Persistent pain six months after breast cancer surgery: a multicentre follow-up study

Abstract

Persistent postoperative pain, defined as surgical site pain lasting beyond 3 months with other causes of pain excluded, can have adverse physical and psychological consequences [1]. We developed a multivariable model for persistent pain at 4 months after breast cancer surgery based on baseline and peri-operative factors [2]. A significant number of women who develop persistent pain 4 months after surgery may continue to have persistent pain at 6 months. Hence, we further evaluated risk factors for persistent pain at 6 months, encompassing patients' characteristics and information available from the 4-month follow-up.

This study is a secondary analysis of a prospective study investigating factors for persistent pain after breast cancer surgery at 4 months at two specialist centres (KK Women's and Children's Hospital, Singapore and Duke University Medical Centre, Durham, NC, USA), from January 2018 to March 2021 [2]. After ethical approval, written informed consent was obtained from English-speaking patients scheduled for elective breast cancer surgery. Patients with a history of intravenous drug abuse, chronic opioid use, receiving chronic corticosteroid therapy or who were pregnant were not studied. Pre-operative baseline information was collected and is available in online Supporting Information Tables S1 and S2. Patients were further assessed for postoperative pain and analgesic use in the post-anaesthesia care unit and at 24–72 h post-surgery. Patients were followed-up via telephone or online survey at 4 and 6 months using a validated pain questionnaire [3].

The primary outcome of persistent pain at 6 months was defined as the presence of one of either a pain score ≥ 3 on a numerical pain rating scale of 0 (no pain) to 10 (worst pain imaginable); or pain affecting daily life activities via indication of ‘yes’ on any of the seven questions that evaluated the impact of pain at 6 months on daily life activities in the Brief Pain Inventory Short-Form [4].

The main objective was to develop a multivariable model for factors associated with persistent pain 6 months after breast cancer surgery. We also identified factors associated with persistent pain at both 4 and 6 months. The multivariable model was finalised using a stepwise variable selection method, incorporating clinically relevant variables with a p value < 0.20 from the univariable logistic regression analyses.

Among the 233 recruited patients, 209 completed the follow-up at 6 months with an incidence of persistent pain of 57.4% (95%CI 50.4–64.2, n = 120). By selecting factors with p < 0.20 in the univariable analysis, we identified the optimal multivariable model for the primary outcome, which contained the following independent risk factors: pain at 4 months after surgery; history of hypertension; axillary surgery; lower pain pressure threshold; and greater pre-operative perceived stress (AUC (95%CI) 0.81 (0.76–0.87)) (Fig. 1a). Among those patients who responded to the 6-month follow-up, 96 (45.9%) reported having persistent pain at 4 and 6 months. The multivariable model for pain at 4 and 6 months comprised pre-operative (perceived stress score, EQ-5D-3L dimension on pain or discomfort, and pain score at other sites) and postoperative (pain score at 72 h with movement) factors (AUC (95%CI) 0.78 (0.71–0.84)) (Fig. 1b).

Figure 1

Open in figure viewerPowerPoint

Multivariable logistic regression model for persistent pain after breast cancer surgery at (a) 6 months after surgery and (b) at both 4 and 6 months after surgery.

Pre-operative perceived stress was the only common independent risk factor in the models of persistent pain at 4 [2] and 6 months after surgery. This may be indicative of altered interaction of stress and nociceptive stimuli, resulting in aberrant stress response mechanisms and central pain processing [5]. Given the use of cognitive-behavioural interventions for reducing perceived stress in patients with breast cancer, future studies should explore their potential in improving peri-operative stress [6]. With a high AUC of 0.81 showing strong predictive capability, our study is novel in using two distinct post-surgical time-points (pain at 6 months, and at both 4 and 6 months) for persistent pain, providing information that could be potentially helpful to avert pain progression through timely intervention. The factors influencing pain at 4 months are related primarily to immediate medical conditions in the initial healing phase (e.g. age, pre-operative pain); while at 6 months, a shift is observed with more indicators of chronic conditions (e.g. previous surgery, persistent pain at 4 months, pressure pain threshold). This suggests a solidification of pain behaviour into a chronic pain pattern, possibly caused by the aberrant whole-brain global reorganisation and homeostasis [7]. Alternatively, the different risk factors could reflect the fragility of models in those settings. The present study has several limitations, including a small sample size; no follow-up beyond 6 months; and the lack of inclusion of additional psychological and contextual factors (e.g. fear-avoidance, lifestyle and social support) in the model. Future studies should validate the model and develop a sparse point-based scoring system to identify at-risk patients for timely monitoring and pre-emptive interventions.