Introduction

Parkinson’s disease (PD) is the second most crucial neurodegenerative disease worldwide. Globally, nearly 10 million persons live with PD, and its prevalence is expected to increase in the upcoming years¹.

Since the seventies, the relationship between impaired glucose metabolism, elevated insulin levels, and PD has been studied in cellular, animal, and human models. Thus, suggesting a link between PD and type 2 diabetes mellitus (T2DM)². Clinical evidence shows that T2DM in people living with PD (PwP) is associated with a more aggressive disease course³,⁴. In addition, in such cases with T2DM diagnosed before PD, PwP had more significant impairment in consequent motor symptoms⁵. Furthermore, persons with PD-DM have an earlier presentation of motor complications⁶.

Studies reporting the relationship between PD and T2DM are vast. However, evidence is scarce when addressing the effect of T2DM onset on the age of PD onset. This study aims to identify whether T2DM contributes to a different age of PD onset.

Methods

An observational, cross-sectional, and analytical study was carried out. The study protocol was approved by the Institutional Review Board and by the Local Ethics Committee (121/19). All participants gave their written Informed Consent.

Consecutive PwP attending the Movement Disorders Clinic at the National Institute of Neurology and Neurosurgery in Mexico City from 2018 to 2020 was recruited. The International Parkinson and Movement Disorders Society clinical criteria were used for diagnosing PD⁷. Only patients with a disease duration between 2 and 4 years were included. This range was selected based on the disease progression model proposed by Holford et al.,⁸ allowing to study the disease when a more predictable progression was expected, and symptom overlap between the two conditions was less problematic.

T2DM diagnosis was defined by at least one of the following: a positive personal history of T2DM recorded in the medical files, a previous diagnosis reported by the subject, or the use of a hypoglycemic drug for glycemic control. The age of T2DM onset was determined as recorded in the medical files or as reported by the subject.

Patients with incomplete demographic or clinical data were excluded from the study. A 2:1 ratio randomization from the non-DM sample was performed. This randomization aimed to reduce the risk of errors resulting from comparing groups with highly unequal sample sizes, especially when parametric assumptions were violated. Unequal sample sizes can lead to unequal variances between samples which affect the assumption of equal variance in some statistical tests, thus increasing the risk for Type 1 error as well as loss of statistical power⁹. On the other hand, increasing the control-to-case ratio in unmatched case-control settings, results in a gain of statistical power until a ratio of 1:4 and then stabilizes thereafter¹⁰; in our study, samples were matched according to a disease duration between 2 and 4 years, thus a matching ratio higher than 2:1 may still have substantial power loss given that T2DM was rare (< 15%) in an under-lying cohort¹¹.

The age of PD onset represented the age when each subject perceived their first motor symptom. For clinical assessment, a movement disorders specialist evaluated motor symptoms. The Movement Disorder Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) was employed for assessing non-motor experiences of daily living (Part I), motor experiences of daily living (Part II), motor evaluation (Part III), and motor complications (Part IV)¹². The Hoehn and Yahr (HY) stages were used to classify PD as mild (stages 1 and 2), moderate (stage 3), and severe (stages 4 and 5).

414 PwP were initially recruited. Seventy-six subjects were categorized as PD-DM, 14 of which were excluded due to T2DM diagnosis after PD onset. Consequently, a total of 62 PD-DM participants were incorporated into the analyses.

From the remaining 338 non-DM PD subjects, 124 were randomly assigned to match a 2:1 ratio, as described above. For the randomization, the function Random Sample of Cases from SPSS was used (Data > Select Cases > Random Sample of Cases). Description of sociodemographic data was done with measures of central tendency (modes, medians, and means) and dispersion ranges (standard deviations and variances). For the inferential analyses, the tests used were as follows. The Pearson’s and Spearman’s coefficients were used to determine correlations. Comparisons of quantitative variables were performed using independent samples Student’s t-test, Welch’s t-test (unequal variances), or Mann-Whitney test, as needed. The Chi-square test and Fisher’s test were used when comparing qualitative variables. p < 0.05 was considered for statistical significance. The statistical package SPSSv25.0 was used.

Results

Overall, 186 PwP (54.3% male) were analyzed. The mean age of the whole sample was 63.4 ± 11.9 years, with a mean PD duration of 3.4 ± 0.8 years. In the PD-DM group, the mean duration of T2DM was 12.4 ± 6.8 years, and T2DM was diagnosed 9.2 ± 6.8 years before the PD onset. No statistical differences in demographic and clinical variables between included and excluded subjects were found.

Regarding T2DM therapy, metformin was the most commonly used treatment (74%), both as monotherapy or as an add-on to glibenclamide (11%). In addition, 14% of the subjects were on a premixed insulin regimen.

Table 1 compares the main variables between the non-DM and the PD-DM groups. In summary, subjects in the PD-DM group were older (mean difference 5.7 ± 1.8, 95% CI: 2.2-9.3 years) and had an older age of onset (mean difference 5.9 ± 1.6, 95% CI: 2.7-9.0 years) in comparison to the non-DM group. No statistically significant differences between groups were found in motor and non-motor scores or disease duration. A statistically significant but weakly positive correlation was found between the T2DM course and the age of PD onset (rs = 0.27, p = 0.03).

Discussion

A link between PD and insulin resistance has been formerly described, suggesting a common neurodegenerative pathway¹³. Animal models have demonstrated that hyperglycemia inhibits dopaminergic neuron activity and lessens levels of extracellular dopamine¹⁴.

On the other hand, T2DM drugs such as glucagon-like peptide-1 receptor (GLP-1R) agonists, thiazolidinediones, and dipeptidyl-peptidase 4 (DPP4) inhibitors have been proposed as potential neuroprotectors or disease modification therapies in PD based on epidemiological studies and in vitro models¹⁵.

Epidemiological studies had reported T2DM as a risk factor or a protective factor for PD, depending on the study design. Prospective studies have shown an increased risk for PD¹⁶, while case-control studies (retrospective) describe a protective role¹⁷.

In this study, the age of PD onset was older in the DM-PD group, which may suggest a neuroprotective role in a certain stage of the pathogenesis of the T2DM. This finding was unexpected due to the substantial evidence suggesting hyperglycemia and insulin resistance as a catalyst of mitochondrial dysfunction, oxidative stress, and inflammation leading to neurodegeneration¹⁸.

A recent meta-analysis on T2DM as a determinant of PD risk and progression failed to find age as a relevant factor. Interestingly, only one of the seven cohort studies included in the meta-analysis T2DM was required to be developed before PD, with the remaining studies also including incident T2DM cases. Age was not investigated in the two case-control studies analyzed¹⁹.

A possible factor that could be responsible for the older age at the PD onset in the PD-DM group could be the hypoglycemic treatment. In our cohort, the most used therapy was metformin. Metformin has shown a neuroprotective role in several neurodegenerative diseases, as well as in PD²⁰. Metformin may have a potential role in almost every aspect of PD physiopathology, resulting in a possible protective factor for the development of PD²¹.

In the present study, neither motor nor HY differences were found. The fact that our groups were controlled by a rather short PD progression may explain these findings.

Several limitations can be listed. First, subjects on the PD-DM were older. Arguably, the age difference might create a bias translating into the age of PD onset. While this cannot be ruled out, the PD duration in both groups is similar. Therefore, subjects in the PD-DM did not have a longer follow-up due to a longer PD duration leading to a bigger chance of developing diabetes. Age-matching between groups using the whole sample before randomization was attempted and was not feasible. Second, no biomarkers such as insulin or HbA1c were collected. In consequence, the relationship between insulin metabolism or glycemic control and its implication in the age of PD onset could not be appraised. Third, the fact that three-quarters of the PD-DM group were on metformin and that there were no patients on GLP-1R agonists, thiazolidinediones, or DPP4 inhibitors does not allow to address the possible effect of drug treatment on the age of PD onset. Finally, due to the study, design recall bias is expected.

Furthermore, the comorbidity burden of T2DM may include other conditions that have been associated with the risk of developing PD. For instance, it has been reported that the use of statins in the context of dyslipidemia decreases the risk of PD²², as well as the use of beta-blockers commonly used to treat hypertension²³. Other commonly sees metabolic disturbance seen in metabolic syndrome is in serum uric acid levels; low uric acid has been associated with morbidity, severity progression, non-motor symptoms, and motor complications of PD²⁴. Due to the study design, these potential confounders or effect modifiers were not assessed and future studies must also consider them.

Likewise, considering that T2DM is one of the most common illnesses among our population, as well as the increasing prevalence of PD worldwide, additional longitudinal research should be conducted to determine if this phenomenon on age of PD onset might be attributed to T2DM metabolic pathogenesis, but also taking into account the role of T2DM treatment and glycemic control periodical measurements. In the future, these could provide valuable information that would furthermore help prevent or delay PD onset.

Conclusion

Patients with DM-PD had an older age at PD onset, suggesting a potential T2DM role in delaying the PD age of onset. Further, research should be conducted to identify if this phenomenon can be attributed to T2DM metabolic pathogenesis but also to T2DM treatment and glycemic control.