This study employed a single-center retrospective cohort design to systematically screen and analyze the complete clinical records of 120 patients who received Non-invasive Positive-Pressure Ventilation (NPPV) treatment at our hospital between January 2024 and October 2025. Patients were divided into two groups based on sedative selection: Group C (ciprofol group, n = 56) and Group P (propofol group, n = 64). All patients received NPPV support for acute respiratory failure and required sedation to alleviate patient-ventilator asynchrony. Through strict matching of baseline characteristics, the two groups were comparable with no significant differences in age, gender, BMI, or distribution of underlying diseases (P > 0.05). The detailed process of study screening is depicted in Figure 1.

Researchers developed multidimensional inclusion and exclusion criteria to ensure data quality. Inclusion criteria include: ① Adult patients (≥18 years old); ② Confirmed acute respiratory failure receiving non-invasive ventilation support; ③ Complete documentation of sedation treatment using ciprofol or propofol; ④ RASS score ≥ +1 prior to medication administration, with agitation requiring sedation; ⑤ Complete medical record documentation, including baseline vital signs, vasoactive drug usage, and hepatic/renal function laboratory data. Exclusion criteria include: ① Early conversion to invasive ventilation (within 1 h of drug administration); ② Missing key clinical data (e.g., RASS score, vital signs); ③ Confirmed history of drug allergy; ④ Concurrent severe hemodynamic instability, end-stage organ dysfunction, or special physiological conditions (pregnancy/lactation); ⑤ Simultaneous participation in other clinical trials that may interfere with outcome assessment.

All patients received standardized non-invasive ventilation support, with ventilator parameters dynamically adjusted based on arterial blood gas analysis results. A unified analgesia and sedation protocol was implemented: first, all patients received continuous infusion of remifentanil as background analgesia. Remifentanil was prepared by diluting 2 mg in 50 mL of normal saline (final concentration 40 μg/mL), with an initial infusion rate set at 3–6 mL/h (approximately equivalent to 0.0017–0.0034 mg/kg/h). One minute after the start of remifentanil infusion, a loading dose of the sedative drug was administered intravenously—patients in the C group received ciprofol 0.4 mg/kg, while those in the P group received propofol 1.5 mg/kg. Maintenance infusion of the sedative was initiated immediately thereafter. Both ciprofol and propofol were administered as ready-to-use intravenous solutions. In this study, ciprofol was used at a concentration of 2.5 mg/mL, and propofol at 10 mg/mL. Based on the drug characteristics and individual patient response, the infusion rates were individually titrated: the initial maintenance infusion rate for ciprofol in the C group was typically 2–8 mL/h (approximately equivalent to 0.05–0.20 mg/kg/h), and for propofol in the P group, 2–12 mL/h (approximately equivalent to 0.20–1.20 mg/kg/h). The infusion rates of all drugs were dynamically titrated according to patient response, with the ultimate goal of maintaining the patient’s Richmond Agitation-Sedation Scale (RASS) score between −2 and 0. Sedation depth, respiratory and circulatory parameters, and adverse reactions were continuously monitored throughout the treatment period.

The primary outcome focused on sedation efficacy and was assessed using commonly adopted ICU standards: sedation induction time (defined as the time from the initiation of the study drug infusion to the first time the patient achieved the target sedation depth). Sedation depth was evaluated using the Richmond Agitation-Sedation Scale (RASS), a validated and reliable tool for assessing consciousness and sedation levels in ICU patients. The RASS score ranges from −5 (unarousable) to +4 (combative), with 0 representing an alert and calm state (Cn et al., 2002). In this study, the target sedation depth was defined as a RASS score between −2 and 0.

Secondary outcomes encompassed a multidimensional evaluation, including: ① Safety indicators: the incidence of hypoxemia (defined as SpO₂ <90% for >1 min) and hypotensive events (defined as a mean arterial pressure [MAP] <65 mmHg or a decrease of >30% from baseline); ② Recovery quality indicators: time to awakening (from discontinuation of sedative infusion until RASS score ≥ −1) and time to full recovery (until RASS score returned to 0); ③ Respiratory function indicators: the temporal changes in respiratory rate (RR), oxygen saturation (SpO₂), oxygenation index (PaO₂/FiO₂), and partial pressure of carbon dioxide (PaCO₂); ④ Hemodynamic indicators: the temporal changes in mean arterial pressure (MAP) and heart rate (HR); ⑤ Other adverse reactions: the incidence of bradycardia, injection pain, nausea/vomiting, and delirium (defined by an Intensive Care Delirium Screening Checklist [ICDSC] score ≥4). Data for all time-sequenced indicators were collected at four fixed time points: before sedation initiation (baseline, T1), and at 1 h (T2), 6 h (T3), and 24 h (T4) after initiation.

As a retrospective analysis, the sample size of this study was constrained by the availability of case data within a specific timeframe; therefore, a formal prospective a priori sample size calculation was not performed. To evaluate the statistical power of the current sample size for detecting a difference in the primary outcome (sedation induction time), a simulation-based estimation was conducted.

The parameters for this simulation were set as follows: a between-group difference of 5 s, which was defined as the prespecified “minimum clinically important difference” for this study—a standardized approach used to determine whether an observed difference in clinical outcomes holds practical significance (Ja et al., 2021); and a standard deviation of 10 s, serving as a conservative estimate of variability during the induction phase. The 5-s threshold was informed by evidence from a recent systematic review indicating no statistically significant difference in sedation induction time between ciprofol and propofol (Armengo et al., 2026). The 10-s standard deviation was contextualized against the considerably larger inter-individual variability (standard deviation of approximately 390 s) reported in a study of a similar ICU population (Liu et al., 2023); the use of a substantially smaller value in our simulation aimed to establish a stringent testing scenario, demonstrating that the sample size would remain adequate even under assumptions that underestimate variability.

Based on these conservative parameters, the simulated required sample size was in the same order of magnitude and closely aligned with the 120 patients actually included in this study. This indicates that, even under the strictest assumptions, the present sample size provides reasonable and acceptable statistical power for detecting the prespecified difference in the primary outcome.

This study employed SPSS 27.0 for statistical analysis. Quantitative data were verified as normally distributed via the Shapiro-Wilk test and described as mean ± standard deviation (x ± s). Intergroup comparisons were performed using the independent samples t-test. Qualitative data were expressed as frequency (percentage), with intergroup comparisons conducted using the chi-square test. All tests were two-tailed, with p < 0.05 indicating statistically significant differences.

A total of 120 patients undergoing treatment with NPPV were included in the study. The cohort comprised 72 males and 48 females, with a mean age of 48.0 ± 14.8 years and a mean body mass index (BMI) of 18.5 ± 3.1 kg/m². There were no significant differences in general characteristics such as age, gender, and BMI between the two groups (p > 0.05), indicating comparability (Table 1).

The primary outcomes for both groups are presented in Table 2. Regarding sedation induction, there was no statistically significant difference in the induction time to achieve the target sedation depth (RASS -2 to 0) between the C group and the P group (t = −1.256, p = 0.212). This result indicates that when used for sedation in NPPV patients, ciprofol has a comparable speed of onset to propofol, and the two drugs show similar efficacy in terms of initial sedative effect.

This retrospective cohort study systematically compared the performance of ciprofol versus propofol for sedation in patients receiving non-invasive positive pressure ventilation. The core findings demonstrate that ciprofol provides sedation equally effective to propofol, as evidenced by no significant difference in induction time. Building on this equivalent efficacy, ciprofol exhibited significant advantages in safety and recovery quality: it was associated with a significantly lower incidence of both hypoxemia and hypotension, shorter times to awakening and full recovery, and a markedly reduced incidence of injection pain. Collectively, these results suggest that ciprofol is an effective and safer alternative for sedation during NPPV therapy.

The findings of this study corroborate previous research and provide important additional evidence regarding the safety advantages of ciprofol within a real-world clinical setting. The relevance of our results to existing evidence is analyzed below from several perspectives.

Our study demonstrates the efficacy and safety of ciprofol for sedation during NPPV. However, several limitations should be considered. First, the retrospective design may not fully preclude confounding factors despite strict inclusion and exclusion criteria. Second, the sample size, though adequate for the primary outcome as contextualized by our simulation (which employed conservative prespecified parameters: a 5-s minimal clinically important difference and a 10-s standard deviation), remains relatively limited. This may constrain the statistical power to detect differences in rare adverse events. Third, the observation period was confined to the sedation and immediate recovery phases, precluding evaluation of long-term outcomes. Finally, although sedation induction time as the primary endpoint directly reflects the drug’s onset of action, future studies could employ more comprehensive metrics—such as time to stable sedation or NPPV tolerance scores—to holistically assess sedation quality.

These limitations highlight the necessity for validating the current conclusions through prospective, multicenter, randomized controlled trials with larger sample sizes. Beyond confirming clinical efficacy, future research should also incorporate pharmacoeconomic evaluations, which are critical for optimizing resource allocation in the high-cost intensive care unit setting (Ss et al., 2012).