Background
Autoinflammatory diseases (AID) are rare conditions with high disease burden; many AID are caused by pathogenic gene variants leading to excessive production of pro-inflammatory cytokines [
1,
2]. Typical AID symptoms include recurrent fevers, inflammation of joints, eyes, skin and serous membranes coupled with increased inflammatory markers [
3]. However, the AID spectrum continuously expands and atypical presentations can result from somatic mosaicisms or low-penetrance variants [
4,
5]. Genetic testing has been established in recent years, allowing a reliable and definitive diagnosis in patients with pathogenic gene variants together with clinical AID symptoms. For patients with clinical AID symptoms and negative or inconclusive genetic testing it remains still challenging to make a diagnosis.
Effective AID treatment is crucial to prevent morbidity, mortality, reduce health-related quality of life and high psychosocial burden. Clinically and genetically confirmed AID can be effectively treated with targeted cytokine inhibition [
6]. In contrast, for patients with clinical AID symptoms without pathogenic gene variants or atypical AID symptoms, no treatment recommendations are available. This is particularly challenging for patients living in low-income countries.
Colchicine is approved and established as effective, safe and low-cost first-line therapy in Familial Mediterranean Fever (FMF) [
7]. Furthermore, colchicine might be an effective treatment for children with clinical AID symptoms but negative or inconclusive genetic testing. In the past years several reports illustrated that colchicine might be beneficial in the therapeutic management of patients with other AID than FMF, particularly in unclassified AID [
8‐
11]. However, more data on effectiveness and safety for colchicine in children with a clinical AID diagnosis without pathogenic gene variants are needed to support clinicians in their treatment decisions.
Therefore, the aims of this study were 1) to describe demographical, clinical and laboratory characteristics of children presenting with a clinical AID diagnosis without pathogenic gene variants and 2) to report the effectiveness of colchicine monotherapy in these patients.
Patients and methods
A single-centre pilot cohort study of consecutive children with a clinical diagnosis of AID without likely pathogenic or pathogenic gene variants was performed between 01/2009 and 12/2018. Children and adolescents ≤18 years of age were included, if they had a clinical AID diagnosis after exclusion of other differential diagnoses based on standardized assessments of fever, constitutional symptoms and classical AID organ manifestations paired with elevated inflammatory markers in flares. The diagnosis was supported by diagnostic or classification criteria where available [
12‐
14]. Patients not meeting a single specific set of AID criteria were considered unclassified AID. Furthermore, for study inclusion they had to fulfil the following criteria: 1) in the AID gene panel test (i) no detected gene variant including patients with periodic fever, aphthous stomatitis, pharyngitis and adenitis (PFAPA), or (ii) a gene variant of (likely) benign/uncertain significance according to the American College of Medical Genetics and Genomics [
15], and 2) evidence of active disease. A genetic AID panel test for variants in the
MEFV, MVK, TNFRSF1A, NLRP3, NOD2, PSTPIP1 and
LPIN2 gene was performed in all patients in a certified laboratory. The AID gene panel was expended by several newly discovered genes associated with AID, such as
IL10RA, IL1RN, IL36RN, ADA2 during the study period. Children were excluded, if they had 1) evidence of amyloidosis or organ damage, 2) previous or ongoing biologic therapy, 3) met criteria for paediatric Behçet disease [
16], 4) discontinued colchicine treatment or 5) were found to have significantly elevated liver enzymes (>3x upper limit of normal). Children with elevated liver enzymes were excluded from the study, because elevation of liver enzymes may be a side effect of colchicine therapy. Data was captured in the designated, institutional web-based Arthritis and Rheumatism Database and Information System (ARDIS) including standardized assessments of outcome measures at all visits [
17]. Ethic approval was obtained from the University of Tuebingen Institutional Review Board (012/2017BO2). A waiver of patients informed consent for the study was obtained. The study was performed in compliance with the Helsinki Declaration.
Demographics, clinical and laboratory features
Demographical data included clinical AID diagnosis, gender, self-reported ethnicity, median age at AID diagnosis and at colchicine treatment start. Furthermore, clinical symptoms and treatment at referral to the study center were analyzed. Flare frequency was defined in categories 0 to 4 with 0 = no flares, 1= > every six weeks, 2 = every five to six weeks, 3 = every three to four weeks and 4 = every one to two weeks; flare duration in days was defined in categories 0 to 3 with 0 = no flares, 1 = one to two days, 2 = three to four days and 3 = five or more days; fever during flares was defined as a body temperature ≥ 38°Celsius measured rectally, orally, axillary or at the ear. AID symptoms were captured using a symptom diary similar to the Autoinflammatory Disease Activity Index (AIDAI) [
18]. At each visit, a complete physical examination and evaluation of possible AID complications were performed. The inflammatory markers C-reactive protein (CRP), serum amyloid A (SAA) and whole blood count, liver enzymes and kidney function tests were performed at each visit. Urine for proteinuria was analysed when urine sampling was possible at visit. The study defined three distinct time points: 1) baseline defined as time of colchicine treatment start, 2) first follow-up as visit after colchicine start and 3) last follow-up as last study visit.
Colchicine therapy
Colchicine treatment was started at 0.5 to 1.0 mg/day. The therapeutic effect was monitored at three to six month visits. Based on disease activity and side effects colchicine dose was adjusted by 0.5 mg/day steps. During disease flares co-medication with non-steroidal anti-rheumatic drugs (NSAID) was allowed. Colchicine safety monitoring included gastrointestinal symptoms, whole blood count, liver and renal function. Colchicine was stopped if adverse events or intolerance occurred at the minimal dose of 0.5 mg/day. If the maximum tolerated colchicine dose did not result in improvement of disease activity anti-interleukin-1(IL-1) treatment was indicated.
Definitions of disease activity and response
Disease activity was defined as physician global assessment (PGA) recorded on a 10 cm visual analogue scale (VAS) with 0 representing no disease activity and 10 maximum disease activity. Disease activity was categorized as mild (PGA ≤2), moderate (PGA > 2 ≤ 5) and high (PGA > 5). Laboratory evidence of disease activity included CRP > 0.5 mg/dL and/or SAA > 10 mg/L.
Colchicine response was defined as complete, partial and no response based on clinical disease activity and inflammatory markers. Response criteria were selected based on previous studies [
19‐
22]. Complete response (CR) was defined as PGA ≤2 plus CRP ≤0.5 mg/dL and/or SAA ≤10 mg/L; partial response (PR) as PGA > 2 ≤ 5 plus CRP > 0.5 mg/dL ≤5 mg/dL and/or SAA > 10 mg/L ≤ 50 mg/L and no response (NR) as PGA > 5 and/or CRP > 5 mg/dL and/or SAA > 50 mg/L.
Outcome
The primary outcome was CR to colchicine at last follow-up. Secondary outcomes included colchicine responses: 1) PR and NR at last follow-up, 2) CR, PR and NR at first follow-up; flare responses: 3) flare frequency, 4) flare duration, 5) fever during flares; colchicine impact of flares: 6) overall flare improvement defined as the composite of reduced flare frequency and shortening of flare duration and lower/no fever during flares, 7) any impact on flares defined as either reduced flare frequency or shortening of flare duration or lower/no fever during flare, 8) requirement for anti-IL-1 treatment and 9) colchicine safety.
Analysis
Baseline demographics were analysed using descriptive statistics; median values and ranges, mean values and interquartile ranges were computed. Comparative analyses were conducted using parametric and nonparametric methods as appropriate. R (version 3.5.1; R Development Core Team, Vienna, Austria, (
http://r-project.org)) was used for data analysis (CRP, PGA, SAA) and visual graphics.
Discussion
This is the first study to systematically evaluate colchicine monotherapy for children with a clinical diagnosis of AID without pathogenic gene variants including children with a clinical diagnosis of FMF, CAPS, PFAPA and those with clinically unclassified AID. Colchicine was found to be effective and safe. At baseline, children displayed moderate disease activity (PGA 4) and elevated inflammatory markers (CRP 12.1 mg/dL; SAA 289.2 mg/L); frequent, severe febrile flares were documented in 97% of the patients. More than half of the patients experienced CR to colchicine at last follow-up including minimal to no disease activity and normalized inflammatory markers. An additional 35% of children had a PR. Overall, the study documented a significant decrease in disease activity (p < 0.0001) and decrease of inflammatory markers (SAA: p < 0.0001, CRP: p < 0.005) due to colchicine. In 93% of patients flares improved and 58% reported no febrile flares at last follow-up. These data suggest that colchicine should be considered for children with a clinical diagnosis of AID in the absence of pathogenic gene variants.
Colchicine was effective across different AID in reducing clinical and laboratory disease activity. Overall, high response rates were documented for FMF (100%), CAPS (88%), PFAPA (85%) and unclassified AID (100%).
In this pilot study, colchicine controlled disease activity in PFAPA patients and had an excellent effect on flare prevention. A total of 83% reported a positive impact of flare characteristics at last follow-up. Importantly, 75% of PFAPA patients reported less frequent flares. In 2008, Tasher et al. reported that eight out of nine PFAPA patients (89%) treated with colchicine had a significantly decreased flare frequency [
23]. Similarly, Butbul Aviel et al. observed significantly reduced flares in eight children with PFAPA treated with colchicine compared to pre-treatment and to ten untreated controls [
24]. These findings support our results, highlighting that colchicine is particularly effective in preventing PFAPA flares. However, more studies including larger cohorts of PFAPA patients are needed, to confirm these findings.
Colchicine was effective in children with CAPS and low-penetrance
NLRP3 variants. All included patients were symptomatic requiring treatment. Patients with low-penetrance
NLRP3 variants were previously shown to be at low risk for severe organ damage [
5,
25,
26]. Overall, 88% of CAPS patients showed a colchicine response (four CR, three PR). Importantly, all patients (100%) reported a positive impact on disease flares resulting in less frequent flares and shorted flare duration. To date, colchicine therapy has not been systematically studied for CAPS and low-penetrance variants. A recent case report documented colchicine effectiveness in an elderly women with cold-induced urticarial-like rush and a pathogenic
NLRP3 variant (A439V) and an additional
MEFV variant (E148Q) [
27]. Among 94 CAPS patients analysed in the Eurofever Registry none was treated with colchicine [
28]. To date, treatment recommendations include anti-IL-1 treatment for the whole spectrum of CAPS [
6]. The results of this pilot study suggest that colchicine should be considered in the proposed CAPS treat-to-target approach, which is anchored in the presenting disease phenotype, disease activity and risk of organ damage [
29]. CAPS patients with mild to moderate disease activity and low risk of organ damage may benefit from colchicine monotherapy as first-line agent. In addition, the low-cost colchicine therapy may provide an important option for low-income countries.
Colchicine was effective in patients with unclassified AID. All patients reported a positive impact on flares, there were no non-responders. Similarly, Papa et al., reported a positive colchicine effect in 78% variant-negative patients with undifferentiated recurrent fevers [
10]. In addition, Chandrakasan et al. demonstrated a positive effect of colchicine in 15 patients with clinical periodic fever syndromes and negative or inconclusive genetics [
11].
Not surprisingly, colchicine was effective in the studied children with FMF. All pathogenic gene variant-negative FMF patients showed colchicine response (63% CR, 37% PR). A positive effect on flares was reported by 100%. Colchicine is an established first-line therapy in FMF, preventing and aborting flares effectively [
7,
30,
31]. Among 121 FMF patients analysed in the Eurofever Registry 79% had two
MEFV variants, 17% had one
MEFV variant, three patients had no variants and two were not tested [
28]. All received colchicine and 62% achieved CR and 36% PR whereas two failed to respond [
28]. Colchicine effectiveness in controlling disease activity is observed in variant-negative and variant-positive FMF patients [
32]. These findings support our data that variant-negative FMF patients have good colchicine response rates.
Colchicine doses of 0.5 to 1 mg/day were effective and safe in the majority of children with clinically diagnosed AID without pathogenic gene variants. A total of 91% of the patients were commenced on 0.5 mg and 9% on 1 mg of colchicine daily. At first follow-up a significant reduction of inflammatory markers and disease activity was documented. Dose adjustments were made to optimize effectiveness and safety. At last follow-up, 44% of patients received 0.5 mg/day, 31% 1 mg/day and 25% 1.5 mg/day; in five children a dose reduction from 1 to 0.5 mg/day due to gastrointestinal side effects was required.
All PFAPA patients in this study cohort were commenced on 0.5 mg/day colchicine at baseline. At last follow-up, more than half of children had required dose adjustment with 31% of patients receiving 1 mg/day and 23% 1.5 mg/day. Recommended daily colchicine dose ranges between 0.5 to 1 mg for flare prevention in PFAPA and should be considered in patients, in whom corticosteroids have resulted in an increased flare frequency [
33,
34]. Recently, the CARRA PFAPA working group has published consensus treatment plans, which should be evaluated in future pilot studies [
35]. For the prophylaxis arm colchicine at 0.5 to 1.25 mg/day was recommended [
35].
The majority of CAPS patients in this pilot study was started on colchicine 0.5 mg/day, only one received 1 mg/day. At last follow-up, 38% were treated with 0.5 mg, 38% with 1 mg and 24% with 1.5 mg/day. One child with unclassified AID was started on 0.5 mg/day of colchicine the other on 1 mg/day. At last follow-up both received 0.5 mg/day of colchicine. Currently, there is a complete lack of colchicine dosing recommendations for CAPS and for unclassified AID.
Most FMF patients were started on 0.5 mg/day of colchicine; one received 1 mg/day. At last follow-up, a third remained on 0.5 mg/day, another third received 1 mg/day and 1.5 mg/day respectively; none required dose increase to 2 mg/day. FMF treatment recommendations suggest age adjusted colchicine dosing regimens: for children < 5 years a starting dose of ≤0.5 mg/day (or ≤ 0.6 mg/day when the only available tablets contain 0.6 mg) and for 5–10 years starting doses of 0.5–1.0 mg/day (or 1.2 mg/day) [
7,
30]. Children > 10 years and adults should be started at colchicine doses of 1.0–1.5 mg/day (or 1.8 mg/day) [
7,
30]. The maximum daily colchicine dose should not exceed 2 mg in children and 3 mg in adults [
7,
30]. An estimated 5 to 10% of colchicine treated patients may experience side effects such as diarrhoea, vomiting and nausea [
36]. Dose adjustments may be necessary to reduce individual toxicity. Furthermore, dosing requirements may be related to the genotype [
32,
37]. Ben-Zvi et al. suggested that control of disease activity in gene variant-negative FMF patients can be achieved with significantly lower colchicine doses compared to homozygous variant-positive FMF patients (M694V) [
32]. Knieper et al. showed that there were no significant colchicine average dose differences between homozygous and compound heterozygous variant-positive FMF patients (M694V; M680I; M694V/M680I; M694V/V726A) [
37]. In contrast, patients with the M694V/E148Q genotype or any heterozygous variant had significantly lower average colchicine doses [
37]. Dose increases may be beneficial in variant-negative FMF patients and may result in reduced flare frequency, in particular when rather low colchicine doses are increased [
37].
In this study the starting dose of colchicine varied between 0.5 mg/day (91%) and 1 mg/day (9%) resulting in a significant decrease in disease activity and inflammatory markers at first follow-up. This data suggests to start colchicine in children with a clinical diagnosis of AID without pathogenic gene variants irrespective of age with 0.5 mg/day. Dose adjustments might be required over time.
The study has several limitations. The sample size was small. However, the study aimed to include a very specific group of children with clinically diagnosed AID without pathogenic gene variants. Tuebingen is a reference center particularly providing support for AID patients with high disease activity, comorbidities, need of biologic therapy and unclear phenotype. The reference center takes care for PFAPA patients with high disease activity, non-response to corticosteroids on demand, unclear phenotype or suspicion of an overlap to another AID. Therefore, the PFAPA patient cohort is rather small compared to other AID at the centre. Furthermore, a comprehensive clinical and genetic work-up, including advanced genetic and functional testing helps to make a clinical and genetic AID diagnosis in several previously undiagnosed children with AID characteristics. Consequently, only few patients fulfilled the very strict inclusion criteria of no ongoing/previous biologic therapy, no organ damage, and no genetically confirmed AID diagnosis. Moreover, there was missing data as a real life cohort was studied. However, standardized assessment of the included patients by using the AIDAI and advanced laboratory testing combined with standardized outcome evaluation resulted in comparable high-quality data captured in the ARDIS.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.