A total 78 children with JIA were eligible for the study, 52.6% females, mean ± SD age = 9.94 ± 4.92 years, and age at diagnosis = 7.44 ± 4.52 years. They were followed up for a mean ± SD [range] disease duration = 2.93 ± 2.70 [6 months; 15 years]. The most frequent types of JIA were systemic (33.3%), oligoarticular (30.8%) and polyarticular negative RF (26.9%). Other parameters included positive ANA in 41.0%, positive RF in 7.7% and uveitis in 9.0%. The most frequent treatment was methotrexate (64.1%), followed by biological therapy (47.4%) (Table 1).

SD: Standard deviation; RF+: positive rheumatoid factor; RF−: negative rheumatoid factor; NSAIDs: non-steroid anti-inflammatory drugs; ANA: antinuclear antibody.

Growth data were available for 67 (85.9%) children only. Analysis of weight, height and weight-for-height percentiles showed 46.2%, 46.2% and 39.7% breaks of the respective growth curves between time of diagnosis and last follow-up (Table 2). In all three growth parameters there were shifts towards lower percentiles from time of diagnosis to last follow-up, which was observed in both genders (Figures 1-3).

¹Growth pattern was assessed according to the change in percentile rank from time of diagnosis to last follow-up, which was classified into 3 categories: regression = drop to a lower percentile; stable = maintaining the same percentile; progression = change to a superior percentile.

Correlation of height-for-age growth curve with demographic and clinical factors showed that children who regressed in the percentile rank from T0 to T1 had lower age at diagnosis than those who did not regress (mean ± SD age = 6.69 ± 4.02 versus 8.54 ± 4.88 years, respectively; p = 0.093) and longer disease duration (3.31 ± 2.21 versus 1.97 ± 1.28 years; p = 0.004, respectively). Analysis of the other factors, such as gender, JIA sub-type, treatment or biological data showed no significant difference between the two groups.

Regarding weight-for-age, children who regressed had longer disease duration than those who did not regress (3.20 ± 2.10 versus 2.10 ± 1.57 years; p = 0.020, respectively). No difference was observed in other factors; except a higher proportion of prednisolone use among children who had growth retardation versus who had normal growth (68.0% versus 32.0%), however, this result was not statistically significant (p = 0.071) (Table 3).

Freq.: frequency, *: statistically significant result (p < 0.005); RF+: positive rheumatoid factor; RF−: negative rheumatoid factor; Exd. Oligarticular: extended olilgarticular; §: significance calculated using Fisher’s exact test.

Age at diagnosis and disease duration were analyzed as predictors for growth retardation severity in children with JIA, which was indicated by the number of percentile ranks lost from diagnosis to last follow-up. Regarding height-for-age, severity of growth retardation was predicted by low age at diagnosis and long disease duration in both univariate and multivariate models (Table 4), showing significant correlations with the number of percentile ranks lost from diagnosis to last follow-up (Figure 4(a), Figure 5(a)). Regarding weight-for-age, severity of growth retardation was only predicted by disease duration (Table 4; Figure 4(b) and Figure 5(b)).

Growth retardation and developmental abnormalities are common complications of JIA and are associated with significant impact on patient’s physical and psychological health and overall quality of life [25] [26] [27] . This retrospective study showed high incidence of growth retardation among the local patients. Almost 1 child in 2 had breaks in growth curve, in at least one of the three growth parameters including weight-for-age, height-for-age and weight-for- height. In other studies, pattern of growth varies according to the study population and methodology and to other associated risk factors. Some authors report 10% to 20% of growth retardation in children with severe forms of JIA [28] , while others reported up to 40% in all sub-types [21] . A prospective case-control study from India found no significant difference between JIA children and healthy controls. Authors compared weight, height, body mass index and growth velocity over 6 months of children with JIA versus healthy children [29] . This shows relatively high proportion of growth impairment among our study population, which points towards the existence of other probable risk factors.

Pathophysiology of growth retardation in children and adolescents with JIA is

*Statistically significant result (p < 0.05); OR: odds-ratio; 95% CI: 95% confidence interval for OR.

mainly related to excessive cytokine levels and their pro-inflammatory action; distinctive of severe forms of JIA [30] [31] . There is strong evidence indicating the existence of systemic and local modulating effect of cytokines (especially IL-6) on growth plate of long bones [30] [32] [33] . In addition, cytokines have indirect action involving insulin-like growth factor-I, which was observed to be reduced in serum of patients with systemic JIA and correlated to excessive production of IL-6 [27] . Number of other hormonal and metabolic factors such as parathyroid dysfunction, sex steroids and vitamin D metabolites contribute in growth retardation by modulating growth hormone/insulin-like growth factor-I axis, resulting in impaired bone growth [8] . Growth retardation may also be induced by prolonged use of corticosteroids, especially when initiated in the young age [8] [27] . Other pathophysiological mechanisms include other cytokine- induced epigenetic changes [34] and malnutrition [35] .

Age at diagnosis and disease duration was the only significant factors correlated with growth retardation. Average age at diagnosis of the population was 7.44 ± 4.52 years, which was in line with other studies [29] [36] Both the occurrence and severity of height retardation were predicted by the low age at diagnosis of the patient; while the no effect was observed on weight growth. This is generally supported by literature; and may be explained by the level of growth hormone secretion in the childhood, which is physiologically lower than in adolescent, resulting in greater impact of the disease activity (cytokines) on insulin-like growth factors-I secretion in the young; and consequently a delayed long bone growth [8] [37] .

We demonstrated that disease duration was a significant predictor for the occurrence of growth retardation in both height and weight. In addition, severity of growth retardation, as indicated by the number of percentile ranks lost, was linearly correlated with the number of years of disease duration. In other words, children with long disease duration represent the most frequent and most severe cases of growth retardation among children with JIA. These observations are concordant with data from literature showing greater growth delay in children with long disease duration; which is a common point between all children chronic inflammatory diseases [8] [31] [38] .

No difference was found between different JIA sub-types in any of the analyzed growth parameters; whereas Mondal et al. reported greater impact on growth velocity in children with polyarticular RF+ form, while children with systemic JIA had the greatest impact on height and weight by comparison to those with other sub-types [29] . Similarly, Okumus et al. reported significantly smaller height in systemic JIA [39] . Severe cases of growth retardation are generally reported in polyarticluar forms with multiple joint involvement or systemic forms with extensive damage [21] .

This study demonstrated no significant effect of treatments on growth, be it positive of negative. This may be explained by a relatively short disease duration (2.93 ± 2.70 years), which may be insufficient to observe an effect of treatments. The use of corticosteroids has been demonstrated to induce or exacerbate growth retardation, in a duration- and dose-dependent manner [21] . Furthermore, children treated by corticosteroids in a young age are highly exposed to delayed puberty than those on other treatment regimen [40] . On the other hand, biological therapy, such as anti-TNF-α and anti-IL-6 have been shown to restore growth, both by reducing disease activity and limiting the use of corticosteroids [41] . Another study demonstrated a strong growth-restoring effect of anti- TNF-α in children with polyarticluar JIA, which was correlated to the decrease in disease activity, independently from corticosteroids effect [42] . However, in systemic JIA, growth restoring effect of biologic treatments seems to be less remarkable [43] . Analysis of these observations, among other therapeutic outcomes, justified the current trend of early initiation of aggressive treatment, using several combination of different agents to improve allover disease outcomes including growth [44] . On the other hand, the use of growth hormone has shown good results in restoring growth of JIA patients and should be considered in the management of these patients [45] .

None of the other disease-related parameters including ANA, RF and uveitis were significantly associated with growth retardation.

The major limitation of this study was a small sample size, which limited the power of sub-groups analysis; in addition to growth data being missing in number of files, which further reduced the sample size. One other notable limitation was the short follow-up duration of the patients, which prevented from observing significant effects of treatments and other factors on growth. In addition, other growth parameters, such as sexual maturation and bone density were not assessed in this study, both showing to be impaired in JIA in other studies [40] [46] .

Despite these limitations, this study provided a sound epidemiological picture of growth retardation among children with JIA and highlighted the importance of systemic and careful assessment of growth parameters in these children.

Future prospective, multicenter studies are warranted to provide a more accurate picture of the growth pattern among JIA children in Saudi Arabia; and investigate further population-specific risk factors associated with this high prevalence. Such study should also assess other parameters including sexual maturation, bone density and local growth abnormalities.

Preventing growth impairment and restoring growth velocity should be among priorities of therapeutic goals. It is achieved through effective decrease of time and severity of disease flare-ups and enhancement of remissions. This requires appropriate use of pharmacological treatments along with systemic, close monitoring of physical development and interdisciplinary management involving pediatricians, rheumatologists and clinical anthropologists [21] .