A total of 219 patients who underwent sedated gastroscopy at our hospital between April and May 2026 were selected; all were classified as ASA Class I - II and aged between 18 and 74 years. Patients were enrolled consecutively and were allocated to three age groups to achieve equal sample size: the young group (aged 18 - 44 years) comprising 73 patients; the middle-aged group (aged 45 - 59 years) comprising 73 patients; and the young elderly group (60 - 74 years) comprising 73 patients.

Inclusion criteria:

1) Age 18 - 74 years;

2) Scheduled for sedated gastroscopy;

3) ASA classification I - II;

4) Informed consent from the patient and their family.

Exclusion criteria:

1) Severe peripheral vascular disease, such as Raynaud’s disease or occlusive arteriosclerosis;

2) Skin lesions on the fingers affecting perfusion index (PI) monitoring;

3) Severe hepatic or renal insufficiency, or coagulation disorders;

4) Allergy to anaesthetic agents such as propofol or sufentanil;

5) Impaired consciousness or inability to cooperate with the investigator.

This study was approved by the hospital’s Ethics Committee, and all patients signed informed consent forms.

Patients were fasted for 8 hours and had nothing to drink for 4 hours prior to surgery. Upon admission to the operating theatre, a peripheral vein was established and 500 mL of compound sodium chloride injection was administered. Oxygen was administered via a face mask at a flow rate of 6 L/min for at least 5 minutes.

Anaesthetic induction: Administer 0.1 - 0.2 μg/kg of sufentanil intravenously; 30 seconds later, slowly inject 1.5 - 2.5 mg/kg of propofol. Begin endoscopy once the MOAA/S score reaches ≤1.

Anaesthetic maintenance: Administer propofol 20 - 50 mg intermittently, as required, based on patient movement and swallowing reflexes, to maintain a stable level of sedation and vital signs.

Monitor and record data at the following time points:

T0: Pre-anaesthesia (5 minutes after entering the room in a supine position).

T1: Immediately after propofol injection.

T2: Upon completion of propofol injection.

T3: Loss of eyelash reflex.

T4: When the gastroscope passes the pyriform sinus.

T5: During the examination (recorded every 3 minutes, taking the average).

T6: Upon awakening (responsive to verbal stimuli, able to open eyes).

Primary analytical time points for this study were T0, T2, T3, and T6. T1, T4, and T5 were recorded for auxiliary observation but were not included in the main statistical analysis because T1 was transient and T4 - T5 belonged to the maintenance phase rather than the induction phase.

Parameters to be recorded: PI, SBP, DBP, HR, SpO₂.

Calculate the maximum change during the induction phase:

ΔPI = Maximum PI value − Baseline PI value

ΔSBP = Baseline SBP value − Minimum SBP value

ΔDBP = Baseline DBP value − Minimum DBP value

Criteria for hypotension: SBP < 90 mmHg or a decrease of >30% from baseline.

Criteria for bradycardia: HR < 50 beats/min.

If any of the above conditions occur, administer vasoactive drugs in accordance with standard protocols.

SPSS 29.0 statistical software was used. Continuous variables are expressed as mean ± standard deviation; within-group comparisons were performed using repeated measures analysis of variance (ANOVA), whilst between-group comparisons were performed using one-way ANOVA with Bonferroni post-hoc correction for multiple comparisons. Categorical variables are expressed as percentages (%), and the chi-square test was used for analysis. Correlations were assessed using Pearson’s correlation analysis. ROC curves were plotted to determine the optimal threshold for predicting hypotension using PI [9]. A p-value of <0.05 was considered statistically significant.

There were no statistically significant differences (P > 0.05) in gender, BMI, ASA classification, time of examination, or total propofol dosage among the three groups; the groups were therefore comparable.

Table 1. Comparison of general data among three groups ( x ¯ ± s , n (%)).

At T0, there were no statistically significant differences in PI, SBP, DBP, or HR among the three groups except for baseline SBP, which was significantly higher in the young elderly group (129.6 ± 12.1 mmHg) than in the young adult group (106.5 ± 10.2 mmHg) and the middle-aged group (108.3 ± 11.4 mmHg) (P < 0.05). Following induction of anaesthesia, PI rose rapidly, peaking between T2 and T3, before gradually declining and returning to near baseline levels during the recovery period [10]. SBP and DBP decreased significantly after induction, exhibiting a trend opposite to that of PI (P < 0.05). The changes in ΔPI, ΔSBP, and ΔDBP increased progressively with age: the young elderly group had the largest fluctuations, followed by the middle-aged group, and the young adult group had the smallest (P < 0.05). There were no significant intergroup differences in HR changes.

The maximum PI during the induction phase was significantly negatively correlated with ΔSBP and ΔDBP (r = −0.721 and −0.684, respectively; P < 0.01); ΔPI was significantly positively correlated with the incidence of hypotension (r = 0.653; P < 0.01); There was no significant correlation between PI and HR (P > 0.05).

As shown in Figure 1, the ROC curve showed: AUC for PI in predicting hypotension during the induction phase of anaesthesia = 0.824 (95% CI: 0.765 - 0.883); Optimal cut-off: 3.65%; Sensitivity: 78.3%; Specificity: 81.5% [11].

In the young-elderly group, the optimal threshold was 3.2%, with a sensitivity of 82.0% and a specificity of 79.5% [12].

Young group: 12.3%.

Middle-aged group: 17.8%.

Young-elderly group: 34.2%.

The differences between groups were statistically significant (P < 0.05).

Note on time points: T1, T4, and T5 were recorded but excluded from the main analysis because T1 was transient and T4 - T5 belonged to the maintenance phase rather than the induction phase. Anaesthetic depth was stable after T3, and PI did not change substantially thereafter.

Table 2. Comparison of the perfusion index (PI) across age groups at different time points (%, x ¯ ± s ).

Note: Legend: *Compared with the young group, P < 0.05; ^#Compared with the middle-aged group, P < 0.05. T0: Pre-anaesthesia; T2: End of propofol injection; T3: Disappearance of the eyelash reflex; T6: Upon awakening. ΔPI: Maximum change in perfusion index; ΔSBP: Maximum decrease in systolic blood pressure; ΔDBP: Maximum decrease in diastolic blood pressure. Same applies to all tables below.

Table 3. Comparison of SBP (systolic blood pressure) across age groups at different time points (mmHg, x ¯ ± s ).

Table 4. Comparison of DBP (diastolic blood pressure) across age groups at different time points (mmHg, x ¯ ± s ).

Table 5. Comparison of changes in induction-phase parameters and incidence of hypotension across age groups.

PI reflects peripheral vasodilation and sympathetic tone [13]. Propofol inhibits the sympathetic nervous system and dilates peripheral blood vessels, thereby enhancing the pulsatile blood flow signal and causing a rapid rise in PI. This study shows that the rise in PI precedes the drop in blood pressure, suggesting that PI may reflect changes in vascular tone at an earlier stage, which is consistent with the findings of domestic and international studies [14].

During highly stimulating phases of gastroscopy, such as passage through the pyriform sinus, there is a transient excitation of the sympathetic nervous system and mild vasoconstriction, causing a slight decrease in PI; during the recovery phase, as anaesthetic drugs are metabolised and vascular tone is restored, PI gradually returns to baseline levels. Changes in PI throughout the procedure are highly synchronised with anaesthetic depth, drug effects, and procedural stimuli [15].

This study confirmed that the magnitude of PI elevation is significantly negatively correlated with the magnitude of blood pressure reduction, suggesting that PI can serve as an indicator for assessing vasodilatory reserve. Patients with high baseline PI have lower vascular tone and are more prone to hypotension following anaesthesia [16]. In elderly patients, with reduced vascular elasticity and diminished autonomic regulatory capacity, PI fluctuations are more pronounced and the risk of hypotension is higher, which is consistent with clinical practice [17].

There was no significant correlation between PI and heart rate, suggesting that PI primarily reflects peripheral vascular tone rather than cardiac time-dependent function, and can serve as an independent supplementary indicator to conventional circulatory monitoring [18].

This study established that a PI of 3.65% can serve as the warning threshold for hypotension during painless gastroscopy in adults, with this threshold reduced to 3.2% for elderly patients [19]. Clinicians may use these values to implement stratified management:

1) PI ≥ 3.65%: Administer appropriate pre-procedural fluid expansion, slow the rate of propofol infusion, and reduce the induction dose;

2) Elderly patients with PI ≥ 3.2%: Exercise heightened vigilance and administer vasoactive agents prophylactically, if necessary;

3) Patients with lower PI values: Circulation is relatively stable; standard dosing may be administered.

PI monitoring requires no additional equipment, incurs no extra costs, and is simple to perform, making it suitable for implementation in endoscopy centres at all levels of hospital [20].

This study found that, when grouped by age, elderly patients exhibited more sensitive PI fluctuations and more fragile circulation, suggesting that anaesthetic management should be individualized [21]. Adopting lower warning thresholds, slower administration rates, and simplified anaesthetic protocols for patients aged 60 years and over can significantly improve safety, aligning with the principles of precision anaesthesia and patient-centred care [22].

PI is susceptible to factors such as finger temperature, limb movement, and sensor placement [23]; this study was single-centre with a limited sample size; future multicentre studies could be conducted to include specific populations such as those with diabetes or hypertension, thereby further refining the application criteria for PI [24].