Intraindividual Variability of Nonmotor Fluctuations in Advanced Parkinson’s Disease

INTRODUCTION

Nonmotor symptoms (NMS) are reported in the majority of Parkinson’s disease (PD) patients and have strong impact on health-related quality of life (hr-QoL) [1]. Fluctuations of NMS can occur with oscillations in response to dopaminergic medication [2–4] and are present in up to 100% of fluctuating PD patients [3, 5]. This study complements preceding reports of the NonMotor Fluctuations in PD (NoMoFlu-PD) study, a multicenter cross-sectional study to assess NMS fluctuations (NMF) and their correlation with motor fluctuations and hr-QoL [6, 7], which used semi-structured interviews in combination with visual analogue scales (VAS) and VAS-based home diaries to characterize NMF in relation to motor states. We found complex heterogeneous patterns of NMFs with particularly frequent and severe fluctuations in psychiatric NMS compared to other NMS domains [6, 7]. However, it remains unclear whether and how NMF vary in an individual subject. The repeated evaluation of ten key NMS over five motor On and five Off periods reported in this manuscript allowed us to estimate the intraindividual (within-patient) stability of NMF. The present analyses thus aim to investigate intraindividual NMS variability of symptom frequency and severity between multiple motor On/Off states based on repeated home diary assessments in fluctuating PD.

SUBJECTS AND METHODS

The study population and assessments have been described elsewhere [6]. In brief, PD patients of all ages and disease severities were included in the NoMoFlu-PD study, if they experienced documented motor fluctuations and in addition did not show any signs of atypical parkinsonism, psychosis, dementia or other relevant conditions interfering with the study protocol. Subjects rated a total of ten NMS (Table 1) during five self-perceived motor On and Off states at home as “present” or “absent” and subsequently the NMS severity using a visual analogue scale (VAS) with a symptom range from 0 (not present) to 100 (maximum symptom intensity). During the initial study visit, patients were trained in how to use a visual analogue scale and their diaries. In addition, the study visit included a levodopa challenge test, so patients were asked to arrive in the outpatient clinic after a 12 hour medication withdrawal [6]. During the visit, patients experienced both a defined motor On and Off state and were asked to complete their home diaries while experiencing similar motor states.

Statistical comparisons between motor states and NMS intensity ranges or patterns, respectively, were calculated using Wilcoxon rank test, McNemar test, Student's paired and unpaired t-test, χ ², Fisher’s exact test, and Friedman non-parametric repeated-measure ANOVA as appropriate. In addition, we calculated Mahalanobis distances (MD), a measure of the distance between individual patients' NMS severity ranges and mean NMS ranges across all the NMS. If not stated otherwise, data are displayed as mean ± standard deviation (SD), absolute numbers or numbers (%). Two-tailed P <  0.05 were deemed statistically significant. Due to the explorative character of the study, α adjustment of P-values has not been carried out.

RESULTS

Demographic characteristics of the overall NoMoFlu-PD study cohort have been reported previously [6]. Complete home diary datasets were available from 38 (38%) patients (24 men (63%); mean age ± SD: 65.6 ± 8.2 years; disease duration: 10.3 ± 7.0 years; Hoehn and Yahr stage in On: 2.4 ± 0.9 and Off state: 3.1 ± 1.0; P <  0.0001 [Wilcoxon test]; UPDRSIII motor score in On: 15.3 ± 11.9 and Off: 31.1 ± 12.0; P <  0.0001 [paired t-test]). Patients with complete home diary datasets had younger age, shorter disease duration and milder PD symptoms compared to patients without complete home diary data (see Supplementary Table 1 for complete statistics). All subsequent analyses were performed for the former patient cohort in order to prevent bias due to skewed adherence to diary records. Most of the diaries were completed in the morning hours before 12:00 am for both motor states (medians [maximal interquartile ranges]: 9:30 to 11:30 [8:30–14:30] for the five On and 8:30 to 11:30 [7:00–16:00] for the five Off states) with no significant differences between the various motor states (P >  0.05; Friedman ANOVA).

For the assessment of intraindividual variability of NMF, diary data were dichotomized to assign patients to one of two groups within each motor state: “instable” fluctuating patients reported a respective NMS within 1–4 of 5 assessments of a given motor state, whilst “stable” fluctuating subjects presented with a specific NMS during either all or none of the investigated motor state. The percentage of instable fluctuating PD patients was used to assess NMS recurrence in the respective motor state (Table 1). We found a significantly higher proportion of instable NMS fluctuators within motor Off states solely for depression (P = 0.007), anxiety (P = 0.031) and bladder urgency (P = 0.0.049). All other NMS did not differ significantly between motor states. We did not detect any differences of demographic and clinical characteristics between stable and instable NMS fluctuators.

To further characterize NMS presentation in individual patients, we analyzed NMF patterns within both motor states. Patterns were defined according to the number of presentation of individual NMS over the total of five motor states ranging from 0/5 (symptom absent in all five assessments) to 5/5 (symptom present in all five explorations). All NMS except problems with concentration/attention, bladder urgency, excessive sweating and dysphagia demonstrated a significantly different distribution of NMF between the two motor states with a shift towards a more instable and more frequent occurrence of NMS in motor Off state compared to motor On state (Supplementary Table 2).

Next, we assessed the variability of severities of these ten NMS between different On states and different Off states on the individual subject level. We therefore used the range of home diary VAS scores from a total of five self-perceived motor On as well as Off states in each individual as a measure of motor symptoms and NMS severity variability with higher ranges representing higher variability (Table 2). Although the motor function was rated worse in Off compared to On state (63 ± 17% versus 39 ± 22% on VAS, P <  0.001, Wilcoxon test) [6], the intraindividual variability of motor dysfunction between the motor On and Off states did not show significant differences. Except for dysphagia, all NMS demonstrated a significantly higher intraindividual variability (higher ranges) in symptom severity during motor Off states. Using the Mahalanobis distance (MD) as a multidimensional measure of data distribution of all NMS severity ranges within each individual subject, we found that intraindividual NMS severity ranges were not significantly different between motor On and Off states (P = 0.392). Similar results were obtained for both motor states when using only morning ratings from before 12:00 am (not shown). Normalization of NMS severity to motor function of the respective motor state led to increased intraindividual severity variability of all NMS on both motor states (Supplementary Table 3). We did not detect differences between the patients with complete and incomplete diary.

DISCUSSION

Although severity and variability of NMF and their correlation with motor complications have been described before [3, 6], there are no reports on intraindividual variability of NMS presentations within defined motor states. By using a systematic assessment of key NMS during a series of five motor On and five Off periods [6], we detected a moderate intraindividual variability of NMS frequencies and severities between multiple motor On and Off states, with 13–52% instable fluctuators for motor On and 24–47% instable fluctuators for motor Off state. Frequencies of instable fluctuators were higher for motor Off compared to motor On states for anxiety, depression and bladder urgency, but not for other NMS.

The factors influencing intraindividual NMF variability remain largely enigmatic. Similar to the findings in group analyses showing no or only weak correlations of NMF with motor function or disease state and no associations with demographic data [6–9], we did not detect any associations of intraindividual NMF instability and demographic or clinical characteristics. Indeed, normalization of NMS severity to motor function for each motor On or Off state even increases NMF variability in the present study. NMF variability in the present study seems to be largely independent from circadian influences, since the ratings were performed mainly in the morning and no differences in the time of rating were detected between the various motor states. Future studies systematically investigating the timing of NMS and NMF are thus warranted to further characterize the circadian rhythm of NMS and their fluctuations.

Several limitations of the present study need to be addressed. First, NMS in our study were rated during self-perceived motor states, hence we cannot exclude that these motor states were different from investigator-controlled On and Off states. Nevertheless, this study reflects highly relevant clinical data on everyday life of PD patients since self-reported outcomes are more likely to provide a realistic impression of individual NMS burden. Second, complete diary datasets were available from only 38% of patients, most likely due to diary fatigue or higher stress burden by diary entries in the older and more advanced patients [10]. Although we did not detect any relevant differences between completers and non-completers, our data should be interpreted only for the younger and less advanced fluctuating PD population (see Supplementary Table 1. Third, we confined our report to motor On state without distinguishing between “motor On” and “On with dyskinesias” [6], since most patients have difficulties judging their hyperkinetic conditions [11], which are known to influence at least some NMS, such as, e.g., dysphagia [12] and pain [5]. The addition of a caregiver diary to adequately evaluate motor states might thus be of additional benefit in future studies. Fourth, we used the range of VAS values from 5 different time points as an aggregate value for the intraindividual variability of NMS. Although the test-retest reliability of VAS is reported to be high for most symptoms [13–15], we cannot exclude bias due to limited intra-rater agreement for the severity analyses of some NMS.

Our study reports moderate intraindividual variability in NMF with respect to defined motor states. Psychiatric NMF were found to be particularly instable concerning symptom frequency and severity, particularly in motor Off states. The variability and thus unpredictability of NMF might contribute to the huge impact of NMF on hr-QoL [3, 6]. Further research on NMFs in PD therefore needs to address NMS on a longitudinal basis to fully encompass their impact on hr-Qol and to draw conclusions for necessary treatment adjustments such as continuous dopaminergic stimulation.

CONFLICT OF INTEREST

The authors have no conflict of interest to report.