Long-term efficacy of ustekinumab in patients with moderate-to-severe psoriasis treated for up to 5 years in the PHOENIX 1 study


  • Conflict of interest
    A.B. Kimball is an investigator and consultant for Amgen, Abbott and Janssen; she also has fellowship funding from Janssen; K.A. Papp has received grants and/or honoraria as a consultant, speaker and/or advisory board member from Abbott, Amgen, Anacor, Astellas, Celgene, Celtic Pharma, Dow Pharma, Eli Lilly, Galderma, J&J, Merck, Novartis, Pfizer and/or UCB; R. Bissonnette has been a consultant, speaker and/or investigator and has received honoraria and/or grants from Janssen, Abbott, Amgen, Astellas Pharma, Isotecknika Pharma, Celgene, Pfizer, Novartis and Tribute Pharmaceuticals; H. Sofen has been a paid consultant and investigator for Janssen; and K.B. Gordon has served as a consultant/investigator for Abbott, Amgen, Centocor and Merck and as a consultant for Eli Lilly and Pfizer. Y. Wasfi, D. Chan, N. Yeilding, S. Li, and P. Szapary are employees of Janssen Research & Development, LLC, Spring House, PA and own stock in Johnson and Johnson.

  • Funding
    This study was funded by Janssen Research & Development, LLC, Spring House, PA, USA.


Background  Ongoing evaluation of biological agents in patients with moderate-to-severe psoriasis is needed to support their long-term use.

Objective  To evaluate long-term efficacy and safety of ustekinumab through 5 years in the PHOENIX 1 study.

Methods  Patients were randomized to placebo or ustekinumab (45 mg or 90 mg) at Weeks 0, 4 and every-12-weeks thereafter; placebo patients crossed-over to ustekinumab at Week 12. Clinical response through Week 244 was evaluated using the Psoriasis Area and Severity Index (PASI) in the Overall Population (i.e. patients receiving ≥1 dose of ustekinumab), Initial Responders (i.e. PASI 75 responders [Weeks 28/40] re-randomized at Week 40 to continue every-12-week maintenance) and Partial Responders (i.e. <PASI 75 responders adjusted to every-8-week maintenance at Weeks 28 or 40). Safety endpoints were evaluated through Week 264 for the Overall Population.

Results  Overall, 68.7% (517/753) of ustekinumab-treated patients completed treatment through Week 244. Initial clinical responses were generally maintained through Week 244 (PASI 75: 63.4% and 72.0%; PASI 90: 39.7% and 49.0%; PASI 100: 21.6% and 26.4%) for patients receiving 45 mg and 90 mg, respectively. Similarly, PASI 75 responses were generally maintained among Initial Responders [79.1% (45 mg) and 80.8% (90 mg)] and Partial Responders [57.6% (45 mg) and 55.1% (90 mg)]. With 3104 patient-years of follow-up, rates of overall adverse events (AEs), serious AEs, serious infections, malignancies and major adverse cardiovascular events were generally consistent over time and comparable between doses.

Conclusions  Through 5 years of continuous treatment, ustekinumab demonstrated stable clinical response and a safety profile consistent with previous reports.


The management of patients with moderate-to-severe psoriasis requires long-term maintenance treatment. Although studies evaluating the efficacy of biological agents beyond 3 years of follow-up are limited, currently available evidence suggests that biological agents may be appropriate long-term maintenance treatments for psoriasis.1,2

The Phase III clinical development programme for ustekinumab (Stelara; Janssen Biotech, Inc., Horsham, PA) in psoriasis included two studies (PHOENIX 1 and PHOENIX 2), which were prospectively designed with long-term extensions (LTEs) of up to 5 years of continuous exposure and follow-up. To our knowledge, these studies represent the longest evaluation of efficacy and safety among psoriasis studies of biological agents conducted to date. In this report, we evaluate the durability of clinical response and summarize observed safety outcomes in patients treated with ustekinumab through up to 5 years of follow-up in the PHOENIX 1 study. Results from these analyses will help dermatologists determine the likelihood of treatment success in psoriasis patients who may be candidates for ustekinumab treatment.


Patients and study design

PHOENIX 1 was a Phase III, randomized, double-blind, placebo-controlled, multicentre trial designed to evaluate the efficacy and safety of ustekinumab 45 and 90 mg during four study periods over the course of 5 years: a 12-week placebo-controlled period, a subsequent 28-week placebo crossover and active treatment period, a randomized withdrawal and retreatment period starting at Week 40, and a LTE period beginning after Week 52, for an additional 4 years (Fig. 1). This study was initiated on December 15, 2005, and the last patient completed the final study visit on May 11, 2011. Eligible patients were at least 18 years old, had a diagnosis of plaque psoriasis for at least 6 months, were candidates for phototherapy or systemic therapy, had a baseline Psoriasis Area and Severity Index (PASI) score of 12 or higher and had at least 10% body surface area involvement with psoriasis. No concomitant treatments for psoriasis were allowed through Week 76; however, low and mid-potency topical corticosteroids (Class III to VII), calcipotriene and tazarotene were permitted after Week 76. Details of the study design and patient entry criteria have been reported previously.3 This report will focus on the LTE period through the end of the study (Year 5).

Figure 1.

 Study schema of PHOENIX 1 from baseline through up to 5 years of treatment.

Efficacy and safety evaluations

The PASI,4 Physician’s Global Assessment (PGA) and Dermatology Life Quality Index (DLQI)5 were used to evaluate disease severity, treatment response and patient-reported outcomes, respectively, through Week 244 (the last Year-5 dose). Safety parameters, including adverse events (AEs) and laboratory evaluations, were monitored through Week 264 (the final Year-5 visit). Serum samples collected through Week 264 were tested for antibodies to ustekinumab using a bridging immunoassay.6

Statistical methods

Efficacy data were analysed through Week 244 for three patient groups: (i) Overall Population, including all patients treated with at least one dose of ustekinumab; (ii) Initial Responders, comprised of PASI 75 responders, (i.e. achieving at least 75% improvement in PASI) at Weeks 28 and 40 who were randomized to ustekinumab at baseline and re-randomized at Week 40 to continue maintenance dosing every-12-weeks; and (iii) Partial Responders, defined as patients achieving PASI 50 (at least 50% improvement) to <PASI 75 at Week 28 or less than PASI 75 at Week 40, who were permitted to adjust their dosing interval to every-8-weeks (Fig. 1). Patients who were withdrawn from ustekinumab at Week 40 and subsequently re-treated after experiencing a loss of therapeutic effect were previously evaluated,2,3 and these patients are included in the Overall Population.

To evaluate long-term efficacy in the Overall Population, dichotomous (PASI and PGA response) and continuous (% PASI improvement) efficacy measures were reported from Weeks 0 through 244 in all patients treated with ustekinumab 45 mg or 90 mg, including those who crossed-over from placebo, Initial Responders, Partial Responders and Week 28-non-responders (i.e. patients who did not achieve PASI 50) who discontinued treatment per protocol. Week-28 non-responders had their Week-28 data carried forward and were considered non-responders through Week 244.

To describe long-term efficacy in the Initial Responders population, the proportions of patients achieving PASI 75, 90 (at least 90% improvement) and 100 (complete clearance) in the ustekinumab 45 mg and 90 mg groups were reported from Week 40 through 244. To further assess the maintenance of response over time, a reverse cumulative distribution function (RCDF) of the per cent improvement from baseline in PASI was generated for each of the following time points: Week 40 (time of re-randomization to continuous maintenance), Week 52 (Year 1), Week 100 (Year 2), Week 148 (Year 3), Week 196 (Year 4) and Week 244 (Year 5). In addition, improvement in health-related quality of life (i.e. mean change in DLQI from baseline at Weeks 76 and 244) was assessed.

Treatment failure rules were applied to each efficacy analysis. Patients who discontinued treatment due to unsatisfactory therapeutic effect or an AE of worsening psoriasis and those who used a protocol-prohibited medication/therapy were considered treatment failures and were included in all subsequent analyses as non-responders. For treatment-failures, baseline values were assigned for continuous endpoints (i.e. considered to have not improved) or non-response status was assigned for dichotomous endpoints regardless of observed data. Patients with missing Week-12 data were considered non-responders at Week 12. Otherwise, missing data were not imputed after the application of treatment failure rules (e.g. discontinuation for reasons unrelated to efficacy). To assess the impact of missing data, a sensitivity analysis was performed in the Initial Responder population using a last observation carried forward (LOCF) procedure to impute missing efficacy data after the application of treatment failure rules.

Safety outcomes were described and analyses were conducted based on actual treatment received in all patients who received at least one dose of ustekinumab. All safety events occurring from the first ustekinumab exposure through up to 5 years of follow-up (Week 264) were included. The impact of potential dose-related toxicity was evaluated by reporting cumulative rates of key safety endpoints [i.e. overall AEs, AEs leading to discontinuation, serious AEs (SAEs) and AEs of interest [i.e. serious infections, malignancies and major adverse cardiovascular events (MACE; defined as cardiovascular death, myocardial infarction or stroke)] per 100 patient-years (PY) of follow-up for patients receiving ustekinumab 45 mg and 90 mg. All SAEs that could represent a MACE were selected, and retrospectively adjudicated in a blinded manner by an independent Clinical Events Committee at the Cleveland Clinic Coordinating Center for Clinical Research.7 The effect of duration of exposure to ustekinumab on key safety endpoints was assessed by year from Year 1 (Week 52) to Year 5 (Week 264), with patients receiving 45 mg and 90 mg analysed as a combined group.


A total of 766 patients were randomized to placebo or ustekinumab 45 mg or 90 mg.3 Of the 753 patients who received at least one dose of ustekinumab, 68.7% (n = 517) completed study agent through the last Year-5 dose at or before Week 244 (Table 1; Fig. 2). Generally similar proportions of patients discontinued study agent in the 45 mg and 90 mg groups. Approximately, 8% of patients each discontinued due to reasons related to safety (n = 63) and efficacy (n = 62), and approximately 4% of patients (n = 33) were lost to follow-up. A total of 10.4% of patients reported other reasons for discontinuing ustekinumab; the majority had either withdrawn consent or were considered non-compliant with study protocol (e.g. use of prohibited medications, inability to complete study visits).

Table 1. Proportion of patients who discontinued ustekinumab treatment and reasons for discontinuation through Year 5
 Ustekinumab 45 mg*Ustekinumab 90 mg*Combined
  1. Values are presented as number of patients (%).

  2. *Includes patients originally randomized to placebo after they crossed-over to receive ustekinumab.

  3. Includes patients who discontinued due to either lack of efficacy or an adverse event of psoriasis, as reported by the investigator.

  4. ‡Includes patients who did not achieve PASI 50 at Week 28 and discontinued from the study per protocol.

  5. ††One additional death occurred in a patient (ustekinumab 90 mg) after study completion.

Patients treated, n378375753
Patients who completed study agent259 (68.5)258 (68.8)517 (68.7)
Patients who discontinued study agent119 (31.5)117 (31.2)236 (31.3)
Reasons for discontinuation
 Efficacy-related40 (10.6)22 (5.9)62 (8.2)
 Unsatisfactory therapeutic effect†18 (4.8)16 (4.3)34 (4.5)
 Week-28 non-responder‡22 (5.8)6 (1.6)28 (3.7)
 Safety-related29 (7.7)34 (9.1)63 (8.4)
 Adverse event28 (7.4)31 (8.3)59 (7.8)
 Death††1 (0.3)3 (0.8)4 (0.5)
 Lost to follow-up14 (3.7)19 (5.1)33 (4.4)
 Other36 (9.5)42 (11.2)78 (10.4)
 Withdrawal of consent19 (5.0)18 (4.8)37 (4.9)
 Non-compliance7 (1.9)9 (2.4)16 (2.1)
 Patient re-location/site closure3 (0.8)10 (2.7)13 (1.7)
 Planning pregnancy4 (1.1)4 (1.1)8 (1.1)
 Miscellaneous3 (0.8)1 (0.3)4 (0.5)
Figure 2.

 Patient disposition in PHOENIX 1 from baseline through up to 5 years of treatment.

Patient demographic and clinical characteristics were similar across treatment groups at baseline (Table 2). Patients who completed 5 years of ustekinumab treatment had similar characteristics compared with all patients who were treated with at least one dose of ustekinumab, indicating that those who remained in the study were generally representative of the original randomized population.

Table 2. Baseline demographic and clinical characteristics of all patients treated with at least one dose of ustekinumab compared with patients who completed ustekinumab treatment through Year 5
 All patients treated with ustekinumabAll patients who completed ustekinumab treatment
Ustekinumab 45 mgUstekinumab 90 mgCombinedUstekinumab 45 mgUstekinumab 90 mgCombined
  1. Values are presented as number of patients (%) or mean ± standard deviation.

  2. *Conventional systemic agents included psoralen plus ultraviolet A, methotrexate, acitretin and cyclosporine.

  3. †Biological agents included adalimumab, alefacept, efalizumab, etanercept and infliximab.

  4. ‡Latent TB infection was defined as a positive Mantoux tuberculin skin test result in a patient with no signs or symptoms suggestive of active TB by medical history, physical examination and chest radiograph.

  5. BSA, body surface area; DLQI, Dermatology Life Quality Index; PASI, Psoriasis Area and Severity Index; PGA, Physician’s Global Assessment; PsA, psoriatic arthritis; TB, tuberculosis.

Patients treated, n378375753259258517
Age (years)45.0 ± 12.045.7 ± 11.545.3 ± 11.744.9 ± 11.546.1 ± 10.845.5 ± 11.1
Male268 (70.9)255 (68.0)523 (69.5)187 (72.2)180 (69.8)367 (71.0)
Weight (kg)93.7 ± 23.494.3 ± 24.094.0 ± 23.791.5 ± 22.194.5 ± 22.793.0 ± 22.4
Duration of psoriasis (years)20.3 ± 11.819.6 ± 11.220.0 ± 11.521.2 ± 11.521.1 ± 11.221.1 ± 11.4
Involved BSA, %28.0 ± 17.825.1 ± 15.226.5 ± 16.626.2 ± 16.924.0 ± 14.525.1 ± 15.8
PGA, marked or severe167 (44.2)162 (43.2)329 (43.7)112 (43.2)121 (46.9)233 (45.1)
PASI score (0–72)20.8 ± 9.019.6 ± 7.520.2 ± 8.319.8 ± 8.119.1 ± 7.019.5 ± 7.5
DLQI score (0–30)11.3 ± 7.211.4 ± 7.011.4 ± 7.111.1 ± 7.011.1 ± 6.811.1 ± 6.9
Relevant medical history
 PsA119 (31.5)134 (35.7)253 (33.6)74 (28.6)91 (35.3)165 (31.9)
 Diabetes45 (11.9)44 (11.7)89 (11.8)29 (11.2)25 (9.7)54 (10.4)
 Hypertension106 (28.0)116 (30.9)222 (29.5)69 (26.6)79 (30.6)148 (28.6)
 Hyperlipidaemia78 (20.6)88 (23.5)166 (22.0)49 (18.9)59 (22.9)108 (20.9)
Previous treatment
 Topical agents361 (95.5)354 (94.4)715 (95.0)246 (95.0)245 (95.0)491 (95.0)
 Phototherapy248 (65.6)236 (62.9)484 (64.3)175 (67.6)165 (64.0)340 (65.8)
 Conventional systemics*210 (55.6)204 (54.4)414 (55.0)142 (54.8)141 (54.7)283 (54.7)
 Biological agents†199 (52.6)186 (49.6)385 (51.1)132 (51.0)132 (51.2)264 (51.1)
Patients with latent TB‡15 (4.0)10 (2.7)25 (3.3)15 (5.8)3 (1.2)18 (3.5)


Overall population

The proportions of patients in the Overall Population who achieved PASI 75 at Week 76 (45 mg: 61.2%; 90 mg: 72.4%) were similar to those observed at Week 244 (45 mg: 63.4%; 90 mg: 72.0%), suggesting that most patients achieved and maintained clinical response over time (Fig. 3a). The dose response initially observed earlier in the study remained evident throughout the LTE, with an approximate 10 percentage point difference in PASI 75 response rates between patients who received 45 mg and those who received 90 mg. The median per cent improvement in PASI (Fig. 3b) and the proportion of patients achieving a PGA score of cleared (0) or minimal (1) (Fig. 3c) also remained stable with maintenance dosing from Week 76 to Week 244. Notably, a substantial proportion of patients continued to achieve very high levels of skin clearance at Week 244, with responses of PASI 90 (45 mg: 39.7%; 90 mg: 49.0%) and PASI 100 (45 mg: 21.6%; 90 mg: 26.4%). Among the sub-group of Partial Responders (26% of Overall Population) who were adjusted to every-8-week dosing at Weeks 28 or 40, approximately half of the patients achieved PASI 75 by Week 60 and maintained response through Week 244 (45 mg: 57.6%; 90 mg: 55.1%); corresponding values for PASI 90 were 27.2% (45 mg) and 27.5% (90 mg).

Figure 3.

 Clinical responses in the Overall Population through up to Week 244 of treatment: (a) proportion of patients achieving at least 75% improvement from baseline in the Psoriasis Area and Severity Index (PASI 75), (b) median per cent improvement in PASI response, and (c) proportion of patients achieving a Physician’s Global Assessment (PGA) score of cleared (0) or minimal (1). *Patients initially randomized to placebo who crossed-over to ustekinumab were included, beginning at Week 24 (i.e. 12 weeks after switching to ustekinumab based on an approximate time-to-response of 12 weeks). Efficacy was not evaluated between Week 40 and Week 76 (Randomized Withdrawal period) because patients withdrawn from treatment at Week 40 per protocol comprised a large portion of the Overall Population (320/753; 42.5%); analyses were resumed at Week 76, when psoriasis recurred in about half of the withdrawn patients and ustekinumab was re-initiated for at least 12 weeks. ‡Patients who re-initiated treatment for 12 weeks or more were re-included in the analyses through Week 244.

Initial responders re-randomized at Week 40 to continue every-12-week maintenance treatment

The high levels of clinical response achieved initially were generally maintained from Week 40 through Week 244 among Initial Responders (Fig. 4a,b). At Week 244, PASI 75 (79.1% and 80.8%), PASI 90 (47.8% and 58.9%) and PASI 100 (31.3% and 38.4%) responses were achieved in the 45 mg and 90 mg groups, respectively. Results were similar when data were analysed using the LOCF approach [PASI 75: 79.2% (45 mg); 78.8% (90 mg) at Week 244]. As with clinical response, improvements in health-related quality of life were consistently maintained. The mean changes in DLQI from baseline at Week 76 (45 mg: −9.1 ± 7.0; 90 mg: −9.6 ± 6.3) and Week 244 (45 mg: −7.2 ± 6.9; 90 mg: −8.1 ± 5.8) were comparable.

Figure 4.

 Proportion of Initial Responders who were re-randomized at Week 40 to receive continuous maintenance therapy every-12-weeks and who achieved at least 75%, 90% or 100% improvement from baseline in the Psoriasis Area and Severity Index (PASI 75, PASI 90, or PASI 100) from Week 40 through Week 244 by ustekinumab dose: (a) 45 mg and (b) 90 mg.

To further evaluate the stability of clinical response in Initial Responders through Week 244, an RCDF of PASI improvements was generated at Weeks 40, 52, 100, 148, 196 and 244 (Fig. 5). As per the study design, 100% of the patients who were re-randomized at Week 40 to continue maintenance dosing every-12-weeks were PASI 75 responders at Week 40. For approximately 80% of patients, the distribution of responses remained generally stable at each time point examined, with at least 50% of patients sustaining PASI 90 response. Among the 20% of patients with less than PASI 75 response, variable proportions (range, 8.6–17.1%) at each time point were Partial Responders (PASI 50 to <PASI 75) and the proportion of patients with less than PASI 50 response was low (range, 2.5–9.3%). At the last efficacy assessment (Week 244), 13 of 140 (9.3%) patients were characterized as having responses below PASI 50. Of these 13 patients, two had less than a PASI 50 response and four were discontinued due to lack of efficacy per investigator assessment. Loss of response in these six patients did not result from decreasing ustekinumab concentrations or development of anti-ustekinumab antibodies (data not shown). Although the other seven patients had achieved PASI 75, they had initiated medications not permitted per protocol [i.e. topical corticosteroids for psoriasis (n = 2) or systemic corticosteroids for other conditions (n = 5)] and were, therefore, considered treatment failures (i.e. assigned a PASI per cent improvement of 0).

Figure 5.

 Reverse cumulative distribution function (RCDF) of percentage improvements in the Psoriasis Area and Severity Index (PASI) from baseline through Week 244 in Initial Responders who were re-randomized at Week 40 to receive continuous maintenance therapy every-12-weeks. *Each point of observation on the RCDF consists of percentage improvements in PASI, which represents an individual patient or group of patients reporting the same response.PASI 50, Psoriasis Area and Severity Index 50; PASI 75 Psoriasis Area and Severity Index 75; PASI 50 Psoriasis Area and Severity Index 50.


Safety outcomes were summarized for 753 patients (3104 PY) who received at least one dose of ustekinumab. To evaluate potential dose-related toxicity, safety endpoints were evaluated by ustekinumab dose received through 5 years (Table 3). The cumulative rates of overall AEs, AEs leading to discontinuation, SAEs, infections, malignancies and MACE were generally comparable between patients receiving ustekinumab 45 and 90 mg (Table 3), suggesting no dose effect. There were 32 serious infections [13 (45 mg) and 19 (90 mg)] reported in 30 patients. Non-melanoma skin cancers (NMSCs) were reported in 14 patients [10 (45 mg) and 4 (90 mg)], with a ratio of basal cell carcinoma to squamous cell carcinoma of 13 : 1. Other malignancies were reported in 15 patients [9 (45 mg) and 6 (90 mg)]. There were 10 cases of MACE [8 (45 mg) and 2 (90 mg)] reported in 10 patients. All patients with MACE had at least three established cardiovascular risk factors. Five deaths [1 (45 mg), and 4 (90 mg)] were reported during the study: sleep apnoea presumed to be cardiovascular in origin (45 mg), complications from bilateral pneumonia, perforated bowel resulting from trauma, suicide and cervical vertebral fracture secondary to a fall. One additional death was reported after the study ended in a patient (ustekinumab 90 mg) who had previously discontinued study treatment due to metastatic pancreatic cancer.

Table 3. Cumulative rates of key safety events per 100 patient-years of follow-up through Year 5*
 Ustekinumab 45 mgUstekinumab 90 mgCombined
  1. *All events for each individual patient were counted separately, except for AEs leading to discontinuation, NMSC and other malignancies for which the number of patients with at least one event rather than the number of events was counted.

  2. †Includes infections reported by the investigator on the case report form.

  3. ‡Includes MACE, as adjudicated by an independent panel of blinded cardiologists at C5.

  4. ¥Includes five patients with prostate cancer, three with melanoma (two in situ and one invasive), and one patient each with breast cancer, colon cancer, lymphoma, metastatic pancreatic carcinoma, head and neck cancer, thyroid cancer, and transitional cell carcinoma.

  5. AE, adverse event C5, Cleveland Clinic Coordinating Center for Clinical Research; MACE, major adverse cardiovascular event (cardiovascular death, myocardial infarction or stroke); NMSC, non-melanoma skin cancer.

Patients treated, n378375753
Patient-years of follow-up (years)1539.91564.23104.2
Average duration of follow-up (weeks)211.8216.9214.4
AEs leading to study agent discontinuation2.032.262.13
Serious AEs5.265.435.35
 Infections requiring treatment28.4430.3729.41
 Serious infections0.841.211.03
Malignant neoplasms1.240.640.93
 Other malignancies (excluding NMSC)¥0.590.380.48

To evaluate the effect of duration of exposure to ustekinumab, safety endpoints were evaluated by year of follow-up (Fig. 6). In this analysis, in which patients receiving ustekinumab 45 and 90 mg were analysed as a combined group, the rates of all safety events were stable, with a consistent trend over time from Years 1 through 5 (Fig. 6). An exception was the gradual decline over time in the rates of AEs leading to discontinuation. Overall, there was no indication of cumulative toxicity with increased duration of exposure to ustekinumab.

Figure 6.

 Summary of key safety findings [events per 100 patient-years (PYs) of follow-up] from Year 1 through Year 5 in all ustekinumab-treated patients. AEs→D/C, adverse events leading to discontinuation; MACE, major adverse cardiovascular event (cardiovascular death, myocardial infarction or stroke); NMSC, non-melanoma skin cancer; SAE, serious adverse event.*Includes MACE, as adjudicated by an independent panel of blinded cardiologists.

No cases of active tuberculosis or other infections of interest (e.g. atypical mycobacterial, systemic fungal or salmonella) were reported through Year 5. One potential opportunistic infection (disseminated cutaneous herpes zoster) was reported early in the study, as described previously.6 No cases of anaphylaxis or serum sickness-like reactions associated with ustekinumab were observed.

Immunogenicity rates remained low through Year 5, with antibodies to ustekinumab detected in 5.2% of patients (39/746 patients with evaluable samples). Titres were primarily low (67% ≤ 1:40), and most antibodies were neutralizing (64%). Only one patient developed antibodies to ustekinumab between Week 76 and Year 3, as previously reported,2 and no additional patients developed antibodies to ustekinumab between Years 3 and 5. Development of antibodies was not associated with injection site reactions.


The management of patients with moderate-to-severe psoriasis frequently requires long-term, continuous treatment with systemic therapies. Our findings from the PHOENIX 1 study substantiate that ustekinumab markedly improves moderate-to-severe psoriasis and is generally well tolerated for up to 5 years. To our knowledge, this study reports the longest duration of continuous treatment and cumulative exposure (3104 PY) in the prospective evaluation of a biological agent in the setting of a single clinical trial. Notably, nearly 70% of patients completed ustekinumab treatment through Week 244 in this 5-year study, which represents a retention rate that compares favourably with other shorter-term studies.8–10

In the Overall Population analysis, we evaluated all patients who were eligible to receive ustekinumab treatment in our study, including Initial Responders, Partial Responders and those who did not respond to therapy. Clinical responses achieved during initial periods of follow-up3 were generally maintained through Week 244. This stability of response was evident using various measures of efficacy (PASI, PGA, % PASI improvement), consistently confirming the effectiveness of stable maintenance dosing. Furthermore, among Partial Responders who did not achieve PASI 75 response on their original treatment regimen, over half did so after dose-interval adjustment from every-12-weeks to every-8-weeks. Two recent meta-analyses evaluated the comparative effectiveness of systemic and biological agents through 12 weeks11 and 60 weeks12 of treatment. Results from these meta-analyses are comparable to those observed in patients treated with ustekinumab for similar durations of treatment. However, a direct comparison of our experience from PHOENIX 1 with studies of other biological agents is limited, as this is the first study to evaluate the long-term efficacy of a biological agent through up to 5 years of treatment. Overall, our results suggest that long-term disease control is effectively maintained through the inhibition of interleukin-12 and interleukin-23 signalling in patients with psoriasis.

Patients who initially respond well to treatment are likely to continue long-term treatment, and it is of clinical importance to anticipate the likelihood of continued treatment success over time. In this study, assessment of clinical response by various efficacy measures (i.e. PASI response, % PASI improvement) over time indicated that approximately 80% of Initial Responders receiving either the 45 mg or 90 mg dose every-12-weeks continued to sustain at least a PASI 75 response through Week 244. Among the remaining 20% of patients, some may have experienced fluctuations in their clinical response over time, reflecting the biology of psoriatic disease. Less than 10% of Initial Responders were characterized as non-responders (<PASI 50) at Week 244, the majority of whom were considered non-responders based on treatment failure rules (e.g. use of prohibited medications for indications other than psoriasis) but had achieved a PASI 75 response. Although the causes of actual loss of response are unknown, variability in serum ustekinumab levels and immunogenicity status did not seem to play a role.

Safety observations presented here through up to 5 years of follow-up were generally consistent with previous reports of the same population at earlier follow-up.2,3 There was no evidence of an increase in the number of AEs (including events of interest, such as serious infections, malignancies and MACE) with increased duration of exposure. As previously reported,2 a slightly higher rate of serious infections was reported in the 90 mg group vs. the 45 mg group, though the disparity between the groups has narrowed with additional follow-up. Regardless, the number of infection events reported in this study remained low overall, which precludes a definitive conclusion regarding a potential dose response. Our findings demonstrated stable rates of NMSCs and other malignancies over time, indicating no apparent increase in malignancy risk after 5 years of follow-up in this study population. The association of psoriasis with an increased cardiovascular risk has been well documented.13 Data reported here for the PHOENIX 1 study population did not demonstrate evidence of an increased risk of MACE over time; however, the impact of interleukin-12/23 inhibition on cardiovascular safety continues to be investigated.14–16 Although the evaluation of safety outcomes from this single study is informative, further evaluation of the long-term safety of ustekinumab treatment in psoriasis is needed. An aggregate analysis based on a much larger population pooled from four Phase II and III studies in psoriasis has been conducted to provide a more thorough safety assessment, including the evaluation of potential AEs that may occur only rarely and in certain subpopulations; results of this analysis will be reported elsewhere. PSOLAR, an ongoing 12 000-patient, prospective, observational registry,17 will report safety outcomes in patients treated with ustekinumab and other systemic and biological agents in the actual use setting. Although observational registries have known limitations, potential safety signals can be detected and investigated further in properly designed randomized controlled trials.

Several limitations should be acknowledged in the interpretation of our results. The relevance of the clinical response observed in the PHOENIX 1 population to the general psoriasis population treated in clinical practice may be limited by some of the study design features (e.g. Partial Responders were adjusted to every-8-week dosing, which may not be an option in all regions, and patients were not dosed according to body weight [i.e. 45 mg for ≤100 kg and 90 mg for >100 kg], as is common in actual use). Patients enrolled in clinical trials may not be representative of the general psoriasis population, as they are selected based on strict inclusion and exclusion criteria and are required to receive more thorough monitoring and medical follow-up in a controlled environment. This study was not designed to evaluate hypotheses on safety endpoints, and the absence of a control group beyond the initial, 12-week, placebo-controlled period limits the interpretability of outcomes observed during the long-term period.

In summary, our results substantiate previous shorter-term reports of ustekinumab2,3 as an effective and well-tolerated long-term treatment in patients with moderate-to-severe psoriasis. Our current findings demonstrate that in patients with moderate-to-severe psoriasis, clinical response is maintained with up to 5 years of continuous maintenance therapy with ustekinumab and the safety profile through extended follow-up is consistent with that reported previously.


The authors thank Cynthia Arnold (Janssen Services, LLC, Spring House, PA, USA) for her writing support and editorial assistance and Yin You (Jannsen Research & Development, LLC, Spring House, PA, USA) for her statistical and programming support.