Why fibromyalgia hurts when nothing shows on tests: a clinical account of central sensitization

What fibromyalgia is, and what it is not.
The diagnosis-of-exclusion problem.

Fibromyalgia is a chronic pain syndrome characterised by widespread musculoskeletal pain lasting longer than three months, accompanied by fatigue, sleep disturbance, cognitive symptoms commonly called “fibro fog”, and frequent comorbidity with mood disorders. Prevalence in the general population is 2-4%, with a female-to-male ratio of about 8 to 1, and the average diagnostic delay from first symptoms to confirmed diagnosis is six to seven years (Häuser et al., 2015).

The diagnostic label most often attached to it is “diagnosis of exclusion”. That phrasing is accurate in a procedural sense: standard inflammatory markers are normal, joint imaging is clean, autoimmune panels are negative, neurological examination unremarkable. The clinician concludes there is no detectable pathology of the conventional kind, and the diagnosis of fibromyalgia is assigned by default.

The phrasing also creates a problem. “Exclusion” is heard by patients as “absence”, and “absence” sounds like “imagined”. The interpretive gap is wide enough that many fibromyalgia patients spend years cycling through specialists, accumulating clean test results, and developing a parallel suspicion that their pain might be psychogenic in the dismissive sense of that word. The ACR 2016 criteria (widespread pain index combined with a symptom severity scale) reduced some of this drift, but the framing problem persists because the underlying signature of fibromyalgia is not the kind of signature standard pathology looks for. The disease is real and measurable. It is measurable in a different layer.

The diagnostic criteria themselves have moved over three decades, which partly explains the six-to-seven-year delay. The 1990 ACR criteria required a clinical examination locating eleven of eighteen specific tender points, which made diagnosis dependent on a single physical test that many primary care physicians never learned. The 2010 revision shifted to a self-report instrument and dropped the tender-point requirement, which broadened recognition but also widened the diagnostic envelope. The 2016 update tightened the criteria again and removed the requirement to rule out other disorders that might better account for the pain, formally acknowledging that fibromyalgia can coexist with other conditions rather than only fill the diagnostic void left when other conditions are absent. The shift from exclusion to coexistence is recent enough that many clinicians still operate on the older logic, which adds years to the typical diagnostic path.

The pain is real because the signal is real.
Central sensitization explained.

That layer is the central nervous system, and the specific phenomenon is called central sensitization. The mechanism is now well-described in the pain neuroscience literature (Sluka & Clauw, 2016). When nociceptive input from peripheral tissue arrives at the dorsal horn of the spinal cord, NMDA receptor activity is upregulated, producing long-term potentiation in the pain pathway. The system, in effect, learns to feel pain. Stimuli that previously sat below the pain threshold begin to cross it. A repeated stimulus produces an escalating response rather than a stable one, which is the phenomenon clinicians call “wind-up”.

Functional MRI confirms this at the level of brain activation. In fibromyalgia patients, sub-threshold mechanical or thermal stimuli activate the insula, primary somatosensory cortex, and anterior cingulate cortex at intensities equivalent to suprathreshold stimuli in controls (Sluka & Clauw, 2016). The brain is not generating phantom signals. It is responding accurately to signals that have been amplified upstream.

This distinction matters because it changes the whole frame of the conversation. The fibromyalgia patient who reports that a light touch is painful is not exaggerating. The light touch is, by the time it reaches perception, no longer a light touch.

What standard medicine struggles with is that “louder system” is not a category most diagnostic pathways accommodate. There is no blood test for gain. There is no scan that quantifies sensitization at clinical resolution. The pathology is functional, distributed, and architectural, not anatomical and localised.

Why analgesics keep failing.
Treating the message, not the volume.

Standard analgesics are built for a different target. NSAIDs reduce prostaglandin-mediated inflammation, which is the upstream driver of pain in arthritic, post-injury, and infectious contexts. In fibromyalgia there is no inflammation to reduce. The NSAID lowers nothing relevant, and the patient experiences either partial placebo response or none at all. Opioids work briefly because they dampen central transmission of everything, including the amplified signal, but tolerance builds quickly and the longer-term effect of opioids on chronic pain is often paradoxical, with opioid-induced hyperalgesia adding to the original sensitization.

Gabapentinoids such as pregabalin, and serotonin-norepinephrine reuptake inhibitors such as duloxetine and milnacipran, do operate centrally. They reduce gain modestly by lowering neuronal excitability and modulating descending inhibition. These are the medications with the best evidence base for fibromyalgia, and they help some patients meaningfully. They do not, however, change the underlying gain setting. They cap the output of a system that is still set high.

This is the architecture of the years-long medication trial that many fibromyalgia patients live through. Drug after drug is added, doses titrated, side effects accumulated, and the pain reduces partially or temporarily and then returns. The patient is not failing to respond to treatment. The treatment is addressing the message rather than the volume.

The trauma-informed approach.
What changes the volume.

Two parallel findings from the last two decades reframe the question of where the volume comes from. The first is the high overlap between fibromyalgia and a history of complex or childhood trauma. In a study of fibromyalgia patients seen in rheumatology clinics, 56% met criteria for post-traumatic stress disorder, compared with much lower rates in matched controls (Cohen et al., 2002), and subsequent studies have replicated and extended this finding across different populations. The second is the consistent observation that HPA-axis regulation in fibromyalgia resembles the pattern seen in chronic stress and PTSD: blunted morning cortisol, flattened diurnal curve, and altered cortisol response to stressors (McEwen, 2007).

These two findings, taken together, suggest a structural reading of central sensitization. The nervous system that has spent years predicting threat learns to keep the gain dial elevated as a default state. The dial does not lower spontaneously when external circumstances change, because the prediction model itself has not changed. The body keeps preparing for the input it learned to expect, and chronic pain is one of the forms that preparation takes when no acute threat is present to direct it.

The clinical implication is that work which addresses the prediction model directly often reduces fibromyalgia pain without targeting pain as such. EMDR, IFS, and structured trauma-reprocessing protocols are not analgesic interventions in the pharmacological sense, but they alter the upstream condition that keeps sensitization in place. In a small clinical-phenomenological study (Laugman, 2026), clients presenting with comorbid PTSD and fibromyalgia reported reductions in pain intensity of 40 to 60% over the course of trauma work, with the pain reduction tracking alongside reductions in PTSD symptoms rather than ahead of them.

Sequencing matters. Trauma reprocessing without prior stabilisation often increases central sensitization in the short term, because reactivating threat networks pushes the gain dial up before it comes down. The same approach with a stabilisation phase in front of it (resourcing, somatic anchoring, careful titration of activation) produces the opposite trajectory, with the gain dial settling lower than baseline as the prediction model updates. The order is not optional. Trauma work that skips the stabilisation phase in a fibromyalgia presentation tends to fail in a particular way: the client reports feeling worse for weeks, attributes it to the therapy, and withdraws. The intervention was not wrong. It was sequenced wrong.