Patients with Parkinson's disease (PD) treated with oral levodopa have a higher prevalence of chronic, prevalently sensory, usually mild axonal polyneuropathy (PNP). Several studies showed a positive association among PNP, cumulative levodopa dosage, low serum B12 and high-homocysteine and methylmalonic acid level. Anecdotal severe acute or subacute PNPs thought to be Guillain-Barré syndrome have been reported in patients receiving continuous intraduodenal infusion of levodopa/carbidopa intestinal gel (LCIG). We report an additional acute case and by a systematic literature search we also reviewed the clinical and laboratory features of 13 other acute and 21 subacute PNP cases occurring during LCIG treatment. In series with at least nine patients, the mean frequency of acute and subacute PNP is 13.6% and the mortality rate at 6 months in acute cases is 14%. The great majority of PNP cases displayed axonal sensory-motor and reduced vitamin B12 levels, and alterations of metabolites of 1-carbon pathway were found in most patients. We discuss the possible role of high-levodopa dosage, vitamin B12, B6 and folate deficiency and accumulation of homocysteine and methylmalonic acid in the pathogenesis to conclude that there is enough, although circumstantial, evidence that alterations of 1-carbon pathway are implicated also in acute and subacute PNP during LCIG usage. There is no solid proof for a dysimmune pathogenesis and in our opinion acute, subacute and chronic PNP, either after oral levodopa or LCIG, represent a continuum. Finally, we propose recommendations for prevention and management of PNP occurring during LCIG treatment.
- PARKINSON'S DISEASE
- GUILLAIN-BARRE SYNDROME
Statistics from Altmetric.com
In recent years, several studies pointed out a higher frequency of clinical and/or electrophysiological findings of polyneuropathy (PNP) in 38–55% of patients with Parkinson's disease (PD) compared to 8–9% of controls.1–5 PNP was correlated with levodopa (L-dopa) dosage, age of patients, low-serum vitamin B12 and high-homocysteine (Hcy) and/or methylmalonic acid (MMA) levels.1–5 PNP was also detected in 5–12% of patients with PD not treated with L-dopa and epidermal nerve fibre loss was reported in treated and untreated patients, although Meisnerr corpuscle loss was present only in treated patients.4–7 These findings may suggest that neuropathic involvement is just another aspect of the primitive neurodegenerative disease. However, in a recent study in a large population, the evidence of a strong association with the duration of L-dopa exposure but not with disease duration or disease severity supports the role of L-dopa as the main risk factor in the development of PNP.5
Studies of PNP in PD mainly focused on patients treated with usually moderate oral L-dopa dosage and exhibiting chronic, mild, prevalently sensory or even subclinical neuropathy.
On the other hand, in patients receiving continuous intraduodenal infusion of levodopa/carbidopa intestinal gel (LCIG) anecdotal cases of severe acute or subacute PNP thought to be or resemble Guillain-Barré syndrome (GBS) have been reported.8–11 We describe an additional acute patient and have reviewed the reported cases to determine whether PNPs occurring during LCIG treatment are a dysimmune disorder, such as GBS, or whether they are a continuum with the cases occurring during oral L-dopa treatment. We also examine the possible pathogenic mechanisms and propose recommendations for prevention and management.
We made a systematic literature search in PubMed using multiple combinations of the following search terms: “Parkinson disease”, “levodopa”, “duodenal levodopa infusion”, “levodopa/carbidopa intestinal gel”, “LCIG”, “polyneuropathy”, “peripheral neuropathy”, “acute neuropathy” “subacute neuropathy”, “Guillain-Barré syndrome”, “vitamin B12” and “homocysteine”. We reviewed abstracts for relevance and then searched the reference lists of appropriate articles to obtain other pertinent references. We defined PNP as acute when the nadir was reached within 4 weeks and subacute when the nadir was reached between 4 weeks and 8 weeks or if explicitly indicated by authors.
A woman with type 2 diabetes had been diagnosed with PD for 18 years. Under treatment with oral L-dopa and different combinations of dopamine agonists, she developed severe on-off fluctuations. LCIG was started with 1650 mg L-dopa equivalent. For 2 months before initiating treatment the patient had mild distal lower limb paresthaesias. Neurological examination showed no weakness or sensory abnormalities and deep tendon reflexes were normal except for absent ankle jerks. An electrophysiological study carried out 3 months after the onset of LCIG showed normal compound muscle action potential (CMAP) amplitudes, normal motor conduction velocities (CVs) and not recordable sensory action potentials (SAPs) (table 1). These results were similar to an electrophysiological study carried out before the onset of LCIG. Four months after the onset of LCIG and 2 weeks after a gastroenteritis attack the patient had acute worsening of sensory symptoms with ascending paresthaesias and numbness, developed ataxia with inability to walk and, in 10 days, flaccid tetraparesis with absent tendon jerks. Cranial nerves were normal. Cerebrospinal fluid (CSF) proteins were increased (148 mg/dL). Electrophysiology performed 1 week after the onset of weakness showed low amplitude distal CMAPs with distal motor latencies (DMLs) and CVs, which did not fulfil the criteria for acquired demyelination (table 1).12 In the peroneal nerve there was an abnormal amplitude reduction of CMAP from below the fibular head compared with CMAP from ankle stimulation, which was ascribed (as after 3 weeks CMAPs were not recordable) to pseudoconduction block due to advancing axonal degeneration. Needle electromyography 4 weeks after the onset of weakness showed spontaneous activity in upper and lower distal muscles with a reduced recruitment pattern.
B12 was 218 pg/mL (211–911), Hcy 8 μmol/L (<14), folate 12.3 pg/mL (3.1–17.5). IgG and IgM to gangliosides GM1, GD1a and GM2 were negative. A provisional diagnosis of GBS was made, LCIG was halted and the patient was treated with six plasma exchanges, and with B12, B1 and folate supplementation. The patient gradually recovered her strength in 5 months but distal paresthaesias and numbness persisted. Neurophysiology showed progressive improvement of CMAP amplitudes in 6 months (table 1). This was the only patient out of 15 treated with LCIG who developed an acute neuropathy.
Acute and subacute PNP during LCIG treatment
The clinical and laboratory findings of the reported patients are summarised in online supplementary table S2. Reports were single cases or small series sometimes only in abstract form. Overall, during LCIG treatment 14 patients developed acute and 21 subacute PNP. Considering the series included at least nine patients the mean frequency of acute and subacute PNP was 13.6% (range: 6.6–28%).8 ,13–16
All patients, except the one we describe, were reported to be free of neuropathic symptoms before LCIG treatment. The L-dopa equivalent dose ranged from 1100 to 3800 mg/day. The duration of treatment with LCIG at the onset of PNP ranged from a few weeks to 29 months.
Antecedent infections were explicitly reported to be absent in 4 acute and 14 subacute patients.14 ,17 ,18 The patient we report had gastroenteritis 2 weeks before onset of PNP. Weight loss preceding or concomitant to the onset of PNP was reported in 1 acute and 3 subacute patients with PNP.15 ,17 ,19 One patient, 2 weeks after the onset of distal paresthaesias and numbness, fell and underwent surgery under general anaesthesia with nitrous oxide inhalation.9 On the 14th postoperative day she became acutely confused and developed increasing weakness in all her limbs.
In nine patients, clinical features were thought to be compatible with GBS or an acute inflammatory neuropathy and in another patient a severe acute sensory–motor PNP was described (current report).8 ,10 ,14 ,20 In one patient, ascending weakness with distal tingling, rapidly progressing to flaccid tetraparaesis, was accompanied by confusion and decreased consciousness.9 Among subacute patients, 17 had a sensor-motor PNP predominantly distal at lower limbs.11 ,13 ,14 ,17 ,18 ,20 Two patients had a clinically exclusive sensory neuropathy.16 ,18
In subacute patients, B12 was decreased in 3/10, folate in 2/3 and B6 in 7/7 patients; Hcy was elevated in 5/6 and MMA in 1/3 patients.11 ,13 ,14 ,16–18 ,20 Mancini et al14 did not report the disaggregated data of the 14 patients with acute and subacute PNP, however, vitamin B12 and folate levels were significantly lower and Hcy levels were significantly higher in patients treated with L-dopa (either oral or duodenal infusion) compared with patients treated with other dopaminergic drugs. Patients with PNP showed significantly higher Hcy levels than patients without PNP.
In five acute and four subacute patients, antibodies to gangliosides were absent (current case).15 ,17 In two subacute patients, antibodies to gangliosides were reported to be present although the specific antibodies, the Ig isotype and the titre were not indicated.11 CSF examination showed increased protein content with normal cells in 5/6 acute and in 4/10 subacute patients (present report).8 ,9 ,11 ,13 ,14 ,16–18
In six acute patients, electrophysiology was indicative of a severe axonal sensory–motor neuropathy.9 ,15 ,20 In the patient we report, SAPs were already not recordable before the onset of acute weakness and in one patient the axonal involvement was limited to sensory fibres.15 Four other acute patients were reported to show a mixed axonal and demyelinating pattern with a decrease in motor conduction velocities <70% of lower limit of normal values and increased F wave latencies. However, only the mean values were shown.14 In subacute patients, electrophysiological studies were reported to be consistent with mild-to-severe axonal sensory–motor PNP in 16 patients.14 ,17 ,18 ,20 In two patients, mixed axonal and demyelinating features were reported but examining the values motor CV was decreased in the demyelinating range in only one nerve, not fulfilling the commonly employed criteria for acquired demyelination.11 ,12 Follow-up at 3–4 months showed decreased CMAP amplitudes with normal distal motor latencies indicative of axonal degeneration. Only in one patient did nerve conductions show prolonged distal motor latencies, partial motor nerve conduction blocks in peroneal and ulnar nerves and reduced conduction velocities of motor and sensory nerves in the demyelinating range.13
Sural nerve biopsy in two acute patients showed axonal degeneration with mild inflammatory infiltrates in one.15 Two subacute patients showed a marked reduction in myelinated nerve fibre density and considerable endoneurial oedema.18 Myelin debris was found in endoneurial macrophages indicating recent active nerve fibre degeneration. Inflammatory infiltrates and indirect signs of inflammatory neuropathy such as focal loss of nerve fibres were absent.
Treatment and outcome
In the great majority of acute patients LCIG was withdrawn. Seven patients were treated with plasma exchange, high-dose intravenous immunoglobulins (IVIgs), or steroids with ‘some benefit’ in two.8 ,14 Four patients were treated only with vitamin supplementation: one patient had a rapid improvement of strength in 2 weeks and was able to walk with a cane after 3 months, one had some benefit and in the other two PNP did not worsen.9 ,15 ,20 Two acute patients died of concurrent disease after 3–6 months.14
Ten patients with subacute PN and vitamin B12/Hcy/folate imbalance were treated only with vitamin supplementation and the majority showed clinical improvement within 3 months.14 In five patients, LCIG was continued and vitamin supplementation given with improvement or disappearance of weakness over several months or not worsening whereas numbness and tingling remained unchanged in two.17 ,18 ,20 Both subacute patients described by Galazky11 with increased CSF protein content and antibody to gangliosides received IVIg. Owing to progression one patient was also treated with methylprednisolone, plasmapheresis and vitamins. Immunomodulatory treatments did not affect neuropathy, which progressed to severe tetraparesis in both patients. At this stage LCIG was discontinued. During the following 3 months, PNP stabilised and subsequently both patients slowly recovered with mild residual motor deficits. In the patient with a demyelinating neuropathy LCIG was interrupted and B1, B12 and folate given although no alterations of B12, Hcy and MMA were found.13 A partial recovery of clinical symptoms was observed after 1 month, and a gradual improvement of nerve conduction studies was found in the following 5 months. One patient developed a subacute PNP despite B12 supplementation.16 B12 level was normal but Hcy and MMA were not tested. A few weeks after the interruption of LCIG the patient showed mild reduction of PNP symptoms.
Acute PNP during oral L-dopa treatment
Two patients receiving high oral L-dopa dosages acutely developed a severe, sensory ataxic PNP.21 Electrophysiology showed absent SNAPs in both patients. One patient showed reduced CMAP amplitudes with slightly reduced CVs. The other had prolonged DMLs and conduction block in the ulnar nerve with reduced CVs. In both patients’ B12 levels were still in the normal range, and Hcy and MMA levels were not checked. Sural nerve biopsy in both patients showed severe active axonal degeneration. The patient showing some demyelinating features at electrophysiology was treated with IVIg and PNP stabilised but did not improve. The other patient received no specific treatment and continued to worsen over 9 years.
Chronic PNP during LCIG treatment
Patients during LCIG also develop chronic neuropathic symptoms and signs. Eleven out of 15 patients with PD treated with LCIG had mild to moderate distal hypoesthaesia and six had severe neuropathic pain interfering with daily activities.22 At electrophysiology all 15 patients had at least one abnormal SAP and 10 patients had at least one abnormal CMAP with normal CV. Electrophysiological abnormalities correlated with L-dopa dosage and weight loss since initiation of therapy.
In phase 3 clinical trials, DUODOPA was administered to 416 patients.23 PNP, with unspecified course, was reported in 4.1% of patients. This review shows that acute and subacute, usually severe, PNPs occur in 13.6% of patients in series with at least nine patients under LCIG treatment. In acute patients mortality was 14% within 6 months from the onset of PNP. The picture resulting from the analysis of PNP cases during LCIG treatment is quite confused. There are several issues:
On the role of L-dopa on deficiency of vitamin B12, B6, and folate or increase of Hcy and MMA in the pathogenesis of PNP;
Whether acute and subacute PNP under LCIG treatment have some relationship with GBS;
Alternatively, whether there is a connection between acute and subacute, often severe, PNP under LCIG treatment and the chronic usually mild PNP occurring during oral L-dopa treatment;
Finally, how acute PNP during LCIG treatment might be prevented and managed.
Biochemical relationships between L-dopa, B12, Hcy, MMA, folate and B6
Patients of PD with acute and subacute PNP under LCIG treatment, as well as patients with chronic PNP during oral L-dopa treatment, show alterations of B12, Hcy, MMA, folate and B6 levels.
The metabolism of L-dopa and the 1-carbon pathway may explain, at least in part, these alterations (figure 1).24 In presence of high dosages of L-dopa there is an increased request of S-adenosyl-methionine as the methyl-group donor with consequent elevations of both Hcy and MMA and depletion of B12 body stores. This scenario may also explain, at least in part, the reduction of folate and B6 during L-dopa treatment.
One-carbon pathway alterations and polyneuropathy
Reduced B12, folate and B6 levels and increased Hcy and MMA levels may all induce PNP in patients with PD treated with L-dopa.
Vitamin B12 participates in two enzymatic reactions: conversion of l-methylmalonyl coenzyme A into succinyl coenzyme A, and methylation of homocysteine to methionine. It is not precisely known how B12 deficiency induces neuropathy. Succinyl coenzyme A enters the citric acid cycle and participates in gluconeogenesis. This pathway, linking carbohydrate and fat metabolism, plays a role in myelin synthesis.25 Methyl-B12 methylates RNA, slowing mRNA degradation. Impaired methylation of RNA may explain the involvement of long axons in vitamin B12 deficiency, since RNA turnover in neurons is very high, and rapid degradation of mRNA might limit production of proteins essential for axon maintenance.26 The neuropathy of B12 deficiency is mainly sensory and, when abnormal, electrophysiology indicates axonal involvement.27 CSF protein content is increased in 65–68% of patients with neurological complications of B12 deficiency.25 ,28 One-third of B12 deficient patients with PNP have an acute onset of PNP and the sudden onset of PNP following exposure to nitrous oxide, which transform B12 into a biologically inactive form, is well described.27 This mechanism could be involved in the pathogenesis of one acute case of PNP under LCIG treatment.9 The prevalence of B12 deficiency increases with age.29 This may explain the positive correlation between PNP and age in patients with PD.4 ,5
The possibility that isolated folic acid deficiency may induce (as B12) subacute combined degeneration of the cord and neuropathy, although acknowledged since the 1970s, has rarely been reported and large scale confirmatory studies are needed.30 Methionine synthetase reaction is thought to be the basis for the neurological consequences of both vitamin B12 and folate deficiency, suggesting a common pathogenic pathway (figure 1).
PNP associated to B6 deficiency is an axonal sensory–motor neuropathy and was commonly seen in patients treated with the antitubercolosis drug isoniazid.31 On the other hand megadoses of pyridoxine may induce a severe ataxic sensory neuropathy.32
Hcy, shown to be greatly elevated in LCIG-treated patients with PD,33 may be neurotoxic through several mechanisms such as depletion of methyl-groups that protect against oxidative stress by increasing vulnerability to mitochondrial toxins, by inducing inflammatory reactions and also by impairing DNA repair mechanisms.34 L-dopa-associated increased Hcy levels were correlated in an electrophysiological study to axonal loss in the sural nerve.35
The fact that only a fraction of patients with PD on L-dopa therapy develop neuropathy could be due to genetic susceptibility. One study showed that one-third of patients with PD with Hcy elevation have a more inefficient form of methylenetetrahydrofolate reductase (figure 1).36
Isolated elevated MMA is so far not known to have any specific effect on the peripheral nerve.
Some of the acute and subacute PNP in LCIG-treated patients were diagnosed to be GBS or GBS-like disorders (current case)8 ,10 ,11 and, therefore, received immunomodulatory treatment. GBS is an acute, postinfective, dysimmune polyneuropathy differing in clinical manifestations, pathology (primary demyelinating or axonal), association with antiganglioside antibodies and pathogenesis.37 Two-thirds of GBS cases are preceded by symptoms of upper respiratory tract infection or diarrhoea. The most frequently identified infectious agent is Campylobacter jejuni and, at least for the axonal subtypes, there is evidence of molecular mimicry between ganglioside-like lipooligosaccharide in the outer membrane of C. jejuni and gangliosides contained in the human peripheral nerves.37 At present, no biomarkers are available to diagnose GBS. It is possible that by chance patients with PD develop GBS during their lifetimes.38 However, considering the incidence of GBS (0.89–1.89 cases per 100 000 persons-year) the 14 acute PNPs reported from 2007 to 2014 are out of proportion with the expected GBS in the limited LCIG-treated population. We deem that acute–subacute neuropathies occurring under LCIG treatment are unlikely to be GBS because: (1) an antecedent infection was reported in only one case; (2) albuminocytological dissociation is not specific of GBS and has been described in 2/3 of patients with neurological complications of B12 deficiency; (3) the great majority of PNPs under LCIG treatment were characterised electrophysiologically and pathologically by a sensory–motor axonal neuropathy that, in western countries, is the rarest GBS subtype; (4) antiganglioside antibodies were absent in acute patients; the not better specified antigangliosides antibodies reported in two subacute cases could be an epiphenomenon; (5) in the great majority of cases investigated, reduced vitamin B12 level or alterations of other components of 1 carbon pathway were found.
The possibility that some inflammatory mechanism may be involved in PNP occurring during LICG treatment cannot be completely dismissed. Mild inflammatory infiltrates and intraneural oedema were reported in one acute and two subacute patients and perivascular lymphocytic cuffing was reported in one chronic neuropathy patient in treatment with oral L-dopa.15 ,18 ,21 It is interesting to point out that acute polyneuropathies resembling GBS, sometimes associated with encephalopathy, have been described in patients receiving gastrojejunal bypass with gastric restriction for morbid obesity.39 ,40 In these patients, CSF protein concentration may be elevated; electromyography and nerve conduction studies showed sensory–motor axonal PNP and nerve biopsy showed, besides active axonal degeneration, various degrees of perivascular inflammation.39 Although the pathogenesis remains obscure, the neurological complications after bariatric surgery are thought to be caused by thiamine, B12 and copper deficiency.
To reconcile the vitamins and inflammatory pathogenetic hypotheses are recent studies showing new effects and functions of B12 not related to its coenzyme role. B12 deficiency may determine an acute imbalance between myelinotoxic and trophic factors, with a subsequent increase of proinflammatory cytokines, toxins and free radicals, and a decrease of growth factor synthesis.41
Prevention and management of PNP in LCIG-treated patients
The question arises as to why patients with PD sometimes develop PNP right after starting LCGI, while they have previously been receiving oral L-dopa without problems. It is possible that increased dosage, availability and metabolism with consequent increased consumption of B12, possibly associated with reduced vitamins supply or absorption due to the gel delivered into jejunum, all play a role.
To patients using LCIG we recommend the evaluation of body mass index, nutritional status, conditions leading to gastrointestinal malabsorption and current medications, such as proton pump inhibitors, that are significantly associated with B12 deficiency.42 Baseline laboratory assessments should include vitamins B12, B6 and folic acid. With as many as 50% of cobalamin-deficient patients found to have normal serum cobalamin, it is recommended, to those patients with levels of less than 300 pg/mL, that they get measurements of Hcy and MMA (the latter is more specific). A neurological examination, nerve conductions and needle electromyography should be performed to detect a pre-existing neuropathy, even of different aetiology, that might represent a risk factor for further deterioration. In presence of laboratory abnormalities, but not symptoms or signs of PNP, supplementation therapy should be started before LCIG and continued during treatment. Patients who have already PNP symptoms and signs should receive supplementation and the indication to initiate LCIG should be well pondered.
If there are no neuropathic abnormalities LCIG can be started. As L-dopa metabolism is influenced by catechol O-methyltransferase (COMT), the administration of LCIG plus a COMT inhibitor might reduce hyperhomocysteinaemia and offers the possibility of reducing the LCIG dose.24 As vitamin status can change rapidly, and PNP has developed even within a few weeks after LCIG onset, B12 (and eventually Hcy and MMA) and folic acid should be tested monthly for the first 6 months, then every 3 months and, if abnormal, supplementation started. For the time being this approach seems more correct than the ‘pragmatical substitution’ that is often empirically carried out from the beginning of LCIG treatment. Moreover, vitamin supplementation did not prevent a subacute PNP in two patients.16
Patients developing acute and subacute PNP represent an emergency. LCIG should be halted, oral L-dopa possibly at lower dose equivalent/day reintroduced, vitamin status assessed and supplementation immediately started to minimise the amount of axonal degeneration, which is associated with poor and late recovery. No standardised recommendations are available for B12 replacement. A common approach is 1000 μg intramuscularly for 5–7 days followed by monthly 1000 μg intramuscular injection. Folate may be administered orally at the dosage of 5 mg/day. More problematic is the supplementation of B6, as the neuropathy from excessive intake of B6 occurs with doses as low as 200 mg/day and patients with polyneuropathy from any cause may be sensitive to toxicity from B6. Moreover, the supplementation of high B6 dosages is not advisable because it may contrast the action of decarboxylase inhibitors in inhibiting the conversion of levodopa to dopamine in the periphery. We do not think that there is evidence that immunomodulatory treatment is effective in acute LCIG-associated PNP.
Differential diagnosis with GBS may be sometimes problematic, as in the patient we report. Investigation on infective antecedents, serology for recent C. Jejuni infection and dosage of IgM and IgG isotypes of antiganglioside may help in the differential diagnosis. Involvement of cranial nerves, never described in acute PNP associated with LCIG treatment, suggests GBS. Serial nerve conduction studies confirm the nature of the neuropathy and, when clearly demyelinating, point more to GBS. CSF examination should be performed to rule out other causes of acute neuropathies characterised by pleocytosis such as Lyme disease, but does not help in the differential diagnosis with GBS. Moreover, alternative aetiologies of acute neuropathies, such as vasculitis, toxic neuropathy, tick paralysis or porphyria, should be excluded based on history or testing.
There is no solid proof that acute and subacute PNP under LCIG have a GBS-like dysimmune pathogenesis.
Although the pathogenesis of neuropathies during LCIG is far from being completely understood, we deem that there is enough, although circumstantial, evidence that B12 deficiency and alterations of 1-carbon pathway are implicated.
Acute, subacute and chronic PNP either after oral L-dopa or LCGI are a continuum in pathogenesis, course and severity.
Multicentre prospective studies are necessary to assess the incidence of PNP in LCIG-treated patients, its relationship to the specific modality of L-dopa delivery, the role of vitamins, homocystein, genetic factors and the usefulness of preventive supplementation treatment.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Contributors AU contributed with manuscript conception, literature search, wrote the first draft, critically reviewed each draft and approved the final version. RE contributed with the case report, critically reviewed each draft and approved the final version. MO contributed to literature search, critically reviewed each draft and approved the final version.
Competing interests None.
Ethics approval Ethics Committee of Santa Maria della Misericordia University Hospital, Udine.
Provenance and peer review Commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.