This study was performed in the East of England, a mixed urban and rural geographical area of 20,000 km2 with a population of approximately 6.4 million people [25]. HEMS provide prehospital critical care to the patients of the East of England Ambulance Service NHS Trust, and operate from five bases: two East Anglian Air Ambulance (EAAA), two Essex & Herts Air Ambulance (EHAAT), and one Magpas Air Ambulance (Magpas). All teams utilise a combination of rotary wing aircraft and ground-based rapid response vehicles dependant on time of day, patient location, and aviation restrictions.

A physician-critical care paramedic (CCP) model is employed by all three HEMS. Physicians are predominantly emergency medicine (EM) or anaesthetic specialists at the level of consultant or senior registrar (at least five years post-graduation), with a minimum of six-months in-hospital anaesthesia training. CCPs have at least three years’ post-registration experience. Physicians and CCPs undergo additional specialist critical care training followed by supervised practice and formative assessment by a prehospital consultant (physicians) or senior CCP prior to autonomous practice. A programme of continued training supports skill retention.

To standardise practice across the region, all three HEMS use a shared PHEA standard operating procedure (SOP). This SOP includes a recommended PHEA induction regime: ketamine (1–2 mg kg− 1), rocuronium (1 mg kg− 1), ± fentanyl (1–3 mcg kg− 1) at the discretion of the attending clinician to attenuate the hypertensive response to laryngoscopy; modified according to patient age, weight, haemodynamics, and suspected injury [18]. Maintenance of anaesthesia is recommended as either infusion: ketamine (1 mg/kg/hr); propofol 1% (2-5 mg/kg/hr) or as bolus: ketamine (up to 0.25 mg kg− 1); midazolam (up to 0.025 mg kg− 1 ); fentanyl (0.5 mcg kg− 1 ) adjusted according to patient physiology. Laryngoscopy is typically performed 45–60 s after administration of rocuronium. Direct laryngoscopy (DL) was the primary method of airway visualisation until 2017 when videolaryngoscopy (VL) (McGrath® Video Laryngoscope, Aircraft Medical, Edinburgh, UK) was introduced for use at the preference of the intubating clinician at EAAA and Magpas. All HEMS utilise a ‘challenge and response’ pre-PHEA checklist to optimise physiology and conditions prior to induction [3]. Positive pressure ventilation (6–8 ml kg− 1 ideal body weight) with a positive end-expiratory pressure (PEEP) of 5 cmH2O, and rate to achieve normocapnia are standard practice. HEMSbase (MedicOne Systems Ltd, UK) is the electronic medical record software used by each HEMS to capture data and patient information.

In this retrospective observational study, we included a consecutive sample of all trauma patients aged ≥ 16 years attended by EAAA and EHAAT between 1st January 2015 and 31st December 2022 and Magpas between 1st November 2015 to 31st December 2022 (owing to the later adoption of the common electronic medical record system) who received a PHEA, defined as a drug-assisted intubation.

Patients who were intubated in cardiac arrest, patients undergoing inter-facility transfer, patients intubated by a non-HEMS team provider, and patients with an injury mechanism other than trauma (injury through transfer of kinetic energy): hanging, drowning, overdose, asphyxia, burns, electrocution, and any other similar non-traumatic mechanism were excluded. Patient records were reviewed by one of the study authors to identify exclusions. Patients without any pre- or post-PHEA SBP readings were excluded following interrogation.

Routinely collected pseudonymised patient data from the electronic medical record were extracted and collated in a single data spreadsheet (Microsoft® Excel for Mac, v16.45) and stored in an organisational secure data environment.

Extracted data items included: demographics (age, sex, weight), suspected injury pattern, mechanism and type of injury (blunt or penetrating), Glasgow Coma Scale total (GCS-T) score, Glasgow Coma Scale motor (GCS-M) score, PHEA indication, fluid administered pre-PHEA, intubation attempts, use of direct or video-laryngoscope, time interval from HEMS team arrival to PHEA (minutes), drug dosage: fentanyl (mcg kg− 1), ketamine (mg kg− 1), and rocuronium (mg kg− 1).

Physiological data recorded routinely by time-calibrated monitors (EAAA – X Series®, ZOLL Medical Corporation, Runcorn, UK; EHAAT & Magpas – Tempus Pro, Philips Electronics UK Ltd, Farnborough, UK) and automatically uploaded to HEMSbase at two-minute (EAAA, EHAAT) or three-minute (Magpas) intervals were collected and interrogated. Manual interrogation was undertaken to record the following pre-PHEA physiological variables based on the closest time-point preceding the recorded PHEA induction time: heart rate (HR), respiratory rate (RR), systolic blood pressure (SBP), and diastolic blood pressure (DBP). Post-PHEA data (HR, SBP, and DBP) were documented at two-minute intervals (or closest time point) up to ten minutes. Apparently erroneous data were reviewed by two data interrogators and excluded if deemed explicitly erroneous.

PHEA induction drug doses were collected; fentanyl (mcg), ketamine (mg), and rocuronium (mg), and divided by the estimated patient weight (kg). These were rounded to the nearest integer and expressed as a rounded dose per kg, with 0,1,2,3 (mcg kg− 1) categories used for fentanyl and 0,1,2 (mg kg− 1) categories used for ketamine and rocuronium. Patients who did not receive the drug were coded as 0, those receiving > 0.5 per kg but < 1.5 per kg were rounded to 1. Patients who received greater than the maximum dose were included in the maximum category for the drug.

The primary outcome was post-PHEA critical hypertension defined as at least one new episode of SBP > 180mmHg ≤ 10 min of induction, or a > 10% increase if SBP was > 180mmHg pre-PHEA [10, 13].

Data management and statistical analyses were performed in R statistical programming language (R Core Team [2018]; R: A language and environment for statistical computing [R Foundation for Statistical Computing, Vienna, Austria]). A pre-defined significance value of p < 0.05 was used throughout. The sample characteristics were described using number (percentage) for categorical variables and mean (± standard deviation (SD)) or median [interquartile range (IQR)] for continuous variables as appropriate.

Logistic regression was used to test the association between the variables and post-PHEA hypertension and coded as a binary outcome (1 = hypertension, 0 = no hypertension). A purposeful model was used to select variables [26]. First, each variable was tested individually to document its unadjusted association with the outcome. A multivariable model was then built using variables with a p-value of < 0.25 from the univariate analysis or of known clinical importance. Backward stepwise elimination was used to find a reduced model that best fits the data, defined at the highest likelihood ratio test for goodness of fit. The final model was assessed to ensure it met the assumptions necessary for logistic regression: no multicollinearity in the variables, linear relationships in the logit of the outcomes, and no unduly influential values. Plausible interactions between variables were tested by including interaction variables in the model and assessing significance and goodness of fit.

For variables grouped in categories, the group containing the most cases was used as the baseline reference in the regression model. Patients were divided into four age groups: 16–34 years; 35–54 years; 55–74 years; 75 + years. Pre-PHEA SBP was grouped as: Low (< 90), Mid (90–140), High (141–180), Critical (> 180) mmHg; heart rate was grouped as: Low (< 60), Mid (60–100), High (> 100) beats/minute; and respiratory rate was grouped as: Low (< 10), Mid [10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25], High (> 25) breaths/minute [27]. For all drug dosage categories, 1 mg/mcg kg− 1 was used as the reference group.

Ethical approval for the study was granted by Allied Health, Nursing & Midwifery and Medicine School Research Ethics Panel at Anglia Ruskin University (ETH2223-3743). The STROBE (Strengthening the Reporting of Observational studies in Epidemiology) reporting guideline was followed [28].