Portal de Periódicos da CAPES

Henoch-Schönlein Purpura in Children from Northwestern Spain

2001; Wolters Kluwer; Volume: 80; Issue: 5 Linguagem: Inglês

10.1097/00005792-200109000-00001

1536-5964

Maria C Calviño, Javier Llorca, Carlos García‐Porrúa, JOSE L. FERNÁNDEZ-IGLESIAS, Pilar Rodríguez‐Ledo, Miguel Á. González‐Gay,

Renal Diseases and Glomerulopathies

Introduction Abbreviations used in this article: ACR, American College of Rheumatology, CI, confidence interval, ESR, erythrocyte sedimentation rate, HSP, Henoch-Schönlein purpura, ICAM-1, intercellular adhesion molecule-1, IgA, immunoglobulin A, URT, upper respiratory tract Henoch-Schönlein purpura (HSP) is a systemic vasculitis characterized by purpuric skin lesions unrelated to any underlying coagulopathy, gastrointestinal manifestations, arthritis, and renal involvement (25,51). Infiltration of small blood vessels with polymorphonuclear leukocytes and the presence of leukocytoclasia are typical pathologic findings in this vasculitis. In addition, immunofluorescence staining of tissues usually discloses the presence of immunoglobulin A (IgA)-dominant immune deposits in the wall of the small vessels and in the renal glomeruli (21,30). Henoch-Schönlein purpura is mainly a childhood disease: it is the most common type of vasculitis in children and an infrequent condition in adults (26). The first description of this syndrome was given by Heberden (27), who reported a child with joint pain and painful subcutaneous edema, abdominal pain, vomiting, bloody stools and urine, and "bloody points" over the skin of his legs. However, the names of Schönlein and Henoch are most commonly used as the designation of the syndrome. Initially, Schönlein (52) described the association between arthralgia and purpuric cutaneous lesions in a child using the term "purpura rheumatica." Some years later, his pupil Henoch (28) described a syndrome of purpura, severe abdominal colic, and melena. A few years later, Henoch referred to nephritis as a complication of this syndrome (29). As recently pointed out by Stone (56), there has been a relative paucity of clinical and epidemiologic studies on HSP in children over the past few years (50). Of note, the long-term morbidity and mortality of HSP are predominantly attributable to renal involvement. In some studies HSP in children has been reported to account for 5%–15% of patients entering end-stage renal failure (10,37). Most of these studies were done on series of selected patient populations with kidney dysfunction attended in reference centers. Two unselected series (33,55), however, suggested a good prognosis for HSP nephritis in children—a more optimistic outcome than most published series based on more selected groups of patients. To investigate this syndrome further, we examined the incidence, clinical spectrum, and prognosis of HSP in an unselected population of children diagnosed at the single reference hospital for a defined population in northwestern Spain over a 20-year period. Special interest was focused on the outcome of these patients and, in particular, on the risk factors implicated in the development of permanent renal damage. Patients and Methods A retrospective study of the case records of all patients diagnosed with primary cutaneous vasculitis in the Division of Pediatrics of the Hospital Xeral-Calde (Lugo, Spain) from January 1980 through December 1999 was undertaken. In Lugo patients aged 14 years and younger are seen at the Pediatrics Division. The Hospital Xeral-Calde is the only referral center for a mixed urban (40%) and rural population of almost 250,000 people living in the Lugo region of northwestern Spain. The main characteristics of the Lugo population have been reported previously (18,19,23,24). The population is relatively static, has its own regional language, and is considered to be descended from Celtic origin (15). During the past 2 decades the population has barely changed, and migratory flows have been relatively low. Inclusion criteria Classification criteria have been proved to be useful in epidemiologic studies of HSP (50,57). The American College of Rheumatology (ACR) criteria and those proposed by Michel et al do not require a biopsy, and IgA deposits were not examined in those studies (39,40). In Lugo, patients with primary cutaneous vasculitis were classified as having either HSP or hypersensitivity vasculitis following the criteria put forward by Michel et al (39). Patients were classified as having HSP if they met 3 or more of the following criteria: 1) palpable purpura, 2) bowel angina, 3) gastrointestinal bleeding, 4) hematuria (macroscopic or microhematuria), 5) aged ≤20 years at the onset of disease, and 6) absence of previous history of medications before the onset of the vasculitis. Patients who fulfilled fewer than 3 criteria were classified as having hypersensitivity vasculitis and therefore were excluded from this study. In children from Lugo, a diagnosis of cutaneous vasculitis was considered in most cases without skin biopsy if they had typical nonthrombocytopenic symmetric palpable purpura involving the lower extremities. For the diagnosis of primary cutaneous vasculitis in children other conditions, such as connective-tissue diseases and infections, in particular meningitis, also had to be excluded (4). Clinical definitions As reported in former studies on adults with primary cutaneous vasculitis (8,18,20,24), drug-intake history and upper respiratory tract (URT) infections before the onset of the vasculitis were considered precipitating events. These events were considered to be present if there was a close temporal relationship (less than 7 days) between the start of drug intake and the onset of HSP or if there was a URT infection shortly before the onset of disease that might explain the occurrence of this condition. A URT infection in children was considered if a cold, influenza, or pharyngitis was observed within the week before the onset of cutaneous vasculitis. We defined the clinical features as follows: Nephritis (renal involvement) was defined as the presence of any renal event that happened over the whole course of the disease. It was defined as previously reported (8,18,20,24,45,49) : 1) "Mild nephropathy" if hematuria (≥5 red blood cells/high power microscopic field) and/or proteinuria (>300 mg/24 hours), without nephrotic range, was present, 2) "Severe nephropathy" if nephrotic syndrome (that is, >40 mg/m 2 body surface per hour or >50 mg/kg per day or >2 g/day proteinuria with plasma albumin ≤25 g/L), with or without edema and/or acute nephritic syndrome (that is, hematuria with at least 2 of the following: hypertension, elevated plasma urea or creatinine, and oliguria) was present, 3) "Renal insufficiency" if the plasma creatinine concentration was more than 125% the upper limit of normal, 4) "Renal sequelae" were considered to be present if at last follow-up (at least 1 year) a patient had any of the renal complications described above. However, a patient who had proteinuria, hematuria, or renal insufficiency during the course of the disease and at last follow-up did not have any abnormality in the urine and had normal renal function was not classified as having renal sequelae. Therefore, the "renal sequelae" group was patients who had permanent renal involvement at last follow-up (20). Joint manifestations: if patients complained of well-defined arthralgia, or if synovitis was observed on examination (18). Gastrointestinal manifestations: 1) Bowel angina: diffuse abdominal pain that worsened after meals or bowel ischemia usually including bloody diarrhea, 2) gastrointestinal bleeding: melena, hematochezia, or positive test for occult blood in the stool (8,18,20,24). Laboratory data: 1) Anemia: hemoglobin lower than 11 g/dL. 2) Leukocytosis: white blood cell count higher than 11,000/mm 3 . 3) Elevated erythrocyte sedimentation rate (ESR) (Westergren): if values were higher than 20 mm/1st hour. 4) High serum IgA: as normal range of serum IgA depends on the age of the patient, a child was considered to have high serum IgA if the IgA value was above the normal range for his/her specific age at the time of diagnosis. Relapse: if a patient diagnosed with HSP and asymptomatic for at least 1 month presented a new flare of skin lesions or other systemic complications (8,18,20). Data collection Clinical and laboratory data at the time of diagnosis of the patients with HSP were extracted from their clinical records according to a specifically designed protocol and stored in a computerized file. In addition to demographic features, data on the following items were analyzed: season and year of diagnosis, initial symptoms of disease, delay to diagnosis from the onset of symptoms, precipitating events (etiologic factors), clinical manifestations (including location of skin lesions; joint, gastrointestinal, and renal manifestations), interval from the onset of symptoms to the development of renal manifestations, laboratory abnormalities (including anemia, leukocytosis, platelet count, elevated ESR, increased IgA and creatinine serum levels, presence of hematuria, proteinuria, nephrotic syndrome, renal insufficiency, throat culture, positive antinuclear antibodies, and decrease of C3 and/or C4 serum levels), and corticosteroid therapy requirement. To examine follow-up and renal sequelae, between June and December 2000, patients classified as having HSP were asked by telephone interview to attend the hospital for further evaluation. Those patients who agreed to come to the hospital for further evaluation were examined by a pediatrician (MCC) and a rheumatologist (MAG-G). Clinical history asking especially for relapses, physical examination, and laboratory analyses, including full blood cell count and routine blood and urine biochemistry tests, were performed again. Those patients with less than 1 year's follow-up who did not attend the hospital for further evaluation (n = 9) were excluded from the analysis of the follow-up of the disease. Finally, relapses and renal sequelae were only examined in those children with at least 1 year's follow-up. Statistical analysis For incidence rate calculations, populations with 0–4, 5–9, and 10–14 years, respectively, were obtained from the 1970, 1981, and 1991 censuses, and the 1999 calculated population. Intercensal populations were estimated by exponential interpolation. To investigate if there were nonrandom peaks in the annual incidence, the actual number of annual cases was compared with the number of cases predicted by a Poisson regression. Age-standardized incidence rates were calculated by the direct method using the European population as standard; the European standard was preferred to the World standard because it better represents a developed country's population. We computed the incidence rates for the whole period 1980–1999 and for the periods 1980–1989 and 1990–1999 using the number of new cases with HSP in each period as the numerator and the estimated population of our area as the denominator. The average annual incidence rate for the whole period 1980–1999 was calculated as the incidence for the 20-year period divided by 20. The average annual incidence rates for the 2 periods, 1980–1989 and 1990–1999, were calculated as the incidence for each 10-year period divided by 10. Rates for HSP were reported as cases for 100,000 population aged 14 years and younger in Lugo. Ninety-five percent confidence intervals (CI) were calculated assuming a Poisson distribution. Continuous data were described as mean and standard deviation (mean ± SD), and categorical variables as percentages. Means and incidence rates were compared using the Student t test. To compare categorical data we performed the chi-square test. When the minimum expected value was less than 5, the chi-square test with Yates correction was used. The predictive model for renal sequelae was obtained by discriminant analysis. We chose a discriminant analysis rather than logistic regression because some variables could not be included in a logistic regression model. Predicted values were computed and receiver operating characteristic (ROC) curves were drawn. Statistical significance was defined as p < 0.05. Calculations were performed with the statistical package Stata Intercooled, release 6.0 (Stata Corp., College Station, TX). Results Between January 1980 and December 1999, 78 children were classified as having HSP. Epidemiologic features of children with HSP The main epidemiologic characteristics of patients with HSP are summarized in Table 1. The median age at the onset of symptoms was 5.5 years. Girls outnumbered boys (53.8%) but the difference of gender was not statistically significant. Cases occurred more commonly in fall and winter. In 28 of 78 cases (35.9%) a history of URT infection before the onset of the vasculitis was found. Drug intake, in particular analgesics (acetaminophen) or antibiotics (beta-lactams, generally amoxicillin), was observed in 23.1% of the patients. In most cases, patients were diagnosed within the first week after the onset of symptoms.TABLE 1: Epidemiologic data and etiologic factors in 78 children with Henoch-Schönlein purpuraIn boys, cases of HSP occurred more commonly between the ages of 4 and 6 years (Figure 1). In girls, in contrast, the onset of disease was more common between the ages of 3 and 7 years. Differences in the distribution of cases by age between boys and girls were not statistically significant (p = 0.46).Fig. 1: Distribution of cases of Henoch-Schönlein purpura by age and gender at the onset of disease.In Figure 2 the annual incidence rate by gender is shown. Peaks of incidence were observed in 1985 and 1999; however, no rhythmic pattern in the annual incidence of HSP in children was observed.Fig. 2: Annual incidence rate by gender during the period of study. Incidence is per 100,000 inhabitants aged 14 years and younger.The overall average annual incidence rate for the 20-year period was 10.45/100,000 people aged 14 years and younger (boys, 9.67; girls, 11.23). The average annual incidence rate during the period 1980–1989 was 8.49/100,000 people aged 14 years and younger (boys, 6.08; girls, 10.92), and during the period 1990–1999, the rate was 12.64/100,000 people aged 14 years and younger (boys, 13.01; girls, 12.25). The adjusted rates for the population aged 14 years and younger by age and sex are shown in Tables 2 and 3. During the period 1990–1999 there was a significant increase in the incidence of children with HSP. Such a significant difference (p = 0.01) was mainly due to the significant increase in the number of patients with HSP seen at the youngest age (aged 0–4 years: p = 0.003), in particular boys, during the period 1990–1999 (see Table 3).TABLE 2: Average annual incidence rates for children with HSP in Lugo, Spain (Rates per 100,000 inhabitants aged 14 years and younger for the period 1980–1999)TABLE 3: Average annual incidence rates for children with HSP in Lugo, Spain (Rates per 100,000 inhabitants aged 14 years and younger for the periods 1980–1989 and 1990–1999)Initial features In 24 of 78 children (30.8%) skin lesions were not the initial manifestations of the disease (Table 4). Abdominal pain was the initial manifestation, which preceded the onset of purpura by 1–10 days (mean, 3.3 d) in 13 patients (16.7%). Arthralgia and/or arthritis was the only initial manifestation in 9 cases (11.5%). In these cases joint manifestations occurred before the onset of purpura by 1–14 days (mean, 5.1 d). Also, in 2 children abdominal and joint manifestations were the initial symptoms, beginning 3 days before the onset of skin lesions. In the remaining 54 patients (69.2%) palpable purpura alone (24 patients) or associated with abdominal and/or articular manifestations (30 patients) was the presenting manifestation (see Table 4).TABLE 4: Initial features of the disease in 78 children with HSPClinical features All children developed nonthrombocytopenic palpable purpura during the course of the disease. In all cases the purpuric lesions concentrated on the buttocks and lower extremities. The upper extremities and trunk were less commonly involved (Table 5). Besides the typical purpuric rash, other skin lesions, generally macular, papular, or, more rarely, urticarial or vesicular ones, were observed in 34 (43.6%) patients. In all cases, however, palpable purpura was the predominant cutaneous lesion. Joint manifestations, in particular arthritis, were common during the early course of the disease; they were observed in 61 of 78 (78.2%) children. In most cases joint involvement was an oligoarthritis involving feet and ankles or knees and, less commonly, upper extremity joints. In Table 6 the joints involved in the series of patients with arthritis are shown. Fifty-seven children (73.1%) suffered abdominal pain. Gastrointestinal bleeding, which was manifested by the presence of grossly bloody stools in 12 cases, occurred in 30.8% of the children (see Table 5). A colonoscopy confirmed the presence of vasculitis in 2 children; in addition, 1 patient had a radiographically confirmed intussusception. Renal manifestations were observed in 42 (53.8%) children within the first 3 months after the onset of symptoms. Nephritis generally occurred early in the course of the disease (Figure 3). All but 1 of the patients with renal manifestations had hematuria, microscopic in 31 of 41 cases and gross hematuria in the remaining 10 children (Table 7). Nineteen of the 41 (46.3%) patients with hematuria also had proteinuria. Isolated proteinuria without hematuria was exceptional, observed in only 1 case. From a clinical point of view, renal involvement was mild in most cases (see Table 5). Six patients of the 42 (14.3%) patients with nephritis developed severe nephropathy consisting of nephrotic syndrome within the first 3 months after the onset of disease. Another 2 patients were hypertensive and had acute renal insufficiency with increase of serum creatinine levels early during the course of the disease (see Table 5). These 2 patients also developed nephrotic syndrome within the first 3 months after the onset of disease. Finally, 3 patients had testicular involvement and 1 patient suffered bullous skin lesions. Corticosteroid therapy (in most cases 1–2 mg/kg per day as initial dose) was prescribed in 18 (23.1%) patients, generally due to gastrointestinal manifestations or severe renal complications. A girl and a boy were also treated with azathioprine and cyclophosphamide, respectively, due to severe nephritis.TABLE 5: Main clinical features in 78 children with HSPTABLE 6: Distribution of arthritis in 78 children with HSPTABLE 7: Laboratory findings * in children with HSPFig. 3: Time from onset of symptoms to development of nephritis in weeks.Laboratory abnormalities Forty-one of 78 patients had leukocytosis at the time of diagnosis. An almost similar proportion (63.9%) of patients had an elevated ESR. Anemia, however, was observed in only 6 children. Platelet counts were generally normal; however, they were increased in a few patients with severe gastrointestinal manifestations. A streptococcal infection was confirmed by throat culture in 8 of 22 patients. Serum IgA concentration was found to be increased in 20 of the 35 (57.1%) children in whom it was tested. Hematuria was present in more than 50% of patients (see Table 7). Of note, focal and proliferative glomerulonephritis with IgA mesangial deposits was confirmed in the 4 cases in whom renal biopsy was performed. Outcome Sixty-nine of 78 children with HSP had at least 1 year's follow-up. With the exception of those patients with nephritis, HSP in children from Lugo was generally acute and self-limited, lasting a few weeks. After a median follow-up of 7 years, 8 (11.6%) of 69 patients had persistent hematuria with proteinuria, although none of them had developed renal insufficiency. In this regard, those 2 patients who had suffered renal insufficiency early in the course of the disease also had improved filtration function. Relapses were observed in 10 (14.5%) cases, especially in those who developed renal sequelae (Table 8).TABLE 8: Outcome of 69 children with HSP and a follow-up of at least 1 yearEpidemiologic and clinical differences between children with and without renal sequelae No significant differences in age at disease onset were found. Renal sequelae were not associated with the onset of the vasculitis in any season in particular. Also, a history of URT infections or drugs before the onset of disease was not associated with a bad outcome. There was a slightly longer delay to diagnosis, from the onset of symptoms until the time of diagnosis at the hospital, in the group of patients who developed renal sequelae (permanent renal involvement) (5.1 days versus 3.9 days in those without renal sequelae) (Table 9).TABLE 9: Epidemiologic differences between children with HSP with and without renal sequelae at last follow-up (at least 1 year)As expected, hematuria at the onset of disease or renal manifestations within the first 3 months after the onset of symptoms of vasculitis were significantly more common in the group of patients with renal sequelae. Of note, the presence of nephrotic syndrome during the course of the disease was generally associated with permanent renal involvement (renal sequelae) (p < 0.001) (Table 10). Also, relapses of the vasculitis were significantly more common in children with renal sequelae (p < 0.001).TABLE 10: Clinical and laboratory differences between children with HSP with and without renal sequelae at last follow-up (at least 1 year)Age-related differences To assess whether the age of onset was associated with different disease severity and outcome, children were stratified in 3 different age-groups: 0–4 years, 5–9 years, and 10–14 years, according to the age at the onset of vasculitis. No epidemiologic or clinical differences were observed; thus, the frequency of severe gastrointestinal manifestations and the proportion of patients who developed nephritis was similar in these 3 groups (data not shown). To investigate further whether the age of disease onset was associated with a different outcome, characterized by a higher frequency of relapses and, especially, the presence of permanent renal involvement, those patients who had at least 1 year's follow-up were also stratified by age (Table 11). We found that the age at the onset of HSP was not associated with a different outcome in terms of permanent renal involvement.TABLE 11: Differences in outcome by age in 69 children with HSP and a follow-up of at least 1 year *Predictive model for renal sequelae As discussed in the Methods section, to establish the best clinical variables associated with renal sequelae a discriminant analysis was undertaken. Based on that, the presence of nephrotic syndrome during the first 3 months after the onset of symptoms was the best predictor for renal sequelae. Thus, most children who suffered this complication had persistent renal involvement at the end of the present study (Table 12). The odds ratio for renal sequelae in patients who had developed nephrotic syndrome was 22.9 (95% CI: 5.53–94.7), p < 0.0001. Using this model and considering as positive all patients who had developed nephrotic syndrome, we obtained the following results: sensitivity 75% (6 of 8), specificity 96.7% (59 of 61), positive predictive value 75% (6 of 8), and positive likelihood ratio 22.7. With this model 65 of 69 (94.2%) patients with a follow-up of at least 1 year were classified correctly for the risk of developing renal sequelae. In Figure 4 the ROC curve for nephrotic syndrome as predictor of renal sequelae is shown.TABLE 12: Influence of the presence of nephrotic syndrome during the course of disease as predictor of renal sequelae in 69 children with HSP and a follow-up of at least 1 year *Fig. 4: Receiver operating characteristic (ROC) curve for nephrotic syndrome as a predictor of renal sequelae.Discussion In a former report we summarized the epidemiology of the systemic vasculitides in adults from northwestern Spain (24). In the present report, we have focused on the analysis of HSP in children in the same area. In addition to providing new data on the incidence and the clinical spectrum of manifestations of this syndrome in southern Europe, we examined particularly the outcome of unselected children with this condition, that is, the frequency of renal sequelae. We also searched for a predictive model for persistent renal involvement. Although HSP may occur from age 6 months to adulthood, 75% of cases are seen in children under 10 years: the disease is observed predominantly in children between the ages of 2 and 10 years, and the median age of onset is approximately 4 years (50). In keeping with former reports, in the present study only 13 of 78 (16.7%) children were 10 years or older at the onset of symptoms, and the median age was 5.5 years. Most series have described a higher incidence in boys, with a ratio of boys to girls of more than 1.3:1.0 (1,46,50,54). In northwestern Spain, however, there was a predominance of girls, with a ratio of girls to boys of 1.2. The annual incidence rate of HSP in children generally ranges between 13.5 and 18 per 100,000 (42,46,55). In northwestern Spain the average annual incidence rate of HSP in the population aged 14 years and younger during the second 10-year period of study (1990–1999) was similar to the incidence reported before in other countries. The etiology of HSP remains unknown. Some reports have shown an increased number of new cases in spring, fall, and winter months (1,2,8,50). Almost 70% of the cases occurred in fall and winter. This higher incidence in the coldest months may speak in favor of an infectious trigger in the pathogenesis of the disease. Indeed, infections, in particular those from URT, were reported as a precipitating event in at least 50% of the cases reported in some series (16). In children from Lugo, a history of URT infection before the onset of disease was found in more than 30% of the cases. Among the URT infections, pharyngitis or rhinopharyngitis and tracheobronchitis have been frequently implicated in the development of the disease. Other foci of infection such as the skin have been considered uncommon. In a controlled study, a frequency of throat infection in 52% of the patients with HSP and only 22% in the control group was observed (16). In HSP, the organism most frequently isolated is the beta-hemolytic streptococcus (7,16). Other organisms and vaccinations against typhoid, paratyphoid A and B, measles, cholera, and yellow fever have also been implicated in the development of HSP (12,43,53,57). Although with a lower frequency than in hypersensitivity vasculitis, medications have sometimes been considered to be precipitating factors for the development of HSP in children (8). IgA has been thought to play an important role in the pathogenesis of HSP (57). IgA-dominant immune deposits affecting small vessels are frequently observed in this syndrome. In this regard, the Chapel Hill Consensus Conference definition of HSP is based on the presence of IgA deposits (30). However, few children with clinical features of HSP undergo immunohistochemical analyses (11). Thus, the absence of these data prevents a definitive diagnosis to be made based on the Chapel Hill Consensus Conference definition. In our case, as in other series (8), pediatricians from the Lugo region did not routinely undertake skin biopsies on children with typical purpuric symmetrical palpable purpura involving the lower extremities (4). For these reasons, classification criteria were used in the present study. In addition to the typical purpuric rash involving mainly buttocks and legs, gastrointestinal complications have been described in up to 75% of the series (6,34). Abdominal pain is frequently colicky, caused by peritoneal or visceral purpura leading to submucosal and mucosal extravasation of blood and edema fluid, which may further lead to ulceration of the bowel mucosa and eventual bleeding into the lumen. Stools may show occult blood. In addition, melena and hematemesis have been reported in up to 50% and 15% of the patients, respectively (1). As in our series, acute intussusception is uncommon (1,14,46). Perforation is exceptional; however, severe cases may lead to hemorrhage and shock (1,9). Other uncommon complications such as malabsorption, exudative enteropathy, and hemorrhagic ascites have been reported (5,31,48). As with gastrointestinal complications, joint manifestations may precede the development of palpable purpura by several days in up to 25% of patients with HSP (44). The 5-day average delay between the onset of joint manifestations and the occurrence of cutaneous lesions seen in Lugo was the same as that reported by Farley et al (16). During the course of the disease, joint manifestations usually occur in more than 50% of cases. As in former studies (50), in the patients from Lugo arthralgia or arthritis, involving mainly the large joints of the lower extremities and often associated with edema, was generally transient, and permanent deformities were not seen. Renal involvement, generally manifested as hematuria either isolated or in association with proteinuria, constitutes the most serious feature of HSP. Its occurrence has been reported in 30%–80% of the cases (32). This complication is lower in children younger than 2 years and higher in older children. However, renal involvement may be even more common as minor degrees of focal mesangial proliferation have been described in the biopsies of patients without urinary abnormalities (58). As previously reported (50), in the present study renal involvement occurred within the first 3 months after the onset of disease. Recurrences of hematuria may be observed in association with new relapses of skin lesions. Seldom do recurrences of hematuria occur a long time after other manifestations of HSP have been resolved (25). As observed in 2 patients in the present series, the most severe clinical presentation is that of mixed nephritic and nephrotic syndromes with hematuria, hypertension, and renal insufficiency associated with severe proteinuria and hypoalbuminemia (38,41). The long-term morbidity and mortality of HSP are almost completely related to renal involvement. However, many of the previous series that reported a relatively high frequency of chronic renal failure in children and adults were based on selected patient populations, usually sent to referral centers because of kidney dysfunction (13,17,35,38). In this regard, based on an unselected population of children and adults with HSP, Blanco et al (8) reported a more severe clinical syndrome with a higher frequency of renal involvement in adults than children. Nevertheless, the final outcome in both age-groups was similar, with a complete recovery in most patients. Henoch-Schönlein purpura is generally a self-limited disease. The average duration of the disease is 4 weeks (1); however, in both children and adults, relapses of the disease are not uncommon. The frequency of relapses varies from series to series. In the present study, relapses were seen in almost 15% of the patients with more than 1 year's follow-up, and they were significantly more common in those patients who developed renal sequelae. As discussed above, comparisons among different series are difficult because many of the previous series were based on selected patient populations. However, as in the present study, in 1999 Saulsbury (50) reported the clinical spectrum of HSP in children from a region in Charlottesville, Virginia, in the United States, also using the ACR classification criteria for HSP. This fact allowed us to compare the clinical features of HSP in the Lugo region of southern Europe with those from the United States (Table 13). In both series the mean age at the onset of disease was similar (5.9 ± 2.9 yr in Charlottesville, Virginia, and 6.2 ± 3.1 yr in Lugo), and the disease was more common in winter and fall months. The only difference between the series was a lower frequency of boys in Lugo. This difference, however, was not statistically significant (p = 0.15). In both series the frequency of high IgA serum levels was similar (see Table 13).TABLE 13: Comparative analysis of the main clinical features of HSP in 2 different regions of the world *We have recently observed an HLA-DRB1*01 association with HSP in northwestern Spain (4). However, patients with severe gastrointestinal or renal sequelae did not have any specific HLA-DRB1 association other than the underlying association with the HLA-DRB1*01 allele. Also, although in patients with HSP from Lugo the intercellular adhesion molecule-1 (ICAM-1) gene polymorphism alone was not associated with the development of the vasculitis, genetic polymorphisms of ICAM-1 at codon 241 have been associated with the susceptibility to severe gastrointestinal complications (3). A major matter of concern for clinicians who see children with HSP is to predict the risk of permanent involvement, characterized by renal sequelae. At present our group is engaged in basic studies to examine possible genes implicated in the development of this complication. From a clinical point of view, few studies have tried to address this problem. Kraft et al (34) observed that the presence of proteinuria and hematuria at the onset of disease was associated with progression to renal insufficiency. Bunchman et al (10) observed that a failure to reach a creatinine clearance of more than 70 mL/min per 1.73 square meters at 3 years after the onset of HSP predicted progression to end-stage renal failure. In an unselected series of 141 children with HSP, Koskimies et al (33) reported persistence of abnormal urinary sediment for at least 1 month in almost 30% of their patients. However, after a mean follow-up of 7.2 years, only 0.7% and 1.4% of the children had progressed to end-stage renal failure or developed chronic glomerular disease, respectively. In another unselected series of 270 children with HSP, Stewart et al (55) observed initial evidence of renal involvement in 20% of patients. A new examination at an average time of 8.3 years after the onset of disease disclosed a good overall prognosis in that series with mortality lower than 1% overall and long-term morbidity of only 1.1%(55). These 2 unselected series suggest a good prognosis for HSP nephritis in children. It is noteworthy that, based on a longer-term follow-up of more than 20 years in 78 children with HSP nephritis, Goldstein et al (22) observed that the severity of the clinical presentation correlated well with the outcome. Those children with the most severe clinical presentation had the worst outcome. In their series, almost 50% of the children who had nephritic syndrome, nephrotic syndrome, or both at the onset of disease developed long-term impairment of renal function. Furthermore, in 16 of 44 girls who later had full-term pregnancies, proteinuria and/or hypertension was observed during the pregnancies, even in the absence of previous active renal disease. Due to this, Goldstein et al (22) suggested that HSP nephritis in children should receive long-term follow-up, especially during pregnancy. Based on the morphologic classification of HSP glomerulonephritis evolved by the International Study for Kidney Disease in Children (13), Goldstein et al (22) also correlated the outcome with the severity of the lesion on the initial renal biopsy. More recently, in 1998 Kaku et al (32) studied the relationship between the progression of renal involvement in children with HSP and various factors. They observed renal involvement in 32% of their 194 children after a period of time between 3 days and 17 months from the onset of disease. Age at onset of more than 7 years, the presence of persistent proteinuria, and a decreased activity of coagulation factor XIII increased the risk of developing renal insufficiency. Using Cox regression model analysis these authors observed that severe abdominal symptoms, persistent purpura, and decreased activity of coagulation factor XIII increased the risk of renal involvement in HSP with hazard ratios higher than 2. In contrast, the use of corticosteroids reduced the risk of nephritis. These authors concluded that patients with risk factors for renal involvement should be treated with corticosteroids to avoid renal impairment. Further prospective and controlled studies are required to support the use of corticosteroids for preventing renal involvement in HSP. In 1999, Mata-Arnaiz et al (36), in an unselected population in which children and adults with HSP were analyzed together, described a higher risk of renal sequelae in those patients with previous antibiotic intake. In our series of children a history of drug intake shortly before onset of disease was not associated with a higher risk of renal sequelae. Of note, in our series of adults with HSP, permanent renal involvement was present in 36% of patients (20). The presence of hematuria at disease onset and the persistence of renal manifestations during the course of the disease were significant indicators of the possible development of renal sequelae in adults. These manifestations along with other features, such as onset in summer, anemia at disease onset, or relapses of the disease, allowed us to predict the development of renal sequelae in most of our patients (20). In the present study, although hematuria and proteinuria during the first 3 months after the onset of the vasculitis and relapses during the course of the disease were also more common in patients with renal sequelae, nephrotic syndrome was the best predictor for renal sequelae. Indeed, most children who developed this complication early on during the course of the disease had persistent renal involvement at the end of the study. In summary, HSP is a common vasculitis, frequently related to a URT infection. In children from northwestern Spain HSP is generally a benign and self-limited condition, although some children develop permanent renal involvement. This fact supports the need for close follow-up. Long-term follow-up studies in unselected patients from other countries are required for better understanding of the nature and outcome of this syndrome. Summary Henoch-Schönlein purpura (HSP) is a systemic vasculitis characterized by purpuric skin lesions unrelated to any underlying coagulopathy, gastrointestinal manifestations, arthritis, and renal involvement. It is the most common type of vasculitis in children and an infrequent condition in adults. The long-term morbidity and mortality of HSP are predominantly attributable to renal involvement. Most studies of renal outcome were done on series of selected patient populations with kidney dysfunction attended in reference centers. There has been a relative paucity of clinical and epidemiologic studies on HSP in children over the past few years. To investigate this syndrome further, we examined the incidence, clinical spectrum, and outcome, and, in particular, the risk factors implicated in the development of permanent renal damage in an unselected population of children diagnosed with HSP at the single reference hospital for a defined population in northwestern Spain over a 20-year period. Seventy-eight children were classified as having HSP. The median age at the onset of symptoms was 5.5 years. Girls outnumbered boys (54%). Cases occurred more commonly in fall and winter. A history of upper respiratory infection before the onset of the vasculitis was not uncommon. The overall average annual incidence rate for the 20-year period (1980–1999) was 10.45/100,000 people aged 14 years and younger. Abdominal pain and/or joint manifestations preceded the onset of purpura in 31% of the children. All children developed nonthrombocytopenic palpable purpura during the course of the disease. Arthritis, generally involving the lower extremities, was seen in 65% of patients. Gastrointestinal bleeding occurred in 31% and renal manifestations in 54% of children within the first 3 months after the onset of symptoms. All but 1 of the patients with renal manifestations had hematuria, associated with proteinuria in 19 of 41 cases. Sixty-nine of the 78 cases were followed for at least 1 year. After a median follow-up of 7 years, 8 (11.6%) of the 69 patients had persistent hematuria with proteinuria; however, none of them had developed renal insufficiency. Hematuria at the onset of disease or renal manifestations within the first 3 months after the onset of symptoms of the vasculitis and relapses were significantly more common in the group of patients with renal sequelae. Of note, the presence of nephrotic syndrome during the course of the disease was generally associated with permanent renal involvement (renal sequelae) (p < 0.001); however, the age at the onset of disease was not associated with a different disease severity and outcome. Using discriminant analysis the best predictor for renal sequelae was the presence of nephrotic syndrome during the first 3 months after the onset of symptoms. The odds ratio for renal sequelae in patients who had developed nephrotic syndrome was 22.9, p < 0.0001. Using this model, 65 of 69 (94.2%) patients with a follow-up of at least 1 year were classified correctly for the risk of developing renal sequelae. In conclusion, in northwestern Spain, HSP is a common, generally benign, and self-limited vasculitis, although some children, in particular those who develop nephrotic syndrome, suffer permanent renal involvement. Long-term follow-up studies in unselected patients from other countries are required for better understanding of the nature and outcome of this syndrome. Acknowledgments We are indebted to the members of the Pediatrics Division of the Hospital Xeral-Calde, Lugo, Spain. We are in debt to Dr. Victor Alcalde, Dr. Mercedes Lueiro, and Dr. Maria Luisa Fernández (Dermatology Division) for their outstanding help in the evaluation and diagnosis of the patients with cutaneous vasculitis. Also, the authors are grateful for the collaboration given by Dr. Teresa Armada and Dr. Jaime Capellá from the Medical Record Department of the Hospital Xeral-Calde.