Impact on postoperative pain of long-lasting pre-emptive epidural analgesia before total hip replacement: a prospective, randomised, double-blind study


Lorenzo Quinzio


Clinical studies on pre-emptive analgesia have produced inconsistent results. We conducted a clinical study investigating the effect of long-lasting pre-emptive epidural analgesia on consumption of analgesics and acute pain. Forty-two patients scheduled for elective hip replacement for osteo-arthritis were randomly assigned to receive, on the day before operation, either 5 ml.h−1 ropivacaine 0.2% (study group, n = 21) or 5 ml.h−1 saline (control group, n = 21). Postoperative analgesia was achieved in both groups by patient-controlled epidural analgesia (PCEA) with ropivacaine 0.2%. The main outcome measure was consumption of local anaesthetics. Additional parameters included visual analogue pain scale (VAS) scores, consumption of rescue analgesics, requests for PCEA boluses, and side-effects. The pre-operative parameters and pain scores were similar in the two groups. Epidural blocks provided sufficient operative analgesia in all patients. Pre-emptive analgesia was continued for 11–20 h and led to significantly decreased pain scores before surgery. The consumption of local anaesthetics was decreased postoperatively in the study group (194 mg vs. 284 mg in the postoperative period). Furthermore, bolus requests occurred more frequently in the control group. VAS scores did not differ significantly between groups. Long-lasting ‘pre-emptive’ epidural analgesia decreases postoperative pain with improved pain control.

Chronic pain is considered one of the most catastrophic outcomes of surgery. One of the most relevant predictors for this complication is postoperative pain of high intensity [1]. Accordingly, attenuation of postoperative pain is not only an ethical issue, but may also reduce the number of patients suffering from longer-lasting postoperative pain.

The beneficial effects of pre-emptive analgesia upon postoperative pain ratings have been shown in numerous experimental studies. However, in clinical settings, results have not been convincing in the long run [2]. This may in part be due to lack of consensus regarding the terminology of pre-emptive analgesia. In several studies the term is used to describe application of an analgesic medication before pain onset. However, as spinal hyperexcitability is thought to play a major role in the pathophysiology of this entity, incomplete afferent blockade with non-steroidal anti-inflammatory drugs (NSAIDs) may not be sufficient to achieve this goal. Other possible explanations for inconsistent results of clinical studies include the timing of the pre-emptive medication. In most studies, analgesics or narcotics have been applied just before start of surgery or at the end of surgery [2]. Furthermore, the effectiveness of pre-emptive analgesia may vary due to the type of surgery. In abdominal surgery – in contrast to limb surgery – a possible effect could not be observed [3].

A small number of published studies have concentrated on the prevention of postoperative pain in patients with a high risk of development of chronic pain (e.g. limb amputation) [4,5]. In these studies, patients did not suffer from significant pain before surgery. The situation may be different in patients who are undergoing surgery to reduce pain. There are no studies on the prevention of acute pain in patients who at the time of surgery already suffer from significant pain. The indication for total hip replacement operations in patients with osteo-arthritis is generally reduction of pain or disability. Our study was designed to investigate the effects of prolonged analgesic prophylaxis with continuous epidural blockade with local anaesthetics on postoperative consumption of analgesics, quality of postoperative pain control, and side-effects.


After approval from the local ethics committee and written informed consent, 42 adult patients consecutive scheduled for total hip replacement were entered into this double-blinded, randomised, controlled, clinical study. Exclusion criteria included contraindications to epidural anaesthesia, neurological or coagulation disorders, analgesic or narcotic abuse, and an ASA classification of 4 or higher.

On the day before surgery, patients were made familiar with the use of the patient-controlled epidural analgesia pump (PCEA, Fresenius, Germany) and the visual analogue pain scales (VAS). The VAS model used for this study was 100 mm in length, a widespread and valuable instrument for measuring pain [6]. For study purposes, patients were advised to consider pain only in the hip planned for operation. All acute pain assessments were done during rest. However, before inclusion in the study, pain during movement was also assessed.

All routine medication, including analgesia, was continued until the evening before the operation. On the afternoon of the day preceding the operation, all patients were transferred from the general ward to the intermediate care ward of the Department of Orthopaedics and stayed there until the start of surgery. After insertion of an intravenous line, all patients received an epidural catheter placed in the L3/4 epidural space. After a test dose consisting of 10 ml ropivacaine 0.2% and confirmation of the correct position with pinprick and cold spray, patients were randomly assigned to one of two groups via a computer system.

Patients in the study group (‘pre-emptive’, n = 21) then received a continuous infusion of ropivacaine 0.2% at a rate of 5 ml.h−1. If analgesia was incomplete, the epidural pump rate was increased to a maximum of 10 ml.h−1. If there was motor blockade, the rate was reduced to a minimum of 3 ml.h−1. The study protocol called for pre-emptive analgesia lasting for at least 12 h prior to surgery.

After confirmation of the correct epidural position, patients in the control group (‘placebo’, n = 21) received a continuous infusion of saline 0.9% at a rate of 5 ml.h−1.

In both groups, the epidural infusion was continued until patients were transferred to the operating room. All patients received standard oral premedication with midazolam 7.5 mg. For operative analgesia, epidural ropivacaine 1% was used to achieve a sensory blockade to T8. Central venous, arterial and urinary catheters were inserted as deemed appropriate by the attending anaesthetist (S.G.), who was unaware of the treatment allocation of the patient.

After completion of the surgical procedure, patients in both groups received a patient-controlled epidural analgesia pump and were returned to the intermediate care unit. The pump was set to deliver boluses of 5 ml ropivacaine 0.2% with a lock-out interval of 15 min without an upper dose limit. If the analgesia was insufficient, patients could request a rescue analgesic consisting of intravenous piritramide (7.5 mg). Epidural catheters were removed at the request of the patients. Patients stayed in the intermediate care unit until the morning of the first postoperative day.


The main outcome parameter was the amount of epidural local anaesthetic required. Additional parameters included VAS pain scale ratings related to the hip operated on, the number of requests for epidural boluses, the need for additional narcotics, vital signs (heart rate, mean arterial blood pressure, capillary oxygen saturation) and side-effects. Additionally, biometric data (age, gender, height, weight), ASA classification, and operation- and anaesthesia-related variables (duration of surgery, blood loss, need for transfusion of packed red blood cells, sedation during operation, need for additional anaesthetic procedures) were recorded. All study-related measurements were taken by the same person (M.H.), who was not aware of the treatment allocation of the patients.

Points of measurement

All measurements were taken before insertion of the epidural catheter on the day before operation (M1), on arrival in the operating room (M2), after completion of surgery (M3), 4 h after completion of surgery (M4), on the evening of the operating day (M5), and every morning and evening until the epidural catheter was removed (M6–11).

Sample size calculation was performed on the assumption that the intervention might reduce the consumption of ropivacaine by 33%, from 300 mg to 200 mg with a standard deviation of 50 mg. With a p < 0.05 and a power of 0.8, a sample size of at least 16 patients per group would have been necessary. Statistical evaluation was achieved with analysis of variance for biometric, anaesthesia, and operation-related data. All other parameters were tested with multifactorial analysis of variance. Results where p < 0.05 were considered statistically significant.


Both groups were comparable regarding biometric, operation- and anaesthesia-related data. However, there was a trend towards younger age in the ‘pre-emptive’ group (mean [SD] 57 [12] vs. 65 [6] years) (Table 1). Pain ratings at rest and during movement did not differ significantly before operation (VAS values in the ‘pre-emptive’ group were 31 at rest and 66 during movement; in the control group, values were 32 at rest, and 60 during movement). At the time of inclusion, chronic therapy with non-steroidal analgesics was being used by 12 ‘pre-emptive’ and nine control patients (non-significant difference). Epidural anaesthesia was sufficient for surgical block in all cases. In both groups there were three patients who required general anaesthesia with laryngeal mask airway because of restlessness in spite of sufficient analgesia. In these cases, only propofol and no opioids were used intra-operatively (Table 1).

Table 1.  Biometric, operation and anaesthesia-related data, expressed as mean (SD).
 Pre-emptive (n = 21) Placebo (n = 21)
  1. NSAID, non-steroidal anti-inflammatory drug. ASA, American Society of Anesthesiologists.

Gender; % female6670
Age; years65 (6)57 (12)
Height; cm166 (6)165 (8)
Weight; kg79 (12)78 (18)
No. of patients on NSAID therapy129
ASA classification
Duration of operation; min107 (30)105 (22)
Total amount of midazolam; mg10.3 (2.5)10.9 (3.3)
Need for general anaesthesia33
Propofol; mg263 (269)264 (229)
Intra-operative ropivacaine consumption191.4 (150–220)191.9 (140–210)

Preoperatively, epidural infusion was maintained in both groups for 11–20 h. The mean (range) total amount of ropivacaine 0.2% in the ‘pre-emptive’ group was 140 (80–180) mg. Pre-emptive analgesia proved to be very effective. VAS values fell in the ‘pre-emptive’ group to 0 in 8/21 patients. The mean values of pain at rest were reduced from 32 to 6 in this group, whereas in the control group values remained stable at about 31 (p < 0.001) (Table 2).

Table 2.  Dose, duration, and clinical effect of ‘pre-emptive’ analgesia, expressed as mean (range).
 Preemptive (n = 21)Placebo (n = 21) 
  1. NS, nonsignificant.

Duration of pre-emptive analgesia; h13.8 (11–18)14.3 (11–20)NS
Amount of ropivacaine before surgery; mg140 (80–180)0 
VAS before pre-emptive analgesia30.7 (5–72)32.1 (10–64)NS
VAS after pre-emptive analgesia6.2 (0–20)31.2 (5–78)< 0.001

For operative analgesia, patients in the ‘pre-emptive’ group received a mean (SD) dose of 191 (39) mg of ropivacaine 1% as compared to 191 (43) mg in the control group. After surgery, patients in the ‘pre-emptive’ group required 194 (90–480) mg of ropivacaine, whereas patients in the control group needed 284 (100–550) mg. A total of 385 mg ropivacaine were used in the ‘pre-emptive’ group and 476 mg in the control group, a statistically significant difference (p = 0.0013). Patients in the control group also requested significantly more boluses, indicating insufficient analgesia (38 vs. 24). There were no differences between the groups in their consumption of additional analgesics or narcotics. However, in the control group, seven doses of narcotic analgesics (piritramide 7.5 mg i.v. each) were administered as compared to only one dose in the ‘pre-emptive’ group. Eight doses of diclofenac suppositories (100 mg each) were administered in the ‘pre-emptive’ group as compared to seven in the control group. The same was true for i.v. metamizol: four doses (1 g each) in the ‘pre-emptive’ group and three doses in the control group (Table 3).

Table 3.  Patterns of postoperative PCEA use, expressed as mean (range). Visual analogue pain scale scores (VAS) did not differ significantly between groups throughout the postoperative study period.
 Preemptive (n = 21)Placebo (n = 21) 
  1. NS, nonsignificant.

Postoperative ropivacaine consumption188.1 (90–480)284.6 (100–550)p = 0.0013
Total postoperative requests for boluses24.4 (10–54)38.3 (10–85)p < 0.01
Duration of PCEA47.8 (36–64)49.1 (37–65)NS
VAS postoperatively  NS
 VAS 41 (0–20)0 
 VAS 538.4 (20–60)39.9 (10–68) 
 VAS 635 (5–55)44.2 (10–72) 
 VAS 729 (10–50)38.4 (15–86) 
 VAS 822.2 (10–45)23.6 (5–52) 
 VAS 911.7 (0–40)13.8 (0–35) 
 VAS 1015 (0–32)9.1 (5–20) 
 VAS 1116.7 (5–22)5 (2–8) 
Additional piritramide doses 0.05 (0–1) 0.33 (0–4)NS
Additional NSAID doses 0.57 (0–3) 0.47 (0–3)NS

The time course of the VAS values did not differ significantly between groups (Table 3).

Vital signs remained stable throughout the study period in both groups without significant differences. There were no differences in blood loss, peri-operative haemoglobin concentration or need for packed red blood cells. The incidence of side-effects did not differ between the groups.


The efficacy of pre-emptive analgesia was proposed by Crile as early as 1913 [7, 8]. The idea has attracted new interest since 1983 [9, 10]. There is some uncertainty about the terminology of pre-emptive analgesia. Whereas many regimens in which prophylactic analgesia was administered were termed ‘pre-emptive’, this term should strictly only be used if analgesia is started before surgery and continued until significant pain no longer has to be taken into account [11]. By this definition, our study may correctly be described as ‘pre-emptive’ because analgesia was started in advance, targeted to interrupt hyperalgesia due to chronic pain and was continued throughout the postoperative period by patient-control, until patients wanted the epidural catheter to be removed.

In a recent large-scale randomised investigation, both local anaesthetics (bupivacaine) and opioids (fentanyl) have been found to be effective in preventing postoperative pain [12]. However, the concept has been challenged by studies which did not find an antinociceptive effect of pre-emptive analgesia [2]. In some studies, hyperalgesia has even been found to be triggered by preventive use of opioids [13]. Most studies in this field have investigated the effect of administration of local or regional anaesthesia, opioids, or non-steroidal anti-inflammatory drugs just minutes before surgical stimulation. However, a meta-analysis of all published studies has shown that this seems to be ineffective for prevention of postoperative pain [14].

The anatomical origin of long-lasting postoperative pain is thought to be in central neurological structures [15]. The pain is triggered by overwhelming sensory input in the peri-operative setting. It has been proposed that the use of pre-emptive epidural analgesia might lead to decreased pain perception [2]. However, no studies have tried to quantify the effect on postoperative pain of totally interrupting sensory input during chronic pain. Existing studies on long-lasting epidural blockade before surgery did not investigate acute pain, but concentrated on phantom limb pain as an inadvertent outcome of amputation, and conflicting results have been reported [4, 5].

The intention of our study was to achieve total sensory blockade via an epidural catheter for at least 12 h before surgery and to estimate the effect of this procedure on postoperative pain. Due to changes in the operation list, this goal was not achieved in all patients. However, only one patient per group received study medication for only 11 h, with the period of pre-emptive analgesia lasting longer than required in all other patients. We concentrated on acute postoperative pain as measured by consumption of analgesics and VAS ratings because we expected a possible effect in this setting to be the most likely. Our study results suggest that the role that the timing of pre-emptive analgesia plays may have been underestimated in previous studies. We found that long-lasting epidural block with local anaesthetics was effective in reducing postoperative analgesic requirements. ‘Pre-emptive’ patients required significantly less local anaesthesia and requested fewer additional boluses. At the same time, pain ratings, analgesic rescue therapy, and duration of postoperative analgesic requirements did not differ significantly.

The efficacy of the ‘pre-emptive’ regimen may be due to the fact that pre-existing pain in the hip, which was rated moderate to severe by the study subjects, was effectively eliminated by epidural analgesia in the study group. This is demonstrated by the VAS ratings, which in the ‘pre-emptive’ group were reduced before surgery not only statistically but also clinically to a significant extent. Most patients in the ‘pre-emptive’ group rated their residual pain on the day of operation as zero, although many of them had suffered from pain in this body region for a long period.

Generally, for VAS ratings, an intervention threshold of 30 is recommended [16]. By this standard, results of pain ratings in both groups were not optimal, as demonstrated by individual values exceeding 30, particularly on the evening of the operation day (M4). However, the analgesic regimen was a patient-controlled system with the additional option of rescue medication in the case of insufficient effect. As this option was only requested occasionally and patients exhibited high satisfaction with analgesic therapy in a postoperative questionnaire, the value of 30/100 has to be challenged in this context.

The pre-emptive analgesic regimen of our study only seems to be effective for the peri-operative period, and the results do not necessarily imply that pre-emptive analgesia is helpful in improving long-term outcomes of surgery. This is in accordance with Nikolajsen et al. [5], who found no significant difference in rates of phantom limb pain after pre-operative epidural analgesia. Possibly, significant effects on chronic pain could be anticipated in surgical procedures which are known to produce higher pain scores as an inadvertent outcome of surgery (e.g. thoracotomy or breast surgery) [1].

Side-effects of the analgesic regimen occurred more often in the ‘pre-emptive’ group than in the control group, although this difference was not statistically significant. However, all of these side-effects were classified as minor and in no case did any serious problems occur. In particular, haemodynamic and respiratory stability was preserved in all patients without interventions.

One problem with our study, which needs to be discussed, is the fact that blinding of the study patients or the anaesthetists involved may have been biased by the perceived effects of the pre-operative epidural analgesia. As expected, pain ratings before the start of the operation were significantly lower in the pre-emptive group than in the control group. However, this point was addressed in advance during the informed consent procedure with all studied patients, who were informed that they might or might not feel sensations like numbness, motor blockade or altered perception of temperature. Furthermore, as all patients were given the same test dose of local anaesthetic, the initial effect of the epidural block was comparable in both groups. In the placebo group, this effect was expected to wear off rapidly, thus preventing an effect of ‘pre-emptive analgesia’. The blinding of the anaesthetists involved was maintained by strict separation between pre-operative and intra-operative treatment. The anaesthetist caring pre-operatively for the patients (J.K.) was not involved in the intra- or postoperative course of the patients. The intra-operative anaesthetist (S.G.) exclusively cared for the patients in the operating theatre. Upon arrival of the patients in the theatre, the attending anaesthetist gave the first dose of ropivacaine (100 mg) after a negative aspiration test, and tested the effect after 15 min. Additional doses were administered as deemed necessary. However, after administration of the first epidural dose, it was impossible to distinguish which group the individual patient belonged to, and thus blinding was maintained.

All postoperative pain ratings – in contrast to the pre-operative ones – were done at rest. This complies with clinical practice in that patients are not mobilised out of bed before the first postoperative day. However, the highest VAS values were expected, and found, during the first 24 h. After that, pain ratings declined significantly. However, at that time patients still had the epidural catheter in place with additional unrestricted access to analgesics. This suggests that no intergroup differences would have existed concerning pain ratings during movement. However, measurement of postoperative VAS values during movement might have provided additional information.

In summary, our study showed that long-lasting epidural pre-emptive analgesia reduced postoperative consumption of analgesics with improved pain ratings. Further studies in this field should focus on the role of the timing of analgesia.


This study was supported by the Department of Anaesthesiology, Intensive Care Medicine, Pain Management, University Hospital Giessen, Germany. None of the authors or participants has any financial interest in the subject matter, materials, or equipment discussed or in competing materials.