Safety of MRI Examinations Under Sedation: A Nationwide Survey in Japan

ABSTRACT

Background

MRI is performed under sedation for patients requiring immobility or for those already sedated for clinical management. Although specialist guidelines advocate for anesthesiologist‐led care, non‐anesthesiologist‐led adult sedation is a common practice worldwide. Furthermore, the increase in risk attributable to sedation has not been quantified.

Purpose

To investigate MRI sedation practices in Japan to quantify associated risks and to identify key safety vulnerabilities.

Study Type

Survey.

Subjects

Data were collected from 549 Japanese medical institutions.

Assessment

A web‐based questionnaire was distributed to members of major Japanese radiological societies to assess institutional policies, sedation management practices, safety measures, and historical adverse events. The survey primarily concerned deep sedation and general anesthesia in adults but also encompassed questions on minimal‐to‐moderate sedation and pediatric sedation.

Statistical Tests

Chi‐squared test, Welch's t‐test, Fisher's exact test were used. Odds ratios (ORs) were calculated based on the cumulative number of adverse events over 2 years to examine the increased risk from sedation. p < 0.05 was considered significant.

Results

Adult MRI sedation management is typically led by attending physicians (84%) with limited consultation from anesthesiologists for sedation (9.0%) or radiologists for scan indication (5.1%). Safety infrastructure was often inadequate, with MRI‐compatible ventilators (48%) or syringe pumps (29%) and capnometers (24%). In an analysis combining both pediatric and adult data, the ORs for the total sedated group versus non‐sedated group were 62 (95% confidence interval, 54–73) for peripheral capillary oxygen saturation drop, 48 (32–72) for respiratory arrest, 15 (8.0–28) for cardiac arrest, 6.7 (2.0–23) for physical trauma, and 8.9 (4.0–20) for projectile accidents.

Data Conclusion

MRI examinations under sedation are associated with a substantially higher risk of adverse events. Developing and implementing a standardized protocol may mitigate these procedural risks.

Level of Evidence

4.

Stage of Technical Efficacy

5.

Plain Language Summary

MRI sometimes requires patients to be sedated to keep them still, but the associated risks in real‐world practice were unclear. This study surveyed 549 facilities in Japan to understand how they manage MRI sedation. The research found that sedation is associated with a substantially higher risk of serious adverse events, such as respiratory and cardiac arrest. It also revealed that sedation exams for adults are often performed without input from anesthesiology or radiology specialists. Developing and implementing a standardized protocol may be an effective way to mitigate these procedural risks and improve patient safety.

1. Introduction

The diagnostic utility of MRI is well established, yet its efficacy hinges on patient compliance with limited to no voluntary movements. To ensure high‐quality, motion‐free scans in individuals unable to cooperate—such as pediatric or anxious adult patients—procedural sedation has become a frequent practice. Although effective, this practice is not without risk. A substantial body of research, particularly in pediatrics, has identified respiratory depression as a primary complication [1, 2]. These risks are amplified by the unique hazards of the MRI suite, where strong magnetic fields and intense acoustic noise create a difficult environment for pharmaceutical administration, patient monitoring, and direct observation.

To manage these high‐stakes conditions, leading guidelines from the American Society of Anesthesiologists [3] and the Association of Anesthetists of Great Britain and Ireland [4] advocate for a specialist‐led model, where sedation is managed by anesthesiologists with specific training for the MRI environment. However, despite these recommendations, a substantial gap often exists between these ideal standards and common global practice. In numerous settings worldwide, sedation for MRI is frequently administered by non‐anesthesiologist physicians [5, 6], often due to resource constraints amid rising demand for non‐operating room anesthesia [7]. In Japan, this reality led to the development of the “Joint Recommendations on Sedation for MRI Examinations,” by Japanese pediatric societies—a comprehensive, practical manual for pediatricians designed to elevate the safety of sedation performed by non‐anesthesiologists [8, 9]. Conversely, a similar standardized framework for adult MRI sedation in these non‐anesthesiologist‐led contexts is notably absent, resulting in practices that are largely determined at the institutional level.

Moreover, while the risks of sedation are well documented, the relative risk increase for adverse events specifically attributable to sedation itself, when compared to non‐sedated MRI examinations, has not been fully quantified. Therefore, to address both the need for clinical standardization and this deficit in the literature, we implemented a nationwide questionnaire survey. This study aims to investigate the real‐world practices of adult MRI sedation in Japan, quantify the associated risks, and identify key contributing factors to enhance patient safety.

2. Methods

This study was conducted from April 21 to May 31, 2025, after receiving approval from the institutional review board of the University of Tokyo (Review Number: 2025002NIe) and permission from the University of Tokyo Hospital. An invitation to participate in a web‐based questionnaire was sent via email to members of the Japanese Society for Magnetic Resonance in Medicine. Cooperation for survey dissemination was also obtained from the Japan Radiological Society, the Japanese Society of Radiological Technology, and the Japan Association of Radiological Technologists. The primary anticipated respondents were radiologists and radiological technologists, who are the main members of these societies, although physicians from other specialties were also on the distribution list and included as potential respondents. As it was presumed that multiple individuals at a single institution would receive the invitation, the instructions requested that recipients consult with relevant colleagues to provide a single, consolidated response for their facility. Upon submission, a copy of the answers was automatically sent to the respondent, who was permitted to edit their response at any time during the survey period mentioned above. The questionnaire was not anonymous, allowing for follow‐up inquiries to clarify any suspected inaccuracies or resolve duplicate responses from the same institution.

The survey (Supporting Information) was designed to collect data on institutional demographics, MRI suite facilities, and operational procedures, including access control and item management, general MRI safety matters, guided by standards such as the American College of Radiology (ACR) Manual on MR Safety [10], the implementation and rules for sedated MRI scans, reasons for considering deep sedation, and historical adverse events. The content primarily concerned examinations under deep sedation and general anesthesia (hereafter, deep sedation) in adults but also encompassed questions regarding minimal‐to‐moderate sedation (formerly referred to as conscious sedation) and pediatric sedation.

The institutional categories are defined based on the Japanese Medical Service Act. “Advanced Treatment Hospitals” are high‐level medical institutions approved by the Ministry of Health, Labour and Welfare that provide advanced medical care and conduct clinical research. “Regional Medical Support Hospitals” are facilities approved by prefectural governors to provide primary care and support local clinics. “General Hospitals” are defined as having 20 or more beds, and “Clinics” have fewer than 20 beds. The “Additional Reimbursement for Imaging Diagnosis Management” is a Japanese billing code used in this study as another indicator of facility scale, as meeting its stricter requirements—such as a high number of specialists and 24‐h support—are generally met only by larger institutions.

The objectives for deep sedation during MRI for both adults and children were broadly categorized into two types: (1) non‐examination purposes, typically involving intensive care unit patients already sedated for reasons unrelated to the MRI, like post‐operative respiratory management, and (2) examination purposes, where sedation was administered specifically to ensure patient immobility due to conditions such as anxiety.

Regarding historical adverse events, the survey collected data over two distinct time frames. First, it asked for the total number of predefined adverse events that occurred within the past 2 years. Second, it gathered information on whether facilities had a lifetime experience with those same events, with responses categorized across four conditions (pediatric/adult and with/without sedation).

3. Statistical Analysis

Statistical analyses were performed using SPSS version 28.0.1.0 (SPSS Inc., Chicago, IL). Descriptive statistics were used to summarize institutional characteristics; continuous variables are presented as mean ± standard deviation, whereas categorical variables are presented as numbers and percentages. The association between categorical variables, such as the number of hospital beds or the billing code and the performance of deep sedation, was assessed using the chi‐squared test. Welch's t‐test was used to compare the means of continuous variables (e.g., MRI units, annual examination volume, number of certified MRI technologists and board‐certified diagnostic radiologists) between facilities that do and do not perform sedation.

To evaluate the risk of adverse events, the incidence rate over the past 2 years was calculated. For this calculation, the total number of sedated examinations was estimated from categorical survey responses. When a frequency was reported as a range (e.g., N to N + 1 times per month), its midpoint (N + 0.5) was used. For the open‐ended category of “4 or more per month,” a conservative value was imputed by assigning a deliberately high estimate of 30 procedures per month. The incidence of adverse events was then compared between the sedated and non‐sedated groups using Fisher's exact test. The magnitude of increased risk was quantified using odds ratios with 95% confidence intervals.

For all inferential tests, a two‐sided p value of less than 0.05 was considered statistically significant. All percentages presented in the results section, as well as figures, are calculated as a proportion of the total number of survey respondents (n = 549) unless otherwise noted.

4. Results

4.1. Institutional Characteristics and Sedation Frequency

4.1.1. Demographics of Responding Institutions

Responses were obtained from 549 institutions nationwide, comprising Advanced Treatment Hospitals (22%, n = 119), Regional Medical Support Hospitals (37%, n = 205), General Hospitals (36%, n = 196), and Clinics (5%, n = 29). The distribution of these facilities by prefecture is shown in Figure S1. Of the respondents, 32% (n = 177) reported that they do not perform MRI examinations under deep sedation or general anesthesia for any patient population. When analyzed by bed count, the proportion of facilities performing deep sedation examinations progressively increased with the number of beds, from 3% in facilities with no beds to 95% in facilities with over 501 beds (Figure 1A). A significant association was found between the number of beds and the performance of deep sedation examinations, as demonstrated by a chi‐square test comparing facility performance across 7 bed count categories (χ ²  = 209.9). Factors related to institutional size, such as the number of MRI units and annual examination volume, were also significantly associated with the performance of deep sedation examinations (Figure 1B). Significant differences were observed in the mean number of MRI units (1.4 ± 0.06 vs. 2.6 ± 0.07) and mean annual examinations (4259 ± 304 vs. 28.997 ± 298) for facilities that do not perform deep sedation versus those that do, respectively (Welch's t‐test: MRI units, t = −11.1; examinations, t = −11.7). Similarly, analysis based on the billing code suggested that the performance of examinations under sedation is associated with its status (χ ²  = 74.9). Additionally, facilities that performed examinations under sedation had a significantly higher mean number of certified MRI technologists (1.1 ± 2.5 vs. 2.1 ± 2.1; t = 4.3) and board‐certified diagnostic radiologists (2.5 ± 15.1 vs. 4.9 ± 5.1; t = 2.1) (Figure 1C).

FIGURE 1.

Association between institutional size and performance of sedated MRI. (A) The number of hospitals that perform deep sedation for any patient population (pediatric or adult) (red) versus those that do not (blue), categorized by bed count, shows a clear trend where facilities with over 200 beds were more likely to perform deep sedation MRI. (B) A scatter plot showing that facilities performing deep sedation (red markers) have a higher number of MRI scanners and a greater annual examination volume. (C) A bar chart demonstrating that facilities with a higher Reimbursement Code—an indicator of a larger and more functionally advanced facility due to stricter requirements—also have a higher percentage of deep sedation performance (gray bars).

4.1.2. Performance of Deep Sedation Examinations by Sedation Objective

The percentage of institutions performing deep sedation examinations varied by sedation objective: 49% for non‐examination purposes, 49% for pediatric examination purposes, and 40% for adult examination purposes (Figure 2A). Regarding the frequency of these examinations, the majority of institutions performed them infrequently. For non‐examination purposes, pediatric examination purposes, and adult examination purposes, 86%, 79%, and 94% of all responding institutions reported conducting 0–3 cases per month, respectively. Conversely, 12%, 21%, and 5.5% reported conducting 4 or more cases per month for each respective purpose (Figure 2B).

FIGURE 2.

Performance and frequency of deep sedation by objective. (A) The percentage of institutions that perform deep sedation, categorized by sedation purpose: For non‐examination purpose (patients are already sedated for reasons unrelated to the MRI), for pediatric MRI, and adult MRI. (B) The monthly frequency of deep sedation cases for each purpose shows that the majority of institutions perform these procedures infrequently (0–3 cases per month).

4.1.3. Availability of MRI Services on Nights and Holidays

A chi‐squared test revealed a significant difference in the availability of MRI services during off‐hours (nights and holidays) between general examinations and those requiring deep sedation (χ ²  = 517). For general MRI, 46% of institutions reported having established conditions or rules for their performance, 38% performed them routinely, 11% generally did not perform them, and 4.9% operated on a case‐by‐case basis. In contrast, for examinations requiring deep sedation, the majority of facilities reported that they are “generally not performed” (64%) or are handled on a “case‐by‐case” basis (20%) (Figure S2).

4.2. Management of the MRI Environment and Equipment

4.2.1. Access Control and Item Management

Regarding access control for the MRI scan room, while a majority of facilities (63%) have a rule where an MR technologist or another staff member performs a check, a substantial portion relies on the individuals entering the room to self‐manage according to rules (30%). Some facilities reported having no rules (6.6%) or having different rules for MRI staff versus non‐staff (0.7%) (Figure S3).

4.2.2. MRI Scan Room Equipment

Regarding equipment, many facilities were well‐equipped with oxygen piping (92%), suction devices (83%), monitoring cameras (97%), a nearby resuscitation area (82%), an emergency cart (86%), and clearly labeled MRI‐compatible/incompatible equipment (91%). However, less than half of the responding institutions had installed waveguides (42%), MRI‐compatible ventilators (48%), or MRI‐compatible syringe pumps (29%) (Figure 3).

FIGURE 3.

Availability of MRI scan room equipment. The bar charts show the percentage of institutions equipped with various devices. Although a majority of facilities have basic equipment such as oxygen piping (92%) and monitoring cameras in 1 or 2 directions (dir) (97%), less than half have specialized, MRI‐compatible equipment, including ventilators (48%), waveguides (42%), and syringe pumps (29%). Additionally, while the majority of facilities report that MR unsafe equipment used in the examination room was “Generally labeled (as MR unsafe)” (67%), a notable portion report that only “Some items are labeled” (24%) or that there was “Almost no labeling” (9%).

4.3. Personnel, Decision‐Making, and Clinical Management in MR Sedation

4.3.1. Deep Sedation MRI: Reasons for Consideration

Among all respondents (n = 549), 52% of institutions consider performing examination‐purposed deep sedation for adult patients. The most common reasons for this consideration (multiple selections allowed) were “Request from the attending physician (due to difficulty remaining still from delirium, intellectual disability, dementia, etc.),” cited by 47% of all survey respondents, and a “Patient request (due to claustrophobia, etc.),” cited by 27% of all respondents (Figure S4). The main reasons for choosing deep sedation or general anesthesia over minimal‐to‐moderate sedation (multiple selections allowed) were: “Request from the attending physician (due to difficulty remaining still)” at 38% of all respondents, “Patient request (due to claustrophobia, etc.)” at 18% of all respondents, and “Policy of the sedationist” at 9.1% of all respondents (Figure S5).

4.3.2. Deep Sedation MRI: Airway Management

Among all respondents, 40% perform examination‐purposed deep sedation in adults. For these examinations, the reported methods of airway management (multiple selections allowed) included “No (intubation‐based) airway management is performed” (cited by 39% of all respondents), “Endotracheal intubation is performed” (7.7% of all respondents), “A laryngeal mask is used” (7.1% of all respondents), and “case‐by‐case” (0.9% of all respondents) (Figure S6).

4.3.3. Deep Sedation MRI: Medications Used

The primary drugs used for deep sedation/general anesthesia in adults (multiple selections allowed) were Midazolam (Dormicum) (reported by 38% of all respondents), Diazepam (Cercine) (26% of all respondents), Propofol (Diprivan) (24% of all respondents), barbiturates (Ravonal) (7.3% of all respondents), and Dexmedetomidine (Precedex) (6.0% of all respondents) (Figure 4A). The primary drugs used for minimal‐to‐moderate sedation in adults (multiple selections allowed) were Etizolam (Depas) (31% of all respondents), Diazepam (Cercine) (26% of all respondents), Hydroxyzine Pamoate (Atarax‐P) (24% of all respondents), Flunitrazepam (Silece) (10% of all respondents), and Thiamylal (Ravonal) (7.6% of all respondents). Other responses included Dexmedetomidine (Precedex) and Chlorpromazine Hydrochloride (Contamin) (Figure 4B).

FIGURE 4.

Drugs used for adult sedation. (A) The top gray bar shows the percentage of all responding institutions that perform any deep sedation for adults (58%). Among these facilities, the most frequently used medications were benzodiazepines such as Midazolam (38% of all respondents) and Diazepam (26% of all respondents), followed by Propofol (24% of all respondents). (B) The top gray bar shows the percentage of all responding institutions that perform minimal‐to‐moderate sedation in adults (69%). Among these, the most common choices were oral benzodiazepines and related drugs, such as Etizolam (31% of all respondents) and Diazepam (26% of all respondents). The “Unknown” category likely represents cases where sedation is managed by each respective clinical department, and the specific agent used is not communicated to others, including the radiology department.

4.3.4. Pediatric Sedation: Consultation & Personnel

For pediatric examinations, the parties consulted for sedation (multiple selections allowed) were the “Pediatrics department” (cited by 45% of all respondents) and the “Attending physician” (7.7% of all respondents). A “case‐by‐case” approach was taken by 6.2% of all respondents, with the anesthesiology department being consulted in 0.9% of all respondent cases (Figure 5A). The personnel responsible for administering deep sedation (multiple selections allowed) were overwhelmingly “Pediatricians” (cited by 53% of all respondents), followed by “Attending physicians from other specialties” (9.8% of all respondents) and “Anesthesiologists” (3.6% of all respondents) (Figure 5B).

FIGURE 5.

Personnel and decision‐making for deep sedation MRI. A series of charts illustrating the management structure for sedation. For sub‐figures (A–D) and (F), the top gray bar shows the percentage of all 549 responding institutions that perform the corresponding type of sedated examination. (A, B) In pediatrics, sedation is predominantly managed by pediatricians, who are the most consulted party (45% of all respondents) and the primary administrators (53% of all respondents). (C, D) In adults, the attending physician is the primary decision‐maker (44% of all respondents) and administrator (49% of all respondents), with limited involvement from anesthesiologists. (E) The indication for the scan is typically decided by non‐radiology physicians for both pediatric and adult cases. (F) The person administering the sedation (“Sedationist”) is also the most common person responsible for monitoring.

4.3.5. Adult Sedation: Consultation & Personnel

For adult examinations, the decision on sedation was most often made by the “Attending physician” (reported by 44% of all respondents). A “case‐by‐case” approach was taken by 24% of all respondents, and “Anesthesiologists” were consulted in 6.2% of all respondent cases. Other responses included consulting the “Emergency Department”, “Psychiatry”, or designated “Hospital Committees” (Figure 5C). The personnel responsible for administering deep sedation (multiple selections allowed) were predominantly the “Attending physician” (cited by 49% of all respondents), followed by “Anesthesiologists” (12% of all respondents) and “Emergency Department” (4.9% of all respondents). Other responses included “Psychiatry”, “Radiology”, “Nurse”, and “Neurosurgery” (Figure 5D).

4.3.6. Indication for Deep Sedation MRI: Radiology Consultation

When planning a deep sedation MRI, the decision on the indication for the examination (multiple selections allowed) was most commonly made by “non‐radiology physicians (attending physician or other consulting doctors)” for both pediatric (cited by 30% of all respondents) and adult cases (28% of all respondents). A “case‐by‐case” approach was the next most common response for both pediatric (21% of all respondents) and adult cases (22% of all respondents). The practice of “always consulting” a radiologist was reported by 3.1% of all respondents for pediatric cases and 5.1% for adult cases (Figure 5E).

4.3.7. Deep Sedation: Monitoring Personnel and Methods

Among facilities performing deep sedation, the personnel responsible for monitoring during deep sedation (multiple selections allowed) were most commonly the “Sedationist” (cited by 38% of all respondents for pediatric and 32% of all respondents for adult cases), followed by the “Referring physician” (20% pediatric, 25% adult), “Nurses” (17% pediatric, 16% adult), and “MR Technologists” (13% pediatric, 14% adult) (Figure 5F). The monitoring methods used (multiple selections allowed) were “SpO₂ (peripheral capillary oxygen saturation) monitor” (cited by 63% of all respondents), “Monitoring camera” (59%), “In‐room direct observation” (34%), “ECG (electrocardiography)” (16%), and “Capnometer” (15%) (Figure S7).

4.3.8. Deep Sedation MRI: Hospital Admission

Among all respondents, 62% performed any examination‐purposed deep sedation. The reported policies regarding hospital admission were as follows: “Hospitalization required” (cited by 27% of all respondents), “No specific rules” (24% of all respondents), and “Hospitalization required if intravenous anesthesia is used” (4.4% of all respondents). Other responses included “For pediatric patients, hospitalization required if deep sedation is performed” and “The anesthesiology department's judgement” (Figure 6A).

FIGURE 6.

Hospital admission policies and locations for sedation induction and awakening. (A) Hospital admission policies for scheduled deep sedation vary, with no single dominant standard. (B) The top bars, outlined in black, indicate that 49% and 40% of all responding institutions conduct scheduled exams under deep sedation for pediatric and adult patients, respectively. The most common locations for inducing deep sedation are the MRI scan room for pediatric cases (32% of all respondents) and a procedure room near the scan room for adult cases (26% of all respondents). Recovery most often occurs in inpatient wards (45% of all respondents) or outpatient procedure rooms (28% of all respondents).

4.3.9. Induction and Recovery Locations

The common location for inducing deep sedation (multiple selections allowed) was the “Scan room” (reported by 32% of all responses for pediatric cases and 13% of all responses for adult cases), a “Procedure room near the scan room” (29% pediatric, 26% adult), the “Inpatient ward” (13% pediatric, 14% adult), and an “Outpatient procedure room” (8.6% pediatric, 5.3% adult). For recovery, the “Inpatient ward” (45% of all respondents) and “Outpatient procedure room” (28% of all respondents) were the most common locations, followed by the “Procedure room near the scan room” and the “Scan room” itself (27% each). A small number of responses indicated the “Operating room” for adult induction and recovery (Figure 6B).

4.3.10. Rules for Ventilator and Syringe Pump Use

Regarding rules for ventilator use (multiple selections allowed), “A physician accompanies the equipment during the scan” (reported by 25% of all respondents) and “No established rules” (16% of all respondents) were the most commonly selected answers. For syringe pumps (multiple selections allowed), “No established rules” (reported by 13% of all respondents) was the most frequently chosen response. Few facilities reported restricting the use of ventilators and syringe pumps to designated locations (8.7% of all respondents for ventilators, 5.5% for syringe pumps) or specific scan rooms (7.5% of all respondents for ventilators, 3.3% for syringe pumps). A small number of facilities (1.1% of all respondents) reported using non‐compatible devices inside the scan room under specific conditions (Figure 7A,B).

FIGURE 7.

Rules for ventilator and syringe pump use. (A) The top gray bar shows the percentage of all responding institutions that have MRI‐compatible ventilators available (48%). Among these, the most common rule is having a physician accompany the device (25% of all respondents). (B) The top gray bar shows the percentage of all responding institutions that have MRI‐compatible syringe pumps available (29%). Among these, the most common response was having “no established rules” for their use (13% of all respondents).

4.3.11. Cabling and Tubing Through the Scan Room Door

When asked if they run cables or tubing through the scan room door (e.g., by leaving it partially open or pinching them) during deep sedation, 38% of all respondents responded “No,” while 18% reported they “Sometimes do,” and 2.7% reported they “Always do” (Figure S8).

5. Adverse Events

5.1. Facilities With Experience of Adverse Events

Although the highest number of facilities reported adverse events during non‐sedated adult examinations—a reflection of the procedure's high volume—a substantial number also reported adverse events during adult scans under deep sedation. Among the 320 institutions that perform any deep sedation on adults, specific adverse events experienced included: SpO₂ desaturation (48 facilities), respiratory arrest (20 facilities), bradycardia (21 facilities), cardiac arrest (19 facilities), burns (4 facilities), ventilator failure/damage (14 facilities), syringe pump failure/damage (5 facilities), and projectile accidents with equipment damage (8 facilities) (Figure 8).

FIGURE 8.

Number of facilities with experience of adverse events. The chart compares the number of facilities that have ever experienced specific adverse events, categorized by patient age (pediatric/adult) and sedation status (sedation/no sedation). The Y‐axis is presented on a logarithmic scale to accommodate the wide range in the number of reporting facilities. This chart displays the number of facilities that have experienced an adverse event and does not reflect the per‐procedure risk.

5.2. Adverse Events in the Past 2 Years

Similarly, analyzing events from the past 2 years, whereas the absolute number of adverse events under sedation for some categories was slightly lower than in non‐sedated exams, the frequency for the combined pediatric and adult population was still substantial. Patient‐related adverse events included 384 cases of SpO₂ desaturation (0.23%), 47 cases of respiratory arrest (0.03%), 55 cases of bradycardia (0.03%), 13 cases of cardiac arrest (0.01%), 3 cases of physical trauma (0.002%), and 1 case of burns (0.001%). Equipment‐related incidents included 11 cases of ventilator failure/damage (0.01%), 6 cases of syringe pump failure/damage (0.004%), 14 cases of monitor failure/damage (0.008%), and 7 projectile accidents with equipment damage (0.004%) (Figure 9). Fisher's exact test showed that the sedated group had a significantly higher incidence rate than the non‐sedated group, indicating a substantial risk increase for all patient‐related and equipment‐related adverse events, except for burns (p = 0.47). The magnitude of this increased risk, quantified by odds ratios, was as follows: 62 (95% confidence interval, 54–73) for SpO₂ desaturation, 48 (32–72) for respiratory arrest, 15 (8.0–28) for cardiac arrest, 6.7 (2.0–23) for physical trauma, and 8.9 (4.0–20) for projectile accidents (Table 1).

FIGURE 9.

Annual rate and number of adverse events in the past 2 years. The chart compares the annual rate (upper panel) and raw number of adverse events (lower panel) that occurred in the last 2 years, categorized by sedation status. Data for pediatric and adult patients are combined within each category. The Y‐axis is presented on a logarithmic scale to visualize the wide range of event counts.

TABLE 1.

Increased risk of adverse events from deep sedation (odds ratios).

Adverse event	SpO2 desaturation	Respiratory arrest	Bradycardia	Cardiac arrest	Physical trauma	Burns	Syringe pump damage	Monitor damage	Projectile accident
Odds ratios (95% CI)	62 (54–73)	48 (32–72)	46 (32–67)	15 (8.0–28)	6.7 (2.0–23)	2.1 (0.28–15)	71 (20–251)	60 (27–132)	8.9 (4.0–20)

Note: The table presents the calculated odds ratios for various adverse events, comparing the risk in the sedated group to the non‐sedated group based on two different calculation methods (experience at each facility and number of events in the past 2 years). The data show a dramatically increased risk for nearly all adverse events when sedation is used.

Abbreviations: CI, confidence interval; SpO₂, peripheral capillary oxygen saturation.

6. Discussion

This nationwide Japanese survey quantitatively demonstrates that MRI examinations performed under sedation are associated with a significantly increased risk of adverse events than non‐sedated examinations. A detailed analysis of the survey findings reveals vulnerabilities in sedation management, clinical procedures, and the physical environment itself, which may contribute to the increased procedural risks observed in sedated patients.

6.1. The Attending Physician‐Led Sedation Management

Our findings clarify that the management system for sedated adult MRI is, in the majority of facilities, handled independently by attending physicians from various non‐anesthesiology specialties. The involvement of anesthesiologists is minimal from the initial decision‐making to the administration of sedation. This situation is not unique to Japan, as a recent survey in France, for example, reported that radiologists perform pediatric CT or MRI sedation without an anesthesiologist in 55% of centers [6]. However, the attending physician‐led model observed in our survey for adults contrasts sharply with other established non‐anesthesiologist models that rely on a robust structure and specialization. A prime example is found in the large US‐based Pediatric Sedation Research Consortium, which reports a safe pediatric model where care is primarily delivered by pediatric specialists trained in acute care, such as pediatric intensivists (49.3%) and pediatric emergency medicine physicians (28.2%), with anesthesiologists accounting for only 9.8% of cases [11]. Similarly, Japan has its own structured system for pediatric sedation, which is overwhelmingly managed by trained pediatricians and has its own established guidelines [8, 9]. Other successful models seen globally include those led by highly specialized and trained MRI sedation nurses [12, 13, 14], whose positive effect on safety has been well established [14]. An important issue identified in our survey for adult sedation, therefore, is not merely the absence of anesthesiologists. Rather, it is the lack of a dedicated, specialized team—like the specialist or nurse‐led models—or consistent management by a designated lead practitioner. This results in sedation being handled inconsistently by individual physicians from various non‐acute care specialties.

The lack of a structured, multidisciplinary approach was also evident in the limited consultation with radiologists regarding the indication for the scan. The lack of multidisciplinary input suggests that the justification for a high‐risk sedated examination—including a thorough consideration of its necessity and potential alternatives—may be inadequately evaluated. This practice contrasts sharply with guidelines from the Association of Anesthetists of Great Britain and Ireland that explicitly mandate that the decision to perform an MRI scan, particularly for a patient in the intensive care unit, should be made by the consultant after discussion with the multidisciplinary team and radiologist [4].

Regarding the choice of sedative agents, our survey found that benzodiazepines are the predominant choice for minimal‐to‐moderate sedation. This finding is consistent with literature reporting the common use of benzodiazepines, which are familiar to non‐specialists and have a reversal agent [15]. However, the infrequent use of other agents like Propofol and Dexmedetomidine for minimal‐to‐moderate sedation might be noteworthy [16, 17]. Propofol, in particular, offers rapid recovery but carries a substantially higher risk of progressing to deep sedation and apnea, requiring advanced airway management skills [18]. Therefore, the most straightforward explanation for its limited use in our survey would be that, in a system predominantly run by non‐specialists, physicians logically opt for more familiar, lower‐risk agents over more potent drugs that require specialist‐level rescue capabilities.

6.2. Clinical Sedation Process

Beyond the management model, our findings highlight notable safety gaps in the clinical process itself. The most alarming of these are airway management and patient monitoring. A majority of facilities performing examination‐purposed deep sedation on adults do so without routinely securing the airway via an endotracheal tube or laryngeal mask airway. As for patient monitoring, despite international standards requiring a dedicated monitor separate from the sedationist [3, 4], the monitoring role was often delegated to personnel with other primary duties, such as MR technologists. Furthermore, even the referring physician—the most suitable candidate for a dedicated monitor—was involved in only about 25% of cases.

This situation is compounded by the low utilization of capnometers, a key tool for the early detection of respiratory depression [1, 19], a gap likely exacerbated by the device's high cost. As for the induction locations, our survey found that deep sedation is induced directly within the MRI scan room in a majority of cases (65%) for pediatric sedation. This practice is likely driven by workflow efficiency, particularly in settings where advanced airway intervention is not routinely planned. However, this approach appears inconsistent with the guiding principle that induction should occur in a “dedicated area” [4] to safely manage this high‐risk phase, when the risk of airway or hemodynamic complications is highest.

Despite these issues, there are also indications that some procedures are performed with caution due to safety concerns. Regarding off‐hours service, our survey revealed a significant discrepancy between routine and high‐risk procedures. Although a majority of institutions provide routine MRI examinations on nights and holidays, the performance of deep sedation MRI during these times is exceptional. This cautious approach to scheduling likely reflects safety concerns over reduced staffing and support out of hours. Such a practice is consistent with the principles of the UK and Ireland guideline, which recommends that only personnel who are experienced and thoroughly familiar with the MRI environment should manage patients, particularly during these more vulnerable, out‐of‐hours periods [4]. Our survey also found that a substantial portion of institutions performing scheduled deep sedation MRI do so on an inpatient basis. This practice may reflect an attempt to mitigate the potential risks of outpatient sedation, such as impaired psychomotor skills due to residual sedative effects [20], in a setting where management protocols for non‐anesthesiologist‐led sedation are not standardized. Indeed, these hazards are highlighted by a case of malpractice involving premature discharge and patients driving under the influence of sedatives, which can lead to serious post‐procedural accidents [21].

6.3. Environmental and Equipment Hazards

The physical environment of the MRI suite presents unmitigated risks, often stemming from a lack of both specialized equipment and protocols to govern its use. Less than half of the surveyed facilities were equipped with MRI‐compatible ventilators or syringe pumps, and many reported having no established rules specifically for the operation of these devices within the MRI environment. This procedural weakness was further exemplified by the number of facilities that improperly run cables through the scan room door—a workaround for an infrastructural deficit, as many lack a proper waveguide. Bypassing the door seal in this manner compromises the integrity of the radiofrequency shield—potentially degrading image quality [22] and posing a direct safety risk—which is contrary to the ACR's guidance on maintaining a closed scan room (Zone IV) [23]. This practice points to a gap in MRI suite design, a topic addressed by the UK‐Ireland guideline, which emphasizes the importance of anesthetic input during the planning phase to ensure a safe and functional environment [4].

Another notable source of risk involves the misuse of devices intended for the MR environment. Our survey identified reports of ventilator failure or damage at numerous facilities. Such incidents can stem from a misunderstanding of safety labels; even devices marketed as “MR Conditional” can contain shielded ferromagnetic components and are at risk of sudden, irreversible failure if brought closer to the magnetic isocenter than specified. Therefore, to mitigate the risk of inadvertent movement toward the magnet, the ACR manual recommends that MR Conditional and all MR Unsafe equipment entering Zone IV be physically tethered to a wall [10].

6.4. Increased Risk of Adverse Events Due to Sedation

A key finding of this study is that deep sedation markedly increases the risk of adverse events during MRI examinations. The calculated odds ratios were significantly high for all events, from patient‐related adverse events like SpO₂ desaturation, respiratory arrest, and cardiac arrest to equipment‐related projectile accidents aside from burns.

The background for this increased risk is likely a complex interplay of factors, including the patient's clinical severity, combined with the issues identified in this survey: an uncoordinated management system with insufficient rules, the inherent difficulty of monitoring within the MRI environment, and inadequate equipment. Under sedation, the number of medical devices and personnel entering the scan room increases, whereas the patient is unable to avoid danger actively. It is presumed that these deficiencies in safety management are directly linked to the occurrence of adverse events.

Comparing our findings on SpO₂ desaturation to previous reports, our results appear broadly comparable when accounting for differences in sedation depth and patient populations. For example, the incidence of desaturation during minimal‐to‐moderate sedation (adult and pediatric) has been reported at 0.17% in one study [14], whereas the Pediatric Sedation Research Consortium reported a rate of 1.7% for pediatric MRI sedation/anesthesia [11]. Our estimated incidence rate for desaturation during deep sedation was within this range. In this study, a notable number of facilities also reported experience in managing serious adverse events in both pediatric and adult patients. For adult examinations, the number of facilities with experience of cardiac arrest was proportionally higher when compared to pediatrics, which reflects differences in patient backgrounds, such as the higher prevalence of comorbidities in adults.

6.5. A Framework for a Safer System

In light of our findings, a fundamental policy to improve safety is to “avoid deep sedation whenever possible,” positioning it strictly as a last resort to be considered only after exhausting a range of alternatives. These alternatives include technological solutions like fast‐scanning sequences and motion correction [24], patient‐centered, non‐pharmacological approaches [25], and less intensive pharmacological options like the use of anxiolytics. The patient‐centered approaches range from established tools like communication devices [26] and music [27] to active, evidence‐based support. Examples include pre‐scan preparation for children [28] and structured communication for adults, where MR technologists actively engage with and verbally support the patient, which has been shown to significantly increase examination completion rates without sedation [29]. The underlying principles of these techniques—reducing anxiety through preparation, distraction, and communication—are adaptable for any patient.

Building a safer system beyond this foundational principle requires a multi‐faceted approach. Key components must include establishing a collaborative consultation system where specialists confirm the indications for both the examination and sedation, and defining clear training and credentialing requirements for all personnel involved. This must be supported by the standardization of monitoring methods, the development of clear rules for medical equipment use in the unique MRI environment, and the implementation of robust emergency response protocols.

Finally, to facilitate the widespread adoption of the safety measures mentioned above, policy‐level incentives could be beneficial. For instance, enhancing reimbursement for sedation management and creating public subsidy programs for the acquisition of necessary safety equipment, such as capnometers and MRI‐compatible pumps, would provide the practical support needed for institutions to invest in a higher standard of patient care.

7. Limitations

First, a precise response rate could not be calculated. The survey was distributed via society membership lists, and a definitive count of the unique medical institutions represented by these members was not available from the source data. This could introduce a selection bias, as facilities with a greater interest in safety management may have been more likely to respond. Second, all data are based on self‐reporting. Therefore, differences in the interpretation of questions or the subjectivity of the respondent may have influenced the results. Even when a consolidated response was provided, the accuracy of historical data may be affected by factors such as incomplete institutional archives for very old events. Third, while the odds ratios calculated in this study are powerful indicators that sedation substantially increases risk, it is worth noting that they are estimates based on event counts from the past 2 years. Fourth, to reduce respondent burden, the survey did not require the number of adverse events to be stratified by patient age (pediatric vs. adult). A consequence of this design is that our risk analysis necessarily combines both populations and does not provide age‐specific odds ratios.

8. Conclusion

This study provides a comprehensive, nationwide assessment of adult MRI sedation, quantitatively confirming that the practice is associated with substantially increased procedural risks. The findings revealed systemic vulnerabilities, including inconsistent management by non‐anesthesiology specialists, a lack of standardized safety protocols, and inadequate equipment. The substantial risks demonstrated by our analysis underscore the need for a more systematic and cautious approach to deep sedation. Establishing and implementing standardized, evidence‐based safety recommendations for adult MRI sedation is likely an important step toward mitigating these identified risks and improving patient care.

Supporting information

Data S1: jmri70149‐sup‐0001‐supinfo.docx.

Amemiya S., Kataoka M., Doi T., Hirai T., Obata T., and Kuroda K., “Safety of MRI Examinations Under Sedation: A Nationwide Survey in Japan,” Journal of Magnetic Resonance Imaging 63, no. 3 (2026): 798–810, 10.1002/jmri.70149.