Long-term safety of ustekinumab in patients with moderate-to-severe psoriasis: final results from 5 years of follow-up

Authors


  • Funding sources
    Funded by Janssen Research & Development, LLC, Spring House, PA, U.S.A.

  • Conflicts of interest
    See Appendix.

Kim A. Papp.
E-mail: kapapp@probitymedical.com

Abstract

Summary Background  Long-term safety evaluations of biologics are needed to inform patient management decisions.

Objectives  To evaluate the safety of ustekinumab in patients with moderate-to-severe psoriasis treated for up to 5 years.

Methods  Safety data were pooled from four studies of ustekinumab for psoriasis. Rates of adverse events (AEs), serious AEs (SAEs) and AEs of interest [infections, nonmelanoma skin cancers (NMSCs), other malignancies and major adverse cardiovascular events (MACE)] per 100 patient-years (PY) of follow-up were analysed by ustekinumab dose (45 or 90 mg) and by year of follow-up (years 1–5) to evaluate the dose response and impact of cumulative exposure. Observed rates of overall mortality and other malignancies were compared with those expected in the general U.S. population.

Results  Analyses included 3117 patients (8998 PY) who received one or more doses of ustekinumab, with 1482 patients treated for ≥ 4 years (including 838 patients ≥ 5 years). At year 5, event rates (45 mg, 90 mg, respectively) for overall AEs (242·6, 225·3), SAEs (7·0, 7·2), serious infections (0·98, 1·19), NMSCs (0·64, 0·44), other malignancies (0·59, 0·61) and MACE (0·56, 0·36) were comparable between dose groups. Year-to-year variability was observed, but no increasing trend was evident. Rates of overall mortality and other malignancies were comparable with those expected in the general U.S. population.

Conclusions  No dose-related or cumulative toxicity was observed with increasing duration of ustekinumab exposure for up to 5 years. Rates of AEs reported in ustekinumab psoriasis trials are generally comparable with those reported for other biologics approved for the treatment of moderate-to-severe psoriasis.

Long-term treatment is required for the management of moderate-to-severe psoriasis. The safety of conventional systemic therapies and phototherapy has been established after several decades of clinical application;1,2 however, long-term use may be limited due to potential end-organ toxicities (methotrexate, ciclosporin)3 and increased risk of skin cancer [psoralen plus ultraviolet A (PUVA)].2 Biologics have been approved for the treatment of psoriasis for 10 years. While the positive benefit–risk assessments demonstrated in the initial short-term studies4–8 are reassuring, long-term safety evaluations are needed to identify adverse events (AEs) that occur rarely or develop over time with cumulative exposure, and to inform appropriately patient management decisions.9

Ustekinumab (Stelara®; Janssen Biotech, Inc., Horsham, PA, U.S.A.) has demonstrated short- and long-term efficacy in the treatment of moderate-to-severe psoriasis.7,8,10–12 Here we evaluated the safety of ustekinumab in patients treated for up to 5 years. Rates of safety events were assessed by ustekinumab dose (45 vs. 90 mg) and by year of follow-up (years 1–5) to evaluate the dose response and impact of cumulative exposure.

Patients and methods

Patients and study design

Patients with moderate-to-severe psoriasis from four randomized, blinded, phase II and III ustekinumab studies were included (Fig. 1). All studies were placebo controlled, except for the ACCEPT active-comparator (etanercept) trial. Patient eligibility criteria and design features have been presented previously.7,8,10,11

Figure 1.

 Number of patients treated with ustekinumab and duration of treatment and follow-up in the pooled ustekinumab psoriasis safety cohort in total and by clinical trial. *Placebo patients who crossed over at week 12 are included in the ≥ 4·5 years but not in the ≥ 5 years exposure category (i.e. ≥ 240 weeks between first and last dose of ustekinumab).

Safety evaluation

Safety endpoints included overall AEs, serious AEs (SAEs) including deaths, AEs leading to discontinuation, AEs of interest [i.e. infections, nonmelanoma skin cancers (NMSCs), other malignancies and major adverse cardiovascular events (MACE)], laboratory abnormalities and antibodies to ustekinumab.13

Infections included: (i) investigator-reported events; or (ii) events categorized within the Medical Dictionary for Regulatory Activities (MedDRA v14.1) Infections and Infestation system organ class. Prior exposure to systemic agents or phototherapy may increase the risk of NMSC; therefore, rates of NMSC by previous treatment received were examined. A blinded, independent Clinical Events Committee at the Cleveland Clinic Coordinating Center for Clinical Research (C5) reviewed all SAEs retrospectively. Events that could represent MACE [i.e. cardiovascular (CV) death, myocardial infarction (MI) and stroke] were adjudicated as MACE or non-MACE by C5 based on standard, prespecified definitions. Deaths of otherwise unknown causes were adjudicated as MACE.

Statistical methods

All patients who received at least one dose of ustekinumab and all events from the first ustekinumab exposure through to the end of the reporting period were included (Fig. 1). Event rates were adjusted for follow-up and expressed as events per 100 patient-years of follow-up (PY). Each event was counted separately, except AEs leading to discontinuation, deaths and malignancies, wherein the number of patients reporting at least one event was reported and each patient was counted only once. Where appropriate, 95% confidence intervals (CIs) were provided based on an exact method assuming the observed number of events followed a Poisson distribution.

Safety events were analysed by (i) ustekinumab dose and (ii) year of follow-up. For analyses by dose, patients originally randomized to placebo/etanercept who crossed over to ustekinumab were included in the appropriate ustekinumab dose group (45 or 90 mg) after crossover. In PHOENIX 2, safety events in patients who dose adjusted (from 45 mg to 90 mg) were summarized according to the dose received at the time of the event. For analyses by year of follow-up, events were summarized by time periods based on onset date relative to the first ustekinumab exposure: year 1 (≤ 48 weeks), year 2 (> 48 and ≤ 96 weeks), year 3 (> 96 and ≤ 144 weeks), year 4 (> 144 and ≤ 192 weeks) and year 5 (> 192 weeks).

In the absence of a long-term control group, previously reported data from the controlled period14 are presented after adjusting for duration of follow-up; selected safety endpoints (overall mortality and other malignancies) were compared with external control populations. For overall mortality, rates observed in ustekinumab-treated patients were compared with those expected in the general U.S. population,15 adjusted for age and sex. For malignancies other than NMSC, rates observed in ustekinumab-treated patients were compared with those observed in the general U.S. population,16 adjusted for age, sex and race. Standardized mortality ratios (SMRs) and standardized incidence ratios (SIRs) for malignancies other than NMSC were calculated by dividing the observed number of ustekinumab-treated patients with events by the expected number of patients with events.

Results

Patients

A total of 3117 patients received at least one dose of ustekinumab. Baseline demographic and patient characteristics were generally similar between groups (Table 1), and characteristics of patients completing the 5-year PHOENIX long-term extensions were comparable to those of the originally randomized populations (data not shown). A total of 8998 PY of follow-up was reported, with at least 4 years of ustekinumab exposure in 1482 patients (including 838 patients with ≥ 5 years) (Fig. 1). The apparent decrease observed between the number of patients with at least 4 and 5 years of exposure was due to the shorter duration of ustekinumab exposure in patients originally randomized to placebo (approximately one-third of all patients), who by study design crossed over to ustekinumab at week 12.

Table 1. Demographic and baseline disease characteristics of treated patients in the pooled ustekinumab psoriasis safety cohort
 PlaceboaUstekinumabb
45 mg90 mgCombined
  1. PASI, Psoriasis Area and Severity Index; PUVA, psoralen plus ultraviolet A; UVB, ultraviolet B. Data are presented as percentage (n/N) unless otherwise indicated. aIncludes only data from the Phase II, PHOENIX 1 and PHOENIX 2 studies because the ACCEPT study did not include a placebo control group; bincludes data from all four studies (Phase II, PHOENIX 1, PHOENIX 2 and ACCEPT); cdata were not collected for the Phase II study.

Patients treated, n732131917983117
Age (years), mean ± SD46·0 ± 12·245·4 ± 12·445·8 ± 12·245·6 ± 12·3
Men70·5 (516/732)67·6 (891/1319)69·2 (1245/1798)68·5 (2136/3117)
White92·5 (677/732)92·9 (1225/1319)91·7 (1649/1798)92·2 (2874/3117)
Body mass index (BMI)
 Normal (BMI < 25)17·9 (131/732)19·1 (252/1318)19·2 (344/1796)19·1 (596/3114)
 Overweight (BMI ≥ 25 and < 30)34·7 (254/732)33·0 (435/1318)33·0 (593/1796)33·0 (1028/3114)
 Obese (BMI ≥ 30)47·4 (347/732)47·9 (631/1318)47·8 (859/1796)47·8 (1490/3114)
Body surface area involvement (%), mean ± SD26·7 ± 17·526·8 ± 17·025·7 ± 16·426·2 ± 16·7
PASI score, mean ± SD19·8 ± 8·019·9 ± 8·119·5 ± 7·519·7 ± 7·7
Relevant medical history
 Psoriatic arthritis28·3 (207/732)26·9 (355/1319)28·0 (503/1798)27·5 (858/3117)
 Depression13·3 (97/732)15·8 (209/1319)14·7 (265/1798)15·2 (474/3117)
 Diabetes mellitus12·6 (92/732)10·4 (137/1319)10·8 (194/1798)10·6 (331/3117)
 Hypertension29·2 (214/732)26·8 (354/1319)27·8 (500/1797)27·4 (854/3116)
 Hyperlipidaemia21·6 (158/732)19·6 (258/1319)20·8 (373/1797)20·3 (631/3116)
 Cigarette smoking, past or currentc60·5 (402/665)62·1 (741/1193)64·1 (1040/1623)63·2 (1781/2816)
Prior medications and exposure
 PUVA27·2 (199/732)27·4 (361/1319)27·5 (494/1798)27·4 (855/3117)
 UVB54·0 (395/732)57·8 (763/1319)54·8 (985/1798)56·1 (1748/3117)
 Ciclosporin14·8 (108/732)14·4 (190/1319)13·2 (238/1798)13·7 (428/3117)
 Methotrexate35·8 (262/732)35·6 (469/1319)35·1 (631/1798)35·3 (1100/3117)
 Etanercept/infliximab/adalimumabc34·4 (229/665)29·5 (352/1193)24·0 (390/1623)26·3 (742/2816)

Overall safety

Placebo-controlled period

After adjusting for duration of follow-up, rates of overall AEs, common AEs and investigator-reported infections remained comparable between ustekinumab- and placebo-treated patients (Table 2). Generally similar rates of SAEs were reported across groups; the most commonly reported were infections (= 8), malignancies (= 7) and cardiac disorders (= 7). Overall, few AEs led to early treatment discontinuation.

Table 2. Summary of adverse events per 100 patient-years for up to 5 years of follow-up; treated patients in the pooled ustekinumab psoriasis safety cohort
 Placebo-controlled perioda (Phase II, PHOENIX 1 and 2)Controlled and uncontrolled periodsb (Phase II, PHOENIX 1 and 2, ACCEPT)
PlaceboUstekinumabUstekinumab
45 mg90 mg45 mg90 mgCombined
  1. aIncludes only data from the Phase II, PHOENIX 1 and PHOENIX 2 studies because the ACCEPT study did not include a placebo control group. bIncludes data from all four studies (Phase II, PHOENIX 1 and PHOENIX 2, ACCEPT). cCommon adverse events are defined as adverse events occurring at an event rate of 2 events per 100 patient-years of follow-up. dData presented are numbers of patients with events per 100 patient-years of follow-up. eSerious adverse events are listed by system organ class adapted from the Medical Dictionary for Regulatory Activities (MedDRA v.14.1) classification system occurring at an event rate of 0·2 events per 100 patient-years of follow-up. All other serious adverse events are included in the ‘Other’ category. fMalignancies listed as serious adverse events include all malignancies, regardless of whether they were rated as serious by the investigator. g‘Other’ category includes: benign neoplasms; blood and lymphatic system disorders; congenital, familial and genetic disorders; ear and labyrinth disorders; endocrine disorders; eye disorders; immune system disorders; investigations; metabolism and nutrition disorders; pregnancy, puerperium and perinatal conditions; reproductive system and breast disorders; skin and subcutaneous tissue disorders; and surgical and medical procedures.

Patients treated, n732790792131920013117
Average weeks of follow-up12·613·413·4148·5136·0150·1
Total patient-years of follow-up177203203376652328998
Adverse events412·1516·2490·6242·6225·3232·6
Common adverse eventsc
 Nasopharyngitis36·236·933·421·020·620·8
 Upper respiratory tract infection20·926·124·617·415·416·2
 Headache25·430·034·47·56·87·1
 Arthralgia12·414·812·85·04·54·7
 Sinusitis6·25·45·94·44·34·3
 Back pain4·59·47·94·64·14·3
 Influenza2·84·93·44·44·04·2
 Bronchitis3·43·92·03·93·73·8
 Hypertension6·26·45·43·63·33·4
 Gastroenteritis5·15·93·43·32·73·0
 Cough6·84·45·42·82·92·9
 Diarrhoea6·810·89·83·22·72·9
 Injection-site erythema1·74·48·42·33·02·7
 Muscle strain4·53·92·92·11·92·0
 Viral upper respiratory tract infection1·73·92·52·41·82·0
 Fatigue8·512·811·82·51·72·0
 Nausea6·26·47·42·01·51·7
Infections, investigator-reported121·5146·7135·789·884·186·5
Adverse events leading to discontinuationd9·85·56·42·42·52·4
Serious adverse events6·87·47·97·07·27·1
Serious adverse events by body systeme
 Infections1·70·52·00·91·21·1
 Malignanciesd,f1·71·01·01·21·11·1
 Cardiac disorders0·00·52·91·11·11·1
 Injury, poisoning and procedural complications0·01·50·00·60·60·6
 Gastrointestinal disorders0·60·00·00·50·60·6
 Nervous system disorders0·61·50·00·40·40·4
 Musculoskeletal and connective tissue disorders0·61·50·00·60·30·4
 Respiratory, thoracic and mediastinal disorders0·60·00·00·30·40·3
 General disorders and administration site conditions0·60·50·00·20·40·3
 Renal and urinary disorders0·00·50·00·50·20·3
 Psychiatric disorders0·60·50·50·30·20·3
 Vascular disorders0·00·50·50·20·20·2
 Hepatobiliary disorders0·00·00·00·20·20·2
 Otherg1·10·02·00·60·80·7

Controlled and uncontrolled period up to year 5

To evaluate the long-term impact of dose, safety event rates up to year 5 in patients receiving ustekinumab 45 and 90 mg were compared. Overall AEs (45 mg, 242·6/100 PY; 90 mg, 225·3/100 PY) and investigator-reported infections (45 mg, 89·8/100 PY; 90 mg, 84·1/100 PY) occurred at comparable rates between doses (Table 2). Consistent with the controlled period, the most commonly reported AEs were nasopharyngitis, upper respiratory tract infection (URTI), headache and arthralgia. Similar rates of SAEs were reported for the 45-mg (7·0/100 PY) and 90-mg (7·2/100 PY) groups (Table 2). Consistent with the controlled period, infections, malignancies and cardiac disorders were most commonly reported. The proportions of patients with abnormal haematology and chemistry laboratory values were generally low and comparable between dose groups (data not shown). Overall, the number of AEs leading to treatment discontinuation was low and comparable between doses (45 mg, 2·4/100 PY; 90 mg, 2·5/100 PY). Twenty deaths (0·22/100 PY) were reported through year 5 (45 mg, 0·13/100 PY; 90 mg, 0·29/100 PY). The causes of death were considered related to CV events (= 5), malignancy (= 5), infection (= 3) and other causes (= 7), with the greatest disparity between the dose groups in the last category (45 mg, gunshot wound; 90 mg, one postoperative haemorrhage, three accidents, two suicides). The observed mortality rate among ustekinumab-treated patients was consistent with that expected in the general U.S. population (SMR = 0·36, 95% CI 0·22–0·55).

To examine the impact of cumulative exposure, safety events were evaluated further by year of follow-up. From year 1 to year 5, rates of overall AEs (data not shown), and AEs leading to discontinuation (Fig. 2) generally decreased. SAE rates demonstrated year-to-year variability with no increasing trend (Fig. 2). The proportions of patients with abnormal haematology and chemistry laboratory values were generally consistent over time (data not shown). Overall, rates of common AEs and SAEs among ustekinumab-treated patients during long-term follow-up were generally consistent with, or lower than, those reported during the controlled period.

Figure 2.

 Number of treatment-emergent adverse events per 100 patient-years of follow-up (PY) (overall rates for 5 years and rates by each year of follow-up) in the pooled ustekinumab psoriasis safety cohort. AE to D/C, adverse event leading to discontinuation; MACE, major adverse cardiovascular event; NMSC, nonmelanoma skin cancer; SAE, serious adverse event.

Adverse events of interest

The same analytical approach that was used to evaluate overall safety was applied to a number of targeted safety events of interest, including infections,17,18 NMSC,19,20 other malignancies21,22 and MACE,23,24 which have been observed at increased rates in patients with psoriasis.

Infections

Through year 5, event rates were comparable between the ustekinumab 45 and 90 mg groups, respectively, for overall infections (89·82 and 84·14/100 PY), infections requiring oral/parenteral antimicrobial treatment (29·0 and 27·6/100 PY) and serious infections (0·98 and 1·19/100 PY) (Table 3). Rates of overall infections decreased over time (data not shown); rates of serious infections demonstrated year-to-year variability with no increasing trend (Fig. 2). Overall, discontinuation of study treatment due to infections was uncommon (0·7%; Table 4). The most frequently reported common infections were nasopharyngitis (28·6%) and URTI (24·7%); the most frequently reported serious infections were diverticulitis (0·38%), cellulitis (0·32%) and pneumonia (0·32%) (Table 4). The proportion of patients reporting herpes zoster remained low (0·9%, 45 mg; 1·4%, 90 mg; 1·3%, overall), and the difference between dose groups [45 mg, 0·32/100 PY (95% CI 0·16–0·56) and 90 mg, 0·54/100 PY (95% CI 0·36–0·77)] narrowed compared with that reported at 4 years of follow-up.25

Table 3. Summary of adverse events of interest per 100 patient-years; treated patients in the pooled ustekinumab psoriasis safety cohort for up to 5 years of follow-up
 Ustekinumab
45 mg90 mgCombined
  1. aIncludes adverse events designated as infection by the investigator on the case report form; bdata presented are number of events per 100 patient-years of follow-up [95% confidence interval (CI)]; cdata presented are number of patients with events per 100 patient-years of follow-up (95% CI); dMACE includes cardiovascular death, myocardial infarction and stroke, as adjudicated by an independent and blinded panel at the Cleveland Clinic Coordinating Center for Clinical Research.

Patients treated, n131920013117
Infectionsa,b89·82 (86·82–92·90)84·14 (81·67–86·66)86·52 (84·61–88·47)
 Infections requiring treatment28·97 (27·27–30·74)27·58 (26·18–29·04)28·16 (27·08–29·28)
 Serious infections0·98 (0·69–1·35)1·19 (0·91–1·52)1·10 (0·89–1·34)
Malignanciesc1·23 (0·90–1·64)1·06 (0·80–1·37)1·13 (0·92–1·37)
 Nonmelanoma skin cancer (NMSC)0·64 (0·41–0·95)0·44 (0·28–0·66)0·52 (0·39–0·70)
 Other malignancies (excluding NMSC)0·59 (0·37–0·89)0·61 (0·42–0·87)0·60 (0·45–0·78)
Major adverse cardiovascular events (MACE)b,d0·56 (0·35–0·85)0·36 (0·22–0·57)0·44 (0·32–0·61)
 Cardiovascular death0·08 (0·02–0·23)0·06 (0·01–0·17)0·07 (0·02–0·15)
 Myocardial infarction0·40 (0·22–0·66)0·31 (0·17–0·50)0·34 (0·23–0·49)
 Stroke0·08 (0·02–0·23)0·00 (0·00–0·06)0·03 (0·01–0·10)
Table 4. Number of patients with infections by type; treated patients in the pooled ustekinumab psoriasis safety cohort for up to 5 years of follow-up
 Ustekinumab
45 mg90 mgCombined
  1. aOccurring in ≥ 5·0% of patients in combined group. bOccurring in ≥ 3 patients in the combined group. A total of 99 serious infection events were reported by 88 patients. cIncludes diverticulitis, diverticular perforation and intestinal haemorrhagic diverticulitis. dIncludes pneumonia, lobar pneumonia, staphylococcal pneumonia and streptococcal pneumonia. eIncludes appendicitis and perforated appendicitis.

Patients treated, n131920013117
Patients with at least one infection, n (%)a1011 (76·6)1354 (67·7)2254 (72·3)
 Nasopharyngitis381 (28·9)540 (27·0)892 (28·6)
 Upper respiratory tract infection350 (26·5)439 (21·9)770 (24·7)
 Sinusitis118 (8·9)162 (8·1)274 (8·8)
 Influenza120 (9·1)158 (7·9)273 (8·8)
 Bronchitis117 (8·9)145 (7·2)257 (8·2)
 Gastroenteritis92 (7·0)111 (5·5)201 (6·4)
Patients with at least one infection requiring treatment, n (%)541 (41·0)707 (35·3)1217 (39·0)
Patients with at least one serious infection, n (%)b33 (2·5)55 (2·7)87 (2·8)
 Diverticulitisc6 (0·45)6 (0·30)12 (0·38)
 Cellulitis5 (0·38)5 (0·25)10 (0·32)
 Pneumoniad4 (0·30)6 (0·30)10 (0·32)
 Appendicitise2 (0·15)4 (0·20)6 (0·19)
 Cholecystitis2 (0·15)2 (0·10)4 (0·13)
 Bronchitis2 (0·15)1 (0·05)3 (0·10)
 Osteomyelitis1 (0·08)2 (0·10)3 (0·10)
 Sepsis1 (0·08)2 (0·10)3 (0·10)
 Viral infection2 (0·15)1 (0·05)3 (0·10)
Patients who discontinued study agent due to infection, n (%)8 (0·6)14 (0·7)22 (0·7)

Nonmelanoma skin cancer

Up to year 5, rates of NMSC were comparable for the ustekinumab 45 mg (0·64/100 PY) and 90 mg (0·44/100 PY) groups (Table 3). Over time, NMSC rates varied year to year with no increasing trend (Fig. 2). Of the 47 patients with NMSCs [including three patients reporting both basal cell carcinoma (BCC) and squamous cell carcinoma (SCC)], 40 had BCC and 10 had SCC (BCC : SCC ratio of 4 : 1) (Table 5). A significantly higher proportion of patients with prior PUVA exposure reported NMSC compared with those with no prior exposure [2·9% (25/85) vs. 1·0% (22/2262); < 0·001]. Prior exposure to other treatments (ultraviolet B, conventional systemics, biologics) demonstrated no impact (data not shown).

Table 5. Number of patients with one or more malignancy by type; treated patients in the pooled ustekinumab psoriasis safety cohort for up to 5 years of follow-up
 Ustekinumab
45 mg90 mgCombined
  1. NMSC, nonmelanoma skin cancer.aPatient was reported to have pre-existing malignant superficial spreading melanoma. bOne patient reported pre-existing cutaneous T-cell lymphoma prior to study entry and one possible case of Hodgkin disease based on autopsy findings in a patient who died from traumatic bowel perforation.26

Patients treated, n131920013117
Patients with NMSC, n242347
 Basal cell carcinoma211940
 Squamous cell carcinoma5510
Patients with ≥ 1 malignancy other than NMSC, n223254
 Prostate8614
Melanoma246
  Melanoma in situ235
  Invasive melanomaa011
 Colorectal235
 Breast314
 Head and neck123
 Renal123
 Bladder202
 Leukaemia022
 Lymphomab022
 Myeloma022
 Pancreatic112
 Adenocarcinoma, unknown primary011
 Cervical011
 Endometrial011
 Oesophageal011
 Lung101
 Ovarian011
 Squamous cell carcinoma, unknown primary011
 Testicular011
 Thyroid101

Other malignancies

Through year 5, 54 patients reported malignancies other than NMSC and rates were comparable between the ustekinumab 45 mg (0·59/100 PY) and 90 mg (0·61/100 PY) groups (Table 3). Over time, there was year-to-year variability, but no increasing trend was evident (Fig. 2). The observed rate of other malignancies in ustekinumab-treated patients was comparable with that expected in the general U.S. population (SIR = 0·98; 95% CI 0·74–1·29) (Fig. 3). For the most commonly reported malignancies (Table 5), SIRs were 1·21 (95% CI 0·66–2·04) prostate, 1·42 (95% CI 0·52–3·09) melanoma, 0·99 (95% CI 0·32–2·31) colorectal, and 0·62 (95% CI 0·17–1·58) breast. Lymphoma (SIR = 0·80; 95% CI 0·10–2·91) occurred in two patients: one pre-existing cutaneous T-cell lymphoma at study entry misdiagnosed as psoriasis, and one possible case of Hodgkin disease based on autopsy findings in a patient who died from traumatic bowel perforation, as previously reported.26

Figure 3.

 Standardized incidence ratio (SIR) and 95% confidence interval for comparisons between malignancies other than nonmelanoma skin cancer in the pooled ustekinumab psoriasis safety cohort and the U.S. National Cancer Institute Surveillance, Epidemiology and End Results (SEER) database. *The SIR was calculated by dividing the observed number of ustekinumab-treated patients with events by the expected number of patients with events in the SEER database.

Major adverse cardiovascular events

Through year 5, the overall rate of MACE was comparable between the 45 mg (0·56/100 PY) and 90 mg (0·36/100 PY) groups. Individual event rates were 0·08 and 0·06 (CV death), 0·40 and 0·31 (MI), 0·08 and 0·00 (stroke), respectively (Table 3). Forty MACE (45 mg, = 21; 90 mg, = 19) were reported in 37 patients who were treated with ustekinumab; the majority were MIs, accounting for 78% (31/40) of all MACE reported. There was year-to-year variability, but no trend towards increasing events over time was noted (Fig. 2). The proportion of patients with CV risk factors was similar at year 5 compared with baseline (data not shown). All patients with MACE in this analysis had at least two established CV risk factors, and previous analyses identified a high proportion of patients as undiagnosed and undertreated at baseline.27

Immunogenicity

Rates of antibodies to ustekinumab ranged from 3·8% to 5·4% across studies, and positive antibody status was not associated with injection-site reactions. No cases of anaphylaxis or serum sickness-like reactions were reported.

Discussion

With approximately 9000 PY of follow-up in 3117 patients treated for up to 5 years, we evaluated the longest duration of continuous follow-up in the largest cohort of patients with psoriasis treated with a biologic agent in a clinical trial setting. Most observed AEs were nonserious, did not result in treatment discontinuation, and no clear signals of dose response or effects of cumulative exposure on safety outcomes were observed. These results confirm that the long-term safety profile of ustekinumab continues to be favourable and consistent with previous reports at earlier follow-up.14,25,26

Through year 5, rates and types of commonly reported AEs, SAEs and AEs of interest remained generally comparable between dose groups, and event rates were consistent with, or lower than, those reported during the controlled period. The difference initially observed in the incidence of serious infections between the 45 and 90 mg groups after up to 3 years of treatment26 continued to narrow with two additional years of follow-up, suggesting little difference in infection risk between the two doses. The decreasing trend observed over time for nonserious AEs and infections may be attributed to reporting fatigue, and longer intervals between study visits during the long-term extensions possibly impacted patients’ ability to recall minor events. The incidence of antibodies to ustekinumab and AEs secondary to laboratory abnormalities were low and consistent over time, with no evidence of cumulative end-organ toxicity. The number of patients who discontinued treatment due to any AE was low, although patients at higher risk for AEs may have terminated early. Overall, these results suggest that the dose of ustekinumab and the longer duration of exposure had no effect on the incidence of overall and targeted safety events after 5 years of follow-up.

The risk of infection is a concern in patients requiring long-term immunosuppressive treatment, and the inherent risk may be elevated in psoriasis.18 With up to 5 years of ustekinumab treatment, no particular infection pattern and no infections related to interleukin (IL)-12/23 deficiency were identified. Patients who are genetically deficient in IL-12/23p40 and IL-12Rβ1 demonstrate increased susceptibility to weakly virulent mycobacterial and salmonella infections,28–30 and IL-23/IL-17A may be implicated in host defence towards certain fungal species such as Candida albicans.31,32 No atypical mycobacterial diseases, disseminated salmonellosis or systemic fungal infections were observed in ustekinumab-treated patients. In addition, no opportunistic infections were reported, except one previously reported case of disseminated cutaneous herpes zoster infection with no evidence of visceral involvement.8 It is possible that, during long-term ustekinumab treatment, IL-12/23 inhibition is incomplete and does not compromise host defence towards these pathogens. The association of active tuberculosis and latent tuberculosis infection (LTBI) reactivation with some antitumour necrosis factor (TNF) agents33–35 was not observed in our analyses, although one case of LTBI reactivation was reported previously in a Taiwanese patient who did not receive antituberculosis prophylaxis.36 Further investigation is needed to determine if the risk of developing tuberculosis is different between biologics. Serious infections occurred infrequently among ustekinumab-treated patients (1·10/100 PY), at a rate generally comparable with those reported for anti-TNF agents in psoriasis (range 0·90–1·82/100 PY).33,35,37,38 The incidences of serious infections reported here and across biologic studies are generally comparable with rates reported in epidemiological evaluations of the general psoriasis population (range 0·91–1·11/100 PY).18,39 Overall, our analyses identified no increased risk for infection after 5 years of treatment with ustekinumab.

Psoriasis may be associated with an increased risk of lymphoproliferative malignancies22 and NMSC,19,20 independent of the potential risk attributable to long-term exposure to systemic treatment.40 Preclinical studies in murine models have associated IL-12/23 inhibition with a theoretical risk of tumour promotion;21 therefore, monitoring malignancy risk during long-term treatment with ustekinumab remains a priority. The spectrum of malignancies observed in our analysis was consistent with that expected in the general population. Prostate cancer was most prevalent, which may be expected given the predominantly male study population. Because trial patients had greater access to routine dermatological care than the typical patient, melanoma may have been detected more frequently and earlier.41,42 Overall, the incidence of malignancies other than NMSCs in ustekinumab-treated patients (0·60/100 PY, SIR = 0·98) was consistent with that expected in the general U.S. population and comparable with that reported for anti-TNF agents in psoriasis (range 0·54–0·72/100 PY).34,35,37,38 Consistent with the literature,2 a higher incidence of NMSC was reported among ustekinumab-treated patients with prior PUVA exposure, supporting a possible predisposition to NMSC with prior PUVA treatment. Overall, the incidence of NMSC in ustekinumab-treated patients (0·52/100 PY) is comparable with, but somewhat lower than, rates reported for anti-TNF agents in psoriasis (range 0·70–1·17/100 PY).34,35,37,38 Whether these observations represent true differences in risk, or are the result of differences between study designs and patient populations, requires further investigation. Reversal of the BCC:SCC ratio as a result of increased rates of SCC have been reported in the setting of profound immunosuppression (e.g. organ transplant).43 The BCC:SCC ratio observed in ustekinumab studies (4 : 1) is consistent with that expected in immunocompetent patients in the general population.43,44 Overall, no evidence of a shift towards a pattern of malignancy (lymphoproliferative cancers and inversion of the BCC:SCC ratio) indicative of severe immunosuppresion,43–45 and no increase in malignancy risk were observed in patients treated with ustekinumab for up to 5 years.

Studies have associated psoriasis with an increased risk of adverse CV outcomes,23,46 including an elevated 10-year risk of MACE.47 The highest relative risks have been associated with younger patients with severe psoriasis.23,48 It has been suggested that certain systemic therapies (e.g. methotrexate,49,50 anti-TNF agents51) may lower the risk of adverse CV outcomes in patients with rheumatoid arthritis; therefore, the impact of psoriasis treatments on CV safety is of interest to the dermatology community. Recent evaluations of the impact of systemic and biologic therapies on CV risk in patients with psoriasis reported inconsistent results,52,53 and differing conclusions as to whether MACE may represent treatment-emergent AEs associated specifically with anti-IL-12/23 agents were reported.54–56 The data presented here do not identify an increased risk of MACE in patients treated for up to 5 years with ustekinumab. The overall rate of MACE in ustekinumab-treated patients (0·44/100 PY) was comparable with those reported for anti-TNF agents in psoriasis (range 0·36–0·84/100 PY)34,38 and, collectively, the reported rates of MACE across studies of biologics are lower than those reported in epidemiological studies of the general psoriasis population receiving nonbiologic systemic treatments (1·64/100 PY, MACE).47 Whether or not biologics in general, and IL-12/23 inhibitors in particular, are associated with an increased or a decreased risk of CV disease in psoriasis remains to be investigated in larger observational registries57,58 and future meta-analyses that will include more patients and sufficiently sized comparator cohorts.

Several limitations should be noted. The lack of randomized dose groups beyond the initial controlled period may have confounded the assessment of differences between the two dose groups. By design, only the PHOENIX studies evaluated the impact of ustekinumab treatment for up to 5 years of follow-up. In the absence of a long-term control group, efforts were made to contextualize our observations using data from studies of other biologics known to have evaluated similar patient populations and data from epidemiological studies of the general U.S. and general psoriasis populations. Clinical trial patients are selected based on specific protocol-defined criteria and receive closer medical follow-up; therefore, comparisons with external general populations may be limited by the number of available confounding variables that could be controlled.

In conclusion, the data presented in this report confirm that the safety profile of ustekinumab remains favourable after up to 5 years of treatment in patients with moderate-to-severe psoriasis. Neither the dose of ustekinumab nor the cumulative exposure to ustekinumab impacted long-term safety outcomes. Results of the current safety evaluation are consistent with shorter-term reports of ustekinumab safety and are generally comparable with observations from studies of other approved biologics in patients with psoriasis. Long-term registries will continue to evaluate the long-term impact of ustekinumab treatment on overall and targeted safety events.57,58

Acknowledgments

The authors would like to acknowledge Cynthia Arnold and Kristin Ruley Sharples, PhD, of Janssen Services, LLC, Spring House, PA, U.S.A., for their editorial assistance and writing support; Yin You of Janssen Research & Development, LLC, Spring House, PA, U.S.A., for her programming and statistical support; and Newman Yeilding, MD, and Shu Li, PhD, both of Janssen Research & Development, LLC, Spring House, PA, U.S.A., for their critical review of the manuscript.

Appendix

Conflicts of interest

K.A.P. has received grants and/or honoraria as a consultant, investigator, member of an advisory board, or speaker from Abbott, Amgen, Anacor, Astellas, Celgene, Eli Lilly, Janssen, Johnson & Johnson, Galderma, LEO Pharma, Merck, Novartis, Pfizer and UCB. C.E.M.G. has received research support and/or has acted as a consultant or lecturer for Abbott, Alliance-Boots, Amgen, Biogen-IDEC, Celgene, Centocor, Galderma, Incyte, Janssen-Cilag, L’Oreal, LEO Pharma, Novartis, Novo Nordisk, Pfizer, P&G, Schering Plough, Stiefel and BioTest. K.G. has served as a consultant for Abbott, Amgen, Eli Lilly, Novartis and Pfizer, and received research (grant) support from Janssen Biotech, Abbott and Amgen. M.L. has received honoraria and/or grant funding as a consultant and/or investigator, and/or speaker from Abbott, Amgen, Anacor Pharmaceuticals, BioLineRX, Celgene, Dermipsor, Eli Lilly, Galderma, GlaxoSmithKline-Stiefel, Janssen, LEO Pharma, Maruho, Novartis, Pfizer, Ranbaxy and Valeant. V.H. has served as an advisor and/or investigator and/or speaker for Abbott, Amgen, Basilea, Janssen, Merck, Novartis and Pfizer. P.D.G. has served as an investigator and/or consultant and/or speaker for Abbott, Amgen, Centocor/Janssen/Johnson & Johnson, MSD, Novartis and Pfizer. B.S. has served as an advisor and/or consultant, and/or speaker for Abbott, Amgen, Celgene, Eli Lilly, Janssen, Maruho, Novartis and Pfizer. K.R. has served as a consultant and/or paid speaker for and/or participated in clinical trials sponsored by companies that manufacture drugs used for the treatment of psoriasis including Abbott, Biogen Idec, Celgene, Centocor, Janssen-Cilag, LEO Pharma, Medac, Merck, MSD (formerly Essex, Schering-Plough), Novartis and Pfizer (formerly Wyeth). P.O.S., Y.W., D.C. and M.-C.H. are employees of Janssen Research & Development, LLC., and own stock in Johnson & Johnson, of which Janssen is a wholly owned subsidiary.

Ancillary