15 patients with PD and 15 healthy age- and sex-matched individuals underwent scintigraphic measurements of gastric emptying after ingestion of a standardized radiolabeled test meal. All participants gave their written informed consent before their participation. The study was approved by the Regional Committee on Health Research Ethics of the Capital Region of Denmark (H-18054923). 14 PD patients and 12 control subjects had previously participated in an experiment, during which postprandial changes in SMA blood flow in response to a standardized liquid test meal intake had been measured with dynamic PC-MRI. These results were recently reported in another publication using a larger cohort of 23 patients with PD and 23 age- and sex-matched healthy control participants (23). We also included one PD patient and three healthy control subjects who had only agreed to participate in scintigraphic assessment of gastric emptying. The recruitment procedures and exclusion criteria have previously been described in detail by Siebner et al. (23). Notably, PD patients and healthy control participants were included regardless of the presence or degree of gastrointestinal symptoms. Patients had to be older than 50 years and be diagnosed with PD by a movement disorder specialist according to the Movement Disorder Society Clinical Diagnostic Criteria for Parkinson's Disease (24).

The demographic and clinical data are summarized in Table 2. Routine dopamine transporter SPECT had been performed, confirming a reduction in striatal dopamine transporter density in all PD patients. All patients received anti-parkinsonian medication. Three patients were only treated with levodopa/decarboxylase inhibitor. In nine patients, levodopa therapy was combined with a dopamine agonist (n = 8) or COMT inhibitor (n = 1). The remaining three patients were treated with a dopamine agonist and MAO-B inhibitor (n = 1), dopamine agonist, levodopa and MAO-B-inhibitor, (n = 1) or dopamine agonist, levodopa and COMT inhibitor (n = 1). Levodopa equivalent daily dose (LEDD) was calculated based on patients' anti-parkinsonian medication using the conversion factors suggested by Tomlinson et al. (25). Nine PD patients and one healthy control participant received prescribed laxatives daily (Macrogol, Sodium Picosulfate), which was continued throughout the study.

Scintigraphic measurements were performed with PD patients in an “ON-medication” state to avoid motor discomfort and to secure comparability with previous PC-MRI measurements. Participants were instructed to fast for solids and liquids for at least 8 hours prior to scintigraphy. After arrival and a couple of minutes of rest participants ingested a standardized meal consisting of 80 g bread, 30 g jam and 120 g omelet, which was radiolabeled with 40 MBq ^99mTc human albumin colloidal particles (Nanocoll), and 120 ml water (26). The participants were given 10 minutes to intake the standardized meal before the first scintigraphic imaging.

Two-minute images were obtained at 0, 30, 60, 120, 180, and 240 minutes after finishing the standardized radiolabeled test meal in the anterior and posterior projections (Figure 1). Dynamic scintigraphy was carried out with a dual-head gamma camera equipped with low energy, general purpose collimators (Symbia Intevo, Siemens Healthcare GmbH, Erlangen, Germany) at the Department of Clinical Physiology and Nuclear Medicine, Copenhagen University Hospital Amager and Hvidovre. Regions of interest for integration of radioactivity were drawn manually on all images by one of two experienced investigators (JLM and MD). The obtained counts were corrected for time decay. Geometric mean values of anterior and posterior counts were used for attenuation correction. For each participant, the gastric retention percentage values measured at a given time point (y-axis) were plotted against time (x-axis). The area under the curve (AUC) was calculated using the trapezoidal rule and expressed as retention percent x hour (27). In this study the AUC was used as summary metric of gastric emptying. Gastric emptying was incomplete after 240 minutes in three participants.

Blood flow and blood glucose measurements were based on data from a subgroup of 14 PD patients and 12 healthy control subjects, which have been published recently in a larger cohort (23). These subjects participated in serial phase-contrast MRI-based measurements of blood flow in the superior mesenteric artery interrupted by interspersed blood glucose measurements in response to a standardized test meal (~400 kcal) intake, and subsequently participated in the present scintigraphic study. The time elapsed between PC-MRI measurements and scintigraphy ranged between 2 weeks and 7 months. For a detailed description of blood flow and blood glucose measurements, we refer to Siebner et al. (23). We tested for between-group differences in magnitude and time to maximum of the postprandial increase in SMA blood flow and glucose levels.

Participants completed the Non-Motor Symptoms Questionnaire (NMS-Quest) describing subjective non-motor symptoms on arrival at the day of MRI measurements. The NMS-Quest consists of 30 items grouped into nine domains (28, 29). The score ranges from minimum 0, indicating absence of non-motor symptoms, to maximum 30, indicating presence of all non-motor symptoms in question. Recall period was the month before participation. The four study subjects, who did not participate in MRI measurements, completed the NMS-Quest on arrival at the day of scintigraphic measurements.

Statistical analysis was performed using R software (RStudio Inc., Version 1.2.1335). Normality of the demographic and clinical group data was tested using Shapiro-Wilk test reported as mean ± standard deviation or median and 10%- & 90%-quantiles as appropriate in Table 2. Unpaired two-samples t-test was used to test for differences between groups in case of normal distribution. Between-group differences were analyzed using Wilcoxon rank-sum test when normal distribution could not be assumed. Statistical significance was accepted at p < 0.05.

To assess whether the postprandial blood flow was associated with gastric emptying, we tested for a relationship between the magnitude and time to maximum of the postprandial increase in SMA blood flow and gastric emptying calculated as AUC using Spearman's rank correlation (p < 0.025, Bonferroni corrected for multiple comparisons). We performed additional correlational analyses for exploratory purposes, testing for relationships between individual gastric emptying and age, blood glucose measurements, NMS-Quest score, disease duration and LEDD, using Spearman's rank correlation and applying an adjusted significance level of p < 0.0083 as statistical threshold after Bonferroni correction.

At the group level, gastric emptying calculated as AUC was not different between patients and healthy controls (p = 0.68, Table 2). The dynamics of gastric retention at the individual and group level are presented separately for PD patients and healthy participants in Figure 2, showing substantial inter-subject variability in both groups. While the individual slopes of gastric retention over time varied considerably, the distribution of gastric retention over time was very similar between groups, resulting in almost identical AUC in PD patients and healthy controls (Table 2). The only exception was a single PD patient who showed a markedly increased gastric retention at all measurements and met the definition of delayed gastric emptying (26). This patient responded well to antiparkinsonian drug treatment, had a 5-year disease duration, was successfully treated with Macrogol for constipation and reported no upper gastrointestinal dysfunction symptoms, why we had no reason to question the diagnosis of probable PD. Apart from this patient, all participants showed <2% gastric retention 4 h after meal intake. A boxplot presentation of the individual gastric retention in PD patients and healthy controls shows the same pattern, confirming a very similar distribution of postprandial gastric retention of the ingested meal in the PD and healthy control groups (Figure 3).

In the 14 PD patients and 12 healthy subjects who had participated in PC-MRI prior to scintigraphic measurements, the absolute postprandial increase in SMA blood flow was attenuated in PD compared to healthy controls (p = 0.006, Table 2). A statistical trend toward an attenuated postprandial increase in blood glucose level was also observed in the patient group compared to the healthy control group (p = 0.08, Table 2). There was no significant between-group difference in time to reach maximal blood flow and time to reach maximum blood glucose level. In summary, the main findings of our previous study in a larger cohort (23) were still present in this subsample.

Our correlation analysis revealed no significant relationships between gastric emptying as indexed by AUC and the magnitude or the time to maximum of the postprandial increase in SMA blood flow (Table 3). Exploratory analyses showed a trend toward significant positive correlation between gastric emptying and NMS-Quest score in PD patients (rho = 0.57, p = 0.03) based on Bonferroni correction. A longer gastric emptying was associated with a higher subjective prevalence of non-motor symptoms. Hence, the longer the individual gastric emptying, the higher the individual NMS-Quest score. Also, exploratory analyses showed that individual gastric emptying did not correlate with age, blood glucose measurements, disease duration or LEDD (Table 3).

When studying gastric function in PD, the influence of the age of participants, disease severity and medication state needs to be considered. Although the available literature is contradictory regarding the effect of healthy aging on gastric emptying, recent studies support age-related slowing of gastric emptying (31, 32). The impact of aging on gastric emptying may also apply to PD patients. A previous scintigraphy study reported a prolongation of gastric emptying in PD patients in an “ON-medication” state, both with response fluctuations after long-term levodopa therapy and without response fluctuations, compared to a healthy control group (13). Since healthy control participants were on average more than 15 years younger than the patients with PD, an effect of age on gastric emptying may account for the between-group difference. Another study using an electromagnetic capsule system showed that gastric transit time was longer in patients with PD compared with healthy subjects who were on average more than 40 years younger (22). The authors suggested that the longer gastric transit time could be due to three outliers but did not exclude an age-related decrease in gastric motility. In fact, a previous study using a similar electromagnetic capsule system showed no difference in gastric transit time in PD compared to age- and sex-matched controls (33), pointing to a relevant influence of age on gastric emptying. Future studies need to systematically examine how healthy aging impacts gastric emptying and how aging influences gastric emptying in PD. In our cross-sectional study, gastric emptying did not correlate with age (Table 3), but the relatively narrow age range of our study population does not allow us to dismiss an effect of aging on gastric emptying. Taken together, there is prevailing evidence against a prolongation of gastric emptying in patients with PD in the “ON-medication” state, when patients are compared to age-matched groups of healthy individuals.

Only one previous scintigraphy study has investigated the influence of disease severity on gastric emptying in PD patients. Hardoff et al. compared gastric emptying between patients with mild PD (Hoehn and Yahr stage 1–2) and patients with moderate PD (Hoehn and Yahr stage 2.5–3.0) and found no significant difference, however including PD patients in different medication states (14). In our study, we included PD patients with Hoehn and Yahr stages 2 and 3 in relatively early stages of the disease (mean disease duration of 3.9 ± 2.2 years; LEDD of 599 ± 256 mg) and only one patient had delayed gastric emptying. As mentioned, we had no reason to question this patient's diagnosis of probable PD. Neither Hardoff et al. nor our study did examine gastric emptying in later disease stages of PD. We can therefore not exclude that impaired gastric emptying might be present in a smaller subgroup of PD patients in the first years after the emergence of parkinsonism. Delayed gastric emptying may be more prominent in later disease stages and needs to be investigated in future studies.

Erratic delivery of levodopa from the stomach to its duodenal and jejunal absorption sites due to delayed gastric emptying has been suggested to be a causative factor for inconsistent response to levodopa and consequently motor fluctuations in PD (1–4, 9, 10). PD patients with response fluctuations after long-term levodopa therapy showed delayed GET compared to non-fluctuating patients (13). Another study found a significant relationship between levodopa pharmacokinetics and gastric emptying and could show that delayed gastric emptying was more common in PD patients with a late plasma levodopa peak (34). In addition, motor fluctuations and dyskinesia in PD patients were considerably improved by continuous jejunal levodopa infusion and thereby bypassing gastric emptying (35). Interestingly, subthalamic deep brain stimulation seems to improve gastric emptying (36, 37). Therefore, the emergence of dysfunctional gastric emptying during disease progression warrants further studies.

We studied PD patients under therapy to secure comparability with previous PC-MRI measurements. Furthermore, we expected the PD patients to be stressed and having more difficulties to cooperate during the MRI session, if they had been in an “OFF-medication” state. Therefore, it remains unclear whether dopaminergic medication affects gastric emptying or vice versa. Several studies showed that a single oral dose of levodopa inhibits gastric emptying in both young and elderly healthy subjects (38–40). However, this might not to be the case in patients with PD, as no difference in breath test GET was found in PD patients before and approximately 1.5 years after the initiation of levodopa treatment (41). Another scintigraphy study observed no significant difference in GET between levodopa-treated and untreated PD patients, and the same was evident when dividing the patient group in mild and moderate PD (14). Lastly, a scintigraphy study showed that PD patients in a defined “OFF-medication” state (withdrawn from anti-parkinsonian medication for 12 h) had significantly faster mean gastric emptying than controls (16). Taken together, there are apparent discrepancies in the reported findings with some studies pointing to an interaction between gastric emptying and levodopa. Even though our current findings are in line with the most recent study, which observed no delayed GET in PD patients in the “ON-medication” state (18), the interaction between levodopa treatment and gastric emptying is still not fully understood and needs further investigation.

Gastric emptying can also be indirectly tested using the ¹³C-octanoate breath test, which measures the concentration of ¹³CO₂ expired in the lungs after intake of a meal containing ¹³C-sodium-octanoate. In contrast to the scintigraphy studies described above, most studies employing the ¹³C-octanoate breath test reported abnormal results in PD patients (11). Since ¹³C-sodium-octanoate is absorbed in the jejunum and metabolized in the liver to ¹³CO₂, the ¹³C-octanoate breath test is not only dependent on mechanical stomach emptying, but also on small intestinal motility, ¹³C-octanoate absorption and hepatic metabolism (11, 42). It is worth noting that small intestinal absorption has been shown to be deranged and small intestinal transit time to be increased in patients with PD (33, 43). Hence, abnormal small intestine absorption and motility rather than delayed gastric emptying may account for the “abnormal” ¹³C-octanoate breath test findings and normal GET as revealed by the most recent gastric scintigraphy study (18).

Post-hoc exploratory analysis revealed that gastric emptying tended to scale positively with the prevalence of non-motor symptoms in the group of PD patients. This might suggest that PD patients with higher subjective prevalence of non-motor symptoms might be in greater risk of suffering delayed gastric emptying than others. However, the NMS-Quest does not include upper gastrointestinal symptoms apart from dysphagia, and gastric emptying was on average not different from healthy controls. Future studies need to validate whether gastric emptying may be useful as a marker of overall non-motor involvement.

This study is a follow-up examination and was prompted by our recent postprandial phase-contrast MRI study (23). Since 14 out of 15 patients and 12 out of 15 healthy controls had been taken part in serial PC-MRI measurements of postprandial blood flow in the SMA, we were able to correlate the gastric emptying measurements with the blood flow measurements of the SMA. Even though we were only able to reexamine a subgroup of the original cohort, we still found the significant attenuation of postprandial increase in SMA blood flow and a trend toward a significant attenuation of postprandial increase in blood glucose level in this reduced sample size of the original cohort. The time from food intake to the postprandial peak in SMA blood flow was not different between groups. We have previously discussed in detail the possible mechanisms behind the impaired functional regulation of postprandial gastrointestinal perfusion in PD (23). We had considered that changes in motility, impaired glucose absorption as well as impaired neural control of gastrointestinal perfusion may be affected in PD and thereby explain the attenuated postprandial response in both blood flow and glucose. This consideration prompted this follow-up examination. Our aim was to correlate the gastric emptying measurements with the previous blood flow measurements of the SMA and thus, to clarify whether changes in gastric emptying could explain the attenuation of the postprandial SMA response. In this sub-cohort, gastric emptying did not correlate with the postprandial blood flow response in the SMA. Gastric emptying correlated neither with postprandial absolute blood flow increase nor with time to reach maximal blood flow in SMA. Albeit our sample size was limited, we tentatively conclude that the attenuated postprandial blood flow in SMA previously seen in PD patients is not related to individual alterations in gastric emptying. In this context, it is important to note, that the SMA supplies the gastrointestinal tract from lower part of the duodenum to left colic flexure (44), but we did not examine small intestinal or colonic gastrointestinal motility in this study. We can therefore not exclude that gastric emptying in PD may be related to changes in postprandial blood flow response in the arteries supplying the stomach. When piloting the PC-MRI study, it was not possible to establish reliable blood flow measurements in the celiac artery or left gastric artery supplying the stomach, spleen, and liver (44). This precluded correlation analyses between postprandial blood flow dynamics from the arteries suppling the stomach with gastric emptying in the current study. Future studies need to measure postprandial blood flow in both the celiac artery with its branches and SMA, and concurrently examine the temporal dynamics of gastric, small intestinal and colonic motility in order to correlate blood flow in the supplying arteries with the temporal dynamics of the respective sections of the gastrointestinal tract.

Given the small sample size and methodological differences in the meal test (e.g., type of meal, body position during measurement), the overall clinical significance of impaired functional regulation of postprandial mesenteric blood flow response in PD remains still to be clarified. Future studies need to clarify whether an attenuated postprandial SMA blood flow response reflects gastrointestinal dysfunction. Longitudinal study designs will be needed to link changes in postprandial SMA blood flow to the course of disease, especially whether it can be useful as progression marker in PD or in prodromal PD.

In this study, gastric emptying function was not affected in a group of PD patients, although gastrointestinal symptoms are frequently reported in PD (1–4). Symptoms may thus be related to impaired function of the middle and lower portion of the gastrointestinal tract. In fact, previous studies showed significantly delayed small intestinal and colonic transit times in PD patients compared to healthy control subjects (18, 33). The impaired gastrointestinal motility in the small intestine and colon could be a contributory mechanism to the attenuated postprandial blood flow in SMA and glucose response seen in PD patients in our study. Also, the abnormal small intestine motility in PD could together with the deranged absorption explain why most ¹³C-octanoate breath tests show delayed gastric emptying in PD in conflict with studies employing gastric scintigraphy (42, 43). Motility impairment of the lower gastrointestinal tract might more be decisive for subjective symptoms. Conversely, the attenuated postprandial blood flow response in SMA may not contribute substantially to clinical gastrointestinal symptoms but might be an independent marker for gastrointestinal involvement in PD.