Mortality related to procedural sedation and analgesia: a 10-year review of a nationwide medical adverse events database

Abstract

Procedural sedation and analgesia (PSA) is generally considered safe, yet fatal events remain poorly characterised across procedures and specialties. We retrospectively reviewed the Japan Council for Quality Health Care nationwide adverse-event database (2012–2021), searching narrative reports with 40 Japanese keywords for sedation terms and drug names. Only reports in which PSA was deemed the primary cause of death were included; non-procedural sedation, planned general anaesthesia and procedure-induced injuries were excluded. Among 10 011 reports identified via keyword search, 805 described deaths and 23 were attributed to PSA. Fatalities clustered in gastroenterology (73.9%), particularly endoscopic retrograde cholangiopancreatography (ERCP; 34.8%). Most cases were high-risk patients aged 70–89 years (65.2%) and retrospectively classified as American Society of Anesthesiologists Physical Status III/IV (87%), yet no anaesthetist involvement was documented. Oxygen desaturation was the earliest sign of deterioration (56.5%), and capnography was not documented in any case. When PSA fatalities in Japan are examined across procedures and specialties, gastrointestinal interventions—especially ERCP—predominate in our cohort; however, the absence of procedure-specific denominators warrants cautious interpretation. Targeted safety measures, in line with current national sedation guidelines, including enhanced presedation evaluation, anaesthetist involvement for high-risk patients and routine capnography, may help to reduce PSA-related mortality.

Introduction

Procedural sedation and analgesia (PSA) is widely employed for endoscopy, bronchoscopy and cardiac catheterisation, often administered by non-anaesthetists (eg, physicians without formal anaesthesia training).¹,³

Minor complications such as transient oxygen desaturation are common, whereas serious outcomes have been considered rare and PSA-related fatalities even rarer.^{4 5} However, a recent report showed that sedation performed outside the operating room had roughly twice the proportion of fatalities compared with anaesthesia performed in the operating room, and about half of these fatal claims involved gastrointestinal procedures.⁶ To date, systematic investigation of PSA-related fatalities across procedures and specialties within a single country or unified healthcare system remains limited.

We therefore aimed to characterise procedure- and specialty-specific patterns of PSA-related fatalities and to identify high-risk settings in which targeted safety improvements are warranted.

Methods

We retrospectively analysed PSA-related deaths reported to the Japan Council for Quality Healthcare (JCQHC) database from January 2012 to December 2021.

The JCQHC,⁷ a public interest incorporated foundation designated by Japan’s Ministry of Health, Labour and Welfare to collect, analyse, and provide medical adverse event information, collects reports from >4600 Japanese medical institutions as of 31 March 2025. Reports are submitted for events that each institution judges to be a medical adverse event, regardless of outcome severity; the system does not capture all deaths. Participation is mandatory for approximately 270 medical institutions subject to reporting requirements, including university hospitals, National Hospital Organisation hospitals, and other designated institutions, and voluntary for others. For each reported event, the institution classifies outcome severity on a standardised scale ranging from ‘no disability’ to ‘death’. Any healthcare professional may submit a standardised report, which the JCQHC anonymises before public release.

In Japan, PSA for routine procedures is not reimbursed as an anaesthesia professional fee (only the drug cost is covered).^{8 9} Consequently, PSA is typically performed by the interventionalist with nursing assistance, and anaesthetist involvement—even for American Society of Anesthesiologists Physical Status (ASA-PS)≥III—is not routine and largely depends on local staffing and reimbursement constraints. Current national guidelines recommend a preprocedure assessment for all cases, consultation with an anaesthetist for ASA-PS≥III, institution-specific records, a dedicated PSA monitor (eg, an appropriately trained nurse or anaesthetist), capnography^{3 8}; however, these elements are not mandated, and national data on their routine implementation are lacking.

PSA-related cases were identified via the JCQHC’s full-text ‘Case Search’ of narrative PDFs using 40 Japanese keywords (online supplemental table S1), covering generic terms (‘sedation’, ‘analgesia’ and ‘anaesthesia’) and brand/generic drug names selected based on the authors’ clinical experience and domestic practice. All keyword hits were retrieved, sorted by outcome classification, and death reports were manually screened.

Cases were included when the study team judged PSA to be the primary cause of death based on the narratives. Adverse events were classified using an established sedation-risk framework to assist case inclusion and categorisation of relevant variables.^{4 10} Exclusions comprised non-procedural sedation (eg, for anxiety), planned airway management (eg, general anaesthesia) or deaths directly caused by the procedure (eg, perforation, haemorrhage and gas/air embolism). One reviewer (HT) screened all fatal reports; four additional reviewers (KF, MH, YMi and KI) independently assessed subsets; discrepancies were resolved with a senior reviewer (TN).

We summarised demographics, comorbidities, departments, procedures, drugs, timing and nature of adverse events and reporters’ preventive recommendations. Because the JCQHC reports lack a structured field for ASA-PS, we assigned ASA-PS retrospectively from the narratives.¹¹

Results

Among 10 011 reports retrieved, 805 described deaths and 23 were attributed to PSA (figure 1). Annual counts showed a modest decline over time; given the small numbers (n=23), no statistical inference is drawn (online supplemental figure S1).

Figure 1. Flow diagram of case selection. PSA, procedural sedation and analgesia.

Deaths clustered in gastroenterology (73.9%), particularly endoscopic retrograde cholangiopancreatography (ERCP; 34.8%) (table 1). Most patients were aged 70–89 years (65.2%) and were retrospectively classified as ASA-PS III/IV (87.0%). No case documented involvement of an anaesthetist.

Table 1. Characteristics of PSA-related deaths.

	Total		Total
Reports, n	23	Sedatives, n (%) *
Patient age, years, n (%)		Midazolam	10 (43.5%)
0–9	2 (8.7%)	Only	6 (26.1%)
30–49	2 (8.7%)	With thiamylal and pentazocine	1 (4.3%)
60–69	4 (17.4%)	With pentazocine	2 (8.7%)
70–79	9 (39.1%)	With pethidine	1 (4.3%)
80–89	6 (26.1%)	Diazepam	4 (17.4%)
ASA-PS classification, n (%)		Only	3 (13.0%)
II	3 (13.0%)	With pentazocine	1 (4.3%)
III	11 (47.8%)	Flunitrazepam only	2 (8.7%)
IV	9 (39.1%)	Propofol only	1 (4.3%)
Major comorbidities, n (%) †		Hydroxyzine with pentazocine	1 (4.3%)
Acute-subacute biliary/pancreatic disease	9 (39.1%)	Unspecified sedative(s)	4 (17.4%)
Sepsis	5 (21.7%)	Analgesics, n (%)
Active pneumonia	4 (17.4%)	Pentazocine	5 (21.7%)
Active upper GI bleeding	4 (17.4 %)	Pethidine (meperidine)	2 (8.7%)
Early-stage gastric cancer	2 (8.7%)	Timing of initial deterioration, n (%)
Intestinal obstruction	2 (8.7%)	During procedure	15 (56.5%)
Old myocardial infarction	2 (8.7%)	Shortly after procedure (0–30 min)	4 (17.4%)
Difficult airway	1 (4.3%)	After procedure (30 min–)	4 (17.4%)
Normal delivery	1 (4.3%)	Initial signs of deterioration, n (%) †
Ornithine transcarbamylase deficiency	1 (4.3%)	Oxygen desaturation	13 (56.5%)
Peripheral pulmonary artery stenosis	1 (4.3%)	Bradypnoea	1 (4.3%)
Rectal cancer with multiple metastases	1 (4.3%)	Hypotension	2 (8.7%)
Severe hypothermia	1 (4.3%)	Tachycardia	2 (8.7%)
Clinical department performing PSA, n (%)		Bradycardia	2 (8.7%)
Gastroenterology	17 (73.9%)	Paradoxical response	2 (8.7%)
Pulmonology	2 (8.7%)	Others	4 (13%)
Paediatrics	2 (8.7%)	Presumed causes of death, n (%)
General Medicine	1 (4.3%)	Pulmonary aspiration syndrome	5 (21.7%)
Obstetrics	1 (4.3%)	Respiratory depression	3 (13%)
Procedures/diagnostic tests, n (%)		Cardiovascular collapse	4 (17.4%)
Endoscopic retrograde cholangiopancreatography	8 (34.8%)	Drug metabolism-related events	3 (13%)
Upper GI endoscopy for bleeding	3 (13%)	Renal/hepatic impairment	1 (4.3%)
Endoscopic submucosal dissection	2 (8.7%)	Drug-drug interactions‡	2 (8.7%)
Bronchoscopy	1 (4.3%)	Unspecified	8 (34.8%)
Catheter angiography	1 (4.3%)	Reporters’ preventive recommendations, n (%) †
Endoscopic pancreatic necrosectomy	1 (4.3%)	Presedation evaluation	11 (47.8%)
Endoscopic transanal decompression	1 (4.3%)	Monitoring duration and modality	11 (47.8%)
Ileus tube placement	1 (4.3%)	Postsedation	5 (21.7%)
Liver biopsy	1 (4.3%)	ECG	3 (13%)
Local anaesthetic thoracoscopy	1 (4.3%)	Capnography	3 (13%)
Percutaneous endoscopic gastrostomy	1 (4.3%)	Communication; interprofessional	9 (39.1%)
Manual placenta removal	1 (4.3%)	Intradepartmental	5 (21.7%)
MRI	1 (4.3%)	Anaesthetist consultation	3 (13%)
Monitoring at the time of deterioration, n (%)		Institutional PSA guideline	9 (39.1%)
ECG and SpO2	7 (30.4%)	Communication; Patient and family	6 (26.1%)
SpO2 only	7 (30.4%)	Additional personnel	5 (21.7%)
No monitoring	2 (8.7%)	Additional supplies	3 (13%)
Monitoring status unknown	7 (30.4%)	Drug–drug interaction alerts‡	2 (8.7%)

^*Percentages do not total 100% because one case involved the use of pethidine alone without a sedative agent.

^†Some cases involved multiple applicable categories.

^‡Suspected cases included serotonin syndrome from pethidine combined with a monoamine oxidase inhibitor and respiratory depression from cimetidine combined with flunitrazepam, prompting consideration of system checks for potentially risky drug interactions.

ASA-PS, American Society of Anesthesiologists Physical Status; GI, gastrointestinal; PSA, procedural sedation and analgesia; SpO₂, oxygen saturation.

Adverse events occurred during the procedure or ≤30 min after in 82.6%; the most common initial sign was a fall in SpO₂ (56.5%). Capnography was not documented for any patient.

Adverse-event reporters suggested that these deaths could have been prevented by enhanced presedation evaluation (47.8%), better monitoring (47.8%) and improved interprofessional communication (39.1%); anaesthetist consultation (13.0%) and capnography (13.0%) were also mentioned.

Medication use varied; key agents included midazolam (43.5%) and pentazocine (21.7%) among others. Dosing details were frequently absent. No cases used remimazolam, which was not available for PSA in Japan during the study period.

Discussion

PSA-related deaths in Japan were reported predominantly during gastrointestinal procedures, particularly ERCP. Although ERCP is recognised as carrying a higher-risk gastrointestinal intervention,¹²,¹⁴ its prominence here should be interpreted with caution because procedure-specific denominators were unavailable. The seventh nationwide survey of gastrointestinal endoscopy in Japan (2024) reported 6424 therapeutic ERCPs, with 104 adverse events (1.62%) and two deaths (0.03%)¹⁴; however, it did not capture PSA-related information (eg, provider, airway management and monitoring), the type of insufflation gas used (carbon dioxide vs air) or presumed causes of death (eg, aspiration and gas embolism), limiting comparability with our dataset. Given ERCP’s risk profile, targeted PSA safety efforts around ERCP are warranted.

Our study showed that deaths mainly involved elderly, comorbid patients (ASA-PS III/IV)—known PSA risk categories.¹,^{515 16} Yet these risks often appeared under-recognised at the time of care: institutional reporters cited inadequate presedation assessment in 47.8% of our cases, and no case documented anaesthetist involvement. As stated above, the Japanese PSA guidelines recommend preprocedure assessment and also advise anaesthetist consultation for ASA-PS≥III.^{3 8} Providers who perform PSA should be familiar with the latest guideline recommendations.

Respiratory complications dominate PSA adverse events^{4 5 17 18}; our cohort aligned with this, with desaturation as the most frequent initial sign. Comorbid pneumonia, sepsis or cardiac disease heightened hypoxaemia risk, suggesting that timelier airway control or tighter respiratory monitoring might have prevented some deaths—consistent with national guidelines, which strongly recommend capnography for moderate-to-deep sedation.^{3 8}

Pentazocine—uncommon internationally¹,³—reflects domestic formulary/regulatory context.¹⁹,²¹ Japanese endoscopy guidance lists it as an adjunct analgesic with limited comparative evidence,⁸ which may limit cross-setting generalisability.

Limitations include the JCQHC’s limited coverage, comprising only a subset of Japan’s >8000 hospitals and >170 000 clinics, potential reporting bias towards safety-aware centres, and absence of denominator data, precluding mortality rate calculation. Narrative, unstructured reporting may have yielded residual misclassification and under-ascertainment of non-PSA, procedure-intrinsic complications despite a priori exclusions. Country-specific practices also limit generalisability. Non-fatal PSA-related morbidity (eg, unplanned admission or intensive care unit admission) was not analysed in this study and should be explored in future studies. Nonetheless, this is the first nationwide, cross-specialty review of PSA-related deaths and underscores the value of national registries for safety improvement.

Conclusion

In Japan’s national adverse-event database, PSA-related deaths were reported chiefly in gastroenterology, particularly during ERCP. Strengthening guideline-based sedation practices, including anaesthetist involvement for high-risk patients and routine capnography, may enhance procedural safety.