For years, academic neuromuscular neurologists have anecdotally observed the overdiagnosis of chronic inflammatory demyelinating polyneuropathy (CIDP) in referred patients – and in the past 5 years, experts have documented the phenomenon and attempted to pinpoint diagnostic pitfalls.
Among the problems they’ve uncovered: A failure to adhere to diagnostic criteria and to focus on symptoms and signs of the disease, interpretation of electrodiagnostic studies that is too liberal or too loose, and an overemphasis on minimally elevated cerebrospinal fluid (CSF) protein levels.
In a 2015 study reported by Jeffrey A. Allen, MD, almost half – 47% – of 59 consecutive patients referred to Northwestern University, Chicago, with a diagnosis of CIDP failed to meet the minimal CIDP diagnostic requirements of the European Federation of Neurological Societies/Peripheral Nerve Society (EFNS/PNS). After reevaluation, these patients were diagnosed with a range of other conditions, including diabetic polyneuropathy, hereditary neuropathy, multifactorial neuropathy, and multifocal motor neuropathy, as well as nonneuropathic disorders such as amyotrophic lateral sclerosis and fibromyalgia (Neurology. 2015;85:498-504).
“The pitfalls in diagnosis go both ways – the missed diagnosis of CIDP, and the diagnosis of CIDP in people who don’t really have CIDP,” said Michael Cartwright, MD, professor of neurology at Wake Forest University, Winston-Salem, N.C. “Overdiagnosis is the more major issue [however]. Physicians and neurologists want to find a neuropathy that can be treated and improved, so there is a lot of hope that we’ll find something like CIDP, which is potentially treatable.”
“These cases are just very complicated though,” he said. “It can be very difficult to figure out if they truly are an immune-mediated neuropathy.”
Michael Cartwright, MD
Kenneth C. Gorson, MD, professor of neurology at Tufts University, Boston, has seen both in his practice and through his research how misdiagnoses can lead to long durations of treatment with intravenous immunoglobulin (IVIg) or lengthy corticosteroid protocols without any evidence of therapeutic benefit. In an editorial published in conjunction with Dr. Allen’s 2015 study, Dr. Gorson noted that he had seen misdiagnosed patients treated with IVIg for up to 12 years (Neurology. 2015;85:488-9).
In a study published several years ago with Dr. Allen and coinvestigator Deborah Galinas, MD, Dr. Gorson reviewed IVIg utilization data from a national specialty pharmacy database and found that only 11% of 65 consecutive patients treated with IVIg met clinical and electrodiagnostic CIDP criteria (Brain Behav. 2018;8:e00932). The patients were being treated by 31 different community neurologists in 14 states, so misdiagnosis “isn’t just a single-center finding or referral bias situation,” Dr. Gorson said.
Experts hope that future revisions to the EFNS/PNS guidelines will achieve more concise clinical definitions, particularly of atypical forms, and more stringent requirements for electrophysiologic testing. Research is also underway to identity biomarkers. In the meantime, however, there are actions neurologists can take to improve diagnostic accuracy and avoid a misdiagnosis of CIDP.
Here are some key tips offered by Dr. Gorson and Dr. Cartwright:
Heed clinical findings – they take precedence. The EFNS/PNS guidelines – the guidelines that are preferred by CIDP experts and are believed to be the most useful set for both research and clinical practice – define clinical and electrophysiological diagnostic criteria as well as concomitant laboratory, imaging, and other investigations to consider (Eur J Neurol. 2010 Mar;17[3]:356-63). In general, all published sets of diagnostic criteria for CIDP similarly base diagnosis on “a combination of clinical, electrodiagnostic, and laboratory features with exclusions to eliminate other disorders that may appear as CIDP,” the ENFS/PNS noted in its guideline report.
Fulfilling both clinical and electrodiagnostic criteria is important, but clinical findings take priority, Dr. Gorson and Dr. Cartwright emphasized. “It’s important to combine the history, the physical, and the electrodiagnostic testing all together,” said Dr. Cartwright. “But clinical findings are the most important, and the electrodiagnostic testing needs to be interpreted in the context of the clinical presentation.”
Diabetes is a perfect example. Diabetic neuropathy can result in slow conduction velocities or apparent conduction blocks that lead to misdiagnosis as CIDP, Dr. Gorson said. “You can surpass the cutoffs and qualify for what would be considered true, demyelinating range abnormalities that meet electrodiagnostic criteria for CIDP, but then you see that the patient has long-standing, poorly controlled diabetes with clinical features of diabetic neuropathy,” he added. “That’s [usually not coexistent] CIDP, notwithstanding the demyelinating features on EMG.”
Most patients with CIDP – 80% – have the classic or typical form that involves sensory and motor dysfunction “that slowly progresses over more than 8 weeks with generalized limb weakness in the arms and legs with proximal muscle predominance,” Dr. Gorson said. “That’s the typical clinical pattern.”
In Dr. Cartwright’s experience, typical CIDP is symmetric and usually involves both proximal and distal parts of the extremities. “There usually are more motor symptoms than sensory symptoms. When we do a physical exam, we detect weakness, and we usually find areflexia, or decreased reflexes.”
Atypical presentations affect an estimated 20% of patients and include predominantly or purely sensory or motor impairment, a focal distribution, a predominantly distal distribution, or asymmetric involvement. “These are the cases in which missed CIDP diagnoses can be an issue,” as well as faulty diagnoses of CIDP, Dr. Cartwright said. “Sometimes CIDP is more focal, with only one limb or two nerves involved. And there are pure sensory variants and distal-predominant variants … and asymmetry, which can be confusing.”
Atypical variants may not meet electrodiagnostic criteria and often require more laboratory, imaging, and other investigations to determine whether CIDP supportive criteria are met. “Supportive criteria include nerve root enhancement and enlargement on MRI, high CSF protein levels, and a nerve biopsy showing demyelination and inflammation,” Dr. Gorson said.
Kenneth C. Gorson, MD
Avoid misclassifying mild slowing of nerve conduction velocity as demyelinating. Documentation of peripheral nerve demyelination is an important part of the diagnostic process, but care should be taken to interpret electrodiagnostic patterns correctly. Another study led by Dr. Allen of CIDP misdiagnosis – this one published in 2018 – documented the role that misinterpretation of electrodiagnostic studies plays in CIDP misdiagnosis. When Dr. Allen and colleagues reevaluated a series of referred patients, they confirmed initial electrodiagnostic conclusions in only 43% of the misdiagnosed cases (Muscle Nerve. 2018 Apr;57[4]:542-9).
Neurologists “weren’t applying the proper cutoff criteria to call something demyelinating,” said Dr. Gorson, who reviewed the study and wrote an accompanying editorial (Muscle Nerve. 2018 Apr;57[4]:517-9). “They overcalled modest conduction velocity slowing, partial conduction blocks, and all sorts of modest electrophysiologic abnormalities that were real but didn’t meet the ED criteria as demyelinating.”
It’s important to appreciate that amplitude-dependent slowing occurs with length-dependent axonal neuropathies or motor neuron disease, Dr. Gorson emphasized. Patients with substantial axonal loss have some degree of conduction slowing, but this slowing is consistent with amplitude reduction – a marker of axonal loss. In such cases, he said, “we’re often seeing slowing that’s not [significant] enough to suggest it’s an acquired demyelinating abnormality independent from the degree of axon loss.”
And as previously discussed, amplitude-independent slowing can occur in diabetic patients who have clinical features of a length-dependent polyneuropathy, making it critical that patients also fulfill clinical criteria for CIDP.
Dr. Cartwright also advises neurologists to ensure that patients are properly warmed for nerve conduction studies. “Patients need to be warmed, otherwise the conductions are slow and can look like demyelination,” he said. “Most labs have [warming] protocols, but sometimes the lab may forget or [may rush]. So if you see someone whose nerve conduction studies are slow and it looks like a demyelinating process, make sure their extremities were properly warmed.”
Look for demyelination at sites not prone to compression. Almost one-quarter of the cases of electrodiagnostic misdiagnosis identified in Dr. Allen’s 2018 study involved focal demyelination at compressible sites. “We need to stop classifying conduction slowing at regions of focal compression or entrapment [as representing demyelinating lesions],” Dr. Gorson said.
“This would be a patient with carpal tunnel syndrome, for instance, who has clinical features [of the disorder] and a prolonged distal motor latency of just the median nerve and just on one side,” he said. “It’s not inflammation from CIDP, but compression or entrapment of [the nerve].”
Attributing conduction velocity slowing at the ulnar nerve at the elbow to CIDP is a similar pitfall, Dr. Cartwright said. “We need to avoid areas of entrapment.”
Don’t limit electrodiagnostic testing to the lower limbs. Current electrodiagnostic criteria require evidence of acquired demyelination in two nerves, but among experts “the general rule is that you should at least study a couple of motor nerves in one arm in addition to studying the legs,” said Dr. Gorson. “Some people suggest at least four motor nerves, of which two should be in the upper limbs. And there have been published papers [that advise] looking at eight motor nerves to increase your yield of finding demyelination.”
In classic CIDP the upper limbs are involved, Dr. Gorson said. “There are rare variants of CIDP that are lower extremity only, but it would be unusual in those cases not to see asymptomatic demyelinating changes in the upper limbs,” he said. “If nerve conduction abnormalities are strictly limited to the legs, and there are no demyelinating features or, for that matter, clinical features in the arms, you should question making the diagnosis.”
Consider ultrasound of the nerves as a complement to EMG. The EFNS/PNS guidelines list MRI of the spinal roots, brachial plexus, and lumbosacral plexus as possible investigation to support a diagnosis, but MRI is “expensive, challenging for the patient, and [produces] images that are of variable quality,” said Dr. Cartwright.
While there no standard ultrasound criteria, academic centers and private practices are increasingly performing nerve imaging with ultrasound to look for multifocal nerve enlargement. It’s done in conjunction with the EMG and is tailored to the patient’s physical, history, and electrodiagnostic findings. “Nerves that demyelinate and remyelinate get bigger. … We can actually measure the cross-sectional area of nerves and compare them to normal values for upper and lower extremity nerves,” said Dr. Cartwright, who has coauthored a textbook and numerous papers on neuromuscular ultrasound. “In CIDP patients, there usually isn’t any question at all as to whether they’re enlarged.”
Recognize that CSF total protein increases with age. Dr. Gorson and other CIDP specialists have cautioned about overreliance on modest elevations in CSF protein levels in making a diagnosis of CIDP, and a 2019 study by Dr. Allen and University of Ottawa coinvestigators showed that use of higher and age-dependent CSF total protein values improves CIPD diagnostic specificity without compromising sensitivity (Muscle Nerve. 2019 Aug;60[2]:180-3). Values gradually exceed 0.60 g/L after age 50, they found.
“If you see a protein level of 100 in someone who’s 60, for instance, that probably will be clinically relevant, whereas a level of 55 in someone who is 80 years old” is likely age related, Dr. Gorson said. Even in the context of a laboratory upper reference limit of 0.45 g/L, this CSF total protein level should not be used to support a diagnosis of CIDP.
On ultrasound, the median nerve (yellow dots) in the upper arm of a 40-year-old woman with CIDP has a significantly enlarged cross-sectional area from a typical 10 mm2 to 70 mm2, and the nerve shows a mixture of enlarged hyper- and hypo-echoic fascicles.