Targeted therapeutic management of diabetes using phytoconstituents: molecular mechanisms, evidence map (2015–2025), and translational outlook

Background: The root cause of diabetes is dysregulated pathways, including those involving AMP-activated protein kinase (AMPK), GLUT-mediated glucose transport, and the PI3K/AKT pathway. There has been a notable increase in research on phytoconstituents as pathway-specific treatments for diabetes; however, the comprehensiveness of this evidence remains unclear. Methods: This systematic review followed PRISMA guidelines and was registered on PROSPERO (CRD420251073083). Databases searched included PubMed, Scopus, Google Scholar, and Europe PMC for experimental studies (in vivo, in vitro, and in silico) published between 2015 and 2024. The final search was conducted in April 2025, and 2025 publications available as “early access” before this date were included. Only English-language studies were included. Animal studies (in vivo) were assessed for risk of bias using the SYRCLE tool, while in vitro studies were evaluated using the ToxRTool, based on test substance characterization, test system description, study design, and data reporting. Narrative synthesis was employed due to the heterogeneity of the data. Results: Out of 3,222 articles, 177 articles met the inclusion criteria. Study types included in vitro (92; 52%), in vivo (66; 37.3%), in silico (15; 8.5%), and other experimental types (4; 2.3%). Phytoconstituents predominantly targeted PI3K/AKT (44.6%), GLUT transporters (19.8%), and AMPK (14.1%) pathways. Rodent models were most used (48.02%). Primary outcomes included improved insulin sensitivity, enhanced glucose homeostasis, and reduced oxidative stress and inflammation. The risk of bias analysis revealed 68.93% of the studies carried a moderate risk, 29.94% a low risk, and 1.13% a high risk. Discussion: Phytoconstituent activity was consistent with the activation of diabetes-relevant signaling pathways, particularly PI3K/AKT, GLUT transporters, and AMPK cascades. However, most evidence was correlative, with limited loss-of-function validation. Methodological irregularities, moderate risk of bias, and limited translational research reduce the strength and generalizability of these findings.