https://orcid.org/0000-0003-3142-4503
Abstract
Current guidelines recommend 6 h of solid food and 2 h of clear liquid fasting for patients undergoing cardiac procedures with conscious sedation. There are no data to support this practice, and previous single-centre studies support the safety of removing fasting requirements. The objective of this study was to determine the non-inferiority of a no-fasting strategy to fasting prior to cardiac catheterization procedures which require conscious sedation.
This is a multicentre, investigator-initiated, non-inferiority, randomized trial conducted in Australia with a prospective open-label, blinded endpoint design. Patients referred for coronary angiography, percutaneous coronary intervention, or cardiac implantable electronic device (CIED)-related procedures were enrolled. Patients were randomized 1:1 to fasting as normal (6 h solid food and 2 h clear liquid) or no-fasting requirements (encouraged to have regular meals but not mandated to do so). Recruitment occurred from 2022 to 2023. The primary outcome was a composite of aspiration pneumonia, hypotension, hyperglycaemia, and hypoglycaemia assessed with a Bayesian approach. Secondary outcomes included patient satisfaction score, new ventilation requirement (non-invasive and invasive), new intensive care unit admission, 30-day readmission, 30-day mortality, 30-day pneumonia.
A total of 716 patients were randomized with 358 in each group. Those in the fasting arm had significantly longer solid food fasting (13.2 vs. 3.0 h, Bayes factor >100, indicating extreme evidence of difference) and clear liquid fasting times (7.0 vs. 2.4 h, Bayes factor >100). The primary composite outcome occurred in 19.1% of patients in the fasting arm and 12.0% of patients in the no-fasting arm. The estimate of the mean posterior difference in proportions with credibility interval (CI) in the primary composite outcome was −5.2% (95% CI −9.6 to −.9), favouring no fasting. This result confirms the non-inferiority (posterior probability >99.5%) and superiority (posterior probability 99.1%) of no fasting for the primary composite outcome. The no-fasting arm had improved patient satisfaction scores with a posterior mean difference of 4.02 points (95% CI 3.36–4.67, Bayes factor >100). Secondary outcome events were observed to be similar.
In patients undergoing cardiac catheterization and CIED-related procedures, no fasting was non-inferior and superior to fasting for the primary composite outcome of aspiration pneumonia, hypotension, hyperglycaemia, and hypoglycaemia. Patient satisfaction scores were significantly better with no fasting. This supports removing fasting requirements for patients undergoing cardiac catheterization laboratory procedures that require conscious sedation.
Seven hundred and sixteen patients randomised to fasting or no fasting prior to cardiac catheterization procedures with the primary composite outcome results. IQR, interquartile range; Vs, verse; CI, credibility interval.
See the editorial comment for this article ‘Fasting prior to percutaneous cardiovascular procedures: is it time to abandon this time-honoured practice?’, by S. Bangalore and M.H. Maqsood, https://doi.org/10.1093/eurheartj/ehae754.
Introduction
Every year, millions of individuals undergo coronary angiography, percutaneous coronary intervention, and cardiac device implantation with physician-guided conscious sedation.1,2 The American Society of Anesthesiologists recommends routine fasting prior to these procedures.3 This involves fasting from solid food for 6 h and fasting from clear liquids for 2 h. Fasting is implemented to reduce aspiration risk, though there is no outcome data to support this practice when using conscious sedation. The rates of emesis were significantly higher with older, high osmolality, contrast agents.4 The risk of aspiration with conscious sedation appears low. An observational study of over 3600 patients undergoing unfasted coronary angiography demonstrated no aspiration events with improved patient satisfaction.5
In real-world practice, patients undergoing procedures fast significantly longer than required.6–8 There are potential harms from fasting, including patient discomfort, poor hydration, contrast-associated acute kidney injury, hypotension, and poor glycaemic control.9 There are potential benefits in removing fasting requirements, including reduced thirst and hunger, improved patient satisfaction, and reduced procedural delays and cancellations.10
Single-centre, randomized trials have suggested non-inferiority of no fasting when compared with fasting before cardiac catheterization procedures.11–14 Observational studies also support the safety of removing fasting requirements in this population.5,6,15 The purpose of this multicentre, randomized study is to assess the non-inferiority of no fasting compared with fasting prior to cardiac catheterization laboratory procedures requiring conscious sedation.
Methods
Trial design
This is an investigator-initiated, multicentre, randomized, non-inferiority trial with a prospective, open-label, blinded endpoint design. The methodology of this study has been published previously.16 Inclusion criteria were age of 18 years and older and referral for coronary angiography, percutaneous coronary intervention, or implantable cardiac device–related procedure. Exclusion criteria were requirement for general anaesthesia, emergent procedures, structural cardiac interventions, planned coronary atherectomy or lithotripsy, electrophysiology studies, and cardiac resynchronization therapy. Patients were not excluded based on obstructive sleep apnoea or body mass index. Recruitment occurred across six sites, both tertiary and regional, with procedures performed at three sites in New South Wales, Australia. Patients were randomized 1:1 to fasting (6 h solid food and 2 h clear liquids) or to no-fasting requirements (Structured Graphical Abstract). This was designed as a pragmatic, ‘real-world’ randomized trial. Proceduralists were permitted to know fasting times. Those in the no-fasting arm were encouraged to eat their regular meals to mimic usual intake but were not mandated to do so. A research member not involved with patient recruitment created a randomly generated allocation table which was uploaded onto REDCAP. Single sequence randomization was performed with stratification by procedure location and type (coronary procedure vs. device procedure). Recruitment occurred from October 2022 to September 2023. The trial was performed in accordance with the Declaration of Helsinki. The first author vouches for the accuracy of the data and fidelity of the trial to the protocol.
Peri-procedural instructions and care
Sodium-glucose co-transporter 2 inhibitors and metformin were withheld for 24 h before and restarted 48 h after. If on insulin, those not required to fast were instructed to administer their usual doses. In the fasting arm, those receiving basal/bolus insulin were instructed to omit morning short-acting therapy, but to continue long-acting insulin at 100% of the usual dose (decreased to 80% if there was concern for hypoglycaemia), while morning mixed insulin doses were halved. Trial investigators were not involved in peri-procedural glucose or blood pressure management unless they had already met the primary outcome. Pre-procedural hydration was not performed routinely. The POSEIDON protocol was recommended for those with glomerular filtration rates <45 mL/min/1.73 m2 without the evidence of fluid overload.17 Sedation was at the discretion of the treating physician and was recommended to involve fentanyl ± midazolam titrated to effect. The procedural contrast used was iohexol (Omnipaque™).
Endpoints
The primary endpoint was a composite of procedure-related aspiration pneumonia, hypotension, hyperglycaemia, or hypoglycaemia (see Supplementary data online, Figure S1). Aspiration pneumonia was defined as a clinical syndrome of tachypnoea/hypoxia/fever/tachycardia with X-ray changes after an aspiration event.18 Hypotension was defined as a systolic blood pressure of <90 mm Hg at any point on the day of the procedure (including during the procedure), or a blood pressure that required intravenous fluid resuscitation or medication administration. Hyperglycaemia was defined as a blood sugar level (BSL) of >11 mmol/L.19 Hypoglycaemia was defined as a BSL of <3.9 mmol/L.20 Blood sugar level was measured within 2 h of scheduled procedure start for all participants. If there was clinical concern, symptoms, or an initial abnormal BSL, repeat BSL testing was performed. If a patient was diabetic, a post-procedure BSL was also performed.
Secondary endpoints included each component of the composite primary outcome, as well as new intensive care unit admission post-procedure, new non-invasive or invasive ventilation requirement post-procedure, 30-day mortality, 30-day readmission, and 30-day pneumonia events. Patients completed quality of life assessments pre-procedure, including the Euro Quality of Life 5-Dimension and a pre-procedure patient satisfaction questionnaire validated for those fasting for cataract surgery.21 This fasting questionnaire includes six questions about fasting symptoms with a five-point ordinal scale ranging from strongly disagree to strongly agree (see Supplementary data online, Table S2). A total score was formed as the sum of the response to each of the six questions, higher scores corresponding to worse symptoms (interpreted as lower levels of satisfaction) with a maximum score of 30. All endpoints were adjudicated by a blinded clinical events committee consisting of an interventional cardiologist, electrophysiologist, and cardiac anaesthetist.
Sample size calculation
Based on data from a similar study, CHOW NOW, our stipulated composite primary outcome was anticipated to occur in 5.9% of the fasting arm compared with 5.1% in the no-fasting arm.14 We assumed that the primary outcome likelihood would follow a binomial distribution within each treatment arm with parameters reflecting the probability of a participant experiencing an outcome. Furthermore, to complete the Bayesian model, we specified that these proportions have informative beta prior distributions (reflecting our assumption that 95% of the probability mass was between 3% and 12%). The posterior distributions for these proportions, therefore, also follow beta distributions and are used to infer the probability of non-inferiority, defined here as the posterior probability that the difference in proportions is no more than the non-inferiority margin of 3%. The decision rule was to declare non-inferiority if this posterior probability exceeded 95%. Through simulations (n = 10 000), we explored the operating characteristics of this trial and determined that a sample size of 300 patients per arm was required to correctly declare non-inferiority on 80% of simulations. We aimed to recruit a total cohort of 700 patients, 600 coronary and 100 device procedures.
Statistical analysis
All statistical analyses were conducted within the Bayesian framework. The posterior distribution of the between-group difference in the proportion of patients experiencing the primary outcome was determined with the model for the primary analysis constructed as described for the sample size calculation. The no-fasting arm was declared to be non-inferior to fasting if there was a 95% or greater probability that the difference between groups was <3% (i.e. an increase of <3% in the composite outcome compared with the fasting group). The 95% credible intervals of the posterior mean difference are presented and were calculated using the highest posterior density method. If the requirement for non-inferiority was met, testing for the superiority for the primary endpoint was performed.
Device and coronary procedure groups were analysed as a single cohort. The primary analysis was performed in the intention-to-treat population, with multiple imputation conducted using the chained regression equations method to manage missing outcome data. Sensitivity analyses of the primary outcome were performed in the per-protocol population with the fasting group defined as those who were randomized to the fasting group and completed fasting requirements. The per-protocol no-fasting group was defined as those who were randomized to the no-fasting group and did not complete 6 h solid food fasting and 2 h clear liquid fasting. Sensitivity analyses were conducted as to the robustness of the result to the choice of prior via the use of uninformative beta priors. Pre-specified post hoc (subgroup) analyses, with tests for interaction, were conducted using Bayesian logistic regression models, adjusted by stratification variables. Subgroup analyses included diagnostic alone vs. percutaneous interventional coronary procedures, coronary vs. device-related procedures, body mass index (<30 vs. ≥30), age (<65 years vs. ≥65 years), inpatients vs. outpatient, and provision of sedation.
Where event rates of >5% were observed, secondary outcomes related to events were investigated using Bayesian logistic regression models. Secondary outcomes corresponding to pre-procedural quality of life assessments were investigated using linear regression models. All models were adjusted for stratification variables.
All Bayesian regression models used for each analysis type had samples obtained from the posterior distribution using a No-UTURN sampler, using a default setting of 4 chains of 2000 iterations with 1000 burn-ins. Weakly informative priors following a normal distribution with a wide variance (1000) were placed on regression parameters. Convergence was confirmed via inspection of the effective sample size and Rhat values for each chain. Posterior predictive checks were used to confirm model fit was appropriate.
Continuous variables are presented as means with standard deviations or medians with interquartile ranges. Categorical variables are stated as absolute and relative frequencies. Exploratory Bayes factor analyses were conducted for selected variables to help fully characterize the intervention effect. A Bayes factor of >100 indicates decisive evidence of a difference between groups. Comparisons of the standardized mean difference were made for pre-procedure fasting questionnaire total score by intervention, length of time fasting (solids and clear liquids) by intervention, length of stay by intervention, and dose of sedative by procedure type. Bayes factor analysis of the independence of provision of sedation to intervention assignment was also undertaken. All statistical analyses were performed using R version 4.3.2 (R Foundation of Statistical Computing).
Results
A total of 716 patients were enrolled in this trial, with 358 allocated to fasting and 358 allocated to no fasting (Figure 1). Baseline characteristics were similar between groups (Table 1). About 65% of participants were male with an average age of 69 ± 11.1 years. Basline medications can be found in Supplementary data online, Table S1. Pre-procedural EQ-5D-5L scores were similar (see Supplementary data online, Table S3). Decisive evidence was found for a difference in average fasting times between groups for both solid food (13.2 vs. 3.0 h, Bayes factor >100) and clear liquids (7.0 vs. 2.4 h, Bayes factor >100; Table 2, Supplementary data online, Tables S6 and S7).
Trial consort diagram
Baseline characteristics stratified by group allocation
| Baseline characteristic | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| Age (years), mean ± SD | 70 ± 11.4 | 69 ± 10.9 |
| Male sex, n (%) | 231/358 (64.5) | 236/358 (65.9) |
| Body mass index,a mean ± SD | 29.6 ± 5.9 | 29.6 ± 5.9 |
| Hypertension, n/total n (%) | 250/358 (69.8) | 249/358 (69.6) |
| Hyperlipidaemia, n/total n (%) | 183/358 (51.1) | 171/358 (47.8) |
| Myocardial infarction, n/total n (%) | 72/358 (20.1) | 89/356 (25) |
| Percutaneous coronary intervention, n/total n (%) | 68/358 (19.0) | 75/358 (20.9) |
| Coronary artery bypass grafting, n/total n (%) | 23/358 (6.4) | 30/358 (8.4) |
| Heart failure,b n/total n (%) | 43/358 (12.0) | 51/358 (14.3) |
| Atrial fibrillation, n/total n (%) | 63/358 (17.6) | 66/358 (18.4) |
| Chronic kidney disease, n/total n (%) | 77/350 (22.0) | 69/354 (19.5) |
| Type 2 diabetes mellitus, n/total n (%) | 97/358 (27.1) | 95/358 (26.5) |
| Current smoker, n/total n (%) | 49/354 (13.8) | 66/356 (18.5) |
| Previous smoking, n/total n (%) | 155/353 (43.9) | 149/352 (42.3) |
| Ischaemic stroke/TIA, n/total n (%) | 20/356 (5.6) | 29/357 (8.1) |
| Haemorrhagic stroke, n/total n (%) | 5/358 (1.4) | 2/357 (.6) |
| Baseline observationsc | ||
| Systolic blood pressure (mm Hg), mean ± SD | 143 ± 22 | 140 ± 21 |
| Heart rate (b.p.m.), mean ± SD | 69 ± 13 | 69 ± 14 |
| Baseline serology | ||
| Creatinine (µmol/L), mean ± SD | 98 ± 71 | 97 ± 70 |
| Glomerular filtration rated (mL/min/1.73 m2), mean ± SD | 70 ± 20 | 71 ± 19 |
| Haemoglobin (g/L), mean ± SD | 137 ± 17 | 137 ± 18 |
| Baseline characteristic | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| Age (years), mean ± SD | 70 ± 11.4 | 69 ± 10.9 |
| Male sex, n (%) | 231/358 (64.5) | 236/358 (65.9) |
| Body mass index,a mean ± SD | 29.6 ± 5.9 | 29.6 ± 5.9 |
| Hypertension, n/total n (%) | 250/358 (69.8) | 249/358 (69.6) |
| Hyperlipidaemia, n/total n (%) | 183/358 (51.1) | 171/358 (47.8) |
| Myocardial infarction, n/total n (%) | 72/358 (20.1) | 89/356 (25) |
| Percutaneous coronary intervention, n/total n (%) | 68/358 (19.0) | 75/358 (20.9) |
| Coronary artery bypass grafting, n/total n (%) | 23/358 (6.4) | 30/358 (8.4) |
| Heart failure,b n/total n (%) | 43/358 (12.0) | 51/358 (14.3) |
| Atrial fibrillation, n/total n (%) | 63/358 (17.6) | 66/358 (18.4) |
| Chronic kidney disease, n/total n (%) | 77/350 (22.0) | 69/354 (19.5) |
| Type 2 diabetes mellitus, n/total n (%) | 97/358 (27.1) | 95/358 (26.5) |
| Current smoker, n/total n (%) | 49/354 (13.8) | 66/356 (18.5) |
| Previous smoking, n/total n (%) | 155/353 (43.9) | 149/352 (42.3) |
| Ischaemic stroke/TIA, n/total n (%) | 20/356 (5.6) | 29/357 (8.1) |
| Haemorrhagic stroke, n/total n (%) | 5/358 (1.4) | 2/357 (.6) |
| Baseline observationsc | ||
| Systolic blood pressure (mm Hg), mean ± SD | 143 ± 22 | 140 ± 21 |
| Heart rate (b.p.m.), mean ± SD | 69 ± 13 | 69 ± 14 |
| Baseline serology | ||
| Creatinine (µmol/L), mean ± SD | 98 ± 71 | 97 ± 70 |
| Glomerular filtration rated (mL/min/1.73 m2), mean ± SD | 70 ± 20 | 71 ± 19 |
| Haemoglobin (g/L), mean ± SD | 137 ± 17 | 137 ± 18 |
SD, standard deviation; TIA, transient ischaemic attack.
aBody mass index was calculated by weight in kilograms divided by the square of the height in metres. Data were missing for 23 patients in the control and 21 patients in the intervention group.
bHeart failure refers to those with an ejection fraction <50%.
cBaseline observation taken prior to procedure.
dThe chronic kidney disease epidemiology collaboration (CKD-EPI) equation was used to calculate glomerular filtration rate.
Baseline characteristics stratified by group allocation
| Baseline characteristic | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| Age (years), mean ± SD | 70 ± 11.4 | 69 ± 10.9 |
| Male sex, n (%) | 231/358 (64.5) | 236/358 (65.9) |
| Body mass index,a mean ± SD | 29.6 ± 5.9 | 29.6 ± 5.9 |
| Hypertension, n/total n (%) | 250/358 (69.8) | 249/358 (69.6) |
| Hyperlipidaemia, n/total n (%) | 183/358 (51.1) | 171/358 (47.8) |
| Myocardial infarction, n/total n (%) | 72/358 (20.1) | 89/356 (25) |
| Percutaneous coronary intervention, n/total n (%) | 68/358 (19.0) | 75/358 (20.9) |
| Coronary artery bypass grafting, n/total n (%) | 23/358 (6.4) | 30/358 (8.4) |
| Heart failure,b n/total n (%) | 43/358 (12.0) | 51/358 (14.3) |
| Atrial fibrillation, n/total n (%) | 63/358 (17.6) | 66/358 (18.4) |
| Chronic kidney disease, n/total n (%) | 77/350 (22.0) | 69/354 (19.5) |
| Type 2 diabetes mellitus, n/total n (%) | 97/358 (27.1) | 95/358 (26.5) |
| Current smoker, n/total n (%) | 49/354 (13.8) | 66/356 (18.5) |
| Previous smoking, n/total n (%) | 155/353 (43.9) | 149/352 (42.3) |
| Ischaemic stroke/TIA, n/total n (%) | 20/356 (5.6) | 29/357 (8.1) |
| Haemorrhagic stroke, n/total n (%) | 5/358 (1.4) | 2/357 (.6) |
| Baseline observationsc | ||
| Systolic blood pressure (mm Hg), mean ± SD | 143 ± 22 | 140 ± 21 |
| Heart rate (b.p.m.), mean ± SD | 69 ± 13 | 69 ± 14 |
| Baseline serology | ||
| Creatinine (µmol/L), mean ± SD | 98 ± 71 | 97 ± 70 |
| Glomerular filtration rated (mL/min/1.73 m2), mean ± SD | 70 ± 20 | 71 ± 19 |
| Haemoglobin (g/L), mean ± SD | 137 ± 17 | 137 ± 18 |
| Baseline characteristic | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| Age (years), mean ± SD | 70 ± 11.4 | 69 ± 10.9 |
| Male sex, n (%) | 231/358 (64.5) | 236/358 (65.9) |
| Body mass index,a mean ± SD | 29.6 ± 5.9 | 29.6 ± 5.9 |
| Hypertension, n/total n (%) | 250/358 (69.8) | 249/358 (69.6) |
| Hyperlipidaemia, n/total n (%) | 183/358 (51.1) | 171/358 (47.8) |
| Myocardial infarction, n/total n (%) | 72/358 (20.1) | 89/356 (25) |
| Percutaneous coronary intervention, n/total n (%) | 68/358 (19.0) | 75/358 (20.9) |
| Coronary artery bypass grafting, n/total n (%) | 23/358 (6.4) | 30/358 (8.4) |
| Heart failure,b n/total n (%) | 43/358 (12.0) | 51/358 (14.3) |
| Atrial fibrillation, n/total n (%) | 63/358 (17.6) | 66/358 (18.4) |
| Chronic kidney disease, n/total n (%) | 77/350 (22.0) | 69/354 (19.5) |
| Type 2 diabetes mellitus, n/total n (%) | 97/358 (27.1) | 95/358 (26.5) |
| Current smoker, n/total n (%) | 49/354 (13.8) | 66/356 (18.5) |
| Previous smoking, n/total n (%) | 155/353 (43.9) | 149/352 (42.3) |
| Ischaemic stroke/TIA, n/total n (%) | 20/356 (5.6) | 29/357 (8.1) |
| Haemorrhagic stroke, n/total n (%) | 5/358 (1.4) | 2/357 (.6) |
| Baseline observationsc | ||
| Systolic blood pressure (mm Hg), mean ± SD | 143 ± 22 | 140 ± 21 |
| Heart rate (b.p.m.), mean ± SD | 69 ± 13 | 69 ± 14 |
| Baseline serology | ||
| Creatinine (µmol/L), mean ± SD | 98 ± 71 | 97 ± 70 |
| Glomerular filtration rated (mL/min/1.73 m2), mean ± SD | 70 ± 20 | 71 ± 19 |
| Haemoglobin (g/L), mean ± SD | 137 ± 17 | 137 ± 18 |
SD, standard deviation; TIA, transient ischaemic attack.
aBody mass index was calculated by weight in kilograms divided by the square of the height in metres. Data were missing for 23 patients in the control and 21 patients in the intervention group.
bHeart failure refers to those with an ejection fraction <50%.
cBaseline observation taken prior to procedure.
dThe chronic kidney disease epidemiology collaboration (CKD-EPI) equation was used to calculate glomerular filtration rate.
Procedural characteristics stratified by group allocation
| Procedural characteristic | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| In-patient procedure, n/total n (%) | 122/358 (34.1) | 125/358 (34.9) |
| Procedure type | ||
| Percutaneous coronary procedure, n/total n (%) | 297/358 (83.0) | 306/358 (85.5) |
| Radial access attempted, n/total n (%) | 288/297 (97) | 293/304 (95.7) |
| Femoral access attempted, n/total n (%) | 17/297 (5.7) | 18/304 (5.9) |
| Diagnostic, n/total n (%) | 207/297 (69.7) | 193/306 (63.1) |
| Coronary intervention,a n/total n (%) | 90/297 (30.3) | 113/306 (36.9) |
| Physiology assessment, n/total n (%) | 19/297 (6.4) | 29/306 (9.5) |
| Coronary stent/balloon, n/total n (%) | 72/297 (24.2) | 83/306 (27.1) |
| Number of stents | 1.4 ± .7 | 1.4 ± .7 |
| Left main intervention | 2/72 (2.8) | 5/83 (6.0) |
| Left anterior descending artery | 31/72 (43.1) | 39/83 (47.0) |
| Right/circumflex artery | 44/72 (61.1) | 51/83 (61.4) |
| Graft intervention | 1/72 (1.4) | 3/83 (3.6) |
| Unplanned calcium modification | 0/72 (0) | 2/83 (2.4) |
| Device procedure, n/total n (%) | 61/358 (17.0) | 52/358 (14.5) |
| Pacemaker, n/total n (%) | 40/61 (65.6) | 28/52 (53.8) |
| Defibrillator, n/total n (%) | 9/61 (14.8) | 4/52 (7.8) |
| Lead revision, n/total n (%) | 5/61 (8.2) | 2/52 (3.8) |
| Generator change, n/total n (%) | 7/61 (11.5) | 18/52 (34.6) |
| Sedation provision | 311/358 (87) | 318/358 (89) |
| Midazolam (mg), mean ± SD | 1.3 ± .75 | 1.2 ± .5 |
| Fentanyl (µg), mean ± SD | 40.3 ± 20.4 | 41.3 ± 24.4 |
| Fasting times | ||
| Solid food fasting time (IQR) | 13.2 (8.5–15.1) | 3.0 (1.8–4.2) |
| Clear liquid fasting time (IQR) | 7.0 (4.2–12.1) | 2.4 (1.2–3.5) |
| Procedural characteristic | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| In-patient procedure, n/total n (%) | 122/358 (34.1) | 125/358 (34.9) |
| Procedure type | ||
| Percutaneous coronary procedure, n/total n (%) | 297/358 (83.0) | 306/358 (85.5) |
| Radial access attempted, n/total n (%) | 288/297 (97) | 293/304 (95.7) |
| Femoral access attempted, n/total n (%) | 17/297 (5.7) | 18/304 (5.9) |
| Diagnostic, n/total n (%) | 207/297 (69.7) | 193/306 (63.1) |
| Coronary intervention,a n/total n (%) | 90/297 (30.3) | 113/306 (36.9) |
| Physiology assessment, n/total n (%) | 19/297 (6.4) | 29/306 (9.5) |
| Coronary stent/balloon, n/total n (%) | 72/297 (24.2) | 83/306 (27.1) |
| Number of stents | 1.4 ± .7 | 1.4 ± .7 |
| Left main intervention | 2/72 (2.8) | 5/83 (6.0) |
| Left anterior descending artery | 31/72 (43.1) | 39/83 (47.0) |
| Right/circumflex artery | 44/72 (61.1) | 51/83 (61.4) |
| Graft intervention | 1/72 (1.4) | 3/83 (3.6) |
| Unplanned calcium modification | 0/72 (0) | 2/83 (2.4) |
| Device procedure, n/total n (%) | 61/358 (17.0) | 52/358 (14.5) |
| Pacemaker, n/total n (%) | 40/61 (65.6) | 28/52 (53.8) |
| Defibrillator, n/total n (%) | 9/61 (14.8) | 4/52 (7.8) |
| Lead revision, n/total n (%) | 5/61 (8.2) | 2/52 (3.8) |
| Generator change, n/total n (%) | 7/61 (11.5) | 18/52 (34.6) |
| Sedation provision | 311/358 (87) | 318/358 (89) |
| Midazolam (mg), mean ± SD | 1.3 ± .75 | 1.2 ± .5 |
| Fentanyl (µg), mean ± SD | 40.3 ± 20.4 | 41.3 ± 24.4 |
| Fasting times | ||
| Solid food fasting time (IQR) | 13.2 (8.5–15.1) | 3.0 (1.8–4.2) |
| Clear liquid fasting time (IQR) | 7.0 (4.2–12.1) | 2.4 (1.2–3.5) |
IQR, median with interquartile range; SD, standard deviation.
aCoronary intervention includes physiologic assessment, stenting, and plain old balloon angioplasty.
Procedural characteristics stratified by group allocation
| Procedural characteristic | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| In-patient procedure, n/total n (%) | 122/358 (34.1) | 125/358 (34.9) |
| Procedure type | ||
| Percutaneous coronary procedure, n/total n (%) | 297/358 (83.0) | 306/358 (85.5) |
| Radial access attempted, n/total n (%) | 288/297 (97) | 293/304 (95.7) |
| Femoral access attempted, n/total n (%) | 17/297 (5.7) | 18/304 (5.9) |
| Diagnostic, n/total n (%) | 207/297 (69.7) | 193/306 (63.1) |
| Coronary intervention,a n/total n (%) | 90/297 (30.3) | 113/306 (36.9) |
| Physiology assessment, n/total n (%) | 19/297 (6.4) | 29/306 (9.5) |
| Coronary stent/balloon, n/total n (%) | 72/297 (24.2) | 83/306 (27.1) |
| Number of stents | 1.4 ± .7 | 1.4 ± .7 |
| Left main intervention | 2/72 (2.8) | 5/83 (6.0) |
| Left anterior descending artery | 31/72 (43.1) | 39/83 (47.0) |
| Right/circumflex artery | 44/72 (61.1) | 51/83 (61.4) |
| Graft intervention | 1/72 (1.4) | 3/83 (3.6) |
| Unplanned calcium modification | 0/72 (0) | 2/83 (2.4) |
| Device procedure, n/total n (%) | 61/358 (17.0) | 52/358 (14.5) |
| Pacemaker, n/total n (%) | 40/61 (65.6) | 28/52 (53.8) |
| Defibrillator, n/total n (%) | 9/61 (14.8) | 4/52 (7.8) |
| Lead revision, n/total n (%) | 5/61 (8.2) | 2/52 (3.8) |
| Generator change, n/total n (%) | 7/61 (11.5) | 18/52 (34.6) |
| Sedation provision | 311/358 (87) | 318/358 (89) |
| Midazolam (mg), mean ± SD | 1.3 ± .75 | 1.2 ± .5 |
| Fentanyl (µg), mean ± SD | 40.3 ± 20.4 | 41.3 ± 24.4 |
| Fasting times | ||
| Solid food fasting time (IQR) | 13.2 (8.5–15.1) | 3.0 (1.8–4.2) |
| Clear liquid fasting time (IQR) | 7.0 (4.2–12.1) | 2.4 (1.2–3.5) |
| Procedural characteristic | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| In-patient procedure, n/total n (%) | 122/358 (34.1) | 125/358 (34.9) |
| Procedure type | ||
| Percutaneous coronary procedure, n/total n (%) | 297/358 (83.0) | 306/358 (85.5) |
| Radial access attempted, n/total n (%) | 288/297 (97) | 293/304 (95.7) |
| Femoral access attempted, n/total n (%) | 17/297 (5.7) | 18/304 (5.9) |
| Diagnostic, n/total n (%) | 207/297 (69.7) | 193/306 (63.1) |
| Coronary intervention,a n/total n (%) | 90/297 (30.3) | 113/306 (36.9) |
| Physiology assessment, n/total n (%) | 19/297 (6.4) | 29/306 (9.5) |
| Coronary stent/balloon, n/total n (%) | 72/297 (24.2) | 83/306 (27.1) |
| Number of stents | 1.4 ± .7 | 1.4 ± .7 |
| Left main intervention | 2/72 (2.8) | 5/83 (6.0) |
| Left anterior descending artery | 31/72 (43.1) | 39/83 (47.0) |
| Right/circumflex artery | 44/72 (61.1) | 51/83 (61.4) |
| Graft intervention | 1/72 (1.4) | 3/83 (3.6) |
| Unplanned calcium modification | 0/72 (0) | 2/83 (2.4) |
| Device procedure, n/total n (%) | 61/358 (17.0) | 52/358 (14.5) |
| Pacemaker, n/total n (%) | 40/61 (65.6) | 28/52 (53.8) |
| Defibrillator, n/total n (%) | 9/61 (14.8) | 4/52 (7.8) |
| Lead revision, n/total n (%) | 5/61 (8.2) | 2/52 (3.8) |
| Generator change, n/total n (%) | 7/61 (11.5) | 18/52 (34.6) |
| Sedation provision | 311/358 (87) | 318/358 (89) |
| Midazolam (mg), mean ± SD | 1.3 ± .75 | 1.2 ± .5 |
| Fentanyl (µg), mean ± SD | 40.3 ± 20.4 | 41.3 ± 24.4 |
| Fasting times | ||
| Solid food fasting time (IQR) | 13.2 (8.5–15.1) | 3.0 (1.8–4.2) |
| Clear liquid fasting time (IQR) | 7.0 (4.2–12.1) | 2.4 (1.2–3.5) |
IQR, median with interquartile range; SD, standard deviation.
aCoronary intervention includes physiologic assessment, stenting, and plain old balloon angioplasty.
The primary composite outcome occurred in 19.1% of the fasting arm vs. 12.0% in the no-fasting arm (Table 3). The requirements for both non-inferiority and superiority were met for the primary composite outcome, with an absolute posterior difference between the no-fasting and fasting group of −5.2% (95% CI −9.6 to −.9; posterior probability >99.5% for non-inferiority and 99.1% for superiority) favouring the no-fasting arm (Figure 2). This result was robust to the use of uninformative priors (see Supplementary data online, Table S16 and Figure S2). The results of the sensitivity analyses of the primary outcome performed in the per-protocol population were consistent with the primary analysis for non-inferiority and superiority. The observed rate of secondary outcomes was similar; this indicated that the difference in the composite primary endpoint was driven largely by hypotension and hyperglycaemia. No clinical aspiration events occurred during the trial. Two patients had pneumonia occurring within 30 days, both in the fasting arm. The blinded clinical events committee adjudicated one event as being ‘not related’ to peri-procedural aspiration, while the other was deemed ‘possibly related’ and was therefore not treated as an aspiration event. The event rate of contrast-associated acute kidney injury was 3.8% (3/78) in the fasting arm and 6.3% (5/80) in the no-fasting arm (Bayes factor of .11 supporting the null hypothesis; see Supplementary data online, Table S15).
Forest plot with credibility intervals for the primary composite outcome. Non-inferiority margin shown as the solid line at 3%
Primary and secondary outcomes
| Outcomes | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| Primary outcome | ||
| Composite of procedure-related aspiration pneumonia, hypotension, hyperglycaemia, or hypoglycaemia, n/total n (%) | 68/356 (19.1) | 43/356 (12.0) |
| Components of the primary outcome | ||
| Hypotension (systolic <90 mm Hg or intervention required), n/total n (%) | 32/358 (8.9) | 22/358 (6.1) |
| Fluid support for blood pressure, n/total n (%) | 7/32 (21.9) | 1/22 (4.5) |
| Medication support, n/total n (%) | 7/32 (21.9) | 4/22 (18.2) |
| Hyperglycaemia, n/total n (%) | 30/356 (8.4) | 23/356 (6.5) |
| Hypoglycaemia, n/total n (%) | 7/356 (2.0) | 2/356 (.6) |
| Aspiration pneumonia, n/total n (%) | 0/358 (0) | 0/358 (0) |
| Secondary outcomes | ||
| Contrast-associated acute kidney injury, n/total n (%) | 3/78 (3.8) | 5/80 (6.3) |
| New ventilation requirement, n/total n (%) | 0/358 (0) | 0/358 (0) |
| New intensive care unit admission, n/total n (%) | 0/358 (0) | 0/358 (0) |
| Readmission within 30 days, n/total n (%) | 28/358 (7.8) | 28/358 (7.8) |
| Death within 30 days, n/total n (%) | 1/358 (.3) | 1/358 (.3) |
| Pneumonia within 30 days, n/total n (%) | 2/358 (.6) | 0/358 (0) |
| Patient satisfaction score,a mean ± SD | 15 ± 4.3 | 11 ± 4.0 |
| Pre-procedure EQ-VAS,b mean ± SD | 70 ± 19 | 70 ± 20 |
| Outcomes | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| Primary outcome | ||
| Composite of procedure-related aspiration pneumonia, hypotension, hyperglycaemia, or hypoglycaemia, n/total n (%) | 68/356 (19.1) | 43/356 (12.0) |
| Components of the primary outcome | ||
| Hypotension (systolic <90 mm Hg or intervention required), n/total n (%) | 32/358 (8.9) | 22/358 (6.1) |
| Fluid support for blood pressure, n/total n (%) | 7/32 (21.9) | 1/22 (4.5) |
| Medication support, n/total n (%) | 7/32 (21.9) | 4/22 (18.2) |
| Hyperglycaemia, n/total n (%) | 30/356 (8.4) | 23/356 (6.5) |
| Hypoglycaemia, n/total n (%) | 7/356 (2.0) | 2/356 (.6) |
| Aspiration pneumonia, n/total n (%) | 0/358 (0) | 0/358 (0) |
| Secondary outcomes | ||
| Contrast-associated acute kidney injury, n/total n (%) | 3/78 (3.8) | 5/80 (6.3) |
| New ventilation requirement, n/total n (%) | 0/358 (0) | 0/358 (0) |
| New intensive care unit admission, n/total n (%) | 0/358 (0) | 0/358 (0) |
| Readmission within 30 days, n/total n (%) | 28/358 (7.8) | 28/358 (7.8) |
| Death within 30 days, n/total n (%) | 1/358 (.3) | 1/358 (.3) |
| Pneumonia within 30 days, n/total n (%) | 2/358 (.6) | 0/358 (0) |
| Patient satisfaction score,a mean ± SD | 15 ± 4.3 | 11 ± 4.0 |
| Pre-procedure EQ-VAS,b mean ± SD | 70 ± 19 | 70 ± 20 |
NIM, non-inferiority margin; SD, standard deviation; VAS, visual analogue scale.
aCumulative score, with higher scores corresponding to decreased patient satisfaction and worse symptoms.
bEQ-5D-5L results can be found in Supplementary data online, Table S3.
Primary and secondary outcomes
| Outcomes | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| Primary outcome | ||
| Composite of procedure-related aspiration pneumonia, hypotension, hyperglycaemia, or hypoglycaemia, n/total n (%) | 68/356 (19.1) | 43/356 (12.0) |
| Components of the primary outcome | ||
| Hypotension (systolic <90 mm Hg or intervention required), n/total n (%) | 32/358 (8.9) | 22/358 (6.1) |
| Fluid support for blood pressure, n/total n (%) | 7/32 (21.9) | 1/22 (4.5) |
| Medication support, n/total n (%) | 7/32 (21.9) | 4/22 (18.2) |
| Hyperglycaemia, n/total n (%) | 30/356 (8.4) | 23/356 (6.5) |
| Hypoglycaemia, n/total n (%) | 7/356 (2.0) | 2/356 (.6) |
| Aspiration pneumonia, n/total n (%) | 0/358 (0) | 0/358 (0) |
| Secondary outcomes | ||
| Contrast-associated acute kidney injury, n/total n (%) | 3/78 (3.8) | 5/80 (6.3) |
| New ventilation requirement, n/total n (%) | 0/358 (0) | 0/358 (0) |
| New intensive care unit admission, n/total n (%) | 0/358 (0) | 0/358 (0) |
| Readmission within 30 days, n/total n (%) | 28/358 (7.8) | 28/358 (7.8) |
| Death within 30 days, n/total n (%) | 1/358 (.3) | 1/358 (.3) |
| Pneumonia within 30 days, n/total n (%) | 2/358 (.6) | 0/358 (0) |
| Patient satisfaction score,a mean ± SD | 15 ± 4.3 | 11 ± 4.0 |
| Pre-procedure EQ-VAS,b mean ± SD | 70 ± 19 | 70 ± 20 |
| Outcomes | Fasting (N = 358) | No fasting (N = 358) |
|---|---|---|
| Primary outcome | ||
| Composite of procedure-related aspiration pneumonia, hypotension, hyperglycaemia, or hypoglycaemia, n/total n (%) | 68/356 (19.1) | 43/356 (12.0) |
| Components of the primary outcome | ||
| Hypotension (systolic <90 mm Hg or intervention required), n/total n (%) | 32/358 (8.9) | 22/358 (6.1) |
| Fluid support for blood pressure, n/total n (%) | 7/32 (21.9) | 1/22 (4.5) |
| Medication support, n/total n (%) | 7/32 (21.9) | 4/22 (18.2) |
| Hyperglycaemia, n/total n (%) | 30/356 (8.4) | 23/356 (6.5) |
| Hypoglycaemia, n/total n (%) | 7/356 (2.0) | 2/356 (.6) |
| Aspiration pneumonia, n/total n (%) | 0/358 (0) | 0/358 (0) |
| Secondary outcomes | ||
| Contrast-associated acute kidney injury, n/total n (%) | 3/78 (3.8) | 5/80 (6.3) |
| New ventilation requirement, n/total n (%) | 0/358 (0) | 0/358 (0) |
| New intensive care unit admission, n/total n (%) | 0/358 (0) | 0/358 (0) |
| Readmission within 30 days, n/total n (%) | 28/358 (7.8) | 28/358 (7.8) |
| Death within 30 days, n/total n (%) | 1/358 (.3) | 1/358 (.3) |
| Pneumonia within 30 days, n/total n (%) | 2/358 (.6) | 0/358 (0) |
| Patient satisfaction score,a mean ± SD | 15 ± 4.3 | 11 ± 4.0 |
| Pre-procedure EQ-VAS,b mean ± SD | 70 ± 19 | 70 ± 20 |
NIM, non-inferiority margin; SD, standard deviation; VAS, visual analogue scale.
aCumulative score, with higher scores corresponding to decreased patient satisfaction and worse symptoms.
bEQ-5D-5L results can be found in Supplementary data online, Table S3.
Subgroup analyses showed that the effect size was similar between the coronary and device subgroups (see Supplementary data online, Figure S3 and Table S12). Decisive evidence was found for improved patient satisfaction scores in the no-fasting arm compared with the fasting arm (11 ± 4.0 vs. 15 ± 4.3 points, Bayes factor >100; see Supplementary data online, Tables S2 and S4). There was no difference in length of stay (see Supplementary data online, Tables S10 and S11).
Discussion
This multicentre trial, across regional and tertiary hospitals, demonstrated the superiority of no fasting prior to cardiac catheterization procedures for the primary composite outcome of aspiration pneumonia, hypotension, hyperglycaemia, and hypoglycaemia with an intention to treat analysis. There were also improvements in pre-procedural patient satisfaction scores. These results suggest that the removal of fasting requirements prior to cardiac catheterization laboratory procedures that require conscious sedation is safe and improves patient centred outcomes.
Anaesthetic guidelines have recommended fasting prior to procedures that require conscious sedation. Patient safety is the well-intentioned purpose behind this recommendation, though evidence to support this practice is poor. Though there are theoretical benefits to fasting there are also potential harms, and both must be considered. Several other studies support reducing fasting requirements prior to cardiac catheterization procedures.11–14,22,23 This is the first multicentre study of its kind. We suggest that the weight of the evidence has shifted the balance towards removing fasting requirements in this setting.
Though patient adherence to the assigned protocol was reasonable, the median time from clear liquid intake to procedure was 2.4 h within the no-fasting group. This complies with current fasting guidelines. While we encouraged regular meals and fluids, we did not mandate intake in this pragmatic trial. In the real-world intake would not be required just prior to procedures. An element of fasting occurs due to timing of the procedure in relation to meals. Moreover, fasting times in our trial are similar to published observational studies which also removed fasting requirements completely.15 In addition, the authors felt it was unethical to enforce food or water intake. It is reasonable to allow patients to moderate their own intake as they feel comfortable. Moreover, only 8.9% of patients within the no-fasting arm were ‘fasted’ according to guideline recommendations (see Supplementary data online, Table S5).
There was a trend towards higher rates of hypotension, hyperglycaemia, and hypoglycaemia occurred in the fasting arm (see Supplementary data online, Table S8). The increased rate of hypotension may be related to poor hydration status. Hyperglycaemia may have been seen more commonly in those who fasted as individuals with diabetes have abnormal circadian hormonal patterns which increase hepatic glucose production with resulting hyperglycaemia.24 The Somogyi effect or ‘dawn phenomenon’ has also been described: rebound hyperglycaemia following hypoglycaemia.25 Disruption of usual patterns of life and insulin administration by imposing fasting in those with diabetes may upset the balance in an individual's glucose control.
Sedation for implantable cardiac devices was considered clinically similar to percutaneous coronary interventional procedures. The device cohort demonstrated a comparable effect size to interventional procedures. The procedural sedation used was similar to the interventional cohort, though a higher midazolam dose was given in those with device-related procedures (see Supplementary data online, Tables S9 and S13). A prior randomized trial for elective cardiac implantable electronic devices showed no food-related adverse events in 100 patients assigned to a reduced fasting time of 1 h.13
The overall primary composite event rate was higher than predicted from the CHOW NOW trial. CHOW NOW has only been published in abstract form and endpoint definitions are not available. Differences in outcome definitions may explain the discrepancy between our cohorts. At our centres, we do not routinely administer intravenous fluid in those who are awaiting coronary angiography without another indication, i.e. renal impairment. This may have accentuated differences in blood pressure outcomes between groups.
There are several limitations to this trial. Pre-hydration was recommended in those with impaired renal function. Routine intravenous hydration in the absence of renal dysfunction was not routine. This may have accentuated differences in hydration status between groups. Fluid administration apart from that used to support blood pressure was not recorded. Due to funding limitations, only a fraction of the total cohort received serological testing for renal function at 48–72h. A total of 110 patients did not complete the patient satisfaction questionnaire. The patient satisfaction questionnaire used has not been validated for cardiac catheterization procedures and it is unclear how these results inform overall patient satisfaction. Trial participants and healthcare staff could not be blinded to group allocation, and this is likely to have influenced patient-reported satisfaction scores. Proceduralists were permitted to know group allocations which may have influenced treatment, resulting in performance bias. The quantity and content of solid and liquid intake was not recorded. Vomiting events were not recorded, though anti-emetic prescriptions were (see Supplementary data online, Table S14). Device procedures contributed 15.8% of the total population which may limit the generalizability of the results to this cohort. Those with planned calcium modification were excluded from the trial, though unplanned calcium modification did occur. Finally, a large proportion of patients in the no-fasting arm (57%) did not report clear liquid intake within 2 h of the procedure, though only 8.9% of the no-fasting group were fasted according to guidelines.
Conclusions
No fasting was non-inferior and superior to fasting prior to cardiac catheterization coronary and implantable cardiac device procedures for a primary composite outcome of aspiration pneumonia, hypotension, hyperglycaemia, and hypoglycaemia. Those allocated to no fasting had significant improvements in patient-reported satisfaction. This trial supports the safety of removing fasting requirements prior to cardiac catheterization laboratory procedures which require conscious sedation.
Supplementary data
Supplementary data are available at European Heart Journal online.
Declarations
Disclosure of Interest
Nothing to declare.
Data Availability
De-identified patient level data will be provided upon reasonable requests which have obtained ethics approval.
Funding
Funding for statistical analysis was gained from the John Hunter Hospital Charitable Trust and the Hunter Medical Research Institute Medical Research Support Program. A.S. is supported by the National Heart Foundation of Australia Future Leader Fellowship (award ID 106025) and the NSW Health Cardiovascular Capacity Building Grant.
Ethical Approval
Ethics approval was gained from the Hunter Research Ethics Committee prior to trial initiation and at all participating sites (Ethics Reference: 2021/PID02660).
Pre-registered Clinical Trial Number
This clinical trial was registered on the Australia New Zealand Clinical Trials Registry. The registration number is ACTRN12622001455752.
