ABSTRACT Objectives This study aimed to assess biochemical differences in peri‐implant crevicular fluid (PICF) using ATR‐FTIR spectroscopy and multivariate analysis, through a between‐participant comparison of healthy and peri‐implantitis (PI) sites, and a within‐participant comparison of PI‐PICF and GCF from contralateral periodontal healthy teeth. Materials and Methods Samples were categorized into three groups based on peri‐implant status and sampling site: PICF from peri‐implant healthy sites, PICF from PI‐affected sites, and GCF from contralateral healthy teeth in PI patients, with n = 20/group. Samples were collected and analyzed through ATR‐FTIR spectroscopy. Chemometric models were applied for cluster/outlier identification and discrimination. Regression coefficient vectors of PLS‐DA models identified key spectral differences between groups. Results A total of 40 implants were analyzed; participants had a mean age of 62 years, predominantly female (59%), non‐smokers (88%), and most received regular maintenance therapy (82%). PCA revealed subtle cluster formations without clear group separation. Comparing PICF from patients with and without PI, the most important wavenumbers for the PLS‐DA model were within 1700–1680 cm −1 , 1190–1130 cm −1 , and 1050–980 cm −1 , associated with protein‐related signals and nucleic‐acid content. Comparing the GCF and PICF from PI patients, using a split‐mouth design, the most important wavenumbers were 3006 cm −1 , 2982 cm −1 , and 2900 cm −1 , related to lipid structures. In terms of accuracy, between‐participant assessment achieved 75%, whereas the split‐mouth assessment reached 83.4%. Conclusions FTIR‐spectroscopy combined with chemometric modeling effectively discriminates peri‐implant health and PI in both inter‐ and intra‐subject comparisons. Higher within‐subject accuracy supports FTIR's potential as a site‐specific, non‐invasive diagnostic tool for peri‐implantitis.
Maligno et al. (Thu,) studied this question.