Abstract

Introduction

Parkinson’s disease (PD) is a neurodegenerative disease characterized by intracellular inclusions of misfolded α-synuclein¹. The clinical diagnosis of PD is based on the cardinal motor symptoms bradykinesia, rigidity, and/or rest tremor², associated with a critical loss of dopaminergic neurons in the substantia nigra. Degeneration of the nigrostriatal dopaminergic system can be confirmed in vivo by brain imaging with ¹²³I-N-ω-fluoropropyl-2β-carbomethoxy-3β-{4-iodophenyl}nortropane (¹²³I-FP-CIT) single photon emission computed tomography (SPECT) (dopamine transporter (DAT) SPECT) or with ι-3,4-dihydroxy-6-[¹⁸F]fluorophenylalanine positron emission tomography (¹⁸F-FDOPA PET). Patients with PD typically have an asymmetric onset of motor features which start in the bodyside contralateral to the predominant dopaminergic deficit. This asymmetry becomes less prominent over the course of the illness³. However, approximately 20% of de novo PD patients have a symmetric disease onset⁴.

In PD, brain changes extend beyond the nigrostriatal dopaminergic pathway and consequently involve several neuronal networks, causing various non-motor symptoms⁵. Functional changes in these networks can be examined using ¹⁸F-Fluorodeoxyglucose PET (¹⁸F-FDG PET). ¹⁸F-FDG PET studies have identified a PD-related pattern (PDRP) characterized by relatively increased glucose metabolism in the pallidum, thalamus, pons and cerebellum, and relative hypometabolism of the premotor cortex, supplementary motor area, and parietal association regions⁶. PDRP expression can be quantified and has been observed to increase with disease progression and decrease with symptomatic treatment^7,8. Only modest correlations exist between PDRP expression and striatal binding ratios measured by ¹⁸F-FDOPA PET or DAT SPECT imaging, indicating that the PDRP reflects more widespread functional brain changes^9–11. In contrast to dopaminergic changes, PDRP expression shows no asymmetry, even in PD patients with strictly unilateral symptoms¹².

Despite clinical asymmetry being well documented, its origin and role in disease pathology are unclear. Some studies have suggested that the asymmetric onset of PD is not random but is directed by brain lateralization (i.e. hand dominance)^13,14. Explanations have been sought amongst structural and biochemical hemispheric differences that may cause specific unilateral vulnerabilities for PD pathology¹⁵.

The issue of asymmetry is a focal point in the recently proposed “brain-/body-first” (BBF) hypothesis by Borghammer and colleagues¹⁶. This hypothesis delineates two primary subtypes of PD. In the body-first subtype, the aggregation of α-synuclein is posited to initiate in the gastrointestinal tract, progressing via the vagus nerve to the brainstem, substantia nigra, and eventually the cerebral cortex. This subtype demonstrates an extended prodromal phase due to the involvement of multiple synapses in the gut-to-brain progression. During this phase, REM sleep behavior disorder (RBD) emerges, coinciding with the involvement of the locus coeruleus/subcoeruleus complex in the brainstem. The presence of RBD prior to motor symptoms or dementia onset is considered a reliable indicator of the body-first subtype. Conversely, the brain-first subtype is proposed to originate unilaterally, for instance, in the amygdala¹⁷. Notably, this progression involves substantial damage to the substantia nigra before observable damage to the autonomic system or the onset of RBD symptoms. RBD may manifest in later stages of brain-first PD, following the spread of the pathology from one hemisphere to the brainstem (top-down propagation). Borghammer and colleagues introduced and expanded upon the BBF hypothesis, proposing the α-Synuclein Origin site and Connectome (SOC) hypothesis in 2021^16,17. An essential premise of this updated BBF hypothesis is that α-synuclein propagation is contingent upon the structural brain connectome. Consequently, the SOC hypothesis seeks to predict and explain the (a)symmetrical presentation of clinical symptoms and dopaminergic degeneration in PD. Specifically, in the body-first subtype, the pathological spread is theorized to be symmetrical, facilitated by the overlapping motor innervation of the right and left dorsal vagal nerves within the gastrointestinal tract. This concept draws support from anatomical evidence in animal studies¹⁸. Conversely, in the brain-first subtype, the pathology primarily spreads to ipsilateral brain structures, guided by intra-hemispheric brain connections, which account for 90% of total brain connections. This pattern results in a more asymmetric propagation of the pathology.

Cross-sectional neuropathological data^19,20, along with multi-modality in-vivo imaging²¹ have provided support for the BBF hypothesis. However, a limited number of studies have tested the validity of its predictions on asymmetry in distinct clinical cohorts of patients presumed to exhibit brain-first or body-first PD. Using ¹⁸F-FDOPA PET imaging, Knudsen et al. examined individuals with iRBD alongside de novo PD patients, both with and without RBD. They classified the patients as body-first and brain-first subtypes based on the presence or absence of iRBD. Thus, the iRBD group (n=21) was labeled as prodromal body-first PD. Among PD patients, those with RBD at least two years before the onset of motor symptoms (PDRBD+, n=11) were classified as the body-first subtype, while those without RBD (PDRBD-, n=22) were considered the brain-first subtype²². The authors found a more symmetric nigrostriatal degeneration in iRBD and PDRBD+ compared to PDRBD-, in line with the BBF hypothesis. In a similar design, Banwinkler et al. (2022)²³ used a questionnaire-based assessment of iRBD presence and (a)symmetric nigrostriatal dopaminergic degeneration (measured with ¹²³I-FP-CIT SPECT) as proxies to distinguish between brain-first and body-first subtypes in a cohort of 255 de novo PD patients. Their investigation focused on comparing hemispheric gray matter loss between these subtypes. Surprisingly, the study did not uncover discernible patterns of brain atrophy between the presumed brain-first and body-first PD subtypes. Additionally, no correlation was observed between the asymmetry in putaminal dopaminergic denervation and the asymmetry in gray matter volume. These results challenge the notion that the spread of pathology throughout the brain aligns with patterns observed in the dopaminergic system, thus questioning the concept of the spread of pathology based on brain connectome principles.

It is tempting to use a degree of asymmetry as a marker to identify PD subtypes in clinical and research settings. However, first, it is crucial to investigate whether asymmetry is indeed associated with the suggested brain/body-first subtypes. In this retrospective study, we compare the degree of asymmetry in putamen DAT binding (¹²³I-FP-CIT SPECT), PDRP expression, regional brain glucose metabolism and motor scores between presumed body-first and brain-first PD. Based on the SOC theory, individuals without RBD, indicating the brain-first subtype (PDRBD-), would exhibit greater asymmetry in putamen DAT binding, hemispheric PDRP expression, and regional glucose metabolism (particularly in the amygdala). Conversely, we anticipate that individuals with RBD as a prodromal phase (body-first subtype, PDRBD+), as well as those presenting with iRBD (presumed prodromal phase of body-first PD), would reveal more symmetric findings in these imaging markers.

Results

Putamen DAT binding

Lowest specific to non-displaceable binding ratio (SBR), putamen (relative) asymmetry index (AI_put), and the absolute difference between left and right SBR are shown in Fig. ​Fig.1.1. Lowest putamen SBRs were similar in PDRBD+ compared to PDRBD− (p=0.7, median 0.93 vs. 0.82). There was no significant difference in AI_put between PDRBD+ and PDRBD− (p=0.97, median 0.13 vs. 0.14).

Open in a separate window

Fig. 1

Dopaminergic and motor asymmetry.

Lowest putamen SBR (a) and putamen DAT asymmetry index (b) plotted per group, with medians indicated by horizontal bars.One outlier for the putamen in each PD group is indicated by a black star, AI_put=1.06 for the PDRBD+ and AI_put=0.69 for the PDRBD− patient. The iRBD patients with an abnormal DAT scan (n=9) have been plotted in dark blue, with the rest of the iRBD patients shown in light blue (n=8). Additionally, c shows the absolute difference between left and right putamen SBR, and d the difference in left and right UPDRS-III motor scores (here, the total iRBD group, n=20, is shown).

17 out of the 20 iRBD patients underwent DAT SPECT imaging, where the group was split based on the DAT scan being evaluated as either normal (iRBD_norm, n=8) or abnormal (iRBD_path, n=9). The median lowest putamen SBR was 2.4 in iRBD_norm and 1.6 in iRBD_path, with the difference being significant (p<0.0001). We did not directly compare the SBRs between the iRBD and PD groups due to differences in camera and processing pipelines. Furthermore, we did not compare the relative asymmetry index AI_put between groups with significantly (or apparently) differing SBRs. The motivation for this is further treated in the Discussion, see also Supplementary Fig. 3 in the Supplementary Information. Instead, we statistically evaluated the differences in absolute asymmetry.

Considering the absolute difference between the highest and lowest putamen SBR (Figure 1c), the median was 0.33 for PDRBD+, 0.22 for PDRBD-, 0.23 for iRBD_path and 0.05 for iRBD_norm. There was, again, no significant difference between PDRBD− and PDRBD+ considering this alternative measure for asymmetry (p=0.4). Notably, iRBD_path was not significantly different from either PDRBD+ (p=0.08) or PDRBD− (p=0.93) with this measure. iRBD_norm had a significantly lower absolute difference than PDRBD+ (p=0.002) but was not significantly different from PDRBD− (p=0.06) or iRBD_path after Bonferroni correction (p=0.03).

We used an exploratory cutoff of AI_put=0.08 to estimate the fraction of patients showing symmetric versus asymmetric degeneration in each group, corresponding to the upper quartile reported in AI_put for HC subjects of the PPMI database in ref. ²². Considering this cutoff, all subjects in iRBD_norm fell below the threshold, while 33% in iRBD_path, 68% in PDRBD+ and 63% in PDRBD− were asymmetric.

We repeated the analysis considering only those PDRBD+ patients with a reported onset of RBD at least two years before the onset of motor symptoms (n=14). This subgroup had a median AI_put of 0.11 and a median lowest putamen SBR of 0.98, none of which were significantly different compared to the PDRBD− group (p=0.9 and 0.8, respectively). There was also no significant correlation between AI_put and duration of RBD symptoms or motor symptoms for any of the PD groups (p>0.2, see Supplementary Fig. 4 of the Supplementary Information).

There was no difference between men and women in AI_put, either with the groups considered together or separately (p≥0.05). There was a weak correlation between AI_put and age for the PDRBD+ group (ρ=0.51, p=0.02). The absolute difference in left and right putamen SBR was not correlated with age for any group (p>0.1). For PDRBD- and iRBD there was no correlation with age (p≥0.3). The absolute difference between left and right putamen SBR did not correlate with the lowest putamen SBR considering the PD patients together (p=0.4), while the relative AI_put correlated with the lowest putamen SBR by ρ=−0.63,p=0.00002.

Hemispheric PDRP expression

There was no significant difference in AI_PDRP between the iRBD, PDRBD+ and PDRBD− groups (p≥0.1 for all; see left panel of Fig. ​Fig.2).2). Each disease group also did not have a higher value of AI_PDRP than their corresponding HC cohort (p>0.05 for all). Similarly, there were no significant between-group differences in the absolute value of AI_PDRP (p≥0.1 for all), emphasizing between-group similarity also in the asymmetric tails of the distributions.

Open in a separate window

Fig. 2

PDRP asymmetry.

a z-scored difference between left and right hemispheric PDRP expression, defined as AI_PDRP. The median for each group is indicated by a horizontal bar. b z-scored PDRP expression in the hemisphere with higher and lower putamen DAT binding (least and most affected hemisphere).

There was no significant difference in PDRP z-scores between the most affected hemisphere (MAH, lowest putamen SBR) and the least affected hemisphere (LAH, highest putamen SBR) (p>0.5 for all groups, see Fig. ​Fig.2,2, right panel). LAH and MAH PDRP z-scores correlated by ρ=0.93,p=1e⁻³¹.

While most patients had a symmetric PDRP expression in each hemisphere, some could be considered asymmetric: AI_PDPR>2 for 3 iRBD (17%), 4 PDRBD+ (18%), and 3 PDRBD− patients (19%). This corresponded to an absolute difference between LAH and MAH z-scored PDRP expression of at least 1.1 for the iRBD patients and 0.8 for the PD patients.

Motor asymmetry

The distributions of motor asymmetry index AI_motor, as described by the absolute difference between left and right scores of the UPDRS-III scale can be seen in Figure ​Figure1d.1d. There was no difference in AI_motor between PDRBD+ and PDRBD− (median 2 vs. 4, p=0.3). Both PD groups were more asymmetric than the iRBD patients (p<0.02 for both).

A summary of the asymmetry indices in the disease groups is shown in Table ​Table33.

Table 3

Summary of median asymmetry indices in our patient groups

AI	iRBDa	PDRBD+	PDRBD−
AIput	0.01 [0.00, 0.02]/0.07 [0.04, 0.08]	0.13 [0.07, 0.23]	0.14 [0.07, 0.23]
SBRL−SBRR	0.05 [0.02, 0.08]/0.23 [0.15, 0.29]	0.33 [0.14, 0.44]	0.22 [0.11, 0.44]
AIPDRP	0.84 [0.50, 1.47]	0.55 [0.34, 1.50]	1.09[0.41, 1.72]
AImotor	1 [0.0, 1.0]	2 [1.0, 4.8]	4 [2.8, 5]

Open in a separate window

The values are reported as median [IQR].

^aFor the DAT SPECT-related asymmetry measures, the results are reported separately for the iRBD patients with normal versus abnormal DAT scans.

AI_motor correlated with AI_put (ρ=0.36,p=0.007) considering the full set of patients. AI_put did not correlate with AI_PDRP, apart from for the PDRBD− group (ρ=0.56,p=0.02), even though this significance seemed to be driven by a single asymmetric outlier. The p-value changed to p=0.26 when removing the outlier. The same p-value for PDRBD+ was 0.8. AI_put did not correlate significantly with any AI_ROI after the Bonferroni correction.

Discussion

Our data does not support the hypothesis that degeneration of the dopaminergic system is more symmetric in body-first PD compared to brain-first PD¹⁷. We studied 22 de novo PD patients with RBD (PDRBD+; confirmed with video-polysomnography) who reported dream-enactment behavior at least six months before the onset of motor symptoms. We also repeated the analysis for a subset of patients (n=14) who reported RBD symptoms at least 2 years before the onset of motor symptoms. According to the SOC theory, these patients belong to the proposed body-first subtype of PD. Sixteen de novo PD patients without RBD were included to represent the proposed brain-first PD subtype (PDRBD−). Asymmetry indices of striatal DAT binding were not significantly different between the body-first (PDRBD+) and brain-first (PDRBD−) groups. In fact, the majority of PD patients (roughly two-thirds) had asymmetric putamen DAT binding, irrespective of RBD status.

Individuals with iRBD are considered to have a prodromal stage of body-first Lewy body disease. According to the SOC-theory^16,17, iRBD patients should have symmetric degeneration of the nigrostriatal system. In our study, nine out of 17 iRBD patients had an abnormal DAT scan (their putamen DAT binding ratio was 2 or more standard deviations below age-matched control values). At the group level, these nine iRBD subjects were not different from PDRBD+ or PDRBD− in terms of the absolute difference between left and right putamen SBR. Moreover, three had clearly asymmetric DAT scans (AI_put>0.08), and one subject in particular showed very asymmetric putamen DAT binding (AI_put=0.34, which corresponds to the top 20% of asymmetric PD patients in our cohort).

Our findings contradict the results from the study by Knudsen et al.²², who found significantly more asymmetry in their PD^−RBD (n=22) versus their PD^+RBD (n=11) group, using ¹⁸F-FDOPA PET. The authors also found that DAT SPECT scans in iRBD patients (n=25) were significantly more symmetric compared to de novo PD (n=419, no distinction between subtypes) from the PPMI database. There are a few plausible explanations for the difference in findings between the two studies. First, a possible contributing factor is the difference in sample size. Considering the substantial spread we observed in AI_put for both our PD groups, small sample sizes may not be sufficiently representative of the whole population. Second, Knudsen et al. combined their 11 PD^+RBD patients with 12 iRBD patients with a pathological ¹⁸F-FDOPA PET scan to increase the sample size, forming a combined RBD+ group. This could also have artificially decreased the mean AI_put in their RBD+ group, as illustrated by a decrease in median AI_put compared to PD^+RBD, and an increase in p-value from p=0.001 when comparing RBD+ to PD^−RBD to p=0.049 when comparing PD^+RBD to PD^−RBD.

Third, it can be noted that the DAT asymmetry index AI_put, being a relative index, features exponential growth with lower SBR (explaining the correlation we observed between AI_put and lowest putamen SBR). This advises caution in its interpretation. Consider two patients with an identical absolute difference in putamen SBR. The patient with overall lower binding ratios will receive a higher AI_put compared to the patient with overall higher binding ratios. We illustrate this in Supplementary Fig. 3 in the Supplementary Information. In the study by Knudsen et al.²², mean putamen SBR was significantly lower in the PD^−RBD group compared to the PD^+RBD group. This could have overestimated AI_put in their PD^−RBD group, compared to considering an absolute difference. Therefore, we additionally decided to consider the absolute difference between left and right for putamen SBRs. This measure, as well, showed no between-group differences for our PD groups.

Additionally, differences in clinical characteristics between the two studies should be considered. Our PDRBD+ patients exhibited a shorter total RBD duration, ranging from IQR=[2.3,6.5] years, compared to Knudsen’s cohort, IQR=[4.8,16.3] years. A shorter duration of (prodromal) RBD symptoms could imply that some of our PDRBD+ patients might have experienced RBD symptoms and early motor changes more or less simultaneously. This makes the body-first status of our PDRBD+ patients less certain. That being said, we also repeated the analysis for a subset of patients (n=14) who reported RBD symptoms at least 2 years before the onset of motor symptoms, and our results did not change. Additionally, there was no correlation between total RBD duration and AI_put for the PDRBD+ group (see Supplementary Fig. 4). We emphasize that in our PDRBD− group, in which the absence of RBD was confirmed carefully with video-PSG, we also observed cases with symmetric dopaminergic degeneration. We believe that this reinforces our conclusion.

Furthermore, the median duration of motor symptoms was 12 months for both our PDRBD+ and PDRBD− groups, while it was 17 and 21.5 months, respectively, for Knudsen’s²². Studies show that dopaminergic asymmetry becomes less prominent as the disease progresses^3,26,27. Here, Cao et al. found that the degeneration in subregions of the striatum was faster in PD patients with probable RBD, compared to PD without probable RBD, except for in the most affected side of the putamen. This implies that if the putaminergic asymmetry would be similar around the time of onset of motor symptoms, PD patients with RBD possibly progress faster into a relatively more symmetric state. Still, a difference of less than a year in the duration of motor symptoms most likely does not have any major effect on the asymmetry analysis between our study and Knudsen’s²².

Finally, the two studies used different imaging modalities. In the present study, DAT SPECT images were analyzed automatically by validated software without anatomical co-registration to individual magnetic resonance imaging (MRI) scans. This is in contrast to Knudsen et al., who used high-quality imaging with ¹⁸F-FDOPA PET and MRI co-registration²². The present study may suffer from lower spatial resolution and signal-to-noise ratio (SPECT vs. PET), and a larger partial volume effect (automatic delineation of volumes of interest). However, it seems unlikely that this will largely influence the determination of the asymmetry index. In fact, many DAT SPECT studies over the last 25 years have been able to capture the asymmetric striatal DAT binding, which is typical for the majority of PD patients^28–32. To illustrate, we have provided examples of individual DAT SPECT scans from both cameras in Supplementary Figs. 5 and 6 in the Supplementary Information.

For both the cohort in Knudsen et al. and ours, the PDRBD+ group was significantly older than PDRBD−. Striatal DAT binding is known to decrease with age³³, and the lowest putamen SBR was indeed negatively correlated with age in our cohort. AI_put was, in turn also correlated with age for the PDRBD+ group. However, this leads directly from the dependence of AI_put on the lowest putamen SBR. The absolute difference in left and right putamen SBR was not correlated with age, and we, therefore, assume that the difference in age between our PDRBD+ and PDRBD− groups did not have a major effect on the overall dopaminergic asymmetry analysis.

In addition to very specific changes in the nigrostriatal dopaminergic system, we investigated the degree of asymmetry in a more global measure of brain function: cerebral ¹⁸F-FDG uptake. The radiotracer ¹⁸F-FDG is analogous to glucose and provides an index for the first step of the cellular glycolytic pathway. This is often considered a proxy for neuronal and synaptic function, but the cellular source of ¹⁸F-FDG uptake is complex and reflects the metabolic activity of astrocytes and neurons³⁴, and perhaps microglia³⁵. In neurodegenerative diseases, the changes in ¹⁸F-FDG brain uptake occur in a disease-specific pattern, following the brain networks that are affected. Such patterns can be identified with spatial covariance analysis, e.g. the PDRP³⁶.

Contrary to the local, asymmetric dopamine depletion in the striatum, the PDRP represents broader functional changes in the underlying pathophysiological pathways of PD. We calculated PDRP scores in each hemisphere in our iRBD, PDRBD+, and PDRBD− groups. Because brain-first PD is suggested to have a more asymmetric intra-cerebral pathology propagation in early disease stages, we expected the metabolic consequence of those changes (PDRP expression) to be asymmetric as well. This would be in contrast to body-first PD subjects, which would have a more symmetric magnitude of PDRP expression in both hemispheres. This was not the case. There was no significant difference between presumed brain- and body-first PD subjects in terms of asymmetry in PDRP expression. At the group level, we also did not find any significant differences between the most- and least affected hemisphere (stratified by the lowest- and highest putamen SBR, respectively). These results are in line with the study by Tang et al.¹², who described symmetric PDRP expression in PD patients in both early (H&Y stage 1) and more advanced (H&Y 2–3) stages. In line with Huang et al.³⁷, PDRP expression in our study was also symmetric in the iRBD group. It must be noted, however, that we found a large degree of spread in the data, both in controls and patients. Although most patients had a similar degree of asymmetry as controls, there were subjects in each patient group that had a significantly higher PDRP expression in one hemisphere.

Because the PDRP is a sum score of all voxels, small but relevant asymmetries may have been obscured. We, therefore, also investigated which regions deviated from the natural asymmetry that was present in controls. Contrary to our hypothesis, we did not find any deviating regions in the PDRBD- (brain first) group. We did find regional deviations in the iRBD group and, to a lesser extent, in the PDRBD+ group, but the meaning of these findings is difficult to interpret. The regions that were found (Table ​(Table2)2) are all components of the PDRP network. How and why natural (a)symmetry is lost in these regions is unclear.

Our study can be considered in relationship to recently published results by Zhou et al.³⁸. They combined clinical and imaging data and identified two distinct trajectories in PD using a data-driven approach. The obtained trajectories closely resemble the “brain-first" and “body-first" subtypes, showing specific early and advanced manifestations for each, whereas the subtype resembling “body-first" featured RBD as an early symptom. Considering the modest accuracy of the RBDQ-HK questionnaire used, it would be interesting to see a similar (future) study using PSG-based RBD markers. The study found no significant differences in the asymmetry of free water in the basal forebrain, amygdala, entorhinal cortex, and hippocampus or in the integrity of the substantia nigra and locus coeruleus (using neuromelanin-sensitive MRI measures). These results support that the BBF theory effectively explains various clinical features and disease progression stages in PD. However, the enigma of asymmetry in brain pathology remains unresolved, warranting further investigation.

This study has limitations. Our data is cross-sectional and only describes a snapshot of the disease evolution. Longitudinal data is needed to verify whether the asymmetry concept of the SOC theory may hold at a specific disease stage. It is theoretically possible that PDRBD− patients would have been more asymmetric than PDRBD+ cases had both groups been scanned at an earlier disease stage. Longitudinal data of ¹⁸F-FDG PET and DAT SPECT scans will become available in an ongoing longitudinal study of RBD patients³⁹. However, the prodromal stage of PDRBD− is ill-defined, and it would be challenging to identify such subjects in the general population. This study also shows that studying asymmetry in neuroimaging data is challenging. In the previous paragraph, we have indicated that relative asymmetry indices are nonlinear and should probably not be compared between groups with differing degrees of degeneration. Furthermore, there is also a natural asymmetry in ¹⁸F-FDG uptake in the majority of brain regions in healthy controls. With regard to our healthy control group, we recognize the absence of biomarkers such as amyloid quantification, capable of stratifying healthy controls based on the risk of developing a neurodegenerative disease. Future studies should consider incorporating such measures to exclude controls that might be in a prodromal phase of neurodegenerative disease, thereby mitigating potential bias in the analyses. This is particularly important when biomarkers of nonspecific neurodegeneration, such as ¹⁸F-FDG, are involved.

Furthermore, we did not have other imaging markers available to further verify the brain or body-first status of our PD subjects due to the retrospective nature of this study²¹. An important example is ¹²³I-metaiodobenzylguanidine (MIBG) scintigraphy, which assesses the sympathetic nervous system. Reduced MIBG uptake in the heart confirms the involvement of the autonomic nervous system (i.e. ’the body’). We assessed autonomic symptoms using the SCOPA-AUT questionnaire and found no significant differences between the two groups, consistent with previous studies²¹. This reinforces the importance of biologically quantifying the peripheral nature of the degeneration (e.g. MIBG scintigraphy).

Finally, we consider that iRBD subjects have prodromal PD or DLB. However, a minority of iRBD patients (<5%) develop multiple systems atrophy (MSA)⁴⁰. Theoretically, our iRBD cohort could include MSA cases. Here we presented the baseline data of our iRBD cohort as a reference³⁶. To date, none of our patients have developed MSA, but longer-term follow-up would be needed to substantiate that all our iRBD patients indeed had prodromal Lewy-body disease. Given the small percentage of MSA-converters, we do not believe that the inclusion of MSA cases would significantly alter the results.

It should be noted that iRBD patients develop PD or DLB in approximately equal numbers⁴⁰. In turn, most DLB patients suffer from RBD, and they typically have abnormal MIBG scans and more severe autonomic dysregulation. Therefore, Borghammer et al. posit that DLB and PD are fundamentally the same diseases and that DLB patients fall in the body-first PD category¹⁶, and further suggest a role for Alzheimer’s co-pathology to understand different subtypes within the brain-first/body-first framework²⁰. Two studies have shown that dopaminergic degeneration is indeed more symmetric in DLB compared to PD^41,42, but these subjects did not confirm body/brain-first status with PSG or other markers. Whether this is indeed related to the routes of propagation of pathology, as suggested by the SOC-theory, remains to be investigated. Results of a study by Arnaldi et al. call this into question⁴³. They indicate that patients with iRBD and asymmetric DAT binding in the caudate nucleus had a higher risk of developing DLB, whereas those with more symmetric caudate nucleus DAT binding were at a higher risk of developing PD.

It remains unclear what aspects of asymmetry are important, how they relate to PD pathology, and how we can meaningfully interpret numerical indices of asymmetry. This study found no relationship between AI_put and age, sex, duration of RBD or motor symptoms, or absolute difference between left and right and lowest putamen SBR. It must also be noted that presynaptic dopaminergic imaging and cerebral glucose metabolism are not direct markers of the underlying pathology (i.e. α-synuclein aggregation) but reflect the functional brain changes in response to pathology.

In summary, our results show that there is no significant difference in asymmetry between presumed brain-first and body-first PD in terms of presynaptic dopaminergic imaging and cerebral glucose metabolism. Dopaminergic changes are asymmetric in most patients (roughly two-thirds) but symmetric in some. PDRP expression is bilateral in most iRBD, PDRBD+ and PDRBD− patients, but some cases are asymmetric. The SOC hypothesis is compelling and may provide a framework for the pathophysiology of PD. It should be emphasized that our data do not falsify this model but call into question the assumption that body-first PD is more symmetric than brain-first PD. The most important implication of this study is that the degree of (a)symmetry on presynaptic dopaminergic imaging alone should not be used as a proxy to classify patients as brain- or body-first PD in future studies.

Methods

Acknowledgements

Author contributions

Contributions (1) Research project: A. Conception, B. Organization, C. Execution; (2) Statistical analysis: A. Design, B. Execution, C. Review and critique; (3) Manuscript preparation: A. Writing of the first draft, B. Review and critique, (4) A. Final approval of the completed version, B. Accountability for all aspects of the work. S.S.L.: 1A, 1B, 1C, 2A, 2B, 3A, 4A, 4B. G.C.: 1A, 1B, 1C, 2C, 3A, 3B, 4A, 4B. B.O.: 1B, 1C, 3B, 4A, 4B. M.B.: 2C, 3B, 4A, 4B. D.A.: 1A, 1B, 1C, 3B, 4A, 4B. P.M.: 1B, 1C, 3B, 4A, 4B. A.J.: 1B, 1C, 3B, 4A, 4B. E.S.: 1B, 1C, 3B, 4A, 4B. S.M.: 1A, 1C, 3B, 4A, 4B. J.B.: 1C, 3B, 4A, 4B. W.H.O.: 1A, 2B, 3B, 4A, 4B. K.L.L: 1A, 2B, 3B, 4A, 4B. S.K.M: 1A, 1B, 1C, 2C, 3A, 3B, 4A, 4B. S.S.L. and G.C. developed the methodology and main parts of the research project together.

Data availability

Our data is available upon request. Researchers can contact Sanne Meles ([email protected]) concerning the data collected at the University Medical Center Groningen and Dario Arnaldi ([email protected]) for the data collected at the IRCCS Ospedale Policlinico S. Martino.

Competing interests

J.B. is a consultant at GE Healthcare and received research grants (but not for the present study) from GE Healthcare. S.M. received speaker honoraria from GE Healthcare, Eli-Lilly, and Life Molecular Imaging. D.A. received speaker honoraria from Idorsia, Bioprojet, Bruno, Italfarmaco, Lundbeck, PIAM, Fidia, and Valeas. W.H.O. has received speaker’s honoria on educational symposia sponsored by Abbvie, the International Movement Disorders Society, and Stada Pharma. He acts as a consultant for Lario Therapeutics and is a member of advisory boards with Intrabio and MODAG and holds stock options with Intrabio and MODAG—all unrelated to this manuscript`s research. All other authors declare no competing interests.

Footnotes

References