Rheumatoid arthritis and growth retardation in children: treatment with human growth hormone

Growth and the Growth Hormone-Insulin Like Growth Factor 1 Axis in Children With Chronic Inflammation: Current Evidence, Gaps in Knowledge, and Future Directions

Growth failure is frequently encountered in children with chronic inflammatory conditions like juvenile idiopathic arthritis, inflammatory bowel disease, and cystic fibrosis. Delayed puberty and attenuated pubertal growth spurt are often seen during adolescence. The underlying inflammatory state mediated by proinflammatory cytokines, prolonged use of glucocorticoid, and suboptimal nutrition contribute to growth failure and pubertal abnormalities. These factors can impair growth by their effects on the GH-IGF axis and also directly at the level of the growth plate via alterations in chondrogenesis and local growth factor signaling. Recent studies on the impact of cytokines and glucocorticoid on the growth plate further advanced our understanding of growth failure in chronic disease and provided a biological rationale of growth promotion. Targeting cytokines using biological therapy may lead to improvement of growth in some of these children, but approximately one-third continue to grow slowly. There is increasing evidence that the use of relatively high-dose recombinant human GH may lead to partial catch-up growth in chronic inflammatory conditions, although long-term follow-up data are currently limited. In this review, we comprehensively review the growth abnormalities in children with juvenile idiopathic arthritis, inflammatory bowel disease, and cystic fibrosis, systemic abnormalities of the GH-IGF axis, and growth plate perturbations. We also systematically reviewed all the current published studies of recombinant human GH in these conditions and discussed the role of recombinant human IGF-1.

Insulin-like growth factor-I levels predict weight, height and protein catabolism in children and adolescents with cystic fibrosis

Multiple reports have demonstrated the benefit of growth hormone (GH) treatment in children with cystic fibrosis (CF) and previous studies have demonstrated low to normal insulin-like growth factor-I (IGF-I) levels in these patients. Most biological effects of GH are mediated by IGF-I; however, the relationship between height, weight and rate of growth has not been systematically studied in CF. We conducted a retrospective analysis of 52 patients (including control volunteers with CF) who had participated in previous studies of GH treatment to determine the relationship between levels of IGF-I and growth in children with CF. In a subset of these patients, we also evaluated the relationship between protein catabolism and IGF-I. Baseline IGF-I levels and IGF-I z-scores were correlated with same day measures of height, weight, height and weight z-scores. In a subset of patients, IGF-I levels were also correlated with leucine rate of appearance (a measure of protein catabolism). IGF-I levels were obtained every six months during our studies and were correlated with same day height, weight and protein turnover. Height and weight velocity were calculated every six months from study baseline and were correlated with IGF-I levels. In all patients, whether treated with GH or controls, we found a positive linear correlation between IGF-I levels and height (r = 0.66, p < 0.0001) and weight (r = 0.61, p < 0.0001), as well as height and weight velocity. There was also a strong relationship between leucine rate of appearance and IGF-I. These results suggest a strong correlation between IGF-I and height, weight and protein catabolism and emphasize the need to normalize IGF-I levels in children with cystic fibrosis.

Growth hormone treatment in children with rheumatic disease, corticosteroid induced growth retardation, and osteopenia

BACKGROUND

In children with severe rheumatic disease (RD), treatment with corticosteroids (CS) is frequently needed and growth retardation and osteopenia may develop. A beneficial effect of human growth hormone (hGH) has been reported but mostly in trials without a control group.

AIMS

To study the effect of hGH on growth, bone mineral density (BMD), and body composition, taking the disease activity and CS use into account.

METHODS

Randomised controlled trial on 17 prepubertal RD patients with growth retardation and/or decreased BMD. The hGH group (n = 10) received treatment with hGH 4 IU/m2/day (approximately 0.045 mg/kg/day) during two years. The controls (n = 7) received no GH treatment.

RESULTS

During the two year study period the disease activity, and use of CS and methotrexate (MTX) did not differ between the groups. There was a significant mean increase in height standard deviation score (HSDS) in the hGH group (0.42+/-0.16 SDS) and a non-significant decrease in the controls (-0.18+/-0.11 SDS). Change in BMD did not differ significantly between the groups, although the increase in BMD for lumbar spine within the hGH group was significant. Lean body mass improved significantly in the hGH group compared to controls (0.64+/-0.19 SDS versus -0.20+/-0.17 SDS), while the decrease in percentage fat was not significant.

CONCLUSIONS

There was a significant effect of hGH on growth and lean body mass, but a longer duration of treatment might be necessary to evaluate the effect of hGH on BMD.

Reduced bone mineral density and radial bone growth in young rabbits treated with dexamethasone eye drops

OBJECTIVE

To investigate the effect of dexamethasone eye drops on bone metabolism in newborn rabbits.

METHODS

Thirty-four 3-week-old rabbits had unilateral clear lens extraction and were randomized into three groups. Postoperatively, group 1 received high-dose and group 2 low-dose dexamethasone eye drops (average doses 0.27 and 0.10 mg/kg body weight/day, respectively). These rabbits also received a postoperative subconjunctival injection of betamethasone. Group 3 (control) received vehicle eye drops only. After 8 weeks of treatment, all animals were killed and the left femurs were isolated and subjected to peripheral quantitative computerized tomography (pQCT) and dual X-ray absorptiometry (DXA) analyses.

RESULTS

DXA showed that rabbits treated with either a high or low dose of dexamethasone eye drops had significantly reduced areal bone mineral density (BMD), area and total bone mineral content (BMC) of the femur. Measurements with pQCT demonstrated a dose-dependent reduction in cortical BMC, cortical volumetric BMD and cortical area. These effects were associated with an inhibition of radial femur growth, cortical thickness and periosteal and endosteal circumferences.

CONCLUSION

Dexamethasone eye drops have systemic effects affecting several bone parameters in young rabbits. Any long-term systemic effects of ocular glucocorticoids need to be further studied.

Growth Hormone/Insulin-Like Growth Factor–I System and Connective Tissues: Basic Aspects and Clinical Implications

This article reviews basic and clinical aspects of the effects of growth hormone and insulin-like growth factor-I on the connective tissues.

GH therapy in juvenile chronic arthritis: results of a two-year controlled study on growth and bone

Disturbance of growth frequently occurs in children suffering from juvenile chronic arthritis (JCA). Recognition of growth impairment is important because reduced final height is one of the permanent consequences. The aim of this study was to evaluate the efficacy and safety of human GH (hGH) in growth-retarded prepubertal children with JCA. Thirty-five children were tested for GH deficiency (GHD) and randomly assigned to a study and an untreated control group; five were GH deficient and were part of the GHD group. All received glucocorticoids. The study group was treated with 1 IU/kg BW.wk hGH; the GHD group was given 0.5 IU. During 2 yr of hGH treatment growth velocity and height SD score increased compared with baseline values. There was a marked increase in growth velocity in the treated groups, but also some increase in the control group. Plasma levels of IGF-I and IGF-binding protein-3 increased with GH treatment. These results suggest that hGH might be useful in the treatment of growth impairment in JCA. GH may counteract the adverse effects of glucocorticoid therapy, but its effect is dependent on the disease activity. Long-term controlled studies are needed to determine the risks and benefits of GH therapy in JCA.

The influence of the hormonal system on pediatric rheumatic diseases

The role of NEI interactions in children with chronic inflammatory rheumatic disorders has not been systematically studied to the same extent as in adults. The data that are currently available suggest that NEI mechanisms are involved in the pathophysiology of some of the diseases. These include JRA, JSLE, and JAS. Prolactin has been most extensively investigated, showing interesting parallels with findings in adult rheumatologic diseases. Limited data on cortisol suggest a deficiency of production in JRA, a situation similar to that in adult RA. These findings suggest that there is a proinflammatory hormonal bias in children with JRA and JSLE. The data in children with chronic autoimmune inflammatory disorders seem to be identical to those seen in adults with RA and SLE, but a clear delineation of the role of the neuroendocrine-immune system in disease pathophysiology is still required. The neuroendocrine aspects of pediatric rheumatologic disease observed to date suggest a number of avenues for further research in the field of neuroendocrine immunology, which may open up novel therapeutic options.

Efficacy of recombinant human growth hormone in children with juvenile rheumatoid arthritis and growth failure

Ten patients with juvenile rheumatoid arthritis (JRA) and growth failure were treated with recombinant human growth hormone (GH) for 1 to 3 years at a dosage of 0.57 IU/kg/wk. All the patients had been on prednisone at a mean dosage of 4.12 mg p.o. daily. GH was low in one patient, two patients had a borderline level and seven patients had adequate response to provocative tests or post-sleep measurement. Serum IGF-I was found to be low in five of six patients. Mean growth velocity increased from 2.45 cm/yr to 4.79 cm/yr after 1 year's treatment with GH (P<0.004). Six patients continued on GH treatment for a second year and continued to have a better growth velocity, with a mean of 5.43 cm/yr (P<0.014). Two patients entered puberty during the second year of GH treatment. This study demonstrates the potential beneficial effect of GH treatment in patients with JRA with growth failure of systemic onset or polyarticular onset who are on prednisone. Further study is needed to determine the long-term effect of GH treatment on ultimate height.

Neuroendocrine immune features of pediatric inflammatory rheumatic diseases

Juvenile rheumatoid arthritis (JRA) and juvenile systemic lupus erythematosus (JSLE) are the most common autoimmune rheumatic diseases in children associated with high levels of autoantibodies and immune reactivity. JRA and JSLE are more common in girls. Disease activity is worse in the morning, improves during the daytime and worsens at night suggesting that neuroendocrine immune mechanisms are involved in disease pathophysiology. Adult patients with RA and SLE have excessive levels of prolactin (PL) while cortisol (CS) production is down-regulated for the degree of ongoing inflammation. PL has potent proinflammatory properties. Normal to low levels of cortisol have been observed in children with active JRA despite the high serum levels of IL-6, IL-1 beta, and TNF-alpha, which activate the hypothalamic-pituitary-adrenal axis (HPA). The CS levels are in fact subnormal because inflammatory stress activates the HPA. Normal serum PL levels were seen in children with JRA, most of whom were not active with higher levels in those with active ANA +ve JRA complicated by uveitis. A trend toward high PL levels was seen in 33 children with JSLE. High serum PL levels are seen in patients with active juvenile ankylosing spondylitis (JAS) only. Growth retardation is a feature of JRA. Patients with JRA have low to normal levels of growth hormone (GH) and low levels of insulin-like growth factor 1 (IGF-1). IGF-1 mediates the effects of GH. The observation of low IGF-1 in JRA raises the therapeutic possibility with IGF-1. Overall, high levels of follicle stimulating hormone and luteinizing hormone are found in children with JSLE while the levels in JRA tend to be normal. Testosterone levels are low in patients with JRA. No significant differences in estrogen levels have been found between patients with JRA and those with JSLE and matched controls. There is evidence that the autonomic nervous function is defective in patients with JRA.

Growth hormone treatment of short stature and metabolic dysfunction in juvenile chronic arthritis

Severe growth retardation and profoundly altered body composition are observed in children treated with glucocorticoids for systemic forms of juvenile chronic arthritis. The aim of this study was to assess the effects of 1 year of treatment with growth hormone (GH) on height velocity and body composition, and the potential effects of such treatment on glucose tolerance. Fourteen children receiving steroid therapy for juvenile chronic arthritis were treated with GH, 1.4 IU/kg/week (0.47 mg/kg/week), for 1 year and were then followed up for 1 year after cessation of treatment. Baseline GH secretion and plasma levels of insulin-like growth factor I (IGF-I) and IGF-binding protein-3 (IGFBP-3) were at the lower limit of normal, but increased with GH treatment. During the year of GH treatment, mean height velocity increased from 1.9 to 5.4 cm/year (p < 0.001), mean lean body mass increased by 12% (p < 0.01) and mean fat mass fell by 20% (p < 0.01) compared with baseline. Decreased glucose tolerance and increased glycosylated haemoglobin levels were also observed. This may have been a result of insulin resistance, as indicated by induced hyperinsulinaemia. Following the 1-year GH treatment period, height velocity fell to pretreatment values, and the height SDS at the end of the second year was lower (p < 0.01) than before treatment. Weight and fat mass increased markedly after cessation of GH treatment. These results suggest that GH may partially counteract the adverse effects of glucocorticoids on growth and metabolism in patients with chronic inflammatory disease. Long-term controlled studies are needed to determine the risks and benefits of GH therapy in such patients.

Dexamethasone impairs growth hormone (GH)-stimulated growth by suppression of local insulin-like growth factor (IGF)-I production and expression of GH- and IGF-I-receptor in cultured rat chondrocytes

Growth depression as a side effect of glucocorticoid therapy in childhood is partially mediated by alterations of the somatotropic hormone axis. The mechanisms of interaction between glucocorticoids and somatotropic hormones on the cellular and molecular level are poorly understood. In an experimental model of primary cultured rat growth plate chondrocytes, basal as well as GH (40 ng/ml) or insulin-like growth factor (IGF)-I (60 ng/ml)-stimulated growth was suppressed dose dependently (10(-l2)-10(-7)M) by dexamethasone (Dexa). An IGF-I antibody specifically and dose dependently inhibited the GH- but not the basic fibroblast growth factor (bFGF)-stimulated cell proliferation. GH increased the IGF-I concentration in conditioned serum-free culture medium; this was reversed by concomitant Dexa. Dexa time dependently suppressed the transcription of GH receptor (GHR) messenger RNA (mRNA) and down-regulated the basal and GH-stimulated expression of GHR. Whereas no suppressive effect on basal type I IGF-receptor (IGFR) was observed, Dexa blocked the IGF-I induced increase of IGF binding. These results were confirmed by GHR and IGFR immunostaining. We conclude that Dexa impairs the GH-induced stimulation of local secretion and paracrine action of IGF-I and reduces the homologous increase of IGFR and GHR expression. The above experiments give further insight on the interaction between GH and glucocorticoids on the cellular and molecular level of growth plate chondrocytes.

Juvenile rheumatoid arthritis. Effects of disease activity and recombinant human growth hormone on insulin-like growth factor 1, insulin-like growth factor binding proteins 1 and 3, and osteocalcin

OBJECTIVE

To investigate possible mechanisms of growth impairment in children with juvenile rheumatoid arthritis (JRA).

METHODS

Eighteen prepubertal children with JRA and growth retardation received recombinant human growth hormone (rHuGH) for 1 year. Growth hormone profiles over 24 hours were obtained before treatment in 12 patients; the levels did not differ from those in "short normal" children. Levels of insulin-like growth factor 1 (IGF-1), IGF binding proteins (IGFBPs) 1 and 3, insulin, osteocalcin, and C-reactive protein (CRP), as well as the erythrocyte sedimentation rate were measured serially. Pretreatment levels were compared with control levels.

RESULTS

In JRA patients, IGF-1, IGFBP-3, and osteocalcin levels were significantly lower and insulin levels significantly higher than those in controls, but there was no significant difference in the level of IGFBP-1. With rHuGH treatment, height velocity and mean levels of IGF-1, osteocalcin, and insulin increased significantly, while mean levels of IGFBP-1 fell significantly. Levels of IGFBP-3 correlated with those of IGF-1. The height velocity correlated positively with IGF-1 and osteocalcin, and negatively with IGFBP-1. Levels of IGFBP-1 were inversely related to those of insulin and IGF-1. There was a significant negative correlation between the CRP and height velocity, IGF-1 level, and osteocalcin level.

CONCLUSION

IGF-1 production is impaired in children with active JRA. Treatment with a therapeutic dose of rHuGH can rectify the IGF-1 deficiency within 4 days, but its effect is adversely influenced by the acute-phase response, as reflected by an elevated CRP level.

Endocrine aspects of paediatric rheumatic diseases

Compared to the now numerous studies on the endocrinology of rheumatic diseases in adults, only a small number of studies has been published on children with rheumatic diseases. Prolactin has been most extensively investigated, showing interesting parallels with the findings in adults with rheumatological diseases. Thus, analogous to adult RA most forms of JRA or JCA (with the exception of ANA-positive JRA with uveitis) appear to show, if anything, low to normal levels of prolactin. Since the prolactin levels in adult RA depend on the inflammatory activity, and the physiological prolactin secretion decreases in chronic stress (especially sleep disorders), these results are most likely to be explained as reactive non-specific mechanisms in the stress of the disease. However, specific mechanisms are also being discussed to explain the low prolactin levels in adult RA. The results of prolactin measurements in juvenile SLE, juvenile ankylosing spondylitis and ANA-positive JRA with a raised incidence of uveitis, contrast with this. These conditions sometimes show significantly higher prolactin levels compared to healthy controls. A correlation of the increase of prolactin concentration with the inflammatory activity has been described for juvenile ankylosing spondylitis. These results correlate well with those of adult forms such as diseases of the seronegative spondyloarthropathies type, SLE and iridocyclitis. Raised prolactin concentrations are also found in these diseases. The inflammation promoting and immunostimulatory effects of prolactin found especially in animal experiments are confirmed clinically in these diseases by reports of successful treatments with the prolactin inhibitor, bromocriptine. The results available up to now for human growth hormone in JRA and JCA tend to be comparable with the results for prolactin in these form of paediatric rheumatological diseases. Besides normal values above, all lowered concentrations are measured for this hormone. Apart from other non-specific factors, its diminished secretion is mainly determined by the inflammatory activity of the disease. Low levels of growth hormone are likely to be a significant factor in the growth retardation in children with inflammatory rheumatological diseases. Up to now, the small number of investigations on gonadotrophins and the sex hormones in juvenile SLE and various forms of JRA published have not as yet yielded unequivocal results. The endocrine aspects of paediatric rheumatological diseases are thus still incompletely elucidated. However, there are many promising avenues for further fruitful research in this field.

Skeletal response to recombinant human growth hormone (rhGH) in children treated with long-term corticosteroids

Corticosteroid therapy causes osteopenia and growth retardation in children; such changes are associated with diminished rates of bone formation and turnover. Since growth hormone activates bone remodeling, the biochemical and skeletal responses to rhGH were evaluated in four pediatric patients, aged 12.8 +/- 3 years, with long-term corticosteroid use (5 +/- 2 years). Recombinant human growth hormone (rhGH), 0.125 mg/kg, was given 3 times/week by subcutaneous injection for 12 months. Iliac crest bone biopsies were obtained after double tetracycline labeling before and at the end of rhGH therapy; serum levels of calcium, phosphorus, alkaline phosphatase, parathyroid hormone (intact), 25-hydroxyvitamin D, 1,25-dihydroxyvitamin D3, osteocalcin (BGP), and insulin-like growth factor-1 (IGF-1) were measured every 3 months during the treatment period. The average dose of prednisone was 0.24 +/- 0.05 mg/kg/day initially, and this did not change during the study. Serum calcium, phosphorus, alkaline phosphatase, 25-hydroxyvitamin D, 1,25-dihydroxyvitamin D3, and BGP were unchanged during the rhGH therapy, but the serum IGF-1 level increased by 71%, p < 0.01. Eroded bone perimeter and cancellous bone area did not change significantly during rhGH therapy. Bone formation rates rose from 423 +/- 475 to 781 +/- 407 microns2/mm2/day, p < 0.05, and the length of double tetracycline-labeled bone perimeter increased by 85%, p < 0.05. The bone formation rate in the growth hormone group exceeded the values of an age-matched reference group (14.3 +/- 3 years), 780 +/- 407 microns2/mm2/day versus 411 +/- 479 microns2/mm2/day, p < 0.05.(ABSTRACT TRUNCATED AT 250 WORDS)

Conventional and nonconventional uses of growth hormone

Although GH has been available as a therapeutic agent for the GH-deficient child for more than 30 years, the conditions of its use have yet to be optimized. The availability of biosynthetic material has provided researchers with the opportunity to develop the protocols necessary to begin to finally answer the most fundamental questions pertaining to dose, frequency, and duration of treatment. It has also permitted the initiation of prospective trials in a large number of conditions that result in childhood short stature, with the expectation that some or many of them will be treated effectively and safely. Finally, it has opened the door to an entire spectrum of potentially new uses of GH and other growth factors for so-called nonconventional indications. That these have implications that range from the short-term rapid healing of a burn graft site, to the more efficient induction of ovulation, to the long-term preservation of lean body mass has excited the interest of investigators in many fields of medicine and physiology. Thus, the recent progress reported in this paper is really the beginning of the new research that will take place with GH and growth factors.

Insulin-like growth factor and growth hormone secretion in juvenile chronic arthritis

Insulin-like growth factor (IGF-1) concentrations were determined in a series of 23 children with juvenile chronic arthritis in conjunction with anthropometric assessment. When standardised z scores were used significant decreases in height and weight were shown in comparison with the normal age/sex matched means. Severe growth disturbance was seen, particularly in those with prolonged disease duration, which was independent of corticosteroid treatment, indicating disease activity itself is a major factor in the growth retardation. Eight children had low IGF-1 z scores--that is, less than -2.00 from age/sex matched mean. Low IGF-1 z scores were associated with low weight z scores but not with low height z scores. Overnight growth hormone secretory profiles were determined in 10 patients, including seven with low IGF-1, and showed generally normal secretion in all but one patient, who subsequently attained normal concentrations coincident with catch up growth. Increased pulse frequency of overnight secretion was commonly seen. Low IGF-1 concentrations probably result from varying factors, particularly nutritional, but do not reflect marked endocrinological abnormalities in most patients.

Neurohormonal immunoregulation

The immune system may be divided into primary lymphoid organs (bone marrow, bursa of Fabricius, and thymus), which produce mature leukocytes and secondary organs (spleen, lymph nodes, tonsils, Peyer's patches, etc.), which are concerned with specific immune responses. In the primary organs, stem cells proliferate and differentiate into various subsets of polymorphonuclear and mononuclear cells. Evidence is increasing that cell proliferation in the primary lymphoid organs is dependent on pituitary growth hormone (GH) and prolactin (PRL), which control the expression of growth regulatory genes (protooncogenes) such as c-myc and also induce essential growth factors (insulinlike growth factor, thymic and bursal hormones, etc.) and, possibly, their receptors. The adrenocorticotropic hormone-adrenal axis serves as an inhibitory pathway, antagonizing the action of PRL and GH on primary lymphoid tissue. The effect of glucocorticoids is especially forceful on thymocytes through the activation of the genetically programmed suicide pathway. Sex hormones also regulate the primary lymphoid organs, but their mechanism of action remains to be clarified. Thymus-derived feedback signals toward the pituitary gland have already been described. The pituitary gland exerts a similar regulatory influence on mature lymphocytes during their antigen-driven differentiation. PRL or GH is required for primary immune reactions; however, the secondary immune response may be less dependent on these hormones. Once the immune system is primed, antigen itself becomes a primary regulator. Exposure of memory cells to antigen leads to the production of growth factors (interleukins) and to the expression of their receptors. Therefore, antigen appears to fulfill, at this stage, a role that is originally played by GH or PRL in the primary lymphoid organs and, to some extent, also during antigen-driven differentiation. During immune reactions, interleukin-1 and tumor necrosis factor activate the adrenocorticotropic hormoneadrenal axis, which plays an important role in setting upper limits to and terminating responses. Lymphocytes have receptors for and react to numerous hormones, neurotransmitters, and mediators derived from a number of organs and tissues. Therefore, ultimately the reaction of a lymphocyte will be the vector of all positive and negative signals received. A hierarchy and sequential system of signals exists. Primary regulatory signals (competence signals) represent the most powerful regulators (e.g., PRL, GH or antigen) of lymphoid cells. The delivery of a competence signal is the prerequisite for subsequent lymphoproliferation, which is regulated by growth factors that are specific for a certain developmental stage of the lymphoid cell and act sequentially. Hormonal factors that promote growth and differentiation deliver the second regulatory signals. Competence factors and growth and differentiation hormones regulate gene expression in lymphocytes. The third class of signals modulate the function of mature effector cells (e.g., locomotion, secretion, phagocytosis, cytotoxicity). Neuro-transmitters appear to function as secondary signal modulators and tertiary functional regulators.

Prognosis in juvenile chronic arthritis. Follow-up of 433 chronic rheumatic children

A total of 433 children, hospitalized in the Rheumatic Children's Hospital Garmisch-Partenkirchen, were followed on average for 15 years (range 10-22 years) after the onset of definite juvenile chronic arthritis. This paper reveals clearly that the different subgroups of chronic rheumatic disease in childhood with their different complications have a definite relationship to prognosis. Severe limitation with inability to attend normal school or employment occurred in 13.4% of children with systemic polyarticular arthritis (Still's syndrome) and 11.4% of the non-systemic cases, but not in the pauci-articular group. In the latter group 82.5% of the children remained without disability or with only slight impairment: this is significantly better than in the systemic or non-systemic polyarticular groups. Although the high incidence of chronic rheumatic iridocyclitis is common in children with pauci-articular arthritis, none of this group had been handicapped by severe eye complications or blindness, in contrast to several early cases with systemic or non-systemic polyarthritis. This might be due to regular eye checks in the pauci-articular group. Of the children in the systemic polyarticular group 10% were dwarfed. Mortality in the systemic group was 13.8%, in contrast to 1% in the non-systemic polyarticular and 0% in the pauci-articular arthritis group. Secondary amyloidosis was the most important cause of death, mainly in systemic cases. Of the children whose amyloidosis had been verified, 44% died in their second or third decades, mostly with uraemia.