A retrospective, observational, single-centre, cohort database analysis of the haemodynamic effects of low-dose spinal anaesthesia for hip fracture surgery

Background Careful administration of either spinal (intrathecal) or general anaesthesia probably has a greater impact on outcomes after hip fracture surgery than which method is used per se. Intraoperative hypotension is associated with poorer outcomes, but appears less prevalent using lower doses of spinal anaesthesia. Methods In this observational single-centre study, intraoperative noninvasive blood pressure data were analysed from 280 patients undergoing unilateral hip fracture surgery after the administration of hyperbaric spinal bupivacaine 0.5%, 1.3 ml (0.65 mg). Results Mean cohort mean arterial pressure (MAP) remained within 10% of baseline (spinal injection) MAP for 97/98 (99.0%) subsequent aggregated 1-min recording intervals. The prevalences of lowest MAP <70 mm Hg and <55 mm Hg were significantly lower than historical equivalents (Anaesthesia Sprint Audit of Practice 1 and 2) (52.9% and 10.4% vs 71.9% and 23.8%, respectively, both <0.0001). The proportions of 10 551 MAP readings <70 mm Hg and <55 mm Hg were 6.7% and 0.4%, respectively. Forty-five (16.1%) patients had relatively persistent hypotension (MAP ≤70 mm Hg for five or more intraoperative readings), and were statistically more likely to be frail (Nottingham Hip Fracture Score ≥7/10, 37.8% vs 19.6%, P=0.0109) and be taking alpha-/beta-blockers (44.4% vs 24.3%, P=0.0099) than the remaining ‘normotensive’ cohort. Surgical anaesthesia remained effective for up to 190 min, with only one patient requiring supplemental local anaesthesia during skin closure. Conclusions Low doses of hyperbaric spinal 0.5% bupivacaine (1.3 ml, 6.5 mg) are associated with minimal reductions in blood pressure during surgery and provide adequate duration of surgical anaesthesia. Randomised comparisons of lower vs higher/standard doses of spinal anaesthesia are now required to confirm outcome benefits in this vulnerable patient group. Clinical trial registration NCT05799300.

Approximately 35 000 people who fracture a hip in the UK each year are administered spinal (intrathecal) anaesthesia for surgical repair, 1 of whom at least 30% develop intraoperative hypotension (depending on definition 2,3 ), which is associated with a greater risk of death within a month of their surgery. 4ecent meta-analysis 5 indicates that the mode of anaesthesia (i.e.'spinal' or 'general') is probably of much less importance than how carefully that anaesthesia is administered (i.e. in a standardised manner, aiming to remobilise, re-enable, and rehabilitate people shortly after surgery 6 ).This supports existing national 7 and international 8 guidelines that recommend using lower volumes of local anaesthetic for spinal anaesthesia (e.g.bupivacaine 0.5%, 1.5 ml) than are commonly used (bupivacaine 0.5%, 2.5 ml 4 ) to reduce the likelihood of perioperative hypotension.However, those recommendations were based on historical research selectively involving younger, fitter people having hip fracture surgery, 9e11 in whom blood pressure was recorded neither particularly accurately nor often enough.
Recently, a non-proprietary method for transferring vital signs data electronically from anaesthetic monitors to storage computers has been described. 12Although intended for medicolegal purposes (e.g.Coronial investigations), analysis of these data could also enable accurate description of how blood pressure changes around the time of surgery, and among subsets of hip fracture patients who are normally excluded from prospective research (e.g. the very old, the very frail, people with dementia 13 ).Furthermore, by comparing these data with published national data (e.g.Anaesthesia Sprint Audit of Practice, ASAP, 1 3 and 2 4 ), it would become possible to determine whether lower dose spinal anaesthesia is associated with a significantly lower prevalence of hypotension during hip fracture surgery.
The primary objective of this research, therefore, is to describe the relative changes in noninvasive systolic, mean, and diastolic arterial blood pressures collected prospectively at regular 2-min intervals from a cohort of consecutive patients undergoing hip fracture repair surgery who were administered low-dose spinal anaesthesia (hyperbaric bupivacaine 0.5%, 1.3 ml without additional opioid).
The secondary objectives of this research are (1) to analyse the significance of any differences in the prevalence of intraoperative hypotension (variably defined) using low-dose spinal anaesthesia for hip fracture repair compared with historical national 'all-dose' and 'mode-dose' (i.e.hyperbaric bupivacaine 0.5%, 2.5 ml) spinal anaesthesia data from England and Wales, 3,4 (2) to describe the proportion of blood pressure recordings that fell below variably defined thresholds for hypotension, and (3) to determine the effective duration of surgical anaesthesia after low-dose spinal administration.

Trial design and participants
This retrospective, observational, single-centre, cohort database analysis was approved after proportionate ethical review by East Midlands -Derby Research Ethics Committee (IRAS 279652) on 15 June 2022.The study is registered on ClinicalTrials.gov(NCT05799300).Individual consent was not sought from study subjects.

Trial participants
All people undergoing unilateral surgical hip fracture repair (cemented/uncemented hemiarthroplasty, dynamic hip screw, cortical screws, proximal femoral nail) administered anaesthesia by the author at Princess Royal Hospital, Hayward's Heath, UK, for whom contemporaneous vital sign (noninvasive systolic/diastolic/mean arterial blood pressure, heart rate, oxygen saturation, ventilatory frequency) data were recorded, were eligible for inclusion in this analysis.Data were available from 10 March 2017 and analysed up to 13 January 2020 (i.e. when hospital services were reconfigured in light of COVID-19).
Vital sign data were captured from anaesthetic monitors (AS3, GE Healthcare, Hatfield, UK) using hardwire connection to a double password-protected laptop computer, and transferred electronically (via NHS email) to double password-protected NHS desktop computers, before deletion from the laptop.These records are normally held for medicolegal purposes/Coronial investigation. 12seudo-anonymisation (hospital number, date of surgery) was used to link these data with patient physical and injury characteristics (sex, age, American Society of Anesthethesiologists [ASA] physiological status score, Nottingham Hip Fracture Score [NHFS 14 ], Jiang risk score 14 ), abbreviated mental test score (AMTS), pre-fracture level of care/mobility/comorbidity (hypertension/valvular heart disease/diabetes)/medications (alpha-with or without beta-adrenoceptor blockade/ other antihypertensive/anticoagulation), process (timings, anaesthesia and surgical interventions), and outcome (died during admission/within 30 days of surgery, length of postoperative stay [LOPS]) data held on the Brighton Hip Fracture database, 15 before data were fully anonymised for analysis.Data were held and analysed in compliance with the requirements of the General Medical Council's Good Clinical Practice, the Data Protection Act 2018, and the European Union General Data Protection Regulations 2016/679.
All study participants received broadly consistent institutional and anaesthetic interventions.During their hospital admission, patients followed a standardised, multidisciplinary hip fracture care pathway, co-ordinated by senior orthogeriatricians.Hip fracture patients requiring surgery (>99.8%) were identified and discussed at a pre-theatre trauma meeting, having usually been admitted to hospital within the previous 24 h (longer if further management was required, particularly of their anticoagulation).

Anaesthetic techniques
After transfer from the hip fracture unit direct to the anaesthetic room, patients were checked in according to institutional and World Health Organization (WHO) criteria.With venous access having been routinely established on the hip fracture unit, patients were commenced on 1 L i.v.crystalloid and connected to routine noninvasive electrocardiogram, pulse oximetry and blood pressure monitoring, the latter set to record every 2 min.After correct side confirmation, a suprainguinal fascia iliaca block was administered (using an antiseptic technique, a linear ultrasound probe, and infiltration of plain bupivacaine 0.375%, 20 ml þ lignocaine 0.5%, 20 ml via an in-plane 50 mm 30 facet bevel stimulating single-shot nerve block needle (Stimuplex®, B Braun, Sheffield, UK), unless the patient had received a fascia iliaca block within the previous 4 h.After commencement of 2e3 L min À1 oxygen via a CO 2 sampling-enabled facemask (adult Sentri TM EcoLite TM , Intersurgical, Wokingham, UK), 20e40 mg (30e60 mg in younger, fitter patients) propofol sedation was administered before positioning the patient comfortably in the lateral decubitus position with the operative side dependant.Using a strictly aseptic technique, lignocaine 1%, 2e3 ml was infiltrated subcutaneously and into the interspinous space via a 23 G hypodermic needle, before accessing the (estimated) L2e3 intrathecal space via a 25 G pencil-point spinal needle (Sprotte NRFit®, Pajunk UK Medical Products Ltd., Newcastle upon Tyne, UK), confirming cerebrospinal fluid flow and administering hyperbaric bupivacaine 0.5%, 1.3 ml by low-pressure injection over 30 s to 1 min.Having confirmed a satisfactory blood pressure 2e3 min after spinal injection, patients were transferred to the operating theatre, where the effectiveness of their block was confirmed: during transfer from bed to operating table, during surgical skin preparation (abduction of the hip for hemiarthroplasty, distraction of the fracture for internal/external fixation) and by pre-incision sharp surgical stimulus.Boluses of propofol sedation (20e40 mg) were administered i.v. according to patient agitation.Boluses of vasopressors (metaraminol 0.5 mg, ephedrine 6 mg) were administered to maintain a subjectively 'suitable' mean arterial pressure (MAP) and heart rate.Hemiarthroplasties were performed in a lateral decubitus position, and fixation in a supine position, by senior (consultant/supervised specialist trainees) surgeons.I.V. antibiotic prophylaxis (teicoplanin 400e600 mg þ gentamicin 160 mg) and tranexamic acid (1 g) were administered before incision.Red blood cell salvage was not used routinely.After surgery, patients were disconnected from i.v.fluids and oxygen (unless indicated clinically) and admitted to the post-anaesthesia care unit, sitting up at 45e60 in their ward beds, before discharge back to the orthogeriatric hip fracture unit care.
Data from the ASAP 2 study were collected and handled as described previously. 3,4

Outcomes
The primary outcome measure was mean (standard deviation, SD) systolic, mean, and diastolic noninvasive blood pressures (NIBPs) taken at 2-min intervals from all patients included in the study.

Sample size calculation
The sample size was determined pragmatically.Previously, the largest prospective trials comparing hypotension after different doses of spinal anaesthesia have reported NIBP recording intervals of 5 min in 10e37 patients.9e11 From historical Brighton Hip Fracture Database data (for the author), patients spend ~20 min in the anaesthetic room, then 15 min being prepared for surgery and 60 min undergoing surgery.In order to calculate a stable mean (SD) for the primary outcome measure, the study aimed to record NIBP data from at least 37 patients for every 2min recording cycle during their estimated (20þ15þ60¼) 95 min stay in the operating theatres, generating ~4162 data points in total (i.e. three systolic/mean arterial/diastolic blood pressures Â 37 patients Â 37.5 recordings per patient), enabling sufficient descriptive graphical representation of the primary outcome measure (i.e..
Low dose spinal anaesthesia for hip fracture repair -3

Trial population
Between

Primary outcome
The mean relative decreases in systolic, mean, and diastolic blood pressures over time compared with baseline (spinal anaesthesia infiltration at t 0 [i.e. the primary objective]) are shown graphically in  S1).

Secondary outcomes
The prevalence of the lowest blood pressure values recorded that fell below various definition thresholds for 'hypotension' among the 280 sets of patient data analysed (and compared with national data from ASAP) are shown in Table 2.
The proportion of blood pressure readings recorded that fell below various definition thresholds for 'hypotension' among the patients' data analysed are shown in Table 3.
The median (IQR [range]) duration of theatre time and outcomes are shown in Table 1.One patient required additional local anaesthesia infiltration to the superior margin of their surgical wound during closure.
During data analysis, it became apparent that a subgroup of the patient cohort had recorded several low intraoperative MAPs.Arbitrarily defining patients in this group by having MAP 70 mm Hg for five or more readings (equivalent to 10 min duration of hypotension), the prevalence of potential causative factors in this group compared with the remainder of the cohort are shown in Table 4.In the low-pressure subgroup, five of the six patients who died during admission (and four of the five patients who died within 30 days of surgery) were 'frail' and adrenoceptor-blocked before operation.

Discussion
In this observational database study, anaesthesia involving low-dose spinal hyperbaric bupivacaine 0.5% (1.3 ml [¼6.5 mg], without additional opioid) was associated with relative normotension.Mean cohort MAP remained within 10% of baseline MAP for 97 of the 98 (99.0%) aggregated 1-min recording time intervals (Fig 1, Supplementary Table S1).The nadir mean cohort MAP reading occurred ~5 min after spinal administration, which anaesthetists should account for when transferring patients between the anaesthetic room and operating theatre.Although nadir blood pressures occurred relatively abruptly, they approximated to only a 10% relative decline from baseline at spinal anaesthesia administration, and normotension was restored equally abruptly, with only a third of patients requiring (low dose) vasopressor administration (median metaraminol 0.75 mg, ephedrine 6 mg).
Within the mean cohort data, there was considerable variation in how much the blood pressure changed in individual patients (relative median change MAP 21% (IQR 20e23% [range 7e36%], Supplementary Table S1).Proportionately, however, relatively few of the systolic and mean blood pressures recorded fell below thresholds that have previously been associated with increased mortality and which have occurred with comparably greater prevalence in both hip fracture 3,4  Of note, greater mean relative decreases in SBP and MAP (~30%) were observed before spinal anaesthesia administration, confirming earlier observational data that challenge the definition of a reference 'baseline' on which subsequent anaesthetic interventions might be administered, 18e20 and indicating important potential hypotensive effects related to both pain (during ward/theatre transfer)/analgesia (nerve block) and anxiolysis/sedation with propofol (for lateral positioning).
Interestingly, the 45 hip fracture patients with relatively persistent hypotension (i.e.MAP 70 mm Hg for five or more readings) were statistically more likely to be frail (i.e. have an NHFS !7/10), be taking alpha-/beta-blockers, experience longer preoperative delays, and to die during hospital admission than the 'normotensive' remainder of the study cohort (Table 4), indicating future avenues for research in identifying potentially 'endotypical' patients at mortal/morbid risk of intraoperative hypotension, and, again, alerting anaesthetists to perhaps more intensive management of such patients. 21,22 contrast to the few randomised controlled trials (RCTs) reporting hypotensive events after variable doses of spinal bupivacaine, 9e11,23 a comparative strength of this study concerns its involvement of a properly representative hip fracture population.It might be expected that including data on older, frailer, co-morbid patients might be associated with more   Mean relative decreases in systolic, mean, and diastolic blood pressures over time compared with baseline (spinal anaesthesia infiltration at t0) for 280 people administered hyperbaric bupivacaine 0.5%, 1.3 ml for unilateral hip fracture repair (excluding total hip arthroplasty).For clarity, data points averaged from fewer than 10 patients are excluded (Supplementary Table S1).DBP, diastolic blood pressure; MAP, mean arterial pressure; SBP, systolic blood pressure.prevalent hypotensive episodes, greater absolute falls, and wider variations in blood pressure, related to impaired homeostasis.However, the data reported in this observational study are quantitatively similar to those reported in previous RCTs 9e11 and observational studies, 3,4 indicating similarly prevalent absolute and relative hypotensive episodes (Table 2) and magnitude of pressure falls after low-dose spinal administration (~10% vs 15 9e11 e20% 10 ), that statistically are significantly smaller than those occurring after 'normal' dose spinal anaesthesia administration (i.e.hyperbaric bupivacaine 0.5%, 2.5 ml [12.5 mg]) (Table 2). 3,4,23ndeed, those quantitative similarities between the data presented and historical data reported for low-dose spinal anaesthesia should reassure readers that the confounding variables normally associated more frequently with observational, compared with RCT, data are unlikely to be influencing the results presented here to a significant degree that might alter the conclusions drawn from those data.The main weakness of this study is that potentially vasoactive stimuli may have influenced blood pressure changes other than attributable to spinal bupivacaine administration, including analgesia (immobility, preoperative nerve block, fracture type 24 ); other stimulation (movement, surgical percussion, hip relocation); bone cement implantation 25 ; active warming; sedation (propofol dose/frequency/total dose); and vasopressor use (alpha-/beta-blocker dose/timing/frequency/total dose).However, previous RCTs also did not take such confounding variables into account, excepting the use of objective vasopressor administration below predefined blood pressure thresholds (SBP <90 mm Hg 9 e100 mm Hg, 11 or a decrease of SBP >20% 10 /MAP >25% from baseline 9 ).
Despite the lower dose of spinal bupivacaine used, surgical anaesthesia remained effective for up to 190 min, with only one of 280 patients requiring supplemental local anaesthesia to the upper margin of their surgical skin incision during closure.These data should allay understandable concerns about low-dose spinal bupivacaine methods providing sufficient duration of surgical anaesthesia, 26,27 particularly when co-administered with an ultrasound-guided fascia iliaca block.
Similarly, these data should re-assure clinicians that using low-dose spinal anaesthesia does not seem to be associated with adverse patient outcomes.Indeed, both the observed incidence of 30-day postoperative mortality and the observed median LOPS compared favourably with time-equivalent national figures (5.4% vs 6.6%, 15.0 vs 15.5 days, respectively).
In conclusion, the observational data presented here suggest that low doses of hyperbaric spinal bupivacaine (0.5%, 1.3 ml, 6.5 mg) are associated with minimal reductions in blood pressure during surgery in a representative sample of hip fracture patients, compared with historical data.These data validate those from RCTs which support the preferential use of lower doses of hyperbaric spinal bupivacaine than are currently used (typically 0.5%, 2.5 ml, 12.5 mg), with the aim of reducing both the prevalence and duration of hypotensive episodes intraoperatively (and so, potentially, reducing poor mortal and morbid outcomes).Low-dose spinal bupivacaine provided adequate surgical anaesthesia in this study.
The mode of anaesthesia probably has little influence on commonly measured outcomes after hip fracture surgery.5e8 Instead, researching the improved delivery of both spinal and general anaesthesia within intraoperative care would seem more likely to improve patients' recovery trajectories.In Low dose spinal anaesthesia for hip fracture repair -7 the first instance, this will require future research focus being redirected away from general vs regional anaesthesia dualism towards comparisons of appropriate outcomes (hypotension, pain, confusion, mobility) after lower vs higher dose anaesthesia.

Declaration of interests
SW convened the International Fragility Fractures Network Consensus Group on the principles of anaesthesia for patients with hip fracture.SW was a member of the Association of Anaesthetists Working Party on the management of hip fractures, and is a scientific advisor to the World Hip Trauma Evaluation (WHiTE).No other interests and no external funding declared.

Fig 2 (
numerical mean [SD] are provided in Supplementary Table Time relative to spinal anaesthesia infiltration baseline (min)Blood pressure relative to spinal infiltration baseline blood pressure (%)

Fig 2 .
Fig 2. Mean relative decreases in systolic, mean, and diastolic blood pressures over time compared with baseline (spinal anaesthesia infiltration at t0) for 280 people administered hyperbaric bupivacaine 0.5%, 1.3 ml for unilateral hip fracture repair (excluding total hip arthroplasty).For clarity, data points averaged from fewer than 10 patients are excluded (Supplementary TableS1).DBP, diastolic blood pressure; MAP, mean arterial pressure; SBP, systolic blood pressure.

Table 1
14tient characteristics, perioperative and outcome data for the 280 people with hip fracture included in the analysis.Cemented hemiarthroplasty ¼ Exeter prosthesis; uncemented hemiarthroplasty ¼ Austin-Moore prosthesis.AMTS, Abbreviated Mental Test Score (out of 10); ASA, American Society of Anesthesiologists physiological status; IQR, inter-quartile range; NHFS, Nottingham Hip Fracture Score (out of 10)14Jiang,14; SD, standard deviation.*One patient required local anaesthesia infiltration to the cephalad end of their surgical wound during closure.Normally distributed data (timings, vasopressor doses administered) are described by their mean and SD, and compared using unpaired t-tests.Skewed or non-continuous data (age, NHFS, AMTS, LOPS) are described by their median (inter-quartile range, IQR [range]), and compared using twotailed Fisher's exact tests.Binary logistic regression calculations were modelled in Excel, using Solver (both Microsoft 2011, Redmond, WA, USA).
10March 2017 and 13 January 2020, 1753 patients underwent surgery to repair their hip fracture.

Table 2
Prevalence (%) of patients with lowest blood pressure below defined 'hypotensive' threshold, by volume of hyperbaric bupivacaine 0.5% spinal anaesthesia administered.Compared with data from the 280 study patients, ' 1.3 ml' from the ASAP data defines an equivalent volume, ' 1.5 ml' the volume above which SBP fell by more than the 'least worst' definition of hypotension (i.e. a relative decrease >20%) and 2.5 ml the mode volume administered in England and Wales.P-values indicate statistical significance differences between ASAP cohorts compared with current study cohort.ASAP, Anaesthesia Sprint Audit of Practice; MAP, mean arterial pressure; SBP, systolic blood pressure.

Table 3
Proportion of 10 551 blood pressure readings below various definition thresholds for 'hypotension'.MAP, mean arterial pressure; SBP, systolic blood pressure.

Table 4
16mparison of potentially causative factors, vasopressor administration, and outcomes between hip fracture patient 'lowpressure' subgroup (i.e.MAP 70 mm Hg for five or more readings), and 'normal-pressure' remainder of the cohort.Age !86 yr and NHFS !7/10 (as a surrogate marker for frailty) have been previously identified as risk factors for increased perioperative mortality/ institutionalisation after hip fracture.16'Likelyaortic stenosis' identifies patients with a clinical/echocardiographic diagnosis of aortic stenosis with or without an audible ejection systolic murmur before operation during cardiac auscultation.A delay of LOPS !36 h represents potentially two episodes of preoperative starvation with or without fluid replacement.LOPS length of (in)patient stay; MAP mean arterial pressure; NHFS, Nottingham Hip Fracture Score; SD, standard deviation.