This randomized controlled trial enrolled patients with moderate-to-severe plaque psoriasis (PASI ≥12) who were diagnosed by dermatologists at the Hospital of Skin Diseases, Chinese Academy of Medical Sciences, between October 2020 and November 2024.

Inclusion criteria were as follows: (1) age 18–80 years, male or female; (2) diagnosis of moderate-to-severe plaque psoriasis, defined as PASI ≥12 or body surface area (BSA) involvement ≥10%; and (3) inadequate response to topical therapies, phototherapy, or both, with no prior use of acitretin, cyclosporine, or biologic agents within the preceding 4 to 6 weeks.

Exclusion criteria were as follows: (1) severe systemic diseases or uncontrolled hypertension; (2) clinically significant hepatic or renal dysfunction; (3) pregnancy, lactation, or planning to become pregnant during the study period; (4) use of systemic glucocorticoids, other conventional immunosuppressants, or biologic agents within 4 weeks prior to enrollment; and (5) known hypersensitivity to acitretin or cyclosporine.

The study protocol was approved by the Ethics Committee of the Hospital of Skin Diseases, Chinese Academy of Medical Sciences, and written informed consent was obtained from all participants prior to enrollment. The trial was registered at the Chinese Clinical Trial Registry (ChiCTR-OPN-17013383).

Eligible participants were randomly assigned in a 1:1:1 ratio to one of three treatment groups: the combination therapy group (acitretin plus cyclosporine), the acitretin monotherapy group, or the cyclosporine monotherapy group. Randomization was performed centrally using a computer-generated random sequence with variable block sizes concealed from investigators, prepared by an independent statistician who was not involved in patient enrollment or treatment. Allocation concealment was maintained through sealed, opaque, and sequentially numbered envelopes. Because of the distinct dosing schedules and monitoring requirements of acitretin and cyclosporine, participants and treating investigators were necessarily aware of treatment allocation (open-label design). However, all efficacy assessments were conducted by independent dermatologists who did not participate in patient care and who remained blinded to treatment assignments throughout the study.

Treatment duration ranged from 12 to 16 weeks, with a follow-up period extending to 24 weeks post-treatment initiation. All patients concurrently received topical glucocorticoid ointment, calcipotriol ointment, or calcipotriol scalp solution as adjunctive therapies. The initial dosage of acitretin was 0.4 mg/kg/day, while cyclosporine was administered at 3 mg/kg/day (divided into two doses daily). Dose escalation was permitted to 0.5 mg/kg/day for acitretin and up to 5 mg/kg/day for cyclosporine if the PASI score reduction was less than 25% after 4 weeks of treatment. Medication doses could be reduced or discontinued at any point if adverse events (AEs) occurred, in accordance with established guidelines (12, 13). Patients received only topical therapies or phototherapy for 4 to 6 weeks during the screening period; no additional active systemic therapies were permitted during treatment, except emollients and topical treatments. Additional active treatments were allowed during the follow-up period, if clinically necessary.

To evaluate cyclosporine exposure, venous blood samples were collected immediately before the morning dose on Day 4 (i.e., prior to the 7th total dose, under twice-daily administration), representing the trough concentration (C0) at steady state. This timing was based on the pharmacokinetic profile of cyclosporine: with a half-life of approximately 6.3 h in healthy subjects (Sandimmun^®, Novartis drug label), steady-state levels are expected to be reached within 4–5 half-lives (14, 15). Therefore, by Day 4, steady-state is considered to have been achieved for most patients. Cyclosporine concentrations were measured in both the combination therapy and cyclosporine monotherapy groups. Quantification was performed using liquid chromatography–tandem mass spectrometry (LC-MS/MS) with an AB SCIEX Triple Quad^™ 4500MD system. The process of patient enrollment, treatment and follow-up is shown in Figure 1.

The primary outcomes were the proportions of patients achieving at least a 75% (PASI75) and 90% (PASI90) reduction from baseline in the PASI. The key secondary outcome was the mean change in PASI score from baseline at each follow-up visit. Additional secondary outcomes included BSA, DLQI scores, and the incidence of AEs. All outcomes were assessed by independent dermatologists who were blinded to treatment allocation.

Differences in mean PASI, BSA, and DLQI scores among the three treatment groups from baseline to week 24 were assessed using one-way analysis of variance (ANOVA) or non-parametric tests, as appropriate. Proportions of patients achieving PASI75 and PASI90 responses were compared between groups using the chi-square test or Fisher’s exact test, as appropriate. All statistical analyses were performed with SPSS software, version 20.0 (IBM Corp., Armonk, NY, United States). A two-sided p-value of <0.05 was considered statistically significant.

A total of 351 patients were screened for eligibility, of whom 6 were excluded (2 owing to revised diagnoses not meeting inclusion criteria, 3 who declined to provide informed consent, and 1 who refused study medication). Consequently, 345 patients were randomized in a 1:1:1 ratio into the combination therapy group (n = 115), the acitretin monotherapy group (n = 115), and the cyclosporine monotherapy group (n = 115).

During follow-up, 6 patients in the combination therapy group, 8 in the acitretin group, and 6 in the cyclosporine group were lost to follow-up. In addition, 7 patients in the combination therapy group, 5 in the acitretin group, and 8 in the cyclosporine group discontinued treatment (primarily because of adverse events or by switching to biologic therapy). Details of patient disposition are presented in Figure 2. Ultimately, 305 patients completed the study and were included in the final analysis: 102 in the combination therapy group, 102 in the acitretin group, and 101 in the cyclosporine group. Baseline characteristics of the patients are summarized in Table 1, showing no statistically significant differences in disease duration or baseline PASI scores among the groups.

Figure 3 illustrates the changes in PASI scores from baseline (week 0) through week 24 across the three treatment groups. Between weeks 2 and 8, patients in the combination therapy group exhibited significantly lower mean PASI scores compared with those receiving either cyclosporine A or acitretin monotherapy (p < 0.001). From week 12 onward, overall differences in mean PASI scores among the three groups were attenuated and no longer statistically significant.

Regarding categorical outcomes, PASI75 responses showed early and sustained differences (Table 2 and Figure 4). By week 4, over 60% of patients in the combination therapy group had achieved PASI75, compared with fewer than 25% in either monotherapy arm (p < 0.001). The advantage of combination therapy was maintained through week 8 and week 12, when nearly 90% of patients in the combination group achieved PASI75, significantly higher than the acitretin and cyclosporine groups (p < 0.001). At week 24, PASI75 response rates remained highest in the combination therapy group (91.18%) compared with acitretin (77.45%) and cyclosporine (85.29%) (p = 0.022).

For PASI90 (Table 2 and Figure 5), early separation was also observed. By week 4, one-fifth of patients in the combination therapy group had already achieved PASI90, whereas responses were rare in the monotherapy groups. These early advantages were maintained at week 12, two-thirds of patients in the combination group reached PASI90 compared with ~15–30% in the monotherapy groups (p < 0.001). At week 24, PASI90 responses remained highest in the combination therapy group (77.45%), compared with 34.31 and 58.82% in the acitretin and cyclosporine groups, respectively (p < 0.001).

At baseline, mean BSA involvement was 0.55 ± 0.19 in the acitretin group, 0.63 ± 0.24 in the combination group, and 0.57 ± 0.20 in the cyclosporine group (p = 0.018) (Table 1), indicating that patients in the combination group had slightly greater disease burden at study entry. From week 4 onward, BSA improved significantly in all three groups, with the combination group showing greater reductions compared with both monotherapy groups (week 4: 0.265 ± 0.162 vs. 0.298 ± 0.147 and 0.314 ± 0.158, respectively; p < 0.001). These between-group differences in BSA were maintained through week 8, consistent with the pattern observed for PASI responses. By weeks 12–16, as clinical symptoms improved and many patients tapered or discontinued the study drug, BSA scores converged and no longer showed significant differences among groups (Figure 6).

Similarly, DLQI scores decreased markedly in all groups during treatment, reflecting improved quality of life. The greatest improvement was observed in the combination therapy group during the first 8 weeks, but differences among groups were not statistically significant after weeks 12–16 of treatment and remained comparable during follow-up to week 24 (p = 0.260) (Figure 7).

Dose escalation after 4 weeks was required for 13 patients in the acitretin group, 10 in the cyclosporine group, and 5 in the combination group. After 24 weeks of follow-up, relapse occurred in 15 patients in the acitretin group, 12 in the cyclosporine group, and 6 in the combination group.

Plasma cyclosporine concentrations were significantly lower in the combination therapy group (104.23 ± 37.26 ng/mL) than in the cyclosporine monotherapy group (161.67 ± 63.68 ng/mL, p < 0.001). The mean acitretin dose was also lower in the combination group (0.287 ± 0.061 mg/kg/day) compared with the acitretin group (0.411 ± 0.092 mg/kg/day, p < 0.001), as was the mean cyclosporine dose (2.815 ± 0.462 vs. 3.387 ± 0.584 mg/kg/day, p < 0.001) (Table 1).

At least one AE was reported by 49 patients in the acitretin group (48.04%), 34 patients in the cyclosporine group (33.66%), and 37 patients in the combination therapy group (36.27%) (Table 3). Most AEs were mild to moderate and resolved within 1–2 weeks after treatment discontinuation.

Hypertension was significantly more common in the cyclosporine group (13 of 101, 12.87%) than in the acitretin (2 of 102, 1.96%) and combination groups (10 of 102, 9.80%; p = 0.015). Three patients required antihypertensive treatment, with a mean cyclosporine dose of 3.10 ± 0.60 mg/kg/day in both cyclosporine-containing groups. Blood pressure returned to normal after cyclosporine discontinuation, and antihypertensive medications were concurrently withdrawn.

Dry lips, dry eyes, or dry skin were significantly more frequent in the acitretin group (12 of 102, 11.77%) compared with the cyclosporine (1 of 101, 0.99%) and combination groups (1 of 102, 0.98%; p < 0.001). Dyslipidemia occurred in 18 patients in the acitretin group (17.65%), compared with 7 in the cyclosporine group (6.93%) and 7 in the combination group (6.86%; p = 0.013). Hepatic abnormalities were observed in 15 patients (14.71%) in the combination group, 12 patients (11.77%) in the acitretin group, and 9 patients (8.91%) in the cyclosporine group (p = 0.427). These events were transient, rarely required additional medication, and resolved after drug discontinuation.

This study has several limitations. First, the relatively small sample size and the exclusion of patients with erythrodermic psoriasis, pustular psoriasis, and psoriatic arthritis limit the generalizability of the findings. The absence of radiographic assessments and evaluation of nail involvement further restricts the scope of disease characterization. In addition, the dose and potency of topical glucocorticoid ointments were not evaluated, which may have influenced treatment outcomes.

Second, the open-label design, while necessary for practical reasons, introduced potential performance bias and detection bias in patient-reported outcomes. To mitigate this, the primary efficacy endpoints were assessed by blinded dermatologists, reducing the risk of detection bias.

Finally, baseline BSA differed across treatment groups, with patients in the combination therapy arm presenting with greater disease involvement at study entry (p = 0.018). Although randomization was otherwise balanced and the differences were relatively small, this imbalance may have introduced bias when comparing treatment effects. To address this, efficacy was primarily assessed using relative measures of improvement (e.g., PASI75 and PASI90 response rates), which are less affected by baseline severity than absolute scores. Nonetheless, the possibility that baseline BSA differences influenced the magnitude of between-group comparisons cannot be excluded, and our conclusions should therefore be interpreted with this limitation.