A study exploring the role of intercostal nerve damage in chronic pain after thoracic surgery

Reports in the literature find the prevalence of chronic pain after thoracic surgery to be 9–80% for thoracotomy and 5–33% for VATS (video-assisted thoracoscopic surgery) [1–14]. The prevalence may be influenced by a number of factors. The patients’ perception of post-operative pain may be influenced by the style of explanation of the operation and inpatient stay that prepares the patient for surgery. It may be influenced by differences in intra-operative technique, method of anaesthesia, method of post-operative analgesia and the experience of the operating surgeon. Of particular relevance is how the authors defined chronic pain, how long after surgery it was assessed, and how the data was collected, analysed and presented. Some studies have also reported the impact the pain has on patients’ lives and how it limits their daily activity [1–3]. Chronic pain after thoracic surgery is clearly a significant problem that requires further investigation. The cause of the chronic pain has not been established. A study has shown that high levels of immediate post-operative pain is associated with an increase in the likelihood of chronic pain, which suggests it is a process that starts around the time of the operation rather than secondary to tissue healing [4]. Rogers describes numerous possible causes for pain after thoracic surgery [15]. There is evidence to suggest chronic pain results from nerve damage at the time of surgery: studies have shown intercostal nerve damage by demonstrating loss of intercostal nerve conduction during surgery and the loss of abdominal reflexes after thoracic surgery is associated with pain [16,17]. A strong neuropathic component is also indicated by chronic pain being frequently accompanied by abnormal cutaneous sensation and often treated with antidepressants and antipsychotics rather than opiates. Our aim was to investigate the prevalence of the nerve damage during thoracic surgery and its association with chronic pain. We also explored the intra-operative factors that may influence nerve damage and chronic pain. The study was performed in the Thoracic Surgery Unit in Nottingham City Hospital and ran from December 2004 to September 2005. Local ethics committee approval was obtained prior to commencement of the study. Patients were selected from the planned operating lists for elective thoracotomy or thoracolaparotomy. Only one eligible patient per operating list was studied. List construction was made independently of this study. The eligible patient that was earliest on the operating list was selected for the study. Children (under 18 years old) were not involved in the study. Patients were excluded if they had previous thoracic wall trauma such as multiple rib fractures or previous thoracic wall surgery involving a thoracotomy or VATS (video-assisted thoracic surgery) because of the possibility of existing intercostal nerve injury. Patients were also excluded if they had a known neurological disease with peripheral manifestations including neuropathy associated with diabetes because their disease may have made interpretation of nerve function measurements difficult. There was a potential risk that the equipment used to stimulate intercostal nerves could affect battery-powered medical devices and therefore patients with a cardiac pacemaker, cardiac defibrillator or cochlear implants were excluded. Refusal of consent excluded the patient from the study. Patients were invited to participate in the study in pre-assessment clinic or on admission to the ward. The study was explained verbally and an information sheet supplied. Written informed consent was always obtained. Prior to surgery patients completed a pain questionnaire: the Short-form McGill Pain Questionnaire (SF-MPQ) which consists of three pain scales: word descriptors with an intensity scale (pain score 1, maximum 45), a visual analogue scale (pain score 2, results presented as a proportion) and a tick-box overall intensity score (pain score 3, range 0–5) [18]. A Hospital Anxiety and Depression Scale questionnaire (HADS), maximum score 42, was also completed [19–22]. We attempted to measure intercostal nerve function in fully conscious participants pre-operatively using a magnetic nerve stimulator (magstim). The nerve stimulator coil was applied perpendicular to the skin overlying the proximal region of the intercostal nerve just lateral to the vertebral body. Two cutaneous button electrodes were placed on the anterior chest overlying the distal portion of intercostal muscle and used to record muscle-evoked potentials (MEPs). A reference cutaneous electrode was placed on the contralateral border of the sternum. The magstim delivered a short magnetic pulse that excited the intercostal nerve, which in turn stimulated the intercostal muscle and the MEP recorded. The patient felt muscle contraction though this was never uncomfortable in any cases. A nerve stimulator/recording device was used to provide precise impulse delivery and MEP measurements (Neurosign 800, Magstim Company, Whitland, Carmarthenshire). Intra-operative intercostal nerve conduction studies were performed, based on the method reported by Rogers [17]. Under general anaesthetic, with the ribs and intercostal muscles exposed, a hand-held monopolar electrical nerve stimulator probe was used to excite the intercostal nerve just lateral to erector spinae. A ground electrode was placed in erector spinae muscle. A pair of recording needle electrodes, mounted 1 cm apart on a rubber-shod right-angled hand-held clamp, was inserted into the most anterior portion of the corresponding intercostal muscle, and MEPs recorded. A reference electrode was placed in serratus anterior muscle. The stimulus was increased from 3.5 mA (with a 300 μs pulse width) upto a maximum of 10 mV, until a supramaximal stimulus was achieved and then the MEP recorded. The latency between stimulus and MEP was deduced and the velocity calculated by dividing the distance between stimulating probe and recording electrodes by the latency. A nerve conduction measurement was made at each of the intercostal spaces one and two above the intercostal incision and one and two spaces below the incision. Measurements were made at four stages during the operation: (i) before the intercostal muscle space incision; (ii) after the intercostal muscle incision but before insertion of the rib retractor; (iii) following the lung or oesophageal resection just after the rib retractor was removed; (iv) after closure of the intercostal space. The surgeon was blinded to the results of the MEP measurements, although he could see whether the intercostal muscle visibly twitched or not. These measurements added 10–20 min onto the length of operation. In order to perform these intercostal nerve conduction studies the method of anaesthesia had to be modified to reduce variations in factors that may influence nerve conduction and to prevent complete blockade of the neuromuscular junction. Core and peripheral temperatures were monitored and the patient cooled or warmed externally to maintain the intercostal nerve temperature as steady as possible. Induction of anaesthesia was performed using 1 μg/kg remifentanil bolus and 1.5–2.5 mg/kg propofol. Anaesthesia was maintained using desfluarane (1.0–1.3 minimum alveolar concentration) and an intravenous infusion of remifentanil. Relaxation was achieved using atracurium: a 0.5 mg/kg bolus followed by an infusion which was adjusted to maintain a minimum two visual twitches of adductor pollicis upon a train-of-four (TOF) stimulation of the ulnar nerve. Immediate post-op analgesia was provided by an epidural infusion of local anaesthetic and opiate or a paravertebral infusion of local anaesthetic with an intravenous morphine infusion under patient control (PCA), though no local anaesthetic was infused prior to the last nerve conduction measurement. Operations were performed by one of three teams each led by a consultant (consultants A, B, C). For lung resections all used similar posterolateral thoracotomy skin incisions, divided latissimus dorsi muscle and approached through the 5th or 6th space. Consultants B and C always entered the pleural space through sub-periostial rib resection and closed the intercostal space with a continuous intercostal muscle suture. Consultant A entered using diathermy along the upper border of the rib, without rib resection, and closed using three interrupted sutures that ran over the upper border of the rib above and under the lower border, but above the intercostal bundle, of the rib below (‘pericostal’ closure). For oesophageal operations consultant B used a thoracotomy incision with phrenotomy and consultants A and C used a thoracolaparotomy incision without rib resection. In each case the maximum spread of ribs at the thoracotomy and the length of time the retractor was in situ were recorded. Patients were blinded to the results of the intra-operative nerve conduction studies. At the 6-week outpatient review the participants completed a SF-MPQ pain questionnaire. Their chest wall was examined at that time for pin-prick sensation around the surgical wound and drain sites. This area was marked with ink and traced onto paper where it was measured using paper marked with 1 mm and 1 cm gridlines. If patients were unable to attend the base hospital (in Nottingham) the pain questionnaire was posted to them and included a question asking them to confirm or not the presence of altered skin sensation or numbness around the scar. At the 3-month outpatient review the pain questionnaire and examination were repeated, as was the psychological HADS questionnaire. Postal contact was made with non-attendees. Data analysis was performed using the SPSS statistical package. Univariate analysis was performed using Chi square (χ2) and independent T-test as appropriate. Fifty patients were initially recruited. Seventeen were lost because of technical difficulties obtaining intra-operative readings, withdrawal of consent post-op and death. Intra-operative problems were complete neuro-muscular blockade, often despite the presence of two twitches from train-of-four stimuli, patient's temperature too low, and insufficient time to complete the measurements. Data from 33 patients were analysed for inclusion in the study. Their characteristics are shown in Table 1. The pre-op pain score 2 was significantly higher than zero (mean 4.2) because this included indigestion type pain and any discomfort perceived as being caused by the cancer, such as chest ‘tightness’ or pleurisy. Continuous data is presented as a mean and standard deviation and categorical data as a proportion of the study group. Of the 24 patients with a primary lung carcinoma 16 had lobectomy, five pneumonectomy, two sleeve lobectomy and one was exploratory. When three drains were used, two were sized 28-36FG but the third was always a soft 19FG drain. Patient characteristics and study findings Although some excellent recordings were achieved, many recordings were not interpretable. If we had repeated the measurements on the patients post-operatively as planned we would be unable to discern that any changes in the MEPs recorded would be due to nerve damage rather than irregularities in the method of measurement. This technique was therefore abandoned. There was no nerve damage detected after thoracotomy or thoracolaparotomy but prior to insertion of the retractor, except for a single nerve in one patient. Total intra-operative nerve damage is the proportion of all the nerves studied in all subjects that failed to conduct impulses along their length and cause intercostal muscle contraction. The number of nerves damaged after intercostal space closure for each individual shows a roughly normal distribution (Fig. 1). Conduction studies revealed nerve injury occurs either as a discrete block at the site of the retractor only, but conducts either side of this point, or fails along the whole length of the nerve. Analyses of these patterns of nerve injury were investigated by comparing these two groups, excluding those with more nerve injury after closure. Nerves damaged along their whole length (n = 5) are associated with significantly longer retractor times (mean 141 vs 95 min, p = 0.04) compared to nerves with discrete blocks (n = 10), though there is no association with pain scores, numbness or analgesia use. Analyses comparing those with further nerves damaged by intercostal space closure (n = 7) with those without such damage (n = 25) showed no significant association with pain scores, numbness or analgesia use. ‘Numbness’ represents any abnormal cutaneous sensation, not just reduced or absent sensation. Although the reported incidence of numbness at 6 weeks and 3 months is the same (18 out of 26 participants for both), the measured area of abnormal sensation in the 20 subjects who were assessed at both times generally shrinks (Fig. 2). Analgesia use and all pain scores fall with time, as did the mean HADS score. The number of injured nerves after intercostal space closure. Measured area of abnormal cutaneous sensation (‘numbness’) at 6 weeks and 3 months post-op (n = 20). For investigation into the factors that may predict intercostal nerve injury analyses were performed comparing those with zero to two nerves damaged (n = 23) with those with three to four nerves damaged (n = 10). Results are expressed as means for continuous data and proportions for categorical, with p-values (significance given as p < 0.05) (Table 2). This shows that nerve damage is not significantly associated with most factors except for technique of closure of the intercostal space: interrupted pericostal sutures are associated with 78% of the low nerve damage group but only 40% of the high nerve damage group (p = 0.03). Although thoracotomy is associated with more nerve damage, this is not significant. Predictive factors for intra-operative nerve damage Analyses were performed to try to identify factors associated with chronic pain, using pain scale 3 in the SF-MPQ as the defining factor for determining the presence of chronic pain (Table 3). This shows that the presence of chronic pain is not associated with most factors including the number of injured intercostal nerves. The only factors that significantly predict chronic pain are a higher pre-op pain score 1 (p = 0.05) and higher 6 week post-op pain scores (p = 0.008, 0.001 and 0.002 for scores 1, 2 and 3, respectively). Chronic pain is also significantly associated with higher post-op HADS scores (p = 0.008) though not pre-op scores. There was no association between pain and the number of drains, the size of the drains or the number of intercostal spaces through which drains were placed. Predictive factors and associations for chronic pain at 3 months post-op Factors that may be associated with the presence of abnormal cutaneous sensation at 3 months post-op were analysed (Table 4). This shows that most factors, including the number of nerves damaged at operation, are not significantly associated with ‘numbness’. However, the presence of a rib fracture is the only factor associated with an absence of numbness (p = 0.006) and numbness at 6 weeks post-op does predict its presence at 3 months (p < 0.001). Predictive factors for the presence of numbness at 3 months The aim of this study was to explore the role of intercostal nerve injury in the development of chronic post-thoracotomy pain. The impetus was the study by Rogers et al. [17] that demonstrated intercostal nerves are damaged during thoracic procedures, though an association with subsequent chronic pain was not reported. Our study involved only elective patients since these planned operations would allow sufficient time for the patient to decide whether or not to participate. Our use of a pre-op pain and psychological scores allowed us to analyse whether pre-existing chest symptoms or feelings of anxiety or depession influenced the development of chronic pain. It was disappointing that the magstim (magnetic stimulator) nerve conduction studies in this setting were not sufficiently accurate to be of practical use. The inconsistency in the readings was caused by difficulty both in locating the intercostal nerve root for stimulation and also in placement of recording electrodes for recording. Precise placement of button skin electrodes over the relevant anterior intercostal muscle was sometimes impossible because fat, breast tissue and pectoralis muscle made it difficult to identify the relevant intercostal space, which meant we were unsure whether readings were from the intercostal muscle we thought was producing the MEP or from other chest wall muscles. There was no visible twitching of individual intercostal muscles, which may have helped identify activity. Had this technique been successful it would have acted as an excellent baseline measurement prior to the influence of any anaesthetic agents, and which could have been repeated in the outpatient setting to verify and the intra-operative nerve studies. The intra-operative studies had to on from the to a There was on two twitches after train-of-four (TOF) stimulation of adductor pollicis to the of However, in some this did not to be some patients had a despite no and some to have complete thoracic blockade despite three or four This difficulty is in the high number of patients who we were unable to study. the of the intercostal nerves to conduct impulses was by of the ribs and intercostal muscles. Although we were of the function of the nerves the surgical we were not to the more proximal or distal of the nerve out of which damaged through injury. In this we may be the of nerve injury. Although some showed more nerves had been injured from the time the retractor was to the of intercostal space closure in the pericostal closure group and three in the intercostal muscle closure there were also some where the nerve conduction had over the same from each closure This may be due to the nerve injury or to the of the neuromuscular If some or all of the nerves with time, which is then the intra-operative studies the of nerve injury. of nerve such as due to may cause damage, other may not. as many as possible of the technique would be and some between In this study three and their were or more for of operation. although we have measured some of the operation that could affect intercostal nerve there are an number of The of two patterns of nerve injury is with the association with retractor This may two of on the nerve by the retractor a of trauma that is to on the nerve causes a injury the whole nerve, due to of the The of these patterns of injury are not since there are no significant associations with the we The fall in analgesia use and pain scores with time was as was the in area of abnormal cutaneous sensation as cutaneous nerves upon the damaged We had increased thoracotomy increased retractor spread and increased retractor time to all significantly increase the incidence of intercostal nerve though this was not by study. Although thoracotomy did to cause more nerve damage compared to thoracolaparotomy this was not significant. increased retractor time to reduce nerve which the authors The technique used by surgeon A along the of the rib and use of interrupted pericostal sutures to to in nerve injury at the time of operation and may a method was used nerve injury caused by closure did not affect any measured Although the pain scale 3 of the SF-MPQ was used to chronic pain at 3 the results of analyses using the other pain was The study findings did not that intercostal nerve injury would be associated with chronic pain. This may be because intra-operative measurements were not for intercostal nerve injury as or because some nerve only cause numbness rather than a or because there is a more significant cause for chronic pain other than intercostal nerve injury. The association of chronic pain with a significantly higher pre-op pain score 1 suggests that pre-existing pain experience may increase the risk of post-op pain. This may be an of psychological or because of a of the at a to and a known as as by The association of chronic pain with significantly higher post-op scores the between chronic pain and and is an high pain scores 6 associated with chronic pain previous study findings [4]. This may be it to of post-thoracotomy pain. The incidence of numbness at 3 months for both pain and no pain is and suggests that nerve damage intercostal or cutaneous detected this is not associated with pain of factors that may predict altered cutaneous sensation at 3 months a of association with intra-operative nerve which was The lower and the absence of rib fractures in the numbness group be explained by the This study did not to measure the of damage to the which may be a factor in pain due to or and to post-operative pain caused by intra-operative studies could investigate whether to reduce the patterns of intercostal nerve damage at operation in chronic pain. damage by the retractor on the nerve may be reduced by retractor to increase the and reduce the of contact with the damage due to injury may be reduced by of the retractor after a of time to allow the of through the under The cause of chronic pain after thoracic surgery has not been though some the role of intercostal nerve Our study two patterns of nerve injury and also demonstrated differences between surgical with rib resection in more nerve damage than diathermy along the of the rib and pericostal closure. However, intercostal nerve damage at the time of operation is not associated with chronic pain or altered cutaneous sensation at 3 months that either the of intra-operative nerve damage is not of nerve damage or that there is a more significant cause for chronic pain other than intercostal nerve injury. to thoracic from Magstim for the nerve conduction equipment and in

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.