Results

A total of 31 patients completed the study, with 15 in the PBM group and 16 in the control group. The PBM group had a mean age of 46.8 ± 9.3 years, while the control group presented a similar age profile, with a mean age of 48.4 ± 8.5 years. The mean body mass index (BMI) was also comparable between groups, with 26.5 ± 2.5 kg/m² in the PBM group and 26.2 ± 2.0 kg/m² in the control group. Regarding ethnicity, 81.2% of participants in the PBM group and 86.7% in the control group self-identified as white. Tobacco exposure (current or former smoking) was reported by 18.8% of participants in the PBM group and 40.0% in the control group. The average length of follow-up with the surgical team was 25.3 ± 16.4 months in the PBM group and 27.3 ± 16.0 months in the control group.

No adverse events, complications, or PBM-related side effects were observed during the study period.

Pain perception was evaluated on postoperative days 2 and 7. On day 2 (T1), the PBM group reported a mean pain score of 2.7 ± 2.1, compared to 3.8 ± 2.7 in the control group. By day 7 (T2), mean pain levels had decreased in both groups, reaching 1.2 ± 1.3 in the PBM group and 1.4 ± 1.4 in the control group.

Descriptive statistics for each biomarker are presented in Table 1. Among the analyzed salivary biomarkers (TNF-α, IL-10, nitrite, and TBARS), no statistically significant changes were observed within groups over time for TNF-α, IL-10, or nitrite, nor between groups at any specific time point. However, a significant difference was detected for TBARS between the PBM and control groups at 48 h postoperatively (T1) (p = 0.034) in the unadjusted analysis. Baseline TBARS levels did not differ significantly between the PBM and control groups (Mann–Whitney U = 74.0, p = 0.072), although numerically higher values were observed in the PBM group. To account for this baseline variability, TBARS levels at T1 were further evaluated using a regression model adjusted for baseline TBARS (T0). In the adjusted model, baseline TBARS was a significant predictor of TBARS at T1 (β = 0.79, 95% CI: 0.52–1.07, p < 0.001), whereas the effect of PBM was not statistically significant (β = −0.053, 95% CI: −0.242 to 0.135, p = 0.567). Additionally, the change in TBARS from baseline to 48 h (ΔTBARS) did not differ between groups (Mann–Whitney U = 114.0, p = 0.828).

Biomarker | Timepoint | Group | Mean ± SD | Median [IQR]
TNF | T0 | PBM (n = 15) | 19.29 ± 8.35 | 14.35 [12.42–25.95]
TNF | T0 | Control (n = 16) | 20.45 ± 7.83 | 20.75 [12.74–22.79]
TNF | T1 | PBM (n = 15) | 18.10 ± 6.67 | 13.57 [12.65–23.12]
TNF | T1 | Control (n = 16) | 32.19 ± 54.66 | 19.09 [12.44–22.82]
TNF | T2 | PBM (n = 15) | 17.09 ± 6.48 | 12.58 [12.40–22.66]
TNF | T2 | Control (n = 16) | 18.26 ± 5.38 | 19.23 [12.61–21.53]
IL10 | T0 | PBM (n = 15) | 6.13 ± 8.77 | 4.00 [3.60–4.39]
IL10 | T0 | Control (n = 16) | 6.48 ± 10.07 | 3.77 [3.59–4.18]
IL10 | T1 | PBM (n = 15) | 6.06 ± 8.82 | 3.76 [3.63–4.29]
IL10 | T1 | Control (n = 16) | 7.22 ± 12.33 | 4.06 [3.54–4.69]
IL10 | T2 | PBM (n = 15) | 9.04 ± 13.47 | 3.88 [3.49–4.38]
IL10 | T2 | Control (n = 16) | 3.92 ± 0.67 | 3.68 [3.53–4.12]
Nitrite | T0 | PBM (n = 15) | 2.03 ± 1.77 | 1.67 [1.37–2.13]
Nitrite | T0 | Control (n = 16) | 2.99 ± 4.69 | 1.74 [1.31–2.54]
Nitrite | T1 | PBM (n = 15) | 2.36 ± 1.79 | 2.11 [1.81–2.28]
Nitrite | T1 | Control (n = 16) | 2.56 ± 1.77 | 2.30 [1.43–2.83]
Nitrite | T2 | PBM (n = 15) | 1.99 ± 0.77 | 2.08 [1.39–2.37]
Nitrite | T2 | Control (n = 16) | 2.05 ± 0.83 | 1.99 [1.37–2.45]
TBARS | T0 | PBM (n = 15) | 0.51 ± 0.45 | 0.23 [0.22–0.81]
TBARS | T0 | Control (n = 16) | 0.25 ± 0.11 | 0.22 [0.22–0.23]
TBARS | T1 | PBM (n = 15) | 0.50 ± 0.49 | 0.24 [0.22–0.61]
TBARS | T1 | Control (n = 16) | 0.24 ± 0.05 | 0.22 [0.22–0.23]
TBARS | T2 | PBM (n = 15) | 0.36 ± 0.40 | 0.22 [0.22–0.25]
TBARS | T2 | Control (n = 16) | 0.24 ± 0.07 | 0.22 [0.22–0.24]

The Friedman test showed no significant temporal variation in TNF-α and IL-10 levels (p > 0.05). TBARS levels showed a small reduction from T0 to T1 in the PBM group (p < 0.05), although the magnitude of this change was small and between-group differences were not significant after baseline adjustment. Pain scores significantly decreased over time in both groups (p < 0.05).

Between groups (PBM vs. Control), effect sizes were generally small for all biomarkers across the three time points (Cohen’s d ranging from 0.04 to 0.36), supporting the absence of clinically relevant differences between groups. Within-group analysis showed that TBARS exhibited reductions of small magnitude between T0 and T1 in both groups (PBM: dz = − 0.04; Control: dz = − 0.11), indicating that the paired contrast alone did not demonstrate a consistent effect, despite the global significance detected by the Friedman test.

In contrast, postoperative pain showed large reductions between T1 and T2 in both groups (PBM: dz = − 1.25; Control: dz = − 1.32; overall effect dz = − 1.24), consistent with a robust clinical improvement over time. These findings indicate that salivary oxidative stress biomarkers showed only minor changes during the early postoperative period, while pain scores decreased substantially over time in both groups, consistent with the expected postoperative recovery trajectory.

Correlation analysis

Spearman correlation analysis revealed no statistically significant associations between baseline biomarkers and postoperative pain scores (p > 0.05). However, some moderate, non-significant trends were observed. A moderate positive correlation was identified between TNF-α and IL-10 (r = 0.34; p = 0.061). Additionally, weak-to-moderate positive correlations were observed between TBARS and pain on postoperative day 2 (r = 0.24; p = 0.188) and on day 7 (r = 0.26; p = 0.151).

Pain scores on days 2 and 7 were significantly correlated with each other (r = 0.64; p < 0.001), suggesting consistency in individual pain perception throughout the postoperative recovery. Overall, these findings indicate that inflammatory and oxidative stress biomarkers at baseline are not strongly associated with short-term clinical outcomes such as postoperative pain.

No correction for multiple comparisons was applied due to the exploratory pilot design; therefore, findings with marginal p-values should be interpreted cautiously. These findings should therefore be interpreted as exploratory observations rather than confirmatory evidence of treatment effects.

Multiple linear regression analysis

Multiple linear regression analyses were conducted to evaluate whether age, BMI, previous surgeries, bariatric surgery, and physical activity influenced salivary levels of TNF-α, IL-10, TBARS, and nitrite at baseline (T0).

For TNF-α, the model had low explanatory power (adjusted R² = −0.037), with no statistically significant predictors (p > 0.05). Although not statistically significant, BMI (β = 0.89; p = 0.225) and physical activity (β = 3.30; p = 0.361) showed positive trends.

The IL-10 model demonstrated an adjusted R² of − 0.139, and none of the predictors reached statistical significance. Age showed a weak and non-significant negative association (β = −0.15; p = 0.493).

For TBARS, the adjusted R² was − 0.043, again indicating limited explanatory value of the predictors. BMI (β = 0.05; p = 0.119) and age (β = −0.0067; p = 0.402) showed weak, non-significant trends.

In contrast, the model for nitrite yielded a higher adjusted R² (0.045), and physical activity was a statistically significant negative predictor (β = −3.09; p = 0.047). Participants who reported engaging in physical activity within the previous six months had significantly lower nitrite levels compared to sedentary individuals.

Overall, the models suggest that demographic and clinical variables had limited ability to explain the variance in inflammatory and oxidative stress markers. However, physical activity may play a role in modulating nitrite metabolism, potentially reflecting more efficient regulation of nitric oxide pathways.