Incidence and risk factors of ipsilateral shoulder pain in patients after thoracic surgeries

ABSTRACT

Background and Aims:

Ipsilateral shoulder pain (ISP) post-thoracotomy impairs the recovery in early postoperative period, the aetiology of which is unclear. We studied to find out the incidence and risk factors associated with ISP.

Methods:

We did a prospective observational study, wherein 296 patients scheduled for thoracic surgeries were enroled. Pain in the shoulder during activity was assessed using American Shoulder and Elbow Surgeons standardised assessment method. All potential predictors were analysed in a multivariable penalised logistic regression model, using ISP as the outcome variable.

Results:

Of the 296 patients, 118 (39.8%) patients developed ISP. Of the 296 patients, 170 patients underwent thoracotomy and 110 underwent video-assisted thoracoscopic surgeries. The incidence of ISP was higher in thoracotomy patients (45.29%) compared to video-assisted thoracoscopic surgeries (32.7%). Majority of patients (43.2%) were aged more than 65 years, which was statistically significant as per univariate analysis (P = .007). The incidence of ISP was the highest at 41.89% among those who had lung cancer (n = 74), with disease involving right upper lobe and left upper lobe, 29% and 25.8%, respectively. The pain severity was moderate in 27.1% of patients during shoulder movements. Among the patients who had ISP, 77.1% expressed it as dull aching, whereas 21.2% described it as stabbing in nature.

Conclusion:

The incidence of ISP in those who underwent thoracic surgery was high and dull aching in nature, of mild to moderate intensity, commonly located on the posterior aspect of the shoulder. It was more common in those who underwent thoracotomy and more than 65 years of age.

INTRODUCTION

Thoracic surgeries are associated with significant operative trauma.[1] While thoracic epidural analgesia may help control the incisional component of the pain, an excruciating post-thoracotomy ipsilateral shoulder pain (ISP) could undermine pain management in the post-thoracotomy patient.[2,3] Thoracic epidural analgesia is ineffective in eliminating this pain.[4] Post-thoracotomy shoulder pain reduces patient function and postsurgical rehabilitation after thoracotomy.[4] The incidence of ISP ranges from 21% to 97%[4,5] and it impairs respiration, mobility, and physical therapy in the early postoperative period.[6]

The aetiology of ISP is unclear, with theories linking it to musculoskeletal origin or referred pain.[4] Several hypotheses on the possible causes of ISP have been proposed, such as transection of a major bronchus, ligament distraction by surgical retraction, shoulder joint strain as a result of intraoperative positioning, pleural irritation due to the thoracotomy tube, and referred pain from irritation of the pericardium or mediastinal and diaphragmatic surfaces.[2,7] ISP is defined as the pain occurring on the operated side of thoracotomy in the immediate postoperative period as early as one hour after surgery[3]; dull aching, stabbing, burning, electric, or throbbing in character, of moderate to severe intensity; and resistant to treatment. ISP is usually located in the deltoid region or the posterior or superior aspect of the arm, at times above the lateral one-third of clavicle, or the anterior surface of the chest, lasting for 3-4 days.[2,7,8]

General treatment approaches consist of regular administration of nonsteroidal anti-inflammatory drugs (NSAIDs) and acetaminophen. Other approaches such as intraoperative injection of local anaesthetic into the pericardial pad of fat, interscalene brachial plexus block, stellate ganglion blockade, and direct blockade of the phrenic nerve have also been tried.[4] We aimed to find out the incidence of ISP; identify the risk factors associated; and to investigate characteristics, locations, and severity of ISP in our patient population.

METHODS

A prospective, observational study was conducted in a tertiary cancer care teaching hospital over a one-year period (August 2017 to August 2018) after obtaining approval from Institutional Ethics Committee and registering with Clinicaltrials.gov (NCT03249766). Two hundred and ninety six American Society of Anesthesiologists (ASA) I, II, and III patients aged 18 years or more, scheduled for various elective thoracic surgeries we enroled in the study after obtaining an informed consent. Surgical approaches were either standard posterolateral thoracotomy or video-assisted thoracoscopic surgery (VATS) or robotic depending on the surgeon’s preference. Patients with preoperative shoulder pain, preoperative shoulder pathology, use of analgesics for more than a week preoperatively, inability to understand the numeric rating scale scoring system for pain assessment, and patients shifted on ventilator were excluded from the study.

All patients who underwent open thoracotomy received a thoracic (T5-T11) epidural catheter before induction of anaesthesia and an epidural infusion of bupivacaine 0.1% + fentanyl 2 mg/ml for intraoperative and postoperative analgesia which is our standard of care. All of them received intravenous (IV) paracetamol 1g at the end of the procedure. Other postoperative analgesia modalities used were IV or oral NSAIDs, fentanyl via IV patient controlled analgesia (PCA) pump, oral anticonvulsants or antidepressants, depending on the severity of pain and adequacy of pain relief, or as rescue analgesia for incisional pain as per the protocols at the institute and the decision of the pain consultant.

Pain in the shoulder and the thorax (incisional site) during rest and activity was assessed on a 11-point numerical rating scale (0 = no pain, 10 = worst imaginable pain) or visual analogue scale (0 = no pain, 1-3 = mild pain, 4-6 = moderate to severe pain, 7-9 = very severe pain, 10 = worst pain possible) separately by an independent assessor at 1 hour, 6 hours, and 12 hours during the patient’s stay in the postanaesthesia care unit.

Pain in the shoulder during activity was defined as that occurring with 90° abduction of the shoulder ipsilateral to the surgery and thoracic pain during activity was defined as pain during coughing and breathing. Patients were asked about shoulder pain with movement (maximum passive forward elevation, internal and external rotation, adduction, as per the American Shoulder and Elbow Surgeons standardised assessment method).[9]

Data including age, gender, body mass index (BMI), ASA physical status, diagnosis, side and type of surgery, surgical approach, positioning during surgery, duration of surgery, type of regional anaesthesia, and level of catheter placement in patients with epidural anaesthesia were recorded.

Demographic data were summarised with descriptive statistics. Continuous data were represented as mean (standard deviation) or median [interquartile range] and categorical data were reported in counts (percentage), respectively. The Shapiro-Wilk test was used to check normality of each variable. Descriptive analysis was performed to identify distribution of variables under study. The univariate analysis using Chi-square test with Pearson’s and Fisher’s exact test was used to associate demographic and clinico-demographic variables with the outcome, viz ipsilateral shoulder pain. All potential predictors were analysed in a multivariable penalised logistic regression model, using ISP as an outcome variable. Backward stepwise deletion based on the Wald test was applied. For all tests, P values were two-sided and a P value less than 0. 05 was considered statistically significant. As the data were collected over a period as mentioned above, there was no sample size calculated.

RESULTS

Of the two hundred and ninety six patients who underwent thoracic surgeries, 118 (39.8%) developed ISP. A total of 170 of the 296 patients underwent thoracotomy, 77 (45.29%) of them developed ISP. Furthermore, among 110 patients and 13 patients who underwent VATS and robotic thoracic procedures, respectively, 36 (32.7%) patients in the VATS group and 5 (38.4%) in the robotic group developed postoperative ISP. None of the 3 patients who underwent surgery via trans-hiatal approach complained of ISP. Patient demographics are shown in Table 1. The overall results were not statistically significant [Table 2, P = .093]. There was no statistically significant contribution toward development of ISP with respect to gender, BMI, and ASA physical status, except for the age. Among the patients who developed ISP, majority of them, 51 (43.2%) patients were aged more than 65 years, while the incidence was lower at 21.2% in those aged less than 45 years, which was statistically significant as per univariate analysis [Table 2, P = .007]. However, multivariate analysis using penalised logistic regression model of the same showed no statistically significant influence of age on the development of ISP [Table 3, P = .502].

Table 1.

Patient characteristics, duration of surgery, duration of anaesthesia, duration of supine position, and duration of lateral decubitus position

Variables	Ipsilateral shoulder pain Mean±SD
	No	Yes	Total
Age (years)	49.55±14.86	51.59±12.66	50.36±14.04
Gender (male/female)	119/59	67/51	296
BMI	23.16±4.36	23.39±4.62	23.25±4.46
ASA (1/2/3)	93/75/10	52/58/8	296
Duration of surgery (min)	228.48±111.31	289.31±100.94	252.73±111.20
Duration of anaesthesia (min)	286.60±122.24	357.25±110.46	314.76±122.50
Duration of supine position (min)	138.03±107.14	176.57±108.12	153.40±109.00
Duration of lateral decubitus position (min)	146.26±85.48	180.42±94.96	159.88±90.78

Data are shown as mean±SD, or number (%) of patients. SD, Standard deviation; BMI, Body Mass Index; ASA, American Society of Anesthesiologists

Table 2.

Univariate analysis: Data are shown as number (%) of patients

	ISP		P
	Yes	No
Age (years)
<45	25 (21.2)	60 (33.7)	0.007
45-65	42 (35.6)	37 (20.8)
>65	51 (43.2)	81 (45.5)
Gender
Male	67 (56.8)	119 (66.9)	0.079
Female	51 (43.2)	59 (33.1)
BMI
Under weight	16 (13.6)	27 (15.2)	0.959
Normal weight	64 (54.2)	93 (52.2)
Over weight	29 (24.6)	46 (25.8)
Obese	9 (7.6)	12 (6.7)
ASA
Grade 1	52 (44.1)	93 (52.2)	0.386
Grade 2	58 (49.2)	75 (42.1)
Grade 3	8 (6.8)	10 (5.6)
Side
Midline	1 (0.8)	6 (3.4)	0.227
Left	29 (24.6)	51 (28.7)
Right	86 (72.9)	114 (64)
Both	2 (1.7)	7 (3.9)
Approach
Transhiatal	0 (0)	3 (1.7)	0.093
Thoracotomy	77 (65.3)	93 (52.2)
VATS	36 (30.5)	74 (41.6)
Robotic	5 (4.2)	8 (4.5)
Regional_analgesia
None	5 (4.2)	10 (5.6)	0.002
Epidural	105 (89)	128 (71.9)
Intercostal	5 (4.2)	34 (19.1)
Paravertebral	3 (2.5)	6 (3.4)
Epidural block level
None	13 (11)	50 (28.1)	0.003
T5-T6	7 (5.9)	7 (3.9)
T6-T8	21 (17.8)	35 (19.7)
Below T8	77 (65.3)	86 (48.3)
Duration of surgery (min)
<207.50	21 (17.8)	85 (47.8)	<.001
>=207.5	97 (82.2)	93 (52.2)
Duration of lateral decubitus position (min)
<185	67 (58.8)	130 (77.8)	0.001
>=185	47 (41.2)	37 (22.2)
Shoulder pain onset after_surgery (min)
Early onset (15-60)	83 (70.3)	2 (1.1)	<.001
Delayed onset (>60)	35 (29.7)	2 (1.1)

ISP, Ipsilateral Shoulder Pain; BMI, Body Mass Index; ASA, American Society of Anesthesiologists; VATS, video-assisted thoracoscopic surgery

Table 3.

Multivariate analysis: Penalised logistic regression

	Odd ratio	Lower bound	Upper bound	P
Age (years)
<45	Ref
45-65	3.14	0.11	88.59	0.502
Gender
Female	Ref
Male	4.39	0.08	241.11	0.469
Surgical approach
Transhiatal	Ref
Thoracotomy	0.06	0.00	2.86	0.155
VATS	0.04	0.00	2.46	0.128
Regional Analgesia
None	Ref
Epidural	60.65	0.00	19,00,352.59	0.437
Intercostal	2.43	0.03	217.87	0.698
Epidural_level
None	Ref
T5-T6	1.26	0.00	12,457.11	0.961
Shoulder pain onset after_surgery (min)
No	Ref
Early onset (15-60)	236.17	5.61	9,936.70	0.004
Duration of surgery (min)
<207.50	Ref
>=207.5	0.97	0.02	40.47	0.989
Duration of lateral decubitus position (min)
<185	Ref
>=185	0.11	0.00	2.80	0.182

VATS : Video-assisted thoracoscopic surgery

In addition, 13 (4.4%) patients complained of contralateral shoulder pain.

The incidence of ISP was the highest at 41.89% among those who had lung cancer (n = 74). Those with disease involving right upper lobe and left upper lobe showed higher incidence, 29% and 25.8%, respectively, when compared to other areas [Figure 1].

Figure 1.

Incidence of ipsilateral shoulder pain (ISP) (n) in patients with Cancer lung according to lobes involved. CA: Cancer

An epidural catheter was inserted for postoperative analgesia in 233 (78.7%) patients. One hundred seventy one (57.7%) patients required intravenous opioids administered via PCA devices for surgical pain. Twenty six patients also received co-analgesics such as anticonvulsants or tricyclic antidepressants for the surgical pain.

Furthermore, among the patients who developed ISP, majority (97 patients, 82.2%) had surgery which lasted over 207.5 minutes. Patients who were in lateral decubitus position during the surgery for less than 185 minutes showed lower incidence of ISP (67 patients, 34%) [Table 2]. Among the patients who had ISP, 91 (77.1%) expressed it to be of a dull aching, whereas 25 (21.2%) as stabbing in nature [Figure 2]. The pain was located in the posterior, anterior, and superior aspect of the shoulder in 66 (55.9%), 16 (13.5%), and 10 (8.47%) patients, respectively [Figure 3]. Among those with ISP, at rest, in 2 (1.7%) patients, it was of moderate intensity while the rest [116 (98.3%)] had mild intensity of pain. On movement however, 32 (27.10%) patients complained of ISP of moderate intensity and in 86 (72.9%) the intensity was mild.

Figure 2.

Nature of ipsilateral shoulder pain (ISP) as complained by the patients (%)

Figure 3.

Different locations of ipsilateral shoulder pain (ISP) as complained by the patients (%)

DISCUSSION

In our study conducted over a period of one year, we found a 39.8% incidence of ISP in patients undergoing thoracic surgeries. Prior studies[7,8,10-15] have reported a 31%-85% incidence rate.This wide variation in incidence may be due to the differences in surgical and anaesthetic technique varying in each institute. Majority (65.3%) of those with ISP had undergone thoracotomy. This may be attributed to the fact that thoracotomy was the surgical approach in majority (57.4%) of our study population. Except for 3 surgeries performed via trans-hiatal approach which did not have any ISP, the incidence of ISP was not very different in those who underwent open thoracotomy (45.29%), VATS (32.7%), and robotic-assisted thoracic surgery (38.4%). These results were also statistically not significant. Manzoor S et al.[16] in their study on phrenic nerve infiltration on ISP had shown an incidence of 43%, at 12 hours postsurgery, in the patients who had not received phrenic nerve block.In a similar study on 205 patients by Bunchungmongkol N. et al.,[17] the incidence of ISP was found to be 47.3%, slightly higher than our findings. However, they too found a higher incidence of ISP in patients who underwent thoracotomy (58.7%) as opposed to those who underwent a VATS procedure (20. 9%).

Ohmori A et al.,[10] in their study, found a significantly higher incidence of postoperative shoulder pain following segmentectomy. In our study, we found a higher incidence of ISP in those who underwent oesophagectomy, especially via the transthoracic approach. This may be probably due to the fact that a large number (43%) of our study population consisted of patients posted for oesophagectomies. But otherwise no significant association could be found, probably due to the small cohort. However, among the patients posted for lung surgeries, higher incidence was seen in those with carcinoma involving upper lobes.

This suggests that the brachial plexus, especially the lower trunk might get injured due to its close proximity to the field of surgical resection and leads to shoulder pain, or if it is a superior sulcus tumour (nonsmall cell lung carcinoma), the tumour itself can invade the plexus and also the stellate ganglion, which causes pain in the shoulder and arm. In our study, 77.11% of the patients described ISP as dull aching in nature and in about half of them, was located at the posterior aspect of the shoulder. Also, when the arm was kept at rest, the intensity of pain was always mild. Similarly, Bunchungmongkol N. et al.[17] reported the pain as dull aching in most patients (87%) and the pain was located at the posterior aspect of the shoulder in half of their patients. However, 67% patients complained of pain of moderate to severe intensity, while in our study, none had pain of severe intensity. In 62% of their patients, pain was aggravated with shoulder movement. Similarly, in our study we found that, in 27.1% patients, the pain aggravated with shoulder movement. The relatively lower incidence could be attributed to variations in definitions and pain measurements, different surgical mix, and also the ethnicity which is known to affect pain perception.

Majority of our patients (78.7%) received thoracic epidural analgesia, 39 patients an intercostal block, while 9 patients a paravertebral block. Fifteen patients did not have regional analgesia in the postoperative period. Irrespective of this, all our patients received multimodal analgesics round the clock. 84.4% of them received NSAIDS and all received paracetamol. About half of them also received intravenous opioids through IV-PCA. Despite the above analgesia, a third continued to have ISP. Gerner P et al.[18] do mention that ISP following thoracic surgeries was relatively resistant to intravenous opioids and only partially relieved by NSAIDs. Similar to our results, Bunchungmongkol, N et al.[17] in their study reported 73% patients developed ISP despite effective epidural analgesia. BlichfeldtEckhardt MR et al.[19] in their editorial reported that either interscalene or suprascapular or phrenic nerve block approaches should be studied for the management of ISP.

Furthermore, we found that ISP was most common (43.2%), although not statistically significant, in patients aged >65 years. Also, no association of ISP either with age or gender could be found.

Bunchungmongkol N et al.[17] identified potential risk factors for ISP as surgery using the thoracotomy approach (P = .014) and surgical duration >120 minutes. However, these potential risk factors were not found to be statistically significant in our study. Even a prolonged duration of lateral decubitus position during surgery did not translate into a statistically significant incidence of ISP. It is possible that our study was not powered adequately to demonstrate an association of ISP with these possible risk factors.

Misiolek H et al.[20] in their study found statistically significant differences between the groups (group I without ISP and group II with postoperative ISP) for BMI (24.67 and 27.68, respectively; P = .049), type of surgery (24% for thoracotomy and 0% for VATS, P = .026), and level of epidural catheter placement (4.35% for catheters placed at the level of T5 or higher and 40.47% for catheters placed below T5; P = .003). No such differences were noted in our study. No epidural catheter was placed above T5 in our study; however, incidence of ISP for epidural catheters placed at above T6 was lower (5.9%) when compared to below T6 (T6-T8,17.8%; <T8, 65.3%). But no statistically significant association could be found. Since there is a dearth of data of ISP in Indian population, we wanted to study the incidence and risk factors of ISP in the Indian cancer patient population.

There were several limitations in our study. One of the main limitations was that patients were followed up to 12 hours after surgery; a longer follow-up to 48 hours or beyond would have helped know the natural course of ISP.[3] Long-term studies with larger cohort and suitably powered would have possibly helped identify risk factors for ISP. The duration of surgery and associated factors would depend on the skill and years of experience of the surgeons and hence, the results of this study may not apply to other centres.

CONCLUSION

Our study demonstrated 39.8% incidence of ipsilateral shoulder pain in patients undergoing thoracic surgeries. ISP was chiefly dull aching in nature, of mild to moderate intensity, most commonly located on the posterior aspect of the shoulder. It was more common in those who underwent thoracotomy and more than 65 years age group. There was no association of ISP with possible risk factors. Much larger cohort studies which are designed and powered to demonstrate a possible association are required in the future.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.