Peri-operative mental health and pain after surgery: cause, consequence or coincidence?

Mental health in the peri-operative period has been gaining attention in the past decade. With the emergence of safer surgical and anaesthetic techniques, the critical metrics of successful surgery are no longer solely focused on physical outcomes such as death, bleeding, surgical site infection and duration of hospital stay. Metrics from a more holistic perspective of quality of life are considered instead, including measures of mental health, pain, and cognitive and social functioning.

Attention is given to the cause vs. consequence discussion in the context of peri-operative mental health. One focus is around the mental health conditions of patients having surgery. Surgery is often a life-changing event, and multiple personal, social, environmental and cultural factors contribute to the mindset and mental reserve with which patients face and experience the surgical event. Numerous studies have shown that mental health conditions such as depression or anxiety before surgery may contribute to poorer outcomes such as delirium, pain, falls and increased readmission rates [1–5]. This is a lens through which multiple researchers have looked at surgical outcomes, while primarily focusing on recovery trajectory, early mobilisation and return to physical functioning.

The second focus is about mental health as the outcome of surgery. Fewer studies have carefully investigated these outcomes, but some have pointed out that depression and anxiety commonly affect surgical patients. Worsening depression, for example, is associated with poorer physical and social recovery outcomes after cardiac surgery [6]. If we think about a 76-year-old independently functioning adult who undergoes emergency cardiac or hip surgery that requires prolonged recovery outside their natural environment, it is logical that worsening depression, for example, would substantially contribute to social isolation, poorer function and major overall decline in quality of life. As such, understanding who might be at risk, and what contributes to poorer mental health trajectory after surgery, is of growing importance, considering the approximately 300 million surgical procedures that are performed globally every year [7].

Persistent, or chronic, pain after surgery is one of the most common long-term surgical complications. In certain surgeries such as thoracic, breast and spine procedures, more than 30% of patients report persistent pain ≥ 3 months after surgery [8–10]. Several studies have investigated how different risk factors are associated with the development of persistent pain after surgery, many of them pointing to psychological factors, with depression being one of them [10–12].

Lee et al. report the results of a systematic review and meta-analysis on the impact of pre-operative depression on postoperative pain outcomes [13]. The authors, by reviewing 60 studies, report that pre-operative depression among patients having surgery was most strongly associated with worse postoperative pain severity immediately after surgery (< 72 h), as well as in the later phases of recovery (> 6 months to 1 year after surgery). Pre-operative depression was also associated with more patient-controlled analgesia requests in the acute postoperative period, but not necessarily consistently related to number of postoperative opioid prescriptions. Interestingly, no significant association between pre-operative depression and pain was found in the intermediate time interval at 3–6 months after surgery, suggesting a U-shaped curve in the association between pre-operative depression and its effect on postoperative pain outcomes. Notably, the degree of change in pain scores from the pre-operative period to 1–2 years after surgery did not significantly differ between patients with and without depression, raising additional questions about the strength of the association, the temporal dynamics between the two phenomena and causality.

Despite the important findings emerging from this work, several questions remain unanswered. The authors dichotomised depression to a binary outcome, due to the heterogeneity of tools to measure peri-operative depression. While this may be operationally convenient to allow meta-analysis or future patient stratification, it is more likely that a severity spectrum exists for depression, with different implications on a patient’s psychological response to the surgical event. The non-uniform use of assessment tools for the presence and severity of depression impedes meaningful integration of research findings across the field.

Another, somewhat similar, challenge associated with accurately capturing subjective patient symptoms is the translation of pain outcomes to a single scale of pain severity, which is unlikely to capture a patient-centric view of how much pain may be interfering with their function and quality of life. In this context, the clinical impact of the standardised mean difference of 0.45 on a 0–10 scale between the groups with and without pre-operative depression, as reported by Lee et al., may be difficult to estimate.

Given these challenges in the accurate assessment of outcomes in therapeutic areas where no objective measures of disease severity and impact exist, one emerging approach is involving individuals with lived experience as partners in research, to improve the patient-centricity of the conducted research and relevance of study outcomes. An example of this is partnering with patients who have experienced depression and pain in the peri-operative period. This approach can help develop useful, inclusive and equitable metrics and help apply them universally in clinical studies, subsequently allowing meaningful aggregation and meta-analysis. Recommendations on how to incorporate these approaches in pain studies are emerging [14,15].

Outside of accurate capturing of meaningful outcomes, one particularly important challenge is the inability of a meta-analysis to account for confounding variables within each of the included studies. Depression may be associated with postoperative pain outcomes, but in some studies [16–18], when confounders such as age, sex, pre-operative pain, anxiety and pain catastrophising are accounted for, depression no longer emerges as an independent prediction factor, raising the question of causality vs. coincidence in the context of pain.

One way to think about causality is through the examination of peri-operative interventions for improving depression, and their effect on pain outcomes. In this context, Table 1 summarises results from nine systematic reviews on peri-operative behavioural interventions on pain outcomes published in the past five years or so. As the table shows, there is substantial heterogeneity in interventions and results, suggesting that in some settings some interventions may be helpful, but highlighting the need for more accurate approaches in identifying patients at risk, collecting robust outcomes and delivering precise interventions.

Table 1.

Summary of systematic reviews of peri-operative behavioural interventions on postoperative pain

Review	Surgical setting	Number of studies	Types of interventions	Effect on pain-related outcomes	Authors’ conclusion
Phang et al. [22]	Total knee arthroplasty	8 RCTs	CBT	Reduction in pain (n = 3)	Current evidence does not support the efficacy of CBT for pain, as literature too heterogenous
Scarone et al. [23]	Spinal fusion	13 (11 RCTs)	CBT (n = 9), mindfulness-based therapy (n = 2), relaxation and psychoeducation (n = 1)	Significant reduction in postoperative pain (n = 6) and in disability scores (n = 8)	CBT is likely to result in reduction in pain short-term (immediately after surgery to 3 months), with inconclusive evidence for long-term effects
Nadinda et al. [24]	Mixed surgeries	21 RCTs	CBT (n = 12), ACT (n = 3), psychoeducation, mindfulness-based stress reduction, expectation manipulation, pain management,  relaxation and guided imagery, stress management training (n = 1 each)	Significant reduction in (sub)acute pain and chronic postsurgical pain in the meta-analysis	Overall significant effect, with small effect sizes
Gorsky et al. [25]	Mixed surgeries	22 RCTs	Relaxation (n = 1), music and relaxation therapy (n = 5), music (n = 3), hypnosis (n = 4), psychoeducation (n = 3), CBT (n = 3)	Reduction in both postoperative opioid use, pain, or both (n = 13). No significant reduction in pain or opioid use (n = 7). Increase in postoperative pain or opioid use (n = 2)	Literature not robust to support a strong recommendation.  Certain interventions can reduce pain, possibly by reducing pre-operative anxiety
Janssen et al. [26]	Lumbar spine surgery	15 (12 RCTs)	CBT interventions	CBT interventions no more effective than usual care for all outcomes, including for back pain and leg pain	Very low to low certainty evidence of no additional effect of CBT interventions on pain
Villa et al. [27]	Major abdominal surgery	9 (8 RCTs) (only six with pain as outcome)	Relaxation therapy (n = 5), CBT (n = 1), coping strategies (n = 2), hypnosis (n = 1)	Significant improvement in postoperative pain (n = 5 out of 6)	Peri-operative psychological interventions feasible and can reduce pain and anxiety
Tong et al. [28]	Orthopaedic surgery	19 (some focused on pain outcomes, some on anxiety outcomes)	Relaxation, CBT, hypnosis, emotional counselling and mixed psychotherapies	Insufficient evidence to confirm reduction in post-operative pain (n = 7).	Psychological interventions may primarily reduce anxiety in the postoperative period
Whale et al. [29]	Total knee arthroplasty	12 RCTs	Music therapy (n = 5), CBT (n = 2), guided imagery, hypnosis, progressive muscle relaxation with biofeedback, pain coping skills, postoperative management program (n = 1 each)	Due to intervention heterogeneity and poor reporting, impossible to make definitive statements about effectiveness of psychology interventions for post- total knee arthroplasty pain outcomes	Future work is needed
Ziehm et al. [30]	Open heart surgery	23 RCTs	Psychoeducation (n = 12), CBT methods (n = 1), relaxation (n = 2), or a combination of (n = 15)	No reduction in pain intensity in the short-, medium-, or long-term interval	There is a lack of evidence to support or refute psychological interventions to reduce postoperative pain after open heart surgery

RCT, randomised controlled trial; CBT, cognitive behavioural therapy; ACT, acceptance and commitment therapy.

In addition, not all behavioural interventions target depression exclusively. There is a substantial overlap among neural circuits responsible for depressive symptoms vs. pain symptoms. As such, the favourable effectiveness of behavioural (as well as pharmacological) antidepressant approaches on various chronic pain syndromes may occur directly via modulation of pain neural circuits, making it challenging to draw conclusions on causality between depression and pain.

While most of the reported research on depression and postoperative pain has tested associations between a single measure of pre-operative depression and some measures of pain after surgery, there is a lack of studies closely measuring depression and anxiety trajectories postoperatively, as well as their association with pain. It is possible that individuals whose depressive symptoms linger after surgery tend to report more pain, while those with improved symptoms report less pain. It is difficult to ascertain the relationship without carefully monitoring both pain and depression levels. The prevailing use of traditional assessments for depression at the interval of 1–2 weeks, such as the Beck Depression Inventory (BDI) and Hospital Anxiety and Depression Scale (HADS), may be challenging to employ in the peri-operative period given the possibility of rapid symptom changes, multiple transitions of care and time required for recovery-related needs.

As a viable alternative, the popularity of smartphone-based ecological momentary assessment methods increases [19–21], and these can capture changes in pain and mood at higher frequency, along with other factors such as sleep and physical function. This is of particular importance in the acute and subacute postoperative period, as symptoms may fluctuate substantially, and measurements at one or two time-points often lead to a loss of valuable information and potential misrepresentation and over-simplification of a complex dynamic process.

The findings from Lee et al. [13], underscore the importance of screening for depression in patients having surgery, but uncertainty remains around the optimal timeframe in which the peri-operative screening should occur. When feasible, it would be helpful to identify patients who have depressive symptoms as early as possible before surgery. However, it is not yet clear that interventions that start pre-operatively are more effective than those started postoperatively, and this needs to be rigorously tested.

As evidence on peri-operative interventions targeting mental health emerges, future studies testing their impact on acute and chronic outcomes will help to establish the causality between depression and pain, and potentially improve pain-related outcomes after surgery. Developing robust prediction models that incorporate key mental health measures that determine the risk of postoperative pain can facilitate the stratified inclusion of high-risk patients to clinical trials that test peri-operative interventions. This approach can open the doors for a more precise and personalised delivery of mental health care to improve surgical outcomes.