Predicting Postoperative Pain by Preoperative Pressure Pain Assessment

Patients scheduled to undergo abdominal total hysterectomy or myomectomy with general anesthesia were eligible to participate in this study. The study was approved by our institutional review board (Mackay Memorial Hospital, Taipei, Taiwan), and signed informed consent was obtained from all subjects. Patients with a history of psychiatric disease, alcohol or drug abuse, chronic nonsteroidal antiinflammatory drugs use, chronic pain, or use of opioids preoperatively were excluded. Forty women, aged 20–55 yr and having an American Society of Anesthesiologists physical status of I or II, were enrolled.

A Chinese version of the State-Trait Anxiety Inventory (STAI) was used in the preoperative holding area to assess patients' anxiety before they entered the operating room. The STAI is a 40-item self-reported questionnaire that measures trait (20 items) and state anxiety (20 items). The latter is the transitory emotional state induced by a particular situation. For the purposes of this study, only the state anxiety portion of the STAI was used to assess preoperative anxiety. Raw scores (20–80) from the STAI were transformed to percentiles (0–100) for further analysis.28The anxiety level was classified using the highest quartile of percentile ranks,29with 1 = mildly anxious (≤ 75th percentile) and 2 = highly anxious (> 75th percentile). Stratification was based on Spielberger's manual.28 

An electronic pressure algometer (Somedic AB, Sollentuna, Sweden) was used to determine the pain threshold and pain tolerance pressure. A probe with a surface area of 1 cm²was applied to the pulp of the third finger of the right hand, and the pressure was increased at a speed of 30 kPa/s. Patients were asked to press a button on a patient-operated switch when they started to feel pain (pain threshold) and when they could no longer stand the pain (pain tolerance). The algometer recorded the pressure at each point. To familiarize the patients with the assessment method, a training session was given in the preoperative holding area. After patient was transferred to the operating room, the baseline pain threshold and tolerance pressures were recorded with the patient lying on the operating table with eyes closed. Fentanyl, 2 μg/kg, was then administered intravenously, and the pressure pain assessment was repeated 4 min after the injection. Fentanyl sensitivity was defined as the percent increase in the pressure pain tolerance after fentanyl administration.

Standard monitoring, including noninvasive blood pressure monitoring, electrocardiography, pulse oximetry, and capnography, was begun when patients entered the operating room. An auditory evoked potential monitor, A-Line® (Danmeter A/S, Odense, Denmark), was used to monitor the depth of anesthesia. No premedication was given because it might have interfered with the preoperative pain sensitivity testing. After completing the postfentanyl pain assessment, general anesthesia was induced with 0.5 mg/kg lidocaine, 2%, and 1.5–2 mg/kg propofol. Tracheal intubation was facilitated with 0.8 mg/kg rocuronium. After intubation, 8 mg dexamethasone and 4 mg ondansetron were given to prevent postoperative nausea and vomiting. During the operation, general anesthesia was maintained with 7–10% desflurane, 0.5 l/min air, and 0.5 l/min oxygen. No more fentanyl was given. The depth of anesthesia was maintained at an auditory evoked potential index value of 15–20 by titration of the inspired desflurane concentration in 1–2% increments. No local anesthetic was given for skin closure.

A handheld slide rule–type VAS with values from 0 to 100 was used to assess the immediate postoperative pain as soon as the patient arrived in the postanesthesia care unit (PACU). If the patient had difficulty manipulating the VAS slide rule, she was asked to state the degree of pain using a numeric rating scale from 0 to 100.30,31The pain assessment was repeated using the VAS 24 h postoperatively with the patient at rest in the ward. All subjects received morphine by intravenous patient-controlled analgesia (PCA) for postoperative pain control. The PCA pump (Abbott APM; Abbott Laboratories, Chicago, IL) was programmed to give a loading dose of 3 mg, a bolus dose of 1 mg, a lockout interval of 5 min, and a 4-h limit of 20 mg. The PCA pump recorded the time when the patient activated it. The number of analgesic requests (demand) made and the number of those requests that resulted in successful deliveries (delivery) were recorded. The demand/delivery ratio was used to measure the quality of analgesia.32This detailed record was downloaded from the PCA pump to a personal computer for further analysis.

Statistical analysis was performed with SPSS software (version 11.5; SPSS Inc., Chicago, IL). Data distribution was evaluated for normality by the Kolmogorov-Smirnov test. The relation between preoperative pain threshold and tolerance and the postoperative VAS pain score and morphine consumption was analyzed with the Pearson correlation test. The Bonferroni correction was used for multiple comparisons, and statistical significance was set at P < 0.016 when preoperative data were compared with postoperative data in the PACU and at 24 h in the ward. The Student t  test was used to compare postoperative pain and analgesic consumption in highly anxious and mild anxious patients. Stepwise multiple regression analysis was used to determine the independent variables age, body weight, duration of operation, STAI Anxiety score, pain sensitivity, and fentanyl sensitivity that were predictive for the dependent variables VAS pain and morphine consumption in the PACU and at 24 h in the ward. Values are reported as mean ± SD. P ≤ 0.05 was considered to be statistically significant.

The 40 consecutive patients enrolled in this study had a mean body weight of 55 ± 6.1 kg and mean age of 41 ± 6.2 yr. The average operation time was 122 ± 42 min, with an estimated blood loss of 405 ± 342 ml (table 1). Procedures performed included either abdominal total hysterectomy or myomectomy. The incisions were made on the low abdominal transverse line. All subjects had approximately the same size surgical wound in the same location.

The mean preoperative pressure pain threshold and pain tolerance were 140 ± 65 kPa and 223 ± 62 kPa, respectively. After fentanyl, the pressure pain threshold increased to 171 ± 67 kPa, and the pain tolerance increased to 272 ± 68 kPa.

The VAS pain score in the PACU before morphine administration was 81 ± 24. At 24 h postoperatively, it was 31 ± 10. The mean morphine consumption was 20.6 ± 6.0 mg (range, 9–34 mg) in the first 24 h postoperatively.

Twenty-eight of 40 patients (70%) rated their pain as 70 or greater on the VAS in the PACU after emergence from general anesthesia. There was no statistically significant relation between the preoperative pain threshold and tolerance levels and the immediate postoperative VAS pain score (P > 0.05). However, the preoperative pressure pain tolerance was significantly correlated with the VAS result 24 h postoperatively (P < 0.001, r =−0.52; fig. 1). Pressure pain tolerance after fentanyl administration was significantly correlated with the total PCA morphine consumption in the first 24 h after surgery (P < 0.002, r =−0.48; fig. 2).

The mean raw STAI score for state anxiety before surgery was 49 ± 11, equivalent to the 66th percentile ± 23. Of the 40 patients, 21 were classified as highly anxious, and the remaining 19 patients were classified as mildly anxious. The immediate postoperative mean VAS pain score of the highly anxious patients was 87 ± 25, significantly higher than that of the mildly anxious patients, 65 ± 28 (P < 0.05). The highly anxious patients also had a significantly higher demand/delivery ratio than the mildly anxious patients (8.8 ± 6.6 vs.  4.7 ± 2.6). However, at 24 h postoperatively, the mean VAS pain scores did not differ significantly between the two groups, nor was there a significant correlation between preoperative anxiety status on morphine consumption either immediately after surgery or at 24 h (P > 0.05; table 2).

Multiple linear regression analysis has shown that postoperative pain in the PACU can be estimated by using STAI percentile rank (β= 0.67, P < 0.0001) and postoperative pain at 24 h in the ward by pain tolerance (β=−0.09, P = 0.01). Postoperative morphine consumption at 24 h in the ward can be estimated by using pain tolerance after fentanyl (β=−0.49) and fentanyl sensitivity (β=−0.23, P = 0.00006; table 3).

This study demonstrates that preoperative pressure pain assessment may predict the level of postoperative pain. Our results show that preoperative pressure pain tolerance after fentanyl analgesia and fentanyl sensitivity may predict the amount of morphine consumption. We also found that patients with higher anxiety levels reported more pain in the immediate postoperative period.

Acute postoperative pain is most often caused by tissue and nerve damage, which may lead to prolonged central and peripheral hyperexcitability in the nociceptive pathways.33Human experimental pain models cannot mimic the clinical circumstance of extensive tissue damage or the emotional component of pain related to the threat of disease. However, pain models are useful in that they generate a painful stimulus under fully controlled and standardized conditions. This allows for semiobjective investigation of an admittedly subjective experience.

Several groups have used the pressure pain model to study pain mechanisms and the effect of analgesics.17–24However, ours is the first investigation of the relation between preoperative pressure pain levels and postoperative pain. Clinical application of the pressure pain model using algometers has been validated for evaluating pain sensitivity.25,34Although different nerve fibers transmit different type of painful stimuli, such as heat, cold, and pressure, measurement of pressure has been shown to be an adequate surrogate to evaluate the results of pain-relieving modalities such as anesthetic blocks, heat, manipulation, and antiinflammatory agents as well as documenting the long-term effectiveness of treatment.25Algometers also have the advantage over heat pain or cold pain models of being portable.

To explore the relation between pressure pain assessment and postoperative surgical pain, it is necessary to control for other factors that may influence postoperative pain, such as sex and type of surgery. That is why we designed our study to include only female patients undergoing lower abdominal gynecologic surgery. We excluded patients with cancer because they have been shown to have higher preoperative anxiety than patients with benign disease.35 

The VAS is the most commonly used tool in clinical research to assess pain in the perioperative period and to evaluate outcome.36In general, VAS scores of 70 or more are regarded as indicative of severe pain.30In our study, 28 of 40 patients (70%) rated their pain as 70 or greater on the VAS immediately after they emerged from general anesthesia using the short-acting anesthetic desflurane. Previous investigation has shown that the relation between increments in noxious stimuli and VAS pain scores is best described by an exponential function.37It is possible that immediate postoperative pain was so severe that it skewed the distribution of VAS scores in our subjects to the right, so that we could not find a correlation between preoperative pressure pain assessment and immediate postoperative VAS pain scores. However, the preoperative pressure pain tolerance was predictive of morphine consumption in the first 24 h after surgery. Our results are comparable to those of Granot et al. ,15who used a heat pain model. Our study also shows that pain tolerance is more reliable in assessing analgesic effects than is pain threshold. This is consistent with previous notions that experimental pain thresholds are not increased by opioids to the same extent as is pain tolerance. The thresholds are therefore of limited utility in defining pain sensitivity.38–40Although pain tolerance is sensitive in assessing analgesic effects, it is highly dependent on the motivation of the subject. Coghill and Eisenach41have suggested the use of a psychophysical rating of a fixed intensity stimulus to include in the study of pain sensitivity if devices are capable of delivering stimulus in a well-controlled fashion.

Several studies have examined the relation between preoperative state anxiety and postoperative pain, but their findings are contradictory. Some investigators have found that increased preoperative state anxiety is associated with increased postoperative pain, increased analgesic requirements, or both,42–44whereas others have found no such effect.45,46Our study supports the contention that preoperative state anxiety is associated with immediate postoperative pain level. The difference in the demand/delivery ratio suggests that highly anxious patients have different coping behaviors when they experience a stressful, painful situation. Although we found that preoperative state anxiety affected immediate postoperative pain, there was no correlation with VAS pain scores 24 h after surgery. This difference may relate to where and when both state anxiety and VAS pain scores were assessed. Subjects were isolated both when they were reporting their anxiety scale in the holding area and when they were recovering in the PACU. The 24-h postoperative pain assessment was performed after they had settled back into a room in the ward. It is somewhat surprising that we found no significant differences between highly and mildly anxious groups in terms of pressure pain threshold and tolerance. However, the anxiety questionnaire was given in a different location than the pain sensitivity testing, which may have influenced the results.

Patient-controlled analgesia generally provides adequate postoperative pain control with minimal side effects. The flexibility of PCA allows patients to titrate their own opioid dose. There can be considerable interpatient variability in PCA morphine doses. Macintyre and Jarvis47found up to 10-fold differences in PCA morphine consumption. Our patients' morphine use in the first 24 h ranged from 9 to 34 mg, approximately a 3-fold difference. The total dose consumed correlated with pressure pain tolerance and the reported level of postoperative pain. Although PCA is designed to be easily titrated by the patient, small changes in the dose available with each press of the button may result in inadequate analgesia. One implication of our study is that individual assessment of pressure pain tolerance might be useful in individualizing the PCA protocol, although this theory must be tested.

In conclusion, we have demonstrated that preoperative pressure pain assessment can be a predictor of postoperative pain and analgesic consumption in women undergoing lower abdominal gynecologic surgery. However, conclusions from this study should be drawn with caution because a significant part of variability cannot be explained by our model. Efforts must continue to be made to improve the pain prediction model, and whether the pressure pain model can be generalized to all surgical patients remains to be studied. We believe it is worth pursuing, because the assessment is simple to perform with an easily available, portable algometer.

The authors thank Hwe-Chu Yeh, R.N., M.S. (Research Assistant, Department of Anesthesiology, Mackay Memorial Hospital, Taipei, Taiwan), for hard work and help.