Efficacy and safety of various drug combinations in treating plaque Psoriasis: A meta-analysis

2.1 Data search and extraction

We performed a systematic literature study using Mendeley software to identify randomized controlled trials that compared the efficacy and safety of different drug combinations for plaque psoriasis and searched several stand-alone databases, including Google Scholar, PubMed, Pub Chem, Wiley library, MEDLINE, JDD Library, and the Cochrane Library. Considering the rapid evolution of knowledge and methodological standards in psoriasis research, particularly over the past two decades, we chose to limit our search to studies published after 2004. This ensures that our analysis reflects the most recent and methodologically sound evidence relevant to the current clinical practice.

Treatment regimens between the time period of 8 and 24 weeks were chosen based on the understanding that studies conducted over shorter durations (<8 weeks) may not capture the full spectrum of treatment effects, potentially leading to biased conclusions. Conversely, long-term studies (>24 weeks) may introduce confounding variables or dropout rates, thereby compromising the reliability of the findings. Additionally, by encompassing a moderate timeframe, research within the 8–24-week range enables a balanced evaluation of both short-term and intermediate-term outcomes. This timeframe minimizes bias and ensures a comprehensive understanding of the drug’s efficacy and safety profile over a meaningful treatment period.

Some of the keywords included meta-analysis, safety, PASI 75, (psoriasis OR plaque psoriasis) AND (combination therapy OR drug combination OR poly-therapy) AND (randomized controlled trial OR clinical trial OR randomized trial).

As per the protocol, the reference lists of relevant articles and their reviews were thoroughly analysed and shortlisted for additional studies. Duplicates were eliminated using the Mendeley method.

Desktop v1.19.8 and screening the titles and abstracts of the remaining records for eligibility.

The PICO framework was used to formulate the research question and define the inclusion and exclusion criteria for the studies. The PICO elements were:

2.2 Inclusion criteria

The years of publication were less than 20 years. Plaque Psoriasis (which also includes Palmo-planter psoriasis and psoriasis vulgaris), randomized controlled trials, parallel group, 8-24 weeks treatment regimes, and combination drugs. Articles written in English. Body surface involvement was ≥5%. We included studies that reported the proportion of patients who achieved at least 75% improvement in the Psoriasis Area and Severity Index (PASI 75) and the incidence of adverse events; studies with comparable baseline characteristics were included.¹⁶

2.3 Exclusion criteria

Abstract/title with psoriatic arthritis, guttate arthritis, erythrodermic psoriasis, non-randomized controlled trial papers (observational studies, retrospective studies, case reports, and review articles), >24 weeks treatment regimes, involvement of paediatric patients, non-English articles, studies that did not compare drug combinations with placebo or monotherapy.

We excluded studies that did not report the proportion of patients who achieved at least 75% improvement in the Psoriasis Area and Severity Index (PASI 75), studies without comparable baseline characteristics, large sample sizes leading to heterogeneity, and trials that included patients with ongoing previous treatment regimens.

Indirect comparisons were performed to compare the efficacy and safety of different drug combinations for plaque psoriasis because there were insufficient trials that directly compared them with placebo. We used trials that evaluated the combination of drugs A and B versus drug B monotherapy as a surrogate for drug A combination versus placebo. The following assumptions were made for indirect comparisons:

Five reviewers independently extracted data from the included studies using a standardized form. We extracted the following information: study characteristics (authors, year, country, design, duration, and sample size), intervention details (drugs, doses, frequency, and route of administration), and outcome measures (PASI 75, adverse events). The Cochrane risk of bias tool (RoB 2) was used to evaluate the methodological quality of the studies. Any discrepancies were resolved by discussion or consultation with other reviewers. We mailed the authors of the original studies for missing or unclear data.

We used WebPlotDigitizer v4.7 and Python Plotly v5.18.0 to extract the PASI 75 data from the Morita paper, which did not report the exact data numbers. We calculated the odds ratio (OR) and 95% confidence interval (CI) for each study by using the extracted data. We performed a meta-analysis using a random-effects model to estimate the pooled OR and 95% CI for the comparison of drug combinations with placebo or monotherapy. We assessed heterogeneity across studies using the I² statistic and the chi-square test.

We explored the potential sources of heterogeneity using subgroup analyses, based on the following variables: disease severity, including indices such as PASI 75, and risk of bias inherent in the studies under review. To evaluate the safety profile of the medications, we gathered data on adverse events (AE), including the prevalence of common side effects, such as abnormal liver enzymes, nausea, headache, and fatigue among the patient cohorts.

Publication bias was assessed using funnel plots. RevMan v5.4.1 , was used for all statistical analyses. We followed the PRISMA guidelines for reporting systematic reviews and meta-analyses by creating the PRISMA checklist.

3.1 Literature results

A total of 154 records were initially identified through a comprehensive search across the following databases: PubMed (n = 15), JDD (n = 40), Cochrane Library (n = 20), Wiley Online Library (n = 30), and Google Scholar (n = 49). After removal of duplicates, 100 unique records were screened, leading to the exclusion of 69 articles for reasons such as not being randomized controlled trials (RCTs), didn’t utilize PASI as a way to measure efficacy,^17,18 lacking relevance to psoriasis, or focusing on drugs not covered in this review. Subsequently, 27 articles were included in the full-text review, of which 17 met our inclusion criteria and were consequently incorporated into this meta-analysis (refer Figure 1). The details of the materials are listed in Table 1.

Figure 1. PRISMA flowchart.

3.2 Risk of bias

In most areas, most ITT trials had a low risk of bias, indicating a high level of scientific rigor. Nonetheless, a few issues were observed in terms of deviations from the intended interventions, indicating potential areas for future investigations. The fact that no domain was predominantly related to a considerable risk of bias demonstrated the overall reliability of the included RCTs. Furthermore, one study that used PPS reported a low likelihood of bias. Because of the prevalence of open-label and single-blinded trials, the randomization domain may be more prone to bias than the others. However, because meta-analyses are intrinsically heterogeneous,¹⁹ they emphasize the importance of carefully assessing diversity because they do not automatically imply biased conclusions (refer Figures 2, 3, 4 and 5).

Figure 2. Risk of Bias 2 graph of ITT studies.

Figure 3. Risk of Bias 2 graph of PPS study.

Figure 4. Risk of Bias 2 Summary of ITT study.

Figure 5. Risk of Bias 2 summary of PPS study.

3.3 Funnel plot interpretation

Our funnel plot analysis revealed a symmetric distribution along the x-axis, indicating that the studies were evenly distributed across the range of standard errors (SE) of the logarithm of the odds ratio (log [OR]). However, upon closer examination, we found that the studies were grouped together and had higher SE (log [OR]) values. The observed clustering suggests heterogeneity among the included studies or variability due to differences in sample size (refer Figure 6).

Figure 6. Funnel Plot analysis for included studies.

3.4 Efficacy analysis

Our analysis used the Mantel-Haenszel random-effects model, which we considered well suited for achieving dichotomous outcomes, as well as better adjusting for study size, data stratification, and generalization across diverse populations. As is well known, the PASI 75 response was regularly used to assess the severity and prognosis of psoriasis patients, and it fulfilled therapeutic expectations in the vast majority of these cases. Therefore, PASI 75 was utilized in this meta-analysis to evaluate the therapeutic efficacy of pharmaceutical combinations.

Our investigation found that combination therapy significantly increased the likelihood of positive outcomes by 2.53 times compared with controls (OR = 2.53, 95% CI: 1.94-3.30, p < 0.00001), refer Figure 7. The test for subgroup differences revealed no statistically significant variations in treatment effects (p = 0.14). However, the I-squared value (34%) indicated some heterogeneity across subgroups, necessitating further investigation using larger datasets.

Figure 7. Forest Plot of efficacy of treatments in combination therapy verses Control therapy in patients with Plaque Psoriasis.

Pioglitazone emerged as the frontrunner among the tested medication combinations, with an impressive odds ratio of 4.92 (95% CI: 2.19-11.05) and a very significant p-value of 0.0001. This study revealed a strong link between pioglitazone combination drugs and better outcomes compared with controls.

Several additional drug combinations showed promising results, with odds ratios of > 2 for methotrexate, cyclosporine, clobetasol propionate, and fumarate. These findings require further examination to confirm their potential benefits owing to data-size constraints.

Data availability for Acitretin and Cetirizine combinations was limited, with only a single study examining each. Although their respective odds ratios (0.98 and 1.00) did not reach statistical significance, drawing definitive conclusions based on single studies would be premature. This underscores the need for further research with larger sample sizes to assess their efficacy comprehensively.

Forest plots and subgroup analysis findings for the efficacy results are provided for easy reference.

3.5 Safety analysis

In our study exploring different drug combinations, we closely examined the side effects they might cause, such as abnormal liver enzymes, headaches, nausea, and fatigue.

3.5.1 Abnormal liver enzyme findings

In our meta-analysis, we found that the overall odds ratio (OR) was 1.89 (95% CI: 0.69-5.22), indicating a nearly two-fold increase in risk; however, when we looked at the overall effect of the drug combinations on abnormal liver enzyme findings, the results were not statistically significant (P = 0.22). This suggests that based on the available evidence, we cannot definitively conclude that the drug combinations had a significant impact on abnormal liver enzyme readings. However, there was significant heterogeneity between trials (I² = 60%, P = 0.03), which could be related to variations in the study techniques, patient demographics, or treatment durations. It is also worth mentioning that²⁰ this study found the highest OR of 6.16, indicating a six-fold increase in ALE risk among patients receiving combination drugs (refer Figure 8).

Figure 8. Forest Plot of Elevated/Abnormal liver enzymes in Combination Therapy verses control therapy patients with Plaque Psoriasis.

3.5.2 Nausea findings

In our investigation, we found that the experimental group had a slightly higher incidence of nausea (6.9%, 46 of 667 patients) than the control group (5.0%, 33 of 653). However, the odds ratio of 1.28, with a 95% confidence interval of 0.77 to 2.14, and a p-value greater than 0.05, which is above the customary threshold of 0.05 for statistical significance, indicate that this difference is not statistically significant. As a result, our findings are insufficient to establish that the incidence of nausea differed considerably between the two treatment groups. This data highlights the need for more research to identify whether experimental treatment is related to a higher incidence of nausea than control treatment (refer Figure 9).

Figure 9. Forest Plot of Nausea occurrence in Combination Therapy verses control therapy patients with Plaque Psoriasis.

3.5.3 Headache findings

The incidence of headache appeared to be equivalent between the experimental and control groups. In the experimental group of 746 patients, 7.4% (n = 55) complained of headaches, whereas in the control group of 741 patients, 8.1% (n = 60) reported comparable events. Overall, there appeared to be no statistically significant difference in the frequency of headaches between the combination therapy and control groups (refer Figure 10). This is seen by the overall odds ratio (OR) of 0.90 (95% CI: 0.60, 1.33), which is very near to 1.0. Furthermore, the p-value for the test of the overall effect is 0.58, which is significantly higher than the widely accepted significance level of 0.05. Furthermore, the heterogeneity statistic was noteworthy at (I² = 0%).

Figure 10. Forest Plot of Headache occurrence in Combination Therapy verses control therapy patients with Plaque Psoriasis.

3.5.4 Fatigue findings

Our study included eight trials that examined the incidence of fatigue with different medication combinations for the treatment of plaque psoriasis. The results revealed that the experimental group experienced considerably less fatigue than the control group did. The odds ratio for fatigue was 0.89 (95% confidence interval, 0.41 to 1.90), indicating that patients in the experimental group were 11% less likely to feel weariness than patients in the control group. This difference was not statistically significant (P = 0.75). There was no significant heterogeneity among the studies (I² = 0%), indicating consistent results across all the investigations (refer Figure 11).

Figure 11. Forest Plot of Fatigue occurrence in Combination Therapy verses control therapy patients with Plaque Psoriasis.