An ultrasound‐guided serratus anterior plane block with continuous local anaesthetic infusion and epidural analgesia for rib fracture pain

We compared analgesia with an ultrasound (US)‐guided serratus anterior plane block (SAPB) to thoracic epidural analgesia (EA) with continuous local anaesthetic infusion in patients with unilateral multiple traumatic rib fractures. EA often carries contraindications in patients with multiple rib fractures (MRFs), whereby having alternative effective methods to treat rib fracture pain remains important to patient care. Thus, we hypothesised that both regional anaesthetic techniques would provide effective pain relief.


| INTRODUCTION
Multiple rib fractures (MRFs) are severe and common injuries in patients with blunt trauma and often cause moderate to severe pain. 1,2The aim of effective analgesia in patients with MRFs is not only to treat the pain to achieve patient comfort but also to prevent severe complications such as atelectasis and pneumonia. 3Epidural analgesia (EA) has become the gold standard in treating the pain caused by MRFs. 4 Yet, many patients with severe blunt thoracic injury have other injuries, typically injuries to the head, extremities, spine, abdomen and pelvis, and may have contraindications for EA. 5 Therefore, it is important to have alternative, effective analgesic methods available to treat rib fracture pain.
An ultrasound (US)-guided serratus anterior plane block (SAPB) is a fascial block used in the treatment of rib fracture pain. 6,7][10][11] Prospective comparative randomised trials regarding the effect of SAPB in MRF analgesia remain lacking, however.
To our knowledge, EA and SAPB with continuous local anaesthetic infusion have not been compared in a MRF analgesia in a prospective randomised study.However, US-guided SAPB with continuous local anaesthetic infusion has been compared to thoracic EA for thoracotomy pain finding no superiority between methods. 12As such, US-guided SAPB might have potential advantages in the treatment of pain caused by MRFs.][10][11] SAPB is a superficial block and, thus, safer than EA in cases involving disturbed haemostasis.At times, a US image can be challenging to interpret due to a possible pneumothorax, subcutaneous emphysema or a potential chest tube.When the anatomical structures are identified, the risk for pneumothorax is low. 13is study aimed to compare analgesia using US-guided SAPB with a continuous local anaesthetic infusion to thoracic EA in patients with unilateral MRFs.We hypothesised that SAPB with a continuous local anaesthetic infusion would provide equivalent pain relief when compared to EA.Our primary outcome was pain relief measured using the numerical rating scale (NRS).

| Inclusion and exclusion criteria
This was a single-centre randomised comparative trial conducted in Finland at Helsinki University Hospital's (HUH's) Trauma Unit.Before commencing, the trial was registered in the European Union Clinical Trials Register (EudraCT number 2018-001657-27).We were granted study permission by the Finnish Medicine Agency (FIMEA) on 23 July 2018, and the study protocol was approved by the HUH ethical board on 15 October 2018.All participants provided their written informed consent before inclusion.
The HUH trauma unit is a tertiary trauma clinic with a catchment area of approximately 2.15 million inhabitants as of 2018.All patients with blunt trauma and abnormal vital signs in the catchment area are treated in the HUH trauma unit.The HUH trauma unit treats approximately 90 patients with multiple traumatic rib fractures annually, with most such patients presenting with a unilateral injury. 14 this study, the inclusion criteria were unilateral MRFs (at least three fractured ribs) and a pain score ≥4 due to rib fractures at rest or upon movement.Pain scores were recorded using a numeral rating scale (NRS) score ranging from 0 to 10, with 0 representing no pain and 10 representing the worst pain imaginable.Other inclusion criteria were a patient aged 18-75 years, a body mass index (BMI kg/m 2 ) ≤38 and a spontaneous airway.Patients could also present with injuries other than MRFs.Patients with one to three contralateral but not dislocated rib fractures were included if the clinical judgement was that no regional analgesia would be necessary for the pain caused by these fractures.If clinically indicated, a chest tube was inserted in the emergency room (ER).
The exclusion criteria included a contraindication for either of the regional analgesic techniques (SAPB or EA) or a patient's inability to use a patient-controlled analgesia (PCA) device (due to a lack of co-operation, cognitive impairment, hemodynamic instability, sleep apnoea or history of severe alcohol or drug misuse).The contraindications for SAPB and EA included a disturbed haemostasis, signs of infection at the puncture site or general limitations to the positioning of the patient for procedures.We also excluded patients with an aortic valve stenosis, a traumatic brain injury, impaired consciousness, pregnancy, a language barrier, an allergy to any of the drugs used in the study protocol and patients taking anticoagulant medication.
Patients who had fractures only in the first, second and third ribs were excluded.
Nordic guidelines for neuraxial blocks in disturbed haemostasis from the Scandinavian Society of Anaesthesiology and Intensive Care Medicine were followed in all randomised patients. 15Patients had INR (international normalised ratio) <1.6 or partial thromboplastin time ≥50, and thrombocytes ≥80 Â 10 9 *L À1 .

| Randomisation and study intervention
Between the study period from 30 November 2018 through 28 August 2020, 117 eligible patients were evaluated for randomisation.Before randomisation, we excluded 57 patients.A total of 59 patients were recruited for the trial: 30 to the SAPB group and 29 to the EA group.
One randomisation envelope was lost (see Figure 1).
Upon a patient's arrival at the hospital, we secured their written informed consent to take part in the study.The anaesthesiologist performing the block opened the envelope once the patient provided their consent to participate in the study.
NRS at rest and upon movement (deep breathing), the percentage of forced expiratory volume in 1 s of expiration (FEV1%), blood pressure, heart rate, pulse oximetry values and method of supplementary oxygen administration were recorded before the block and at 30 min, 1, 2 and 12 h and on Days 1 and 2 after the block was administered.
As a surrogate marker for pain relief, we measured FEV1 and FEV1% using PulmoLife ® (MD Spiro, a Micro Direct Company), a handheld spirometer, which is simple to use at bedside on the ward.
The PulmoLife ® unit measures FEV1 and FEV1% of the predicted value (FEV1% predicted) based on a patient's gender, age, height and race.

| Block performance
All anaesthesiologists performing the blocks were experienced in the use of both types and familiar with the study protocol.The blocks were performed in five cases by another anaesthesiologist than the investigators.In most cases, the block was performed in the blockspecific operating room (OR).For patients admitted to intensive care, the block was performed in the intensive care unit (ICU).Patients not treated in the ICU were monitored in the block room for 1 h following the block before being transferred to the ward.
For the SAPB, the patients were placed in lateral decubitus position with the injured side up.We used Pajunk ® Tsui Catheter-Over-Needle (18 gauge) under ultrasonographic guidance.A linear US probe was used in a transverse plane to identify latissimus dorsi (LD) and serratus anterior (SA) muscles in the midaxillary line at the level of the fourth to fifth ribs. 6,16,17The needle was inserted using the in-plane technique and the catheter placed over the needle.Catheters were fixed with glue (Histoacryl B Braun ® ) and adhesive tape.

An initial bolus of local anaesthetic (30 mL Ropivacaine Fresenius
Kabi ® 2 mg*mL À1 for patient ≥50 kg and 20 mL for patient <50 kg) was given through the needle and a continuous infusion of ropivacaine 2 mg*mL À1 was started in the catheter at a rate of 1 mL/10 kg/h with maximum rate of 10 mL/h.The optimal level for thoracic epidural was estimated using trauma computer tomography (CT) images.The epidural catheters were placed in the lateral decubitus or sitting position using a Portex ® 18-gauge Tuohy needle.Epidural catheters were fixed with adhesive tape.In the EA group, an initial bolus of 5 mL ropivacaine 2 mg*mL À1 was given after testing dose of 3 mL Lidocaine 20 mg*mL À1 with adrenalin 5 μg*mL 1 (Lidocaine cum adrenalin Orion ® ).A continuous epidural infusion with ropivacaine 2 mg*mL À1 was initiated at a rate of 5-12 mL/h and titrated according to the NRS score and hemodynamic factors.
We asked patients to estimate their pain using NRS during positioning and when performing the block.Lidocaine 1 mg*mL À1 (Lidocaine Baxter ® ) was used to diminish pain at the puncture site.
Patients received intravenous fentanyl and diazepam boluses upon assessment by the treating anaesthesiologist during positioning.We also measured the time for preparations before inserting the block and the time for the procedure.Patient satisfaction with the pain relief for 1 h after the block was administered was recorded using a verbal rating scale from 1 to 5 (with 1 representing very dissatisfied, 2 dissatisfied, 3 cannot say, 4 satisfied and 5 very satisfied).
Patients in both groups received a rescue PCA with intravenous oxycodone boluses.Doses were set to 0.05 mg*kg À1 with a maximum single dose of 5 mg for patients <65 years and to 0.03 mg*kg À1 with a maximum single dose of 3 mg for patients ≥65 years.The lock-out period was set to 10 min and doses were limited to a maximum of four doses/h.Patients were advised to use the PCA device if the NRS was ≥4 at rest or upon movement.Patients were also given 1000 mg acetaminophen thrice daily orally (Para-Tables 1000 mg Orion ® ) or intravenously (Paracetamol B.Braun 10 mg*mL À1® ).The ropivacaine infusion was continued until the managing physician determined it was unnecessary.We recorded total PCA oxycodone consumption at 24 and 48 h following the block.

| Primary outcome
Our primary outcome was change in NRS at rest and upon movement from baseline through Day 2 following block performance at multiple time points, measured at 30 min, 1, 2 and 12 h, and before noon on Days 1 and 2 after the block.We considered an NRS reduction ≥2 to be clinically relevant.

| Secondary outcomes
The secondary outcomes were (1) FEV1% at the same time points as NRS was measured; (2) total PCA oxycodone consumption 24 and 48 h after the block performance; and (3) patient satisfaction at 1 h after the block.

| Statistics and power calculation
The sample size calculation was based on the hypothesis that NRS reduction ≥2 from baseline is clinically relevant.With a significance level of 5% and statistical power of 80%, 25 patients in each group would be needed to reach the required statistical power.Thus, we opted to randomise 30 patients in each group to compensate for any potential dropouts.We considered p values < 0.05 as statistically significant.For the continuous normally distributed variables, which consisted of NRS, FEV1%, blood pressure, heart rate and pulse oximetry values, we compared treatment groups using uncorrected t-tests and reported results as means and standard deviations (SDs).The consumption of oxycodone was not normally distributed; we, thus, used the Mann-Whitney U-test in our analysis, reporting results as the median (and range).Categorical values were compared using Fischer's exact test, for which we reported the number of observations and percentages.All statistical analysis was performed using Microsoft ®

| RESULTS
In total, 59 patients were randomised: 30 to receive SAPB and 29 to receive EA.There was failure to set an epidural in two cases, and in the SAPB group, one patient needed mechanical ventilation shortly after the block.These patients' outcomes were excluded from further analysis.
Patient characteristics at baseline were similar between groups (Table 1).In the SAPB group, 17 (57%) patients had chest tube versus 10 (35%) in the EA group.However, the difference was not statistically significant.2.
The block performance-related factors did not differ between the groups, except for the use of diazepam as a sedative (Table 3).
In the SAPB group, one patient needed mechanical ventilation 1 day after the block.Another patient in the SAPB group received an epidural 12 h after SAPB due to unsatisfactory analgesia.This patient had anterior rib fractures and a rupture to ventral abdominal wall muscles.Distinguishing the pain from the rib fractures from that caused by the abdominal muscle injury which also affected breathing proved difficult.
Five patients in both groups had to be operated under general anaesthesia due to concomitant injuries during the study period.
One patient with SAPB and operable clavicular fracture was a triathlete with pre-existing bradycardia.He experienced a short pulseless electric activity (PEA) period at the end of the operation when under general anaesthesia 1 day after the block.As such, he received cardiac compressions a few times and, thereafter, a normal sinus rhythm resumed.PEA, in this case, was assumed to be caused by a vagal reflex, but his serratus catheter was withdrawn to avoid any possible speculations for causality.This event caused no further consequences to the patient.Postoperatively, the patient was disappointed with the catheter removal given the effective pain relief achieved with SAPB.
In both groups, the local anaesthetic infusion was continued on Four epidural catheters were detached.Two of these patients received a new epidural catheter, one patient of which was in the SAPB, and one patient who no longer needed regional analgesia.In the EA group, two infusions were stopped due to unsatisfactory analgesia.Good analgesia was achieved for one of these patients with SAPB, while the other was achieved with an erector spinae block.In the EA group, two patients exhibited signs of an infection at the puncture site after the 48-h study period on Days 6 and 10, respectively.
We recorded two epidural failures due to anatomical reasons, one in a patient who received SAPB with excellent pain relief.The majority of patients were treated on the ward and received the block in the block OR after arrival from the ER.In the EA group, two patients needed vasoactive medication for hypotension following the test dose of lidocaine and an initial bolus of ropivacaine.A total of seven patients (23%) in the SAPB group and eight patients (28%) in the EA group were admitted to the ICU. or upon movement (SAPB 6.8 [SD 1.7] vs EA mean 6.9 [SD 1.9], respectively).An NRS reduction ≥2 within 1 h upon movement was achieved in 24 out of 29 patients in the SAPB group and 25 out of 27 in the EA group, a difference that was not statistically significant ( p = .42).However, EA was associated with a better reduction in the pain scores at rest and upon movement during the first 12 h (see Fig- ures 2 and 3).We detected no statistically significant difference in pain scores on Days 1 and 2 after the block when comparing the SABP and EA groups.

| Secondary outcomes
The mean increase in FEV1% within 1 h was one percentage point in the SAPB group and 14 percentage points in the EA group (Figure 4).
FEV1% was lower in the SAPB group than in the EA group before block, although the difference was not statistically significant (p < .051).We also found no significant difference in FEV1% on Days 1 and 2 after the block.We did find a wide interindividual variation in the PCA oxycodone consumption.However, we found no statistical difference between groups.Median PCA oxycodone consumption was 46 mg (range 0-140 mg) at 24 h and 96 mg (range 19-260 mg) at 48 h in the SAPB group and 34 mg (range 0-125 mg) at 24 h and 68 mg (range 0-338 mg) at 48 h in the EA group, respectively.We observed no difference in patient satisfaction with pain relief at 1 h when comparing groups.

| Explorative outcomes
Pulse oximetry values, heart rate or method of supplementary oxygen did not differ significantly between groups.In addition, blood pressure remained stable following both types of blocks.The mean arterial pressure (MAP) reduction 30 min after the block was 5 mmHg in the SAPB group.In the EA group, MAP fell slightly more (10 mmHg on average), although this difference was not clinically significant.block performance, likely leading to an increase in FEV1%.We also found that oxycodone consumption did not differ between groups.

| DISCUSSION
However, opioid consumption was not a perfect indicator for thoracic pain relief in this randomised trial given concomitant injuries.Patients in the EA group received more sedating medication during positioning and when performing the block, most likely because of the more challenging position.We found no significant difference in the amount of fentanyl administered during the procedure.Thus, we assume that fentanyl did not impact the NRS values at 1 h following the block.The lower blood pressure shortly after the epidural bolus was expected because of the vasodilatation caused by EA.
One strength of our study was that we had a relevant clinical problem and a realistic patient Most of our patients arrived at the hospital in the evening or at night, and we recruited them upon arrival to the ER.The blocks were performed by only a few experienced anaesthesiologists, thereby diminishing variation across procedures.We measured NRS and FEV1% during the first 2 days following the block because we aimed to compare analgesic methods during the acute phase, at a time when patients often have contraindications for neuraxial analgesia.A standardised, objective method does not exist to assess the pain caused by rib fractures.As such, NRS served as a subjective measurement for pain.In addition, the aim of the analgesia was to improve respiratory function.Thus, we chose the FEV1% measurement using a PulmoLife ® minispirometer for its simple and repeatable use on the ward.Patients appeared to learn to use the device easily.However, its clinical significance should be evaluated in future.
The majority of local anaesthetic infusions were stopped as planned in both groups (77% in the SAPB group and 62% in the EA group).We still observed various problems with catheters.Catheters sometimes migrate out during mobilisation and a displaced catheter causes unsatisfactory analgesia.Early mobilisation remains one of our primary goals in treating patients with MRFs.Two patients with EA showed signs of an infection at the puncture site.These patients were obese and had an epidural catheter, both long-established risk factors for infection.
A recent (2020) database study from the UK evaluated the quality of analgesia provided by serratus anterior plane (SAP) catheters compared with thoracic epidural or paravertebral catheters in 354 patients with MRFs. 8In that study, the primary outcomes were a change in dynamic pain scores and change in inspiratory volume.Pain was reduced and inspiratory volumes increased in all groups.Moreover, the researchers observed no difference in the decrease in pain scores and increase in inspiratory volumes between the EA, paravertebral or SAP groups.
Many trauma patients have contraindications for EA, especially during the first post-trauma days.Establishing an epidural block for a trauma patient with MFRs can be technically challenging due to concomitant injuries and positioning limitations.Patients, particularly in older age groups, may be prescribed anticoagulant medication, a contraindication for EA.Furthermore, trauma patients are at high risk of thromboembolic complications, and it is important to initiate thromboprophylaxis quickly after trauma. 18The timing for the manipulation of the epidural catheter in patients treated with low molecular-weight heparin must also be taken into consideration.
In this randomised trial, we administered continuous local anaesthetic infusion in an SA catheter.It would be interesting to further investigate serratus catheters with local anaesthetic infusions compared with regular boluses at different concentrations.

| Limitations
Some limitations and weaknesses were identified.First, we could not blind the block to patients or clinicians.A placebo-controlled study design would have proved unethical because it would have proved painful to patients.The patient population was rather heterogeneous with multiple injuries, possibly influencing rescue opioid use.
We included in this trial spontaneously breathing patients who were able to answer our questions and estimate static and dynamic pain scores.
This naturally excluded the most severely injured intubated patients.

| CONCLUSIONS
EA is effective for static and dynamic pain caused by MRFs but has its limitations in many patient groups.According to our findings, SAPB with a continuous local anaesthetic infusion is an effective alternative method for unilateral rib fracture analgesia, particularly, when EA is contraindicated.

AUTHOR CONTRIBUTIONS
Anna Lundén and Pekka Tarkkila were responsible for the study design and planning, patient recruitment, data collection and analysis and manuscript writing.

average for 5
days (1-8 days in the SAPB group and 1-12 days in the EA group).There were different reasons for the unplanned interruption of continuous local anaesthetic infusion.During the 48-h study period, we recorded unplanned interruptions to the local anaesthetic infusion in five patients in the SAPB group and six patients in the EA group, respectively.In the SAPB group, two infusions were stopped due to a change in the patient status (mentioned above) and one catheter was detached.This patient received a new SABP without delay.
SABP with a continuous local anaesthetic infusion diminished the NRS scores significantly during the study period.Pain relief in the EA group was significantly more effective during the first 12 h following F I G U R E 2 Numerical rating scale (NRS) at rest.Lines represent the 95% confidence interval.Asterisk (*) indicates p < .05.Twentyfour hours represents the first post-block day and 48 h represents the second post-block day.F I G U R E 3 Numerical rating scale (NRS) upon movement.Lines represent the 95% confidence interval.Asterisk (*) indicates p < .05.Twenty-four hours represents the first post-block day and 48 h represents the second post-block day.F I G U R E 4 Forced expiratory volume in 1 s as a percentage (FEV1%).Asterisk (*) indicates p < .05.Twenty-four hours represents the first post-block day and 48 h represents the second postblock day.
Other injuries than rib fractures.Factors relating to the block performance.
T A B L E 3Note: Values, mean [standard deviation]; numbers (percent of total within group).Abbreviation: NRS, numerical rating scale.