POEMS syndrome is one of medicine’s stranger plasma cell disorders. A tiny, often barely-detectable clone can wreak havoc across the nervous, vascular, endocrine, pulmonary, and hematologic systems, producing a disease burden far out of proportion to its limited presence in the bone marrow.
That mismatch is exactly why it gets missed. A patient shows up with neuropathy and gets worked up for CIDP. Someone ends up in endocrinology for an unexplained hormonal issue, or in pulmonology for pleural effusions, long before anyone connects the dots. Getting the diagnosis right matters because the actual target is the plasma cell clone.
What the Acronym Doesn’t Tell
POEMS, polyneuropathy, organomegaly, endocrinopathy, monoclonal plasma cell disorder, and skin changes, is a helpful but incomplete acronym. Patients don’t need every letter represented, and some of the most useful diagnostic clues (elevated VEGF, sclerotic bone lesions, papilledema, thrombocytosis, volume overload) aren’t in the name at all.
The underlying clone is almost always lambda-restricted, and the disease behaves nothing like conventional multiple myeloma: minimal marrow plasmacytosis, no destructive lytic lesions, little cast nephropathy, and often only a whisper of monoclonal protein. Sclerotic bone lesions show up instead of lytic ones. None of this correlates well with severity.
The Role of VEGF in POEMS Syndrome
Vascular endothelial growth factor is the biomarker most tied to disease activity, and it explains many of the clinical features. Increased vascular permeability and angiogenesis account for the edema, ascites, effusions, papilledema, and cardiopulmonary complications. But it’s clearly not the whole mechanism. Directly blocking VEGF hasn’t worked well as therapy, while hitting the plasma cell clone itself does, a strong hint that VEGF is a downstream signal, not the disease’s engine.
Interleukin-6 seems to matter more when POEMS overlaps with Castleman disease. There’s also evidence of restricted immunoglobulin light-chain variable gene usage and assorted chromosomal abnormalities, but nothing here has matured into something that changes treatment or risk stratification yet.
Making the Diagnosis
There’s no single confirmatory test, diagnosis rests on recognizing a pattern. Both major criteria are mandatory:
- Polyneuropathy, typically demyelinating
- A clonal plasma cell disorder, almost always lambda-restricted
At least one of:
- Elevated VEGF
- Sclerotic bone lesions
- Castleman disease
Plus at least one minor criterion:
- Organomegaly
- Extravascular volume overload (edema, ascites, pleural effusion)
- Endocrinopathy
- Characteristic skin changes
- Papilledema
- Thrombocytosis or erythrocytosis
The real matter is knowing when to even look. Any acquired demyelinating neuropathy paired with a lambda monoclonal protein, thrombocytosis, papilledema, or unexplained fluid overload should trigger suspicion.
Useful cutoffs: plasma VEGF above 200 pg/mL runs about 95% specific, and serum VEGF above 1920 pg/mL about 98% specific (serum values run 10-50x higher than plasma, mostly from platelet release during clotting, which is why the two aren’t interchangeable). A newer marker, N-terminal propeptide of type I collagen, has been proposed at a cutoff of 70 ng/mL, though it isn’t yet standard.
The CIDP Trap
Neuropathy usually dominates the presentation, and it can look a lot like CIDP, progressive weakness, sensory loss, demyelinating changes on nerve conduction studies. There are some concrete discriminators worth knowing rather than just a general sense of overlap. A high platelet count shows up in roughly 54% of POEMS patients vs 1.5% of CIDP patients.
Nerve conduction studies in POEMS tend to show more slowing in intermediate nerve segments rather than distal ones, and conduction block, common in CIDP, is rare. Nerve biopsy in POEMS typically lacks the large onion bulbs characteristic of CIDP, and shows less inflammation overall, with what inflammation there is located in the epineurium rather than spreading into the endoneurium.
A 2026 cost-effectiveness analysis found that checking VEGF at initial workup, rather than waiting for a trial of IVIG to fail, cut both misdiagnosis rates and downstream healthcare spending. IVIG and plasma exchange rarely give durable benefit in POEMS. If a patient gets a partial or temporary neurologic response to IVIG but other findings remain unexplained, that’s not a reason to rule POEMS out.
Why the Clone Hides
Clonal burden tends to be low. Serum protein electrophoresis may show only a faint M-spike, and free light chains can be elevated with a normal kappa/lambda ratio, both of which make standard myeloma screening insensitive here. A routine iliac crest biopsy can come back clean simply because the clone is sitting inside a sclerotic lesion somewhere else in the skeleton.
Workup should combine serum and urine immunofixation, marrow assessment, VEGF measurement, and imaging capable of catching sclerotic or metabolically active lesions, with targeted biopsy of a suspicious lesion when the marrow itself is uninformative.
On the marrow biopsy itself, immunohistochemistry outperforms standard flow cytometry, because it preserves the architecture, plasma cells often “rim” lymphoid aggregates in a pattern that’s essentially specific to POEMS (it shows up almost nowhere else outside lymphoplasmacytic lymphoma). Only about 1 in 8 biopsies comes back with no clonal, architectural, or megakaryocytic clue at all.
Because this is a multisystem disease, the full evaluation extends well past hematology, to establish baseline organ involvement. On the endocrine side, hypogonadism is the most frequently affected axis, followed by thyroid and glucose abnormalities, with adrenal insufficiency last but most dangerous to miss, subclinical adrenal insufficiency in particular, since an unrecognized case can tip into adrenal crisis.
Thrombotic Risk in POEMS Syndrome
Thrombosis is another risk worth building into the initial workup. Close to 30% of patients have an arterial or venous thrombotic event over their disease course, with arterial events outnumbering venous roughly two to one before treatment starts.
Seven to ten percent present with a cerebrovascular event, on average about two years after neuropathy onset, and the risk factors, elevated hemoglobin/hematocrit, volume overload, splenomegaly, thrombocytosis, marrow plasmacytosis, overlap heavily with the disease’s own features.
In practice this usually means at least aspirin prophylaxis, with anticoagulation considered case by case against fall risk. It’s also worth noting lenalidomide has been linked to ischemic cerebrovascular events in this population, which is why some clinicians get a baseline head MRA before starting it.
The Relationship With Castleman Disease
Castleman disease, or Castleman-like lymph node changes, shows up in a subset of POEMS patients, which muddies the picture, lymph node histology alone doesn’t establish classic POEMS. If a patient lacks polyneuropathy or a clonal plasma cell disorder, they’re better classified as a Castleman disease variant. The distinction isn’t academic: cytokine profile, clinical course, and treatment approach can all differ between the two.
Risk Assessment: What Actually Predicts Outcome
There’s no validated molecular or genetic risk marker, so prognosis comes down to clinical phenotype and organ dysfunction. Older age, low albumin, pleural effusion, pulmonary hypertension, severe renal impairment, respiratory involvement, and marked volume overload all track with worse outcomes.
Coexisting Castleman disease appears to shorten both progression-free and overall survival, and elevated IL-6, seen in around 40% of patients, marks a particularly steep drop-off, with 8-year OS of 64% vs 97% in those with normal levels.
What doesn’t predict outcome: how many POEMS features a patient checks off. Patients who reach a complete hematologic response have substantially longer progression-free survival than those who don’t.
You can also read about
Radiation for Localized Disease
Radiation is first-line when there’s an isolated plasmacytoma with no clonal plasma cells on iliac crest biopsy, and it can still work with two or three small, localized lesions absent broader marrow involvement. Response is slow and follows a predictable order: edema, papilledema, and skin changes tend to improve first, while neuropathy can keep recovering for years.
In a Mayo Clinic series of 35 radiation-treated patients, four-year OS was 97%, though four-year failure-free survival was only 52%. Longer follow-up in an expanded cohort showed six-year PFS of 62% and 10-year OS of 70%. A UK retrospective study found a bigger gap than “might help” suggests: three-year PFS was 100% in patients who got short-course systemic therapy alongside radiation for a solitary plasmacytoma, vs 65% with radiation alone.
Systemic Therapy for Disseminated Disease
Systemic therapy is the call when clonal plasma cells turn up in the marrow, sclerotic lesions are diffuse, or multiple skeletal lesions are present, regardless of how low the plasma cell percentage actually is. Because randomized trials in POEMS are essentially nonexistent, most regimens are borrowed from myeloma and light-chain amyloidosis. Alkylating agents, lenalidomide, and autologous stem cell transplant have the most clinical mileage behind them. Steroids alone can help temporarily but aren’t a real treatment.
Lenalidomide plus dexamethasone is a well-established option, particularly outside the transplant-eligible population. In a prospective study of 41 newly diagnosed patients treated for 12 cycles, 43% achieved a complete VEGF response, 46% a complete hematologic response, and 59% saw neurologic improvement, with three-year PFS and OS of 75% and 90%.
Melphalan plus dexamethasone also performs well: in a 31-patient study, 81% achieved a hematologic response, every evaluable patient had a VEGF response, and all patients saw at least some neurologic improvement. Follow-up was limited, and prolonged melphalan exposure can compromise later stem cell collection, worth weighing if transplant is still considered.
Autologous Stem Cell Transplant
Transplant can offer durable control in eligible patients, though the supporting data skews toward younger, fitter cohorts, which likely inflates how good the results look. Patients with severe fluid overload, pulmonary hypertension, respiratory compromise, or poor functional status carry more peri-transplant risk, a short induction course can start controlling the clone, and may lower complications.
Mobilization typically uses G-CSF, with plerixafor added as needed. Engraftment syndrome should be closely monitored, as it appears to be more severe in POEMS than after standard myeloma transplantation, and careful management of volume status and cardiopulmonary function is essential through this window.
Proteasome Inhibitors and Monoclonal Antibodies
Bortezomib has real activity against the clone, but its neuropathy risk has historically kept it on the sidelines here. Weekly, subcutaneous dosing may soften that risk, but it still needs to be tailored to the baseline neuropathy.
The numbers back up the interest: in a 69-patient series using weekly subcutaneous bortezomib with dexamethasone, 46% reached a complete hematologic response, 44% a complete VEGF response, 84% had a clinical response overall, and two-year OS was 96%, with only two patients showing bortezomib-related neuropathy worsening, which improved after stopping the drug.
Ixazomib, an oral proteasome inhibitor, looks promising in early data: in a study of 12 newly diagnosed patients on ixazomib, cyclophosphamide, and dexamethasone, all seven evaluable patients hit a complete VEGF response and 83% improved neurologically. The cohort is too small for firm conclusions, but it makes a case for further work on lower-neurotoxicity proteasome inhibitor regimens.
Daratumumab-based combinations are showing encouraging results in case series, hematologic, VEGF, and clinical responses all included, and use is growing, especially in relapsed disease or when standard options don’t fit. Prospective data is still thin. Anti-BCMA therapy has appeared in scattered case reports, but its role isn’t defined yet.
Supportive Care Shouldn’t Be an Afterthought
Physical therapy and ankle-foot orthoses help with mobility, fall prevention, and avoiding fixed contractures while nerves slowly recover. Gabapentin, pregabalin, or tricyclics can manage neuropathic pain in the meantime.
Respiratory muscle weakness raises real risks, hypoventilation, infection, worsening pulmonary hypertension, and some patients need CPAP or BiPAP. Depression is worth screening for too: prolonged disability and loss of independence take a toll that goes beyond the disease itself.
Why Response Assessment Doesn’t Follow Myeloma Rules
Standard myeloma response criteria fall apart here because the M-protein is often small and marrow involvement minimal to begin with. Free light chains are elevated in many patients, but the ratio usually stays normal, which limits its use as a standalone response marker.
Some patients improve substantially with no measurable drop in monoclonal protein at all.
VEGF can start falling within three months of treatment, but the absolute number doesn’t always track with how the patient is actually doing, trends matter more than any single value.
Recovery timelines vary wildly by organ system. Edema, effusions, papilledema, and skin changes tend to respond early. Neurologic improvement usually isn’t obvious until around six months, and full nerve recovery can take two to three years. PET/CT abnormalities can linger for 6-12 months even when treatment is working.
Follow-up runs roughly every three months, benchmarked against a solid pretreatment baseline and weighted toward whatever caused the most morbidity in that particular patient.
Survival and Causes of Death in POEMS Syndrome
Ten-year overall survival now approaches 80% in contemporary series, largely thanks to earlier recognition and better access to plasma-cell-directed therapy. Patients who reach a complete hematologic response do especially well, five-year PFS of 88%, vs 50%.
Mortality also looks different from myeloma. In a Mayo Clinic analysis spanning 2000-2022, no deaths were attributed to overwhelming plasmacytosis or cast nephropathy. Of the deaths with a documented cause, about 45% were classified as POEMS-related, 22% were second malignancies (roughly half MDS/AML), 10% were treatment complications, and the remainder were unrelated to the diagnosis.
Within the POEMS-related group, the dominant pattern was progressive capillary leak, persistent ascites, effusions, or anasarca eventually driving hypotension, renal failure, and cardiopulmonary collapse. A subset of these patients died within months of diagnosis, the rest developed this pattern more than five years in, often with elevated IL-6 in that terminal window.
Clinical Takeaways
Think POEMS when demyelinating neuropathy shows up alongside a lambda plasma cell clone, sclerotic bone lesions, elevated VEGF, thrombocytosis, papilledema, or fluid overload that doesn’t otherwise add up. Localized disease can often be handled with radiation alone, disseminated disease needs systemic therapy and, for eligible patients, transplant. Response should be tracked across hematologic, VEGF, imaging, and organ-specific measures, recognizing that nerves recover on their own clock, sometimes taking years to show the full benefit of treatment.
VEXAS Syndrome: Is It More a Matter of Inflammation or Hematopoietic Clonality?
FAQ
Is POEMS syndrome a form of cancer?
POEMS syndrome arises from a clonal plasma cell disorder, but it behaves differently from conventional MM. The plasma cell burden is usually small, while cytokine-driven vascular and multisystem effects account for much of the illness.
Can POEMS syndrome occur with a normal bone marrow biopsy?
Yes. A routine iliac crest biopsy may miss the clone when plasma cells are concentrated within a sclerotic bone lesion. If clinical suspicion remains high, imaging and targeted biopsy of a suspicious lesion may help establish the diagnosis.
Is an elevated VEGF level enough to diagnose POEMS syndrome?
No. VEGF supports the diagnosis but is not specific enough to be used alone. Diagnosis still requires demyelinating polyneuropathy, a clonal plasma cell disorder, at least one additional major criterion, and at least one minor criterion.
Can POEMS neuropathy improve after treatment?
Yes, although recovery is slow. Neurologic improvement may not become evident for approximately six months, and maximal recovery can take two to three years. Persistent neuropathy shortly after treatment does not necessarily indicate treatment failure.
Can localized POEMS syndrome be cured with radiation?
Radiation can provide long-term control and may be curative when the plasma cell disorder is confined to an isolated plasmacytoma without disseminated marrow involvement. Improvement often develops gradually, with neuropathy recovering later than edema, papilledema, or skin changes.
Why is thrombosis common in POEMS syndrome?
Thrombotic risk appears to reflect several overlapping features, including thrombocytosis, erythrocytosis, splenomegaly, volume overload, and plasma cell activity. Both arterial and venous events occur, although arterial events appear to be more common before treatment.
Can POEMS syndrome relapse after successful treatment?
Yes. Relapse may appear through rising VEGF, recurrent fluid overload, worsening neurologic or pulmonary function, increasing monoclonal protein, or new active bone lesions. Because these markers may change at different times, follow-up should assess the disease across several clinical and laboratory domains.