Pronociceptive Effects of Remifentanil in a Mouse Model of Postsurgical Pain: Effect of a Second Surgery

Male Swiss CD1 mice weighing 25–28 g at the beginning of the experiments were used. Animals were housed five per cage and maintained in a room with a 12-h light–dark cycle (light between 8:00 am and 8:00 pm), at controlled temperature (21°± 1°C) and humidity (55 ± 10%). Food and water were available ad libitum  except during behavioral evaluation. All procedures and animal handling met the guidelines of the International Association for the Study of Pain and the European Communities directive 86/609/EEC regulating animal research. The protocol used in the study was endorsed by the ethics committee of our institution (ComitèÈEtic d'Experimentació del Parc de Recerca Biomèdica de Barcelona, Barcelona, Spain).

Remifentanil (Ultiva®; GlaxoSmithKline, Madrid, Spain) and sevoflurane (Sevorane®; Abbott Laboratories SA, Madrid, Spain) were supplied by the Department of Anesthesiology at the Hospital del Mar (Barcelona, Spain). Remifentanil was dissolved in saline (0.9% NaCl) and infused subcutaneously at the nape of the neck over a period of 30 or 60 min using a KD Scientific pump (KD Scientific Inc., Holliston, MA). All infusions were made during sevoflurane anesthesia (3.0–3.5% vol/vol), an inhalational anesthetic drug that we have previously shown has no effect on nociceptive thresholds.4An intravenous catheter (22 gauge) was inserted in the posterior aspect of the neck and carefully pushed forward subcutaneously approximately 1 cm. After removal of the needle, the catheter was loosely fixed around the neck of the mice with adhesive tape. Because mice were immobile during sevoflurane anesthesia, the catheter remained in place during the procedure.

In all instances and regardless of the remifentanil dose, the infusion rate was kept constant at 0.8 ml/h.

We used a mouse model of postoperative pain previously described in our laboratory.4In a sterile operating room, mice were anesthetized with sevoflurane (3.0–3.5% vol/vol) plus a constant infusion of remifentanil or saline, administered during a period of 30 min. A 0.7-cm longitudinal incision was made with a number 20 blade through the skin and fascia of the plantar surface of the right hind paw, starting 0.3 cm from the proximal edge of the heel and extending toward the toes. The underlying plantaris muscle was then exposed and incised longitudinally, keeping the muscle insertions intact. After hemostasis with slight pressure, the skin was closed with two 6-0 silk sutures and the wound covered with povidone-iodine antiseptic ointment. After surgery, animals were allowed to recover in cages with sterile bedding. Control animals underwent a sham procedure (sham incision) that consisted of the administration of sevoflurane plus saline for 30 min, without remifentanil or incision. When a second surgery was performed 27 days later, the same experimental protocol described above was used.

Hyperalgesia to noxious heat stimulation and to mechanical punctuate stimulation were determined in each experimental condition. Before the experiments, animals were habituated to the environment (testing equipment without nociceptive stimulation) for 3 days. We used the following nociceptive tests.

Heat hyperalgesia was evaluated as previously described.15Paw withdrawal latency in response to radiant heat was measured using the Hargreaves test equipment (Ugo Basile, Varese, Italy). Briefly, mice were placed in methacrylate cylinders (30 cm high, 9 cm in diameter; Servei Estació, Barcelona, Spain) positioned over a glass surface. Animals were habituated to the environment for 2 h before testing. The heat source was then positioned under the plantar surface of the hind paw and activated with a light beam intensity set to elicit baseline latencies of 9–11 s in control mice. A cutoff time of 20 s was used to prevent tissue damage in the absence of a response. The mean paw withdrawal latencies for both hind paws were obtained from the average values of three separate trials, taken at 5- to 10-min intervals, to reduce the possible influence of thermal sensitization on the response.

Mechanical nociceptive thresholds were evaluated measuring the hind paw withdrawal response to von Frey filament stimulation.16Animals were placed in methacrylate cylinders (30 cm high, 9 cm in diameter) with a wire grid bottom, through which the von Frey filaments were applied (bending force range from 0.008 to 2 g; North Coast Medical, Inc., San Jose, CA). To minimize stress during the experimental procedure, animals were allowed to habituate for 2 h before testing. The filament of 0.4 g was first used; then the strength of the next filament was increased or decreased according to the response (up–down method16). The results of the evaluation were obtained in grams, which is a continuous variable that can be analyzed with parametric methods. The upper limit value (2 g) was recorded even if there was no withdrawal response to this force. Clear paw withdrawal, shaking, or licking were considered nociceptive-like responses. Both hind paws were alternatively tested.

Special care was taken to reduce interindividual variability while using the smallest number of animals per group. Before the study, the animals were habituated by the same investigator for 3 days, and mechanical and thermal thresholds were determined daily during 3 additional days to obtain baseline values. All experimental groups received the same inhaled concentration of sevoflurane (3.0–3.5% vol/vol) during a remifentanil or saline infusion (rate of 0.8 ml/h).

In all experiments, the investigator recording the data was blinded to the treatment and the doses of remifentanil administered.

To establish whether the infusion dose (μg · kg⁻¹· min⁻¹), the duration of infusion (time), or the total dose administered over time would alter the extent and duration of postoperative hyperalgesia, we performed the following experiments.

Remifentanil was administered to different groups of mice (8–10 animals/group) at total doses of 20, 40, 80, or 100 μg/kg infused over a period of 30 min (corresponding to infusion doses of 0.66, 1.33, 2.66, or 3.33 μg · kg⁻¹· min⁻¹). Higher doses of remifentanil could not be used because of motor impairment and/or respiratory depression. Control mice received saline. Nociceptive thresholds (thermal and mechanical hyperalgesia) were determined 2, 4, 7, and 10 days after the procedure.

To establish the effect of the dose and time of infusion on the pronociceptive effects of remifentanil, we performed two sets of experiments: First, we determined whether a nonhyperalgesic dose of remifentanil, infused over an extended period of time (60 min), would induce significant hyperalgesia. We infused 0.66 μg · kg⁻¹· min⁻¹remifentanil for 60 min (total dose 40 μg/kg) and compared the effects with those observed after the administration of the same total dose (40 μg/kg) infused over a 30-min period (infusion dose of 1.33 μg · kg⁻¹· min⁻¹, positive control group). In both groups, the volume of infusion was kept constant at 0.8 ml/h. The dose of 40 μg/kg was selected because induced a distinct pronociceptive effect lasting several days when performing the dose–response curves. To ensure uniformity of the volume infused, the positive control group received a 30-min infusion of remifentanil (1.33 μg · kg⁻¹· min⁻¹in 0.4 ml), followed by a 30-min infusion of 0.4 ml saline. Nociceptive thresholds were determined 2, 4, 7, and 10 days after remifentanil.

A second set of experiments was performed to establish whether increasing the infusion time would enhance the magnitude and duration of a pronociceptive dose of remifentanil. We compared the effects of remifentanil at the same infusion dose (1.33 μg · kg⁻¹· min⁻¹) but administered during 60 min (total dose 80 μg/kg) or 30 min (total dose 40 μg/kg). The volume of infusion was kept at 0.8 ml/h. Nociceptive thresholds were determined 2, 4, 7, and 10 days after the opioid.

Experiments evaluating the effects of a single treatment (incision, remifentanil administration, or their combination), have been previously reported by our group, using the same strain of mice and experimental protocol.4In the current experiments, all animals received two treatments (first and second), separated by a period of 27 days, a time when nociceptive thresholds were completely recovered. The following treatments were used:

1. Sham incision (without surgery) and saline infusion (control group, sham incision + saline)

2. Incision and saline infusion performed during sevoflurane anesthesia (incision + saline)

3. Remifentanil infusion and sham incision (sham incision + remifentanil)

4. Incision performed during remifentanil anesthesia (incision + remifentanil)

After each treatment (first and second), nociceptive thresholds were measured at baseline and 1, 7, 10, 14, 18, 21, and 25 days later. When a second incision was performed, it was always on the same surgical site as the first one. The dose of remifentanil used was 2.66 μg · kg⁻¹· min⁻¹infused over a period of 30 min, except when testing a nonhyperalgesic dose of remifentanil (0.66 μg · kg⁻¹· min⁻¹) on repeated surgery (table 1).

Under these experimental conditions, we assessed whether repeated surgery during remifentanil anesthesia would enhance postoperative hyperalgesia. The observed effects were compared with those obtained after a repeated incision or a repeated remifentanil infusion (each separated by 27 days). (table 1). The following groups of experiments were performed:

1. Control group, where thermal and mechanical hyperalgesia were measured in nontreated animals; served as reference for the other groups (sham incision + saline)

2. Repeated incision (incision + saline)

3. Two surgeries performed each during 2.66 μg · kg⁻¹· min⁻¹remifentanil (incision + remifentanil)

4. Two surgeries performed each during 0.66 μg · kg⁻¹· min⁻¹remifentanil (incision + remifentanil)

5. Repeated remifentanil infusion at 2.66 μg · kg⁻¹· min⁻¹(sham incision + remifentanil)

We also evaluated whether either a previous incision or an infusion of remifentanil would modify the hyperalgesia induced by surgery with or without remifentanil anesthesia performed 27 days later (table 1), with the subsequent groups of experiments:

• Sham incision + remifentanil infusion (as first treatment), followed by incision + saline infusion (second treatment)

• Sham incision + remifentanil, or incision + saline (as first treatment), followed by incision + remifentanil (second treatment)

The mean area above the time–effect curves (AACs; 0–10 days) obtained with the different doses of remifentanil were plotted in figure 1to show the correlation between the infused doses and their overall effects. In this figure, each graph includes the equation corresponding to the represented linear regression, where y is the overall effect and x is the infused dose. Pearson coefficients, R  ², are included as a measure of the relation between both variables. The ED⁵⁰values of remifentanil for thermal and mechanical hyperalgesia were calculated by nonlinear regression analysis with a sigmoidal dose–response equation (variable slope) using GraphPad Prism 4 (GraphPad Software Inc, San Diego, CA).

All data presented in the time-course graphs (figs. 2–5) are expressed as mean values ± SD of 5–10 mice. For each mouse and time point, the responses in seconds (Hargreaves test) or grams (von Frey) are expressed as the changes with respect to the baseline values, normalized (subtracted) to the mean value of the corresponding control group (represented in the figures by a broken line). This calculation facilitates the graphic representation and interpretation of the data, where negative values indicate net pronociceptive effects and positive values indicate antinociception.

The time course of the effects of the infusion dose, the duration of infusion, and the total dose were analyzed using a linear mixed model with two factors, the experimental condition (infusion dose, infusion duration, or total dose) and the time of evaluation (day), as well as their interaction. A random intercept was considered, but random effects were not included. For the covariance structure of the repeated measures, a diagonal matrix was chosen. If the interaction between experimental condition and time was statistically significant, multiple Tukey post hoc  comparisons for repeated measures were performed at each time point. Analyses were performed with R (The R Foundation for Statistical Computing, Vienna, Austria), using its libraries nlme17and multcomp.18In figure 2, mean values ± SDs for the AACs are represented as histograms to illustrate the overall nociceptive sensitization corresponding to each experimental condition. Because the AAC data are reduced to a single value per mouse, the comparison of the AACs was performed with the nonparametric Wilcoxon Mann–Whitney test for independent samples, using SPSS version 13.0 (SPSS Inc., Chicago, IL).

To analyze the effects of the first and second treatment at each time point, nonparametric multiple test procedures were applied19using the R npmc library.20AACs and interquartile ranges (table 1) were calculated to illustrate the overall nociceptive sensitization corresponding to the first and second treatments. The comparison of the AACs was performed with the nonparametric Wilcoxon Mann–Whitney test for repeated measures, using SPSS version 13.0. Because of the duration of the pronociceptive effects observed after surgery, the AACs of the time effect were calculated for a period of 0–25 days (figs. 3–5).

Baseline thresholds to thermal and mechanical stimuli obtained before remifentanil administration were similar in all groups of study, with mean values of 11.28 ± 1.05 s and 1.21 ± 0.14 g, respectively. Saline administration (control group) did not induce significant changes in nociceptive thresholds over the 10-day period of evaluation. In contrast, increasing the dose of the remifentanil infusion induced dose-dependent pronociceptive effects in both the Hargreaves and von Frey tests. Maximal pronociceptive effects were observed on day 2, and these results are shown in tables 2 and 3.

In the Hargreaves test, the magnitude and duration of remifentanil-induced thermal hyperalgesia increased in a dose-dependent manner (P < 0.001, for the dose and time of evaluation; table 2). At this time point, the highest infusion doses (2.66 and 3.33 μg · kg⁻¹· min⁻¹) induced 34 ± 21% and 34 ± 9% decreases in thermal latency (P < 0.01 vs.  saline) that remained statistically significant up to 10 days after remifentanil infusion (26 ± 21% to 33 ± 14% decrease; P < 0.05). At the same time point (day 2), the lower doses of remifentanil did not induce significant pronociceptive effects when compared with saline.

Comparison of the overall effects of remifentanil induced by the different doses was achieved using the AACs of the time–effect curves over the 10-day period of evaluation. Figure 1Ashows a linear positive correlation between the different doses of remifentanil and the magnitude of thermal hyperalgesia, with a coefficient R  ²= 0.9676; the calculated ED⁵⁰of remifentanil was 1.7 (95% confidence interval, 1.3–2.1) μg · kg⁻¹· min⁻¹. These values could be underestimated because of the impossibility to test higher doses of remifentanil.

In the von Frey test, remifentanil also induced long-lasting mechanical hyperalgesia in a dose-dependent manner (P < 0.001). On day 2 after administration, the highest doses (2.66 and 3.33 μg · kg⁻¹· min⁻¹) induced 40 ± 15% and 42 ± 19% decreases in mechanical thresholds, respectively (P < 0.01 vs.  saline; table 3). Mechanical hyperalgesia remained statistically significant 10 days after treatment (P < 0.05 compared with the saline group), with residual mean threshold reductions of 24 ± 18% and 27 ± 9%. At day 2, the dose of 0.66 μg · kg⁻¹· min⁻¹did not induce significant hypersensitivity to mechanical stimuli. A linear correlation between the dose and the extent of mechanical hyperalgesia (AAC) was also obtained, with a correlation coefficient R  ²= 0.9496 (fig. 1B) and a calculated ED⁵⁰of 1.26 (95% confidence interval, 1.0–1.6) μg · kg⁻¹· min⁻¹.

In these experiments, we first assessed whether remifentanil at nonhyperalgesic doses could induce hyperalgesia when infused over an extended period of time. We infused the same total dose of 40 μg/kg remifentanil over a 60-min (0.66 μg · kg⁻¹· min⁻¹) or 30-min period (1.33 μg · kg⁻¹· min⁻¹, positive control group). The results in the Hargreaves test show that the infusion of 40 μg/kg remifentanil in 30 min induced significant thermal hyperalgesia at days 2 and 4 (P < 0.05 compared with saline; fig. 2A, left). On the contrary, when the same total dose was infused over a period of 60 min (0.66 μg · kg⁻¹· min⁻¹), no pronociceptive effects were observed. The AACs of the time–effect curves were 2.7 ± 6 and 14.5 ± 12 for the 0.66- and 1.33-μg · kg⁻¹· min⁻¹doses (P = 0.089). These results are represented as histograms on the right side of the plantar latency graph shown in figure 2A.

Similar results were obtained in the von Frey test. The pronociceptive effects of 40 μg/kg remifentanil infused over 30 min (1.33 μg · kg⁻¹· min⁻¹) were maximal on day 2 and lasted approximately 7 days (fig. 2A, right). When the same total dose was administered over 60 min (0.66 μg · kg⁻¹· min⁻¹), no significant pronociceptive effects were observed. The AACs were 0.51 ± 0.8 and 2.97 ± 0.7 for the 0.66- and 1.33-μg · kg⁻¹· min⁻¹infusion doses (P < 0.05). This demonstrates a significant overall pronociceptive effect of the 1.33-μg · kg⁻¹· min⁻¹dose (P < 0.001 compared with saline) but no significant effect of the 0.66-μg · kg⁻¹· min⁻¹infusion dose (fig. 2A, histogram). The results show that administration of remifentanil at a low infusion dose over an extended period of time does not induce significant pronociceptive effects.

We also assessed whether a dose of remifentanil that induced pronociceptive effects would increase hyperalgesia when infused over a prolonged period of time. We compared the effects of the same dose of remifentanil (1.33 μg · kg⁻¹· min⁻¹) infused over a period of 60 min (total dose 80 μg/kg) or 30 min (total dose 40 μg/kg). Both groups displayed significant thermal and mechanical hyperalgesia when compared with the saline group (fig. 2B; P < 0.05 for each test). In both groups, thermal hyperalgesia was statistically significant on days 2 and 4 when compared with saline (P < 0.01), lasting approximately 4 days. Mechanical hyperalgesia was observed on days 2, 4, and 7. For the 60-min group, P < 0.05 was obtained at day 2, when compared with saline, whereas for the 30-min group, differences were statistically significant from saline on days 4 and 7 (P < 0.05) and lasted approximately 7 days. In these experiments (fig. 2B, graphs), no significant differences between groups receiving remifentanil could be established at any time point or when comparing the AACs (fig. 2B, histograms). The results show that the increase in the infusion time did not alter the magnitude and duration of the pronociceptive effects of remifentanil.

In all of the groups, baseline nociceptive thresholds were similar, with mean values of 11.82 ± 0.8 s in the Hargreaves test and 1.16 ± 0.2 g in the von Frey test.

After a first surgery, significant thermal hyperalgesia was observed in the operated paw in both the incision + saline group and the incision + remifentanil group. Hyperalgesia was of similar magnitude (on day 1) but longer lasting in the latter (surgery + remifentanil [2.66 μg · kg⁻¹· min⁻¹]; fig. 3A). On the first day after surgery, thermal nociceptive thresholds decreased 62 ± 19% and 67 ± 15% in the saline and remifentanil groups, respectively (P < 0.01 for each compared with untreated mice), with no significant differences between groups at this time point. In the incision + saline group, thermal hyperalgesia disappeared 7 days after surgery, but lasted up to 10 days in mice receiving incision + remifentanil anesthesia (28 ± 25% decrease at day 10; P < 0.05). Moreover, the comparison of the AACs showed significant differences between the incision + saline and incision + remifentanil groups after the first surgery (P = 0.031; table 4).

A second surgery performed 27 days later also induced sustained thermal hyperalgesia in the incision + saline and incision + remifentanil groups, with no statistical differences between groups on the first day after surgery (fig. 3A). The AAC showing the overall thermal hyperalgesia was significantly greater in the incision + remifentanil group than in the incision + saline group (P = 0.01). For each group, no significant differences in the magnitude and duration of thermal hyperalgesia were observed when comparing the first and second surgeries (fig. 3Aand table 4).

In the von Frey test (fig. 3Band table 1), significant mechanical hyperalgesia was observed after the first surgery in the incision + saline and the incision + remifentanil [2.66 μg · kg⁻¹· min⁻¹] groups (P < 0.01 vs.  untreated mice), and was more severe and long lasting in the incision + remifentanil group. On the first day after surgery, mechanical thresholds decreased 43 ± 7% and 64 ± 11% in the incision + saline and incision + remifentanil groups, respectively (P < 0.05). Thresholds returned to baseline 7 days after surgery in the incision + saline group, but persisted up to 10 days in the incision + remifentanil group (35 ± 20% decrease on day 10; P < 0.05 vs.  control). The comparison of the AACs showed significant differences between the incision + saline and incision + remifentanil groups (P < 0.01).

The second surgery also induced pronounced mechanical hyperalgesia in both groups. On day 1 after the second surgery, the decrease in mechanical thresholds was enhanced (72 ± 16% and 81 ± 13% for each group). The effects lasted approximately 7 days in the incision + saline group but were prolonged up to 14 days in the incision + remifentanil group (42 ± 19% decrease at day 14; P < 0.01 vs.  control). Comparison of the AACs also demonstrated significant differences between the incision + saline and incision + remifentanil groups (P < 0.01). Therefore, in our experimental mouse model, the administration of a pronociceptive dose of remifentanil during two consecutive surgeries significantly increased the magnitude of postoperative mechanical (but not thermal) hyperalgesia after the second surgery (tables 1 and 4).

The intraoperative administration of a nonhyperalgesic dose of remifentanil (0.66 μg · kg⁻¹· min⁻¹) during two consecutive surgeries (fig. 3Cand table 1) induced a significant mechanical hyperalgesia after each surgery when compared with untreated mice (sham incision + saline), and the second surgery produced a higher degree of hyperalgesia (table 1). The results also show that on day 1 after surgery, mechanical hyperalgesia thresholds decreased 48 ± 16% and 55 ± 18% after the first and second surgeries, respectively, lasting 7 and 10 days (P < 0.05 compared with the untreated group; fig. 3C). Table 1shows the values of the AACs in the following groups: incision + saline, incision + remifentanil (2.66 μg · kg⁻¹· min⁻¹), and incision + remifentanil (0.66 μg · kg⁻¹· min⁻¹). A one-way analysis of variance followed by the Tukey test revealed a significant effect of the dose of remifentanil for both the first and the second surgeries (P < 0.05). Moreover, incision + nonhyperalgesic doses of remifentanil induced a longer-lasting hyperalgesia than incision + saline after the first and second surgeries (fig. 3C). Therefore, surgery performed during low doses of remifentanil still enhances postoperative hyperalgesia when compared with surgery + saline performed during sevoflurane anesthesia, supporting the results showing a dose-related hyperalgesic effect of remifentanil (fig. 1).

We also evaluated the effects of the administration of two consecutive doses of remifentanil (2.66 μg · kg⁻¹· min⁻¹, 30 min) separated by a period of 27 days (table 1and fig. 4, repeated sham incision + remifentanil). Mechanical hyperalgesia was similar on day 1 after each treatment (44 ± 20% and 50 ± 14% decreases for the first and second administrations). However, sensitization lasted 10 days after the first exposure (21 ± 28% decrease on day 10; P < 0.05 compared with controls) and 14 days after the second (32 ± 27% decrease; P < 0.05). The comparison of the AACs shows that overall hyperalgesia was greater after the second exposure (table 1).

In another group of experiments (table 1), we assessed whether a previous remifentanil exposure could enhance incision-induced hyperalgesia. In figure 5, we have plotted the effects of a first incision + saline treatment with and without a previous remifentanil infusion. The figure shows that surgery after remifentanil exposure induces a 51 ± 19% decrease in mechanical thresholds on day 1 that lasted up to 10 days (18 ± 25% decrease; P < 0.05 compared with control). On the contrary, hyperalgesia disappeared on day 7 when mice were not previously exposed to remifentanil. These results indicate that a previous exposure to remifentanil increases incisional pain. It is interesting to note that in table 1, when the overall pronociceptive effects of sham incision + remifentanil (as first treatment) are compared with those of the incision + saline (as a second treatment), they induce similar pronociceptive effects.

Finally, we assessed whether a previous surgery (incision + saline) or a remifentanil infusion (sham incision + remifentanil) would distinctly change postoperative hyperalgesia induced by a subsequent surgery performed during remifentanil anesthesia (table 1). In both groups, the second treatment induced greater hyperalgesia than the first one, but the extent of mechanical hyperalgesia after the second procedure was similar regardless of whether animals received an infusion of remifentanil or a surgical incision.

The current study shows that the pronociceptive effects of remifentanil are determined by the dose rather than by the duration of infusion. Therefore, regardless of the time of exposure, drug concentration at the μ-opioid receptor effector sites seems to be the critical factor for the development of remifentanil-induced nociceptive sensitization. The study also shows, for the first time, that when a second surgery is performed after nociceptive thresholds are restored, a significant increase in postoperative mechanical (but not thermal) hyperalgesia is observed, regardless of the type of anesthesia. In all instances (first and second surgeries), the extent and duration of postoperative pain sensitization is significantly greater when surgery is performed during remifentanil anesthesia. Moreover, a previous exposure to remifentanil enhances the duration of incision- and remifentanil-induced hyperalgesia.

In the same mouse model of postoperative pain, we have previously reported that remifentanil induces delayed hyperalgesia and enhances postincisional pain when infused during surgery.4,21In the current investigation, we tried to assess whether the mode of remifentanil administration (infusion dose, time, total dose) could be a determinant of its pronociceptive effects. The objective was to provide answers to unsolved clinical questions that may help to reduce postoperative pain in patients undergoing surgery during remifentanil anesthesia. Remifentanil has been reported to induce dose-dependent pain sensitization in several experimental and clinical studies in humans.6–9However, these studies assessed simultaneously the effect of a given infusion dose of remifentanil (μg · kg⁻¹· min⁻¹) and the effect of the total dose of drug administered, which is proportional to the time of infusion when the infusion dose is kept constant. Consequently, the likely pronociceptive effects associated with each factor independently could not be established. In the current study, we were able to show that remifentanil-induced pronociceptive effects correlate positively with the infusion dose used, whereas the time of administration or the total dose has no major effects. However, because remifentanil plasma levels after subcutaneous administration were not determined in our study, a different relation among dose, concentration, and time after intravenous administration cannot be excluded.

The acute administration of phenanthrene derivatives (morphine, methadone) by different routes induces delayed nociceptive sensitization lasting up to 2 days,11–13whereas hyperalgesia induced by piperidine derivatives such as fentanyl has been reported to persist up to 5 days.22In our experimental conditions, remifentanil-induced sensitization persisted for approximately 10 days, which is the longest period of time reported after a short exposure (30 min) to an opioid. However, because no direct comparison between the effects of matched doses of the different opioids was attempted in our study, no definite conclusions regarding the possible longer duration of the pronociceptive effects of remifentanil can be derived. The comparison of the postoperative pain sensitization induced by opioids following different schedules of administration has not been fully investigated, even though it could be a relevant factor in the development of chronic postsurgical pain. Although conclusions from animal studies cannot be precisely applied to humans, our results strongly suggest that the administration of low infusion doses of remifentanil in clinical practice would reduce the hyperalgesic effects regardless of the duration of infusion. To support this assumption, a recent study reported greater postoperative pain 1 month after breast surgery in patients who received high doses of opioids in the postanesthesia care unit.23 

In our study, remifentanil had more effect decreasing mechanical than thermal thresholds, and lower doses were needed to induce mechanical (ED⁵⁰1.26 μg · kg⁻¹· min⁻¹) than thermal hyperalgesia (ED⁵⁰1.7 μg · kg⁻¹· min⁻¹). This is consistent with the greater sensitivity to opioids of pathways activated by mechanical rather than by thermal stimuli, as suggested in several studies.3,4,12,24,25In our model, the administration of 40 μg/kg remifentanil (1.33 μg · kg⁻¹· min⁻¹) did not induce cold allodynia in the acetone drop test26(data not shown), a result supported by clinical studies evaluating cold sensitivity after remifentanil anesthesia.13 

The results demonstrate that a given remifentanil infusion dose induces the same degree of hyperalgesia regardless of the duration of infusion. However, prolonged infusion times may induce intraoperative acute antinociceptive tolerance to remifentanil,27–29an aspect that was not evaluated in the current investigation. It is likely that increasing remifentanil infusion doses to compensate the decrease in efficacy (acute tolerance) would enhance postoperative hyperalgesia.

The absence of cumulative effects of remifentanil when infused over extended periods of time is probably related to the rapid degradation of the opioid by plasma and tissue esterases1; this property would favor steady state levels of μ-opioid receptor occupancy during the infusion, partially explaining that the duration of anesthesia does not alter delayed remifentanil hyperalgesia. Time of infusion might be a more relevant factor when assessing the pronociceptive effects of opioids with slower metabolism/longer action.30–32This would imply that low doses of remifentanil could be infused for prolonged periods of time without inducing postoperative pain sensitization, and also that a single bolus of a high dose of remifentanil (i.e. , during the induction of anesthesia) could cause significant and long-lasting hyperalgesia in the postoperative period. Currently, we are testing these assumptions in our mouse model.

Establishing the effective doses of remifentanil that induce pronociceptive effects in our model was essential to select the optimal doses to study its effects after repeated surgery. Based on the current results, we were able to select the hyperalgesic (2.66 μg · kg⁻¹· min⁻¹) and nonhyperalgesic doses (0.66 μg · kg⁻¹· min⁻¹) used in the second part of the study.

The extent and duration of postoperative pain sensitization after a plantar incision were significantly increased in mice anesthetized with remifentanil at hyperalgesic doses, corroborating previous findings from our laboratory.4,21After complete recovery, a second surgery performed at the same site and in analogous experimental conditions induced more prominent and persistent changes in mechanical thresholds in remifentanil-treated mice, regardless of the remifentanil dose (table 1). Interestingly, repeated surgery performed during a low nonhyperalgesic dose of remifentanil increased the duration of postoperative hyperalgesia when compared with incision alone (fig. 3Cand table 1). From these experiments, we could conclude that the magnitude and duration of the pronociceptive effects of remifentanil, when used during surgery in mice, are dose dependent, supporting the results of the first part of the study.

The repeated administration of two identical doses of remifentanil (2.66 μg · kg⁻¹· min⁻¹) separated by a 27-day interval increased hyperalgesia after the second dose, corroborating the results obtained after repeated administration of other opioids (heroin,33morphine,34or fentanyl22). Moreover, either an incision or an exposure to remifentanil increased the duration of hyperalgesia after a subsequent incision (figs. 3 and 5), regardless of surgery being performed during sevoflurane or remifentanil anesthesia. These results support previous studies performed in rats showing that chronic morphine administration increases incision-induced hyperalgesia and local cytokine release.35,36However, other studies show that acute administration of morphine reduces peri-incisional cytokine expression and neutrophil infiltration.37The reduction of opioid-containing leukocytes around the wound could increase incision-induced hyperalgesia.38In addition, the reported effects of opioids delaying wound healing39–41could also favor the perpetuation of postoperative pain.

Taken together, our results suggest that, after nociceptive thresholds return to baseline values, a previous surgery and/or remifentanil infusion induces long-lasting persistent neuroplastic changes in nociceptive pathways that facilitate mechanical pain sensitization in future situations (pain memory42) A long-lasting imprint induced by acute nociceptive stimuli in the nervous system has been reported in models of inflammatory pain induced by carrageenan,42–45although substantial differences between postincisional and other pain models are likely to be present.46–48 

Multiple central and peripheral mechanisms have been implicated in pain sensitization after tissue injury including C-fiber sensitization,42protein kinase C,49,N -methyl-d-aspartate receptors,43nitric oxide,4and spinal dynorphin,28,50among others. Surprisingly, all of these mechanisms have also been implicated in opioid-induced hyperalgesia.3Our results in a postoperative pain model in mice put forward the clinical need to take preventive measures during surgical anesthesia, to avoid latent sensitization in future surgeries.

In conclusion, the current study illustrates that the infusion dose, but not the duration of infusion or the total dose administered, determines the pronociceptive effects of remifentanil. The study also shows for the first time that a second incision performed at the same surgical site during high- or low-dose remifentanil anesthesia increases postoperative mechanical hyperalgesia in mice. Moreover, either a previous exposure to remifentanil alone or a first surgery (performed with or without remifentanil) significantly enhances postoperative mechanical hyperalgesia after a second surgery. Although preclinical studies sometimes do not translate to the clinical bedside, the current results may be useful to design clinical studies testing the effects of remifentanil after repeated surgery in humans.

The authors thank Klaus Langohr, Ph.D. (Statistician, Grup de Recerca Clínica en Farmacologia Humana i Neurociències, Programa de Recerca en Neuropsicofarmacologia, Institut Municipal de Investigacions Mèdiques–Hospital del Mar, Barcelona, Spain), for his excellent assistance in the statistical analysis; Marta Pulido, M.D. (Servei d'Assessorament en Edició I Publicació Biomèdica, Institut Municipal de Investigacions Mèdiques–Hospital del Mar, Barcelona, Spain), for editorial assistance; and José Cerezo (Freelance Graphic Designer, Carcaixent, Valencia, Spain) for his help with the figure design.