Henoch-Schönlein purpura (HSP) is a non-thrombocytopenic systemic hypersensitivity vasculitis of childhood. Its aetiology remains largely unknown, and the prognosis, although generally very good, depends primarily on the extent of renal involvement.

The primary long term complication is renal disease, which develops in 5% of patients. Renal involvement may manifest itself with haematuria, with or without proteinuria, and occasionally with nephritic or nephrotic syndrome, or with renal failure sometimes associated with a rapidly progressive glomerulonephritis. Urinalysis and blood pressure measurements are mandatory when evaluating a child with HSP. Patients who develop renal involvement generally do so within three months of the onset of rash.¹ Although patients with only haematuria do not develop end-stage renal disease (ESRD), the association of proteinuria and haematuria may be associated with a 15% risk and the combination of nephritic-nephrotic syndrome with a 50% risk of progression to ESRD.^2,3 Children with overt renal disease at presentation are therefore managed early by paediatric nephrologists and may require follow up for at least five years.⁴ However, as most of these studies have originated in secondary referral centres on selected populations of children with established nephritis, these recommendations may not necessarily apply to unselected populations of children, most of whom would have no active nephritis on diagnosis.

Furthermore, the follow up of those without initial nephritis, nephrotic syndrome, or renal failure is still recommended to detect renal involvement, and although it is suggested that it need not be five years, the exact duration of follow up remains undetermined.⁵ With an incidence of HSP in the population estimated at 14 per 100 000 and with 75% of cases occurring in children aged 2–11 years, with the majority not developing long term renal impairment, a more judicious evidence based approach to selective follow up is needed if resources are to be used appropriately.^6,7

If the risk of long term renal impairment could be estimated by the presence or absence of renal involvement at diagnosis, and if the time period when renal involvement is very unlikely after HSP diagnosis could be determined, that evidence could be used to identify which children with HSP (without initial nephritic or nephrotic syndrome or renal failure) will need follow up, and if so, for how long. To try to answer these questions we undertook a systematic review of the published literature.

METHODS

RESULTS

DISCUSSION

In this review, children with normal urinalysis had no long term renal impairment, confirming previous studies, including a report on a selected cohort of children with normal urinalysis at diagnosis of HSP.^5,8 As the great majority (97%) of children who will have abnormal urine findings would develop them within six months of the initial presentation, even when urinalysis is normal at presentation, urine testing needs to be continued for that period.¹⁸ The proportion of children presenting with IHP was less in this analysis (37%) than in previous studies (67%) and may be related to the grouping of several large unselected cohorts of affected children.⁶

The study also confirmed that the risk of long term renal impairment was low (1.6%) in those with only isolated proteinuria or haematuria, but was much higher (19.5%) if the initial presentation was complicated by nephritic or nephrotic syndrome, with the risk being 2.5 times greater in females than males.^3,8–,¹¹

The strength of analysing several large unselected cohort studies with good follow up for outcomes is tempered by some potential limitations. Some studies did not specify if the urinary or renal abnormalities observed during the follow up period were all related to the first presentation of HSP or may have been associated with relapses, which may render inaccurate the calculations made based on the findings on the initial presentation. It is impossible to be certain that some children with renal involvement at presentation were not followed up more thoroughly and for longer than the others; ascertainment bias cannot therefore be excluded, nor can the theoretical possibility that some patients with no renal involvement at presentation may have developed some urinary abnormalities at a later stage. The definitions of haematuria, proteinuria, nephritic syndrome, nephrotic syndrome, and renal failure were not identical across all studies, but it is unlikely that minimal differences in case definition would have induced significant bias in some reports. The fact that some studies did not differentiate whether haematuria was isolated or was associated with proteinuria could jeopardise the analysis of the observed outcomes, as the presence of proteinuria carries a more severe prognosis on the renal function than isolated haematuria.³⁴ Similarly, other studies did not differentiate whether nephrotic syndrome occurred with or without nephritis, making the interpretation of outcome fraught with a degree of imprecision. As many children with HSP remain undiagnosed and/or do not present to healthcare workers, all those analysed reports which originated from medical establishments may have inflated the risk, and the true risk of long term renal impairment in children with normal or isolated urinary findings may be even lower than calculated in this review. However, as it is very unlikely that children with overt renal disease do not present for medical care, the calculated risk in those with established nephropathy is probably a reflection of the true risk. As most analysed reports did not specifically report any therapy for renal involvement on diagnosis, it could be assumed that the findings represent the natural history of this vasculitis, but we cannot rule out the role, if any, of any therapy which had not been reported. A main limitation in this review is the unavailability, for each individual case, of the time periods for urinary abnormalities observed after the diagnosis of HSP and the time for the observed outcome, precluding an acturial or “survival”-type data analysis and only allowing simple aggregation of the data according to the blocks of outcome-time given in the studies. In addition, the variable duration of follow up in the studies and the only remaining option of consolidating outcome data available at very different follow up periods does not truly reflect the prognosis, as the absence of nephropathy after two years in one study has a completely different meaning than the same outcome in another study spanning 20 years. Only a large prospective cohort study with a long follow up period could provide a more reliable survival analysis and address these inherent weaknesses.

Despite these limitations, and until large prospective studies have been completed, a practical algorithm may be suggested as a result of this review and the reports on which it was based. At diagnosis and at each HSP recurrence, blood pressure measurement and urinalysis need to be carried out. If haematuria and/or proteinuria is found, serum urea and creatinine determination will be needed. In case of normal urinalysis or if there is IHP without NNP, a periodic check of urinalysis will be needed up to six months after diagnosis. There is no need to follow up after the first six months those whose urinalysis remains normal, but measurements of serum urea and creatinine need to continue in the presence of continued IHP. In case of NNP, a paediatric nephrologist needs to be consulted as urgent evaluation is necessary and long term renal follow up will be definitely needed, even after resolution of the nephritis, especially in affected girls as they will need monitoring of renal function and blood pressure during and after future pregnancies.⁸