Clinical Features and Prognosis of Henoch-Schönlein Purpura in Children and Adults: A 13-Year Retrospective Study at a Single Centre

Article information

Abstract

Introduction

In children, Henoch-Schönlein purpura (HSP) is a common form of systemic vasculitis that is known to generally have a favorable prognosis [1,2]. Conversely, in adults, the prevalence of HSP is low, but is also more likely to show renal involvement, resulting in a poor prognosis [3,4]. However, although reports have compared the differences between children and adults with respect to HSP, the studies were limited by the lack of cases and short observation periods, and there have been few reports on experimentally or clinically specific differences and prognoses [5-7]. In the present study, we retrospectively reviewed 469 cases of HSP in children and adults, over a period of 13 years, and examined differences in their clinical characteristics, examination outcomes, and prognoses.

Material and methods

We performed a retrospective study using medical records from a total of 469 inpatients and outpatients who had been diagnosed with HSP at the Department of Pediatrics, Department of Nephrology, Department of Rheumatology, or Department of Dermatology at the Wonju Severance Christian Hospital, between June 2003 and February 2016. The patients were divided into a child group aged 18 years or younger (353 patients) and an adult group aged over 18 years (116 patients). In accordance with the criteria of the European League against Rheumatism/Paediatric Rheumatology International Trials Organization/Paediatric Rheumatology European Society (EULAR/PRINTO/PRES), all patients were diagnosed with HSP if they showed palpable purpura and had at least one of the following accompanying conditions: a) abdominal pain; b) arthritis or arthralgia; c) biopsy of affected tissue demonstrating predominant immunoglobulin A (IgA) deposition; and/or d) renal involvement [8].

Predisposing factors were classified as upper respiratory tract infection (URI), cancer, or drugs. In terms of clinical manifestations, we investigated skin lesions, joint pain, and gastrointestinal (GI) symptoms. The laboratory studies performed were erythrocyte sedimentation rate (ESR), hemoglobin, white blood cell (WBC) count, C-reactive protein (CRP), blood urea nitrogen (BUN), uric acid, C3/C4, PCR for eight strains of respiratory virus (adenovirus, corona virus, influenza A and influenza B, respiratory syncytial virus (RSV), metapneumovirus, rhinovirus, parainfluenza virus), antistreptolysin O (ASO), Mycoplasma pneumoniae-immunoglobulin M (IgM), and IgA, performed at the time of initial admission to the hospital. Test results were classified as demonstrating an increase or a decrease, based on the reference values in the 20th edition of the Nelson Textbook of Pediatrics and the 19th edition of the Harrisons Principles of Internal Medicine [9,10].

Renal involvement was categorized in terms of microscopic/ gross hematuria, proteinuria, nephrotic syndrome, and normal, based on urinalysis performed throughout the period of care. Renal outcome was classified as normal urine, persistent hematuria±proteinuria, or chronic kidney disease (CKD; denoted by continually elevated Cr, or by the patient being a recipient of dialysis or kidney transplant). Persistent urinary abnormalities or CKD were defined as poor renal outcome, and a binary logistic regression analysis was performed.

SPSS 18.0 (IBM Corp., Armonk, NY, USA) was used for statistical analysis, and chi-square test or Fisher’s exact test was used to compare categorical variables. A binary logistic regression analysis was performed. P<0.05 was defined as a statistically significant result.

Results

1. Patients characteristics

There were a total of 469 patients included in this study. The mean age was 7.62±3.76 years in the child group (353 patients; 1-18 years) and 47.8±18.0 years in the adult group (116 patients; 19-87 years). In terms of sex ratio, male patients were more common in both groups (child vs. adult; 53.8% vs. 52.6%, P=0.905). The mean follow-up duration was 392.5 days for the child group and 565.1 days for the adult group.

In terms of predisposing factors, URI was the most common in the child group (45.9% vs. 23.3%, P<0.001).

Drug use was reported in 17 adults solely (colchicine, dapsone, triamcinolone, antidiabetic/antihypertensive drugs, 13.8%), and cancer was reported in three adults only (Hodgkin’s lymphoma, adenocarcinoma, ovarian cancer, 2.6%). Spring was the highest frequency for both group (30.3% vs. 27.6%, P=0.66), antithetically summer was the lowest frequency for both group (14.7% vs. 19.0%, P=0.347).

In terms of skin lesions, purpura was observed on the lower limb in 85.5% of child patients and 59.5% of adult patients, and could be seen frequently in other locations in adults (P<0.001). Joint pain was significantly more frequent in the child group than in the adult group (54.4% vs. 27.6%, P<0.001). Overall GI symptoms were more frequent in the adult group (36.8% vs. 37.1%, P=0.510) than in the child group, but there was no significant difference between two groups. In terms of specific symptoms, abdominal pain was more frequent in the child group (31.2% vs. 26.6%, P=0.411) than in the adult group, while nausea/vomiting (2.0 % vs. 2.6%, P=0.711), and hematochezia (1.4% vs. 7.1%, P=0.007) were more frequent in the adult group, differences of hematochezia were statistically significant (Table 1).

Table 1.

Clinical Characteristic between Child and Adult Group

2. Laboratory outcomes

The mean test results for platelet count (372,030±120,726 ×10⁶/L vs. 278,786±108,779×10⁶/L, P<0.001) were significantly higher in the child group. Compared with the child group, the adult group showed a greater increase in anemia (7.5% vs. 16.4%, P=0.006), C3 hypocomplementemia (1.1% vs. 10.2%, P=0.002), C3 hypercomplementamia (3.6% vs. 4.1%, P=1.0), and C4 hypocomplementemia (1.8% vs. 2.1%, P=0.302). While the child group were common in C4 hypercomplementemia (11.5% vs. 4.2%, P=0.033) than the adult group. Uric acid showed a higher mean value in adult patients than in child patients (4.09±1.2 mg/dL vs. 5.50±1.5 mg/dL, P<0.001); the number of patients showing hyperuricemia was also higher in the adult group than in the child group (3.1% vs. 12.1%, P=0.042) (Table 2).

Table 2.

Laboratory Results between Child and Adult Group

Mycoplasma antibody was detected in 34.4% of cases in the child group. The positive rate of polymerase chain reaction (PCR) for respiratory viruses were 8.8%, adenovirus, RSV type A, and rhinovirus were detected in child.

High ASO (24.7% vs. 2.9%, P<0.001) was more common in the child group. The level of high IgA (30.0% vs. 50.0%, P <0.001) was greater in the adult group than that in the child group.

4. Univariate and multivariate logistic regression and odd ratio for renal involvement and outcomes

A univariate logistic regression analysis was used to analyze several factors that might influence renal involvement. Of these, the factors with a statistically significant effect on renal involvement were age (OR 1.09, P=0.004 in the univariate analysis OR 1.10, P=0.045 in the multivariate analysis), high uric acid (OR 10.50, P=0.034 in the univariate analysis OR 5.03, P=0.041 in the multivariate analysis in child group). There was no significant risk factor on renal involvement in adult group (Table 3).

Table 3.

Results of Univariate and Multivariate Logistic Regression Analysis of the Risk of Renal Involvement in HSP Patients

Another univariate logistic regression analysis was used to analyze the factors that might influence renal outcomes. The factors significantly associated with renal outcomes in this analysis were then subjected to an additional multivariate logistic regression analysis. The child group were presented nephrotic syndrome (OR 258.9, P=0.013 in the univariate analysis and OR 249.71, P=0.016 in the multivariate analysis). And adult group show gross hematuria (OR 49.3, P=0.014 in the univariate analysis and 100.63, P=0.006 in the multivariate analysis) (Table 4).

Table 4.

Results of Univariate and Multivariate Logistic Regression Analysis of the Risk of Poor Renal Outcome in HSP Patients

Discussion

Since this study included a total of 469 HSP patients, it can be considered a large-scale study as compared with previous research. With respect to the clinical observations made involving adults and children from several prior studies on fewer patients, the present study generally shows a greater difference in prognoses [11,12].

Among predisposing factors in the present study, the incidence of URI was significantly high in the child group. This is similar to trends seen in previous studies. The etiology of URI was only tested in some child patients in this study; adenovirus, RSV type A, and rhinovirus were detected in one case each, and the overall virus positive ratio was 8.8%. There was a suspicion that these infections were preceded, but there was no relation with the rate of kidney involvement. According to previous studies, the main causes of URI are viruses (parvovirus, rhinovirus) [13] and bacteria (staphylococci, streptococci) [14].

Medication with drugs prior to HSP onset was only common in adult patients, and that didn’t show an effect on renal prognosis in our study. The drugs taken included colchicine, dapsone, triamcinolone, and antidiabetic/antihypertensive drugs. In previous studies, quinolones [15] and clarithromycin [16] were reported to be associated with HSP. Cancer was observed in only three cases involving adult patients; the types of cancer were Hodgkin’s lymphoma, adenocarcinoma, and ovarian cancer, and these had an effect on renal prognosis. Other studies have reported lung (nonsmall-cell) cancer [17], multiple myeloma [18], prostate cancer [19], Hodgkin's lymphoma [20] and non-Hodgkin's lymphoma [21].

In terms of purpura distribution, a study by Kang [11] reported significantly higher involvement of the upper and lower limbs in the adult group; the present study also found that purpura was not limited to the lower limbs in 40.5% of cases for adults (vs. 14.4% for children). Joint pain was known to be more frequent in children in previous studies, and was also frequently observed in child patients in the present study. GI symptoms did not show any major differences between in the child vs. the adult patients in previous studies. Our study also found no differences in overall GI symptoms. But in terms of specific symptoms, we observed only abdominal pain was common in child patients, whereas abdominal pain with hematochezia more frequently in adult patients.

Renal involvement itself is a controversial issue, but generally previous studies have shown a pattern of more involvement in adults. Carlos [5] and Hung et al. [11] reported that only microscopic hematuria and renal insufficiency were more common in adults. In contrast, the present study found that microscopic hematuria had not difference of frequency, whereas gross hematuria, nonnephrotic proteinuria, and nephrotic syndrome were more common in adults. Previous studies documented the risk factors of renal involvement in child were increased level of plasma creatinine and hypertension [22]. Whereas we found that risk factors for renal involvement include older age and high level of serum uric acid in the child group. The adult group didn’t show any risk factors for renal involvement in this study.

Renal outcomes showed significantly better prognoses for children (89.5% vs. 56%, P<0.001) than for adults. Previous studies have reported proteinuria, hypertension, and anemia as major negative factors [5,11,23]. Other studies asserted that the risk factors of renal outcomes were presence of nephrotic syndrome, severe case of histologic findings of kidney in child group [24,25]. Butani and Morgenstern claimed that no apparent risk factor of poor renal outcomes in children [26]. Other risk factors reported to be associated with worse prognosis in adults are severe renal symptoms at onset, female sex (during and after pregnancy), and increased serum creatinine level and proteinuria on renal biopsy [27-29]. We found that factors affecting persistent or proceeding renal dysfunction were nephrotic syndrome in child group and gross hematuria in adult group.

A previous study demonstrated that hyperuricemia in IgA nephropathy patients was a predictive factor for progressive nephropathy and that allopurinol could improve hypertension [30]. Our study show hyperuricamia was risk factor of renal involvement in child group, but that was not relation to poor renal outcomes. We postulate that the reason for the difference between the results obtained in this study and those reported in previous studies is that we performed uric acid studies in fewer patients in the adult group and the prognosis of renal outcomes in the child group was generally better than that in the adult group.

There have been reports that blood IgA level is elevated in more than half of HSP patients, and that renal involvement is associated with higher IgA [31-33]. In our study, rates of increased IgA level were 32.2% on all patients and we also found a clear increase in mean IgA levels in the adult group. The rate of kidney involvement was also more common in adults, that was consistent with previous studies. But high IgA level was not risk factor of renal involvement in the multivariate analysis.

Studies on complements in children have reported that some patients show a temporary decrease in C3, but that complement levels are not a prognostic factor for renal involvement [34,35]. Kawasa et al. [36] reported that terminal complement complex showed increased levels in the blood in 83% of patients in the acute phase, and that it was associated with relapse during follow-up, but that C3 and C4 showed normal or elevated levels. However, there have also been reports that it is impossible to prove that complement activation is involved in the pathogenic mechanism of HSP [37]. We found that increased C4 was more common than decreased C4 regardless of age. However, it was neither a risk factor for renal involvement, nor a prognostic factor for renal outcomes.

M. pneumoniae infection shows non-lung symptoms in 25% of cases, and sometimes causes a rash. M. pneumoniae IgM-positive HSP patients have been reported previously [38]. In the present study, 34.4% of child patients were positive for M. pneumoniae , that was more frequent than positive ratio of high ASO titer (22.4%). Nonetheless, given that other research organizations have also shown an interest in this area, we believe that further discussion is required about the potential significance of macrolide antibiotics in HSP treatment.

This study had some limitations. First, as a retrospective study, there were cases lost to follow-up. Second, the treatment of HSP was mostly conducted at the Department of Pediatrics for child patients, and at various departments for adult patients. Hence, the level of consistency in examinations and medication administration is not clear, and so the comparison between child patients and adult patients is not perfect. Third, the small number of adult cases as compared with the number of child cases could have caused statistical bias.

In conclusion, in our study, HSP showed different clinical patterns in children and in adults. Risk factors for renal involvement were age and high level of uric acid in child group. Risk factors for renal outcomes were nephrotic syndrome in the children group, and gross hematuria in the adult group. In addition, further long-term follow-up observations and large-scale studies are required in the future to investigate the relationship between M. pneumoniae and HSP.

Article information Continued