OUP user menu

The Association Between Preoperative Pain Catastrophizing and Postoperative Pain Intensity in Cardiac Surgery Patients

Reenam S. Khan BSc, Petros Skapinakis MD, MPH, PhD, Kamran Ahmed MBBS, MRCS, Demetrios C. Stefanou MD, Hutan Ashrafian MBBS, MRCS, Ara Darzi FRCS, FACS, KBE, Thanos Athanasiou PhD, FRCS, FETCS
DOI: http://dx.doi.org/10.1111/j.1526-4637.2012.01386.x 820-827 First published online: 1 June 2012


Objective. Postoperative pain is associated with delayed discharged and recovery, reduced patient satisfaction, and increased costs. The aim of this study was to investigate the short-term association between preoperative psychological variables (pain catastrophizing, anxiety, and depression) and postoperative pain in a sample of cardiac surgery patients.

Design. This is a prospective epidemiological study.

Setting. This study was carried out at two Imperial College Healthcare National Health Service Trust Hospitals (St. Mary's Hospital and Hammersmith Hospital, London, UK).

Subjects. Sixty-four cardiac surgery patients completed the “pain catastrophizing scale (PCS),” the “hospital anxiety and depression scale,” and the “verbal rating scale” (VRS) for pain intensity preoperatively and at 48 hours postoperatively. Analgesia consumption was recorded. Data on demographic, operative, and clinical characteristics were obtained from medical records.

Outcome Measures. Pain intensity at 48 hours postoperatively.

Results. Scores on the anxiety, depression, and PCSs were not significantly different between the pre- and postoperative period. In contrast, patients reported a higher level of pain intensity postoperatively (P < 0.001). In the fully adjusted multiple regression analysis, postoperative pain intensity was predicted by a higher level of preoperative pain intensity (dichotomized above median; β = 2.00, 95% confidence interval [CI]: 0.28–3.72) and a higher score on the preoperative PCS (dichotomized above median; β = 1.87, 95% CI: 0.53–3.21).

Conclusions. Pain catastrophizing can predict postoperative pain intensity in cardiac surgery patients, independently of the presence of anxiety, depression, or preoperative level of pain. Future studies should aim to establish the role of pain catastrophizing in longer-term outcomes in cardiac surgery.

  • Catastrophizing
  • Pain
  • Cardiac Surgery
  • Anxiety
  • Depression


Acute postsurgical pain is associated with slower recovery, wound infection, and cardiovascular and respiratory complications [1,2]. It reduces patient satisfaction and increases health care costs [3–5]. Acute pain predisposes to the development of chronic pain [6]. Chronic pain can reduce quality of life and result in physical disability [7]. Studies suggest that pain is often inadequately managed [6]. Even for the same surgical procedure, where the surgical stimulus is presumably uniform, there is a significant variability in reported pain and the quantity of opioid analgesia needed for pain relief [8,9]. This supports evidence that suggests a role for demographic, genetic, personality, and psychological factors in pain perception [10]. Preoperative modification of risk factors for severe pain has the potential to alleviate postsurgical pain and prevent complications.

Previous literature suggests that psychological factors are important in pain perception [11,12]. Traditionally, anxiety and depression have been the main variables of interest. Some studies have shown that a higher level of preoperative anxiety is associated with heightened postoperative pain [13,14], while others reported an association between state but not trait anxiety with postoperative pain [15,16]. Other studies, however, have not confirmed this association [17]. Depression has also been associated with heightened postsurgical pain severity and analgesia consumption in some studies [18], but in others no effect was found after controlling for anxiety [19].

It has recently emerged that another psychological factor, “pain catastrophizing,” may be a better predictor of pain than anxiety and depression [20]. Catastrophizing is a type of negative and distorted thinking (“maladaptive cognition”) that can be defined as “an exaggerated negative mental set brought to bear during an actual or anticipated painful experience.” According to a previous research, catastrophizing has three correlated but independent dimensions: magnification (e.g., “I'm afraid that something serious will happen”), rumination (e.g., “I can't stop thinking about how much it hurts”), and helplessness (e.g., “There's nothing I can do to reduce the intensity of the pain”) [20]. Catastrophizing has been associated with psychiatric and personality variables including fear of pain, coping strategies, anxiety, and depression [20–23].

A relationship between catastrophizing and pain intensity has been shown in several studies, mostly for chronic pain (pain lasting for >6 months) [24,25]. A few studies have considered catastrophizing in the context of acute postsurgical pain. Studies in patients undergoing breast surgery, knee surgery, elective abdominal surgery, ear–nose–throat (ENT) surgery, and cesarean sections have shown that high-preoperative catastrophizing scores tend to predict heightened postsurgical pain [9,10,21,26–28]. This has been attributed to the finding that higher catastrophizing scores are associated with greater attention to pain and awareness of bodily sensations [29–31]. Most of these studies, however, have not adjusted for anxiety or depression. For catastrophizing to be conceptually useful, it should be shown to contribute to the pain experience independently of anxiety and depression [20].

Pain has been highlighted as a primary concern of patients undergoing major surgical procedures [32]. No studies have established the role of catastrophizing as a predictor of pain severity in patients undergoing major abdominal, thoracic, and cardiac procedures. These procedures are associated with relatively more pain than those included so far in studies of catastrophizing [33,34]. The present study aims to investigate the hypothesis that pain catastrophizing (measured preoperatively) will be associated with postoperative pain independently of anxiety and depressive symptoms in patients undergoing cardiac surgery involving median sternotomy.


Study Design and Population

This prospective study was carried out within the Imperial College National Health Service Hospitals (St. Mary's Hospital and Hammersmith Hospital, London, UK). Patients eligible for participation were those scheduled for open heart surgery involving median sternotomy from August 2009 to June 2010. Patients were approached by a member of the clinical team. Patients were excluded if they did not give consent to participate or could not complete the questionnaire either due to cognitive impairment or language barriers. We also excluded patients with prolonged intensive care unit (ITU) stay, alcoholism, and drug abuse. The study was granted ethical approval by the Imperial College Research Ethics Committee and informed consent was obtained from all participating subjects (Ethics number: 09/H0712/10). Patients were informed of their right to withdraw at the outset.


Validated questionnaires were administered to measure psychological variables (catastrophizing, anxiety, and depression) and pain intensity. Data on demographic, clinical, and operative characteristics were obtained from patients' notes.

Pain-Related Catastrophic Thinking

Levels of catastrophic thinking were assessed using the pain catastrophizing scale (PCS) [20]. This is a 13-item questionnaire relating to the type of thoughts or cognitions when the subjects are in pain. Ratings are made on a 5-point Likert-type scale with the endpoints 0 (“not at all”) and 4 (“all the time”). The scores range from 0 (no catastrophizing) to 52 (severe catastrophizing), with no clear “cutoff” distinguishing high and low catastrophizing. The PCS yields a total score as well as three subscale scores assessing the three dimensions of pain catastrophizing, namely rumination, magnification, and helplessness.

Anxiety and Depression

Symptoms of anxiety (A) and depression (D) were assessed by the hospital anxiety and depression scale (HADS) [35]. This is a 14-item brief self-assessment scale to evaluate anxiety and depressive symptoms. Each item is rated on a 4-point scale from 0 (absence of symptom or presence of positive feature) to 3 (presence of symptom or absence of positive feature). The possible score for each subscale (A or D subscales) ranges from 0 to 21. A score of 0–7 is considered normal, 8–10 indicates a subthreshold case, and a score ≥11 indicates a “clinical” case. The total score of each subscale can also be used as a continuous variable in analyses.

Pain Intensity

Patients provided pain intensity ratings on a 10-point (1 = no pain, 1–10 = unbearable pain) verbal rating scale (VRS). Patients were asked to rate pain both pre- and postoperatively. Postoperative pain intensity was assessed at 48 hours after the operation. An “overall pain intensity” score was obtained using an average of the reported pain intensity at rest and on movement. The numerical VRS has construct validity and statistical power that is similar to the visual analog scale, but it is easier to administer and score than the visual analog scale [36]. Patients were also given the VRS preoperatively on the day before surgery (see the Procedure section).

Analgesia Consumption

Regarding the management of postoperative pain, all patients received a loading dose of morphine 1 mg/mL (a total of 10 mg morphine in 10 mL saline) and were prescribed regular paracetamol (1 g p.o. every 6 hours) and aspirin (75 mg p.o. daily). They received standard care in terms of subsequent dose adjustments or additional pain medication (“rescue analgesia”) if deemed necessary. These medications were morphine and tramadol. Morphine was given intravenously via patient-controlled pumps (PCA), while tramadol was given orally. The PCA pumps were programmed to deliver a 1 mg dose of morphine as a bolus, with a subsequent 5-minute lockout and a limit of 10 doses per hour (10 mg).

Information was recorded from the drug charts relating to analgesia usage every 6 hours for the first 48 hours postoperatively. In the analysis, we adjusted for the total amount of morphine or tramadol used at 36 hours postoperatively, i.e., up until the 12-hour period prior to the assessment of pain intensity. It should be noted that we did not adjust for preoperative analgesia use as only one patient received a single dose of paracetamol 1 g in that period.

Demographic and Medical Information

Demographic (age, gender, ethnicity, height, weight, and occupation) and medical information (baseline clinical information, perioperative data, and postoperative complications) were obtained from the patients' medical records. The European system for cardiac operative risk evaluation (EuroSCORE) is a method of calculating predicted operative mortality for patients undergoing cardiac surgery [37]. The logistic EuroSCORE for each patient was calculated [38]. Some of the data (relating to demographic variables, cardiovascular risk factors, comorbidities, and postoperative complications) required were also available from an institutional database, which was used to verify the recorded information. The comorbidities for each patient were coded using the World Health Organization International Classification of Diseases, Clinical Modification (ICD-10) system [39].


On the day before surgery, patients were asked to complete the pain catastrophizing scale (PCS), the HADS, and the pain VRS. All patients received general anesthesia. Anesthesia was induced using propofol 1–2 mg/kg, pancuronium 0.1 mg/kg, and fentanyl 8–15 µg/kg, and maintained by air and oxygen and propofol 2–3 mg/kg/h. Normothermia was maintained with warm i.v. fluids, a heating mattress, a humidified airway, and a warm operating theater. Pain intensity was again assessed at 48 hours postoperatively using the pain VRS. Six-hourly analgesia consumption postoperatively was recorded. Patients were asked to fill in the questionnaires (PCS and HADS) again at 48 hours postoperatively. Postdischarge, the medical records for all of the patients were obtained. The information gathered from the medical records was compared with that from an institutional database to validate the results.

Statistical Analysis

All statistical analyses were performed using Stata (version X, Stata Corp, College Station, TX, USA). All data were evaluated for normal distribution using the Shaprio–Wilks goodness-of-fit test. Continuous variables are presented as mean with standard deviation (SD) in brackets and categorical variables as numbers and percentages.

Scores on the PCS were dichotomized using the median value as the cutoff (40). We also carried out analyses using the continuous scores to check the consistency of our results. Demographic data, clinical data, and questionnaire scores for each group were compared to identify possible confounding variables. In our analysis of the HADS scores, we used the continuous scores of the two subscales. Differences between pre- and postoperative values of the psychological variables of interest and pain intensity were evaluated using the Wilcoxon signed-rank sum test. Spearman's correlation coefficients were calculated to assess the relationships between anxiety, depression, and catastrophizing. All hypothesis testing was two-tailed and P values of less than 0.05 were considered to indicate statistical significance.

To investigate the association between preoperative pain catastrophizing and postoperative pain intensity, we carried out a series of linear regression models using the scores on the postoperative pain intensity scale as the dependent variable. We first ran a simple regression model of postoperative pain intensity, with age and gender as the independent variables (model 1). Then, we entered into the model ratings of preoperative pain intensity, preoperative anxiety, and preoperative depression (model 2). Finally, in model 3, we additionally entered the preoperative pain catastrophizing scores (as a binary variable using the median as the cutoff). Models 2 and 3 were also adjusted for total “rescue” analgesia used at 36 hours postoperatively, i.e., 12 hours prior to the assessment of pain intensity.


Baseline Characteristics of the Patients

Of 80 patients scheduled for cardiac surgery involving median sternotomy, 10 were excluded because they were unable to comprehend the questionnaire due to linguistic difficulties. Five patients were excluded because they did not give written consent to participate. Sixty-five patients completed the preoperative questionnaire. One was excluded postoperatively due to a prolonged postoperative ITU stay. Postoperative questionnaires were obtained for the remaining 64 patients. The medical records of all 64 patients were obtained. However, the postoperative drug charts were unavailable in five cases. This meant that the second part of the analysis (multiple regression) was performed using 59 sets of data.

In all patients, a midline sternotomy was performed and the sternum was closed using sternal wires. There were 54 men and 10 women. The average age of the participants was 65.73 (SD: 11.13) years. Table 1 summarizes the baseline demographic, clinical, and operation data obtained from the medical records. Table 2 summarizes the mean pre- and postoperative scores on the PCS, HADS-Anxiety, HADS-Depression, and VRS for pain-intensity scales. It can be seen that pre- and postoperative differences for anxiety, depression, and pain catastrophizing were not significant (although there was a trend for a higher level of depression postoperatively, P = 0.10). In contrast, pain intensity was significantly higher postoperatively (P < 0.001).

View this table:
Table 1

Baseline demographic and clinical characteristics of patients (N = 64)

Preoperative Variables
Demographic VariablesMean (SD) or N (%)
Age at operation65.73 (11.13)
Female gender10 (16%)
Caucasian ethnicity47 (74%)
Married42 (66%)
Clinical characteristics
  Current smoker9 (14%)
  BMI (kg/m2), mean27.62 (4.52)
  Logistic EuroSCORE4.26 (3.89)
  Diabetes mellitus18 (28%)
  NYHA dyspnea class1.84 (0.792)
  CCS angina class1.76 (0.657)
  Comorbidities5.34 (2.57)
Operation Characteristics
Operation Characteristics
Elective operation42 (66%)
Type of procedure
  CABG only41 (64%)
  CABG + valve procedure15 (23%)
  Valve procedure only8 (13%)
On-pump47 (73%)
  • BMI = body mass index; CABG = coronary artery bypass grafting; CCS = Canadian Cardiovascular Society; HADS = hospital anxiety and depression scale (A) anxiety (D) depression; NYHA = New York Heart Association; SD = standard deviation; VRS = verbal rating scale (pain).

View this table:
Table 2

Pre- and postoperative mean scores (SD) and median scores (range) for psychological variables and pain intensity

Anxiety/Depression ScoresPreoperative Mean (SD), Median (Range)Postoperative Mean (SD), Median (Range)
HADS anxiety score4.72 (4.52), 4 (0–21)5.64 (5.05), 3.5 (0–17)
z = 1.29, P = 0.20*
HADS depression score3.50 (3.54), 3 (0–17)4.47 (4.85), 3 (0–21)
z = 1.66, P = 0.10*
Pain catastrophizing score
  PCS score11.69 (11.1), 9 (0–50)10.84 (11.37), 6 (0–44)
z = 1.49, P = 0.14*
Pain intensity score
  VRS score1.70 (1.72), 1 (1–9)4.26 (2.84), 4 (1–10)
z = 5.77, P < 0.001*
  • * Wilcoxon signed-rank sum test for the difference between pre- and postoperative values.

  • HADS = hospital anxiety and depression scale; PCS = pain catastrophizing scale; SD = standard deviation; VRS = verbal rating scale for pain.

Correlation Between Preoperative Pain and Psychological Variables and Postoperative Pain Intensity

The correlation coefficients for the relationship between preoperative pain intensity, preoperative psychological variables (PCS score and HADS scores), and postoperative pain intensity (VRS score) are shown in Table 3. The pain intensity reported preoperatively positively correlated with pain intensity postoperatively (r = 0.33, P = 0.008). Preoperative total PCS scores were significantly positively correlated with scores on the pain VRS scale assessed postoperatively (r = 0.41, P < 0.001). Anxiety and depression were not significantly correlated with postoperative pain intensity, although there was a positive trend for anxiety (r = 0.20, P = 0.12). More interestingly, however, there was a significant positive correlation between preoperative anxiety and depression and PCS scores (r = 0.44, P < 0.001 and r = 0.47, P < 0.001, respectively), indicating the potential for confounding in the association of PCS with pain intensity.

View this table:
Table 3

Correlation (Spearman's rho) between preoperative psychological and pain variables (pain catastrophizing scale, anxiety, depression, preoperative pain intensity) and postoperative pain intensity

Spearman's Rho (P Value)
Postoperative PainPCSHADS AnxietyHADS-DepressionPreoperative Pain
Postoperative pain1.00
PCS0.41 (P < 0.001)1.00
HADS anxiety0.20 (P = 0.12)0.44 (P < 0.001)1.00
HADS depression0.16 (P = 0.23)0.47 (P < 0.001)0.59 (P < 0.001)1.00
Preoperative pain0.33 (P = 0.008)0.20 (P = 0.12)0.43 (P < 0.01)0.20 (P = 0.12)1.00
  • HADS = hospital anxiety and depression scale; PCS = pain catastrophizing scale.

Multiple Regression Analysis

Summarized in Table 4 are the results of the multiple regression analysis. In the fully adjusted model, two variables were significantly associated with postoperative pain intensity: preoperative pain intensity and pain catastrophizing. To better illustrate the association in the models, we have dichotomized these two variables using the median as the cutoff (but similar results were obtained with the continuous scores). For pain catastrophizing (PCS), β was 1.87 (95% confidence interval: 0.53–3.21), i.e., patients with a PCS score above the median were on average having 1.87 more score on the postoperative VRS of pain intensity. Figure 1 shows the levels of pain according to different levels of catastrophizing (grouped in four using the 25th, 50th, and 75th centiles). This shows how the patients showing the higher levels of catastrophizing also tended to report greater postoperative pain intensity.

View this table:
Table 4

Multiple regression analyses predicting postoperative pain intensity

Pain Intensity (VRS) at Day 3 Postoperatively
Independent VariablesModel 1* β (95% CI)Model 2 β (95% CI)Model 3 β (95% CI)
Age−0.13 (−0.08, 0.05)−0.013 (−0.07, 0.04)−0.002 (−0.05, 0.05)
Gender0.01 (−1.96, 1.96)0.27 (−1.44, 1.99)0.70 (−0.94, 2.34)
Preoperative pain intensity1.85 (0.10, 3.59), P = 0.041.69 (−0.13, 3.51), P = 0.072.00 (0.28, 3.72), P = 0.02
HADS (A)na0.06 (−0.15, 0.27)0.008 (−0.19, 0.21)
HADS (D)na0.02 (−0.23, 0.28)−0.04 (−0.29, 0.20)
Preoperative PCSnana1.87 (0.53, 3.21), P = 0.007
  • Note: All β's are standardized regression coefficients.

  • Specification of models: * Model 1: independent variables included age, gender, weight, and preoperative pain intensity (dichotomized above median).

  • Model 2: independent variables additionally included scores on HADS anxiety and depression.

  • Model 3: Model 2 + preoperative pain catastrophizing scale (dichotomized above median).

  • Models 2 & 3 have also been adjusted for total tramadol or morphine 12 hours prior to the assessment of pain intensity.

  • CI = confidence interval; HADS = hospital anxiety (A) and depression (D) scale; na = not applicable; PCS = pain catastrophizing scale; VRS = verbal rating scale.

Figure 1

Box plot showing differences in postoperative pain intensity scores for different levels of the pain catastrophizing scale (PCS).


This study investigates the association between pain catastrophizing, anxiety and depression, and pain intensity in the context of major cardiac procedures. Our results show that pain catastrophizing measured preoperatively can predict postoperative pain intensity independently of the presence of anxiety and/or depressive symptoms. The latter was not associated with postoperative pain and, therefore, pain catastrophizing was the only psychological-type variable that predicted the intensity of pain. Catastrophizing, however, has a complex relationship with anxiety and depressive symptoms as all psychological variables were strongly correlated with each other preoperatively. The predictive value of pain catastrophizing for pain intensity was independent of and equally important with the intensity of preoperative pain.

Our study adds to a small but growing body of literature that pain catastrophizing is associated with increased postoperative pain intensity [21,26–28,40]. The present research extends previous findings in showing that pain catastrophizing predicts postoperative pain in patients undergoing cardiac surgery and that this effect is independent of other potentially confounding variables. Both psychological and neurophysiological mechanisms may explain this relationship. According to cognitive behavioral models of pain, negative appraisals (such as catastrophizing) adversely affect perceived pain intensity [24]. Furthermore, some authors have proposed that catastrophizing interrupts the descending pain inhibition signals to the spinal cord, where it favors neuroplastic changes in response to painful stimuli, causing pain sensitization [41]. This is supported by the finding that higher levels of catastrophizing have been associated with a weaker response to opioid analgesia [27,42].

Levels of anxiety and depression, as measured by the HADS, did not predict postoperative pain intensity. This supports findings by Riddle et al., who found that catastrophizing, but not anxiety or depression, predicted poor outcome following knee arthroplasty [43]. Papaioannou et al. found that preoperative scores on the PCS, HADS-Anxiety, and HADS-Depression scale all correlated with pain severity following elective lumbar fusion surgery in a sample of Greek patients. However, in the fully adjusted model, catastrophizing remained the only significant predictor of postoperative pain [40]. Our findings are inconsistent with some previous research, which suggests that depression is associated with increased pain intensity or severity. However, the vast majority of these studies have been in patients with chronic pain and undergoing minor/intermediate procedure, which may explain the discrepancies [44,45]. A significant relation between preoperative anxiety and postoperative pain has been reported in obstetrics and gynecology surgery [28]. Differences in the measurement of psychological distress may partly explain these discrepancies among studies, but it is also likely that these differences are specific to the type of operation.

In our study, anxiety and depression were correlated with catastrophizing preoperatively, which is in concordance with previous studies [20,23]. A suggested link between anxiety and catastrophizing is that catastrophizing is used by individuals as a coping strategy, and anxiety ensues when coping mechanisms fail. However, the strength of the relationships between catastrophizing and anxiety, and catastrophizing and depression, were only in the moderate range. This is consistent with previous results and adds evidence that catastrophizing is distinct from anxiety and depression [20,23]. In our adjusted analysis, two variables only, preoperative pain intensity and pain catastrophizing, emerged as independent predictors of postoperative pain intensity. This is in concordance with research that pain catastrophizing is the most robust of the three variables (anxiety, depression, and catastrophizing) as a psychological predictor of postoperative pain [20,40]. Therefore, intervention targeting catastrophizing is likely to have a greater positive impact on postsurgical pain outcomes than those addressing other psychological variables [43].

The results of the present study may have some important clinical implications. First, the findings suggest that using validated measuring instruments, such as the PCS, it would be possible to identify patients who are likely to experience more severe postoperative pain and related complications. Second, the possibility of designing interventions with the aim to reduce catastrophizing in acute pain settings should be explored further. So far, such interventions, based on cognitive–behavioral techniques, exist mainly for chronic pain [43]. It would be important to test minimal or brief interventions in surgical settings with the aim to reduce the levels of catastrophic thinking. A reduction in the levels of postoperative cardiac pain could result in a reduced risk of complications, promoting a faster recovery and discharge. A reduction in acute pain may result in a decreased incidence of hospital stay, chronic poststernotomy pain, allowing patients to enjoy a better quality of life.


There were certain limitations to this study and the results need to be interpreted in light of these. These findings are based on a single geographical area with predominantly male population within a restricted geographical area. Whether these findings can be generalized to other populations remains unknown. In addition, the small sample size of women did not allow us to examine gender differences. This study assessed catastrophizing in the context of cardiac surgery; it is therefore not appropriate to extend these results for other types of major surgeries, such as thoracic surgery. Furthermore, in our study, due to the relatively small sample size, we used the PCS as a whole and we did not attempt to investigate differences between the three dimensions because the statistical power of the study would be compromised.


This study was the first to assess pain catastrophizing in the context of cardiac surgery and the results show a positive association with postoperative pain independently of anxiety and depression. Several key questions remain, however, to be answered in relation to catastrophizing in the context of surgery. The most important is to formally assess whether preoperative psychological treatment for catastrophizing is feasible and if so, whether this translates to improvements in pain outcomes. Before such programs are implemented within the cost-constrained health care system, the usefulness of routinely assessing pain catastrophizing preoperatively must also be determined.


  • Conflicts of interest: There are no conflicts of interest related to the article on the part of the authors.


View Abstract