56 Scand J Rheumatol 2006;35:56–60

Predictive factors for nephritis, relapse, and significant proteinuria

in childhood Henoch–Schönlein purpura

JI Shin, JM Park, YH Shin, DH Hwang, JH Kim, JS Lee

Department of Paediatrics, The Institute of Kidney Disease, Yonsei University, College of Medicine, Seoul, Korea

Objective: To identify predictive factors for nephritis, relapse, and significant proteinuria in childhood Henoch–
Schönlein purpura (HSP).
Methods: Two hundred and six consecutive patients with HSP (93 female, 113 male), followed up at a single
Scand J Rheumatol Downloaded from informahealthcare.com by Case Western Reserve University on 10/31/14

centre between 1996 and 2001, were analysed retrospectively. They were regularly monitored for clinical and
laboratory parameters for renal sequelae and relapse.
Results: Nephritis was seen in 78 patients (38%), relapse in 52 (25%), and significant proteinuria in 39 (19%). In
univariate analysis, an older age at onset (w10 years), persistent purpura, severe bowel angina, and relapse were
identified as factors associated with nephritis and significant proteinuria. Relapse-related factors were an older
age, persistent purpura, severe bowel angina, and leucocytosis. Logistic regression analysis showed that nephritis
was significantly associated with an older age, persistent purpura, and relapse, and significant proteinuria was
closely related to severe bowel angina and relapse.
Conclusion: We identified some predictors for nephritis, relapse, and significant proteinuria in childhood HSP,
and close attention should be paid to those patients with the risk factors, such as an older age at onset, persistent
purpura, severe bowel angina, and relapse.
For personal use only.

Henoch–Schönlein purpura (HSP) is a vasculitis
due to immunoglobulin A (IgA)-mediated inflamma-
tion of small vessels, characterized by involvement
of the skin, joints, gastrointestinal tract, and kidneys
(1–3). Renal involvement has been reported to
occur in 30–60% of the patients with HSP (4).
Although the short-term prognosis of HSP is
favourable in general, some patients with severe
nephritis may progress to chronic renal insufficiency
(5–7). There have been some reports on the
risk factors for renal involvement in HSP (4, 8, 9),
but finding predictors for the development
of severe nephritis would be the more important
issue for the clinician caring for children with
HSP and, until now, few studies have reported on
the risk factors for relapse or significant proteinuria
(8, 9).
Therefore, we evaluated the predictive factors
for nephritis, relapse, and significant proteinuria
in children with HSP by using univariate and
multivariate analyses and designed a clinically
Jae Seung Lee, Department of Paediatrics, CPO Box 8044, Yonsei
University College of Medicine, 134 Shinchon-dong, Seodaemun-gu,
Seoul 120-752, Korea.
E-mail: jsyonse@yumc.yonsei.ac.kr
Received 18 January 2005
Accepted 13 May 2005
applicable predictive model for nephritis and
significant proteinuria.
Patients and methods
Study group
Two hundred and six consecutive patients with HSP
(93 female, 113 male), followed up at the Department
of Paediatrics, Yonsei University from 1996 to 2001,
were analysed retrospectively. The diagnosis of HSP
was made according to the criteria of Michel et al
(palpable purpura, bowel angina, gastrointestinal
bleeding, haematuria, age at onset(20 years, no
medication; the presence of at least three criteria
yielding a correct classification of HSP cases of
87.1%) (10). The patients with less than 1 year follow-
up and those referred from other hospitals were
excluded in this study. The age at onset was 7.2¡2.8
years, and the duration of follow-up was 3.1¡2.2
years. During the follow-up the patients were
regularly monitored for clinical and laboratory
parameters for renal sequelae and relapse. A total
of 175 patients (85%) were treated with steroids for
relief of abdominal or joint symptoms, which did not
differ in patients with or without nephritis. This
study was approved by the institutional review board
and the research ethics committee of Yonsei
Severance Hospital.

# 2006 Taylor & Francis on license from Scandinavian Rheumatology Research Foundation
www.scandjrheumatol.dk DOI: 10.1080/03009740510026841
 Nephritis in Henoch–Schönlein purpura
Definitions
Recurrent purpura persisting for 1 month or
more was classified as persistent purpura (4).
Arthritis was defined as swelling of the joints or
painful periarticular soft tissue oedema. The patients
who could not eat any meals due to diffuse
abdominal pain were found to have severe bowel
angina. Nephritis was defined by the presence of
macroscopic or microscopic haematuria (w5 red
blood cells per high-power microscopic field in a
centrifuged specimen) with or without proteinuria
(4). Significant proteinuria was defined by
proteinuria equal or superior to 20 mg/m2/h (5).
Relapse was considered when a patient who was
Scand J Rheumatol Downloaded from informahealthcare.com by Case Western Reserve University on 10/31/14
previously diagnosed as having HSP and who had
been asymptomatic for at least 1 month presented
with a new flare of cutaneous lesions or other
systemic complications (11). Leucocytosis and
thrombocytosis were defined by w156109/L and
w4006109/L, respectively. Elevation of C-reactive
protein required values ofw0.8 mg/dL. The levels of
serum IgA and C3 were measured in 73 patients by
the automated Dade Behring Nephelometer II (Dade
Behring Marburg GmbH, Marburg, Germany).
For personal use only.
Statistics
Using SPSS (version 11.0 for Windows), compar-
isons were made by using the Student’s t-test and x2-
test for univariate analysis. Stepwise logistic regression
analysis was used for multivariate analysis. All
differences were considered significant at pv0.05.
Results
Clinical features of the patients
All patients developed palpable purpura and 41
(20%) manifested persistent purpura. Abdominal
pain was seen in 108 (52%) patients, severe bowel
angina in 26 (13%), and gastrointestinal bleeding
such as guaiac-positive stools or grossly bloody
stools in 47 (23%) of the 206 patients. Arthritis was
present in 129 (63%) patients, and relapse occurred in
52 (25%). Nephritis was found in 78 (38%) patients,
and 75% of them developed nephritis within 3
months after the onset of the disease. The mean
interval from the initial symptoms of HSP to the
onset of haematuria was 38¡19 days. Nephritis
manifested as isolated haematuria in 21 (27%),
haematuria with mild proteinuria (v20 mg/m2/h) in
18 (23%), and significant proteinuria (w20 mg/m2/h)
in 39 (50%) of the 78 patients. Nephrotic-range
proteinuria was diagnosed in 29 patients (37%), and
six (8%) showed acute nephritic features. At the latest
57
observation, 41 patients showed normal urinalysis
and renal function, 34 minor urinary abnormalities,
one active renal disease, and two had progressed to
chronic renal insufficiency.
Univariate analysis of predictive factors
Patients with nephritis were significantly older than
those without nephritis (8.2¡3.1 vs. 6.7¡2.6,
pv0.0001), and data analysis at various cut-off
values of age revealed that an age of more than 10
years was a significant factor for nephritis and
significant proteinuria. Persistent purpura, severe
bowel angina, and relapse were also more common
in patients with nephritis and significant proteinuria
than in those without. The prevalence of arthritis was
lower in patients with nephritis (p50.003).
Gastrointestinal bleeding was not different in
patients with and without nephritis, but was closely
associated with severe bowel angina (46% vs. 19%,
p50.002). Serum IgA mean levels did not differ in
patients with and without nephritis. Relapse-related
factors were an age of more than 10 years at onset,
persistent purpura, severe bowel angina, and leuco-
cytosis (Table 1). In the present study, 17 patients
were hypertensive at the onset of nephritis, and a
univariate analysis revealed that an age more than 10
years (41% vs. 16%, p50.02), severe bowel angina
(29% vs. 11%, p50.046), and gastrointestinal bleed-
ing (47% vs. 21%, p50.029) were significantly related
to the development of hypertension. Hypertension in
14 patients was well controlled by anti-hypertensive
medications and safely tapered off, but three showed
persistent hypertension during the course of the
disease.
Multivariate analysis of predictive factors
In the multivariate analysis (Table 2), an age greater
than 10 years, persistent purpura, and relapse were
identified as factors associated with nephritis. The
existence of arthritis was protective. Relapse was
significantly related to an age more than 10 years,
persistent purpura, and severe bowel angina. In
patients with significant proteinuria, severe bowel
angina and relapse were significant predictive factors.
In addition, an age more than 10 years [odds ratio
(OR) 3.24; 95% confidence interval (CI) 1.10–9.50;
p50.032] and gastrointestinal bleeding (OR 3.61;
95% CI 1.21–10.83; p50.022) were independent risk
factors of hypertension on a multivariate analysis.
Predictive model for nephritis and significant proteinuria
We designed predictive models for nephritis and
significant proteinuria based on the variables with a
www.scandjrheumatol.dk
 www.scandjrheumatol.dk

58
Table 1. Univariate analysis of the factors associated with nephritis, relapse, and significant proteinuria in HSP.

Variable Nephritis Relapse Significant proteinuria

% with % without p-value % with % without p-value % with % without p-value

Scand J Rheumatol Downloaded from informahealthcare.com by Case Western Reserve University on 10/31/14

(n578) (n5128) (n552) (n5154) (n539) (n5167)

Age w10 years 30 12 0.001 31 14 0.008 31 16 0.028
Boys 54 56 ns 52 56 ns 54 55 ns
Persistent purpura 39 9 v0.0001 44 12 v0.0001 44 14 v0.0001
Abdominal pain 55 51 ns 64 49 ns 62 50 ns
Severe bowel angina 26 5 v0.0001 33 6 v0.0001 36 7 v0.0001
Gastrointestinal 24 22 ns 31 20 ns 23 23 ns
bleeding
Arthritis 50 70 0.003 62 63 ns 54 65 ns
Relapse 46 13 v0.0001 – – – 51 19 v0.0001
Preceding infection 26 35 ns 25 34 ns 18 34 ns
Leucocytosis 18 13 ns 27 11 0.006 21 14 ns
Thrombocytosis 32 41 ns 44 36 ns 36 38 ns
Increased CRP 30 33 ns 39 29 ns 28 32 ns
For personal use only.

Increased ASO titre 33 31 ns 35 31 ns 26 33 ns

Serum IgA level (mg/ 223¡90 267¡80 ns 249¡118 221¡70 ns 219¡89 243¡91 ns
dL)
Serum C3 level (mg/ 103¡31 115¡32 ns 107¡36 104¡28 ns 99¡31 112¡31 ns
dL)

CRP, C-reactive protein; ASO, anti-streptococcal O titre.

Table 2. Multivariate analysis of the factors associated with nephritis, relapse and significant proteinuria in HSP.

Variable Nephritis Relapse Significant proteinuria

OR (95% CI) p-value OR (95% CI) p-value OR (95% CI) p-value

Age w10 years 2.97 (1.27–6.93) 0.012 2.39 (1.05–5.44) 0.038 1.93 (0.79–4.74) ns
Persistent purpura 3.44 (1.37–8.63) 0.009 3.49 (1.48–8.22) 0.004 1.86 (0.72–4.81) ns
Severe bowel angina 2.80 (0.85–9.21) ns 3.56 (1.25–10.11) 0.017 4.11 (1.36–12.41) 0.012
Relapse 3.68 (1.66–8.17) 0.001 – – 2.49 (1.07–5.83) 0.035
GI bleeding 0.69 (0.30–1.58) ns 1.22 (0.53–2.79) ns 0.56 (0.21–1.53) ns

JI Shin et al
Leucocytosis 0.96 (0.36–2.54) ns 2.28 (0.91–5.69) ns 1.18 (0.41–3.43) ns
Arthritis 0.35 (0.18–0.70) 0.003 0.91 (0.43–1.92) ns 0.57 (0.25–1.26) ns

OR, odds ratio; CI, confidence interval; GI, gastrointestinal.

 Nephritis in Henoch–Schönlein purpura
p-value lower than 0.05 (12). Variables in these
models included an age of more than 10 years at
onset, persistent purpura, severe bowel angina, and
relapse. The model gave one point to each one of
these four variables associated with nephritis and
significant proteinuria. The risk of nephritis or
significant proteinuria increased on increasing the
sum of scores composed of these predictive factors
(Figure 1).
Discussion
Since the first description by Henoch a century ago,
Scand J Rheumatol Downloaded from informahealthcare.com by Case Western Reserve University on 10/31/14
nephritis has been recognized as the most proble-
matic of all clinical manifestations of HSP (2, 5).
Until now, however, very few studies have examined
predictors of renal sequelae and relapse in childhood
HSP (4, 8, 9).
Kaku et al reported persistent purpura, severe
abdominal symptoms, and decreased factor XIII
activity to be significant risk factors for renal
involvement based on a Cox regression model (4).
Sano et al indicated that an age greater than 4 years
and severe abdominal symptoms were independent
For personal use only.
Figure 1. Predictive model for (A) nephritis and (B) significant
proteinuria in children with Henoch–Schönlein purpura. Variables
included in these models were the following: an age of more than
10 years at onset, persistent purpura, severe bowel angina, and
relapse.
59
risk factors for renal involvement by using logistic
regression. Persistent purpura had a borderline
significance (p50.071) (8). Recently, Rigante et al
demonstrated that nephritis was associated with only
persistent purpua but not with severe abdominal pain
and the age at onset (9).
In the present study, the HSP features that clearly
emerged as predictive factors for nephritis by
univariate statistical analysis—that is, an older age,
persistent purpura, severe bowel angina, and
relapse—were linked to each other and all indicated
a severe disease. They were also independent
predictors for nephritis, relapse, and significant
proteinuria in various combinations on a multi-
variate analysis. The above studies have shown that
persistent purpura is an important predictive factor
in the development of nephritis and our data also
demonstrated that it was an independent predictive
factor for nephritis and relapse. Regarding the age at
onset, an age greater than 10 years showed a
strong association with nephritis in our study,
and was an independent predictive factor for
nephritis and relapse. Although severe bowel angina
was not an independent predictive factor of nephritis,
it was the best predictor of relapse and significant
proteinuria.
Although Sano et al reported that persistent
purpura was the independent risk factor for
significant proteinuria (8); our study demonstrated
that severe bowel angina and relapse were significant
predictive factors. Rigante et al reported that
relapsing disease was significantly related to
persistent purpura, but they could not explain
the association between renal involvement and
relapse (9). We also found that patients with
relapse were more likely to develop nephritis and
significant proteinuria on a multivariate analysis, and
persistent purpura, severe bowel angina, and an older
age were closely associated with the relapse of
disease. Therefore, the relapse of disease during
follow-up reinforces the need for careful renal
monitoring.
A predictive model was designed in this study
based on the risk factors for nephritis and significant
proteinuria (12), and it was very useful to apply to an
individual patient clinically, and we could predict
nephritis or significant proteinuria in most children
with HSP with this model.
In conclusion, we have identified some predictors
for nephritis, relapse, and significant proteinuria in
patients with HSP, and close attention should be paid
to those patients with the risk factors, such as an
older age at onset, persistent purpura, severe bowel
angina, and relapse. Further studies are necessary
to identify whether more aggressive treatment in HSP
children with these factors can prevent severe
nephritis or relapse.
www.scandjrheumatol.dk
 60
References
1. Koutkia P, Mylonakis E, Rounds S, Erickson A.
Leucocytoclastic vasculitis: an update for the clinician.
Scand J Rheumatol 2001;30:315–22.
2. Davin J-C, Weening JJ. Henoch–Schönlein purpura nephritis:
an update. Eur J Pediatr 2001;160:689–95.
3. Lee TH, Lee EY, Cho YS, Yoo B, Moon HB, Lee CK.
Concurrent occurrence of chylothorax and chylous ascites in a
patient with Henoch–Schönlein purpura. Scand J Rheumatol
2003;32:378–9.
4. Kaku Y, Nohara K, Honda S. Renal involvement in Henoch–
Schönlein purpura: a multivariate analysis of prognostic
factors. Kidney Int 1998;53:1755–9.
5. Meadow SR, Glasgow EF, White RHR, Moncrief MW,
Cameron JS, Ogg CS. Schönlein–Henoch nephritis. Q J Med
1972;41:241–58.
Scand J Rheumatol Downloaded from informahealthcare.com by Case Western Reserve University on 10/31/14
6. Szer IS. Henoch–Schönlein purpura: when and how to treat.
J Rheumatol 1996;23:1661–5.
7. Cron RQ, Sharma S, Sherry DD. Current treatment by United
States and Canadian pediatric rheumatologists. J Rheumatol
1999;26:2036–8.
For personal use only.
www.scandjrheumatol.dk
JI Shin et al
8. Sano H, Izumida M, Shimizu H, Ogawa Y. Risk factors
of renal involvement and significant proteinuria in
Henoch–Schönlein purpura. Eur J Pediatr 2002;161:
196–201.
9. Rigante D, Candelli M, Federico G, Bartolozzi F,
Porri MG, Stabile A. Predictive factors of renal
involvement or relapsing disease in children with
Henoch–Schönlein purpura. Rheumatol Int 2005;25:
45–8.
10. Michel BA, Hunder GG, Bloch DA, Calabrese LH.
Hypersensitivity vasculitis and Henoch–Schönlein purpura:
a comparison between the 2 disorders. J Rheumatol
1992;19:721–8.
11. Blanco R, Martı́nez-Taboada VM, Rodrı́guez-Valverde V,
Garcia-Fuentes M, Gonzalez-Gay MA. Henoch–Schönlein
purpura in adulthood and childhood: two different
expressions of the same syndrome. Arthritis Rheum
1997;40:859–64.
12. Garcia-Porrua C, Gonzalez-Louzao C, Llorca J, Gonzalez-
Gay MA. Predictive factors for renal sequelae in adults
with Henoch–Schönlein purpura. J Rheumatol 2001;28:
1019–24.