonpharmacological Therapies
Graded exercise	Low	Aerobic exercise has been best studied but strengthening and stretching have also been shown to be of value	1, A	Worsening of symptoms when program is begun too rapidly	• Counsel patients to “start low, go slow.” • For many patients focusing first on increasing daily “activity” is more helpful before starting exercise
Cognitive-behavioral therapy (CBT)	Low	Pain-based CBT programs have been shown to be effective in one-on-one settings, small groups, and via internet	1, A	No significant side effects of CBT per se but patient acceptance is often poor when they view this as a “psychological” intervention	• Internet-based programs are gaining acceptance and are more convenient for working patients
Complementary and alternative medicine (CAM) therapies	Variable	Most CAM therapies have not been rigorously studied	1, A	Generally safe	• There is emerging evidence that CAM treatments such as Tai Chi, yoga, balneotherapy, and acupuncture might be effective • Allowing patients to choose which CAM therapies to incorporate into an active treatment program can increase self-efficacy
CNS neurostimulatory therapies		Several different types of CNS neurostimulatory therapies have been shown to be effective in FM and other chronic pain states		Headache	These treatments continue to be refined as we learn about optimal stimulation targets and “dosing.”
Pharmacological therapies		Pharmacological therapy is best chosen based on the predominant symptoms, and initiated in low dose with slow dose escalation	Level 5, Consensus		• Some practitioners find that getting patients on a drug regimen that helps improve symptoms prior to initiating nonpharmacological therapies can help improve compliance
Tricyclic compounds		• Nortriptyline 10–70 mg qhs • Cyclobenzaprine 5–20 mg qhs	1, A	Dry mouth, weight gain, constipation, “groggy” or drugged feeling	• When effective, they can improve a wide range of symptoms including pain, sleep, bowel, and bladder symptoms • Taking these drugs several hours prior to bedtime improves side effect profile
Serotonin norepinephrine reuptake inhibitors	Doluxetine is generic, milnacipran not	• Duloxetine, 30–120 mg daily • Milnacipran, 100–200 mg daily	1, A	Nausea, palpitations, headache, fatigue, tachycardia, hypertension	• Warning patients about transient nausea, taking with food, and slowly increasing dose can increase tolerability • Milnacipran might be slightly more noradrenergic than duloxetine and thus potentially more helpful for fatigue and memory problems, but it is also more likely to cause HTN
Gabapentinoids	Gabapentin is generic, pregabalin not	• Gabapentin, 800–2400 mg/day in divided doses • Pregabalin up to 600 mg/day in divided doses	1, A	Sedation, weight gain, dizziness	• Giving most or all the dose at bedtime can increase tolerability
Gamma-hydroxybutyrate	Available for treating narcolepsy, cataplexy	GHB 4.5–6.0 g per night in divided doses	1, A	Sedation, respiratory depression and death	• Shown to be efficacious but not approved by US FDA because of safety concerns
Low-dose naltrexone	Low	4.5 mg/day	2 small single-center RCTs
Cannabinoids	NA	Nabilone 0.5 mg po qhs–1.0 mg bid	1, A	Sedation, dizziness dry mouth	• No synthetic cannabinoid is approved in US for treatment of pain
Tizanidine	Low	Tizanidine 4–8 mg every 6–6 h	One small RCT	Sedation, dizziness	• No long-term safety data
Esreboxetine	NA	Esreboxetine 4 or 8 mg/day	Two RCTs	Hypertension, tachycardia, urinary retention	• Not available in US
Selective serotonin reuptake inhibitors (SSRIs)	SSRIs that should be used in FM (see Pearls) are all generic	Fluoxetine, sertraline, paroxetine	1, A	Nausea, sexual dysfunction, weight gain, sleep disturbance	• Older, less selective SSRIs may have some efficacy in improving pain, especially at higher doses that have more prominent noradrenergic effects • Newer SSRIs (citalopram, escitalopran, desvenlafaxine) are less effective or ineffective as analgesics
NSAIDs		• No evidence of efficacy • Can be helpful to treat comorbid “peripheral pain generators”	5, D	Gastrointestinal, renal, and cardiac side effects	• Use the lowest dose for the shortest period of time to reduce side effects
Opioids		• Tramadol with or without acetaminophen, 50–100 mg every 6 h • No evidence of efficacy for stronger opioids	5, D	Sedation, addiction, tolerance, opioid-induced hyperalgesia	• There is increasing evidence that opioids are less effective for treating chronic pain than previously thought and their risk:benefit profile is worse than other classes of analgesics

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

Determination of Disability Epidural Opioids for Postoperative Pain Myofascial Pain Syndrome Orofacial Pain Pain Management During Pregnancy and Lactation Membrane Stabilizers

Stay updated, free articles. Join our Telegram channel

Join