Comment on “The impact of diet on Parkinson's disease risk: A data-driven analysis in a large Italian case-control population”

To the Editor:

I read with interest the recent article by Gigante et al. examining dietary patterns and Parkinson's disease (PD) risk in a large Italian case-control population. ¹ The authors should be commended for addressing an important and difficult question using a relatively large multicenter sample and a data-driven dietary approach. Their finding that sweets, red meat, and processed meat were positively associated with PD, whereas fruit intake appeared protective, adds to a literature that remains heterogeneous and methodologically challenging. At the same time, I believe several aspects of interpretation warrant greater caution.

First, the authors appropriately acknowledge a core limitation of retrospective case-control research: the difficulty of establishing the temporal sequence between exposure and disease onset. This issue is particularly important in PD, where the prodromal phase may extend for many years before diagnosis^2,3 and where recall-based assessment of presymptomatic diet is inherently vulnerable to bias. ⁴ Under such circumstances, observed associations are valuable for generating hypotheses, but they cannot by themselves support strong causal inferences.

Second, the interpretation of the sweets factor deserves especially careful treatment. The article presents higher sweet consumption as a possible risk factor for PD, but the authors also report that sweet-factor scores correlate positively with disease duration. That observation raises the possibility that increased sweet intake may reflect disease-related behavioral change, at least in part, rather than a pre-diagnostic causal exposure. This interpretation is biologically plausible. Altered reward processing, appetite, and food preference may emerge during prodromal or early PD, and several reviews have noted that craving for sweet foods in PD remains incompletely understood and may relate to underlying disease mechanisms rather than diet acting solely as an upstream cause. ⁵ For this reason, the sweets association seems particularly vulnerable to reverse causality and should be framed more cautiously.

Third, the reported association between general anesthesia and PD risk also requires restrained interpretation. The odds ratio reported for general anesthesia was notably large, yet the discussion itself recognizes an alternative explanation: exposure to anesthesia may be linked to medical or surgical procedures undertaken for prodromal PD-related conditions. This is important because PD is now widely understood to have a prolonged prediagnostic phase characterized by autonomic, gastrointestinal, urinary, and balance-related symptoms.^2,3 Such manifestations could reasonably increase healthcare utilization and the likelihood of surgery before formal diagnosis. In that setting, prior general anesthesia may function less as an independent causal exposure than as a surrogate marker of prediagnostic disease burden. In epidemiologic terms, this is compatible with protopathic bias. Existing literature on anesthesia exposure and later alpha-synucleinopathy has not established a clear causal relationship after consideration of surgical context and indication,^6,7 which further supports a cautious interpretation.

Taken together, these issues suggest that the observed associations may be shaped not only by potentially relevant exposures, but also by the natural history of PD and by the intrinsic limitations of retrospective exposure assessment.^1–4 I therefore suggest that the article's conclusions be interpreted primarily as hypothesis-generating. In particular, the sweets association should be discussed in light of possible reverse causality,^1,5 and the general anesthesia association in light of possible protopathic bias.^2,3,6,7

Future work could help clarify these relationships. Prospective cohort designs, incident-case analyses, and stratification by the timing and indication of surgical procedures would all strengthen causal interpretation. Such analyses would be especially helpful in distinguishing true modifiable risk factors from markers of prodromal disease.^2,3 I believe these steps would sharpen the contribution of this otherwise valuable study and help prevent over-interpretation of findings that remain biologically and temporally complex.