Tapentadol para el dolor musculoesquelético crónico en adultos
Resumen
Antecedentes
El dolor musculoesquelético crónico es un trastorno prevalente y una causa principal de discapacidad y ausentismo laboral en todo el mundo. Los opiáceos se utilizan con frecuencia para el tratamiento del dolor crónico, aunque los efectos adversos a menudo restringen sus efectos beneficiosos a largo plazo. El tapentadol es un opiáceo y un inhibidor de la recaptación de norepinefrina que puede provocar una menor incidencia (y gravedad) de efectos adversos en comparación con otros opiáceos fuertes.
Objetivos
Determinar la eficacia, la seguridad y la tolerabilidad del tapentadol de liberación prolongada para el dolor moderado a grave durante al menos tres meses por cualquier causa musculoesquelética.
Métodos de búsqueda
Se realizaron búsquedas en las bases de datos electrónicas (el Registro Cochrane Central de Ensayos Controlados (Cochrane Central Register of Controlled Trials, CENTRAL), MEDLINE, EMBASE y Web of Science) hasta marzo de 2014, sin restricciones de idioma, así como en los registros de ensayos y las listas de referencias de los estudios recuperados. Se estableció contacto con los fabricantes del fármaco para obtener mayor información.
Criterios de selección
Ensayos controlados aleatorizados (ECA) del tapentadol en pacientes con dolor musculoesquelético crónico, en comparación con placebo o control activo.
Obtención y análisis de los datos
Dos autores de la revisión, de forma independiente, seleccionaron los ensayos para inclusión, evaluaron el riesgo de sesgo de los estudios incluidos y extrajeron los datos. Se realizaron dos metanálisis para las comparaciones de tapentadol de liberación prolongada versus placebo y tapentadol de liberación prolongada versus control activo (oxicodona). Se utilizaron los modelos de efectos aleatorios cuando hubo heterogeneidad y los modelos de efectos fijos cuando no la hubo. Además, se realizaron análisis de subgrupos. El resultado primario de eficacia fue el control del dolor evaluado por el cambio en las puntuaciones de la intensidad del dolor y la tasa de respuesta (al menos un alivio del dolor del 50%). El resultado primario de seguridad fue la tasa de retiros debido a los efectos adversos.
Resultados principales
Cuatro ECA de diseño paralelo de calidad moderada que incluyeron a 4094 pacientes con osteoartritis o dolor lumbar, o ambos, cumplieron los criterios de inclusión. Tres ensayos fueron estudios de fase III con un seguimiento de 12 semanas y el cuarto ensayo fue un estudio de seguridad abierto con un seguimiento de 52 semanas. Todos los ensayos fueron controlados con oxicodona y tres también fueron controlados con placebo. Dos ensayos incluyeron a pacientes con osteoartritis de rodilla, uno evaluó a pacientes con dolor lumbar y uno reclutó a ambos. Todos los estudios informaron la última observación realizada (LOCF, por sus siglas en inglés) como el método de imputación. Se solicitaron a los fabricantes los análisis imputados de la observación inicial realizada (BOCF, por sus siglas en inglés) y cualquier dato no publicado, pero rechazaron la solicitud. Dos de los cuatro estudios controlados con oxicodona y uno de los tres estudios controlados con placebo no proporcionaron datos sobre la tasa de pacientes que respondieron al tratamiento. Se consideró que dos estudios tuvieron un alto riesgo de sesgo.
En comparación con placebo, el tapentadol se asoció con una reducción media de 0,56 puntos (intervalo de confianza [IC] del 95%: 0,92 a 0,20) en la Numerical Rating Scale (NRS) de 11 puntos a las 12 semanas y con un aumento de 1,36 (IC del 95%: 1,13 a 1,64) en el riesgo de responder al tratamiento (número necesario a tratar para lograr un resultado beneficioso adicional [NNTB] 16; IC del 95%: 9 a 57, durante 12 semanas). Hubo una heterogeneidad de moderada a alta en las estimaciones de los resultados de eficacia. El tapentadol se asoció con un aumento de 2,7 veces (IC del 95%: 2,05 a 3,52) en el riesgo de interrupción del tratamiento debido a los efectos adversos (número necesario a tratar para lograr un resultado perjudicial adicional [NNTD] 10; IC del 95%: 7 a 12, durante 12 semanas).
En comparación con la oxicodona, los datos agrupados indicaron una reducción de 0,24 puntos (IC del 95%: 0,43 a 0,05) en la intensidad del dolor a partir del inicio en la NRS de 11 puntos. Los dos estudios que evaluaron la tasa de pacientes que respondieron al tratamiento mostraron un aumento no significativo de 1,46 (IC del 95%: 0,92 a 2,32) en el riesgo de respuesta al tratamiento entre los pacientes tratados con tapentadol. El tapentadol se asoció con una reducción del riesgo del 50% (IC del 95%: 42% a 60%) de interrupción del tratamiento debido a los efectos adversos (NNTB 6; IC del 95%: 5 a 7, durante 12 semanas). El tapentadol también se asoció con una reducción del 9% (IC del 95%: 4% a 15%) en el riesgo general de efectos adversos (NNTH 18; IC del 95%: 12 a 35, durante 12 semanas) y con una reducción no significativa del 43% (IC del 95%: 33% a 76%) en el riesgo de efectos adversos graves. Hubo una heterogeneidad de moderada a alta en la mayoría de las estimaciones de eficacia (excepto en el caso del resultado primario) y de seguridad. El análisis de subgrupos mostró una mejoría mayor con tapentadol entre los pacientes con osteoartritis de la rodilla y entre los resultados agrupados de los estudios de calidad más alta y con un período de seguimiento más corto, aunque no hubo diferencias estadísticamente significativas en el tamaño del efecto entre estos subgrupos.
Conclusiones de los autores
El tapentadol de liberación prolongada se asocia con una reducción en la intensidad del dolor en comparación con placebo y oxicodona. Sin embargo, la significación clínica de los resultados no está clara debido a las siguientes razones: diferencia modesta entre las intervenciones en los resultados de eficacia, alta heterogeneidad en algunas comparaciones y resultados, altas tasas de retiros, falta de datos del resultado primario en algunos estudios y la imposibilidad de utilizar la BOCF como método de imputación. El tapentadol se asocia con un perfil de seguridad y tolerabilidad más favorable que la oxicodona.
PICO
Resumen en términos sencillos
Tapentadol para el dolor musculoesquelético crónico en adultos
El dolor en los huesos, las articulaciones y los músculos es muy común y a menudo puede ser persistente. Se espera que el 50% de la población general presente este tipo de dolor durante al menos tres meses o más a lo largo de la vida. Este trastorno se denomina dolor musculoesquelético crónico.
Los opiáceos son un tipo de fármaco analgésico potente que se utiliza para tratar a los pacientes que presentan dolor crónico moderado a grave. Los pacientes que toman estos fármacos suelen presentar efectos secundarios, incluidos efectos secundarios graves.
Esta revisión Cochrane tuvo como objetivo evaluar la efectividad (reducción de la intensidad del dolor) y la seguridad del tapentadol (un nuevo opiáceo) en el dolor musculoesquelético crónico de moderado a grave, en comparación con placebo (un fármaco falso) y otros fármacos que actúan en dicho dolor.
En marzo de 2014 se realizó una búsqueda bibliográfica de estudios que compararan el tapentadol con placebo u otros fármacos en pacientes adultos con dolor musculoesquelético.
Se encontraron cuatro estudios que compararon el tapentadol con placebo u oxicodona (otro opiáceo) en 4094 adultos.
Hubo evidencia de calidad moderada de que tres de diez pacientes tratados con tapentadol tuvieron al menos una reducción del dolor del 50% (respondieron al tratamiento), mientras que sólo dos de diez pacientes tratados con oxicodona y placebo respondieron al tratamiento.
También hubo evidencia de calidad moderada de que los pacientes tratados con tapentadol tuvieron un mayor riesgo de retirarse del ensayo debido a los efectos secundarios en comparación con placebo (dos de diez pacientes tratados con tapentadol y uno de diez pacientes tratados con placebo). En los pacientes tratados con oxicodona, cuatro de diez se retiraron debido a los efectos secundarios. El estreñimiento, las náuseas y los vómitos y el picor (prurito) fueron menores con el tapentadol que con la oxicodona, pero no hubo diferencias en cuanto a la fatiga, el insomnio, la somnolencia ni el dolor de cabeza.
El efecto clínico beneficioso general del tapentadol en el dolor musculoesquelético crónico moderado a grave encontrado en los ensayos clínicos aleatorizados es relativamente pequeño (una conclusión común encontrada en todos los ensayos de opiáceos para el dolor crónico). Se necesitan más estudios para determinar qué pacientes con dolor musculoesquelético crónico se beneficiarían más de este nuevo opiáceo.
Authors' conclusions
Background
Description of the condition
The International Association for the Study of Pain (IASP) defines chronic pain as pain without apparent biological value that has persisted beyond the normal tissue healing time (usually taken to be three months) (IASP 1986). This definition is widely accepted, although other definitions have also been proposed. The exact time point that differentiates between acute and chronic pain (three to six months) is controversial (Marcus 2000), and it can be argued that every pain that is not chronic is acute. Chronic pain is usually categorised as cancer‐related or non‐cancer‐related in origin, and most treatment guidelines are specifically directed to one or the other. In the case of chronic musculoskeletal pain, it includes various painful local or regional musculoskeletal disorders (such as low back pain and osteoarthritis), but also people with chronic widespread pain.
Musculoskeletal pain is extremely prevalent, with virtually every adult having at least one episode in the last year and about 50% of adults in the last month (Crombie 1999). Most of these episodes are brief but recurrent or chronic pain is also common (Papageogiou 1995; Urwin 1998). The exact prevalence of chronic musculoskeletal pain is difficult to estimate. The prevalence rates reported in individual studies vary widely, from 10% (Buskila 2000) to over 50% (Andersson 1993). This variation is due to different case definitions, methods used to collect data, time periods, and populations studied. The most important risk factors for chronic musculoskeletal pain are age (the most common causes of chronic musculoskeletal pain have a higher incidence and prevalence in older people, sex (some causes are more common in one gender), and the presence of degenerative disease processes (eg osteoarthritis) (Crombie 1999; LeResche 2000).
Moderate‐to‐severe chronic musculoskeletal pain is an important cause of physical disability and work absence, carrying a huge economic and social cost (Breivik 2006; McAlindon 1992; Reisine 2001). Therefore, it represents a relevant problem (although at different levels) to patients, health professionals, the healthcare system, and society.
Description of the intervention
Opioids started to be used for pain relief thousands of years ago when opium analgesic properties were discovered. Nowadays, opioids are well established as pharmacological treatment options for both oncological and non‐oncological chronic moderate‐to‐severe pain (Jadad 1995; Trescot 2008a), and the use of chronic opioid therapy for chronic non‐cancer pain has increased substantially (Chou 2009). Commercially available strong opioids for clinical use include morphine, hydromorphone, levorphanol, oxycodone, and fentanyl. The primary activity of these drugs is at the μ‐opioid receptors (μ‐agonists), which are found primarily in the central nervous system (brainstem and medial thalamus) (Trescot 2008b).
Although the strength of the actual evidence‐based recommendations differ according to the person's condition (e.g. malignant versus non‐malignant pain) and for specific drugs, they all agree that strong opioids should only be considered when other pharmacological options have failed to control pain adequately (Chou 2009; Trescot 2008a). However, up to 50% of people still receive inadequate pain relief. In these situations, pain relief is very challenging due to the variety of pain mechanisms involved and adverse effects of the available analgesics (Freynhagen 2006).
Tapentadol is a novel opioid analgesic that is currently being evaluated in clinical trials for chronic moderate‐to‐severe pain.
How the intervention might work
Tapentadol is a centrally acting opioid analgesic with two complementary mechanisms of action, namely μ‐opioid receptor agonism and noradrenaline re‐uptake inhibition. In comparison with other strong opioids, tapentadol is a weak opioid agonist due to the lower binding affinity to opioid receptors. The synergistic analgesic effect of the noradrenaline re‐uptake inhibition contributes to an opioid‐sparing effect. These properties may reduce the incidence (and severity) of adverse effects associated with μ‐opioid agonism, such as constipation, nausea, and vomiting.
The analgesic effects of tapentadol are independent of metabolic activation and have no known active metabolites. Therefore, there are claims that tapentadol may be associated with a low inter‐individual efficacy variations and drug‐drug interactions, while providing broad pain inhibition due to its complementary modes of action.
Tapentadol should be considered in clinical practice as an option to other strong opioids in the third step of the World Health Organization (WHO) Pain Relief Ladder (WHO 2010).
Why it is important to do this review
The Food and Drug Administration (FDA) approved tapentadol extended‐release (ER) formulation in chronic pain in September 2011 for the US market, following the European Medicines Agency's (EMA) decentralised decision in 2010. Several clinical trials have been conducted and others are in progress, with the aim of assessing the efficacy and safety of tapentadol for non‐oncological chronic pain. Therefore, it is relevant to evaluate in a systematic review the benefits and risks of this new analgesic agent in people with moderate‐to‐severe musculoskeletal chronic pain, which is the most frequent cause of non‐oncological chronic pain.
Objectives
To determine the efficacy, safety, and tolerability of tapentadol extended‐release for moderate‐to‐severe pain for at least three months for any musculoskeletal cause.
Methods
Criteria for considering studies for this review
Types of studies
All randomised controlled trials (RCTs), published and unpublished, where randomisation was explicit and appropriate. We included both parallel and cross‐over studies.
Types of participants
Participants included adults, both male and female, aged 18 years or older. We considered all participants with moderate‐to‐severe chronic musculoskeletal pain of any cause.
We considered any person with pain for at least three months to have chronic pain. We classified people as having moderate pain if they scored "moderate" in the 4‐point verbal categorical rating scale (VRS), 4 to 6 (inclusive) in the 11‐point numerical rating scale (NRS), or 40 to 60 mm (inclusive) in the visual analogue scale (VAS). People were classified as having severe pain if they score "severe" in the 4‐point VRS, 7 or greater in the 11‐point NRS, or more than 60 mm in the VAS (Breivik 2000; Farrar 2001).
Types of interventions
We included all studies comparing tapentadol with placebo, other strong analgesic agents, both, or other active controls.
Two oral tapentadol formulations are available: immediate release and ER. Since the scope of this Cochrane review was chronic musculoskeletal pain, we considered the ER formulation in doses of 100 to 500 mg a day, which corresponds to the dosage range approved by the FDA and EMA.
Types of outcome measures
Primary outcomes
-
Primary efficacy outcome: pain control assessed by change in pain intensity scores and responder's rate (at least 50% pain relief).
-
Primary safety outcome: withdrawal rate due to adverse effects.
For primary efficacy outcome assessment, we considered the VAS, categorical scales, or other validated assessment tools measuring pain intensity.
We evaluated primary outcomes by direct participant report. We considered and extracted data for the intention‐to‐treat (ITT) population for quantitative analysis.
We planned to use baseline‐observation‐carried‐forward (BOCF) imputation method for primary efficacy outcome analysis because this method gives a more accurate reflection of results with zero pain relief in situations of trial withdrawal. However, none of the published studies provided this information and the drug manufacturer denied our requests to access unpublished data (e‐mail correspondence).
Therefore, we considered last‐observation‐carried‐forward (LOCF) imputation method for primary efficacy outcome estimate.
Secondary outcomes
-
Patient Global Impression of Change (PGIC).
-
Quality of life scores using validated scales.
-
Requirements for breakthrough analgesia.
-
Other relevant outcome measure, for example functional health status and well‐being (assessed by the 36‐item Short Form (SF‐36) health survey, the EQ‐5D health survey or the Western Ontario and McMaster Universities Index of Osteoarthritis scale), and sleep evaluation.
-
Withdrawals rate (global).
-
Adverse effects (global, serious, and most frequently reported).
Search methods for identification of studies
Electronic searches
We searched the following electronic databases: the Cochrane Central Register of Controlled Trials (CENTRAL; The Cochrane Library; Issue 2 of 12, 2014); MEDLINE and MEDLINE in Process (Ovid) 1950 to 17 March 2014; EMBASE (Ovid) (1974 to 17 March 2014); Web of Science (SCI‐EXPANDED, SSCI, CPCI‐S, CPCI‐SSH) searched to March 2014. We listed the search strategies we used in Appendix 1. We attempted to identify all relevant studies irrespective of language.
Searching other resources
On the 18 March 2014, we screened the bibliographies of studies and review articles in this field for other potentially relevant studies. We also searched trials registries for details of unpublished and ongoing trials: the metaRegister of controlled trials (mRCT) (www.controlled-trials.com/mrct), clinicaltrials.gov, and the WHO International Clinical Trials Registry Platform (ICTRP) (apps.who.int/trialsearch/).
Data collection and analysis
Selection of studies
Following the search for relevant studies, we read the titles and abstracts of the identified potential eligible trials. If it was clear from the abstract that the trial did not meet the selection criteria we excluded the study. If it was unclear from the abstract whether the trial met the selection criteria, we read the full report. We obtained the full reports of all studies that appeared to meet the selection criteria.
Two review authors independently assessed the studies and, using the above inclusion criteria, decided which articles to include. We resolved any differences in opinion with respect to inclusion of the studies by consensus, and if necessary by consultation with a third review author. We documented the justification for excluding studies (see Characteristics of excluded studies table).
Data extraction and management
We designed a systematic data extraction form to record the following information from the studies.
-
Publication details.
-
Participant population details (including nature of pain).
-
Interventions.
-
Outcome measures.
-
Analgesic results.
-
Adverse effect results.
-
Quality of life scores.
-
Participant preference.
-
Withdrawals.
-
Quality assessment criteria.
Two review authors extracted data and cross‐checked for accuracy. A third review author helped to resolve disagreements by consensus.
Assessment of risk of bias in included studies
Two review authors independently assessed the quality of studies and risk of bias in accordance with the domain‐based evaluation proposed by Cochrane, using Cochrane's 'Risk of bias' assessment tool (Higgins 2011), and the definitions suggested in the Authoring or assessing a Cochrane Protocol, Review, or Review Update for the Cochrane Pain, Palliative and Supportive Care (PaPaS) Review Group (AUREF 2012) as well as the "ACTINPAIN Writing Group of the IASP Special Interest Group on Systematic Reviews in Pain Relief; Cochrane Pain, Palliative and Supportive Care Systematic Review Group Editors. "Evidence" in chronic pain‐establishing best practice in the reporting of systematic reviews (Moore 2010).
Measures of treatment effect
Continuous data
We used the summary statistic of mean difference (MD) and 95% confidence interval (CI) between treatment groups.
Dichotomous data
The summary statistic was the risk ratio (RR) and 95% CI between groups. We also calculated the number needed to treat (or to harm) for an additional beneficial (or harmful) outcome (NNTH/NNTB) (and 95% CI) from meta‐analysis estimates (adjusted odds ratios (OR)) taking into account the baseline risk, defined as the percentage of participants with events in the control arm.
Unit of analysis issues
All included trials used the same dose range for tapentadol and active‐control during treatment maintenance period (tapentadol ER 200 to 500 mg/day and oxycodone controlled release (CR) 40 to 100 mg/day).
We did not identify any cross‐over trials for inclusion.
Dealing with missing data
We contacted the pharmaceutical companies responsible for the development of tapentadol (Johnson & Johnson Pharmaceutical Research and Development L.L.C., and Grünenthal GmbH) for further relevant information.
Assessment of heterogeneity
We used the I2 statistic to detect the presence of heterogeneity. When we found statistical heterogeneity (I2 statistic > 50%), we investigated this by conducting non‐planned subgroup analyses (see Subgroup analysis and investigation of heterogeneity). We performed fixed‐effect or random‐effects meta‐analyses according to the presence (I2 statistic > 50%) or absence (I2 statistic < 50%) of statistical significant heterogeneity.
Assessment of reporting biases
We prepared a matrix containing the major outcomes that were routinely measured in studies assessing opioids for pain relief and compared these with the outcomes that were reported in tapentadol studies. If we identified mismatches, we contacted the trial authors and the drug manufacturers for further information. In these cases, we also searched the clinical trials registers to find trial protocols and look for differences between protocol and trial publications.
We planned to assess the risk of publication bias by comparing the pooled NNTB from present meta‐analysis estimates with the NNTB value of 8 (considered to be the limit of clinical utility) (Moore 2002), and calculate the number of participants enrolled in unavailable trials that was needed to achieve the point clinical non‐utility (NNTB 8 or greater).
Data synthesis
We used Review Manager 5 to pool studies and calculate an estimate of overall treatment effects (RevMan 2014). We performed fixed‐effect or random‐effects meta‐analysis according to statistical heterogeneity. For continuous data, we used the inverse variance method; for dichotomous data, we used the Mantel‐Haenszel method. Data synthesis included the systematical description of the studies' findings and the quantitative data pooling.
Subgroup analysis and investigation of heterogeneity
We performed two meta‐analyses, one for placebo‐controlled and one for active‐controlled trials.
For each meta‐analysis, we planned to conduct the following subgroup analyses independently of whether or not statistical significant heterogeneity was present.
-
Pain intensity (moderate versus severe).
-
Chronic musculoskeletal pain conditions (eg low back pain or osteoarthritis) versus chronic musculoskeletal pain syndromes (eg fibromyalgia).
-
Aetiology of chronic pain conditions.
-
Study quality (poorer quality versus higher quality). We considered a trial to be of higher quality if it had an adequate allocation concealment and rated 'low' or 'unclear' in the risk assessment of all other type of bias according to the Cochrane 'Risk of bias' assessment tool. We considered a trial to be of poorer quality if it had an inadequate or unclear allocation concealment, or if it rated 'high' in the 'Risk of bias' assessment of any other type of bias (Higgins 2011).
For the comparison tapentadol versus placebo, the data available only allowed us to perform one of the pre‐planned subgroup analysis (aetiology of chronic pain conditions). For the comparison tapentadol versus active control intervention, the data available only allowed us to perform two of the pre‐planned subgroup analysis (aetiology of chronic pain conditions and study quality). For the tapentadol versus active control comparison, we decided to perform a post‐hoc subgroup analysis based on the trial follow‐up period because we included studies with different lengths of follow‐up.
Sensitivity analysis
We pre‐planned no further sensitivity analyses besides those described in Subgroup analysis and investigation of heterogeneity: study quality, characteristics of the participants and outcomes (e.g. pain intensity and aetiology), characteristics of interventions (e.g. dosage of control and intervention groups), data analysis (e.g. LOCF and BOCF, data from cross‐over trials), and methods of analysis (e.g. fixed‐effect or random‐effects methods). We did not identify any further issues when conducting the review that should have been subjected to further sensitivity analysis.
Results
Description of studies
Results of the search
Figure 1 shows the flow diagram of study selection (Moher 2009).
Study flow diagram.
The search strategy yielded 291 potentially relevant reports. After applying inclusion/exclusion criteria, we included four RCTs (cited in 32 reports) that enrolled 4094 participants (1876 treated with tapentadol, 992 treated with placebo, and 1226 treated with oxycodone).
We did not identify any trials on‐going or awaiting publication that matched our selection criteria. We were able to include data from all relevant phase III studies. However, the drug manufacturer did not grant access to the final clinical reports of those studies, and, therefore, some potentially useful information was unavailable for meta‐analysis.
Included studies
See: Characteristics of included studies table.
All included studies were RCTs, of which three were double‐blind, active and placebo controlled (Afilalo 2010; Afilalo 2013; Buynak 2010), and one was an open‐label trial comparing tapentadol ER to oxycodone CR (Wild 2010).
Two studies evaluated people with knee osteoarthritis (Afilalo 2010; Afilalo 2013), one trial evaluated people with low back pain (Buynak 2010), and one trial evaluated people with both chronic pain conditions (Wild 2010).
All studies used tapentadol ER in a dose range of 200 to 500 mg/day and oxycodone CR as active comparator in doses ranging from 40 to 100 mg/day. Three studies used a placebo‐controlled arm (Afilalo 2010; Afilalo 2013; Buynak 2010), and one study was an open‐label, long‐term safety trial (Wild 2010).
Three studies aimed to evaluate efficacy as primary outcome (change from baseline in pain intensity score) and had a follow‐up period of 12 weeks (Afilalo 2010; Afilalo 2013; Buynak 2010). The open‐label trial aimed to evaluate safety (primary outcome was the number of participants with treatment emergent adverse effects) and had a longer follow‐up period (52 weeks) (Wild 2010). Despite these differences, the outcome measurements reported in all studies were similar.
Afilalo 2013 was not fully published. We collected data for this trial from ClinicalTrials.gov and review articles reporting results from this trial.
Excluded studies
See: Characteristics of excluded studies table.
We excluded 217 reports after initial screening due to at least one of the following reasons: tapentadol formulation (immediate release), pain condition (acute pain, post‐operative pain, neuropathic pain, etc), trial design (uncontrolled studies, non‐randomised studies, cost‐effectiveness studies, pooled analysis, review articles), and duplicates (see Figure 1).
We excluded five reports, corresponding to three studies (Ashworth 2010; Gálvez 2013; Lange 2010), after full‐text review because of trial design. Two studies were not controlled trials (Ashworth 2010; Gálvez 2013), and one study used pooled analysis (Lange 2010).
Risk of bias in included studies
See: Characteristics of included studies table.
We considered two trials to be at low risk of bias (Afilalo 2010; Buynak 2010). Wild 2010 was at high risk of bias due to its open‐label design. We also judged Afilalo 2013 to be at high risk of bias due to concerns regarding the outcomes selected and because we could not find an allocation concealment description (Figure 2).
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
The drug manufacturer funded all included studies.
Allocation
All studies, except Afilalo 2013, reported that allocation sequences were computer generated, and interactive voice response systems were used to hide allocation (Afilalo 2010; Buynak 2010; Wild 2010). We did not find any allocation concealment description regarding Afilalo 2013.
Blinding
Wild 2010 was an open‐label trial, hence neither the personnel or participants were blinded. In all other studies, participants, investigators, and outcome assessment were blinded until the end of the trial (Afilalo 2010; Afilalo 2013; Buynak 2010).
Incomplete outcome data
All included studies had overall withdrawal rates that were considerably high (ranging from 48% to 56%). These withdrawal rates are not unusual in this clinical condition (chronic musculoskeletal pain) and in studies evaluating opioids. Taking into account the overall withdrawal rate and the fact that only LOCF data were available, we could not exclude the possibility of bias.
There was no evidence of differential overall withdrawal rates.
Selective reporting
We have no reasons to suspect incompleteness of data reporting in the fully published trials (Afilalo 2010; Buynak 2010; Wild 2010). Afilalo 2013 was not published. However, the trial protocol was available and all of the study's pre‐specified primary and secondary outcomes were reported.
Other potential sources of bias
Afilalo 2013 was not published. We obtained its data from ClinicalTrials.gov. The trial efficacy results were generally less favourable than those in the remaining studies, which may be due to some publication bias.
It would be highly desirable to have responder rates published in all studies. Unfortunately, neither Afilalo 2013 nor Wild 2010 reported this outcome.
We did not identify any other potential sources of bias.
Effects of interventions
Tapentadol extended release versus placebo
Pain control
For the primary efficacy outcomes, we were able to retrieve data from three studies (1973 participants) for the outcome 'change in pain intensity from baseline at week 12' (Afilalo 2010; Afilalo 2013; Buynak 2010), and from two studies (1313 participants) for the outcome 'responders rate' (at least 50% pain relief) (Afilalo 2010; Buynak 2010). In comparison with placebo, tapentadol was associated with a higher reduction in pain intensity from baseline at week 12 (MD on the 11‐point NRS of ‐0.56, 95% CI ‐0.92 to ‐0.2; I2 = 65%; Analysis 1.1) and with a higher responder rate (RR 1.36, 95% CI 1.13 to 1.64; I2 = 0%; NNTB tapentadol versus placebo for 12 weeks: 16, 95% CI 9 to 57; Analysis 1.2).
Patient Global Impression of Change
Concerning PGIC (much or very much improved), we included data from three studies (1554 participants) (Afilalo 2010; Afilalo 2013; Buynak 2010). Tapentadol‐treated participants were more likely to consider themselves much or very much improved (RR 1.53, 95% CI 1.28 to 1.82; I2 = 60%).
Quality of life
With respect to functional health status and well‐being scores (EQ‐5D, SF‐36, WOMAC), three studies provided data for EQ‐5D (Afilalo 2010; Afilalo 2013; Buynak 2010), two studies for SF‐36 (Afilalo 2010; Buynak 2010), and two studies for WOMAC (Afilalo 2010; Afilalo 2013). Tapentadol was associated with a higher increase in SF‐36 physical component summary score (MD 2.57, 95% CI 1.69 to 3.44; I2 = 0%). There were no significant differences with respect to placebo for the other functional health status and well‐being scores.
Withdrawals due to adverse effects
The primary safety outcome was trial discontinuation due to treatment‐emergent adverse effects. We were able to retrieve data from all three studies (1975 participants) (Afilalo 2010; Afilalo 2013; Buynak 2010). Tapentadol was associated with a higher risk of withdrawal due to adverse effects (RR 2.68, 95% CI 2.05 to 3.52; I2 = 0%; Analysis 1.3), in comparison with placebo (NNTH 10; 95% CI 7 to 12, for 12 weeks). The overall withdrawal risk was also higher among tapentadol‐treated participants (RR 1.09, 95% CI 0.90 to 1.32; I2 = 73%). The causes for trial discontinuation in placebo‐controlled trials were: withdrawal by participant (30%), adverse effects (27%), lack of efficacy (18%), lost to follow‐up (5%), and trial drug non‐compliance (5%). Tapentadol‐treated participants were at a lower risk of withdrawal due to lack of efficacy (RR 0.36, 95% CI 0.25 to 0.50; I2 = 0%). There were no differences for the remaining withdrawal causes.
Adverse effects
Regarding overall incidence of adverse effects and based on data from two studies (Afilalo 2010; Buynak 2010), tapentadol was also associated with a higher risk (RR 1.25, 95% CI 1.16 to 1.35; I2 = 0%, 1318 participants; Analysis 1.4). Considering the incidence of serious adverse effects, based on data from three studies (Afilalo 2010; Afilalo 2013; Buynak 2010), there were no differences between tapentadol and placebo (RR 1.01, 95% CI 0.47 to 2.16; I2 = 21%, 1974 participants; Analysis 1.5).
Three studies (1974 participants) reported data on specific adverse effects (gastrointestinal disorders, nervous system disorders, fatigue, and pruritus) (Afilalo 2010; Afilalo 2013; Buynak 2010). Tapentadol was associated with a higher risk of constipation (RR 2.43, 95% CI 1.86 to 3.17; I2 = 4% ), nausea (RR 2.81, 95% CI 2.18 to 3.62; I2 = 0%), vomiting (RR 2.77, 95% CI 1.83 to 4.21; I2 = 56%), dry mouth (RR 3.08, 95% CI 1.92 to 4.94; I2 = 0%), somnolence (RR 3.27, 95% CI 2.26 to 4.73; I2 = 19%), dizziness (RR 2.73, 95% CI 2.08 to 3.60; I2 = 13%), and fatigue (RR 2.15, 95% CI 1.48 to 3.11; I2 = 0%). There were no differences for diarrhoea (RR 0.85, 95% CI 0.59 to 1.23; I2 = 0%), headache (RR 1.13, 95% CI 0.91 to 1.40; I2 = 44%), and pruritus (RR 2.67, 95% CI 0.85 to 8.37; I2 = 72%).
Subgroup analyses
We performed subgroup analysis based on pain aetiology and study quality.
Pain aetiology
Regarding pain aetiology, there were no significant differences between estimates from both groups (low back pain and osteoarthritis) for all outcomes. However, the difference between tapentadol and placebo for the primary outcome (change in pain intensity from baseline at week 12) among the subgroup of people with osteoarthritis did not reach statistical significance (MD ‐0.45, 95% CI ‐0.94 to 0.04; I2 = 74%, two trials). This analysis fails to explain, at least partially, the statistical heterogeneity found in the efficacy outcomes.
Trial quality
In subgroup analysis based on trial quality, the 'change in pain intensity from baseline at week 12' (MD ‐0.74, 95% CI ‐1.00 to ‐0.48; I2 = 0%, two trials), the PGIC (RR 1.67, 95% CI 1.45 to 1.92; I2 = 60%, two trials, 1012 participants), and the change in WOMAC score (MD ‐0.27, 95% CI ‐0.42 to ‐0.12; one trial) outcomes were higher in pooled analysis of high‐quality studies, without significant heterogeneity.
Tapentadol extended release versusoxycodone
Wild 2010, a long‐term safety trial of 52 weeks' follow‐up, was also included in this comparison. In general, the size effect estimate in this trial was smaller in comparison with the 12‐week studies (Afilalo 2010; Afilalo 2013; Buynak 2010) for both efficacy and safety outcomes. Therefore, the inclusion of this trial represented a conservative approach. Further issues regarding this decision are dealt with when presenting the results of a subgroup analysis and in the Discussion.
Pain control
For the primary efficacy outcomes, we retrieved data from four studies (2427 participants) for the outcome 'change in pain intensity from baseline' (Afilalo 2010; Afilalo 2013; Buynak 2010; Wild 2010), and from two studies (1327 participants) for the outcome 'responders rate' (at least 50% pain relief) (Afilalo 2010; Buynak 2010).
In comparison with oxycodone, tapentadol was associated with a higher reduction in pain intensity from baseline at week 12 (MD on the 11‐point NRS ‐0.24, 95% CI ‐0.43 to ‐0.05; I2 = 14%; Analysis 2.1), without significant difference in the responders rate (RR 1.46, 95% CI 0.92 to 2.32; I2 = 82%; NNTB tapentadol versus oxycodone for 12 weeks 11, 95% CI 7 to 21; Analysis 2.2).
Patient Global Impression of Change
Regarding PGIC (much or very much improved), we included all four studies (2360 participants) for analysis (Afilalo 2010; Afilalo 2013; Buynak 2010; Wild 2010). There was no significant difference between tapentadol and oxycodone (RR 1.15, 95% CI 0.98 to 1.35; I2 = 70%).
Quality of life
Considering functional health status and well‐being scores (EQ‐5D, WOMAC), we were able to use data from two studies (Afilalo 2010; Afilalo 2013). Tapentadol was associated with a higher increase in EQ‐5D health status index (MD 0.1, 95% CI 0.07 to 0.13; I2 = 0%). The mean change in WOMAC pain subscale was similar between tapentadol and oxycodone (MD ‐0.03, 95% CI ‐0.23 to 0.17; I2 = 76%, two trials).
We retrieved data for trial discontinuation due to treatment‐emergent adverse effects from all four studies (3102 participants) (Afilalo 2010; Afilalo 2013; Buynak 2010; Wild 2010).
Withdrawals due to adverse effects
Tapentadol had a lower risk of withdrawal due to adverse effects compared to oxycodone (RR 0.5, 95% CI 0.42 to 0.60; I2 = 44%; NNTB tapentadol versus oxycodone for 12 weeks: 6, 95% CI 5 to 7; Analysis 2.3). The overall withdrawal risk was also higher among oxycodone‐treated participants (RR 0.76, 95% CI 0.70 to 0.83; I2 = 41%, four trials, 3102 participants). Tapentadol‐treated participant were at a higher risk of withdrawal due to loss to follow‐up (RR 1.73, 95% CI 1.04, 2.89; I2 = 0%, four trials, 3102 participants) and to lack of efficacy (RR 2.23, 95% CI 1.45 to 3.42; I2 = 0%, four trials, 3102 participants). There were no differences for the remaining withdrawal causes. The causes for trial discontinuation in oxycodone‐controlled trials were: adverse effects (49%), withdrawal by participant (24%), lack of efficacy (8%), trial drug non‐compliance (6%), and loss to follow‐up (5%).
Adverse effects
We included three studies for incidence of overall adverse effects of any type (Afilalo 2010; Buynak 2010; Wild 2010). Tapentadol was associated with a lower risk (RR 0.91, 95% CI 0.85 to 0.96; I2 = 57%; 2449 participants; Analysis 2.4).
Considering the incidence of serious adverse effects and based on data from all included studies, there was no significant difference between tapentadol and oxycodone (RR 0.57, 95% CI 0.24 to 1.33; I2 = 64%; 3099 participants; Analysis 2.5) (Afilalo 2010; Afilalo 2013; Buynak 2010; Wild 2010).
For specific adverse effects, four studies (3099 participants) reported data on gastrointestinal disorders, nervous system disorders, fatigue, insomnia, and pruritus (Afilalo 2010; Afilalo 2013; Buynak 2010; Wild 2010), and two studies supplied data on incidence of insomnia (Buynak 2010; Wild 2010). Tapentadol was associated with a lower risk of constipation (RR 0.53, 95% CI 0.47 to 0.61; I2 = 0%), nausea (RR 0.57, 95% CI 0.50 to 0.64; I2 = 0%), vomiting (RR 0.41, 95% CI 0.33 to 0.51; I2 = 18%), dizziness (RR 0.81, 95% CI 0.69 to 0.95; I2 = 0%), and pruritus (RR 0.41, 95% CI 0.26 to 0.65; I2 = 64%). There was no difference regarding diarrhoea (RR 1.13, 95% CI 0.82 to 1.56; I2 = 69%), fatigue (RR 0.93, 95% CI 0.73 to 1.17; I2 = 0%), insomnia (RR 0.94, 95% CI 0.31 to 2.85; I2 = 82%), somnolence (RR 0.81, 95% CI 0.57 to 1.16; I2 = 71%), and headache (RR 1.22, 95% CI 0.99 to 1.50; I2 = 6%). Tapentadol was associated with a higher risk of dry mouth (RR 1.80, 95% CI 1.29 to 2.50; I2 = 0%).
Subgroup analyses
The data available allowed us to perform two of the pre‐planned subgroup analysis (aetiology of chronic pain conditions and trial quality) and a post‐hoc subgroup analysis (trial follow‐up period: 12 versus 52 weeks' follow‐up).
Aetiology of chronic pain conditions
Concerning the subgroup analysis based on pain aetiology (low back pain (Buynak 2010) and osteoarthritis (Afilalo 2010; Afilalo 2013)), tapentadol was associated with a higher improvement in the 11‐point NRS in people with osteoarthritis pain (MD ‐0.40, 95% CI ‐0.66 to ‐0.14; I2 = 0%), but not in people with low back pain. Nevertheless, there were no statistical significant differences in the effect size between these subgroups. Tapentadol was also associated with a higher responder rate (RR 1.85, 95% CI 1.40 to 2.45; I2 = 82.5%, one trial, 686 participants) and improvement in PGIC (RR 1.28, 95% CI 1.12 to 1.46; I2 = 0%, two trials, 918 participants) in people with osteoarthritis than those with low back pain. There were no data available to explore subgroup differences in the other efficacy outcomes. There were no statistical significant differences between these subgroups in safety outcomes.
Trial quality
In subgroup analysis based on trial quality, the risk reduction of adverse effects among tapentadol‐treated participants was significantly higher in better‐quality studies (RR 0.88, 0.83 to 0.93; I2 = 0%, two trials, 1332 participants) than in low‐quality studies (RR 0.95, 95% CI 0.90 to 0.99, one trial, 1117 participants).
In the post‐hoc subgroup analysis based on trial follow‐up period (12 and 52 weeks) tapentadol was associated with a higher improvement in the 11‐point NRS in studies of shorter follow‐up period (MD ‐0.28, 95% CI ‐0.50 to ‐0.07; I2 = 26%, three trials) than in the trial of longer follow‐up period. Nevertheless there were no statistically significant differences in the effect size between these subgroups. Safety outcomes results for this subgroup analysis were similar to those found in the global analysis, that is, tapentadol was associated with a lower risk of withdrawals due to adverse effects and with a lower incidence of adverse effects. These differences with respect to oxycodone were higher in studies of shorter follow‐up period. In this subgroup of studies, tapentadol was also associated with a lower risk of serious adverse effects (RR 0.41, 95% CI 0.19 to 0.87; I2 = 23%, three trials, 1982 participants) and somnolence (RR 0.69, 95% CI 0.54 to 0.87; I2 = 14%, three trials, 1982 participants).
There was no significant heterogeneity among study results for the primary efficacy outcome 'change in pain intensity from baseline'. Subgroup analysis explained, at least partially, the heterogeneity we found for the other main efficacy and safety outcomes.
We have no evidence of publication bias based on the comparison between pre‐specified outcomes in study protocols and published reported outcomes. Studies reported data for the major outcomes that are routinely measured in studies assessing opioids for pain relief. We did not perform a funnel plot analysis because of the low number of included studies.
Discussion
Opioids are commonly used in the treatment of chronic musculoskeletal pain. We have reviewed the efficacy and safety of tapentadol in people with osteoarthritis of the knee and low back pain. Other non‐Cochrane reviews with the same scope have been published since we started this review (Afilalo 2013; Lange 2010; Manchikanti 2011; Riemsma 2011). We performed a meta‐analysis for the most relevant efficacy and safety outcomes.
Summary of main results
We included four parallel‐design RCTs enrolling 4094 participants. Three were phase III efficacy studies with a follow‐up of 12 weeks and one was a safety study with a follow‐up of 52 weeks. All studies had a control‐active group (oxycodone) and three studies were also placebo‐controlled. We judged Wild 2010 and Afilalo 2013 to be at high risk of bias: Wild 2010 due to its open‐label design and Afilalo 2013 due to missing information.
Of the three placebo‐controlled trials, two were in favour of tapentadol regarding primary efficacy outcomes (Afilalo 2010; Buynak 2010), and one trial did not find significant differences between treatment groups (Afilalo 2013). Pooled results from these studies showed a 0.56 point reduction in pain intensity from baseline at 12 weeks in the 11‐point NRS and a 1.36 increase in the risk of respond to the treatment (ie at least 50% pain relief). Tapentadol was associated with a 2.68‐fold increase in the risk of discontinuing treatment due to adverse effects.
The overall withdrawal risk was not significantly higher among tapentadol‐treated participants as the withdrawals due to lack of efficacy favoured tapentadol in comparison to placebo and account for the no significant difference between groups.
We found moderate‐to‐high statistical heterogeneity for the efficacy outcome estimates, although heterogeneity tests can be misleading due to the number and clinical heterogeneity of the studies included in the meta‐analysis.
Subgroup analysis based on pain aetiology did not show differences between the subgroups of people with osteoarthritis of the knee and low back pain.
For the outcomes 'change in pain intensity from baseline at week 12', 'PGIC', and 'WOMAC change', pooled effect size was higher in high‐quality studies, favouring tapentadol, without significant heterogeneity.
Of the four oxycodone‐controlled studies, two were in favour of tapentadol regarding the primary efficacy outcome 'change in pain intensity from baseline' (Afilalo 2010; Afilalo 2013), and two studies found no significant differences between treatment groups (Buynak 2010; Wild 2010). The results of the two studies that evaluated responder's rate were in favour of tapentadol (Afilalo 2010; Buynak 2010). Pooled results from these studies showed a 0.24 points reduction in pain intensity from baseline in the 11‐point NRS and a non‐significant 1.46 increase in the risk of respond to the treatment.
Tapentadol was associated with a 50% reduction in the risk of discontinuing treatment due to adverse effects (primary safety outcome). Tapentadol was also associated with a 9% reduction in the risk of adverse effects and with a non‐significant 43% reduction in the risk of serious adverse effects. Specific adverse effects less frequently reported among tapentadol‐treated participants were constipation, nausea, vomiting, dizziness, and pruritus. Tapentadol was associated with a higher risk of dry mouth.
The overall withdrawal risk was also higher among oxycodone‐treated participants as adverse effects accounted for half of all withdrawals in this group. Tapentadol had a higher withdrawal rate due to lack of efficacy.
Moderate‐to‐high statistical heterogeneity was found for most efficacy (except for change in pain intensity from baseline) and safety outcome estimates. As for the comparison against placebo, the small number of trials included in the meta‐analysis precludes strong conclusions about the significance of statistical heterogeneity, particularly in the case of clinical heterogeneity.
Subgroup analysis showed a higher improvement in reduction in pain intensity from baseline with tapentadol among people with osteoarthritis of the knee, studies of higher quality and shorter follow‐up periods, although there were no statistical significant differences in the effect size between these subgroups. Overall, the main safety outcomes results for all subgroup analysis were similar to those found in the global analysis, that is tapentadol was associated with a lower risk of withdrawals due to adverse effects and with a lower incidence of adverse effects.
Subgroup analyses explained, at least partially, the heterogeneity found in most efficacy and safety outcomes. In summary, low heterogeneity was found among people with osteoarthritis of the knee and between results of studies of higher quality and shorter follow‐up periods.
Overall completeness and applicability of evidence
Although we included a large number of participants (more than 4000) in this Cochrane review, tapentadol is a recent drug and only a relatively small number of studies met the inclusion criteria (four RCTs). Nevertheless, these studies can be considered very large in relation to other studies in chronic pain. The number of studies may have contributed to the high statistical heterogeneity found for some results. In fact, subgroup and sensitivity analysis explained most, but not all, the heterogeneity found. All included studies were well designed, but two studies were rated as having high risk of bias due to an open‐label design and unclear allocation concealment. The drug manufacturer funded all included studies.
The Afilalo 2013 study was never published and its data were collected from the clinical trials site, clinicaltrials.gov. This trial's efficacy results were generally worse than those in the remaining studies, which may point to publication bias. Given the effect size and number of studies, our review findings are considerably susceptible to further publications of opposite direction. Nevertheless, it is likely that we found all of the available information of the completed studies as we searched the clinical trials registers and agencies to find further trials. We searched the publications for differences between them and the original protocols and found all outcomes to be consistently reported.
We were only able to include people with low back pain and knee osteoarthritis. We did not identify any evidence for tapentadol in chronic musculoskeletal pain of other causes.
All included studies had high withdrawal rates, with withdrawal due to adverse effects comprising 27% of the total in the placebo‐controlled arm and 49% of the total in the oxycodone‐controlled arm. For every six people treated with tapentadol, one more presented an adverse effects than if they were treated with placebo, reflecting the high incidence of adverse effects in chronic opioid consumption. This is in line with other studies of opioids in chronic pain. This high withdrawal rate makes the imputation method considerably important. We planned to use BOCF, but had to use LOCF due to data access limitations, as such our results can be slightly overestimated.
Three trials had a follow‐up period of 12 weeks and one trial had a follow‐up of 52 weeks. When considering long‐term use, the review results can be somewhat overestimated as all opioid results are worse with long‐term follow‐up (Nüesch 2009; Trescot 2008c), and, judging by Wild 2010, this is also the case with tapentadol.
As such, opioid treatment in chronic pain conditions remains controversial due to concerns about adverse effects, long‐term efficacy, functional outcomes, and the potential for drug abuse and addiction. These concerns may contribute to the hesitancy to prescribe tapentadol, as may the regulatory and legal barriers in the prescription of controlled substances.
Nevertheless, we believe there is sufficient evidence to justify this meta‐analysis and moderate strength in our recommendations regarding the use of tapentadol as an option in chronic musculoskeletal pain treatment.
Quality of the evidence
See Characteristics of included studies, 'Risk of bias' tables, and 'Risk of bias' summary tables (Figure 2; Figure 3).
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Of the included four studies, we only judged two studies to be of high quality (Afilalo 2010; Buynak 2010). We considered Wild 2010 to be at high risk of performance and detection bias due to its open‐label design. We were unable to obtain any information regarding Afilalo 2013 allocation concealment and so we considered it to be at high risk of bias.
Although enrolling more than 4000 participants, the number of included studies was still limited and we found significant heterogeneity for most efficacy and safety outcomes. Moreover, as is usual in these conditions, the withdrawal rates were high. Therefore, we considered that there is only moderate evidence that tapentadol is more efficacious than placebo and oxycodone in the treatment of people with chronic musculoskeletal pain. There is also moderate evidence that tapentadol has a better tolerability and safety profile than oxycodone.
Potential biases in the review process
The main potential source of bias in this review comes from the changes to the original protocol (see Differences between protocol and review) as we had to use LOCF for primary efficacy outcome analysis. The LOCF method may overestimate treatment effect mostly in studies with shorter follow‐up and high withdrawal rates, as are the studies included in this Cochrane review. We analysed data for pain intensity change from baseline at week 12 and not during the overall treatment maintenance period. This may also overestimate effect size. We were unable to conduct subgroup analyses for pain intensity, and the possibility exists that effect size may be different for moderate versus severe cases of chronic musculoskeletal pain. An additional bias was that we could not obtain data from all studies for all outcomes.
A further limitation of this review was the relatively small number of included studies, although enrolling a large number of participants. In this context, the results of the pooled analysis should be viewed with caution especially in the presence of statistically heterogeneity, as further studies may have an important impact in effect size estimations.
Agreements and disagreements with other studies or reviews
Lange 2010 published a pooled analysis of three studies included in this review (Afilalo 2010; Afilalo 2013; Buynak 2010). The authors of this paper worked at that time for the companies responsible for the development of tapentadol, and had access to the individual participant data from these studies. The imputation issue was central in this paper. This pooled analysis reported that, when compared to placebo, tapentadol effectively reduced pain. When using BOCF instead of LOCF, the results were slightly less positive for tapentadol but still significant in comparison to placebo (change in pain intensity from baseline at week 12 in the 11‐point NRS of ‐0.60 (95% CI ‐0.80 to 0.39) using LOCF and ‐0.30 (95% CI ‐0.47 to ‐0.90) using BOCF). The trial authors did not report results for the comparison tapentadol versus oxycodone. The only possible inference is based on indirect descriptive analysis. Lange 2010's pooled analysis found a significant reduction of ‐0.30 (95% CI ‐0.53 to ‐0.12) in pain intensity at week 12 for oxycodone versus placebo when using LOCF, but not when using BOCF. In fact, when using the BOCF imputation method, oxycodone was not associated with a decrease in pain intensity at week 12 in comparison to placebo (0.30, 95% CI 0.11 to 0.50), as such we would expect tapentadol to perform better compared to oxycodone when using BOCF. This seems to be associated with the fact that tapentadol has fewer withdrawals and withdrawals due to adverse effects when compared to oxycodone. In those participants when using BOCF, there would be no pain reduction from the baseline, as when using LOCF the pain reduction is still considered despite the drop‐out. These results are in accordance with ours and considerations regarding the risk of slight overestimation of results were already made.
Manchikanti 2011 published a systematic review on opioid treatment for chronic non‐cancer pain but did not conduct a meta‐analysis. With respect to tapentadol, the review authors included three studies (also included in this Cochrane review: Afilalo 2010; Buynak 2010; Wild 2010), and concluded that tapentadol has "similar effects as other opioids, and with fewer side effects".
Another systematic review funded by Grünenthal GmbH, Riemsma 2011, compared different opioids in severe and moderate‐to‐severe chronic pain, including neuropathic and malignant aetiology. Based on a network meta‐analysis, the review authors concluded that tapentadol was superior for the primary outcome (mean change in pain intensity) to oxycodone, hydromorphone, and morphine. Tapentadol had fewer gastrointestinal adverse effects in comparison with the other opioids.
Another Cochrane review on oral tapentadol for cancer pain is under way (Wiffen 2015).
Overall, our findings and conclusion are in accordance with these previous studies.
Study flow diagram.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Comparison 1: Tapentadol Extended Release (100 to 250 mg/day) versus placebo, Outcome 1: Change in pain intensity from baseline at week 12 (11‐point numerical rating scale)
Comparison 1: Tapentadol Extended Release (100 to 250 mg/day) versus placebo, Outcome 2: Responder rate (at least 50% pain reduction)
Comparison 1: Tapentadol Extended Release (100 to 250 mg/day) versus placebo, Outcome 3: Study discontinuation due to treatment‐emergent adverse effects
Comparison 1: Tapentadol Extended Release (100 to 250 mg/day) versus placebo, Outcome 4: Adverse effects
Comparison 1: Tapentadol Extended Release (100 to 250 mg/day) versus placebo, Outcome 5: Serious adverse effects
Comparison 2: Tapentadol expended release (100 to 250 mg/day) versus oxycodone controlled release (20 to 50 mg/day), Outcome 1: Change in pain intensity from baseline (11‐point numerical rating scale)
Comparison 2: Tapentadol expended release (100 to 250 mg/day) versus oxycodone controlled release (20 to 50 mg/day), Outcome 2: Responder rate (at least 50% pain reduction)
Comparison 2: Tapentadol expended release (100 to 250 mg/day) versus oxycodone controlled release (20 to 50 mg/day), Outcome 3: Study discontinuation due to treatment‐emergent adverse effects
Comparison 2: Tapentadol expended release (100 to 250 mg/day) versus oxycodone controlled release (20 to 50 mg/day), Outcome 4: Adverse effects
Comparison 2: Tapentadol expended release (100 to 250 mg/day) versus oxycodone controlled release (20 to 50 mg/day), Outcome 5: Serious adverse effects
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 Change in pain intensity from baseline at week 12 (11‐point numerical rating scale) Show forest plot | 3 | 1973 | Mean Difference (IV, Random, 95% CI) | ‐0.56 [‐0.92, ‐0.20] |
| 1.2 Responder rate (at least 50% pain reduction) Show forest plot | 2 | 1313 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.36 [1.13, 1.64] |
| 1.3 Study discontinuation due to treatment‐emergent adverse effects Show forest plot | 3 | 1975 | Risk Ratio (M‐H, Fixed, 95% CI) | 2.68 [2.05, 3.52] |
| 1.4 Adverse effects Show forest plot | 2 | 1318 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.25 [1.16, 1.35] |
| 1.5 Serious adverse effects Show forest plot | 3 | 1974 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.01 [0.47, 2.16] |
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 2.1 Change in pain intensity from baseline (11‐point numerical rating scale) Show forest plot | 4 | Mean Difference (IV, Fixed, 95% CI) | ‐0.24 [‐0.43, ‐0.05] | |
| 2.2 Responder rate (at least 50% pain reduction) Show forest plot | 2 | 1327 | Risk Ratio (M‐H, Random, 95% CI) | 1.46 [0.92, 2.32] |
| 2.3 Study discontinuation due to treatment‐emergent adverse effects Show forest plot | 4 | 3102 | Risk Ratio (M‐H, Random, 95% CI) | 0.50 [0.42, 0.60] |
| 2.4 Adverse effects Show forest plot | 3 | 2449 | Risk Ratio (M‐H, Random, 95% CI) | 0.91 [0.85, 0.96] |
| 2.5 Serious adverse effects Show forest plot | 4 | 3099 | Risk Ratio (M‐H, Random, 95% CI) | 0.57 [0.24, 1.33] |
