Polyneuropathy monitoring in Parkinson's disease patients treated with levodopa/carbidopa intestinal gel

Abstract Objectives Levodopa‐carbidopa‐intestinal‐gel (LCIG) infusion is an effective treatment for advanced PD with motor fluctuations. Polyneuropathy occurs as a complication in 10–15% of patients. We wanted to assess the frequency of polyneuropathy in Finnish advanced Parkinson's disease (PD) patients with continuous LCIG infusion, and the value of different clinical monitoring parameters during follow‐up. Materials and methods Patient records of PD patients started on LCIG infusion at Helsinki University Hospital who received nerve conduction studies at baseline and 6 months after treatment initiation were reviewed for epidemiological information, mini mental state examination, baseline and 6 months’ UPRDS‐III, weight, body mass index, levodopa dose (LD), plasma homocysteine levels, folate, vitamin B6 and B12. Results Out of 19 patients (n = 6 on B‐vitamin substitution), two (10.5%) developed new‐onset polyneuropathy after initiation of LCIG therapy (n = 0 with vitamin substitution). Neuropathy was associated with significant weight loss (BMI reduction > 1.5), but not with other monitoring parameters. Homocysteine rose significantly in patients not substituted with B‐vitamin complex, but not in patients with B‐vitamin substitution. Homocysteine changes correlated with LD changes in the absence of vitamin B substitution. After oral B‐vitamin substitution, both patients’ polyneuropathy remained electrophysiologically and clinically stable. Conclusions Rates of polyneuropathy in Finnish PD patients with LCIG treatment are comparable to previous studies. Patients’ weight should be included in regular follow up monitoring and can be used for patient self‐monitoring. Vitamin B substitution appears to reduce coupling between levodopa dose and homocysteine and may be useful to prevent polyneuropathy related to LCIG.

With increasingly wide use of LCIG infusion, a considerable rate of new onset polyneuropathy has been reported. Cases range from acute onset polyneuropathy to subacute and chronic onset. PD patients have an increased rate of polyneuropathy compared to healthy age matched controls . This could be a feature of the disease itself, that is, caused by alpha synuclein deposition in the peripheral nervous system, or due to cumulative levodopa exposure during treatment, possibly in conjunction with nutritional factors such B-vitamin deficiency, and high serum homocysteine level (see Comi et al., 2014;Müller et al., 2013, for a review and see also Ceravolo et al., 2013;Rajabally & Martey, 2011;Toth et al., 2010;Toth et al., 2008).
The rates reported for new onset polyneuropathy during LCIG infusion therapy range from 4.5% to 19.0% (Merola et al., 2016;Uncini et al., 2015;Zadikoff et al., 2020). A cross-sectional study comparing PD patients on different therapeutic regimens (LCIG infusion vs. oral levodopa vs. other dopaminergic treatment) found that the prevalence of polyneuropathy was highest in the LCIG group, and higher in patients treated with levodopa compared with other dopaminergic therapy . Patients with polyneuropathy had significantly higher daily levodopa intake, and LCIG infusion patients had significantly lower vitamin B12 and folate levels, and higher homocysteine levels than the other groups. Levodopa daily dose correlated with homocysteine levels. A recent post-hoc analysis of phase III data and registry data showed that the occurrence of polyneuropathy is increased in LCIG infusion patients with LD exceeding 2000 mg per day (Zadikoff et al., 2020). Vitamin B1 and B12 supplementation stabilized or improved polyneuropathy symptoms (Merola et al., 2016), and there were no cases of acute or subacute neuropathy in a cohort of patients substituted with vitamin B complex from onset of LCIG therapy (Rispoli et al., 2017). However, 19% of patients developed chronic polyneuropathy despite vitamin supplementation (Rispoli et al., 2017

Electrophysiological assessment
All patients included in the study underwent at least two ENMG assessments (baseline and 6 months after treatment initiation), which were carried out by the same clinical neurophysiologist (J.T.). Sensory nerve action potential (SNAP) and conduction velocity were studied in ulnar, radial, and sural nerves. Motor compound action potential, distal latency, and conduction velocity were studied in median, deep peroneal, and tibial nerves. Electromyography (EMG) was studied from anterior tibial muscle and gastrocnemius muscle. Polyneuropathy was diagnosed if at least two nerves showed reductions in amplitude and/or conduction velocity and the changes were not explained by a common focal compression neuropathy (such as carpal tunnel syndrome, ulnar nerve compression at elbow, peroneal nerve at head of fibula), or by common radicular symptoms. More than 50% reduction in SNAP or CMAP amplitude during follow up, or reduction of amplitude or conduction velocity compared to lower limits of common reference values was considered as an abnormal result in the follow up ENMG, leading to the diagnosis of polyneuropathy. More than 30% reduction in SNAP or CMAP during follow up was defined as at risk of developing polyneuropathy. In EMG, appearance of motor unit loss (rarefication) or fibrillations suggested recent or ongoing motor axonal damage.

Statistics
Statistical analysis was carried out using SPSS version 25 (www.ibm.com). Normality of data distribution was tested using a Kolmogorov-Smirnov test. A paired t-test was used for normally distributed data, and a Wilcoxon signed-rank test for paired samples for non-normally distributed data and samples of <10 subjects. Comparing patients with and without neuropathy, a Mann-Whitney U-test was used. Correlation testing was carried out using the Pearson correlation coefficient for normally distributed data, and Spearman correlation coefficient for non-normally distributed data and samples of <10 subjects.

RESULTS
Over the period from 2017 to 2018, 19 patients were started on LCIG and were followed up with at least 2 nerve conduction studies. Patients' epidemiological data can be found in Table 1

DISCUSSION
In summary, we observed a relatively low rate of new onset polyneuropathy after initiation of LCIG therapy (2/19 or 10.5%) in advanced PD patients who did not have any pre-existing polyneuropathy. Rows in light grey highlight patients with B-vitamin substitution from baseline onward (marked with *), dark grey highlighting marks patients who developed polyneuropathy. BMI, body mass index; PNP, polyneuropathy; n, normal; elev., elevated; red., reduced.

Occurrence of polyneuropathy
The rates of polyneuropathy observed in our cohort are low compared to previous studies. None of our advanced PD patients had pre-existing polyneuropathy, compared to 9-55% in previous studies (Merola et al., 2016;Rispoli et al., 2017;Toth et al., 2008). The incidence of new-onset polyneuropathy (2/19, 10.5%) was relatively low, but in keeping with the reported range from 4.5% to 19.0% (Merola et al., 2016;Uncini et al., 2015;Zadikoff et al., 2020). Disease duration was shorter than in some previous studies (12.2 years on average), so cumulative exposure to levodopa was shorter, which may explain differences in baseline rates of polyneuropathy. Because of the relatively short follow-up time of 6 months, chronic polyneuropathies were possibly missed. Further reasons for differences in both baseline occurrence of polyneuropathy and new-onset polyneuropathy after LCIG initiation may include genetic differences (all patients were of Finnish descent), or dietary differences (Nordic Council of Ministers, 2014). It is known that the Finnish gene pool is quite distinct from other ethnicities (Lek et al., 2016), and previous studies on LCIG and neuropathy were carried out in several European countries and Australia (Zadikoff et al., 2020), Italy Merola et al., 2016;Rispoli et al., 2017), and Germany (Jugel et al., 2013). Epidemiological studies of the prevalence of polyneuropathy in Finland are unavailable but one study reports a rate of polyneuropathy of 1.4% in middle-aged control subjects (Lehtinen et al., 1989).

LCIG infusion, weight loss, and polyneuropathy
We found that the occurrence of polyneuropathy was associated with significant weight loss over the first 6 months (BMI reduction ≥ 1.5 kg/m2). Weight loss in LCIG-treated patients has been found to correlate with time spent with dyskinesia (Fabbri et al., 2019). Data on weight is only available in a few studies on LCIG treatment (Jugel et al., 2013;Klostermann et al., 2012;Zadikoff et al., 2020). The frequency of weight loss was higher in patients with LD ≥ 2000 mg daily (17.0% of patients), as was the rate of polyneuropathy, even though daily hours of dyskinesia were similar between the groups. However, an association between weight loss and polyneuropathy remains unclear (Zadikoff et al., 2020). Two of 20 PD patients treated with LCIG infusion developed acute sensorimotor axonal polyneuropathy with significant weight loss in both cases, as well as reduced vitamin B6 and folate and increased homocysteine (Klostermann et al., 2012). The number of electrophysiologically affected nerves was found to correlate with weight lost since initiation of LCIG infusion (Jugel et al., 2013).

Levodopa dose and homocysteine levels
The role of the so-called one carbon pathway, which is relevant in levodopa metabolism, and its role in polyneuropathy are discussed in a review by Uncini et al. (2015). Homocysteine is a by-product of levodopa metabolism, so high levels of daily levodopa dose raise plasma homocysteine levels, which has been shown before Zadikoff et al., 2020). Vitamin B6, vitamin B12, and tetrahydrofolate are involved in homocysteine metabolism, either by cleaving it into cysteine and methylmalonic acid (vitamin B6), removing it from the homocysteine cycle, or by restoring homocysteine to S-adenosylmethionine (vitamin B12 and folate), the form relevant in metabolizing levodopa (Uncini et al., 2015).

Vitamin B levels, substitution, and polyneuropathy
Our polyneuropathy cases both showed a relative decrease in vitamin B6 levels compared to baseline but had normal vitamin B12 and folate at all times. Case studies have shown lowered vitamin B6, B12, and folate in individual patients with LCIG-associated polyneuropathy (Klostermann et al., 2012). One study found reduced folate and vitamin B12 levels in PD patients on levodopa compared with other treatments, but there were no differences between patients with and patients without polyneuropathy . Another study reports reduced vitamin B6 levels and elevated homocysteine, but normal folate and vitamin B12 levels, in patients with advanced PD with LCIG infusion or oral levodopa therapy (Loens et al., 2017).
Almost all of the patients suffered from polyneuropathy, regardless of therapy.
On the other hand, none of the patients with vitamin B complex substitution developed polyneuropathy in our cohort, in keeping with previous studies (Rispoli et al., 2017). Vitamin B1 and B12 supplementation stabilized or improved polyneuropathy symptoms (Merola et al., 2016). Thus, vitamin B complex is relevant in the pathophysiology of LCIG-associated neuropathy, but not reliable as sole clinical marker.
Is there something special about vitamin B6? As mentioned earlier, vitamins B6, B12, and folate are all required for levodopa metabolism and thus higher LD will deplete plasma levels. However, LCIG contains 20% carbidopa, which irreversibly binds to free vitamin B6 and deactivates vitamin B6-dependent enzymes (Loens et al., 2017). Hence, the presence of carbidopa might further reduce vitamin B6 availability, deactivate B6-dependent enzymes cleaving homocysteine and removing it from the homocysteine cycle, and thus, drive the development of levodopa-associated neuropathy.'

Strengths and limitations
Patients were collected in one center only over a relatively short period

TRANSPARENT PEER REVIEW
The transparent peer review history for this article is available at https: //publons.com/publon/10.1002/brb3.2408

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request and based on our institutional research permission protocol.