Comparative safety and sedation-related outcomes of fospropofol disodium versus propofol for adult sedation outside general anaesthesia induction or maintenance (including ICU and procedural sedation): a systematic review, meta-analysis and trial sequential analysis

BackgroundFospropofol disodium is a water-soluble prodrug of propofol that avoids a lipid emulsion carrier. It may therefore reduce injection pain, cardiorespiratory depression and lipid exposure. In May 2021 it was approved in China (H20210017) for induction of general anaesthesia in adults. Randomised trials have since tested it for sedation outside the operating room, but this setting has not been systematically reviewed. MethodsWe followed PRISMA 2020 and PRISMA-S and searched Europe PMC, PubMed, Embase, Cochrane CENTRAL, Web of Science and the Chinese databases CNKI, Wanfang and VIP (the last accessed through the XJMU Library database proxy) from inception to 2026. Eligible patients were adults receiving sedation for purposes other than general anaesthesia induction or maintenance, including continuous sedation during ICU mechanical ventilation and procedural sedation. These uses are referred to here as sedation outside the operating room. We included randomised controlled trials with propofol as the comparator and excluded trials of general anaesthesia induction or maintenance to avoid overlap with the published induction meta-analysis. Binary outcomes were pooled with a DerSimonian-Laird random-effects model, and REML plus modified Hartung-Knapp estimates were reported alongside to improve interval coverage when few studies were available. Outcomes with a high proportion of zero cells were also analysed with the Peto method. We stratified by setting and blinding and performed leave-one-out analysis, trial sequential analysis (TSA) and GRADE assessment. Hypotension reached the prespecified threshold of at least 10 studies, so publication bias was assessed with the Harbord test, with the originally specified Eggers test reported alongside. The protocol was prospectively registered and took effect on 16 September 2026. All searches were completed on 21 September 2026, and screening, extraction and analysis were performed after registration. ResultsTwelve randomised controlled trials were included. Event counts for all 12 studies were checked against the full-text original publications. For Liu 2026, a preprint, counts were taken directly from its Tables 2 and 3. Compared with propofol, fospropofol disodium reduced injection pain under the DerSimonian-Laird model (8 studies, OR 0.09, 0.03-0.21) and increased pruritus/paraesthesia (4 studies, OR 32.56, 12.89-82.26; RD +28.5%). During procedural sedation, it was associated with fewer respiratory adverse events (5 studies, OR 0.19, 0.11-0.32). This composite includes respiratory depression, hypoxaemia and related events reported by individual trials and should not be extrapolated to all adult sedation settings. For hypertriglyceridaemia, the DerSimonian-Laird estimate was OR 0.29 (0.13-0.65), but all four studies were conducted in ICU settings and the REML plus modified Hartung-Knapp estimate was 0.29 (0.08-1.08), with a confidence interval crossing 1. Certainty was very low, so this finding remains hypothesisgenerating. Hypotension differed by setting: OR 0.24 (0.13-0.42) for procedural sedation and 0.90 (0.53-1.55) for continuous ICU sedation, with a significant interaction (P < 0.001). The overall estimate (OR 0.51, 0.28-0.93) is reported for summary only. Under REML plus modified HartungKnapp it was 0.51 (0.25-1.03), crossing 1, so it does not support a claim that the drug lowers blood pressure and should not be extrapolated to the ICU. Bradycardia did not differ significantly (9 studies, OR 0.45, 0.17-1.18), and the direction reversed in the only double-blind study. Two studies reported slower onset or loss of consciousness with fospropofol disodium (MD +2.22 and +0.92 min), but these were not pooled because heterogeneity was extreme (I2 = 99.6%). Recovery, discharge and successful extubation showed no clear differences, although heterogeneity was high (I2 = 85%-92%) and the evidence cannot establish equivalence. Sedation efficacy was not pooled because definitions differed. Patient-important outcomes (death, delirium, ICU length of stay and duration of mechanical ventilation) were each reported by only one or two small studies with inconsistent definitions and time points, so no reliable conclusion can be drawn. Setting-specific results should take priority over the overall mean. Certainty ranged from very low to moderate. O_TBL View this table: org.highwire.dtl.DTLVardef@10e2859org.highwire.dtl.DTLVardef@7b1ee4org.highwire.dtl.DTLVardef@1290d08org.highwire.dtl.DTLVardef@16a068org.highwire.dtl.DTLVardef@17d2482_HPS_FORMAT_FIGEXP M_TBL O_FLOATNOTable 2:C_FLOATNO O_TABLECAPTIONBinary event counts by study and outcome (fospropofol disodium versus propofol). C_TABLECAPTION C_TBL O_TBL View this table: org.highwire.dtl.DTLVardef@e7e1eeorg.highwire.dtl.DTLVardef@1a1d1bforg.highwire.dtl.DTLVardef@d17650org.highwire.dtl.DTLVardef@1ae7fa2org.highwire.dtl.DTLVardef@13d6798_HPS_FORMAT_FIGEXP M_TBL O_FLOATNOTable 3:C_FLOATNO O_TABLECAPTIONThreshold, recording window and intervention requirement for hypotension in each study (10 studies). C_TABLECAPTION C_TBL ConclusionsDuring procedural sedation, fospropofol disodium may reduce injection pain and respiratory adverse events and may lower the risk of hypotension compared with propofol. This hypotension benefit should not be extrapolated to the ICU. During continuous ICU sedation no data were available for injection pain or respiratory adverse events, so the safety evidence in that setting remains limited. Hypertriglyceridaemia was lower only under the DerSimonian-Laird model. Because the result depends on the method and certainty is very low, it cannot be treated as definitive. The main cost is a clear increase in pruritus and abnormal sensations. Onset may be slower, but the evidence comes from two bolus-dosing studies in procedural sedation, and no such disadvantage was seen during continuous ICU infusion. Evidence on sedation efficacy, recovery time, discharge time and successful extubation is limited, and equivalence between the two drugs cannot be asserted. Because most studies were small and open-label and certainty ranged from very low to moderate, these findings support investigational use only and do not justify a routine clinical recommendation. Research in contextO_ST_ABSEvidence before this studyC_ST_ABSWe searched Europe PMC, PubMed, Embase, Cochrane CENTRAL, Web of Science Core Collection and Chinese databases (CNKI, Wanfang, VIP) by crossing drug terms (fospropofol, fospropofol disodium, Lusedra, Aquavan) with sedation setting terms (ICU, mechanical ventilation, endoscopy, gastroscopy, hysteroscopy, bronchoscopy, procedural sedation). Within the retrieved literature, the only relevant meta-analysis addressed general anaesthesia induction; no synthesis of the sedation setting was identified. Added value of this studyThe present synthesis focuses on sedation outside the operating room, includes 12 RCTs, and stratifies by setting. It reports eight safety and efficacy outcomes (recovery time and other continuous outcomes are reported narratively), together with heterogeneity, leave-one-out sensitivity analysis, trial sequential analysis and GRADE certainty, and it quantifies the costs of fospropofol disodium (pruritus/paraesthesia; slower onset after bolus dosing). Clinical implicationsFospropofol disodium may reduce injection pain during sedation. For hypertriglyceridaemia the result depends on the estimator and certainty is very low. During procedural sedation it may reduce hypotension and respiratory adverse events, whereas the evidence for a hypotension benefit during continuous ICU sedation is weak. Because the approved indication in China is adult general anaesthesia induction only, these uses remain investigational; they do not constitute a routine clinical recommendation. Applicability to specific patient groups remains an inference from pharmacological mechanism.