“Prevalence of chronic pain following resection of pelvic bone tumours: A single centre prospective observational survey”

Abstract

Background

Hemipelvectomy is a major surgery most often performed for pelvic malignancy. These complex surgeries often involve dissection around major neurovascular bundle and resection of tumour being bone along with involved tissues. This may result in short and long term morbidities. There is very little literature about incidence of chronic pain after pelvic resections. We conducted a prospective study at a tertiary cancer hospital to assess the prevalence of chronic pain post hemipelvectomy.

Method

This is a single centre prospective observational study conducted over 30 months. Pain scores were recorded using Brief pain inventory (BPI) and pain detect questionnaire. The quality of life was assessed using musculoskeletal tumour society (MSTS) score. Intra-operative details like extent of surgical resection, nerves spared, details of intra-operative and post-operative analgesia were retrieved from the patient files. Data were analysed using SPSS 21 version.

Results

Neuropathic pain post hemipelvectomy was uncommon. The prevalence of mild to moderate somatic pain was around 30%. Functional limitation was minimal as assessed by BPI and MSTS score. A high incidence of numbness was seen to persist in and around the area of surgical incision (50%).

Conclusion

This is first study to report the incidence of chronic pain post hemipelvectomy done for pelvic tumour resections. Despite the extensive nature of resection involved, there is a low prevalence of neuropathic pain in this population. However, incidence of persistent somatic pain is high and there is a need for further studies for evaluating the causality

Introduction

Primary bone tumours account for a small percentage of all cancers. ¹ Data from Surveillance, Epidemiology, and End Results Program (SEER) of the National Cancer Institute USA estimates a total of 3600 (1 per 100,000 men and women) new cases of bone and joint cancer to occur in 2020. This figure accounts for only <0.2% of all new cancer cases and within this group of primary bone cancers pelvic tumours account for 15–20% of cases.^2,3 However, the percentage of survival at 5 years among these patients now stands at 66%.^1,4

Involvement of the pelvic bones by primary tumours may occur as primary bone tumours or due to soft tissue neoplasms which affect the pelvis in contiguity. ² Over the years, hemipelvectomy has become an established mode of treatment for primary pelvic tumours. ⁵ Hemipelvectomy can be done in two different ways: external (with limb amputation) and internal (with limb preservation). In the last few decades, the use of external hemipelvectomy for the treatment of primary pelvic tumours has declined with greater efforts to preserve the limb by internal hemipelvectomy. ⁵ Nevertheless, both procedures are fraught with complications, involve extensive resections, major blood loss and transfusion and a prolonged recovery phase. ⁶ In some situations, this may involve sacral nerve roots sacrifice or injury.

Though there is literature regarding the oncological, functional and survival outcomes post hemipelvectomy,^5–9 data regarding acute and chronic pain outcomes post hemipelvectomy is sparse. We planned this study with the objective to look into the prevalence and nature of chronic pain in the survivors of primary pelvic tumours. Additionally, we studied whether there was any association between intra-operative nerve handling, perioperative pain management and persistent post-operative pain. We also evaluated the functional outcomes using musculoskeletal tumour society (MSTS) score in these patients.

Materials and methods

This is a prospective observational study carried out at a single centre tertiary care oncology institute over 30 months, from November 2015 to April 2018 (6/11/2015–6/04/2018). After approval from Institutional review board (IRB), screening was done of all consecutive adult patients who attended the follow-up clinic after having completed the primary treatment (resection of primary pelvic tumour) at least 6 months prior to the date of interview. Fifty patients were screened and 40 patients were enrolled in the study. Each patient would be interviewed only once in the course of the trial.

The patients were asked questions with respect to their pain, which was recorded using Brief pain inventory (BPI) and pain detect questionnaire. The functional evaluation was done using musculoskeletal tumour society (MSTS) score. ¹⁰ As the various pain scales are not available in local languages, these questions were administered by the investigators current pain medication if any was recorded. Intra-operative details like extent of surgical resection, sacrifice/injury to major nerves and details of intra-operative and post-operative analgesia were retrieved from the patient files. Any obvious motor weakness on the operated side and gait of patient was recorded. Data were expressed as percentages and in mean (±SD) or median (±IQR) as applicable. Correlation between causality factors and categorical scales (BPI, MSTS scores) was done using Chi square test while t test was used for analysis with continuous scales like pain detect. Data were analysed using SPSS 21 version.

Results

Fifty patients were screened. Four were excluded as they were not adults and six patients had already participated in the trial on a previous follow-up, hence 40 patients were enrolled. Out of the 40 patients that were enrolled, 25 were males and 15 were females. The mean age of patients was 34 years (Range 18–63). The histopathology was chondrosarcoma in 15 patients, Ewing’s sarcoma, ¹¹ Osteosarcomas, ⁵ Giant cell tumour in ⁵ and soft tissue sarcomas in ⁴ patients. Three patients were hypertensive; rest of the patients did not have any significant coexisting medical issues. In 38 patients, the limb was salvaged. Pelvic resections were classified according to Enneking and Dunham into four types; iliac (T1), acetabular (T2), pubis or ischium (T3) and sacral (T4) or combinations of the same. ¹¹ Only in two cases external hemipelvectomy had been performed. Details of surgery are shown in Table 1. In the patients with limb salvage, ischiofemoral arthrodesis was done in four patients, reconstruction with mesh was done in 20 patients, shelf prosthesis was used in one patient, and no reconstruction was done in rest of patients. The primary cause for heterogenicity of the study population is the overall rare occurrence of this cancer and the additional variation due to the underlying histology of the disease. Femoral nerve was involved and hence sacrificed in one patient, femoral nerve was injured in one patient, obturator nerve was sacrificed in five patients, and sciatic nerve was handled in almost all patients. One patient had persistent peroneal nerve palsy post-operatively. Fifteen patients had wound healing issues. Nine of these patients resolved with conservative management, three patients needed re exploration of the wound, one patient underwent implant removal (screws) and two patients needed plastic surgery intervention in the form of flaps. Nine patients received post-operative radiation, one patient received pre-operative radiation. 16 patients received neo adjuvant and adjuvant chemotherapy as per Institute protocol. In addition, two out of five patients with osteosarcoma received only post-operative/adjuvant chemotherapy.

Table 1.

Type of pelvic resections performed based on classification given by Enneking and Dunham.

Type of surgery
External hemipelvectomy		2
Internal hemipelvectomy	Type 1	10
	Type 3	2
	Type 1 + 2	6
	Type 2 + 3	6
	Type 1 + 4	2
	Type 1 + 2 + 3	7
	Type 1 + 2 + 4	4
	Type 1 + 2 + 3 + 4	1

With respect to perioperative pain, five patients were on pre-operative painkillers, all related to the primary disease. These were limited to the use of over the counter available simple analgesics like paracetamol and non-steroidal anti-inflammatory drugs. In the pre-operative period, epidural was inserted in 36 patients. In 29 patients’ epidural analgesia included local anaesthetic infusion while in seven patients, only epidural opioids were given. Four patients received intravenous opioids through patient-controlled analgesia pumps. All patients received an individualized rehabilitation program post-surgery which included exercises and physiotherapy. Nine patients needed analgesics for a week after discharge. There was no correlation between delayed wound healing and need to continue analgesics after discharge (p = .5).

The average duration of follow-up after surgery among the patients studied was 4.6 (2.1–7.1) years. One patient had local disease recurrence, while three patients had distant recurrence. Only three patients were on pain medication on follow-up of which two patients were taking painkillers over the counter and one patient was on tablet pregabalin once a day on irregular basis. One patient had received an interventional block in the past. Thirty percent of patients (30%) were complaining of any pain at time of assessment when pain scores were recorded using BPI. Mild to moderate pain was present in 25 (62.5%) patients in the last 24 hours. Details of pain assessment using the BPI has been summarized in Table 2. Seventeen percent of patients had tingling/pricking sensation around the operated side, while 50% of patients had numbness of slight to moderate degree post-surgery. The median pain detect score was 3 (IQR1-7) with zero scores in 14 patients. In 25 patients, the score was less than 12, suggesting that neuropathic component was unlikely, and in one patient the scores were 17/38 suggesting that neuropathic pain cannot be ruled out. Details of symptoms of pain detect are summarized in Table 3. The mean MSTS score was 24 (±5) (on scale of 0–30, higher scores indicating better function.) (Table 4). We found no correlation between chronic persistent post-surgical pain (CPSP), analysed by BPI scores at measurement and post-operative analgesia technique (p = .4) There was no correlation between chemotherapy (p = .6), radiotherapy (p = .6) and incidence of neuropathic pain. Only two patients underwent external hemiplevectomy. There was no correlation between type of surgery and neuropathic pain scores recorded using Pain detect (p = 1.0) 82.5% of patients experienced modest to no disability in walking ability and gait post-surgery.

Table 2.

Summary of BPI findings.

		Median (IQR)	No of patients as per severity (%)
			Score 0	Mild symptoms (1–3)	Moderate symptoms (4–6)	Severe symptoms (7–10)
Worst pain in last 24 h		1 (10)	25 (62.5)	9 (22.5)	5 (12.5)	1 (2.5)
Least pain in 24 h		0 (5)	31 (77.5)	6 (15)	3 (7.5)	—
Average pain in 24 h		1 (5)	15 (37.5)	20 (50)	5 (12.5)	—
Pain right now		0 (5)	28 (70)	10 (25)	2 (5)	—
Pain affecting activity	General activity	0 (5)	24 (60)	14 (35)	2 (5)	—
	Mood	0 (7)	30 (75)	9 (22.5)	—	1 (2.5)
	Walking disability	1 (0)	18 (45)	19 (47.5)	3 (7.5)	—
	Normal work	0 (5)	21 (52.5)	17 (42.5)	2 (5)	—
	Relation with people	0 (5)	35 (87.5)	4 (10)	1 (2.5)	—
	Sleep	0 (5)	26 (65)	12 (30)	2 (5)	—
	Enjoyment of life	0 (5)	32 (80)	6 (15)	2 (5)	—

Table 3.

Summary of findings of pain detect.

Symptoms	No of patients (%)
	Never	Hardly noticed	Slightly	Moderately	Strongly	Very strongly
Do you suffer from a burning sensation (e.g., stinging nettles) in the marked areas?	30 (75)	3 (7.5)	3 (7.5)	4 (10)	—	—
Do you have a tingling or prickling sensation in the area of your pain (like crawling ants or electrical tingling)?	28 (70)	5 (12.5)	5 (12.5)	2 (5)	—	—
Is light touching (clothing, a blanket) in this area painful?	36 (90)	2 (5)	2 (5)	—	—	—
Do you have sudden pain attacks in the area of your pain, like electric shocks?	31 (77.5)	2 (5)	6 (15)	1 (2.5)	—	—
Is cold or heat (bath water) in this area occasionally painful?	35 (87.5)	4 (10)	1 (2.5)		—	—
Do you suffer from a sensation of numbness in the areas that you marked?	12 (30)	8 (20)	10 (25)	10 (25)	—	—
Does slight pressure in this area, e.g., with a finger, trigger pain?	30 (75)	3 (7.5)	4 (10)	2 (5)	1 (2.5)	—

Table 4.

Summary of MSTS lower extremity.

Median	Description with prevalence expressed as, n (%)
Pain score	0	1	2	3	4	5
Pain score	—	—	5 (12.5)	3 (7.5)	7 (17.5)	25 (62.5)
Function	—	2 (5)	3 (7.5)	7 (17.5)	12 (30)	16 (40)
Emotional acceptance	1 (2.5)	—	3 (7.5)	8 (20)	11 (27.5)	17 (42.5)
Walking ability	—	—	7 (17.5)	10 (25)	12 (30)	11 (27.5)
Galt	1 (2.5)	3 (7.5)	3 (7.5)	19 (47.5)	8 (20)	6 (15)
Supports	1 (2.5)	6 (15)	1 (2.5)	—	2 (5)	30 (75)

Discussion

This prospective survey in patients post pelvic tumour surgery is reassuring as the incidence of neuropathic pain is minimal. However, this study also highlights that CPSP remains a possibility after pelvic bone tumour surgery.

There have been various review articles in the past that study functional and oncological outcomes along with the incidence of surgical/wound complications post hemipelvectomy.^7,12–14 But very few of them report the incidence or prevalence of CPSP.^15,16 In a previous study, the incidence of chronic pain following hemipelvectomy was around 40%; however, diagnosis of chronic pain was made based on presence of pain after 12 weeks. Detail pain assessment was not a part of the protocol, with the study primarily looking at survival rates after hemipelvectomy. ¹⁵ Our study shows a 30% incidence of CPSP in patients undergoing resection surgeries for pelvic tumours. We also report a high incidence of numbness in and around the area of surgical incision (50%) which tends to linger for a long time. In a study conducted by Furtado et al., ¹⁶ amongst 250 patients who underwent lower extremity amputation for bone or soft tissue tumours the incidence of mild and moderate pain was 48.4% and 33.95, respectively. Amongst the patients with severe pain, only one had undergone hemipelvectomy. ¹⁶ Though the incidence of CPSP is high, the severity of CPSP and incidence of neuropathic pain post hemipelvectomy is low. Also, we found minimal functional limitation as assessed by BPI and MSTS score.

Various studies have identified a number of pre-, intra- and post-operative risk factors for the development of CPSP and neuropathic pain.^17–19 Cancer population has multiple risk factors for developing CPSP, which partly explains the incidence of CPSP in our study population. Pre-operative pain and psychological factors have been noted to be strong predictors for CPSP.^20,21 Around 12.5% of our study population reported pre op pain needing analgesics. Unfortunately, as patients were recruited in a post-surgery follow-up clinic, we were not able capture data regarding the pre op psychosocial factors in this cohort. Also, genetic factors affecting chronic pain is a new area of research. A recent study identified CACNG2 gene polymorphism responsible for chronic neuropathic pain in cancer patients post mastectomy. ²² Genetic susceptibility to pain is therefore possibly another factor that could explain these results.

The reported association of CPSP with intra-operative nerve handling and damage has been variable. In surgeries such as mastectomy, thoracotomy and amputation where mechanisms and risk factors for CPSP are better studied, it has been reported that more than 50% of patients with nerve and tissue damage develop CPSP.^23,24 Yet, there are other studies which state that there is no consistency with nerve injury and development of CPSP. ²⁵ In our study, despite the fact that sciatic nerve handling was inevitable in all patients, we have a low incidence of neuropathic pain our patients. This observation that has been replicated in other studies as well. In patients following thoracic surgery with CPSP, only half had significant neuropathic symptoms identified by a validated questionnaire. ²⁶ An interesting observation by Richardson et al. ²⁴ is that rib resection, which results in more intercostal nerve damage, is associated with a reduced incidence of post-thoracotomy neuralgia.

Numbness post-surgery is not unusual. Around 69% of patients who underwent thoracotomy complained of numbness, which was not related to intra-operative nerve handling/damage; also presence of numbness was also not found to be associated with post op development of pain syndromes. ²³ A similar finding is seen in patients following mastectomy. It was seen that intra-operative damage to the intercostobrachial nerve with objective signs of nerve injury (such as numbness) did not essentially result in chronic pain. ²⁷ There are other studies which have found no association between intercostal nerve damage assessed at the time of thoracotomy and development of chronic pain later.^24,28 In our study, though numbness was present in nearly half the patients, neuropathic pain was not seen in this group.

There remains a high incidence of complication after hemi pelvic surgery, with one study reporting a complication rate of 50–80%. ⁶ The most common complication reported is flap necrosis and wound infection. These complications can lead to prolonged hospitalization which could be associated with pain. In our study, population the incidence of wound complication was 35%; however, there was no significant association of wound infection or re-exploration with post-operative pain. Other functional outcomes such as mobility are obviously better after internal hemipelvectomies than external. ⁶ With only 5% of our study patients undergoing external hemipelvectomy, functional outcome in terms of mobility is similar to those of previous reported studies.^6,12 A few studies report an observed higher incidence of acute and chronic post-operative pain in patients post external hemipelvectomy, but there is no proved statistical co relation between the two.^6,15

This study is not without limitations. Annually, around 15–18 pelvic surgeries are performed at our institute. We aimed to include over a hundred patients in the selected time frame. However, after excluding paediatric population we were able to interview only 40 patients in 30 months. We were also unable to capture pre-operative data about psychological factors in the study population such as anxiety, catastrophising; which are known to affect CPSP. ²⁹ Pain detect and BPI have been translated and adapted to various languages (Spanish, Japanese, etc.) studies have proven that when such adaptations are performed under supervision with help of independent translators they continue to be as reliable and valid as original versions.^30,31 Whilst in our study, we did not provide translated versions but the investigators have tried their best to maintain the integrity of questions and maintain minimal bias whilst conducting the interview.

It is however, the first study to exclusively look into the prevalence of pain in this population.

Conclusion

In conclusion, this is the first study to report the incidence of chronic pain post hemipelvectomy done for pelvic tumour resections. It is heartening to see that despite the extensive nature of resection involved and presence of several risk factors associated with the CPSP, there is a low prevalence of neuropathic pain in this population. The prevalence of mild to moderate somatic pain was around 30%. A high incidence of numbness was seen to persist in and around the area of surgical incision (50%). Functional limitation was minimal as assessed by BPI and MSTS score.

There is need to conduct appropriately designed studies to study casualty between persistent somatic pain and perioperative risk factors in this population.