Objective To study efficacy and safety of escalating doses of canakinumab, a fully human anti-IL-1β monoclonal antibody in the severe cryopyrin-associated periodic syndrome, neonatal-onset multisystem inflammatory disease (NOMID).
Methods 6 patients were enrolled in this 24-month, open-label phase I/II study. All underwent anakinra withdrawal. The initial subcutaneous canakinumab dose was 150 mg (or 2 mg/kg in patients ≤40 kg) or 300 mg (or 4 mg/kg) with escalation up to 600 mg (or 8 mg/kg) every 4 weeks. Full remission was remission of patient-reported clinical components and measures of systemic inflammation and CNS inflammation. Hearing, vision and safety were assessed. Primary endpoint was full remission at month 6.
Results All patients flared after anakinra withdrawal, and symptoms and serum inflammatory markers improved with canakinumab. All patients required dose escalation to the maximum dose. At month 6, none had full remission, although 4/6 achieved inflammatory remission, based on disease activity diary scores and normal C-reactive proteins. None had CNS remission; 5/6 due to persistent CNS leucocytosis. At the last study visit, 5/6 patients achieved inflammatory remission and 4/6 had continued CNS leucocytosis. Visual acuity and field were stable in all patients, progressive hearing loss occurred in 1/10 ears. Adverse events (AEs) were rare. One serious AE (abscess due to a methicillin-resistant Staphylococcus aureus infection) occurred.
Conclusions Canakinumab at the studied doses improves symptoms and serum inflammatory features of NOMID, although low-grade CNS leukocytosis in four patients and headaches in one additional patient persisted. Whether further dose intensifications are beneficial in these cases remains to be assessed.
ClinicalTrials.gov identifier NCT00770601.
- DMARDs (Biologic)
- Disease Activity
Statistics from Altmetric.com
Neonatal-onset multisystem inflammatory disease (NOMID) is a member of the cryopyrin-associated periodic syndromes (CAPS) family of disorders caused by autosomal-dominant gain of function mutations in CIAS1 (also called NLRP3, NALP3 or PYPAF1) encoding cryopyrin.1 ,2 CAPS are a spectrum of diseases including in increasing severity, familial cold autoinflammatory syndrome (FCAS), Muckle–Wells syndrome (MWS) and NOMID. These disorders present with fevers, neutrophilic urticaria, arthritis and elevated inflammatory markers. Progressive organ damage is caused by ongoing organ inflammation. Persistent cochlear inflammation leads to hearing loss; persistent aseptic meningitis causes elevated intracranial pressures and optic nerve atrophy that can contribute to cognitive impairment and progressive vision loss, respectively; ocular inflammation manifests as conjunctivitis and uveitis.
Initial proof-of-concept studies with the short-acting recombinant IL-1 receptor antagonist, anakinra, resulted in dramatic improvements in CAPS outcomes, demonstrating the critical role of IL-1 in these diseases.3–8 Canakinumab (Ilaris) is a fully human anti-IL-1β monoclonal antibody. Potential advantages over anakinra include a more convenient dosing schedule and possibly the specificity for IL-1β alone. It is effective in the treatment of FCAS and MWS,8 the less severe forms of CAPS, and was approved by the FDA for their treatment in June 2009. However, optimal dosing and effects in controlling the systemic and organ-specific disease manifestations in patients with NOMID are less well understood.
Seven patients meeting criteria for NOMID were recruited from an ongoing study (NCT00069329) and enrolled at the NIH Clinical Center from January 2009 to February 2011. One patient withdrew prior to undergoing study procedures. Patients had at least two of the following manifestations: neutrophilic urticaria, CNS involvement and epiphyseal or patellar overgrowth on radiographs (table 2 and see online supplementary material).
Study design and treatment
This open-label study was designed to investigate an optimal dosing regimen, safety and efficacy of canakinumab in six NOMID patients at 6 months with possible extension to 24 months. All patients were on anakinra (dose range 2.1–4.5 mg/kg/day) at screening. Because of the severity of flares with anakinra withdrawal, patients were taken off anakinra for only 6–48 h. The predefined baseline visit occurred at the time of drug withdrawal.
The first three patients were treated with an initial canakinumab dose of 150 mg subcutaneously (or 2 mg/kg for patients ≤40 kg). All other patients were treated with an initial dose of 300 mg (or 4 mg/kg for patients ≤40 kg) subcutaneously. Doses were escalated as per online supplementary methods and supplemental table 1.
The study was discontinued on 17 February 2011 due to protracted recruitment of severe patients with NOMID, the lack of intent to expand to additional centres and no new safety signals detected.
Assessment of clinical outcomes
The prespecified primary outcome was the proportion of patients in full remission (inflammatory and CNS) at 6 months. A NOMID-specific daily diary was kept by the patient/parent at home. Inflammatory remission was defined as C-reactive protein (CRP) ≤10 mg/L and global diary score remission (weekly averaged global diary score ≤2). CNS remission was defined as a CSF white blood cell count (WBC) ≤15 cells/μL and headache diary score remission (weekly averaged score ≤0.5).
Clinical at 12, 18 and 24 months
Visual analogue scale (VAS) for pain and overall disease activity were completed by the parent/patient and the physician at each NIH visit. Height, weight and ACR 66/68 tender and swollen joint counts were recorded at each visit. Radiographs were performed at the screening and end-of-study visits.
Erythrocyte sedimentation rate (ESR), serum amyloid A (SAA) and CRP were analysed at a central laboratory.
Lumbar punctures (LPs) were performed. MRIs with gadolinium-enhanced fluid-attenuated inversion recovery (FLAIR) sequences of the brain and inner ear were performed and scored by JB.9
Audiology data were transferred from the NOMID natural history study database (ClinicalTrials.gov identifier: NCT00059748). Hearing assessment included audiological evaluations.9 Outcomes in each ear were categorised as ‘stable’ or ‘worsened’, according to a modification of the American Speech and Hearing Association (ASHA) criteria.10
All patients underwent ophthalmological exams, including visual acuity and field analysis, and dilated eye exams, which were scored by RB.9
Adverse events were recorded and graded according to the NCI common terminology criteria for adverse events (CTCAE) v4.0 after study completion.
Descriptive statistics were used for demographic and baseline characteristics. Means, CIs and standard errors of the mean (SEM) were calculated.
Patient demographics and baseline characteristics
At enrolment, patients were between 11 and 34 years of age, mean 18.7 (8.09). Four patients had genomic and two had somatic CIAS1 mutations (table 1).11 Organ damage was recorded (table 2). All patients showed evidence of a flare at baseline after discontinuation of anakinra—4/6 had elevations of CRP, 2/6 had elevated CSF leucocytes, 3/6 had abnormal global diary scores and 5/6 had increased headache diary scores.
Primary outcome: remission at the primary endpoint (month 6)
All patients exhibited worsening of their clinical symptoms within the first week, and acute phase reactants increased within the first 8 days after initiating canakinumab and resulted in canakinumab dose escalation. At 6 months, three adult patients were on canakinumab, 600 mg subcutaneously every 8 weeks (n=1) and 300 mg every 8 weeks (n=2). Two paediatric patients were on 8 mg/kg every 8 weeks. The last adult patient was enrolled on a faster dose escalation regimen and was at 600 mg every 4 weeks.
All patients demonstrated clinical improvements from baseline with 4/6 patients fulfilling inflammatory remission criteria at month 6. CNS remission as defined by the protocol was not achieved by any of the six patients. Five patients would have been in remission based on headache diary scores alone and the remaining patient had had a normal CSF leucocyte count, but with persistent headaches. Since full remission criteria included inflammatory and CNS remission, no patient met the primary endpoint.
Although not prespecified, inflammatory, CNS and full remission at any time point with available data were assessed. LPs were available at 25/142 follow-up visits (2–6 per patient). Overall, each patient was in both full and CNS remission for 10.3% of these visits (range 0–25%). Using all available CRP values with associated global diary scores, each patient was in inflammatory remission for 70.7% of the visits (range 14.8–93.3%).
Remission of primary measures during the extension phase
To address whether increased doses are necessary to obtain better CNS control, the study was extended in the first five patients. Median duration in the trial was 615 days (232–749 days). By the end of the study, 5/6 patients required the maximal allowable dose, 600 mg every 4 weeks for those ≥40 kg (n=2, starting day 406 and 407, respectively) and 8 mg/kg every 4 weeks for those <40 kg (n=2, starting day 274, 401). The one patient who was not at maximum allowable dose was treated with 600 mg every 6 weeks at day 455, end of study (see online supplementary table S1).
Of the five patients who entered the extension phase, four had LPs as one patient did not undergo LPs after 6 months due to technical difficulties. At 12 months, 4/4 were in inflammatory remission but only 1/4 patients was in CNS remission (2/4 had normal headache diary scores and 2/4 and normal CSF leucocyte counts). At 18 months, one additional patient did not undergo an LP for technical issues. Inflammatory remission was achieved in 3/3 patients and 1/3 was in CNS remission (3/3 had normal headache diary scores and 1/3 a normal CSF leucocyte count). Two patients had 24-month evaluations. Both were in inflammatory remission but neither were in CNS remission (both had normal headache diary scores but had CSF leucocytosis) (table 3).
At 6 months, investigator global VAS score, patient/parent global VAS score and patient/parent global diary scores decreased in 5/6 patients compared with baseline with decreased mean scores (see online supplementary figure S1). Of the six enrolled patients, five had headaches after anakinra withdrawal at baseline—in all five mean headache diary scores decreased at 6 months.
Similar results were seen during the extension phase of the study where further improvements in clinical outcomes were seen in the majority of patients at all the time points. Patient/parent global diary scores improved in 5/6 patients and headache diary scores improved in all patients reporting headaches at baseline. Investigator global VAS score improved from baseline in 3/5 patients at 12 months, 4/5 at 18 months and 2/2 at 24 months. Patient/parent global VAS improved in 3/5 patients at 12 months, 4/5 at 18 months and 2/2 at 24 months (see online supplementary figure 1).
Mean CRP, SAA and ESR values decreased on therapy compared with baseline values. Mean (SD) CRP at baseline was 18 mg/L (16.2), which decreased to 11.5 mg/L (20.4) at the primary endpoint (figure 1). Also, 3/4 patients with elevated CRP levels at baseline showed improvements at 6 months with only one patient showing abnormal values at that time. The remaining patients improved but did not normalise during the extension phase. Five patients had SAA levels tested. These were all abnormal at baseline (>10 mg/L) and decreased from 96.4 (100.3) to 63.2 (118.2) at 6 months with 3/5 patients showing improvements. The two patients with abnormal values at 6 months showed improvements during the extension phase. Mean ESR at baseline was 15.5 (7.8), which decreased to 12 (10) at 6 months. Also, 4/6 patients showed improvements at 6 months with the remaining two patients showing improvement in the extension phase.
See table 2 for a description of baseline organ manifestations and damage. Patients were at the severe end of the NOMID spectrum. All patients had fevers, urticaria-like rashes and elevated serum inflammatory markers.
Mean 4F-PTA hearing values of the entire cohort did not change throughout the study (data not shown). One ear with baseline mild hearing loss further worsened during the extension phase with a 10 dB loss in bone conduction in two consecutive frequencies at 12 months, which worsened to 25 and 30 dB loss at 24 months.
The mean cochlear enhancement score on MRI did not change over the course of the study (data not shown). Cochlear enhancement was stable in the majority of ears at the primary endpoint (n=7), worsened in two ears and improved in three ears.
Visual acuity and field did not worsen in any patient over the course of the study. On low doses of canakinumab, worsening of papilledema developed in one patient and an episode of anterior uveitis in an additional patient. Both resolved with uveitis temporarily treated with steroid eye drops. The episode of uveitis was associated with disease-related conjunctivitis; however, conjunctivitis was not observed in any other patient.
At baseline, 2/6 patients had aseptic meningitis with a mean (SD) CSF leucocyte count of 10.3 (12.4). At the primary endpoint, mean counts increased to 37.2 (28.9) with aseptic meningitis present in 5/6 patients. Mean headache diary scores decreased from baseline to 6 months in all patients reporting headaches (n=5), which persisted throughout the extension phase for all patients (figure 1).
Leptomeningeal enhancement on MRI worsened in 3/6 patients, slightly improved in one patient and remained stable in two patients. The three patients with worsened enhancement showed a higher increase in their CSF WBC values compared with baseline values (mean increase of 51 vs 2.7) consistent with the notion that worsening of leptomeningeal enhancement represents leptomeningitis from CNS inflammation. IQs were stable.
Four children were enrolled under age 18. Three of the four were under the 5th centile of growth at baseline and remained so throughout the study. The other patient was normal height and remained the same percentile. Two adults had reached maximum growth potential without further growth. Joint pain or swelling was not present at baseline and mild swelling and tenderness occurred only twice with resolution at subsequent visits. No new bony lesions occurred on radiographs.
Medication dosing, pharmacokinetics and safety
All patients required dose escalation. Peak serum concentrations per patient ranged from 53.3 to 189 μg/mL. Canakinumab CSF drug concentrations ranged from 0 to 650 ng/mL with a median concentration of 210 ng/mL; corresponding to 0 to 0.011 of the relative serum concentration with a mean ratio of 0.0048 per patient. With IC50 of canakinumab in vitro at 7.1 ng/mL,12 these levels were estimated to be sufficient for a pharmacodynamic effect. CSF drug concentration levels negatively correlated with CSF WBC counts (see online supplementary figure 2).
At baseline, endogenous serum IL-1β was detected in 5/7 patients, while CSF IL-1β was observed in 2/7 patients. Upon treatment with canakinumab, IL-1β levels in serum increased in all subjects, indicating capture of circulating IL-1β by canakinumab. CSF IL-1β levels also increased in all subjects, indicating similar capture of endogenous IL-1β by canakinumab in the CSF.
Canakinumab was overall well tolerated. One serious adverse event, an abscess due to methicillin-resistant Staphylococcus aureus occurred. Twelve infection-related AEs were reported by six patients (table 4). These were single episodes of ear infection, fungal infection, influenza, localised infection of the finger, sinusitis, joint injury, cough, nasal congestion, oropharyngeal pain and acne, as well as two episodes of subcutaneous abscesses due to staphylococcal infection in the same patient.
The clinical manifestations in CAPS are mainly due to excessive IL-1β, which is confirmed by the impressive clinical response to IL-1 inhibiting medications. IL-1 inhibition is the standard of care for patients with CAPS, and several agents are approved for their treatment.3–8 These therapies not only improve clinical signs and symptoms of the disease, but they halt progression of organ damage in the most severe end of disease, NOMID, with appropriate dose adjustments.9 ,13 Canakinumab was FDA approved for the treatment of CAPS in 2009, and this study was designed to investigate whether canakinumab doses from 150 mg (or 2 mg/kg for <40 kg) every 8 weeks up to 600 mg (or 8 mg/kg if <40 kg) every 4 weeks can control the systemic and the CNS manifestations of NOMID.
Canakinumab improves the symptoms and serum inflammatory laboratory findings of disease after a clinical flare induced by IL-1 blockade withdrawal in a severe subset of six NOMID patients. Consistent with a recent report,14 most patients experienced persistent improvements in global measures of disease activity with decreases in serum inflammatory markers. However, no patient achieved full remission criteria at the primary endpoint (6 months), which in our study included absence of CNS inflammation defined as a CSF WBC count of ≤15 cells/μL in CSF and low headache scores.
Recently, elevated levels of another biomarker of inflammation, CSF IL-6, was shown to correlate with poor clinical outcomes in a number of degenerative, traumatic and inflammatory neurologic conditions including lupus and Behcet's disease.15–17 In NOMID patients, CSF IL-6 levels were found to be 6-8 times higher in the CSF than serum and were correlated with the presence of inflammatory cells, specifically monocytes.18 Consistent with the markers of CNS inflammation reported in this study, data obtained in 5/6 during participation in a concurrent study (data not published), showed persistently elevated CSF IL-6 levels while receiving canakinumab.
NOMID patients require high doses of IL-1-blocking agents to control CNS inflammation. We recently reported that increasing doses of IL-1 blocking agents are necessary to control CNS disease. Using a conservative cut-off of 5 cells/μL, we reported that 11/25 patients had CSF leucocytosis after 36 months on anakinra and 5/18 still had leucocytosis at 60 months despite continued dose escalation.9 In this study, increased CSF leucocyte cell counts of greater than 15 cells/μL occurred in all but one patient compared with baseline despite improvements in headache diary scores in 5/6 patients. In three patients, leptomeningeal enhancement increased from baseline with a concurrent increase in CSF WBC counts. No patient exhibited worsening of their IQ; however, the impact of even low-grade inflammation on cognition long term remains unknown.
Canakinumab doses at the beginning of the study were low and subsequently increased, resulting in dose-dependent clinical improvements. Canakinumab has a long half-life (23–26 days), and to achieve steady-state levels in blood (∼4 dosing cycles) it takes several months. Thus, for some measurements, patients may not have reached steady-state levels after dose escalations at the time of assessment.
Study limitations include the small sample size and open-label design. Severe patients with NOMID are rare, making it not feasible to enrol large numbers of patients. An open-label design was necessary as placebo control is unethical when the damaging effects of persistent inflammation on organ damage are well recognised. Different dosing regimens occurred in each patient as protocol amendments allowing for higher doses occurred stepwise. Finally, the ability for further dose escalation to control disease could not be tested.
Overall, canakinumab at the low doses used to treat patients with FCAS and mild MWS was insufficient to control disease activity in severe patients with NOMID. With dose increases up to 600 mg (8 mg/kg) every 4 weeks symptom control was achieved and CRP levels were low, but CSF pleocytosis persisted in 5/6 patients. Recent reports show that progressive dose increases are needed in severe patients with CAPS;15 ,16 however, whether higher canakinumab doses than those used in this study may result in improved control of CNS inflammation remains unknown.
NOMID serves as a model disease for studying the impact of excessive IL-1 signalling in human disease. There may be subtle differences between IL-1-blocking therapies on disease control at the organ level, which need to be further investigated. The impact of drug concentration and tissue penetration may be factors in adequate inflammatory control and possibly prevention of damage progression in the most severe patients. As the importance of IL-1 is becoming increasingly recognised in more common diseases, including gout, diabetes, Alzheimer's and coronary artery disease, lessons learned from the study of IL-1 blockade in NOMID will be instructive in designing and interpreting further treatment studies in these disorders and in addressing the critical issue of tissue penetration for optimal target blockade.
Handling editor Tore K Kvien
Correction notice This article has been corrected since it was published Online First. In the second sentence of the ‘Remission of primary measures during the extension phase’ section of the Results ‘4/5 patients’ has been changed to ‘5/6 patients’.
Acknowledgements This study was funded by Novartis. This study was also supported by the Intramural Research Program of the National Institute of Arthritis and Musculoskeletal and Skin Diseases at the NIH. Portions of this work were also supported by the Intramural Research Programs of the National Cancer Institute, the National Institute on Deafness and Other Communication Disorders, the National Eye Institute and the NIH Clinical Center. Ms Rodriguez-Smith's research was made possible through the National Institutes of Health (NIH) Medical Research Scholars Program, a public-private partnership supported jointly by the NIH and generous contributions to the Foundation for the NIH from Pfizer Inc, The Doris Duke Charitable Foundation, The Alexandria Real Estate Equities, Inc. and Mr. and Mrs. Joel S. Marcus, and the Howard Hughes Medical Institute, as well as other private donors.
Contributors All authors meet ICJME criteria for authorship.
Competing interests AC, SF, LC and AC are employees of Novartis. G-M has received grant support from Regeneron, Novartis and SOBI Inc.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Unpublished data are available to interested parties upon request.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.