J Am Soc Nephrol 13: 12711278, 2002

Henoch-Schonlein Purpura in Adults: Outcome and Prognostic

Factors
EVANGELINE PILLEBOUT,* ERIC THERVET, GARY HILL, CORINNE ALBERTI,
PHILIPPE VANHILLE, and DOMINIQUE NOCHY
*

INSERM U426, Faculte de Medecine X. Bichat, Universite Paris 07, Paris, France; Nephrology and Renal
Transplant Unit, Hopital Saint Louis, Paris, France; Nephrology and Renal Transplant Unit, Hopital
Europeen George Pompidou, Paris, France; Department of Biostatistics, Hopital Saint Louis, Paris, France;

Nephrology and Renal Transplant, Hopital de Valenciennes, France; Pathology Unit, Hopital Europeen
George Pompidou, Paris, France.

Abstract. Henoch-Schonlein Purpura nephritis (HSPN) has

been extensively studied in children but, its natural history in
adults is much less known. A cohort of 250 adults suffering
HSP was retrospectively analyzed for a median follow-up
period of 14.8 yr. All patients had biopsies consistent with HSP
(predominant IgA mesangial deposits) associated with purpura,
bowel angina, and/or abdominal pain. At presentation, palpable
purpura was present in 96% of patients, and arthritis was
reported in 61%, and gastrointestinal involvement in 48%.
Thirty-two percent of the patients showed renal insufficiency
(Creatinine clearance [CrCl] 50 ml/min), usually associated
with proteinuria (99%) and/or hematuria (93%). Endocapillary
glomerulonephritis was the most frequent lesion on renal biopsy
(61%). At the end of follow-up, patient survival was only 74%.
The first cause of death was carcinoma (most of them of respiratory or digestive tract). Regarding renal function, 11% of patients

reached end-stage renal failure, 13% exhibited severe renal failure

(CrCl 30 ml/min), and 14% moderate renal insufficiency (CrCl
50 ml/min). Clinical remission defined as the absence of proteinuria, hematuria, and a normal renal function was achieved in
only 20%. This is a retrospective study; therefore, it is not possible
to demonstrate any steroid and/or cyclophosphamide efficacy in
diminishing the incidence of renal insufficiency. Multivariate
analysis demonstrated that renal function impairment and proteinuria level at presentation and, on renal biopsy, the degree of
interstitial fibrosis, percentage of sclerotic glomeruli, and presence
of glomeruli with fibrinoid necrosis were associated with a poor
renal prognosis. The data indicate that clinical presentation of
HSPN in adults is severe and its outcome relatively poor, worse
than in children. Identification of clinical and histologic prognostic factors may permit the design of appropriate therapeutic prospective studies.

Henoch-Schonlein purpura (HSP) is a leukocytoclastic vasculitis involving small vessels with the deposition of immune
complexes containing IgA. It is characterized by the association of skin, joint, and gastrointestinal manifestations that may
occur in successive episodes (1). In addition to these manifestations, renal involvement is common, and the long-term prognosis depends on its severity.
HSP primarily affects children, and its incidence is approximately 15 cases/100,000 children per yr (2); it is less common
in adults. Although HSP has been extensively studied in children, much less is known about its natural history in adults.
Apart from a recent multicenter Italian study (3,4), data on this
disease in adults are confined to small series with relatively
short follow-up (515). In adults, however, the incidence of

HSP and the severity of its clinical manifestations appear not to

be the same as in children.
The incidence of renal involvement in adults varies from 45 to
85% of cases, depending on the data for patients and the definition
of renal involvement (16). Among cases of glomerulonephritis,
HSP is only responsible for 0.6 to 2% of adult nephropathies
(7,13,16). The risk of progression to renal insufficiency, which
ranges from 5% (17,18) to 15% (19 21) in children, seems to be
much higher in adults, approximately 30% (0 to 50%) (515).
These differences are probably due to the small size of the series,
the presence or absence of renal involvement, and the differences
in the duration of follow-up. No study has yet been able to
demonstrate the ability of any treatment to prevent progression to
renal failure in either children or adults.
We retrospectively analyzed 250 adults suffering from HSP,
with biopsy-proven renal involvement. The goal of this study was
to describe the clinical, laboratory, and histologic characteristics
of this population at presentation. Patient and long-term renal
follow-up were analyzed, and prognostic factors identified.

Received October 23, 2001. Accepted January 28, 2002.

Correspondence to: Dr. Evangeline Pillebout, INSERM U426, Faculte de
Medecine Xavier Bichat, 16, rue Henri Huchard BP416, 75870 Paris cedex 18,
France. Phone: 33-1-44-85-62-70; Fax: 33-1-42-28-15-64; E-mail:
pillebou@bichat.inserm.fr
1046-6673/1306-1271
Journal of the American Society of Nephrology
Copyright 2002 by the American Society of Nephrology
DOI: 10.1097/01.ASN.0000013883.99976.22

Materials and Methods

Inclusion and Exclusion Criteria
This was a retrospective study of 250 patients aged 15 yr or more
suffering from HSP nephritis. All patients underwent a renal biopsy

1272

Journal of the American Society of Nephrology

between 1983 and 2000, in eleven different medical centers. HSP

nephritis was suspected when hematuria, proteinuria, and/or renal
failure were associated with a characteristic purpuric eruption and/or
abdominal or joint pains (at least two of these three clinical signs).
Predominant mesangial IgA immune deposits on renal biopsy were
required for inclusion in the study. We excluded patients suffering
from IgA nephropathy without systemic signs (Bergers disease),
those with other diseases associating nephropathy and purpuric rash,
i.e., systemic lupus erythematosus and cryoglobulinemia, and those
with thrombocytopenia (100,000/mm3).

Evaluation
Clinical and Biologic Data. The characteristics of skin manifestations, gastrointestinal and joint symptoms, and renal involvement
were recorded. Hypertension was defined according to World Health
Organization criteria.
We recorded the levels of aspartate and alanine amino transferases
(ASAT, ALAT), bilirubin, alkaline phosphatase, and -glutamyl
phosphatase for hepatic disorders. Serum IgA levels were recorded
when available.
Renal insufficiency was defined as creatinine clearance (CrCl) 50
ml/min (calculated by Cockroft formula [22]): moderate renal functional impairment as CrCl between 30 and 50 ml/min, and severe renal
failure as CrCl 30 ml/min. Proteinuria was defined as proteinuria
0.1 g/d, and nephrotic syndrome as plasma albumin 30 g/L and
proteinuria 3 g/d. Hematuria was defined as 10 red cells/mm3 in
the urine, and was macroscopic if 1500 red cells/mm3.
Renal Pathology. All renal biopsies were independently examined by two pathologists blind to clinical features. On immunofluorescence, the predominance of mesangial IgA among glomerular Ig
deposits was required. To be adequate, at least ten glomeruli had to be
present. We evaluated semiquantitatively: (1) endocapillary lesions
and their focal or diffuse distribution; (2) extracapillary proliferation,
graded according to the number of glomeruli involved; (3) interstitial
fibrosis (% of the parenchyma involved); (4) interstitial cell infiltration, tubular necrosis, and red cell casts, graded from 0 to 3
depending on the extent of the parenchyma involved; and (5) arteriolar hyalinosis and arteriosclerosis (from 0 to 3). The proportions of
glomeruli involved by crescents, fibrinoid necrosis, and global sclerosis were recorded. All biopsies were then classified according to the
following classification:
1. Mesangiopathic glomerulonephritis. Histologically normal glomeruli by light microscopy or minimal mesangial prominence.
2. Focal and segmental glomerulonephritis. Segmental endo- and
extracapillary proliferation involving less than 50% of the glomeruli. The remaining involved and uninvolved glomeruli were either
normal or exhibited minimal mesangial prominence.
3. Endocapillary proliferative glomerulonephritis, divided into two
subclasses: 3(a), moderate pure endocapillary proliferative lesions;
3(b), severe endocapillary proliferation, possibly with extracapillary proliferation involving less than 50% of the glomeruli.
4. Endocapillary and extracapillary glomerulonephritis. Lesions were
as above, with crescents involving more than 50% of the glomeruli.
5. Fibrotic kidney with global glomerular sclerosis involving more
than 50% of glomeruli.
Outcome. We evaluated the outcome of each patient, as of the
reference date of July 1, 2000. For patients who died, the cause of
death and renal status at death were determined. When patients were
lost to follow-up, their renal status at the last visit was recorded. Renal
outcome was classified as remission (no renal failure, proteinuria, or

J Am Soc Nephrol 13: 12711278, 2002

hematuria); persistent proteinuria and/or hematuria without renal insufficiency; moderate or severe renal insufficiency as defined above;
and end-stage renal failure (ESRF).

Statistical Analyses
Results were expressed as numbers (percentages) for categorical
variables and as median (1st to 3rd quartiles: Q1 to Q3) or mean (
SD), as indicated, for continuous variables. Comparisons were based
on the 2 test for categorical variables and Wilcoxon or KruskallWallis tests for continuous variables. Survival curves from the day of
renal biopsy were computed using Kaplan-Meier estimate. A Cox
model was fit to evaluate the influence of the following variables on
severe renal failure occurrence: age, CrCl, proteinuria level, hematuria, the pathologic classification, the percentages of crescents, necrotic
glomeruli, sclerotic glomeruli, interstitial fibrosis, and the use of
specific treatments. For model selection, we used stepwise forward
selection. All tests were two-tailed, and P 0.05 was considered
significant. Statistical analyses were performed with SAS 6.12 software (SAS Institute, Cary, NC).

Results
Extrarenal Presentation
The median age at HSP onset was 50 (range, 15 to 86 yr). At
the reference date, the median duration of follow-up was 14.8
(13.1 to 20.8) yr. Male/female ratio was 1.7. In 80 (32%) of our
250 cases, there was a recent history of intercurrent infection,
mostly of the upper respiratory tract.
Most patients (96%) had purpura, always localized to the
lower limbs. Other parts of the body were sometimes involved,
but rarely the face or mucosa (5%). Blisters and hemorrhagic
necrotic skin lesions occurred in 35% of the patients. Relapses
of purpura were observed in 48%. Most of them only relapsed
once, but in rare cases, up to ten times. Arthritis occurred in
61% of the cases, was symmetrical, and the joints most often
involved were the knees or the ankles. Gastrointestinal involvement was observed in 120 cases (48%). The main symptom was colicky abdominal pain. Bleeding occurred in 61
cases of gastrointestinal involvement (51%) and was serious,
requiring transfusion or surgery or leading to death, in 13 cases
(11%). Purpura was the first clinical manifestation in 83% of
the cases, arthritis in 9%, and gastrointestinal involvement in
8%. Renal manifestations preceded clinical symptoms in only
two patients (4 and 5 mo before, respectively).
We frequently observed mild, usually brief, liver disorders,
with evidence of liver cell damage (15%) or cholestasis (87%).
Elevated levels of IgA were found in 118 of the 196 patients
tested (60%). Positive anti-neutrophil cytoplasmic antibody
was found in 4 of the 105 patients tested (3.8%).
Clinical and biologic presentations differed among patients
according to age, as shown in Table 1. More patients aged 30
yr had a recent history of intercurrent infection and arthritis
than older patients. By contrast, in those aged 60 yr, purpura
was much more frequently necrotic, IgA levels were higher,
and CrCl was much lower (this last out of proportion to the
differences expected on the basis of age).

Renal Features
Evidence of renal involvement was detected within a median
period of 2 mo (range, 1 to 4 mo) after onset of the first clinical

J Am Soc Nephrol 13: 12711278, 2002

Outcome of Adult Henoch-Scho nlein Purpura Nephritis

1273

Table 1. Clinical and biologic presentation of Henoch-Scho lein purpura (HSP) according to agea

Parameter

Age 30 yr
(n 54)

Age 30 to 60 yr
(n 111)

Age 60 yr
(n 85)

Recent history of infection, n 80

Presence of arthritis, n 153
Necrotic purpura, n 87
Serum IgA level (g/L)
CrCl (ml/min)

54.7%
71.7%
16.7%
3 0.2
99 5

28.4%
67.9%
36.4%
4 0.2
90 3

23.5%
48.2%
44.7%
5 0.3
46 3

0.0004
0.005
0.0001
0.0001
0.0001

CrCl, creatinine clearance.

symptoms. All patients exhibited some evidence of renal abnormality at the time of biopsy. Nineteen patients (36%) had
high BP, but most of them had a past medical history of
hypertension. Of the 250, 169 patients (68%) had normal and
81 (32%) impaired renal function. Renal function impairment
was moderate in 46 cases (18%) and severe in 35 cases (14%).
Proteinuria was generally mild (66% had 3 g/24 h), and 28%
of the patients had nephrotic range proteinuria. Hematuria was
microscopic in 85% and macroscopic in only 25 patients (10%)
(Table 2).

Pathology
The most frequent lesion in renal biopsies was proliferative
endocapillary glomerulonephritis (class 3) (61%). Proliferative
endo- and extracapillary glomerulonephritis (class 4) was only
present in 21 cases (8%). In these cases, the mean percentage
of glomeruli containing extracapillary proliferation was 62
4%. Circumferential crescents involving all glomeruli were
only seen in four biopsies. Fibrinoid necrosis of the glomerular
tuft was present in 48% of biopsies. The mean percent of
glomeruli per biopsy affected by fibrinoid necrosis was below
5% in classes 1, 2, and 3a, 14 2% in class 3b, and 19 5%
in class 4. Inflammatory infiltrates and interstitial fibrosis were

Table 2. Renal data at the time of renal biopsy for patients

with HSP
% of patients (n 250)

Renal function
normal
moderately impaired
severely impaired
Proteinuria
0.1 g/24 h
0.1 to 1 g/24 h
1 to 3 g/24 h
3 g/24 h
nephrotic
Hematuria
no hematuria
microscopic
macroscopic
Hypertension

67.6
18.4
14.0
4
25.6
38.8
31.6
27.9
5.6
84.6
9.8
36

common, present in 50 and 54% of the biopsies, respectively.

The quantity of red cell casts (117 biopsies) was correlated
with the intensity of tubular necrosis (102 biopsies) (P
0.0001). Only two patients exhibited necrotizing and granulomatous angeitis of the interlobular arteries. Arteriolar hyalinosis (13%) was most frequent among patients with diabetes, and
arteriosclerosis (45%) among those with hypertension. They
only coexisted in 8% of cases.
Almost all of the renal lesions were significantly worse in
those aged 60 yr than in those aged 30 yr. This was true not
only for vascular lesions and interstitial fibrosis, as might be
anticipated, but also for tubular necrosis (0.69 0.80 versus
0.37 0.60; P 0.015) and particularly for the presence of
glomerular fibrinoid necrosis (1.9 3.0 versus 0.7 1.2; P
0.008).

Outcome
The patient survival curve is shown in Figure 1A. Sixty-four
patients (26%) died. Median survival time was 15 yr (range, 13
to 21). The mean age at death was 67.3 1.5 yr. The most
frequent cause of death was neoplasia in 17 patients (27% of
deaths), involving lung in nine patients and upper respiratory
and digestive tracts in five. Eleven of the seventeen patients
dying of cancer never received any immunosuppressive treatment, four had been treated by corticosteroids alone, and two
by corticosteroids associated with cyclophosphamide (P
0.05). They thus did not differ significantly from the remaining
patients as regards treatment. The mean delay between renal
biopsy and death was 42 7 mo (range, 5 to 118 mo).
The second cause of death was infection (16%). Eight of the
twelve lethal infections were attributable to immunosuppressive treatment. The mean delay between renal biopsy and death
there was 11 3 mo (range, 2 to 41 mo). Death was secondary
to HSP evolution in seven patients (11%), due particularly to
severe digestive involvement. Cardiovascular diseases were
responsible for the death of six patients (9%). The ratio of
cardiac to cancer deaths in our patients was thus reversed
compared with that in the general population.
The renal survival curve is shown in Figure 1B. Twentyseven patients (11%) developed ESRF. Those who went into
ESRF tended to do so rather rapidly, 13 (48%) of 27 by 3 yr,
and all but 2 of 27 patients by 10 yr. Twenty-five patients were
dialyzed, and 12 were renal transplant recipients. None of these
recipients lost their graft because of a relapse of HSP. Thirty-

1274

Journal of the American Society of Nephrology

J Am Soc Nephrol 13: 12711278, 2002

treatment would be steroids combined with cyclophosphamide

(Figure 2; P 0.0002). Seventy-five patients received converting enzyme inhibitors and/or angiotensin II receptor blockers. There were no differences between patients and regarding
outcome between the patients receiving these agents and other
patients.

Prediction of Renal Outcome

Figure 1. (A) Estimated Kaplan-Meier survival time after diagnosis of

Henoch-Scho nlein purpura (HSP) and its 95% confidence interval
(CI): median, 178 mo (95% CI, 157 mo to not determined). (B)
Estimated Kaplan-Meier time to occurrence of end-stage renal failure
(ESRF) after diagnosis of HSP and its 95% (CI): median, 246 mo
(95% CI, 189 mo to not determined). Analysis is somewhat hampered
by the limited numbers of patients after 180 mo.

two patients (13%) had severe renal failure, and 35 (14%) had
moderate renal insufficiency. In all, 94 (38%) of the 250
patients exhibited significant renal insufficiency.
At the end of the follow-up, only 20% of the patients were
in clinical remission. Among the other patients, 50% had
microscopic hematuria (isolated in 30 cases), 47% had minimal
or moderate proteinuria, and 8% had nephrotic-range proteinuria. Thirty-eight patients (15%) were lost to follow-up after a
median of 1.8 yr (range, 0.9 to 4.1). Seven of them were in
clinical remission at the time of their last examination.

Univariate Analyses. Table 3 shows the clinical and histologic data for patients at the time of renal biopsy categorized
according to their renal outcome. Prognostic factors for severe
renal impairment are shown in Table 4. Age over 50 yr at onset
was a strong predictor of severe renal failure (relative risk [RR]
2.53; P 0.0012); 66% of the patients with severe renal
failure at their last visit were older than 50 yr at initial biopsy,
and similarly, the mean age of patients with severe renal failure
at the end of follow-up was significantly higher than that of
patients with normal renal function (65 versus 43.5 yr). Clinical presentation, including purpura, arthritis, and abdominal
involvement, was not associated with a significant difference
in renal function at the end of the follow-up. The presence of
renal failure at onset was the strongest predictive factor of renal
failure at the end of follow-up (RR 5.86; P 0.0001); thus
69% of the patients with severe renal failure at presentation had
severe renal failure at the end of follow-up. Proteinuria 1 g/L
(RR 2.13; P 0.0012) and the presence of macroscopic
hematuria (RR 2.13; P 0.038) at biopsy correlated the
CrCl at the end of follow-up, but the presence of nephroticlevel proteinuria was not predictive of a poor outcome.
As shown in Table 3, the pathologic process was more
predictive of the renal prognosis than the clinical presentation.
The glomerular classification was predictive of the renal outcome (P 0.0002). Of the 74 patients with class 1 and class
2 glomerulonephritis, 61 (82%) exhibited normal renal function at the final consultation, and only 11 (15%) exhibited
severe renal failure. In contrast, of the 90 patients with severe
endocapillary proliferative glomerulonephritis (biopsy classes
3b and 4), 47 (52%) exhibited normal renal function at the final

Therapy
Therapy was administered to 140 patients (56%). Steroids
alone were given to 93 (37%), associated with cyclophosphamide to 47 (19%), and cyclophosphamide alone to only one
patient. Levels of proteinuria (1.7 0.2 g/24 h versus 3.7
0.3; P 0.0001) and creatinine (125.4 9.3 mol/L versus
154.7 12.6; P 0.05) were significantly higher in the
treated group of patients. Treatments differed according to the
glomerular classification. The worse the lesion, the greater the
probability that the patient would be treated and that the

Figure 2. Number of patients with severe renal failure at the end of

follow-up, according to their glomerular classification (classes 1, 2,
3a, 3b, 4, and 5) and the treatment received. n 250 patients with
Scho nlein-Henoch nephritis. , no treatment; , corticosteroids
alone; , corticosteroids cyclophosphamide.

J Am Soc Nephrol 13: 12711278, 2002

Outcome of Adult Henoch-Scho nlein Purpura Nephritis

1275

Table 3. Renal outcome in relation to clinical, biologic, and histologic data at the time of renal biopsya
Outcomes Observed during Long-Term Follow-Up
Variable at Biopsy

Age (yr)
CrCl (ml/min)
Hematuria (%)
absence (n 18)
microscopic (n 208)
macroscopic (n 24)
Proteinuria (%)
1 g/24 h (n 81)
1 to 3 g/24 h (n 85)
3 g/24 h (n 84)
Biopsy classification (%)b
class 1 (n 28)
class 2 (n 46)
class 3a (n 82)
class 3b (n 69)
class 4 (n 21)
class 5 (n 3)
Glomerular necrosis (%)
Glomerular sclerosis (%)
Interstitial fibrosis (%)
a
b

Normal Renal
Function n 160

Moderate Renal
Failure n 31

Severe Renal
Failure n 59

43.5 (32 to 58)

86.7 (68.6 to 115.4)

64 (59 to 75)
39.6 (28.3 to 51.7)

65 (35 to 73)
34.3 (18.7 to 57.9)

15 (9.4)
132 (82.5)
13 (8.1)

2 (6.5)
27 (87.1)
2 (6.5)

1 (1.7)
49 (83.0)
9 (15.3)

0.2

60 (37.5)
49 (30.6)
51 (31.9)

12 (38.7)
8 (25.8)
11 (35.5)

9 (15.3)
28 (47.5)
22 (37.3)

0.02

26 (16.4)
35 (22.0)
51 (32.1)
36 (22.6)
11 (6.9)
0
0 (0 to 9)
3 (0 to 11)
0 (0 to 10)

1 (3.2)
1 (3.2)
15 (48.4)
11 (35.5)
3 (9.7)
0
0 (0 to 15)
9 (0 to 20)
10 (0 to 20)

1 (1.7)
10 (16.9)
16 (27.1)
22 (37.3)
7 (11.9)
3 (5.1)
6 (0 to 17)
18 (1 to 42)
20 (10 to 30)

0.001
0.001

0.0002

0.3
0.001
0.001

Data are n (%) or median (quartiles). See Materials and Methods (Renal Pathology) for biopsy classification.
n 249; one biopsy was excluded (only one glomerulus).

consultation, and 29 (32%) exhibited severe renal failure. The

proportion of glomeruli affected by fibrinoid necrosis (RR
1.95; P 0.013) influenced the outcome, but the proportion of
crescents in endocapillary and extracapillary glomerulonephritis did not. The number of global sclerotic glomeruli (RR
2.94) and the degree of interstitial fibrosis (RR 5) correlated
strongly with poor outcome (P 0.0001). The median proportion of global sclerotic glomeruli in biopsies of patients
with severely deteriorated renal function at the end of follow-up was 18% (range, 1 to 42%), contrasting with 3%
(range, 0 to 11%) for patients without severe renal failure. The
median proportion of interstitial fibrosis in biopsies of patients
with severely deteriorated renal function at the end of follow-up was 20% (range, 10 to 30%), contrasting with 0% (0 to
10%) for patients without renal failure. Intensity of interstitial
inflammatory cell infiltration (P 0.0004) also influenced the
renal outcome, but the presence of tubular necrosis did not.
Univariate analysis failed to demonstrate any benefit of
treatment on renal function, but the administration of treatment
was frequently associated with both clinical and histologic risk
factors for renal failure (Figure 2). Of the 59 patients with
severe renal failure or ESRF at outcome, 20 (34%) were not
treated, 24 (41%) received steroids alone, and 15 (25%) received steroids combined with cyclophosphamide. Of the 110
patients not treated, 18% had severe renal failure at the end of
follow-up, 16% had moderately impaired renal function, and
66% had normal renal function. Of the 93 patients treated with

steroids alone, 26% had severe renal failure at the end of

follow-up, 11% had moderately impaired renal function, and
63% had normal renal function. Of the 47 patients treated with
steroids and cyclophosphamide, 32% had severe renal failure
at the end of follow-up, 8% had moderately impaired renal
function, and 60% had normal renal function.
Multivariate Analyses. As shown in Table 5, creatinine
and proteinuria levels at the time of renal biopsy and the
proportions of interstitial fibrosis, glomerular fibrinoid necrosis, and glomerular sclerosis were found to be independent risk
factors for severe renal failure.

Discussion
The goal of this study was to define the long-term prognosis
and risk factors for severe renal failure in adults with HSP
nephritis. In contrast to previous outcome studies of comparable adult populations, multivariate analyses could be performed here because of the larger number of patients and the
length of follow-up. We demonstrated that the clinical presentation of HSP is more severe in older adults and that the renal
prognosis for HSP nephritis is poor compared with what is
observed in series in children. It should be stressed that we
studied a group of selected patients whose HSP-exhibited renal
involvement severe enough to warrant renal biopsy.
Our study extends the findings of previous studies (4 14,15)
conducted in smaller series. The patients in our population
were roughly comparable to those previously described by

1276

Journal of the American Society of Nephrology

J Am Soc Nephrol 13: 12711278, 2002

Table 4. Significant prognostic factors for severe renal

failure by univariate analysisa
Variable at Biopsy

Relative Risk

Age 50 yr
Creatinine 120 mol/L
Proteinuria 1 g/24 h
Macroscopic hematuria
Corticosteroids alone
CS Cy
Biopsy classification
class 1
class 2
class 3a
class 3b
class 4
class 5
Presence of crescents
Glomerular necrosis
10%
Glomerular sclerosis
10%
Interstitial fibrosis 10%

2.53
5.86
2.13
2.13
1.98
1.5
1
9.02
6.52
11.7
11.99
64.48
0.78
1.95

95% CI

1.44 to 4.45
0.0012
3.30 to 10.44 0.0001
1.6 to 6.64
0.0012
1.04 to 4.35
0.038
1.14 to 3.44
0.015
0.82 to 2.74
0.19
1
1.14 to 71.29
0.86 to 49.59
1.57 to 87.33
1.44 to 99.75
6.61 to 629
0.28 to 2.16
1.15 to 3.31

0.037
0.07
0.016
0.022
0.0003
0.64
0.013

2.94

1.71 to 5.05 0.0001

2.58 to 9.68 0.0001

a
For biopsy classification, see Materials and Methods (Renal
Pathology). CS, corticosteroids; CY, cyclophosphamide.

Table 5. Significant prognostic factors for severe renal

failure by multivariate analysis
Variable at Biopsy

Relative Risk

95% CI

Creatinine 120 mol/L

Proteinuria 1 g/24 h
Glomerular necrosis
10%
Glomerular sclerosis
20%
Interstitial fibrosis 10%

4.27
2.98
1.83

0.963 to 0.987 0.0001

1.72 to 7.98
0.0045
1.036 to 1.78 0.04

2.12

1.053 to 4.54

0.02

3.83

1.036 to 1.78

0.0008

others, but they were older on average, and more of them were
hypertensive. Note, however, that the World Health Organization has recently lowered the measures that define hypertension
(23).
We show here that the presentation of HSP differs significantly with age, as older patients exhibited much more severe
renal and extrarenal manifestations than younger patients.
Coppo et al. (3) has shown that children differ from adults by
more frequent joint manifestations, and that this difference
persists in adulthood, with decreasing joint manifestations as
patients age, a finding that our study confirms and extends
(Table 1). All the studies of children (24 26) show that cutaneous necrosis is uncommon and affects fewer than 5% of
cases. In our adult patients, by contrast, necrotic purpura was
much more common and became increasingly so with age.
Pathologically, we found that 8% of patients had endocap-

illary and extracapillary proliferative glomerulonephritis (class

4), which is within the previously described range (3 to 23%)
(4). This great disparity in histologic severity is probably due
to different indications for renal biopsy in different countries.
Furthermore, numerous studies were performed before the
International Study of Kidney Disease in Children classification was instituted, and precise data regarding the pathology
were not always available.
In our series, mortality was high and was in most cases
secondary to neoplasia. Cancer was the leading cause of mortality, accounting for 27% of deaths, a frequency far higher
than that for cardiovascular disease (9%), the leading cause of
death in the general population. This was not correlated with
the use of immunosuppressive treatment; in fact, 8 of 17
patients either had their cancers or precursor conditions, such
as emphysema, at the time of diagnosis of HSP. Furthermore,
these cancers were more likely to be neoplasms of the lung
(14%) and of the upper respiratory and digestive tracts (8%),
than those occurring in general French population (4 and 2%
respectively) (27). To our best knowledge, this high proportion
of deaths related to cancer has never been described in prior
cohort studies of HSP nephritis. Most of the neoplasms in our
patients involved the upper respiratory and digestive tracts,
whose mucosa secrete IgA. It is of interest that the pathophysiology of HSP has been attributed to abnormal IgA clearance
(28,29). Here, however, we found no evidence for a possible
paraneoplastic origin of HSP. The proportion of alcohol and
tobacco abusers in our population was quite high (more than
10%); therefore, it might follow that HSP is most often found
in this setting, a possibility that would also favor the development of neoplasms (30).
The renal prognosis for our series was poor, because nearly
25% of patients reached renal insufficiency. Other studies have
shown better renal outcome. However, our study and that of
Fogazzi et al. (14) are the only ones to have a long follow-up.
This may explain the even poorer renal prognosis reported by
the latter authors (renal insufficiency in almost 50% of cases).
In this respect, HSP nephropathy may resemble IgA nephropathy. Many authors indeed had initially reported a favorable
outcome for IgA nephropathy until the point in time when
follow-up was long enough to enable them to conclude that
end-stage renal disease was a frequent complication (3135).
We were able to identify several risk factors for severe renal
failure. As in other studies, extrarenal clinical manifestations
and gender did not affect the renal prognosis. Age, the presence
of macroscopic hematuria, initial renal insufficiency, and proteinuria 1 g/24 h were found by univariate analysis to be
prognostic factors for a decline in renal function; by multivariate analysis, however, only initial renal insufficiency and
proteinuria were associated with renal insufficiency. Regarding
macroscopic hematuria, HSP nephropathy therefore differs
from IgA nephropathy, as several authors have shown that
macroscopic hematuria is associated with better long-term
renal function (36,37). However, other authors (38,39) have
shown that episodes of macroscopic hematuria are frequently
associated with acute renal failure during IgA nephropathy. In
our patients, these episodes were frequently associated with

J Am Soc Nephrol 13: 12711278, 2002

acute renal failure, possibly because of the tubular necrosis

associated with macroscopic hematuria (P 0.00012). This, in
turn, may be responsible for the association of macroscopic
hematuria with chronic renal failure seen on univariate analysis
(Table 4). By contrast, contrary to the general impression,
macroscopic hematuria was not correlated with the number of
crescents.
In agreement with previous studies (3), we show here that
the degree of proteinuria is a significant prognostic factor. As
far as we know, no other study has demonstrated the role of
initial renal function. Some authors (20) have shown that the
presence of a high percentage of crescents has an unfavorable
effect, but this was not confirmed by others (3,12) or by our
own study. Similarly, we found that no single class of glomerular lesions has a prognostic value by multivariate analysis.
Among the acute lesions considered, the only histologic variable that was an independent prognostic factor was the presence of glomerular fibrinoid necrosis. By contrast, we demonstrate that chronicity factors, such as the percentage of
interstitial fibrosis and of global sclerotic glomeruli, are unfavorable risk factors. These variables are usually described in
other glomerulonephritis, particularly IgA nephropathy
(32,35,40,41).
As to the efficacy of specific treatments by steroids alone or
combined with cyclophosphamide, it is difficult to draw any
conclusions; their efficacy in reducing the incidence of severe
renal insufficiency could not be demonstrated. On univariate
analysis, steroids even seem to worsen the course of renal
evolution. As in most previous retrospective studies of HSP
nephropathy, the limitation of the present investigation was
that the distribution of established renal risk factors was different in treated versus untreated patients. No study has so far
been able to demonstrate the efficacy of treatment of this
nephropathy in adults. The importance of irreversible histologic lesions, such as interstitial fibrosis and glomerular sclerosis, as indices of poor prognosis makes the efficacy of
treatment doubtful. Nevertheless, the prognostic value of glomerular fibrinoid necrosis provides some hope of improvement
with steroids. The results of pediatric therapeutic studies (42
44), even though none was prospective and randomized against
placebo, were mostly positive.
In conclusion, although earlier small series have suggested a
relatively poor prognosis in HSP in adults as compared with
children, this is the first large study with long follow-up to
confirm and refine this impression. It further identifies independent prognostic factors, such as initial renal failure, the
level of proteinuria, and the histologic quantification of interstitial fibrosis and glomerular sclerosis. Because of the retrospective nature of this study and despite the absence of positive
impact of corticosteroids and cyclophosphamide treatment, we
cannot definitely reject such treatments until their failure to
benefit is confirmed in a placebo-controlled clinical trial with
randomized treatment allocation.

Acknowledgments
We are grateful to the following pathologists, who provided listings
of the HSP nephropathy patients and slides of their renal biopsy:

Outcome of Adult Henoch-Scho nlein Purpura Nephritis

1277

He le`ne Beaufils (Hopital Pitie Salpe trie`re, Paris, France), Dominique

Droz (Hopital St Louis, Paris, France), Be atrice Mougenot (Hopital
Tenon, Paris, France), and FrancineWalker (Hopital Bichat, Paris,
France). We are also grateful to the following clinicians from Internal
Medicine Services or from the Groupe dEtudes Ne phrologiques dIle
de France (GENIF) for their cooperation: Vale rie Caudwell and
Olivier Kourilsky (Hopital dEvry, Evry, France), Dominique Chauveau, Philippe Lesavre and Jean Pierre Gru nfeld (Hopital Necker,
Paris, France), Michel Delahousse (Hopital Foch, Suresnes, France),
Gilbert Deray (Hopital Pitie Salpetrie`re, Paris, France), Christian
Jacquot and Jean Barie ty (Hopital Europe en Georges Pompidou,
Paris, France), Henri Kreiss (Hopital Necker, Paris, France), Frank
Martinez and Christophe Legendre (Hopital St Louis, Paris, France),
Luc Moulonguet (Hopital A. Pare , Boulogne, France), Jean Charles
Piette (Hopital Pitie Salpetrie`re, Paris, France), Eric Rondeau and Jean
Daniel Sraer (Hopital Tenon, Paris, France), Je rome Rossert and
Pierre Ronco (Hopital Tenon, Paris, France), and Franc ois Vrtovsnik
and Franc oise Mignon (Hopital Bichat, Paris, France).

References
1. Jennette JC, Falk RJ, Andrassy K, Bacon PA, Churg J, Gross
WL, Hagen EC, Hoffman GS, Hunder GG, Kallenberg CG,
McCluskey RT, Sinico RA, Rees AJ, Vanes LA, Waldherr R,
Wilk A: Nomenclature of systemic vasculitides. Proposal of an
international consensus conference. Arthritis Rheum 37: 187
192, 1994
2. Nielsen HE: Epidemiology of Schonlein-Henoch purpura. Acta
Paediatr Scand 77: 125131, 1988
3. Coppo R, Mazzucco G, Cagnoli L, Lupo A, Schena FP: Longterm prognosis of Henoch-Schonlein nephritis in adults and
children. Italian Group of Renal Immunopathology Collaborative
Study on Henoch-Schonlein purpura. Nephrol Dial Transplant
12: 22772283, 1997
4. Coppo R, Amore A, Gianoglio B: Clinical features of HenochSchonlein purpura. Italian Group of Renal Immunopathology.
Ann Med Interne (Paris) 150: 143150, 1999
5. Bernhardt JP, Chatelanat F, Veyrat R: The Schonlein-Henoch
syndrome in the adult. Clinical study of 16 cases, histological
examination of the kidney in 7 cases. Schweiz Med Wochenschr
96: 1228 1229, 1966
6. Cream JJ, Gumpel JM, Peachey RD: Schonlein-Henoch purpura
in the adult. A study of 77 adults with anaphylactoid or Schonlein-Henoch purpura. Q J Med 39: 461 484, 1970
7. Fillastre JP, Morel-Maroger L, Ducroiset B, Richet G: Renal
involvement in rheumatoid purpura in adults. Study of 20 renal
biopsies. Value of the examination of the glomerulus in immunofluorescence. Presse Med 78: 23752378, 1970
8. Debray J, Krulik M, Giorgi H: Rheumatoid purpura (SchonleinHenoch syndrome) in the adult. Apropos of 22 cases. Sem Hop
47: 18051819, 1971
9. Bar-On H: Renal involvement in adults with Henoch-Schoenlein
syndrome. Harefuah 84: 542544, 1973
10. Sinniah R, Feng PH, Chen BT: Henoch-Schoenlein syndrome: A
clinical and morphological study of renal biopsies. Clin Nephrol
9: 219 228, 1978
11. Roth DA, Wilz DR, Theil GB: Schonlein-Henoch syndrome in
adults. Q J Med 55: 145152, 1985
12. Lee HS, Koh HI, Kim MJ, Rha HY: Henoch-Schoenlein nephritis in adults: A clinical and morphological study. Clin Nephrol
26: 125130, 1986
13. Faull RJ, Aarons I, Woodroffe AJ, Clarkson AR: Adult HenochSchonlein nephritis. Aust N Z J Med 17: 396 401, 1987

1278

Journal of the American Society of Nephrology

14. Fogazzi GB, Pasquali S, Moriggi M, Casanova S, Damilano I,

Mihatsch MJ, Zucchelli P, Ponticelli C: Long-term outcome of
Schonlein-Henoch nephritis in the adult. Clin Nephrol 31: 60
66, 1989
15. Blanco R, Martinez-Taboada VM, Rodriguez-Valverde V, Garcia-Fuentes M, Gonzalez-Gay MA: Henoch-Schonlein purpura
in adulthood and childhood: Two different expressions of the
same syndrome. Arthritis Rheum 40: 859 864, 1997
16. Rieu P, Noel LH: Henoch-Schonlein nephritis in children and
adults. Morphological features and clinicopathological correlations. Ann Med Interne (Paris) 150: 151159, 1999
17. Meadow SR, Glasgow EF, White RH, Moncrieff MW, Cameron
JS, Ogg CS: Schonlein-Henoch nephritis. Q J Med 41: 241258,
1972
18. Niaudet P, Habib R: Schonlein-Henoch purpura nephritis: Prognostic factors and therapy. Ann Med Interne 145: 577580, 1994
19. Yoshikawa N, White RH, Cameron AH: Prognostic significance
of the glomerular changes in Henoch-Schoenlein nephritis. Clin
Nephrol 16: 223229, 1981
20. Goldstein AR, White RH, Akuse R, Chantler C: Long-term
follow-up of childhood Henoch-Schonlein nephritis. Lancet 339:
280 282, 1992
21. Scharer K, Krmar R, Querfeld U, Ruder H, Waldherr R, Schaefer
F: Clinical outcome of Schonlein-Henoch purpura nephritis in
children. Pediatr Nephrol 13: 816 823, 1999
22. Cockroft DW, Gault MH: Prediction of creatinine clearance from
serum creatine. Nephron 16: 31 41, 1976
23. Subcomittee G: 1999 World Health OrganisationInternational
Society of Hypertension Guidelines for the Management of Hypertension. J Hypertens 17: 151183, 1999
24. Abdel-Al YK, Hejazi Z, Majeed HA: Henoch Schonlein purpura
in Arab children. Analysis of 52 cases. Trop Geogr Med 42:
5257, 1990
25. Wananukul S, Pongprasit P, Korkij W: Henoch-Schonlein purpura presenting as hemorrhagic vesicles and bullae: Case report
and literature review. Pediatr Dermatol 12: 314 317, 1995
26. Saulsbury FT: Hemorrhagic bullous lesions in Henoch-Schonlein
purpura. Pediatr Dermatol 15: 357359, 1998
27. Menegoz F, Cherie-Challine L: Cancer en France: Incidence et
mortalite , situation en 1995. Paris, Ministe`re de lemploi et de la
solidarite , 1995
28. Roccatello D, Picciotto G, Torchio M, Ropolo R, Ferro M,
Franceschini R, Quattrocchio G, Cacace G, Coppo R, Sena LM,
et al.: Removal systems of immunoglobulin A and immunoglobulin A containing complexes in IgA nephropathy and cirrhosis
patients. The role of asialoglycoprotein receptors. Lab Invest 69:
714 723, 1993
29. Grossetete B, Launay P, Lehuen A, Jungers P, Bach JF, Monteiro
RC: Down-regulation of Fc alpha receptors on blood cells of IgA
nephropathy patients: Evidence for a negative regulatory role of
serum IgA. Kidney Int 53: 13211335, 1998

J Am Soc Nephrol 13: 12711278, 2002

30. Mashberg A, Boffetta P, Winkelman R, Garfinkel L: Tobacco

smoking, alcohol drinking, and cancer of the oral cavity and
oropharynx among U.S. veterans. Cancer 72: 1369 1375, 1993
31. DAmico G, Minetti L, Ponticelli C, Fellin G, Ferrario F, Barbiano di Belgioioso G, Imbasciati E, Ragni A, Bertoli S, Fogazzi
G, Duca G: Prognostic indicators in idiopathic IgA mesangial
nephropathy. Q J Med 59: 363378, 1986
32. Alamartine E, Sabatier JC, Guerin C, Berliet JM, Berthoux F:
Prognostic factors in mesangial IgA glomerulonephritis: An extensive study with univariate and multivariate analyses. Am J
Kidney Dis 18: 1219, 1991
33. Galla JH: IgA nephropathy. Kidney Int 47:377387, 1995
34. Koyama A, Igarashi M, Kobayashi M: Natural history and risk
factors for immunoglobulin A nephropathy in Japan. Research
Group on Progressive Renal Diseases. Am J Kidney Dis 29:
526 532, 1997
35. Radford MG, Jr., Donadio JV Jr, Bergstralh EJ, Grande JP:
Predicting renal outcome in IgA nephropathy. J Am Soc Nephrol
8: 199 207, 1997
36. Bennett WM, Kincaid-Smith P: Macroscopic hematuria in mesangial IgA nephropathy: Correlation with glomerular crescents
and renal dysfunction. Kidney Int 23: 393 400, 1983
37. Beukhof JR, Kardaun O, Schaafsma W, Poortema K, Donker AJ,
Hoedemaeker PJ, van der Hem GK: Toward individual prognosis
of IgA nephropathy. Kidney Int 29: 549 556, 1986
38. Praga M, Gutierrez-Millet V, Navas JJ, Ruilope LM, Morales
JM, Alcazar JM, Bello I, Rodicio JL: Acute worsening of renal
function during episodes of macroscopic hematuria in IgA nephropathy. Kidney Int 28: 69 74, 1985
39. Delclaux C, Jacquot C, Callard P, Kleinknecht D: Acute reversible renal failure with macroscopic haematuria in IgA nephropathy. Nephrol Dial Transplant 8: 195199, 1993
40. Donadio JV, Bergstralh EJ, Offord KP, Holley KE, Spencer DC:
Clinical and histopathologic associations with impaired renal
function in IgA nephropathy. Mayo Nephrology Collaborative
Group. Clin Nephrol 41: 6571, 1994
41. Ibels LS, Gyory AZ: IgA nephropathy: Analysis of the natural
history, important factors in the progression of renal disease, and
a review of the literature. Medicine (Baltimore) 73: 79 102,
1994
42. Mollica F, Li Volti S, Garozzo R, Russo G: Effectiveness of
early prednisone treatment in preventing the development of
nephropathy in anaphylactoid purpura. Eur J Pediatr 151: 140
144, 1992
43. Kaku Y, Nohara K, Honda S: Renal involvement in HenochSchonlein purpura: A multivariate analysis of prognostic factors.
Kidney Int 53: 17551759, 1998
44. Niaudet P, Habib R: Methylprednisolone pulse therapy in the
treatment of severe forms of Schonlein-Henoch purpura nephritis. Pediatr Nephrol 12: 238 243, 1998

Access to UpToDate on-line is available for additional clinical information

at http://www.jasn.org/