To the Editor:
We read the article by Saccomanno, et al, “Predictors of effectiveness of anakinra in systemic juvenile idiopathic arthritis,” with great interest1. They reported that shorter disease duration, fewer active joints, higher ferritin levels, and greater activity of systemic manifestations were independently correlated with achievement of complete clinical response at 1 year (CCR1) in systemic juvenile idiopathic arthritis (sJIA). They proposed a model with 4 variables (disease duration ≤ 3.9 yrs, active joint count ≤ 10, ferritin > 444 ng/ml, systemic manifestation score > 3) to predict response to anakinra in sJIA1.
We have analyzed these variables and tested this model in our cohort of patients with sJIA. The patients (0–18 yrs old) treated with anakinra at Hacettepe University Pediatric Rheumatology Department between January 2006 and January 2018 were included. The patients were classified with sJIA according to the International League of Associations for Rheumatology criteria2. Demographic data, clinical manifestations, laboratory findings, and CCR1 were documented. The study was approved by the ethics committee of Hacettepe University (GO 16/154) and written consent from the patients/parents was obtained according to the Declaration of Helsinki.
These factors were the same as those stated in the article by Saccomanno, et al1: the definition for CCR (absence of fever, negative C-reactive protein, physician’s global assessment ≤ 1, active joint count ≤ 1, and ≥ 75% reduction of corticosteroid dose from baseline), systemic manifestation score, and the inclusion/exclusion criteria for patients to keep the results comparable. The normality of distribution of numerical variables was investigated using visual (histogram, probability plots) and analytic methods (Shapiro-Wilk’s test), and descriptive analyses were presented using medians, minimum, and maximum values. Chi-square or Fisher’s exact test was used to compare categorical variables, and Mann-Whitney U test was used to compare the non-normally distributed variables between 2 groups. A p value of < 0.05 was considered a statistically significant result.
A total of 45 patients with sJIA were included; 28 (62.2%) were complete responders while 17 (37.8%) did not achieve CCR1 (non -responders). The characteristics of the patients are presented in Table 1.
The median active joint count was lower (0 vs 3; p = 0.005) while the ferritin level was higher (620.5 vs 162 ng/ml) and the disease duration was shorter (6 vs 24 mos) at initiation of anakinra treatment in complete responders compared to nonresponders. However, 2 latter differences were statistically insignificant. Significantly more patients were still receiving anakinra (85.7% vs 41.2%; p = 0.002) while fewer patients were taking corticosteroids at 1 year (3.6% vs 82.4%; p < 0.001) in the complete responder group compared to the nonresponders. Flares under anakinra treatment were more frequently observed among nonresponders than complete responders. The model proposed by Saccomanno, et al1 was tested in all patients (Table 2). Significantly higher numbers of complete responders than nonresponders met all 4 variables (67.9% vs 8.3%, respectively; p = 0.001). Of note, only 1 patient among the 12 nonresponders who had complete information for the 4 variables of the model met all variables in this model.
The rate of CCR1 in our sJIA cohort (62.2%) is in the high range among the rates (31–85%) reported in previous studies1,3,4,5,6,7. In our patients, the number of active joints at initiation of anakinra was significantly higher in complete responders compared to nonresponders, and the model proposed by Saccamanno, et al1 seems to be associated with higher rate of CCR1.
Previously reported potential predictors of antiinterleukin (IL)-1 response were older age at disease onset, shorter disease duration, less severe joint disease, elevated white blood cell and neutrophil counts, and anti-IL-1 agents being used as first-line treatment in patients with sJIA naive to disease-modifying antirheumatic drugs (DMARD) or corticosteroids4,5,7,8. In our cohort, there was no significant difference between complete responders and nonresponders regarding white blood cell and neutrophil counts and none of our patients were corticosteroid- or DMARD-naive at initiation of anakinra. There was a trend toward older age at disease onset, longer disease duration, and lower ferritin levels in nonresponders compared to complete responders (Table 1). However, these differences were not statistically significant, probably because of the small sample size.
Verifying and validating the factors predicting anakinra response in sJIA are critical for personalized medicine and choosing the right patient to benefit from anakinra.
REFERENCES
- 1.
- 2.
- 3.
- 4.
- 5.
- 6.
- 7.
- 8.