This meta-analysis followed the protocol registered in the International Prospective Register of Systematic Reviews (PROSPERO) and adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Page et al., 2021) (Figure 1). The PICOS (patients, intervention, comparison, outcomes, and study design) strategy was used to select the studies: P) adult patients diagnosed with GAD. The inclusion criteria were clear and required patients to meet the DSM-IV criteria for GAD using a structured interview, such as the Mini-International Neuropsychiatric Interview (MINI) (Sheehan et al., 1998). On the other hand, exclusion criteria were established to exclude patients with other axis I disorders, except for dysthymia, specific phobia, social phobia, somatization disorder, or a history of major depressive disorder. Patients at risk of suicide, with current severity of suicidal ideation, those with a current diagnosis of seizure disorder, or a history of bipolar disorder, schizophrenia, psychotic disorder, or factitious disorder were also excluded. Patients in the intervention group (I) were treated with pregabalin, while those in the comparator group (C) were treated with benzodiazepines, SSRIs, SNRIs, or placebo. The outcomes of interest (O) were efficacy and safety, evaluated using different scales, and adverse events, respectively. The included studies (S) were randomized clinical trials.

To ensure the quality and coherence of this study, rigorous exclusion criteria were applied. First, non-randomized studies, such as case reports, case series, editorials, cohort studies, case-control studies, cross-sectional studies, and protocols, were excluded. This decision was made to ensure that the data obtained were from randomized clinical trials, which are considered the gold standard in scientific research. Additionally, pediatric populations were excluded, as the focus of this study was on adult patients. Duplicate studies were excluded to avoid data duplication and ensure the integrity of the analysis. In the case of duplicate studies reporting related information in different publications, the decision was made to group them into a single study to avoid duplication of basic characteristics and to ensure that each variable was accounted for only once. However, relevant information was extracted from each duplicate study for inclusion in the analysis. This allowed the utilization of complementary data and ensured a comprehensive evaluation of the available results without incurring unnecessary redundancy. Studies with incomplete data or those in which variables were not comparable were also discarded, as this could have affected the validity and reliability of the results.

Comprehensive searches for information were conducted across multiple sources, including PubMed, EMBASE, SCOPUS, and the Cochrane Library, with no restrictions on date or language. Additionally, a manual search of the reference lists from the included studies was performed, which allowed for the identification of additional studies that might have been missed in the electronic searches.

The search strategy used in this study targeted the keywords “pregabalin” or “Lyrica” and “anxiety” (Supplementary Material 1). Two reviewers independently screened the titles and abstracts of the selected studies. Any discrepancies or disagreements between reviewers were resolved through discussion with a third reviewer.

Two reviewers conducted the screening process, and any disagreements were resolved by consensus with a third reviewer. Baseline characteristic variables extracted from the included studies included study details, region, period, follow-up duration, sample size, age, female participants, duration of GAD, doses of pregabalin or control, conflict of interest (COI), and funding sources. Additional information was obtained regarding the treatment schemes and drug-free periods.

The primary outcome measures were the Hamilton Anxiety Rating Scale (HAM-A) score and adverse events. The HAM-A is a widely used clinical tool for measuring the severity of anxiety symptoms in patients, ranging from 0 to 56, with higher scores indicating greater anxiety severity (Thompson, 2015). It was also the responder rate according to the HAM-A scale, defined as a 50% or greater reduction from baseline in the HAM-A total score or a HAM-A score of 7 or less. The subdomains of the HAM-A scale, namely, psychic and somatic, were also compared. The Clinical Global Impression Improvement Scale (CGI-I) was also extracted, which measures the overall change in a patient’s clinical condition using a 7-point scale, with 1 indicating “very much improved” and 7 indicating “very much worse” (Busner and Targum, 2007). The responder rate to the CGI-I was defined as a CGI-I score of 2 or less (much or very much improved). Furthermore, the number of patients discontinuing treatment in each group, reasons for discontinuation (lack of efficacy or adverse events), costs of the drug, total costs, and quality-adjusted life years (QALYs) were evaluated. Sleep disturbances were assessed using the HAM-A (insomnia subdomain), and the Medical Outcomes Study Sleep Scale (MOS-Sleep) score was also evaluated. It incorporates 12 items covering different aspects of sleep quality and sleep disturbances. These items assess sleep disturbances, sleep adequacy, sleepiness, sleep quantity, snoring, and awakening with shortness of breath or headache (Viala-Danten et al., 2008). Finally, the adverse events were compared.

The risk of bias assessment was conducted by two reviewers using the Cochrane Collaboration tool for randomized controlled trials in the Review Manager 5.4 software. The following six domains were considered: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessments, incomplete outcome data, and selective reporting. Each domain was evaluated and assigned a judgment of a low, high, or unclear risk of bias. The results of the risk of bias assessment are shown in Figure 2; Supplementary Material 2. Additionally, forest plots displayed the risk of bias grading for each trial, providing a comprehensive overview of the potential biases in the included studies.

Meta-analyses were conducted using Review Manager 5.4 software. Odds ratios (ORs) with corresponding 95% confidence intervals (CIs) were calculated for dichotomous variables. Continuous variables were analyzed using the mean differences (MDs) or standardized mean differences (SMDs) with 95% CIs, depending on whether the studies had compatible units or scales. Heterogeneity was assessed using the chi-square statistic and I2 index, where I2 values of 25%, 50%, and 75% represented low, moderate, and high levels of heterogeneity, respectively. A fixed-effects model was used in cases of homogeneity, whereas a random-effects model was applied in the presence of heterogeneity. Precise data points from the study figures were extracted using the WebPlotDigitizer software version 13.1.4. Missing data were addressed following the guidelines outlined in the Cochrane Handbook to ensure a comprehensive analysis (Higgins et al., 2019).

Publication bias was evaluated using funnel plots that were visually generated using Review Manager version 5.4. Funnel plots depict the effect estimates of the analyzed studies on the x-axis, while the standard errors are represented on the y-axis. These plots provide a graphical representation of potential publication bias in the included studies. In an ideal scenario without publication bias, the plot would display a symmetrical funnel shape, indicating that smaller studies with larger standard errors are scattered more widely around the estimated effect size, whereas larger studies with smaller standard errors are clustered closely to the estimated effect size. However, asymmetry in the funnel plot may suggest the presence of publication or other sources of bias. It is important to note that no formal statistical tests were conducted in this assessment of publication bias, and the interpretation was based on visual inspection of funnel plots.

Subgroup analyses were conducted based on the duration of follow-up, categorizing outcomes into short-term (2 and 4 weeks), medium-term (8 and 12 weeks), and long-term (6 months–1 year) outcomes. Subgroup analyses were performed according to dosage. Owing to heterogeneity and variation in the dosage employed across studies, the dosage was divided into low-dose (less than 300 mg/day) and high-dose (greater than 300 mg/day) pregabalin. If studies provided different dosages, the results were reported separately for each dosage type rather than in combination.

Sensitivity analyses were conducted to explore the effects of separate comparisons: pregabalin vs. placebo, pregabalin vs. benzodiazepine, pregabalin vs. selective serotonin reuptake inhibitors (SSRIs), and pregabalin vs. serotonin-noradrenaline reuptake inhibitors (SNRIs).

The Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach utilizing GRADEpro was employed to assess the certainty of the outcomes (Guyatt et al., 2013). GRADE evaluates the quality of evidence based on factors such as the study design, risk of bias, inconsistency, indirectness, imprecision, and publication bias.

The initial search yielded 939 studies. After reviewing the titles and abstracts and excluding non-randomized studies, 806 studies were eliminated, resulting in 133 articles. After excluding studies that did not compare pregabalin with placebo, other drugs for the indication of GAD, or other types of reviews, 99 studies were eliminated, leaving 34 studies. After reviewing the full text, 20 studies were eliminated because of duplication or not being used for GAD, resulting in 14 studies. After examining all references in these studies, no additional studies were added (Figure 1) (Feltner et al., 2008; Montgomery et al., 2008; Silva Miguel et al., 2013; Rickels et al., 2005; Feltner et al., 2003; Kasper et al., 2014; Pande et al., 2003; Hadley et al., 2012; Rickels et al., 2012; Álvarez et al., 2015; Cvjetkovic-Bosnjak et al., 2015; Kasper et al., 2009; Vera-Llonch et al., 2010; Montgomery et al., 2006).

Regarding the risk of bias (Figure 2), the studies were randomized; however, the mode of randomization was not reported in most cases. The evaluators of the results were not blinded, and in approximately half of the cases, patients who did not reach the end of the follow-up were not reported.

Table 1 presents the basic characteristics of the studies included. Fourteen studies and 4,822 patients were included (2,650 in the pregabalin group, 432 in the benzodiazepine group, 792 in the SSRI/SNRI group, and 948 in the placebo group). Most of the studies (7 of 13, 53.8%) were published in the United States, and the follow-up period ranged from 4 to 52 weeks. The mean age ranged from 35.7 to 72.4 years in the pregabalin group. The number of female participants, duration of symptoms, conflict of interests (COI), and funding are presented in Table 1. The therapeutic regimens are shown in Supplementary Table 1.

At 2 weeks, the HAM-A global score showed greater efficacy in the pregabalin group (MD −1.23, 95% CI −1.79 to −0.66; participants = 4,232; studies = 23; I2 = 99%). The higher dose of pregabalin (>300 mg) showed significant differences (MD −1.64, 95% CI −2.32 to −0.96; participants = 3,035; studies = 15; I2 = 100%), but no significant differences were found with doses below 300 mg (MD −0.33, 95% CI −2.18 to 1.51; participants = 1,197; studies = 8; I2 = 96%). Table 2 presents the results comparing pregabalin with placebo, benzodiazepines, and SSRIs/SNRIs. Both high and low doses of pregabalin showed significant improvement compared to placebo. Benzodiazepines demonstrated greater reduction in the HAM-A scale compared to low doses of pregabalin. However, no significant differences were found between benzodiazepines and high doses of pregabalin. However, both doses of pregabalin showed significant superiority compared to SSRIs/SNRIs.

At 4 weeks, the HAM-A global score showed a greater reduction in the pregabalin group (MD −1.12, 95% CI −1.60 to −0.63; participants = 4,168; studies = 23; I2 = 99%), with differences in favor of pregabalin at doses over 300 mg (MD −1.46, 95% CI −2.01 to −0.91; participants = 2,971; studies = 15; I2 = 99%), and no differences were found between groups with doses below 300 mg (MD −0.45, 95% CI −1.91 to 1.01; participants = 1,197; studies = 8; I2 = 95%). Table 2 presents the results comparing pregabalin with placebo, benzodiazepines, and SSRIs/SNRIs separately. Both doses of pregabalin showed significant superiority compared to placebo. No significant differences were found compared with benzodiazepines at either dose. SSRIs/SNRIs were superior to low doses of pregabalin, but no differences were found with pregabalin at high doses.

At 8 weeks, the HAM-A global score showed greater improvement in the pregabalin group (MD −2.50, 95% CI −4.21 to −0.79; participants = 1,359; studies = 5; I2 = 100%). Doses below 300 mg and above 300 mg showed greater benefit in the pregabalin group (MD −2.33, 95% CI −4.24 to −0.42; participants = 1,086; studies = 4; I2 = 100%) and (MD 1.18, 95% CI 0.48 to 1.88; participants = 1,917; studies = 2; I2 = 0%), respectively. Pregabalin was superior to placebo and SSRIs/SNRIs when divided by control type (Table 2).

At 12 weeks, the HAM-A global score showed lower scores in the pregabalin group (MD 0.99, 95% CI 0.35 to 1.63; participants = 2,302; studies = 3; I2 = 0%), with no differences observed with the higher dose of 300 mg (MD 0.10, 95% CI −1.42 to 1.62; participants = 385; studies = 1; I2 = 0%), and pregabalin showing higher scores on the HAM-A global scale compared to the control group with doses below 300 mg (MD 1.18, 95% CI 0.48 to 1.88; participants = 1,917; studies = 2; I2 = 0%). At 12 weeks, no differences were found between pregabalin and benzodiazepines for either dose. SSRIs/SNRIs were superior to low doses of pregabalin (Table 2).

At 6 months to 1 year, the HAM-A score showed significantly more favorable results in the pregabalin group (MD −3.31, 95% CI −4.30 to −2.31; participants = 831; studies = 2; I2 = 50%). Both doses above 300 mg and doses below 300 mg showed significantly greater improvement in the pregabalin group: (MD −5.00, 95% CI −7.55 to −2.45; participants = 106; studies = 1; I2 = 0%) and (MD −3.00, 95% CI −4.09 to −1.91; participants = 725; studies = 1; I2 = 100%), respectively. Pregabalin showed superiority over placebo at high doses and superiority over SSRIs/SNRIs at low doses (Table 2).

Independently from study duration, the response rate on the HAM-A scale was significantly higher in the pregabalin group (OR 1.51, 95% CI 1.31 to 1.75; participants = 3,092; studies = 13; I2 = 6%) (Figure 3). There were significant differences in favor of pregabalin both with doses above 300 mg and doses below 300 mg (OR 1.79, 95% CI 1.45 to 2.22; participants = 1,397; studies = 7; I2 = 0%) and (OR 1.32, 95% CI 1.08 to 1.60; participants = 1,695; studies = 6; I2 = 0%). When compared to different control groups, pregabalin showed significant superiority over placebo at high doses (OR 1.99, 95% CI 1.51 to 2.61; participants = 872; studies = 4; I2 = 0%) and low doses (OR 1.60, 95% CI 1.13 to 2.28; participants = 535; studies = 3; I2 = 0%). It also showed superiority over benzodiazepines at high doses (OR 1.52, 95% CI 1.08 to 2.15; participants = 525; studies = 3; I2 = 0%). However, only the low dose of pregabalin could be compared with SSRIs/SNRIs, and no significant differences were found (OR 1.26, 95% CI 0.97 to 1.65; participants = 902; studies = 1; I2 = 0%).

When the HAM-A scale was divided according to the psychic subdomain, the pregabalin group showed higher overall satisfaction scores at the end of the follow-up (MD −0.92, 95% CI −1.49 to −0.35; participants = 3,361; studies = 18; I2 = 100%). Doses above 300 mg showed significant differences in favor of pregabalin (MD −1.10, 95% CI −1.75 to −0.46; participants = 2,706; studies = 14; I2 = 100%), while doses below 300 mg did not show differences between the groups (MD −0.19, 95% CI −1.36 to 0.98; participants = 655; studies = 4; I2 = 86%). Both doses of pregabalin showed significant superiority over placebo: (MD −2.05, 95% CI −2.56 to −1.53; participants = 1,340; studies = 7; I2 = 99%) and (MD −1.21, 95% CI −1.82 to −0.59; participants = 411; studies = 2; I2 = 12%), respectively. In contrast, high doses of pregabalin showed significant superiority compared to benzodiazepines (MD −0.36, 95% CI −0.59, −0.14; participants, 687; studies, 4; I2 = 87%). There were no significant differences compared with SSRIs/SNRIs (MD 0.17, 95% CI −0.68 to 1.03; participants = 786; studies = 4; I2 = 99%).

In contrast, the HAM-A somatic domain showed better values in the pregabalin group (MD −0.64, 95% CI −1.04 to −0.25; participants, 3,361; studies, 18; I2 = 99%). The group receiving doses above 300 mg showed significant differences in favor of pregabalin (MD −0.79, 95% CI −1.23 to −0.36; participants = 2,706; studies = 14; I2 = 99%), and the group receiving doses below 300 mg did not show significant differences (MD 0.03, 95% CI −0.91 to 0.97; participants = 655; studies = 4; I2 = 77%). Both doses of pregabalin showed significant superiority over placebo: (MD -1.31, 95% CI -1.86 to −0.76; participants = 1,340; studies = 7; I2 = 99%) and (MD −0.80, 95% CI −1.43 to −0.17; participants = 411; studies = 2; I2 = 0%), respectively. In contrast, low doses of pregabalin showed significant superiority compared to benzodiazepines (MD 1.00, 95% CI 0.01 to 1.99; participants = 137; studies = 1; I2 = 0%). There were no significant differences compared with SSRIs/SNRIs (MD −0.14, 95% CI −1.03 to 0.76; participants = 786; studies = 4; I2 = 99%, p = 0.77).

CGI-I at the end of the follow-up showed significant improvement in the pregabalin group overall (MD −0.25, 95% CI −0.38 to −0.12; participants = 4,276; studies = 17; I2 = 100%) (Figure 4A). Doses above 300 mg and below 300 mg showed differences in favor of pregabalin: (MD −0.26, 95% CI −0.40 to −0.11; participants = 2,914; studies = 14; I2 = 100%) and (MD −0.24, 95% CI −0.37 to −0.12; participants = 1,362; studies = 3; I2 = 71%). Both doses of pregabalin showed significant superiority over placebo: (MD −0.50, 95% CI −0.61 to −0.40; participants = 1,304; studies = 7; I2 = 99%) and (MD −0.30, 95% CI −0.33 to −0.27; participants = 266; studies = 1; I2 = 0%). There were no significant differences compared with benzodiazepines (MD −0.07, 95% CI −0.20, 0.06; participants, 2,030; studies, 6; I2 = 98%). Low doses of pregabalin were superior to SSRIs/SNRIs (MD −0.30, 95% CI −0.42 to −0.18; participants = 725; studies = 1; I2 = 0%).

The response rate to CGI-I at the end of the follow-up was significantly higher in the pregabalin group (OR 1.33, 95% CI 1.15 to 1.55; participants = 2,733; studies = 15; I2 = 52%) (Figure 4B). Doses above 300 mg showed significant differences in favor of pregabalin (OR 1.44, 95% CI 1.22 to 1.71; participants = 2,213; studies = 11; I2 = 56%), whereas lower doses did not show significant differences (OR 0.97, 95% CI 0.69 to 1.37; participants = 520; studies = 4; I2 = 0%). High doses of pregabalin showed significant superiority over placebo (OR 1.80, 95% CI 1.43 to 2.25; participants = 1,270; studies = 6; I2 = 49%), whereas no significant differences were found compared with benzodiazepines or SSRIs/SNRIs (OR 1.07, 95% CI 0.81 to 1.41; participants = 783; studies = 5; I2 = 33%) and (OR 0.87, 95% CI 0.59 to 1.29; participants = 418; I2 = 0%), respectively.

The discontinuation rate was significantly lower in the pregabalin group (OR 0.80, 95% CI 0.70 to 0.91; participants = 5,183; studies = 23; I2 = 38%) (Figure 5). There were significant differences in favor of pregabalin for doses above 300 mg and below 300 mg: (OR 0.84, 95% CI 0.73 to 0.96; participants = 4,438; studies = 20; I2 = 37%) and (OR 0.63, 95% CI 0.47 to 0.86; participants = 745; studies = 3; I2 = 21%), respectively. The low dose of pregabalin showed a significantly lower discontinuation rate than placebo (OR 0.61, 95% CI 0.42 to 0.89; participants = 510; studies = 2; I2 = 59%). Additionally, the low dose of pregabalin showed a lower discontinuation rate compared to SSRIs/SNRIs (OR 0.68, 95% CI 0.41 to 1.13; participants = 235; studies = 1; I2 = 0%). There were no differences compared with benzodiazepines (OR 0.77, 95% CI 0.59 to 1.01; participants = 1,010; studies = 5; I2 = 50%).

Discontinuation due to adverse events did not differ between the groups (OR 0.90, 95% CI 0.76 to 1.07; participants = 5,183; studies = 23; I2 = 46%). Doses above 300 mg had a significantly lower discontinuation rate (OR 0.82, 95% CI 0.68 to 0.99; participants = 4,438; studies = 20; I2 = 46%), while there were no differences in doses below 300 mg (OR 1.43, 95% CI 0.92 to 2.21; participants = 745; studies = 3; I2 = 0%). The discontinuation rate due to adverse events with high doses of pregabalin was significantly lower for pregabalin than for SSRIs/SNRIs (OR 0.57, 95% CI 0.41 to 0.79; participants = 1,185; studies = 5; I2 = 0). There were no differences when compared with placebo (OR 1.11, 95% CI 0.86 to 1.45; participants = 2,739; studies = 12; I2 = 28%) or benzodiazepines (OR 0.91, 95% CI 0.62 to 1.34; participants = 1,010; studies = 5; I2 = 51%).

The discontinuation rate due to a lack of efficacy was significantly lower in the pregabalin group (OR 0.58, 95% CI 0.44 to 0.77; participants, 4,845; studies, 22; I2 = 0%). High doses showed a significantly lower discontinuation rate in the pregabalin group (OR 0.54, 95% CI 0.39 to 0.75; participants = 4,100; studies = 19; I2 = 0%), whereas low doses did not show significant differences (OR 0.68, 95% CI 0.41 to 1.13; participants = 745; studies = 3; I2 = 0%). The discontinuation rate due to lack of efficacy was significantly lower with pregabalin than with placebo (OR 0.44, 95% CI 0.31 to 0.63; participants = 2,650; studies = 12; I2 = 0%). There were no differences between pregabalin and benzodiazepines (OR 0.76, 95% CI 0.43 to 1.35; participants = 1,010; studies = 5; I2 = 0%) or SSRIs/SNRIs (OR 1.19, 95% CI 0.57 to 2.46; participants = 1,185; studies = 5; I2 = 0%).

Adverse events are shown in Tables 3, 4. Compared to SSRI/SNRI, high-dose pregabalin (Table 3) showed a significantly lower rate of nausea (OR 0.36, 95% CI 0.24–0.54), blurred vision (OR 0.32, 95% CI 0.15–0.71), asthenia (OR 0.32, 95% CI 0.15–0.71), insomnia (OR 0.34, 95% CI 0.15–0.74), and ataxia (OR 0.36, 95% CI 0.14–0.94). In contrast, SSRI/SNRI showed a significantly lower rate of somnolence (OR 4.17, 95% CI 1.20–14.55) and dizziness (OR 2.35, 95% CI 1.55–3.57). Low doses of pregabalin could not be compared with those of SSRIs/SNRIs.

On the other hand, high doses of pregabalin (Table 4) showed a significantly lower rate of somnolence (OR 0.69, 95% CI 0.53–0.89), blurred vision (OR 0.58, 95% CI 0.36–0.93), and asthenia (OR 0.58, 95% CI 0.36–0.93) compared to benzodiazepines. However, benzodiazepines showed a significantly lower rate of dizziness (OR 2.90, 95% CI, 2.16–3.88), dry mouth (OR 3.67, 95% CI, 1.85–7.30), infection (OR 1.87, 95% CI 1.19–2.94), and constipation (OR 2.15, 95% CI, 1.07–4.32). Low doses of pregabalin (Table 4) resulted in a significantly lower rate of somnolence (OR 0.39, 95% CI 0.18–0.83), asthenia (OR 0.39, 95% CI 0.18–0.86), and confusion (OR 0.15, 95% CI 0.03–0.70) compared to benzodiazepines. In contrast, benzodiazepines showed a lower incidence of headache than low-dose pregabalin (OR 1.54, 95% CI 1.06–2.23).

The total cost was significantly higher in the pregabalin group than in the SSRI/SNRI group (MD 476.23, 95% CI 94.91 to 857.54; participants = 1,231; studies = 3; I2 = 73%) (Figure 6A). This was mainly due to the significantly higher drug costs in the pregabalin group than in the SSRI/SNRI group (MD 478.55, 95% CI 113.77 843.32; participants = 1,231; studies = 3; I2 = 99%) (Figure 6B). However, the cost-effectiveness, evaluated using QALYs, was significantly higher in the pregabalin group than in the SSRI/SNRI group (MD 0.02, 95% CI 0.01 to 0.03; participants = 1,162; studies = 2; I2 = 0%) (Figure 6C).

Publication bias (Figure 7) was high for the HAM-A mean difference at different follow-up periods and adverse events. No publication bias was observed in the remaining variables.

Table 5 presents the results of the GRADE assessment. There was a high level of certainty for the responder rate on the HAM-A and CGI-I scales as well as for the discontinuation rate. Regarding the mean difference in HAM-A scores at 8 weeks, the certainty was moderate due to increased publication bias. The short-term HAM-A variable (2 weeks) showed low certainty owing to high publication bias and wide variability in the results.

Several limitations were encountered during the development of this study that must be considered when interpreting the results. First, missing data had to be handled according to Cochrane guidelines to estimate standard deviations. While studies provided information in graphical form, software (WebPlotDigitizer) was used to extract the necessary data. Additionally, for some variables measured at follow-up, there was a mix of follow-ups conducted between 4 and 8 weeks, requiring adjustments to closer follow-ups. Importantly, the lack of long-term outcomes, for example, beyond 6 months, limits our complete understanding of the long-term effects of pregabalin. Furthermore, some studies did not clearly provide patient numbers in each group, hindering comparative analysis. It was not possible to compare low-dose pregabalin adverse events with SSRI/SNRIs, and some adverse events, such as weight gain, were reported in only one study, preventing their inclusion in the meta-analysis. In addition, studies assessing the effects of different drugs on specific side effects, including weight gain, are insufficient. In conducting subgroup analyses, some subgroups had a limited number of included studies, raising questions about robustness of the results. It would have been interesting to include over 10 studies to enable meta-regressions exploring potential confounders, as in some cases, factors could only be visually studied through graph inspection.

This meta-analysis has several significant strengths that contribute to its robustness and relevance. First, it stands out as the most recent study to date, implying that it includes the most up-to-date information on the topic in question. Additionally, notable inclusion of a large number of studies was achieved, increasing the representativeness and generalizability of the results obtained. This wide study inclusion also allowed for greater control of heterogeneity, thus strengthening the internal validity and reliability of the findings. An additional strength of this meta-analysis is the variety of comparisons made. Beyond placebos, first-line drugs, such as benzodiazepines and SSRI/SNRIs, were included, providing a more complete and practical perspective of pregabalin’s efficacy relative to conventional treatments. Another key strength of this study is the presentation of the results. Detailed information on the primary variable of interest across different follow-up periods was provided, allowing for a more accurate evaluation of the temporal evolution of pregabalin effects. In addition, the inclusion criteria were well defined for GAD, and the definitions of response rates used were homogeneous, allowing for more precise comparability between studies. Additionally, an exhaustive exploration of subgroups considering dose and control drug types was performed. This provides greater granularity in the analysis and allows for the identification of possible patterns or differences in treatment responses based on these variables.