Perioperative pregabalin in pain management after laparoscopic cholecystectomy : A quality of evidence assessment

(1)

Perioperative pregabalin in pain management

after laparoscopic cholecystectomy – version 2

A quality of evidence assessment

Masi Koskinen

Degree project

The School of medicine ¨

Orebro university Sweden 5th _{June 2015}

Supervisor: Maija-Liisa Kalliom¨aki (MD), Anaesthesiology, The School of Medicine, University of Tampere, Finland

(2)

Abstract

Background The usefulness of an estimate depends on the quality of ev-idence for that estimate. The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system offers a means for authors of systematic reviews to assess the quality of evidence. Acute pain after laparoscopic cholecystectomy is a major clinical challenge. Pregabalin, an anticonvulsant belonging to the gabapentinoids, is thought to relieve postop-erative pain by counteracting central sensitisation. In this systematic review, I applied the GRADE approach to effects of perioperative pregabalin in pain management after laparoscopic cholecystectomy.

Methods and materials I performed a database search for relevant ran-domised clinical trials. The found articles were screened for inclusion. I extracted and tabulated study and participant characteristics, as well as re-sults of the following patient-important outcomes: postoperative (1) pain scores, (2) opioid consumption, and (3) incidence and severity of nausea, vomiting, and sedation. The quality of evidence for these effects was rated high unless lowered according to the criteria outlined by the GRADE system.

Results The database search yielded eleven records of which six were in-cluded. A statistically significant reduction in the pain scores and the opioid consumption was reported in four and three trials, respectively. Effects on the nausea, vomiting, and sedation were varied. Flaws in the body of evi-dence lowered the quality of evievi-dence for these effects to very low.

Conclusions The data available from the included articles provide little evidence for effects of pregabalin on the chosen outcomes after laparoscopic cholecystectomy.

(3)

List of Figures

1 Pain pathways . . . 19

2 PRISMA flow chart . . . 20

List of Tables

1 Study characteristics . . . 21 2 Participant characteristics . . . 22 3 Results . . . 23 4 Evidence profile 1 . . . 24 5 Evidence profile 2 . . . 24

(5)

Abbreviations

ASA . . . American Society of Anesthesiologists BMI . . . Body Mass Index

CENTRAL . . . Cochrane Central Register of Controlled Trials CINAHL . . . Cumulative Index to Nursing and Allied Health

Lit-erature

COX-2 . . . Cyclooxygenase-2

GABA . . . Gamma-Amino Butyric Acid

GRADE . . . Grading of Recommendations Assessment, Develop-ment, and Evaluation

IASP . . . International Association of Study of Pain ICTRP . . . International Clinical Trials Registry Platform IQR . . . Inter-Quartile Range

LC . . . Laparoscopic Cholecystectomy

MEDLINE . . . Medical Literature Analysis and Retrieval System Online

NA . . . Not Available

(6)

NSAID . . . Nonsteroidal Anti-Inflammatory Drugs VAS . . . Visual Analogue Scale

VRS . . . Verbal Rating Scale

PACU . . . Post-Anaesthesia Care Unit PCA . . . Patient-Controlled Analgesia

PICO . . . Patient, Intervention, Comparison, and Outcome PMC . . . PubMed Central

PONV . . . Postoperative Nausea and Vomiting QoE . . . Quality of Evidence

RCT . . . Randomised Clinical Trial

SBU . . . Statens Beredning f¨or medicinsk Utv¨ardering, [The Swedish Council on Health Technology Assessment] SD . . . Standard Deviation

(7)

1. Background

1.1 Elective laparoscopic cholecystectomy

Laparoscopic cholecystectomy (LC) is a surgical procedure where the gall-bladder is removed by operating through small incisions made on the ab-domen. The area of surgery is visualised with a scope attached to a video camera, and a workspace around the gallbladder (pneumoperitoneum) is cre-ated by inflating the peritoneal cavity with an inert gas (carbon dioxide in routine use). [1, 2]

A German surgeon, professor Erich M¨uhe, was the first person in the world to perform an LC in 1985 [3]. In the early 1990s, laparoscopy started to gain popularity in the elective treatment of symptomatic cholelithiasis [4].

Cholelithiasis (formation of gallstones) is a common condition. In Sweden, approximately 30 % men and 50 % women aged over 40 years have gall-stones. Although most of the carriers (80 %) are asymptomatic, symptoms and complications lead to over 10 000 cholecystectomies in Sweden every year. [5]

Laparoscopy is preferred to open surgery, unless the patient has an elevated risk of complications [6]. Laparoscopic cholecystectomy in itself is associ-ated with fewer postoperative complications and less pain compared to open cholecystectomy. The pain is nevertheless a major clinical problem during

(8)

the early recovery in the hospital. [2]

1.2 Acute postoperative pain

Pain is defined as ”[a]n unpleasant sensory and emotional experience as-sociated with actual or potential tissue damage, or described in terms of such damage.” [7] According to the International Association of Study of Pain (IASP) more than 80% of patients undergoing surgery experience acute postoperative pain. Most of these patients receive inadequate pain relief. [8] Pain is a subjective experience and difficult to measure and analyse. Pain after LC is assessed with the same scales as postsurgical pain in general: the Visual Analogue Scale (VAS) and the Verbal Rating Scale (VRS). [9] When using the VAS, the patient is asked to rate his or her level of pain on a continuum between no pain (0) and the worst imaginable pain (10) by marking it on a plain 10-cm line with only the anchor points 0 and 10 at the opposite ends. When the VRS is used, the patient rates his or her pain verbally a 5-point scale: no pain (0); slight pain (1); moderate pain (2); severa pain (3); unbearable pain (4). The wording can differ somewhat (for example, slight versus mild, or unbearable versus horrible) [10, 11].

Laparoscopic cholecystectomy gives rise to three sorts of pain: incisional (parietal), deep abdominal (visceral), and shoulder pain (referred visceral). The incisional pain is dominant, and most intense on the day of surgery. Shoulder pain, which is the least intense of the three and usually resolves within the first 24 hours after surgery, is connected with use of pneumoperi-toneum during surgery. [2]

The incisional pain is inflammatory by nature. The damaged tissue releases chemical substances (prostaglandines and cytokines among others) which activate pain reseptors (nociceptors), and also lead to a further release of

(9)

algogenic (pain producing) and inflammatory mediators. The nociceptors convert their activation into neural impulses (transduction). Sensory path-ways (via A-delta and C nerve fibres) transmit these impulses to the dorsal horn of the spinal cord, where the impulses are modulated while being re-layed further (via second-order fibres) to the thalamus and finally to the cortex where the pain is consciously perceived (Figure 1). [12]

The tissue injury can increase the responsiveness of the pain pathways in two ways: by lowering the threshold of the nociceptors (peripheral sensitisa-tion), and by increasing the excitability of the spinal neurons in an activity-dependent manner (central sensitisation) [12]. Both these processes can give rise to persistent pain, which is experienced by 10–50 % of patients after common surgery. Two to ten per cent suffer from severe chronic pain. [13]

1.3 Postoperative analgesia and gabapentinoids

Pain is most effectively controlled by combining several analgesics, each of which targets one of the above-mentioned processes in the sensory pathways (transduction, transmission, modulation, and perception; Figure 1). The ra-tionale behind this multimodal approach is that such a combination relies less on any particular mechanism of action and may avoid the side effects that a high dose of any individual agent would give. [12] The acute postoper-ative pain after laparoscopic cholecystectomy, like postsurgical pain in gen-eral, is treated with a combination of opioids, nonsteroidal anti-inflammatory drugs (NSAIDs), cyclooxygenase-2 (COX-2) inhibitors, acetaminophen, and gabapentinoids [14].

Gabapentinoids are a group of anticonvulsant drugs. The two major gabapenti-noids are pregabalin, and its predecessor gabapentin. They are structural analogues of gamma-amino butyric acid (GABA), but do not act on GABA receptors. Instead, they bind to alpha-2-delta subunits in voltage-gated

(10)

cal-cium channels, and reduce calcal-cium influx into presynaptic neurons. As a con-sequence, the release of several excitatory neurotransmitters (glutamate, no-radrenaline, and substance P) is reduced, which is believed to be the mecha-nism behind their analgesic effects. In particular, gabapentinoids are thought to counteract injury-induced up-regulation of the alpha-2-delta subunits in the dorsal horn neurons, and hence prevent central sensitisation. [15]

Pregabalin has many advantages over gabapentin. Unlike gabapentin, pre-gabalin has highly predictable and linear pharmacokinetics. It is absorbed proportional to dose, and has high bioavailability. Maximal plasma concen-tration is achieved rapidly (after one hour). These properties make prega-balin attractive in clinical practise. [16]

Pregabalin and gabapentin are currently licensed for use as analgesics only in treatment of neuropathic pain [17, 18]. Several studies have showed that gabapentin can be used safely and efficiently for postoperative pain after var-ious types of surgery (including LC) [19]. The results concerning pregabalin are yet inconsistent. [19–21]

1.4 Quality of evidence (GRADE)

Grading of Recommendations Assessment, Development, and Evaluation (GRADE) is a tool for authors of systematic reviews and clinical practice guidelines. It consists of two parts: rating quality of evidence (QoE) and grading strength of recommendation. The latter rests upon the results ac-quired from the preceding quality assessment. [22] GRADE does not neces-sarily give reproducible results but the judgments are explicit and transpar-ent. [23]

Several methods exist for rating quality of evidence [24]. GRADE is outcome centric meaning that rating is made for chosen outcomes rather than for

(11)

individual studies. Hence, two outcomes within a single study can have different quality ratings. [22]

To understand GRADE, it is essential to distinguish an estimate of an out-come from the quality of evidence for that estimate. An anecdotal example gives an illustrative depiction of the difference: A man working in a weather bureau states that he has figured out that there is a 40 % probability of rain, and a 10 % chance that he knows what he is talking about [25]. The point is that poor quality makes an estimate useless.

Several organisations worldwide have adopted the GRADE approach (World Health Organization, The Cochrane Collaboration, and British Medical Jour-nal, to name but a few) [26]. The Swedish Council on Health Technology Assessment (SBU) uses GRADE in their quality assessments, and The Na-tional Board of Health and Welfare (Socialstyrelsen) applies it in guideline development [27, 28]. In Finland, The Finnish Medical Society Duodecim advocates the use of the GRADE approach [29].

1.5 Aims of this study

The explicit question I am trying to answer in this study is the following: What is the quality of evidence for any treatment effect on

- acute postoperative pain,

- cumulative opioid consumption, and

- incidence and severity of postoperative sedation, nausea and vomiting after laparoscopic cholecystectomy, when the patient is given at least one preoperative dose of pregabalin?

(12)

2. Methods and materials

I made a quality of evidence assessment according to GRADE. The process began by defining the Population, Intervention, Comparison, and patient-important Outcomes (PICO) [22]. After a systematic literature search and inclusion of relevant studies [22], I extracted and tabulated study details, demographic characteristics, and results. I did not make any meta-analysis to combine results from individual studies into single outcome estimates. Instead, I focused on examining limitations that would undermine any such estimate.

The PICO for this systematic review was

(P) Adults undergoing laparoscopic cholecystectomy. (I) Dose and frequency of pregabalin administration. (C) Placebo group.

(O) Postoperative pain scores, opioid consumption, sedation, nausea, and vomiting.

2.1 Search of literature

I searched for relevant studies in the following databases: PubMed (PubMed Central [PMC] and MEDLINE), the Cochrane Central Register of Controlled Trials (CENTRAL), Scopus, and Cumulative Index to Nursing and Allied

(13)

Health Literature (CINAHL). No date or language restrictions were used in the search, but only full-text articles in English, Spanish, Swedish, and Finnish were included. The last search was performed on the 15th of April 2015. Additional studies were looked for in the bibliographies of retrieved articles and two recent reviews (Eipe et al. [20], and Mishriky et al. [21]). My search strategy was to start with the term pregabalin and narrow down the scope of the search using the terms cholecystectomy, laparoscopic, post-operative pain, and placebo. The search phrase consisted of plain words sep-arated by ”AND”. No descriptors or qualifiers in form of Medical Subject Headings (MeSH) or similar were attached. Instead, the search engines were allowed to look for matches in all fields to capture even unindexed studies (for example those marked as in process or as supplied by publisher in PubMed). I also searched clinical trial registries (the World Health Organization Inter-national Clinical Trials Registry Platform [WHO ICTRP], ClinicalTrials.gov, and ClinicalTrialsRegister.eu) with the same search words for indications of on-going and past trials that have not led to a publication.

In this systematic review, I included only prospective, randomised, placebo-controlled trials studying postoperative pain or opioid consumption as pri-mary outcome in adults (over 18 years) after a laparoscopic cholecystectomy, where participants in the intervention group(s) were given at least one pre-operative dose of pregabalin. Active-controlled trials were excluded.

2.2 Opioid consumption

The opioid consumption data were converted to morphine equivalents. The conversion ratios from the different opioids to equianalgesic morphine doses were 0.01:1 for fentanyl, 3:1 for ketorolac, and 10:1 for pethidine. [30–32]

(14)

2.3 The GRADE process

The patient-important outcomes for the GRADE process were the pain scores and the cumulative opioid consumption early (approximately 2 h) and late (24 h) after the surgery, and the incidence and severity of postoperative nausea, vomiting, and sedation during the follow-up period.

The quality of evidence for a treatment effect was rated based on the entire body of evidence (trials reporting comparable data after similar interven-tions). When the evidence came from a single trial, the QoE was rated as equal to the quality of that trial.

I examined the five factors that can rate down the quality of evidence: limita-tions to the study design, inconsistency between the results, indirectness of the evidence, impresicion of the reported results, and publication bias. Ran-domised placebo-controlled clinical trials (RCT) gave by default high-quality evidence. Any one of the above-mentioned limitation lowered the rating by one step on scale high, moderate, low, and very low [22].

Limitations: I examined whether the number of recruited participants was based on a power analysis; the participants were followed up and included in the analysis according to the intention-to-treat principle; whether drop-outs were given an explanation; and the randomisation and blinding were adequate (adapted from [24]).

Inconsistency: Similar effects were expected for similar interventions [33]. Differences in effect sizes larger than the confidence intervals were considered a sign of inconsistency (adapted from [34]). Consistency could not be evalu-ated for treatment effects reported by single trials. The unknown consistency was judged as inconsistency.

Indirectness: If the results came from research that included patients, in-terventions and outcomes that were different from those of interest in this

(15)

review, I considered the evidence indirect [35].

Imprecision: Lack of reporting the uncertainty of results was considered a sign of imprecision.

Publication bias: I looked for unpublished trials in the clinical trials reg-istries and for signs of commercial funding in the included trials [36]. I did not take into consideration those unpublished trials, where the recruitment of participants was completed for less than three years ago. I also exam-ined whether conflicts of interest, and financial ties (industry influence) were disclosed [36].

2.4 Ethics

The Declaration of Helsinki outlines ethical principles regarding medical re-search on humans [37]. I examined the included publication to see, whether

- an ethics committee had approved the research protocol (§23),

- all patients had given a written consent to participate in the trial (§25), and

- an adequate rescue analgesic (”the best proven intervention”) was avail-able to all patients regardless of their allocation to the intervention group or the placebo group (§33).

The number in the parenthesis refers to the relevant paragraph in the Dec-laration [37].

(16)

3. Results

3.1 Search of literature

I assessed eleven trials for inclusion in this study. Six [38–43] of them met the inclusion criteria, and were selected for further analysis. The reasons for the exclusion of the remaining five studies is presented in Figure 2. Two [42, 43] of the included trials were recorded in clinical trials registries.

3.2 Study design

All the included studies were labeled as prospective, randomised, double-blind placebo-controlled trials. A randomisation method was presented in all but the work by Chang et al. [41] On the other hand, Chang et al. were the only ones to specify that the group allocation was blinded for the surgeons [41]. Table 1 shows whether the medication and the outcome assessment were blinded.

Two articles reported the location and the setting of the trial [39, 40]. Every article contained a section explaining the use of the statistical methods in analysing numerical data (demographic, pain scores, opioid consumption, incidence of side-effects etc.). All but one study [39] included a power

(17)

anal-ysis.

Three of the six articles do not categorise their outcomes as primary [39,42,43] and secondary [39, 41, 42].

All the studies had an ethics committee approval, and every patient had given a written consent for participation. A rescue analgesic was included in every protocol.

3.3 Demographic characteristics

All but two [39,41] presented a participant flow chart, which shows how many participants were assessed for eligibility, enrolled, randomised, followed up and analysed.

The majority of participants in every trial except for one [38] were females (Table 2). The mean weight among the participants in that trial was signifi-cantly lower than in two other trials [39, 40] that reported the mean weight. Sarakatsianou et al. [43] did not reach their minimum number of participants yielded by the power analysis owing to many drop-outs.

The duration of anaesthesia was mentioned in two studies [38, 40], and the time to extubation in one study [43].

Propofol or thiopental with fentanyl were used for inducing anaesthesia, which was maintained using volatile anaesthetics except for one study [38], where propofol infusion was used. Tracheal intubation was facilitated with a relaxant.

The intraoperative fentanyl consumption differed from study to study. In the studies of Chang et al. [41] and Sakatsianou et al. [43], patients were given 25 mg of pethidine and 1 g of acetaminophen, respectively, at the end of the

(18)

surgery. In one trial patients received a local anaesthetic ”around the gall bladder bed and the laparoscopic port sites” [42].

3.4 Pain and opioid consumption

Statistically significant reductions in the pain scores were observed by Agar-wal et al. (0 to 24 h), Balaban et al. (0 to 2 h), Peng et al. (0 to 90 min), and Sarakatsianou et al. (0 to 24 h) [38, 39, 42, 43].

Statistically significant treatment effects on the opioid consumption were observed by Agarwal et al. (at 2 hours), Balaban et al., and Bekawi et al. (at 24 hours) [38–40]. The last-mentioned reported also the incidence of intake of rescue medicines during the first 24 hours after the operation: In the placebo group all 30 patients needed pethidine, while in the pregabalin group only 8 patients (Number Needed to Treat [NNT] was 1.4).

Patients in the study of Balaban et al. experiences least pain of all both in the treatment group and in the placebo group, and showed a significant reduction in the opioid consumption that was proportional to the given dose of pregabalin (Table 3) [39].

3.5 Side-effects

Bekawi et al. reported less postoperative nausea and vomiting (PONV) in the intervention group (150 mg PGL) compared to the placebo group [40]. None of the other trials found any difference in the incidence or severity of PONV.

Balaban et al. and Chang et al. reported increased postoperative sedation early (at 15 min and 2 h, respectively) in the intervention group (300 mg

(19)

PGL) [39, 41]. No other trial showed any statistically significant difference in the incidence or severity of sedation at any time point.

3.6 Quality of evidence

Table 4 shows the quality of evidence for treatment effects for which the body of evidence came from at least two trials reporting comparable data (as means or medians with a measure of uncertainty) after similar interventions. For any other effect, for which the evidence came from a single trial, the QoE is presented in Table 5.

I lowered the rating from the initial value high to very low based on the following observations:

Limitations. Balaban et al. [39] do not include any power analysis in their report, and Sarakatsianou et al. [43] failed to include the minimum number of patients estimated with their power analysis. Only Agarwal et al. [38] and Peng et al. [42] incorporated the intention-to-treat principle in their analyses. Chang did not disclose the method of randomisation [41], and none of the others mentioned whether the surgeons were blinded for the allocation of the patients.

Imprecision. Sarakatsianou et al. [43] did not report confidence intervals for their results.

Inconsistency. Only two trials reported comparable results on pain and opioids [39, 41]. The early VAS scores in the placebo groups differed between the two trials as did the effect sizes after one dose of 300 mg of pregabalin. The early cumulative opioid consumptions in the placebo groups were within a standard deviation from each other but the effect sizes differed tenfold. The side-effects were not comparable owing to the heterogeneity of the in-terventions and the duration of the follow-up period.

(20)

Publication bias. Two studies disclosed their conflicts of interest and fi-nancial ties: One had none [43] and the other declared having received travel and research grants and an honorarium from Pfizer Canada [42]. The ties to Pfizer in the aforementioned trial [42] and the non-disclosure of funding in four other trials [38–41] increases the risk of publication bias.

The following observations did not alter the quality-rating:

Indirectness. The selection criteria limited the inclusion of the trials to those studying the patients, interventions, and outcomes of interest in this systematic review. Thus, no indirectness was detected.

Publication bias. The search of the clinical trial registries showed up two unpublished trials. One of the trials (IRCT201312041617N8) was expected to complete recruitment of participants by the 21st _{of January 2015, and} the other one (ACTRN12612000447853) was still recruiting (assessed on the 30th _{of April 2015). Thus, there were no signs of publication bias in form of} unpublished trials.

(21)

4. Discussion

In this systematic review, I examined RCTs studying effects of perioperative pregabalin on postoperative (1) pain scores, (2) opioid consumption, and (3) side-effects in patients undergoing LC. In particular, I assessed the quality of evidence for these effects applying the GRADE approach. My assessment showed that using only the data available from the included articles the evidence for any of the effect is of very low quality.

The low quality rating does not mean that no treatment effect exists or have been shown. Four of the six trials reported a statistically significant reduction in the pain scores [38, 39, 42, 43], and three in opioid consumption [38–40] among patients who received pregabalin. Moreover, the incidence and severity of the side-effects was found to differ between the intervention and placebo groups [39–41].

Instead of refuting the existence of any effect, the low quality rating suggests that very little confidence should be put in the estimates based on these trials. The true effects might be significantly different from the reported estimates. [44]

My results can be compared with those of a recent systematic review from Eipe and colleagues [20]. Their review studies the analgesic efficacy of prega-balin in the perioperative setting using the GRADE approach. Their meta-analysis includes a subgroup of abdominal laparoscopic surgery (comprising LC, gynecological and urological surgery, etcetera) for which they show a

(22)

sta-tistically significant reduction in the opioid consumption but no reduction in the pain scores. The quality of evidence for these findings is rated moderate, and low, respectively.

The higher quality of evidence compared to my study might be attributed to the inclusion of the other types of abdominal laparoscopic surgery than LC. In fact, the subgroup comprises of 22 RCTs of which only five [38, 39, 41–43] concern LC. Thus, the review of Eipe et al. presents an extension to a more general population as compared to my study. The review shows that pooling data from this larger population raises the quality of evidence for a positive effect of pregabalin on the opioid consumption but still leads to no certainty about whether any effect on the pain scores exists.

Another recent review from Mishriky and colleagues [21] includes a similar meta-analysis showing a reduction both in the pain scores and the opioid consumption. However, this review does not assess the quality of evidence, which leaves the importance of the results unclear.

In my study, the findings concerning the PONV and sedation were inconclu-sive. Both the above-mentioned reviews show a statistical significant reduc-tion in the PONV among pregabalin-treated patients. Mishriky et al. [21] find also an increase in sedation, while Eipe et al. [20] do not corroborate this finding in their meta-analysis. Neither of the reviews estimates the quality of evidence for these side-effects which, again, leaves the importance of the findings unclear.

The inconsistencies in the results regarding the opioid consumption might be partly due to methodological reasons. McQuay and colleagues have ques-tioned the use of analgesic consumption as an outcome measure for the ef-ficacy of pain management. Although intuitive, they deem this approach a possible source of contradictory result between trials. According to their find-ings, this kind of comparison is valid only when the pain scores in the treat-ment groups are similar. [45] That is, the pain should be managed equally

(23)

well both in the intervention group and the placebo group to make a possible difference in the analgesic consumption a reliable measure of efficacy of the intervention.

On the other hand, it is not straightforward to evaluate when pain levels are similar between patients. According to DeLoach and colleagues, a post-operative perceptual-cognitive impairment caused by the anaesthesia during surgery leads to an imprecision of ±20 mm in any single VAS score [10]. This is a rather wide range, and could complicate any attempt to control for differences in the postoperative pain levels.

There is considerable diversity in the demographic characteristics (age, weight, sex ratio, etcetera) among the included trials (Table 2). Almost all the pa-tients in the trial of Bekawi and colleagues [40] were younger than the av-erage patient in the trial of Balaban and colleagues [39]. About two thirds of the patients in the study of Agarwal and colleagues [38] were men while the patients are mostly women in all the other studies. Also the duration of surgery and the intraoperative fentanyl consumption varied substantially from around 40 to 90 minutes, and 1 to 6 µg/kg, respectively. All this het-erogeneity might add to the difficulty of discerning effects of the individual pregabalin-interventions.

The greatest limitation to my study is probably that it does not have a quan-titative meta-analysis as a basis for the quality assessment. I did not contact the authors of the trials to ask for the original data which, in retrospect, hampered much of my analysis.

To my knowledge, this review is the first one to focus exclusively on laparo-scopic cholecystectomy studying the quality of evidence for effects of prega-balin in the perioperative setting. Other reviews on this topic merge several types of surgery [20, 21, 46–48], and do not include a quality of evidence as-sessment (except for [20]). A recent Cochrane review examines pregabalin and gabapentin as a single category of anticonvulsants [49].

(24)

In conclusion, the data available from the included articles provide evidence of very low quality for effects of perioperative pregabalin on the chosen out-comes after LC. The low quality rating does not rule out pregabalin having beneficial of harmful effect, but rather calls for more rigour in reporting trials, and for conducting a proper meta-analysis focusing on laparoscopic cholecystectomy.

(25)

Appendices

Figure 1: The four processes in the pain pathway: transduction, transmis-sion, modulation, and perception. Nociceptors convert noxious stimulation into neural impulses (transduction). Primary sensory neurons transmit im-pulses to the dorsal horn of the spinal cord, where the imim-pulses are modulated (by decending inhibitory signals) and relayed to second-order neurons which carry the impulses to the thalamus and to the sensory cortex for percep-tion. [12]

(26)

Id e n tifica ti o n Scre e n in g El ig ib ili ty In cl u d e d 27 of records identified through database searching: CENTRAL (n=7), CINAHL (n=3), MEDLINE (n=6), SCOPUS (n=11) 0 of additional records identified through other

sources

11 of records after duplicates removed

6 of full-text articles assessed for eligibility

5 of records excluded: 3 reviews 1 active placebo 1 hemodynamic responses as primary outcome 6 of studies included in qualitative synthesis

Figure 2: PRISMA flow chart showing inclusion and exclusion of retrieved studies. Adapted from [50]

(27)

Table 1: Characteristics of the included trials regarding primary and sec-ondary outcomes, and blinding.Study characteristics

Author Outcomes Blinded medication by hospital pharmacy Blinded outcome assessment

Primary How and when

assessed Secondary

How and when assessed Agarwal (2008) 1. Severity of postoperative pain!

!

2. Postoperative fentanyl requirement How: VAS 100 mm!

rest (static) and! coughing (dynamic).!

!

When: on arrival to

PACU (0 h) and then every 2 h for 24 h. PONV, headache, sedation, and respiratory depression How: Ordinal scales (Ramsay and others)!

!

Respiratory depression defined as ventilation frequency ≤8 bpm and O2 sat. <90%!

!

When: NA Yes Yes Balaban (2012) 1. Severity of postoperative pain!

!

2. Postoperative opioid requirement How: VAS, 0 to 10!

!

When: on arrival to PACU (0) and 15, 30, 60, 90, 120 (min), and 3, 4, 6, 8, 10, 12, and 24 (h) after surgery. Somnolence, diizziness, confusion, and ataxia How: Ramsay Sedation Scale!

!

When: on arrival to PACU (0) and 15, 30, 60, 90, 120 (min), and 3, 4, 6, 8, 10, 12, and 24 (h) after surgery. NA Yes Bekawi (2014) 1. Postoperative pain scores!

!

2. 24-hour pethidine consumption after surgery How: VAS, 0 to 10, Max. score at different time intervals.!

!

When: Recorded at 0, 2, 4, 6, 12, 24 (h) postoperatively. PONV, sedation, dizziness, and somnolence

How: PONV using

a 4-point scale.! When: 6 and 12 h after surgery!

!

How: Sedation using Numeric Sedation Scale (NSS) 5-point scale! When: 2, 6, 12 and 24 h after surgery No, the intervention group was given Lyrica capsules Yes Chang (2009) 1. "overall incidence! of PLSP and [its] VRS score" How: 'VRS', 0 to 10, !

!

When: at 2, 4, 12, 24, and 48 h postoperatively

Surgical pain, times to first rescue analgesia, and total rescue ketorolac.!

!

PONV, sedation, dry mouth, lack of concentration, and blurred vision How: VRS, 0 to 10, !

!

How: Sedation as a score over 2 on a 5-point ordinal scale.!

!

When: at the same

timepoints.

Yes NA

Peng (2010)

1. Postoperative pain score at rest!

!

2. and on active movement (in PACU only) How: NRS, 0 to 10!

!

When: every 30 min

in PACU, but later not specified Analgesic consumption, side-effects, and recovery profil How: Patients completed a diary!

!

When: day 1, 2, 7 Yes NA Saraka-! tsianou! (2013) 1. Postoperative pain score!

!

2. Postoperative 24-h opioid consumption How: VAS, 0 to 10!

!

Postoperative PCA morphine consumption!

!

When: 1, 8, 16, and 24 h postoperatively PONV, sedation, Itching, headache, dizziness, blurred vision, lack of concentration, shoulder pain, and respiratory depression How: PONV, 4-point scale!

!

Ramsay sedation scale!

!

When: NA Yes NA

(28)

Table 2: Characteristics of participants in the included trials. Intervention groups labelled as placebo or as the given pregabalin dose in milligrammes.Patient characteristics

Author Dose (mg)

Patient flow Demographic characteristics

# patients! (recruted / analysed) Explanation for drop-outs Males (%) ASA

status I! Age (years)

Weight (kg) BMI! (kg/m2) Duration of surgery (min) Intraoperative fentanyl (ug)

mean ± SD (min-max) or median (min-max) mean ± SD

Agarwal (2008) Placebo 30 / 29 Conversion to open surgery!

!

Re-exploration 60% NA! 44.6 (22-69) 55.7 ± 9.1 NA 91.3 ± 22.9 212.6 ± 27.4 150 30 / 27 77% 46.6 (25-76) 56.2 ± 10.1 88.7 ± 22.7 218.7 ± 30.3 Balaban (2012)

Placebo 30 / 30 None 20% NA 51.4 ± 15.7 74.9 ± 13.7 NA 38.9 ± 9.4 Induction:! 1.5 ug/kg 150! 300 30 / 30! 30 / 30 23%! 27% 54.5 ± 14.7! 52.7 ± 11.8 77.1 ± 14.5! 74.8 ± 13.9 42.4 ± 12.2! 41.7 ± 7.9 Bekawi (2014) Placebo 30 / 30 None 10% 83% 37.6 ± 7.9 (25-59) 73.5 ± 7.4 (59-87) NA 58.1 ± 5.5 (50-70) Induction:! 1 ug/kg 150 30 / 30 17% 80% 37.6 ± 6.9 (27-52) 77.9 ±10 (60-96) 60.9 ± 7 (50-80) Chang (2009) Placebo 40 / 38 Conversion to open surgery!

!

Inappropriate supplemental analgesics NA NA NA NA NA Recorded but not reported None 300 40 / 39 Peng (2010) Placebo 55 / 46 Protocol violation!

!

Questionnaire not completed or returned 28% NA 47 (21-65) NA 29 ± 4 NA 185 ± 57 50! 75 55 / 48! 55 / 48 29%! 42% 46 (22-65)! 43 (22-65) 27 ± 5! 29 ± 6 187 ± 65! 195 ± 73 Saraka-! tsianou! (2013) Placebo 25 / 20 Convertion to open!

!

40% 55% 54 (20-77) NA 30 (23-35) 45 (30-67) Induction:! 3 ug/kg!

!

300 25 / 20 40% 35% 53 (18-72) 28.5 (22-39) 42 (30-72) 22

(29)

Table 3: Results retrieved from the included trials. Interventions labelled as placebo or as the given pre-gabalin dose in milligrammes. Standard deviations (SD) marked with a star (*) retrieved from the meta-analysis of Mishriky et al. [21]. Analgesics, VAS, opioid consumption, side-effects

Author Dose (mg) Administration time Postoperative analgesics VAS score!

Mean ± SD, Median (IQR)

Cumulative opioid consumption in morphine equivalents (mg)!

Mean ± SD, Median (IQR)

Early (2 h) Effect Late (24 h) Effect Early (2 h) Effect Late (24 h) Effect

Agarwal (2008) Placebo Preop: 1 h!

!

Postop: -Fentanyl (PCA)!

!

Bolus: 20 ug! Lockout: 5 min! max: 2 ug/(kg h) Max, 0-4h:! 4.0 (3.8) Max, 12-24h:! 3.5 (4.0) NA 75.8 ± 9.9 150 Max, 0-4h:! 3.0 (2.0) -1 Max, 12-24h:! 2.0 (2.0) -1.5 55.5 ± 12.5 _{-27 %} Balaban (2012) Placebo Preop: 1 h!

!

Postop: -Fentanyl 25 ug intravenously if VAS ≥ 5 1.0 ± 0.8 NA _{7.3 ± 1.6}_* NA 150! 300 0.7 ± 0.8! 0.3 ± 0.8 -0.3! -0.7 3.3 ± 1.4 *! 1.5 ± 1.0 * -56 %! -80 % Bekawi (2014) Placebo Preop: 2 h!

!

Postop: 12 h and twice daily for 2 days Pethidine 1 mg/kg intramuscularly every 6 hours if VAS ≥ 3 or requested!

!

Diclofenac 75 mg intramuscularly if VAS > 4 or requested At 0 h:! 3.7 ± 1.7 1.1 ± 0.6 NA 7.5! min: 5.9! max: 8.7 150 At 0 h:! 2.4 ± 1.3 -1.3 0.6 ± 0.6 _-0.5 0! min: 0! max: 9.6 -100 % Chang (2009) Placebo Preop: 1 h!

!

Postop: 12 h after first dose

Ketorolac 30 mg intravenously on request 4.9 ± 1.8! 1.8 ± 1.4 5.8 ± 6.4 13.7 ± 13.8 300 5.2 ± 1.9 _+0.3 2.1 ± 1.8 _+0.3 5.4 ± 5.0 _{-8 %} 12.8 ± 13.7 _{-7 %} Peng (2010) Placebo Preop: 1 h!

!

Postop: 12 h and 24 h after first dose Fentanyl bolus of 25– 50 ug every 5-10 min if needed in PACU.!

!

Acetaminophen 325 mg and codeine 30 mg 1–2 tablets orally every 4–6 h as needed (max. 12 tablets per day).

2 (1-5) 4 (2-5)! 3.0 (2.4-3.5) NA 50! 75 2 (1-5)! 1 (0-2)! 0! -1 3 (2-5)! 3 (2-6)! -1! -1 2.3 (1.2-3.4)! 1.6 (1.2-2.8) -23 %! -47 % Saraka-! tsianou (2013)

Placebo Preop: night before and 1 h!

!

Postop: -Morphine (PCA)!

!

Bolus: 0.5 mg Lockout: 15 min Max: 12 mg in 4 h!

!

At 1 h:! 5 (NA) 3 (NA) At 1 h:! 0.5 (NA) 3.75 (NA) 300 At 1 h:! 2 (NA) -3 1 (NA) _-2 At 1 h:! 0 (NA) -100 % 1.5 (NA) _{-60 %} 23

(30)

Table 4: Evidence profile for comparable results after similar interventions (adapted from [22]).

GRADE evidence profile: 300 mg pregabalin 1 hour before laparoscopic cholecystectomy

Quality assessment Summary of findings

# patients

# of studies Limitations Inconsistency Indirectness Imprecision Publication bias Placebo Pregabalin Effect size Quality

Pain at 2 h (assessed on scale 0-10)

2 Yes Yes No No Yes 68 69 Balaban: -0.7!

Chang: +0.3

Very low

Cumulative opioid consumption at 2 h (assessed as morphine equivalents)

2 Yes Yes No No Yes 68 69 Balaban: -5.8 mg (-80 %)!

Chang: -0.4 mg (-8 %)

Very low

Table 5: Evidence profile for all treatment effects reported by single trials (adapted from [22]).

GRADE evidence profile: all treatment effects reported by single trials

Author Quality assessment Quality

Limitations Inconsistency Indirectness Imprecision Publication bias

Agarwal (2008) yes yes no no yes very low

Balaban (2012) yes yes no no yes very low

Bekawi (2014) yes yes no no yes very low

Chang (2009) yes yes no no yes very low

Peng (2010) yes yes no no yes very low

(31)

Bibliography

[1] Brunicardi FC, Andersen D, Billiar T, Dunn D. Schwartz’s Principles of Surgery. 10th ed. McGraw-Hill Education; 2014.

[2] Bisgaard T, Kehlet H, Rosenberg J. Pain and convalescence after la-paroscopic cholecystectomy. The European journal of surgery = Acta chirurgica. 2001 Feb;167(2):84–96.

[3] Reynolds W. The first laparoscopic cholecystectomy. JSLS : Journal of the Society of Laparoendoscopic Surgeons / Society of Laparoendoscopic Surgeons. 2001 Jan;5(1):89–94.

[4] Wilson RG, Macintyre IM, Nixon SJ, Saunders JH, Varma JS, King PM. Laparoscopic cholecystectomy as a safe and effective treat-ment for severe acute cholecystitis. BMJ (Clinical research ed). 1992 Aug;305(6850):394–6.

[5] Läkemedelsverket. Läkemedelsboken - Gallvägs- och pankreassjukdomar [Internet]; 2014. [Accessed 6th May 2015]. Available from: http://www. lakemedelsboken.se/b3_mag_gallvagpankreas_2013fm10.html.

[6] Kullman E. Gallsten [Internet]; 2014. [Accessed 6th May 2015]. Avail-able from: http://www.internetmedicin.se/page.aspx?id=478. [7] Merskey H, Bogduk N, editors. Classification of Chronic Pain. 2nd ed.

(32)

[8] International Association of Study of Pain (IASP). Why acute pain? [Internet]; 2014. [Accessed 6th May 2015]. Available from: http:// www.iasp-pain.org/GlobalYear/AcutePain.

[9] Wills VL, Hunt DR. Pain after laparoscopic cholecystectomy. The British journal of surgery. 2000 Mar;87(3):273–84.

[10] DeLoach LJ, Higgins MS, Caplan AB, Stiff JL. The visual analog scale in the immediate postoperative period: intrasubject variability and correlation with a numeric scale. Anesthesia and analgesia. 1998 Jan;86(1):102–6.

[11] Leino KA, Kuusniemi KS, Lertola KK, Olkkola KT. Comparison of four pain scales in patients with hip fracture or other lower limb trauma. Acta anaesthesiologica Scandinavica. 2011 Apr;55(4):495–502.

[12] Kelly DJ, Ahmad M, Brull SJ. Preemptive analgesia I: physiological pathways and pharmacological modalities. Canadian journal of anaes-thesia = Journal canadien d’anesth´esie. 2001 Nov;48(10):1000–10. [13] Kehlet H, Jensen TS, Woolf CJ. Persistent postsurgical pain: risk factors

and prevention. Lancet. 2006 May;367(9522):1618–25.

[14] The prospect working group. Procedure specific postoperative pain man-agements - Laparoscopic cholecystectomy [Internet]; 2015. [Accessed 10th May 2015]. Available from: http://www.postoppain.org/. [15] Weinbroum AA. Non-opioid IV adjuvants in the perioperative period:

pharmacological and clinical aspects of ketamine and gabapentinoids. Pharmacological research : the official journal of the Italian Pharmaco-logical Society. 2012 Apr;65(4):411–29.

[16] Ben-Menachem E. Pregabalin pharmacology and its relevance to clinical practice. Epilepsia. 2004 Jan;45 Suppl 6:13–8.

(33)

[17] FASS. Lyrica [Internet]; 2015. [Accessed 9th May 2015]. Avail-able from: http://www.fass.se/LIF/product?userType=0&nplId= 20040607000717.

[18] FASS. Neurontin [Internet]; 2015. [Accessed 9th May 2015]. Avail-able from: http://www.fass.se/LIF/product?userType=0&nplId= 20000519000137.

[19] Mitra S, Khandelwal P, Roberts K, Kumar S, Vadivelu N. Pain relief in laparoscopic cholecystectomy–a review of the current options. Pain prac-tice : the official journal of World Institute of Pain. 2012 Jul;12(6):485– 96.

[20] Eipe N, Penning J, Yazdi F, Mallick R, Turner L, Ahmadzai N, et al. THE PERIOPERATIVE USE OF PREGABALIN FOR ACUTE PAIN - A SYSTEMATIC REVIEW AND META-ANALYSIS. [Published Ahead of Print]. Pain. 2015 Mar;.

[21] Mishriky BM, Waldron NH, Habib AS. Impact of pregabalin on acute and persistent postoperative pain: a systematic review and meta-analysis. British journal of anaesthesia. 2015 Jan;114(1):10–31.

[22] Guyatt G, Oxman AD, Akl EA, Kunz R, Vist G, Brozek J, et al. GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. Journal of clinical epidemiology. 2011 Apr;64(4):383–94.

[23] Guyatt G, Oxman AD, Sultan S, Brozek J, Glasziou P, Alonso-Coello P, et al. GRADE guidelines: 11. Making an overall rating of confidence in effect estimates for a single outcome and for all outcomes. Journal of clinical epidemiology. 2013 Mar;66(2):151–7.

[24] Guyatt GH, Oxman AD, Kunz R, Vist GE, Falck-Ytter Y, Sch¨unemann HJ. What is ”quality of evidence” and why is it important to clinicians?

(34)

BMJ (Clinical research ed). 2008 May;336(7651):995–8.

[25] Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, Alonso-Coello P, et al. GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ (Clinical research ed). 2008 Apr;336(7650):924–6.

[26] GRADE working group - Organizations [Internet]; 2014. [Accessed 6th May 2015]. Available from: http://www.gradeworkinggroup.org/ society/.

[27] SBU. Utvärdering av metoder i hälso- och sjukv˚arden: En handbok. 2nd ed. Stockholm: Statens beredning för medicinsk utvärdering (SBU); 2014.

[28] Socialstyrelsen. Hur vi tar fram v˚ara vetenskapliga

un-derlag [Internet]; 2015. [Accessed 5th May 2015].

Available from: http://www.socialstyrelsen.se/

nationellariktlinjerfordepressionochangest/omriktlinjerna/ saharriktlinjernatagitsfram/vetenskapligaunderlag.

[29] Jousimaa J, Liira H, Liira J, Komulainen J. [Judgment of the level of evidence and strength of recommendations in healthcare according to the GRADE working group]. Duodecim; l¨a¨aketieteellinen aikakauskirja. 2010 Jan;126(16):1936–43.

[30] FASS. Fentanyl Meda [Internet]; 2014. [Accessed 9th May 2015]. Avail-able from: http://www.fass.se/LIF/product?userType=0&nplId= 19950322000012.

[31] FASS. Toradol [Internet]; 2013. [Accessed 9th May 2015]. Avail-able from: http://www.fass.se/LIF/product?userType=0&nplId= 19900629000024.

(35)

[32] FASS. Petidin Meda [Internet]. 2013;[Accessed 9th May 2015]. Avail-able from: http://www.fass.se/LIF/product?userType=0&nplId= 19741206000040.

[33] Guyatt GH, Oxman AD, Kunz R, Atkins D, Brozek J, Vist G, et al. GRADE guidelines: 2. Framing the question and deciding on important outcomes. Journal of clinical epidemiology. 2011 Apr;64(4):395–400. [34] Guyatt GH, Oxman AD, Kunz R, Woodcock J, Brozek J, Helfand

M, et al. GRADE guidelines: 7. Rating the quality of evidence– inconsistency. Journal of clinical epidemiology. 2011 Dec;64(12):1294– 302.

[35] Guyatt GH, Oxman AD, Kunz R, Woodcock J, Brozek J, Helfand M, et al. GRADE guidelines: 8. Rating the quality of evidence–indirectness. Journal of clinical epidemiology. 2011 Dec;64(12):1303–10.

[36] Guyatt GH, Oxman AD, Montori V, Vist G, Kunz R, Brozek J, et al. GRADE guidelines: 5. Rating the quality of evidence–publication bias. Journal of clinical epidemiology. 2011 Dec;64(12):1277–82.

[37] World Medical Association Declaration of Helsinki: ethical princi-ples for medical research involving human subjects. JAMA. 2013 Nov;310(20):2191–4.

[38] Agarwal A, Gautam S, Gupta D, Agarwal S, Singh PK, Singh U. Eval-uation of a single preoperative dose of pregabalin for attenEval-uation of postoperative pain after laparoscopic cholecystectomy. British journal of anaesthesia. 2008 Nov;101(5):700–4.

[39] Balaban F, Ya˘gar S, Özgök A, Ko¸c M, Güllapo˘glu H. A randomized, placebo-controlled study of pregabalin for postoperative pain intensity after laparoscopic cholecystectomy. Journal of clinical anesthesia. 2012 May;24(3):175–8.

(36)

[40] Bekawi MS, El Wakeel LM, Al Taher WMA, Mageed WMA. Clinical study evaluating pregabalin efficacy and tolerability for pain manage-ment in patients undergoing laparoscopic cholecystectomy. The Clinical journal of pain. 2014 Nov;30(11):944–52.

[41] Chang SH, Lee HW, Kim HK, Kim SH, Kim DK. An evaluation of perioperative pregabalin for prevention and attenuation of postopera-tive shoulder pain after laparoscopic cholecystectomy. Anesthesia and analgesia. 2009 Oct;109(4):1284–6.

[42] Peng PWH, Li C, Farcas E, Haley A, Wong W, Bender J, et al. Use of low-dose pregabalin in patients undergoing laparoscopic cholecystec-tomy. British journal of anaesthesia. 2010 Aug;105(2):155–61.

[43] Sarakatsianou C, Theodorou E, Georgopoulou S, Stamatiou G, Tzovaras G. Effect of pre-emptive pregabalin on pain intensity and postoperative morphine consumption after laparoscopic cholecystectomy. Surgical en-doscopy. 2013 Jul;27(7):2504–11.

[44] Balshem H, Helfand M, Sch¨unemann HJ, Oxman AD, Kunz R, Brozek J, et al. GRADE guidelines: 3. Rating the quality of evidence. Journal of clinical epidemiology. 2011 Apr;64(4):401–6.

[45] McQuay HJ, Poon KH, Derry S, Moore RA. Acute pain: combina-tion treatments and how we measure their efficacy. British journal of anaesthesia. 2008 Jul;101(1):69–76.

[46] Moore RA, Straube S, Wiffen PJ, Derry S, McQuay HJ. Pregabalin for acute and chronic pain in adults. The Cochrane database of systematic reviews. 2009 Jan;(3):CD007076.

[47] Engelman E, Cateloy F. Efficacy and safety of perioperative pregabalin for post-operative pain: a meta-analysis of randomized-controlled trials. Acta anaesthesiologica Scandinavica. 2011 Sep;55(8):927–43.

(37)

[48] Baidya DK, Agarwal A, Khanna P, Arora MK. Pregabalin in acute and chronic pain. Journal of anaesthesiology, clinical pharmacology. 2011 Jul;27(3):307–14.

[49] Gurusamy KS, Vaughan J, Toon CD, Davidson BR. Pharmacological interventions for prevention or treatment of postoperative pain in people undergoing laparoscopic cholecystectomy. The Cochrane database of systematic reviews. 2014 Jan;3:CD008261.

[50] Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ (Clinical research ed). 2009 Jan;339(jul21 1):b2535.

Perioperative pregabalin in pain management after laparoscopic cholecystectomy : A quality of evidence assessment