The methodology employed for this study involved a systematic search of randomized controlled trials (RCTs) related to the management of chronic pain using tapentadol. The search was performed in the publicly accessible databases of PubMed and EMBASE, with the search timeline extending up to September 30, 2023. The specific search strategy employed, including search terms and their combinations, is thoroughly documented in the Supplementary Section of this paper. The process of literature selection, data extraction, and analysis in this study strictly adheres to the protocols delineated in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, ensuring the transparency and reproducibility of our research.

The primary outcome measure in this study was the reduction in the Numerical Rating Scale (NRS) from baseline. The inclusion criteria were as follows: (1) Randomized controlled clinical trials (RCTs); (2) Studies where tapentadol monotherapy was used as the treatment; (3) Studies involving subjects experiencing various types of chronic pain and diagnosed as functional capacity levels I–III; (4) Studies that reported NRS scores at specific time points for chronic pain; (5) Studies involving participants aged 18 years or older.

The exclusion criteria included the following: (1) Studies involving subjects without a washout period before enrollment. Since tapentadol is primarily used as a second-line treatment, participants typically have received other opioid treatments before joining the study. Without a washout period, previous medications could potentially affect the efficacy assessment of tapentadol; (2) Cancer pain research; (3) Studies where chronic pain was not accompanied by reported NRS baseline values; (4) Studies using a randomized withdrawal design; (5) Crossover study designs that did not report data from the first cycle; (6) Studies involving subjects with concurrent psychiatric disorders.

Data extraction for this study was performed using Microsoft Excel (version 16.38). The extracted data encompassed several categories: (1) Basic literature information, including authors, year of publication, and clinical trial registration number; (2) Trial details such as the formulation and dosage of tapentadol, the etiology of pain, sample size, trial duration, the use of blinding and placebo control; (3) Participant characteristics, which included the mean age, gender ratio, and baseline NRS score; (4) Outcome measures, specifically the change in NRS score from baseline at each visit. To ensure the accuracy and reliability of the data extraction process, it was independently conducted by two researchers. Any discrepancies between their findings were resolved through discussion with a third researcher. Graphical data were extracted using Engauge Digitizer software (version 4.1). If the extraction error between the two researchers exceeded 2%, the data in question were re-extracted, and the average value of the two extracts was used for further analysis.

The risk of bias (RoB) for each included randomized controlled trial (RCT) was independently examined by two researchers utilizing the Cochrane RoB2 tool (13). This tool, specifically the RoB2_IRPG_beta_v9.xlsm, was employed to assess potential bias in five key areas: the randomization process, deviation from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported result. The RoB for each category was designated as “low” if there was a low risk of bias, “high” if there was a high risk of bias, or “unclear” if there was insufficient information or uncertainty about potential bias. In the event of discrepancies in the RoB2 assessments between the two primary researchers, a third researcher was consulted to review the assessments and make the final decision.

A time-effect model was developed using the change in NRS from baseline as the efficacy indicator. Influencing factors such as dosage, formulation, administration frequency, baseline NRS values, male ratio, and year of publication (before or after 2014) were considered, along with whether a placebo control was used. The year 2014 was a critical turning point in the U.S. opioid crisis, characterized by a significant increase in overdose deaths, heightened media attention, proactive interventions from governmental and regulatory bodies, and the beginning of legal challenges against pharmaceutical companies. In response to the crisis, the FDA implemented numerous warnings and updated guidelines that emphasized the need for clearer medication labeling. These concerted efforts increased public awareness of the risks associated with opioids and enhanced strategies to address the crisis. This, in turn, influenced the methodologies used in subsequent clinical research (14–16).

This model quantified the relationship between the change in NRS from baseline and time, and examined the influencing factors. Detailed methods are available in the Supplementary Materials.

A total of 223 articles were retrieved, of which 16 articles (representing 7.2% of the total) were included in the final analysis. These articles comprised 22 treatment arms involving 4,508 participants treated with tapentadol. All the trials were designed using dose titration. The mean age of the participants ranged from 41.4 to 65.5 years (with a median of 58.4 years), and the male proportion varied from 36% to 59% (with a median of 42.4%). The average BMI of the participants ranged from 25.2 to 33.6 kg/m² (with a median of 31.1 kg/m²), and the baseline NRS scores of the participants ranged from 3.9 to 8.4 (with a median score of 7.5). The specific literature screening process can be found in Figure 1, and the included studies and their demographic characteristics are provided in the Supplementary.

The results of the literature quality assessment indicated that four studies were deemed to present a “high” risk, two studies were judged to have a “medium” risk, and 14 studies were considered to bear a “low” risk. Detailed assessments of the literature quality can be found in the Supplementary.

The estimated values of the final model parameters are shown in Table 1. Covariate analysis revealed that age significantly influenced the parameter E_max, while the year of publication (before or after 2014) and the use of a placebo control significantly affected the parameter ET₅₀. The expression for the final covariate model can be seen in Equations 1, 2. Specifically, when the age of the participant was 59 years, the maximum reduction in the NRS from baseline following tapentadol intervention was 56.7%. For participants aged 45 and 65 years, the maximum reductions in the NRS from baseline post-tapentadol intervention were 40.6% and 63.7%, respectively. The latter represented a 23.1% higher reduction compared to the former. When the year of publication was before 2014 (i.e., PY = 0 in Equation 2) and the trial was not placebo-controlled (i.e., Placebo = 0 in Equation 2), the ET₅₀ value was 0.984 weeks. When the year of publication was after 2014 (i.e., PY = 1 in Equation 1), the ET₅₀ value for tapentadol extended by 1.71 weeks. In placebo-controlled trials (i.e., Placebo = 1 in Equation 2), the ET₅₀ value for tapentadol extended by 1.19 weeks.(1)Emax=−56.7×(Age59)1.21(2)ET50=0.984×(1+1.74×PY)+1.19×PLACEBO

The final model exhibited a robust goodness-of-fit without an evident bias, as shown in the supplementary. An internal validation was conducted utilizing 1,000 iterations of bootstrap resampling, which resulted in successful parameter estimation in 924 iterations. Upon analyzing the results from 924 iterations, it was found that the median of the parameter estimates derived from the bootstrap was in alignment with the parameter estimates from the original dataset (Table 1). This correspondence attests to the stability of the model parameter estimation. The VPC results (Figure 2) revealed that the 95% confidence interval (CI) of the model predictions encompassed the majority of the observed values, suggesting the model's strong predictive performance.

Based on the final model, we simulated the change rate in NRS from baseline following tapentadol intervention for trial participants at three age levels (45, 55, and 65 years). During the simulation, the year of publication was fixed as post-2014, and the trial design was set to placebo-controlled. Taking the age of 55 years as an example, the change rates in NRS from baseline at 1 week, 4 weeks, 8 weeks, 12 weeks, and 16 weeks post-treatment were 10.6%, 26.3%, 35.0%, 39.3%, and 41.8% respectively. These represented 17.1%, 50.5%, 67.1%, 75.4%, and 80.2% of their E_max value (52.1%). When the ages of the trial participants were 45, 55, and 65 years, the change rates in NRS from baseline at the 16th week were 32.8%, 41.8%, and 51.2% respectively (Figures 3a,d).

We simulated the change rate in NRS from baseline following tapentadol intervention for studies published before and after 2014. During the simulation, the age was fixed at 55 years, and the trial design was set to placebo-controlled. For studies published before 2014, the change rates in NRS from baseline at 1 week, 4 weeks, 8 weeks, 12 weeks, and 16 weeks post-treatment were 16.4%, 33.7%, 41.0%, 44.1%, and 45.9%, respectively. These represented 31.4%, 64.7%, 78.7%, 84.6%, and 88.1% of their E_max value (52.1%). The efficacy plateau for participants in studies published before 2014 appeared approximately at the 8th week. For studies published after 2014, the change rates in NRS from baseline at 1 week, 4 weeks, 8 weeks, 12 weeks, and 16 weeks post-treatment were 10.6%, 26.3%, 35.0%, 39.3%, and 41.8%, respectively. These represented 17.1%, 50.5%, 67.1%, 75.4%, and 80.2% of their E_max value (52.1%). The efficacy plateau for participants in studies published after 2014 appeared approximately at the 16th week (Figures 3b,e).

This study found that the design of placebo control significantly impacts the efficacy of tapentadol. We simulated the typical efficacy values of tapentadol under both placebo-controlled and non-placebo-controlled conditions. During the simulation, the age was fixed at 55 years, and the year of publication was set as post-2014. In non-placebo-controlled studies, the change rates in NRS from baseline at 1 week, 4 weeks, 8 weeks, 12 weeks, and 16 weeks post-treatment were 14.0%, 31.0%, 38.9%, 42.4%, and 44.5%, respectively. These represented 26.9%, 59.5%, 74.7%, 81.4%, and 85.4% of their E_max value (52.1%). The efficacy plateau for participants in non-placebo-controlled studies appeared approximately at the 12th week. In placebo-controlled studies, the change rates in NRS from baseline at 1 week, 4 weeks, 8 weeks, 12 weeks, and 16 weeks post-treatment were 10.6%, 26.3%, 35.0%, 39.3%, and 41.8%, respectively. These represented 17.1%, 50.5%, 67.1%, 75.4%, and 80.2% of their E_max value (52.1%). The efficacy plateau for participants in placebo-controlled studies appeared approximately at the 16th week (Figures 3c,f).

This study investigated the trend of other factors influencing the efficacy of tapentadol, aside from covariates, through subgroup analysis. In the subgroup analysis, the age of participants was adjusted to 55 years, the year of publication was adjusted to post-2014, and the trial design was adjusted to placebo-controlled to eliminate the influence of heterogeneity at the covariate level on the results of the subgroup analysis. The results showed that a higher titrated doses of tapentadol, a higher baseline NRS level of the subjects, a sustained-release formulation of tapentadol, and a smaller proportion of male subjects tended to have better efficacy. However, due to large variations, the differences were not statistically significant. For example, at 24 weeks with other factors controlled, the percentage change in NRS from baseline varied significantly across different conditions. Tapentadol doses below 175 mg resulted in a change of −48.9 (95% CI: −55.9 to −42.5), while doses of 175 mg or greater showed a change of −52.6 (95% CI: −59.3 to −46.4). Similarly, baseline NRS values below 7.36 led to a change of −48.7 (95% CI: −58.9 to −39.9), compared to −52.9 (95% CI: −59.2 to −46.6) for values of 7.36 or higher. Immediate-release (IR) formulations had a change of −48.7 (95% CI: −54.2 to −43.5), whereas extended-release (ER) formulations showed −54.6 (95% CI: −64.1 to −45.0). When the male ratio was below 42%, the change was −52.8 (95% CI: −63.3 to −42.9), vs. −49.6 (95% CI: −55.6 to −43.9) for a ratio above 42%. Additionally, non-back pain resulted in a change of −41.2 (95% CI: −43.6 to −38.8), compared to −53.1 (95% CI: −58.7 to −47.7) for back pain (Figure 4). These variations underline the influence of dosage, baseline severity, formulation, gender distribution, and type of pain on the effectiveness of tapentadol. Moreover, we found significant differences in the efficacy of tapentadol for different pain locations. The efficacy of tapentadol for lower back pain was significantly better than for non-back pain. For instance, at 52 weeks, the NRS reduction rates for the two groups differed by about 15%.