A Retrospective Real-World Study of the Effectiveness and Tolerability of Tildrakizumab in UK Adults with Moderate-to-Severe Chronic Plaque Psoriasis

This retrospective real-world data study was undertaken at four specialist dermatology departments within NHS hospitals, two in England and two in Scotland. All adults receiving tildrakizumab or having previously received tildrakizumab for treatment of moderate-to-severe plaque psoriasis were included in the study. Data were collected by specialist dermatology nurses or pharmacists. Prior to data collection, ethics approval was obtained and a data-sharing agreement was put in place at each site. The study was performed in accordance with the Helsinki Declaration of 1964 and its later amendments. Consent from patients for the use of their data in the study and its publication was not required as the study comprised purely a retrospective review of selected data from anonymized medical records.

Data collected included demographic data, use of non-biologic and/or biologic treatments prior to tildrakizumab, initiation dose of tildrakizumab and subsequent titrations, and tildrakizumab tolerability, and effectiveness. Because of the real-world retrospective nature of the study, and due to part or all of the timeline being during the recent COVID-19 pandemic, it was not possible to measure with accuracy PASI scores at defined timepoints, e.g., PASI75, PASI90, PASI100. Instead, subjects were defined as responders or non-responders based on whether they were maintained on treatment with tildrakizumab or treatment was withdrawn.

Continuous normally distributed variables were reported using means and standard deviation. Variables were compared between the groups of responders and non-responders using a t test if the normality was met. Otherwise, the Mann–Whitney test was applied. Equality of variances was tested using Levene’s test. Continuous not normally distributed variables were reported using median, first and third quartile (Q1 and Q3). Normality was assessed using the Shapiro–Wilk test. Categorical variables were compared using the chi-squared test (with the Yates correction when necessary). The binary response to tildrakizumab was assessed using a multivariate logistic regression model built for this outcome. The initial model contained the following variables as predictors: sex, age, body mass index (BMI), number of special sites, time from psoriasis diagnosis, number of previous non-biologic systemic treatments (continuous), number of previous biologic systemic treatments (continuous), psoriatic arthritis, hypertension, current or past smoker, dose increase to 200 mg, initial dose, initial DLQI score, and initial PASI score. The final model was obtained using a stepwise backward procedure with a p value for removal of > 0.1. Only the final model was reported. For the predictors from the final logistic regression, an effect plot was presented. The significance level was set to 0.05 but the results with p value equal to 0.05 were also interpreted. All statistical calculations were conducted using R v. 4.02 and R Studio v. 2021.09.2 software.

Characteristics of the study population are presented in Table 1. Patients ranged in age from 17 to 82 years, reflecting the breadth of the adult population. Patients tended to be overweight and exhibited relatively high levels of comorbidities; the median number of comorbidities per patient was two, with a range of 0–6. The median time since diagnosis of plaque psoriasis was considerable, and the history of previous non-biologic and biologic therapy varied accordingly.

For systemic non-biologics, patients had previously received a median of two (range 0–5) treatments during their treatment for plaque psoriasis. For biologics, the number and type of treatments represented in the study population ranged from zero to six biologics from each of the three different biologics treatment classes. The most common biologic that patients had been exposed to previously was adalimumab, with 54/100 (54%) patients having received it. The next most common biologics were etanercept (29/100, 29%), ustekinumab (20/100, 20%), secukinumab (19/100, 19%), brodalumab (8/100, 8%), ixekizumab (4/100, 4%), certolizumab (2/100, 2%), and infliximab (2/100, 2%). Tildrakizumab was used as the first-line biologic in 19/100 (19%) patients.

Males and females receiving tildrakizumab differed in certain characteristics. The median body weight for females was 89 kg versus males 100 kg (p = 0.03) although this did not translate into a difference in BMI, which was 33 for females and 32 for males (p = 0.33). Depression was more frequent in females than males (28 vs. 9%, p = 0.01). Although males had higher baseline PASI scores (median initial PASI of 12 for females and 15 for males, p = 0.05) the baseline DLQI scores were similar (16 for females and 20 for males, p = 0.17). For females and males, median final PASI scores were 1.7 and 0.75 (p = 0.41) and median final DLQI scores were 3 and 0 (p = 0.13), respectively. There was a significant difference in the absolute percentage change from baseline for the DLQI score; the median change for males was equal to 100% and for females 80% (p = 0.02).

A partial response to the previous biologic was the most common reason for patients being initiated on tildrakizumab (61/80, 76%). The other reasons were a full lack of response to the previous biologic (11/80, 14%) or an adverse reaction (5/80, 6%).

At the end of the study collection period, most patients (91/122, 75%) were continuing treatment with tildrakizumab (Fig. 1). Of the remaining patients, 24/122 (20%) were regarded as not having responded fully or partially to tildrakizumab after a mean treatment duration of 33 weeks (SD 13 weeks), with duration of treatment ranging from 9 to 55 weeks. Seven of these patients discontinued treatment earlier than 28 weeks. Treatment was discontinued in an additional 7/122 (6%) patients for reasons unrelated to response including non-attendance at follow-up (n = 2), lymphoma diagnosis (n = 1), lung cancer diagnosis (n = 1), local sensitivity reaction (n = 1), patient choice (n = 1), and an unknown reason (n = 1).

Patients receiving tildrakizumab for more than one year tended to remain on treatment (Fig. 2). The probability of a patient remaining on tildrakizumab after one year was 73%, ranging from 57 to 100% among the four sites.

Median baseline PASI and DLQI scores were 12 (range 0–40) and 20 (range 0–30). Median final PASI and DLQI scores were significantly lower in responders (0.35 and zero) than non-responders (6 and 10) (p < 0.001). The absolute percentage changes in PASI and DLQI from baseline to final assessment were significantly greater in responders (96 and 100%) than non-responders (62 and 58%) (p < 0.001). The duration for which patients received tildrakizumab varied because of the real-world nature of the study. Accordingly, to give a sense of the change over time in PASI scores, Fig. 3 provides the trend, exponential, and weighted by the number of results, in PASI over time for patients receiving tildrakizumab.

The only significant predictor in the multivariate logistic regression model was the number of previous biological treatments: with every single additional treatment, the probability of the response to tildrakizumab decreased by 40% (Table 2). The effect plot for this continuous predictor is presented in Fig. 4.

The 100-mg dose of tildrakizumab was used most frequently in patients under 90 kg, with the 200-mg dose reserved more for patients over 90 kg, particularly those over 120 kg (Fig. 5). Among the ten patients with body weight over 120 kg, eight patients were initiated on a dose of 200 mg and one was escalated from 100 to 200 mg. The dose of the remaining patient was not recorded. Interestingly, eight of these nine patients achieved a good response, with three of them reaching a PASI ≤ 1 and DLQI = 0. In 13 patients, tildrakizumab was escalated from the 100-mg to 200-mg dose, with eight (62%) of them responding to treatment. PASI scores decreased in 83% patients and DLQI decreased in half of them. Confidently attributing the reduction in PASI and DLQI to the dose change was not possible in all cases because of insufficient data being recorded.

Tildrakizumab was generally well tolerated, with suspected adverse drug reactions recorded in 6/113 (5%) of patients. There was only one tolerability-related treatment discontinuation, in a patient with a local sensitivity reaction. Four of 59 (7%) patients receiving 100-mg tildrakizumab each experienced one suspected adverse reaction comprising (1) headache and nausea/fatigue the day after injection, (2) fatigue and widespread joint pain, (3) left-sided chest discomfort (although the chest X-ray and electrocardiogram were normal and the event was thought unlikely to be associated with tildrakizumab), and (4) diarrhea and vomiting after the first dose. Two of 51 (4%) patients receiving 200-mg tildrakizumab each experienced one suspected adverse reaction comprising (1) a local sensitivity reaction, and (2) facial swelling and breathlessness after the first dose. Thus, a good safety profile was observed for both 100-mg and 200-mg doses of tildrakizumab in a real-world setting.