Efficacy of Pregabalin in Acute Postoperative Pain Under Different Surgical Categories

Abstract

The efficacy of pregabalin in acute postsurgical pain has been demonstrated in numerous studies; however, the analgesic efficacy and adverse effects of using pregabalin in various surgical procedures remain uncertain. We aim to assess the postsurgical analgesic efficacy and adverse events after pregabalin administration under different surgical categories using a systematic review and meta-analysis of randomized controlled trials.

A search of the literature was performed between August 2014 to April 2015, using PubMed, Ovid via EMBASE, Google Scholar, and ClinicalTrials.gov with no limitation on publication year or language. Studies considered for inclusion were randomized controlled trials, reporting on relevant outcomes (2-, 24-hour pain scores, or 24 hour morphine-equivalent consumption) with treatment with perioperative pregabalin.

Seventy-four studies were included. Pregabalin reduced pain scores at 2 hours in all categories: cardiothoracic (Hedge's g and 95%CI, −0.442 [−0.752 to −0.132], P = 0.005), ENT (Hedge g and 95%CI, −0.684 [−1.051 to −0.316], P < 0.0001), gynecologic (Hedge g, 95%CI, −0.792 [−1.235 to −0.350], P < 0.0001), laparoscopic cholecystectomy (Hedge g, 95%CI, –0.600 [–0.989 to –0.210], P = 0.003), orthopedic (Hedge g, 95%CI, −0.507 [−0.812 to −0.202], P = 0.001), spine (Hedge g, 95%CI, −0.972 [−1.537 to −0.407], P = 0.001), and miscellaneous procedures (Hedge g, 95%CI, −1.976 [−2.654 to −1.297], P < 0.0001). Pregabalin reduced 24-hour morphine consumption in gynecologic (Hedge g, 95%CI, −1.085 [−1.582 to −0.441], P = 0.001), laparoscopic cholecystectomy (Hedge g, 95%CI, –0.886 [–1.652 to –0.120], P = 0.023), orthopedic (Hedge g, 95%CI, −0.720 [−1.118 to −0.323], P < 0.0001), spine (Hedge g, 95%CI, −1.016 [−1.732 to −0.300], P = 0.005), and miscellaneous procedures (Hedge g, 95%CI, −1.329 [−2.286 to −0.372], P = 0.006). Pregabalin resulted in significant sedation in all surgical categories except ENT, laparoscopic cholecystectomy, and gynecologic procedures. Postoperative nausea and vomiting was only significant after pregabalin in miscellaneous procedures.

Analgesic effects and incidence of adverse effects of using pregabalin are not equal in different surgical categories.

INTRODUCTION

Pregabalin is a structural analogue of gamma-aminobutyric acid that acts as a potent ligand for alpha 2-delta subunits of the voltage-gated calcium channels in the nervous system. Such action results in a reduction in the depolarization-induced influx of calcium, hence a reduction in the release of excitatory neurotransmitters including glutamate, noradrenaline, dopamine, and serotonin.¹ Compared with gabapentin, pregabalin is more potent, is associated with fewer adverse effects, and has a more predictable and linear pharmacokinetic profile.^1,2 Its absorption is extensive, rapid, and proportional to dose.^1,2 Pregabalin is an attractive adjuvant for perioperative analgesia in this regard as it can be taken on an empty stomach, does not lead to gastrointestinal bleeding, and is generally well-tolerated.³

A multimodal analgesic technique is now often employed in acute postsurgical pain management in an attempt to improve analgesic efficacy and decrease requirement for opioids that are associated with undesirable adverse effects.⁴ Uses of pregabalin therefore range from treatment of neuropathic pain to being an adjunct in the multimodal management of postsurgical pain.⁴

The efficacy of pregabalin in treating acute postsurgical pain has been demonstrated in numerous studies. A recent meta-analysis has suggested that pregabalin, at all doses and administration regimens, has opioid-sparing effects and reduces pain scores in the postsurgical setting,⁵ at the expense of increased sedation and visual disturbances; however, the efficacy of pregabalin in providing such in various surgical categories remains uncertain, and it is not known whether the risk : benefit ratio is greater for certain surgical categories. Therefore, the aim of this meta-analysis was to evaluate the analgesic efficacy of pregabalin in reducing postsurgical pain in terms of 2- and 24-hour postsurgical visual analogue scale (VAS) pain scores and 24-hour accumulative morphine-equivalent consumption, in various surgical categories to provide a useful reference in perioperative care.

MATERIALS AND METHODS

Protocol

This review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines for reporting meta-analyses.⁶ Approval by ethics committee or written consent were not required for the extraction of data on studies already conducted for the purposes of this meta-analysis. Before commencing this meta-analysis, all authors agreed on the inclusion and exclusion criteria, which were articles with at least the abstract published in English, and gave data on at least 1 of the primary outcomes. This protocol was not published.

Eligibility Criteria

Studies considered for inclusion in the meta-analysis were randomized, double-blinded, controlled trials (RCTs) that investigated a minimum of 10 subjects in each group⁶ and reported on relevant pain outcomes with intervention or treatment with perioperative pregabalin. These studies had to present data for at least 1 of our prespecified outcome variables, which were 2- or 24-hour postsurgical pain or 24-hour morphine-equivalent consumption.

Systematic Search

A comprehensive search for literature for pregabalin was performed between August 2014 to April 2015, using PubMed, Ovid via EMBASE, Google Scholar, and ClinicalTrials.gov with no limitation on the year of publication or language. Attempts were made at accessing www.clinicalstudyresults.org to identify potentially relevant studies that have not been published in medical journals, but the website is no longer in use. The keywords used in the search included “pregabalin,” “lyrica,” “analgesia,” “acute pain,” “post-surgical pain,” and “post-operative pain.” Identified references were screened using title, abstract, and keywords. Searches of the reference lists of identified studies were also made.

Study Selection and Data Collection

Two primary investigators (D.M.H.L. and S.W.C.) screened the titles independently and removed the studies that did not meet the specified screening criteria. Abstracts, literature reviews, and meta-analyses were excluded. Potentially eligible trials were analyzed in detail on the basis of the full text and disagreements were discussed between D.M.H.L. and S.W.C. Data extraction was performed by the 2 reviewers (D.M.H.L. and S.W.C.) independently and included data on the patient (number of subjects, type of surgery, and type of anesthesia), data on the intervention and control (dose and frequency of pregabalin administered), and data on the outcomes (pain intensity, given as acute pain scores at rest, total opioid-equivalent consumption, and adverse effects including nausea, vomiting, sedation, and visual disturbance). Assessing each study for surgical category was performed by D.M.H.L. and C.-W.C.

Data Extraction

The pain intensity measured by either VAS or numeric rating scale (NRS) was extracted as pain scores. These scales have been shown to correlate well.⁷ The cumulative opioid consumption at the closest time to 24 hours postsurgery was extracted and converted to an equianalgesic dose of parenteral morphine in mg, based on the following scale: 15 : 1 for hydromorphone, 1.3 : 1 for oxycodone, 1 : 100 for fentanyl, 20 : 1 for codeine, 10 : 1 for tramadol, 10 : 1 for pethidine, 4 : 1 for hydrocodone, 1 : 100 for remifentanil, 1 : 1 for piritramide, and 1 : 1 for nalbuphine.^8–11 If results were presented as the number of doses given, data were extracted from the methods section to ascertain the dosage and then converted to equianalgesic dose of parenteral morphine in mg for inclusion in the meta-analysis. Data regarding postsurgical analgesic consumption were not included from studies that did not utilize opioids during the postsurgical period^12–14 or if data were presented as the number of patients who required rescue analgesics, although pain scores and other information from these studies were included in the analysis whenever given.

The primary outcomes of this present study were pain scores at rest at 2 and 24 hours postsurgery, and morphine-equivalent consumption in the first 24 hours postsurgery. Secondary outcomes were sedation at first assessment and adverse effects. Pain scores at 2 hours postsurgery were selected as the first time point for analysis because pain prior to that time point might be reduced by the effects of analgesics administered during anesthesia. Where pain scores were not available at 2 hours postsurgery, the closest time point was used. Pain scores and opioid consumption at 24 hours postsurgery were chosen in this study as most trials assessed here ceased data collection after 24 hours. Where data were presented graphically, the originals were obtained from the authors or extracted from graphs if no response was obtained from the authors. Twenty-eight corresponding authors were emailed for further details regarding data in the published studies. Seventeen of the emailed authors replied with further data not available in the published articles.

Studies were classified according to surgical categories, these were gynecologic, orthopedic (not including spine surgery), spine, ear, nose and throat (ENT), cardiothoracic surgery, and laparoscopic cholecystectomy. Where studies reported cumulative data on several different surgical categories or if the authors were only able to find 1 or 2 studies of that surgical category (eg, eye surgeries and breast surgeries), these were included in a miscellaneous, or >1 surgical category, group.

Assessment of Risk of Bias

The quality of the studies was assessed by 2 investigators (D.M.H.L. and S.W.C.) independently, using the Cochrane Collaboration's tool for assessing risk of bias.¹⁵

Statistical Analysis

Meta-analysis was used to assess the pooled effects of pregabalin 2 hours and 24 hours postsurgery. If the study included different doses of pregabalin, the higher dosage was used in this analysis. Data were analyzed using Comprehensive Meta-Analysis software (version 2.2.064, Englewood, NJ). Meta-regression was not performed in this review as a minimum number of 10 studies per subgroup is required.¹⁶

VAS pain scores or NRS pain scores were extracted from each study. Mean and standard deviation (SD) values were used when available, but when median and range data were presented, the mean was estimated using the median value, or the median value itself was used if the sample size exceeded 25 subjects in each group.¹⁷ In addition to the various different scoring methods used to assess pain, another major consideration was the heterogeneity of the studies, which included different types of patients, different pregabalin regimens in terms of time, dose and frequency, and method of administration. To take into consideration the heterogeneity of the studies, Hedge g standardized mean difference, which is the difference between the 2 means divided by the pooled SD, with a correction for small sample bias, using a random-effects model was computated and reported as the effect size between the pregabalin and the control groups. Hedge g was chosen as most of the studies investigated in this meta-analysis were small (<40 subjects per group). Hedge g is also an index of treatment efficacy independent of the scoring system used to measure efficacy, which is particularly useful in the present study as VAS 0–10, VAS 0–100 and NRS have all been used as pain scoring systems.

With regard to the analysis of adverse effects of pregabalin, in studies that have categorized patients according to a score (eg, sedation score) and if continuous data were available, this was inputted as means (SD). For studies that have categorized patients according to none, slight, moderate, or severe sedation, all patients, except those who had been classified as “none” by the investigators were regarded as being sedated for the purposes of this present meta-analysis, and these data were inputted using dichotomous data handling techniques. A Forest plot was generated for each endpoint and Hedge g with 95% confidence intervals (CIs) were reported. Effects on dichotomous outcomes such as visual disturbances, nausea, vomiting, and postsurgical nausea and vomiting were reported using odds ratio (OR) with a random-effects model. Publication bias was assessed using Funnel plots (Comphrensive Meta-Analysis).¹⁸ Sensitivity analysis was assessed using the 1 study removed technique. For all tests, statistical significance was defined as a 2-tailed P value of < 0.05.

RESULTS

Our primary search strategy identified 1700 publications. Seventy-four studies were included in this meta-analysis (Supplementary Figure 1). Results here were presented as all included studies and then according to the surgical category.

Risk of Bias

The results of the risk of bias assessment are summarized in Supplementary Table 1.

Study Protocols

The study protocols of the included trials varied significantly and led to considerable heterogeneity.

It is important to note that the primary outcomes as defined in this meta-analysis were not necessarily the primary outcomes of the published trials, and therefore those trials might not be powered to detect significant differences for the variables included in this meta-analysis. The primary outcomes of the trials are given in Tables 1–7   .

TABLE 1.

Characteristics of Studies in the Cardiothoracic Surgery Category

TABLE 5 (Continued).

Characteristics of Studies in the Orthopedic Surgery Category

TABLE 2.

Characteristics of Studies in the ENT Surgery Category

TABLE 3.

Characteristics of Studies in the Gynecologic Surgery Category

TABLE 3 (Continued).

Characteristics of Studies in the Gynecologic Surgery Category

TABLE 4.

Characteristics of Studies in the Laparoscopic Cholecystectomy Category

TABLE 5.

Characteristics of Studies in the Orthopedic Surgery Category

TABLE 7.

Characteristics of Studies in the Miscellaneous Surgery Category

TABLE 7 (Continued).

Characteristics of Studies in the Miscellaneous Surgery Category

Effect of Pregabalin on Primary Outcomes in all Surgical Categories

Two-Hour VAS pain scores

A total of 60 studies with a total of 2019 patients taking pregabalin and 2019 patients on the control treatment that reported pains scores at or around 2 hours postsurgery were included. Overall, pregabalin reduced VAS pain scores at 2 hours postsurgery (Hedge g and 95%CI, −0.970 [−1.197 to −0.743], z score −8.389, P < 0.0001), Figure 1.

FIGURE 1.

Forest plot for 2-hour pain scores.

Twenty-Four Hour VAS Pain Scores

A total of 57 studies with a total of 2033 patients taking pregabalin and 2033 patients on the control treatment that reported pains scores at 24 hours postsurgery were included. Overall, pregabalin reduced pain scores at 24 hours postsurgery (Hedge g and 95%CI, −0.442 [−0.665 to −0.220], z score −3.894, P < 0.0001), Figure 2.

FIGURE 2.

Forest plot for 24-hour pain scores.

Subgroup Analysis According to Dosing Regimen

Fifty-five studies that provided information on 24-hour pain scores were categorized according to whether a single dose (prior to surgery) or multiple doses (starting from the night, or days prior to surgery) were administered. There was no difference seen in 24-hour pain scores in these 2 subgroups. Pregabalin reduced pain scores at 24 hours postsurgery regardless of whether a single dose (Hedge g and 95%CI, −0.566 [−0.914 to −0.218], z score −3.191, P = 0.001), or multiple doses were administered (Hedge g and 95%CI, −0.322 [−0.571 to −0.073], z score −2.536, P = 0.011).

Twenty-Four Hour Morphine-Equivalent Consumption

Forty-six studies with a total of 1610 patients taking pregabalin and 1636 patients on the control treatment that reported morphine-equivalent consumption at 24 hours postsurgery were included. Overall, pregabalin reduced morphine-equivalent consumption at 24 hours postsurgery (Hedge g and 95%CI, −0.932 [−1.212 to −0.652], z score −6.519, P < 0.0001), Figure 3.

FIGURE 3.

Forest plot for 24-hour morphine-equivalent consumption.

Effect of Pregabalin on Primary Outcomes in Different Surgical Categories

Cardiothoracic Procedures

There were 4 studies^19–22 with a total of 107 patients taking pregabalin and 110 patients on the control treatment in this group. Pregabalin reduced the pain score at rest 2 hours postsurgery (P = 0.005). No significant difference was seen in pain score at rest at 24 hours postsurgery (P = 0.537) or morphine-equivalent consumption (P = 0.239), Figure 4 (Table 8).

FIGURE 4.

Forest plot for primary outcomes of studies under the cardiothoracic surgery category.

TABLE 6.

Characteristics of Studies in the Spine Surgery Category

TABLE 8.

Summary of Results According to Surgical Type

ENT Procedures

There were 6 studies^{12–14,23–25} with a total of 265 patients taking pregabalin and 266 patients on the control treatment in this group. Pregabalin reduced the pain score at rest 2 hours postsurgery (P < 0.0001) and pain score at rest at 24 hours postsurgery (P = 0.004). No statistically significant reduction in morphine-equivalent consumption was seen (P = 0.568), Figure 5.

FIGURE 5.

Forest plot for primary outcomes of studies under the ear, nose and throat surgery category.

Gynecologic Procedures

There were 17 studies^26–42 with a total of 980 patients taking pregabalin and 730 patients on the control treatment in this group. Pregabalin reduced the pain score at rest 2 hours postsurgery (P < 0.0001), pain score at rest at 24 hours postsurgery (P = 0.001), and the morphine-equivalent consumption (P = 0.001), Figure 6A.

FIGURE 6.

A,B Forest plot for primary outcomes of studies under the gynecologic surgery category.

Due to the heterogeneity within the gynecologic group, a subanalysis was performed on open hysterectomy studies only.^{27–33,36,37,41,42} There were 11 studies with a total of 468 patients taking pregabalin and 485 patients on the control treatment in this group. Pregabalin reduced the pain score at rest 2 hours postsurgery (P = 0.004), pain score at rest at 24 hours postsurgery (P = 0.003), and the morphine-equivalent consumption (P = 0.002), Figure 6B.

Laparoscopic Cholecystectomy Procedures

There were 6 studies^43–48 with a total of 273 patients taking pregabalin and 225 patients on the control treatment in this group. Pregabalin reduced the pain score at rest 2 hours postsurgery (P = 0.003), pain score at rest at 24 hours postsurgery (P = 0.036), and the morphine-equivalent consumption (P = 0.023), Figure 7.

FIGURE 7.

Forest plot for primary outcomes of studies under the laparoscopic cholecystectomy category.

Orthopedic Procedures

There were 12 studies^49–60 with a total of 430 patients taking pregabalin and 642 patients on the control treatment in this group. Pregabalin reduced the pain score at rest 2 hours postsurgery (P = 0.001), pain score at rest at 24 hours postsurgery (P = 0.013), and the morphine-equivalent consumption (P < 0.0001), Figure 8.

FIGURE 8.

Forest plot for primary outcomes of studies under the orthopedic surgery category.

Spine Procedures

There were nine studies^{61–65,67–69,88} with a total of 291 patients taking pregabalin and 332 patients on the control treatment in this group. Pregabalin reduced the pain score at rest 2 hours postsurgery (P = 0.001) and the morphine-equivalent consumption (P = 0.005). No significant difference was seen in pain score at rest at 24 hours postsurgery (P = 0.373), Figure 9.

FIGURE 9.

Forest plot for primary outcomes of studies under the spine surgery category.

Miscellaneous Procedures

There were 20 studies^{41,66,70–87} with a total of 1165 patients taking pregabalin and 884 patients on the control treatment in this group. Pregabalin reduced the pain score at rest 2 hours postsurgery (P < 0.0001) and the morphine-equivalent consumption (P = 0.006). No significant difference was seen in pain score at rest at 24 hours postsurgery (P = 0.422), Figure 10.

FIGURE 10.

Forest plot for primary outcomes of studies under the miscellaneous surgery category.

Common Adverse Effects of Pregabalin

Sedation Effects of Pregabalin

Thirty studies had included data on the sedative effects of pregabalin,^{13,14,21,22,25,29,30,32,37,39,43,50–52,55–57,59,60,62,65–69,77,79,82,85,87,88} with a total of 1147 patients taking pregabalin and 1170 patients on the control treatment. Subgroup analysis was performed on studies according to the surgical categories (number of studies) under cardiothoracic surgery (2), ENT surgery (3), gynecologic surgery (4), laparoscopic cholecystectomy (2), orthopedic surgery (7), spine surgery (6), and miscellaneous surgery (6). Data from George et al³⁰ could not be included in the analysis as there was no difference in sedation between the treatment and control group. With the exception of ENT surgery, laparoscopic cholecystectomy and gynecologic surgery, pregabalin was associated with sedation in all other surgical categories (overall OR and 95% CI, 2.144 [1.640–2.803], z score 5.574, P < 0.0001), Table 8.

Visual Disturbances

Fifteen studies had included data on incidence of visual disturbance (including blurred vision) after pregabalin administration,^{24,34,35,39,46,48,51,62,64,68,77,79,80,85,88} with a total of 491 patients taking pregabalin and 498 patients on control treatment. There were not enough studies under different surgical categories for subgroup analyses to be performed. Overall, pregabalin was found to be associated with an increased incidence of visual disturbances (OR and 95%CI, 6.215 [3.317–11.646], z score 5.702, P < 0.0001).

Nausea

Thirty-one studies had included data on nausea prevalence after pregabalin administration,^{13,21,24,26,28,30,33,37,44,46,47,50–52,54,56,57,59,60,63–65,67,68,70,71,77,85,86,88} with a total of 1067 patients taking pregabalin and 1038 patients on the control treatment. It was found that there was no difference in nausea incidence in cardiothoracic surgery, ENT surgery, gynecologic surgery, laparoscopic cholecystectomy, and spine surgery between pregabalin and control treatment groups. Pregabalin administration was associated with reduced incidence of nausea in miscellaneous surgery (OR and 95% CI, 0.138 [0.073–0.262], z score −6.085, P < 0.0001) and orthopedic surgery (OR and 95% CI, 0.586 [0.377–0.911], z score −2.373, P < 0.018). Overall results showed that pregabalin reduced postsurgical nausea (OR and 95% CI, 0.478 [0.365–0.626], z score −5.364, P < 0.0001).

Vomiting

A total of 22 studies provided information on vomiting incidence after pregabalin administration,^{13,21,23,25,44,46,47,50–52,54,56,59,60,63,65,67,68,70,71,77,85} with 826 patients treated with pregabalin and 816 on control treatment. The different surgical categories were cardiothoracic surgery (1), ENT surgery (3), laparoscopic cholecystectomy (3), miscellaneous surgery (4), orthopedic surgery (7), and spine surgery (4). In subgroup analysis, pregabalin was associated with reduced vomiting only after miscellaneous procedures (OR and 95% CI, 0.163 [0.073–0.368], z score −4.375, P < 0.0001), but pregabalin was found to be associated with reduced postsurgical vomiting in overall analysis (OR and 95% CI, 0.468 [0.328–0.668], z score −4.173, P < 0.0001).

Postsurgical Nausea and Vomiting

A total of 20 studies provided data on postsurgical nausea and vomiting (PONV) incidence after pregabalin administration,^{12,14,27,31,32,34,35,38,40,42,43,45,62,66,69,73,79,80,82} with 638 patients treated with pregabalin and 644 on control treatment. The different surgical categories were ENT surgery (2), gynecologic surgery (8), laparoscopic cholecystectomy (2), miscellaneous surgery (6), and spine surgery (2). In subgroup analysis, pregabalin was associated with reduced PONV in miscellaneous surgery only (OR and 95% CI, 0.528 [0.309–0.902], z score −2.339, P < 0.019) but pregabalin was found to be associated with reduced PONV in overall analysis (OR and 95% CI, 0.592 [0.415–0.845], z score −2.887, P < 0.004).

No evidence of publication bias was seen using funnel plot analysis with regard to 2- and 24-hour pain scores and 24-hour morphine-equivalent consumption (Supplementary Figures 2A to C), or with regard to adverse effects (Supplementary Figures 2D to h).

DISCUSSION

This present meta-analysis shows that perioperative administration of pregabalin significantly reduced VAS pain scores at 2 hours postsurgery in all surgical categories, and at 24 hours postsurgery in all surgical categories with the exception of cardiothoracic and spine procedures. Total morphine consumption at 24 hours postsurgery was significantly reduced in all surgical categories with the exception of cardiothoracic and ENT procedures. Adverse effects include significant sedation after pregabalin in cardiothoracic, orthopedic, spine, and miscellaneous procedures. PONV was significantly reduced after pregabalin in all, except miscellaneous procedures. Taken together, results of this meta-analysis show that pregabalin is useful in reducing postsurgical pain as well as reducing morphine consumption, with concomitant reduction in PONV.

It has long been recognized that different surgical procedures require procedure-specific pain management.^89–92 It is evident that the degree of pain experienced by patients after different surgical procedures is not universal, and even some laparoscopic approaches might result in unexpectedly high levels of postsurgical pain.^93,94 Moreover, the analgesic efficacy of different pain medications might also be different in different types of surgery. The analgesic efficacy of paracetamol is 2-fold less in orthopedic compared with dental procedures.⁹⁵ It has also been found that the analgesic efficacy between nonsteroidal anti-inflammatory agents and paracetamol depends on the magnitude of the surgical procedure.⁹⁶ In addition to differing analgesic effects of the same drug under different conditions, a 50% decrease in pain might have a different clinical relevance depending if it were a reduction from 4 to 2, or 8 to 4 on the VAS pain scale.⁹⁷ Therefore, specific recommendations for surgical procedures including abdominal hysterectomy, laparoscopic cholecystectomy, and total knee arthroplasty have been made.⁹⁸ It is in recognition that pain management should be procedure-specific that provided the insight to take this approach of subgroup analysis for this current investigation.

A previous meta-analysis of 11 RCTs¹⁰ concluded that presurgical pregabalin administration did reduce 2-hour pain scores and postsurgical opioid requirement. The authors divided the studies under investigation by pregabalin dose, <300 or ≥300 mg and found that the higher dose reduced opioid consumption more than the lower dose. Pregabalin also reduced opioid-related adverse effects such as vomiting, but the risk of visual disturbance was greater. Another recently conducted meta-analysis on 55 RCTs⁵ concluded that when all doses and administration regimens were combined, pregabalin was associated with a significant reduction in pain scores at rest and during movement and opioid consumption at 24 hours compared with placebo. Pregabalin was also associated with less postsurgical nausea, vomiting, and pruritus, although it was associated with higher incidence of sedation, dizziness, and visual disturbance. These previous meta-analyses have been criticized for not having investigated surgical specific-opioid consumption as different procedures will result in different opioid requirements.⁹⁹ Hence, this caveat has been addressed in the present meta-analysis. This meta-analysis is the first study to investigate the efficacy of pregabalin when used under different surgical procedures in acknowledgment that different surgical procedures result in variable pain intensity and different opioid requirements,⁹⁴ and that the efficacy of perioperative analgesia varies according to surgical type.⁹⁸ By identifying the types of surgery that would benefit from pregabalin, clinicians can improve efficiency in treating acute postsurgical pain and can better allocate resources.

This present meta-analysis is the first to show that the analgesic effect of perioperative pregabalin is procedure specific. With regard to the cardiothoracic procedure category, pain at 2 hours postsurgery was significantly lower in the pregabalin group, but no difference was seen at 24 hours postsurgery. It should be noted that only 2 studies showed data for 24-hour VAS pain scores, therefore there are insufficient data to draw definitive conclusions, and the only study showing reduction in morphine consumption after pregabalin did not show either 2-, or 24-hour VAS pain scores. No data on PONV were given and significant sedation was seen after pregabalin, so although overall, pregabalin appears to be efficacious for acute postsurgical pain in cardiothoracic procedures, caution should be exercised when deciding to use pregabalin.

In the ENT category, although both 2- and 24-hour postsurgical pain was shown to be reduced in the pregabalin group, there was no difference in total morphine-equivalent consumption at 24 hours between pregabalin and the control group. PONV is more common in patients undergoing ENT, compared with other procedures,¹⁰⁰ and as no difference was seen in either sedation or PONV, pregabalin can be recommended for use in ENT procedures.

There is strong evidence to recommend the use of pregabalin in gynecologic procedures, due to the large effects sizes with regard to pain reduction, and no evidence of increased sedation and PONV.

With regard to laparoscopic cholecystectomy, caution should be exercised when considering pregabalin, as although pain scores at 2 and 24 hours, and morphine-equivalent consumption are reduced, the OR seen for sedation was extremely high, even though, due to the heterogeneity of the studies, this was not statistically significant. Pain scores tend to be low after laparoscopic cholecystectomy procedures (not >5 on the VAS at 2 hours postsurgery according to the studies included here), and as pain reduction at 24 hours postsurgery and total morphine-equivalent consumption is modest in terms of effect-size, the risk–benefit ratio should be carefully considered.

Although pain scores at 2 and 24 hours, and morphine-equivalent consumption are reduced in orthopedic surgery, the reduction of pain scores at 2 hours is modest and sedation was significantly increased in the pregabalin group. The increased risk of sedation might be preferable when weighed with the significantly decreased morphine-equivalent consumed. Considering that many orthopedic procedures are performed in the elderly¹⁰¹ the risk of sedation might outweigh the benefit of modest decrease in pain scores.

With regard to spinal, and also miscellaneous surgeries, a large decrease in pain at 2 hours and total morphine consumption was seen, although there was no reduction in pain at 24 hours postsurgery. Considering the high incidence of sedation, in both spinal and miscellaneous surgical procedures, pregabalin should be used with caution.

It should be noted that although statistically significant reductions in the pain scores were noted in all surgical procedures in this meta-analysis, the magnitude of effect is relatively small. For example, in Bafna et al,²⁶ Balaban et al,⁴⁴ Aydogan et al,⁷² Eskandar and Ebeid,⁵¹ and Lee et al,⁵⁵ statistically significant decreases in pain scores at 2 hours postsurgery were reported, although the standard difference in mean pain scores between pregabalin and the control group was only <1 point on the VAS pain score. Studies on clinically significant decreases in VAS/NRS pain scores have demonstrated that an average decrease in pain score of at least 1.80 points on NRS scores or 1.3 to 2.8 points on VAS pain scores are required for the decrease to be considered clinically meaningful.^102,103 The reduction in pain scores demonstrated in the studies included in this meta-analysis may reach statistical significance, but might be too small to be considered of clinical significance.

An interesting finding from study by Mishriky et al⁵ was that a single preoperative dose was as effective as multiple doses, and that smaller doses (≤75 mg) were as effective as larger (300 mg) doses in terms of reducing opioid consumption. It was beyond the scope of this present meta-analysis to subdivide the studies according to surgical-type as well as dosages and dosing regimens, although an analysis of single versus multiple doses did not reveal any differences in efficacy regarding 2-hour postsurgical pain. Subgroup analyses performed in this present meta-analysis according to whether single or multiple doses of pregabalin were used showed a statistically significant reduction in 24-hour postsurgical pain for both single and multiple dose, contrary to previous studies.¹⁰ In particular, with regard to the gynecologic category, it was noted, that 8 out of 13 studies showed significant reduction in 24-hour postsurgical pain score, of which, 6 studies used a single-dose of pregabalin and 2 used multiple dose. The dose of pregabalin used included low dose (≤75 mg), intermediate dose (100–150 mg), and high dose (>150 mg). Sensitivity analysis data from this present meta-analysis do not show that higher doses were more effective at reducing pain scores when compared with lower doses (data not shown). There is no evidence from this current meta-analysis to recommend multiple dosing, or dosages >75 mg, in any of the surgical procedures that has investigated dosing.

Well established adverse effects of pregabalin are sedation, dizziness, and headache, and so pregabalin should be used with caution in an ambulatory setting.¹⁰⁴ As shown in previous meta-analyses, pregabalin is associated with increased incidence of visual disturbances and sedation; but reduced incidence of PONV.^5,10 Of the 15 studies included in this analysis that showed such an association, only 4 provided information on morphine-equivalent consumption. Two of these 4 studies showed a pregabalin-associated reduction in morphine-equivalent consumption, whereas the remaining 2 showed no reduction. Due to the limited data available, it is not possible to ascertain whether the reduction in incidence of PONV is due to a direct effect of pregabalin or a result of reduced opioid consumption. Opioids are considered the primary analgesic therapy in postsurgical pain, but are associated with many dose-related adverse effects such as sedation, respiratory depression, postsurgical nausea and vomiting, urinary retention, ileus, and constipation.¹⁰⁵ This meta-analysis shows that administration of pregabalin reduced morphine-equivalent consumption in most surgical categories, and looking at the effect size data show that there is up to 30% reduction. These data indicate pregabalin is useful to reduce opioid induced adverse effects, as seen by the reduced incidence of nausea and vomiting.

Meta-analyses have been conducted to assess the effects of perioperative gabapentin on postoperative pain,^106–108 and although all the studies concluded that perioperative gabapentin was able to reduce postsurgical pain and 24-hour morphine consumption, a recent meta-regression on RCTs of perioperative gabapentin that included 133 trials, found that these effects of gabapentin might have been overestimated by statistically significant small study effects.¹⁰⁹ Small study effects may also explain the difference in findings between our current meta-analysis and previously published work on pregabalin.

A problem inherent with meta-analyses using the random-effects model is the assumption that the effects underlying different studies are drawn from a normal distribution.¹¹⁰ This is seldom true, especially in the case of pain scores, which commonly show a skewed distribution. Much data used in this present meta-analysis were drawn from median, rather than the mean values required. Efforts were made to reduce the impact of clinical heterogeneity by analyzing data according to the type of surgery. Some studies are heterogeneous in themselves in that the investigators had included different surgical types in their own analysis.⁸⁴ Methodologic heterogeneity also exists in the assessment of pain and sedation. In addition to the commonly used VAS pain scores and NRS pain scores, which have been shown to correlate well,⁷ Roger Pain Scale was used in 1 study.⁷⁶ Similarly, with regard to assessing sedation, both the Ramsay Sedation Scale and Richmond Agitation Sedation Scale were used. Although these scales have been shown to correlate well,¹¹¹ some studies have neither stated with which method they have assessed sedation, nor at which time postsurgery, was the assessment carried out. Some studies have instead reported on either presence or absence of somnolence and these data were excluded in the analysis for sedation effects. It is noted here although that none of the studies included in this present meta-analysis were powered to assess pregabalin-associated adverse effects, as these were secondary outcomes of the studies.

In the setting of an ideal RCT, subjects are placed in a closely monitored environment, where their pain intensity is regularly assessed. Analgesia is provided on demand by the nursing staff in the form of nursing-controlled analgesia or delivered by the subjects themselves using patient-controlled analgesia (PCA). The pain intensity of both control and treatment group should therefore be titrated to similar levels, although total opioid consumption and time to first analgesic would differ between the 2 groups based on the effectiveness of the treatment. Limitations certainly exist for both nursing-controlled analgesia and PCA in providing adequate analgesia. For the former, inadequacy of nursing staff can result in delay in delivering analgesics; for the latter, malfunctioning, poor initial titration, or incorrect setup of the PCA instruments can also prevent timely delivery of analgesics. Such limitations, however, would apply to both control and treatment group in a well-conducted trial and the pain scores of both control and treatment group will therefore be similar. It is proposed here that pain scores should only be 1 of the primary outcomes in such trials, whereas the more pertinent parameters would be changes in analgesic consumption and in time to first analgesic requirement.

Pain is not only affected by gender, age, and psychologic well-being, but also by polygenetic elements. The current list of genetic polymorphisms that may affect the action of analgesics is growing rapidly, but 1 of the enzyme systems of high relevance to opioids is the cytochrome P450 system.¹¹² As it has been shown that polymorphisms that affect opioid metabolism are found in up to 30% of the general population,¹¹² future clinical trials utilizing opioid consumption as an outcome could take genetic variability into consideration. The fact that none of the studies included here have factored in the genetic variability in opioid metabolism brings in another layer of heterogeneity, especially when an increase in opioid requirement in 1 or 2 patients can have substantial impact in the overall results.

Out of the 74 studies assessed in this meta-analysis, only 12 investigated the effects of pregabalin on chronic (≥3 months) postsurgical pain.^{19,21,29,30,33,41,61,63,65,76,79} Chronic postsurgical pain is an underexplored area and more studies are required to assess the efficacy of pregabalin in this regard.

CONCLUSIONS

In conclusion, the analgesic efficacy and adverse effects of pregabalin might not be similar under all surgical categories. Although sedation may be increased, especially in cardiothoracic, spinal, and miscellaneous procedures, this was not seen in ENT, gynecologic, or laparoscopic cholecystectomy procedures. Two-hour VAS scores were reduced in all procedures, but effect sizes varied greatly. Taken together, this meta-analysis shows strong evidence that consideration for the use of pregabalin in postsurgical pain should be procedure-specific.