G Model BONSOI-4129; No. of Pages 7
ARTICLE IN PRESS Joint Bone Spine xxx (2015) xxx–xxx
Available online at
ScienceDirect www.sciencedirect.com
Original article
Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa Etienne Merlin a,b , Richard Mouy c , Bruno Pereira a , Luc Mouthon d,e , Aurélie Bourmaud f , Jean-Charles Piette g , Judith Landman-Parker h , Patricia Chellun i , Mustapha Layadi j , Caroline Thomas k , Loïc Guillevin d,e , Anne-Marie Prieur c , Pierre Quartier c,e,l,∗ a
Service de Pédiatrie, CHU Clermont-Ferrand, Clermont-Ferrand, France Inserm CIC 1405, Centre de Recherche Chez l’Enfant, Clermont-Ferrand, France c Unité d’Immunologie, Hématologie et Rhumatologie Pédiatrique, Assistance Publique-Hôpitaux de Paris (AP-HP), Hôpital Necker-Enfants Malades, Paris, France d Service de Médecine Interne, Hôpital Cochin, Centre de Référence pour les Vascularites Nécrosantes et la Sclérodermie Systémique, AP-HP, Paris, France e Université Paris-Descartes, Paris, France f Département de Santé Publique, Institut Cancérologique de la Loire, Saint-Étienne, France g Service de Médecine Interne, Université Pierre-et-Marie-Curie, Paris-6, Hôpital Pitié-Salpêtrière, AP-HP, Paris, France h Hôpital Trousseau, AP-HP, Paris, France i Service de Pédiatrie Générale, CHU Niort, France j Service de Pédiatrie, hôpital Mère-Enfant, CHU de Limoges, Limoges, France k Service d’Hématologie-Oncologie, Hôpital Mère-Enfants, Nantes, France l Institut Hospitalo-Universitaire IMAGINE, Paris, France b
a r t i c l e
i n f o
Article history: Accepted 19 January 2015 Available online xxx Keywords: Necrotizing vasculitis Criteria Classification Prognosis Relapse
a b s t r a c t Objective: To assess the prognostic impact of clinical presentation in children with polyarteritis nodosa (PAN). Methods: Children diagnosed between 1986 and 2006 in a tertiary care pediatric rheumatology center were classified as “cutaneous PAN” (group 1), “cutaneous PAN with significant extra-cutaneous features” (group 2) or “visceral childhood PAN” (group 3). Outcome measures: (1) clinical remission off-therapy at last follow-up, (2) requirement and length of glucocorticoid therapy, (3) presence of disease-related sequelae. Results: Twenty-nine children were included. Sixteen met the Ankara criteria for PAN. Nine patients were qualified as group 1, 11 as group 2, and 9 as group 3. At last follow-up, 15 children were in clinical remission off-therapy: 4 from group 1 (44%), 4 from group 2 (36%) and 7 from group 3 (78%). Glucocorticoid therapy was required for 8 (89%), 7 (64%) and 7 (78%) patients from groups 1, 2 and 3, respectively. Seven children did not require any glucocorticoid therapy. Time-dependent probability of achieving glucocorticoid-free clinical remission was similar between the three groups. Three patients (one from each group) had digital ischemia leading to amputation. There were no significant between-group differences in outcome based on the three outcome measures addressed. Conclusion: Outcome was not strikingly predictable from initial presentation in children with PAN. The organ distribution-based distinction between cutaneous and visceral PAN had little prognostic power in this series. © 2015 Published by Elsevier Masson SAS on behalf of the Société Française de Rhumatologie.
1. Introduction
Abbreviations: CRP, C-reactive protein; ENT, ear-nose-throat; ESR, erythrocyte sedimentation rate; HBV, hepatitis B virus; IVIG, intravenous immunoglobulin; NSAIDs, non-steroidal anti-inflammatory drugs; PAN, polyarteritis nodosa. ∗ Corresponding author. Unité d’Immunologie-Hématologie et Rhumatologie Pédiatrique, Hôpital Necker-Enfants Malades, Paris; Tel.: +33 1 44 49 48 28. E-mail address: [email protected] (P. Quartier).
Polyarteritis nodosa (PAN) was first described as a lethal condition characterized by diffuse necrotizing vasculitis that predominantly affects medium-sized arteries [1]. A separate distinction was later made for cutaneous PAN, characterized by a chronic relapsing benign course and an absence of systemic involvement [2]. Although the cutaneous features are similar, there
http://dx.doi.org/10.1016/j.jbspin.2015.01.007 1297-319X/© 2015 Published by Elsevier Masson SAS on behalf of the Société Française de Rhumatologie.
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
G Model BONSOI-4129; No. of Pages 7
ARTICLE IN PRESS E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
2
is emerging consensus that these two diseases should be considered different, since progression from cutaneous to visceral PAN appears extremely rare [2]. The clinical spectrum of pediatric-onset PAN varies from severe visceral involvement to less aggressive disease with cutaneous and subcutaneous involvement and milder systemic symptoms [3,4]. The 2009 Ankara classification criteria offers a valuable framework to distinguish PAN from other primary vasculitides but is unable to distinguish “cutaneous PAN” from “visceral PAN” [5]. It is unclear whether intermediate forms with predominant cutaneous features associated with mild extra-cutaneous signs (such as marked weight loss, abdominal pain or headaches) are to be considered visceral or cutaneous in terms of prognosis and therapeutic approach. We therefore conducted a retrospective study designed to assess whether disease characteristics at diagnosis have the power to predict long-term patient outcome. 2. Methods Retrospective analysis was led on the medical records of patients diagnosed with PAN before age 18 years between January 1986 and December 2006 and followed in a French tertiary care pediatric rheumatology center at Hôpital Necker-Enfants Malades, Paris. Diagnosis of PAN relied on the association of a clinical presentation compatible with a systemic inflammatory process involving ischemic manifestations and histological small or middle-sizedvessel involvement after ruling out other vasculitides. Patients were excluded if they presented with diffuse alveolar hemorrhage, pauci-immune glomerulonephritis, histological evidence of capillary and/or venular vasculitis and presence of anti-neutrophil cytoplasm antibodies (ANCA) specific for proteinase 3 or myeloperoxidase detected by ELISA, granuloma or IgA deposition. The study was approved by the Rhône-Alpes-Auvergne inter-regional institutional review board. 2.1. Data collection Clinical data collected included fever, constitutional symptoms (general malaise, fatigue, weight loss), evidence of specific organ involvement (as defined in the Ankara criteria) [6] such as high blood pressure, renal involvement, peripheral neuropathy, muscle or skin involvement – for all other organs, we collected clinical symptoms compatible with vasculitis. Ear-Nose-Throat (ENT) manifestations preceding (within 21 days) or accompanying PAN were also recorded. We also recorded what other diagnoses were considered. The laboratory data were: complete blood count and differential, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), cryoglobulins and serological hepatitis B virus (HBV) status. All imaging and/or histological investigations performed during follow-up were collected. Any abnormalities were reviewed by an expert clinician to rule them in or out as PAN-related. All treatments were systematically recorded. 2.2. Definitions Visceral involvement was defined as involvement of at least one of the following organs: nervous system, gastro-intestinal tract (documented by histology, endoscopy and/or computed tomography scan), testis (clinically or by histology), heart, lungs, kidney, or any other organ. Functional complaints (abdominal pain, headache, myalgia, arthralgia) and arthritis were not considered as visceral involvement. PAN was classified as “cutaneous PAN” in the absence of visceral involvement, marked weight loss (≥ 10% of body weight), abdominal pain or headache (group 1), “cutaneous PAN with significant extra-cutaneous features” if no visceral involvement was proven but marked weight loss, abdominal pain and/or headache
was present (group 2), and “visceral childhood PAN” if visceral involvement was present (group 3). Histology was considered as diagnostic in cases where fibrinoid necrosis of the arterial wall was straightforward, and non-diagnostic in cases where there was an inflammatory infiltrate composed predominantly of polymorphonuclear and lymphocytic cells within the medium-sized artery wall but without clear fibrinoid necrosis. Based on the data collected, the patients were re-evaluated according to the Ankara criteria [5]. Patient status at the latest follow-up was assessed as follows: clinical remission off-therapypatients presented neither clinical nor physical symptoms for at least 6 months after discontinuation of therapy; treatmentdependency-patients presented neither clinical nor physical signs but were still under therapy; active-disease-patients presented with clinical evidence of active disease. Relapse was defined as a resurgence of one or more clinical symptoms in a patient who had previously been free of any manifestation of the disease [7]. 2.3. Outcome Three main outcome measures were considered: • treatment-free clinical remission at last follow-up; • requirement and length of glucocorticoid therapy; • presence of disease-related sequelae. 2.4. Statistics First, we ran a univariate comparison between the 3 groups. Next, we ran a univariate analysis for each of the five outcome parameters in an attempt to identify prognostic factors. Analyses were led using Fisher’s exact test for qualitative variables or a nonparametric Kruskall-Wallis test for quantitative variables. Given the small number of patients, we did not run multivariate analyses. No adjustment was made for analyzing multiple outcomes [8]. All analyses were performed on R software (version 2.13.0) using the Epicalc package. 3. Results A total 29 children (20 girls and 9 boys) were included in the study. Median age at diagnosis was 6.9 years (range: 2–14). 3.1. Initial features Eight patients had an ENT infection 5 to 22 days prior the first signs of PAN. At first presentation, all children were febrile. The most frequent features involved: skin (24 children), joints (n = 21), muscles (n = 12), gastro-intestinal tract (n = 10), ENT (n = 9), and nervous system (n = 7). Other features are shown in Table 1. The delay between first signs and diagnosis ranged from 5 to 1580 days (median: 37 days). Before PAN was diagnosed, four patients were misdiagnosed as systemic-onset juvenile idiopathic arthritis, 3 as acute rheumatoid fever, 3 as Kawasaki disease, and one as systemic lupus erythematosus. Reason for delayed diagnosis > 30 days was: absence of initial cutaneous involvement (7 children), glucocorticoids or IVIG administered before skin biopsy (6 children), or spontaneous transient remission (1 child). HBV serology was negative in all patients. At least one biopsy was performed in all patients: in 16 cases, histological examination showed diagnostic PAN lesions with destruction of the mediumsized artery walls and fibrinoid necrosis, and in 13 patients the biopsies were non-diagnostic. Angiography was performed in two
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
G Model BONSOI-4129; No. of Pages 7
ARTICLE IN PRESS E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
3
Table 1 Patient characteristics at the time of diagnosis. n = 29 Age at onset (years) Sex ratio (M/F) Fever Cutaneous involvement Nodules Livedo reticularis Necrosis Purpura Muscle involvement Joints (arthralgia or arthritis) Hypertension Renal involvement Nervous system Peripheral Central Abdominal pain Testicular involvement ENT involvement Laboratory Leukocytes (x109 .L−1 ) Neutrophils (x109 .L−1 ) Platelets (x109 .L−1 ) ESR (mm.h−1 ) CRP (mg.L−1 ) 1992 diag. criteria fulfilled Ankara criteria fulfilled
6.9 ± 3.7 (2–14) 9/20 29 (100) 24 (82) 14 (48) 9 (31) 3 (10) 7 (24) 12 (41) 21 (72) 1 (3) 2 (7) 6 (21) 1 (3) 10 (34) 2 (22) 9 (31) 17.9 ± 8.7 (6.1–38) 14 ± 8.5 (3.8–31) 510 ± 239 (199–999) 75 ± 30 (10–160) 139 ± 73 (6–320) 4 (14) 16 (55)
ENT: ear-nose-throat; ESR: erythrocyte sedimentation rate; M: male; F: female; CRP: C-reactive protein. Quantitative data are expressed as means ± SD (range), and qualitative data are expressed as n (%).
patients (one cerebral angiography, one mesenteric angiography) due to seizures and severe abdominal pain, and neither showed any arterial abnormalities. Sixteen patients met the Ankara criteria. The 13 patients not fulfilling the Ankara criteria met clinical criteria but missed the histological diagnostic criterion and had not been angiographied. Clinical and biological presentation did not differ significantly according to whether or not patients met the Ankara criteria.
3.2. Comparison of children diagnosed as cutaneous PAN (group 1), cutaneous PAN with significant extra-cutaneous features (group 2), and visceral PAN (group 3) Eleven children initially presented as cutaneous PAN and 13 as cutaneous PAN with significant extra-cutaneous features. Two patients in each group subsequently developed visceral involvement at 42, 59, 66 and 398 days post-diagnosis, respectively, and were ultimately considered as visceral childhood PAN. Finally, 9 patients qualified as group 1, 11 as group 2, and 9 as group 3. The characteristics of the 3 groups are reported in Table 2. CRP and ESR differed significantly between the three groups (Table 2).
3.3. Treatment As first-line therapy, 18 patients were given glucocorticoids, 8 were given NSAIDs, and three were given IVIG. At last follow-up, 7 patients had received no glucocorticoids but only NSAIDs ± IVIG (Table 3), including two that qualified as group 3, 16 children had received glucocorticoids ± IVIG, and 6 had received glucocorticoids + cyclophosphamide with or without other therapies (adalimumab, rituximab, plasma exchange, cyclosporine, azathioprine, colchicine).
Figure 1. Cumulative incidence of glucocorticoid-free complete remission. Group 1: myalgia ± arthralgia, but absence of visceral involvement apart from mild abdominal pain and headache; Group 2: myalgia ± arthralgia + marked weight loss, abdominal pain and/or headache without proven visceral involvement; Group 3: proven visceral involvement.
3.4. Outcome Median follow-up was 70 months (range: 7–178) and was 62, 59 and 72 months for group 1, 2 and 3 respectively (P = 0.5). Median number of relapses was 1 (range: 0–10). One patient (from group 3) died of colonic perforation 10 years after the onset of severe relapsing PAN. 3.4.1. Clinical remission off-therapy At last follow-up, 15 children were in clinical remission offtherapy (no glucocorticoid and no other therapies): 4 from group 1 (44%), 4 from group 2 (36%) and 7 from group 3 (78%), P = 0.4 (Table 3). 3.4.2. Glucocorticoid therapy Glucocorticoid therapy was required for 8 (89%), 7 (64%) and 7 (78%) patients from groups 1, 2 and 3, respectively (P = 0.5). Seven patients did not require any corticoid therapy. In the other 22 cases, median duration of glucocorticoid therapy was 169, 83 and 132 months, respectively (P = 0.3). Maximal dose of oral glucocorticoids was 2, 1.6 and 1.5 mg/kg/day, respectively (P = 0.5), whereas dose 9 months after diagnosis was 0.16 mg/kg/d (range: 0.05–2), 0.4 mg/kg/d (range: 0.1–2) and 0.3 mg/kg/d (range: 0.1–1), respectively (P = 0.6). Cumulated dose of oral glucocorticoids was 777 mg/kg (range: 212–3270), 651 mg/kg (range: 315–2635) and 551 mg/kg (range: 60–2837), respectively (P = 0.7). Time-dependent probability of achieving glucocorticoidfree clinical remission was similar between the 3 groups (Figure 1). 3.4.3. Sequelae Three patients (one from each group) presented digital ischemia leading to amputation 1, 5 and 8 years after onset of the disease. All three children presented a histological hallmark of arterial necrosis. Overall, there were no significant differences between the three groups on the three selected outcome measures (Table 3). 3.5. Prognostic factors None of age, sex, PAN subgroup, Ankara criteria fulfillment, visceral involvement or leukocyte count pointed to a higher probability of glucocorticoid-free clinical remission.
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
G Model BONSOI-4129; No. of Pages 7
ARTICLE IN PRESS E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
4 Table 2 Characteristics of the three groups.
N Sex (M/F) Age (years) Typical histological findings Previous ENT infection Muscle involvement HTA Renal involvement Peripheral neuropathy Cutaneous features Livedo Nodules Necrosis Purpura Abdominal pain ENT involvement Mucosal involvement Joint Chills Headache Laboratory exams Leukocytes (x10e9/L) PNN (x10e9/L) Platelets (x10e9/L) ESR (mm/h) CRP (mg/l) ASLO > 600 UI/L 1992 diag. criteria fulfilled Ankara criteria fulfilled
Group 1 9
Group 2 11
Group 3 9
1/8 6 (3–13) 5 (56) 4 (44) 2 (22) 0 0 0
3/8 8 (4–13) 6 (54) 3 (28) 7 (63) 0 0 3 (27)
5/4 5 (2–14) 5 (56) 1 (11) 3 (33) 1 (11) 2 (22) 3 (33)
3 (33) 8 (89) 1 (11) 3 (33) 0 5 (56) 1 (11) 8 (89) 0 0
3 (27) 7 (64) 0 3 (27) 7 (64) 2 (18) 0 9 (82) 3 (27) 4 (36)
3 (33) 5 (56) 2 (22) 1 (11) 3 (33) 2 (22) 3 (33) 4 (44) 0 1 (11)
17.2 (7.5–33.1) 10.8 (4.5–31.0) 512 (309–990) 57 (10–78) 63 (6–150) 4 (44) 0 5 (56)
17.0 (6.1–31.5) 16.0 (4.3–30.0) 409 (199–905) 84 (59–101) 132 (59–320) 2 (18) 2 (18) 6 (54)
13.5 (7.3–38.0) 7.6 (3.8–25.0) 439 (307–999) 90 (32–160) 152 (113–270) 0 2 (22) 5 (56)
P
0.03 0.05
ASLO: anti-streptolysin antibodies; ENT: eye-nose-throat; ESR: erythrocyte sedimentation rate; M: male; F: female; CRP: C-reactive protein; PNN: polynuclear neutrophils. Group 1: myalgia ± arthralgia but absence of visceral involvement apart from abdominal pain and headache; Group 2: marked weight loss, abdominal pain and/or headache; Group 3: proven visceral involvement. P-values are shown only if ≤ 0.05. Quantitative data are expressed as median (range), and qualitative data are expressed as n (%).
Patients requiring glucocorticoids had higher leukocyte counts at diagnosis (19.5 ± 8 × 109 /L−1 [range: 7–38] vs 13.2 ± 6 × 109 /L−1 [range: 6–20], respectively, P = 0.06) and higher neutrophil counts (15.6 ± 8 × 109 /L−1 vs 9.3 ± 5 × 109 /L−1 , respectively, P = 0.06) than other patients, and fulfilled Ankara criteria in 14 out of 22 cases versus 2 out of 7 patients who remained glucocorticoid-free (odds ratio 4.38 at 95% CI [range: 0.53–53], P = 0.1). There were no differences on other variables. None of the studied parameters correlated with duration of glucocorticoid therapy or number of relapses. All three patients with ischemic disease-related sequelae fulfilled the Ankara criteria.
4. Discussion In this retrospective survey of 29 patients diagnosed with childhood-onset PAN from a single tertiary care pediatric rheumatology center, patient outcome could not be predicted by a histological necrotizing feature nor by initial disease presentation as none of the initial clinical or biological parameters correlated with patient status at last follow-up. Patients were fairly evenly spread across three groups at diagnosis – a group with visceral involvement, another group with no visceral involvement but marked constitutional symptoms (weight loss, abdominal pain or
Table 3 Treatments and outcome measures.
N Treatments Oral glucocorticoids Bolus methylprednisolone IVIG Cyclophosphamide Other Outcome Follow-up (months) Status Dead Active disease CR on therapy Treatment-free CR Sequelaea Glucocorticoid therapy duration (months) Number of relapses (mean, [median], range)
Group 1 9
Group 2 11
Group 3 9
8 (89) 0 (0) 6 (67) 2 (22) 0
7 (64) 1 (9) 4 (36) 2 (18) 0
7 (78) 2 (22) 4 (44) 2 (22) 2 (22)
62 (25–178)
50 (19–119)
72 (7–157)
0 2 (22) 3 (33) 4 (44) 1 (11) 69 (0–178) 2.4 [1] (0–5)
0 3 (27) 4 (36) 4 (36) 1 (9) 26 (0–119) 1.5 [0] (0–5)
1 (11) 0 1 (11) 7 (78) 1 (11) 39 (0–130) 2.2 [1] (0–10)
CR: clinical remission. Group 1: myalgia ± arthralgia, but absence of visceral involvement apart from mild abdominal pain and headache; Group 2: myalgia ± arthralgia + marked weight loss, abdominal pain and/or headache without proven visceral involvement; Group 3: proven visceral involvement. a Digital necrosis leading to amputation in the 3 cases, as well as diffuse lesions of the brain and testis in the patient from group 3
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
G Model
BONSOI-4129; No. of Pages 7
Organ involvement Author
No pts
PAN denomination
Definition used
Fever
Skin
HTA
Kidney
Joint
GIT
Muscle
Sore throat
Nervous system
Testis
CRP
ESR
ASLO
Ste
Cy
Relapsing pts
Evolution
1993
David
12
Associated with streptococcus
12
12
0
2
12
10
8
12
0
0
55
NA
–
10
2
7
2 proven visceral
1995
Kumar
10
Benign cutaneous
NA
NA
2
–
7
–
NA
NA
NA
–
NA
NA
NA
10
2
NA
7 autoamputations
1997 1998
Till Mocan
2 1
Cutaneous Cutaneous
Vasculitis associated with streptococcus infection Absence of systemic involvement No Absence of major organ system involvement
1 1
2 1
0 0
– –
2 1
1 0
1 0
2 0
0 0
0 0
NA NA
50 NA
2 1
1 1
0 0
2 0
0 0
2002 2004
Bauza Ozen
1 33
Cutaneous Cutaneous
1 14
1 33
0 0
– –
13
1 0
0 5
1 12
0 1
0 0
NA 37
52 60
1 NA
1 25
1 5
0 4
0 0
2004
Ozen
63
Systemic
36
58
27
NA
25
15
45
NA
14
4
NA
NA
NA
63
42
5
0 deaths
2005
Fathalla
4
Cutaneous
4
4
0
–
3
0
NA
2
0
–
NA
NA
3
4
1
3
0
2013
Elfetheriou 69
60
61
11
13
52
28
57
NA
7
4
56
64
21
69
57
24
3 deaths
Systemic
Confined to the skin with no organ involvement Organ involvement, regardless of vessel size. Constitutional symptoms and elevated acute-phase reactants Small and middle-sized vessels of the skin, muscles and joints EULAR PReS, cutaneous PAN excluded
ARTICLE IN PRESS
Year
E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
Table 4 Published series of childhood PAN.
5
G Model BONSOI-4129; No. of Pages 7 6
ARTICLE IN PRESS E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
headache) and a third group without any of these features – yet digital necrosis leading to amputation occurred in each group and there was no striking difference between the groups in terms of glucocorticoid usage or rate of glucocorticoid-free clinical remission. The rationale behind analyzing three groups of patients according to initial presentation was that the distinction between “cutaneous” and “visceral” PAN remains hazy, and intermediate forms could have been hard to classify. The clinical heterogeneity of the same disease is illustrated by the recently described monogenic form of PAN-related to an impaired function of adenosine deaminase 2, where within a same family, the homozygous carriers of the mutation have either cutaneous or visceral PAN [9]. Since cutaneous and visceral PAN share the same histopathological pattern, differentiating them is a challenge that can only be resolved based on clinical and imaging assessment at presentation and over the course of the disease [10]. This distinction is challenging for at least three reasons. First, the spectrum of cutaneous PAN, which includes fever, myalgia, arthralgia, non-erosive arthritis and, in some cases, peripheral neuropathy, is included within the spectrum of visceral PAN [11]. In adults, cutaneous PAN likely appears as an indolent disease, and some authors limit the definition of cutaneous PAN to cutaneous involvement without fever, myalgia, arthralgia or peripheral neuropathy (except if present at the same localization as skin lesions) [12]. In contrast, a review of 119 children with cutaneous PAN found that most children had fever and elevated acute-phase reactants, and that musculoskeletal features, joint involvement and peripheral neuritis were frequent (43%, 58% and 18%, respectively) [13]. Second, cutaneous PAN is frequently though to be associated with streptococcus. An infectious trigger is observed in many cases of either cutaneous or visceral PAN, but in a setting of systemic inflammation the diagnostic value of elevated ASLO is not straightforward for streptococcus infection. A high titer of ASLO or a suspected streptococcus infection can be associated with systemic PAN too [14,15]. Third, cutaneous PAN is thought to have a benign course. However, some cases of cutaneous PAN can prove severe due to digital necrosis and auto-amputations and will subsequently require glucocorticoids and/or other immunosuppressants (Table 4) [16,17]. In our series, four out of 20 cases without visceral involvement needed cyclophosphamide and two ultimately had finger amputations, whereas two patients with biopsy-proven visceral PAN did not require any glucocorticoid therapy. One major conceptual difference between cutaneous PAN and visceral PAN is that only visceral PAN carries a risk of vital organ involvement [4,15,18]. However, the heterogeneity of prognosis beyond organ involvement-based classifications prompts us to think that, within the PAN framework defined by the Ankara classification, prognosis should be evaluated based on specifically-validated tools, such as a “pediatric five-factor score” [19]. It has recently been reported that severe gastro-intestinal involvement correlates to a higher relapse rate [15]. Gastro-intestinal involvement is known to be a poor prognosis factor in adults with PAN [20]. Our sample population is too small to confirm these results, but note that the only patient who died had severe GI ischemia. Among the patients selected here with a diagnosis of pediatric PAN, several did not fulfill the recent Ankara criteria [11], yet showed no significant difference in clinical presentation and outcome to the other patients. The histological hallmark of PAN is the presence of fibrinoid necrosis with complete destruction of medium- and sometimes small-sized artery walls, but in some cases this typical pattern is missing or incomplete [3]. Angioimaging may have allowed some of the patients who missed the histological criterion to fulfill the Ankara criteria, but few patients
were angiographied. Even so, Ankara criteria were more sensitive in our series than the ACR criteria which were first designed in adults [21] then re-adapted to pediatric PAN [22]. All our patients had a clinical presentation that was compatible with Ankara criteria. Histological assessment was performed in all cases but was not always typical, likely related to the fact that PAN is a patchy disease. Consequently, and to better assess disease severity, there is a strong rationale for making wider use of angiographic imaging procedures in children with a suspicion of PAN. Arteriography is usually considered the most powerful imaging exam [23] but is not frequently performed, especially in moderate forms. Other imaging approaches considered for the Ankara classification include CT scan [24], angio-CT scan or angio-MRI. Although less sensitive than arteriography [23], these non-invasive approaches are key to the assessment of a systemic vasculitis and are now gaining in popularity. The treatment of childhood PAN is not well codified. Note that 7 of our patients received neither glucocorticoids nor immunosuppressants. Several patients first received IVIG due to a suspicion of Kawasaki disease and went on to show dramatic but transient improvement. It is unclear whether IVIG usage had a glucocorticoid-sparing effect in some patients [25]. Cyclophosphamide has been shown to reduce the number of relapses [15], but in our series it was only administered in a few patients with a severe disease course. We posit that the significant effect of immunosuppressive drugs on disease activity in several patients here can explain the absence of difference in outcome between patients with visceral forms and patients with non-severe relapsing disease. Given that most patients had been exposed to glucocorticoids for relatively long periods of time, a more modern approach with more frequent early introduction of immunosuppressive drugs might be worth considering. Our study has several limitations. First, the number of patients was too small to allow a fine-grained comparison between different groups of patients, and there may have been a recruitment bias towards more severe forms in our center. Second, the assessment of long-term outcome warrants caution in this retrospective study, as patients were included over a 20-year period in which therapeutic strategies may have differed over time. Nevertheless, this study still represents one of the largest series of pediatric PAN, and median follow-up was relatively long. In conclusion, this study failed to find evidence that initial presentation has predictive value for outcome in patients with pediatric-onset PAN. In particular, when considering a looser definition of “cutaneous PAN”, as also proposed in many other publications [11,13,16,26], we found a few cases where children with “cutaneous PAN” and no visceral involvement went on to develop severe complications, such as digital necrosis, ending up with extended periods of glucocorticoid therapy and immunosuppressive drugs and having long-lasting disease activity. We conclude that the organ distribution-based distinction between cutaneous and visceral childhood PAN had little prognostic power in this series.
Contributors Etienne Merlin, Richard Mouy and Pierre Quartier conceptualized and designed the study, drafted the initial manuscript, and approved the final manuscript as submitted. Bruno Pereira and Aurélie Bourmaud carried out the initial analyses, reviewed and revised the manuscript, and approved the final manuscript as submitted. Luc Mouthon, Jean-Charles Piette, Judith Landman-Parker, Patricia Chellun, Mustapha Layadi, Caroline Thomas, Loïc Guillevin and Anne-Marie Prieur coordinated and supervised the data
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
G Model BONSOI-4129; No. of Pages 7
ARTICLE IN PRESS E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
collection, critically reviewed the manuscript, and approved the final manuscript as submitted. Disclosure of interest The authors declare that they have no conflicts of interest concerning this article. References [1] Kussmaul ARM. Uber eine bisher nicht beschreibene eigenthümliche Arterienerkankung (Periarteriitis nodosa), die mit Morbus Brightii und rapid fortschreitender allgemeiner Muskellähmung einhergeht. Deutsches Arch F Klin Med 1866;1:84–517. [2] Chen KR. Cutaneous polyarteritis nodosa: a clinical and histopathological study of 20 cases. J Dermatol 1989;16:429–42. [3] Dillon MJ, Eleftheriou D, Brogan PA. Medium-size-vessel vasculitis. Pediatr Nephrol 2010;25:1641–52. [4] Ozen S, Anton J, Arisoy N, et al. Juvenile polyarteritis: results of a multicenter survey of 110 children. J Pediatr 2004;145:517–22. [5] Ozen S, Pistorio A, Iusan SM, et al. EULAR/PRINTO/PRES criteria for HenochSchönlein purpura, childhood polyarteritis nodosa, childhood Wegener granulomatosis and childhood Takayasu arteritis: Ankara 2008. Part II: Final classification criteria. Ann Rheum Dis 2010;69:798–806. [6] Ruperto N, Ozen S, Pistorio A, et al. EULAR/PRINTO/PRES criteria for HenochSchönlein purpura, childhood polyarteritis nodosa, childhood Wegener granulomatosis and childhood Takayasu arteritis: Ankara 2008. Part I: Overall methodology and clinical characterisation. Ann Rheum Dis 2010;69:790–7. [7] Hellmich B, Flossmann O, Gross WL, et al. EULAR recommendations for conducting clinical studies and/or clinical trials in systemic vasculitis: focus on anti-neutrophil cytoplasm antibody-associated vasculitis. Ann Rheum Dis 2007;66:605–17. [8] Rothman KJ. No adjustments are needed for multiple comparisons. Epidemiology 1990;1:43–6. [9] Navon Elkan P, Pierce SB, Segel R, et al. Mutant adenosine deaminase 2 in a polyarteritis nodosa vasculopathy. N Engl J Med 2014;370:921–31. [10] Eleftheriou D, Brogan PA. Vasculitis in children. Best Pract Res Clin Rheumatol 2009;23:309–23.
7
[11] Ozen S, Ruperto N, Dillon MJ, et al. EULAR/PReS endorsed consensus criteria for the classification of childhood vasculitides. Ann Rheum Dis 2006;65: 936–41. [12] Nakamura T, Kanazawa N, Ikeda T, et al. Cutaneous polyarteritis nodosa: revisiting its definition and diagnostic criteria. Arch Dermatol Res 2009;301: 117–21. [13] Bansal N-K, Houghton KM. Cutaneous polyarteritis nodosa in childhood: a case report and review of the literature. Arthritis 2010;2010:687547. [14] David J, Ansell BM, Woo P. Polyarteritis nodosa associated with streptococcus. Arch Dis Child 1993;69:685–8. [15] Eleftheriou D, Dillon M, Tullus K, et al. Systemic polyarteritis nodosa in the young: a single centre experience over 32 years. Arthritis Rheum 2013;65:2476–85. [16] Mocan H, Mocan M, Peru H, et al. Cutaneous polyarteritis nodosa in a child and a review of the literature. Acta Pædiatrica 1998;87:351–3. [17] Williams VL, Guirola R, Flemming K, et al. Distal extremity necrosis as a manifestation of cutaneous polyarteritis nodosa: case report and review of the acute management of a pediatric patient. Pediatr Dermatol 2012;29:473–8. [18] Morgan AJ, Schwartz RA. Cutaneous polyarteritis nodosa: a comprehensive review. Int J Dermatol 2010;49:750–6. [19] Guillevin L, Pagnoux C, Seror R, et al. The Five-Factor Score revisited: assessment of prognoses of systemic necrotizing vasculitides based on the French Vasculitis Study Group (FVSG) cohort. Medicine (Baltimore) 2011;90:19–27. [20] Guillevin L, Lhote F, Gayraud M, et al. Prognostic factors in polyarteritis nodosa and Churg-Strauss syndrome. A prospective study in 342 patients. Medicine (Baltimore) 1996;75:17–28. [21] Lightfoot Jr RW, Michel BA, Bloch DA, et al. The American College of Rheumatology 1990 criteria for the classification of polyarteritis nodosa. Arthritis Rheum 1990;33:1088–93. [22] Ozen S, Besbas N, Saatci U, et al. Diagnostic criteria for polyarteritis nodosa in childhood. J Pediatr 1992;120:206–9. [23] Schmidt WA. Use of imaging studies in the diagnosis of vasculitis. Curr Rheumatol Rep 2004;6:203–11. [24] Adaletli I, Ozpeynirci Y, Kurugoglu S, et al. Abdominal manifestations of polyarteritis nodosa demonstrated with CT. Pediatr Radiol 2010;40:766–9, http://dx.doi.org/10.1007/s00247-009-1466-4. [25] Aries PM, Hellmich B, Gross WL. Intravenous immunoglobulin therapy in vasculitis: speculation or evidence? Clin Rev Allergy Immunol 2005;29:237–45. ˜ A, Idoate M. Cutaneous polyarteritis nodosa. Br J Dermatol [26] Bauzá A, Espana 2002;146:694–9.
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
ARTICLE IN PRESS Joint Bone Spine xxx (2015) xxx–xxx
Available online at
ScienceDirect www.sciencedirect.com
Original article
Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa Etienne Merlin a,b , Richard Mouy c , Bruno Pereira a , Luc Mouthon d,e , Aurélie Bourmaud f , Jean-Charles Piette g , Judith Landman-Parker h , Patricia Chellun i , Mustapha Layadi j , Caroline Thomas k , Loïc Guillevin d,e , Anne-Marie Prieur c , Pierre Quartier c,e,l,∗ a
Service de Pédiatrie, CHU Clermont-Ferrand, Clermont-Ferrand, France Inserm CIC 1405, Centre de Recherche Chez l’Enfant, Clermont-Ferrand, France c Unité d’Immunologie, Hématologie et Rhumatologie Pédiatrique, Assistance Publique-Hôpitaux de Paris (AP-HP), Hôpital Necker-Enfants Malades, Paris, France d Service de Médecine Interne, Hôpital Cochin, Centre de Référence pour les Vascularites Nécrosantes et la Sclérodermie Systémique, AP-HP, Paris, France e Université Paris-Descartes, Paris, France f Département de Santé Publique, Institut Cancérologique de la Loire, Saint-Étienne, France g Service de Médecine Interne, Université Pierre-et-Marie-Curie, Paris-6, Hôpital Pitié-Salpêtrière, AP-HP, Paris, France h Hôpital Trousseau, AP-HP, Paris, France i Service de Pédiatrie Générale, CHU Niort, France j Service de Pédiatrie, hôpital Mère-Enfant, CHU de Limoges, Limoges, France k Service d’Hématologie-Oncologie, Hôpital Mère-Enfants, Nantes, France l Institut Hospitalo-Universitaire IMAGINE, Paris, France b
a r t i c l e
i n f o
Article history: Accepted 19 January 2015 Available online xxx Keywords: Necrotizing vasculitis Criteria Classification Prognosis Relapse
a b s t r a c t Objective: To assess the prognostic impact of clinical presentation in children with polyarteritis nodosa (PAN). Methods: Children diagnosed between 1986 and 2006 in a tertiary care pediatric rheumatology center were classified as “cutaneous PAN” (group 1), “cutaneous PAN with significant extra-cutaneous features” (group 2) or “visceral childhood PAN” (group 3). Outcome measures: (1) clinical remission off-therapy at last follow-up, (2) requirement and length of glucocorticoid therapy, (3) presence of disease-related sequelae. Results: Twenty-nine children were included. Sixteen met the Ankara criteria for PAN. Nine patients were qualified as group 1, 11 as group 2, and 9 as group 3. At last follow-up, 15 children were in clinical remission off-therapy: 4 from group 1 (44%), 4 from group 2 (36%) and 7 from group 3 (78%). Glucocorticoid therapy was required for 8 (89%), 7 (64%) and 7 (78%) patients from groups 1, 2 and 3, respectively. Seven children did not require any glucocorticoid therapy. Time-dependent probability of achieving glucocorticoid-free clinical remission was similar between the three groups. Three patients (one from each group) had digital ischemia leading to amputation. There were no significant between-group differences in outcome based on the three outcome measures addressed. Conclusion: Outcome was not strikingly predictable from initial presentation in children with PAN. The organ distribution-based distinction between cutaneous and visceral PAN had little prognostic power in this series. © 2015 Published by Elsevier Masson SAS on behalf of the Société Française de Rhumatologie.
1. Introduction
Abbreviations: CRP, C-reactive protein; ENT, ear-nose-throat; ESR, erythrocyte sedimentation rate; HBV, hepatitis B virus; IVIG, intravenous immunoglobulin; NSAIDs, non-steroidal anti-inflammatory drugs; PAN, polyarteritis nodosa. ∗ Corresponding author. Unité d’Immunologie-Hématologie et Rhumatologie Pédiatrique, Hôpital Necker-Enfants Malades, Paris; Tel.: +33 1 44 49 48 28. E-mail address: [email protected] (P. Quartier).
Polyarteritis nodosa (PAN) was first described as a lethal condition characterized by diffuse necrotizing vasculitis that predominantly affects medium-sized arteries [1]. A separate distinction was later made for cutaneous PAN, characterized by a chronic relapsing benign course and an absence of systemic involvement [2]. Although the cutaneous features are similar, there
http://dx.doi.org/10.1016/j.jbspin.2015.01.007 1297-319X/© 2015 Published by Elsevier Masson SAS on behalf of the Société Française de Rhumatologie.
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
G Model BONSOI-4129; No. of Pages 7
ARTICLE IN PRESS E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
2
is emerging consensus that these two diseases should be considered different, since progression from cutaneous to visceral PAN appears extremely rare [2]. The clinical spectrum of pediatric-onset PAN varies from severe visceral involvement to less aggressive disease with cutaneous and subcutaneous involvement and milder systemic symptoms [3,4]. The 2009 Ankara classification criteria offers a valuable framework to distinguish PAN from other primary vasculitides but is unable to distinguish “cutaneous PAN” from “visceral PAN” [5]. It is unclear whether intermediate forms with predominant cutaneous features associated with mild extra-cutaneous signs (such as marked weight loss, abdominal pain or headaches) are to be considered visceral or cutaneous in terms of prognosis and therapeutic approach. We therefore conducted a retrospective study designed to assess whether disease characteristics at diagnosis have the power to predict long-term patient outcome. 2. Methods Retrospective analysis was led on the medical records of patients diagnosed with PAN before age 18 years between January 1986 and December 2006 and followed in a French tertiary care pediatric rheumatology center at Hôpital Necker-Enfants Malades, Paris. Diagnosis of PAN relied on the association of a clinical presentation compatible with a systemic inflammatory process involving ischemic manifestations and histological small or middle-sizedvessel involvement after ruling out other vasculitides. Patients were excluded if they presented with diffuse alveolar hemorrhage, pauci-immune glomerulonephritis, histological evidence of capillary and/or venular vasculitis and presence of anti-neutrophil cytoplasm antibodies (ANCA) specific for proteinase 3 or myeloperoxidase detected by ELISA, granuloma or IgA deposition. The study was approved by the Rhône-Alpes-Auvergne inter-regional institutional review board. 2.1. Data collection Clinical data collected included fever, constitutional symptoms (general malaise, fatigue, weight loss), evidence of specific organ involvement (as defined in the Ankara criteria) [6] such as high blood pressure, renal involvement, peripheral neuropathy, muscle or skin involvement – for all other organs, we collected clinical symptoms compatible with vasculitis. Ear-Nose-Throat (ENT) manifestations preceding (within 21 days) or accompanying PAN were also recorded. We also recorded what other diagnoses were considered. The laboratory data were: complete blood count and differential, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), cryoglobulins and serological hepatitis B virus (HBV) status. All imaging and/or histological investigations performed during follow-up were collected. Any abnormalities were reviewed by an expert clinician to rule them in or out as PAN-related. All treatments were systematically recorded. 2.2. Definitions Visceral involvement was defined as involvement of at least one of the following organs: nervous system, gastro-intestinal tract (documented by histology, endoscopy and/or computed tomography scan), testis (clinically or by histology), heart, lungs, kidney, or any other organ. Functional complaints (abdominal pain, headache, myalgia, arthralgia) and arthritis were not considered as visceral involvement. PAN was classified as “cutaneous PAN” in the absence of visceral involvement, marked weight loss (≥ 10% of body weight), abdominal pain or headache (group 1), “cutaneous PAN with significant extra-cutaneous features” if no visceral involvement was proven but marked weight loss, abdominal pain and/or headache
was present (group 2), and “visceral childhood PAN” if visceral involvement was present (group 3). Histology was considered as diagnostic in cases where fibrinoid necrosis of the arterial wall was straightforward, and non-diagnostic in cases where there was an inflammatory infiltrate composed predominantly of polymorphonuclear and lymphocytic cells within the medium-sized artery wall but without clear fibrinoid necrosis. Based on the data collected, the patients were re-evaluated according to the Ankara criteria [5]. Patient status at the latest follow-up was assessed as follows: clinical remission off-therapypatients presented neither clinical nor physical symptoms for at least 6 months after discontinuation of therapy; treatmentdependency-patients presented neither clinical nor physical signs but were still under therapy; active-disease-patients presented with clinical evidence of active disease. Relapse was defined as a resurgence of one or more clinical symptoms in a patient who had previously been free of any manifestation of the disease [7]. 2.3. Outcome Three main outcome measures were considered: • treatment-free clinical remission at last follow-up; • requirement and length of glucocorticoid therapy; • presence of disease-related sequelae. 2.4. Statistics First, we ran a univariate comparison between the 3 groups. Next, we ran a univariate analysis for each of the five outcome parameters in an attempt to identify prognostic factors. Analyses were led using Fisher’s exact test for qualitative variables or a nonparametric Kruskall-Wallis test for quantitative variables. Given the small number of patients, we did not run multivariate analyses. No adjustment was made for analyzing multiple outcomes [8]. All analyses were performed on R software (version 2.13.0) using the Epicalc package. 3. Results A total 29 children (20 girls and 9 boys) were included in the study. Median age at diagnosis was 6.9 years (range: 2–14). 3.1. Initial features Eight patients had an ENT infection 5 to 22 days prior the first signs of PAN. At first presentation, all children were febrile. The most frequent features involved: skin (24 children), joints (n = 21), muscles (n = 12), gastro-intestinal tract (n = 10), ENT (n = 9), and nervous system (n = 7). Other features are shown in Table 1. The delay between first signs and diagnosis ranged from 5 to 1580 days (median: 37 days). Before PAN was diagnosed, four patients were misdiagnosed as systemic-onset juvenile idiopathic arthritis, 3 as acute rheumatoid fever, 3 as Kawasaki disease, and one as systemic lupus erythematosus. Reason for delayed diagnosis > 30 days was: absence of initial cutaneous involvement (7 children), glucocorticoids or IVIG administered before skin biopsy (6 children), or spontaneous transient remission (1 child). HBV serology was negative in all patients. At least one biopsy was performed in all patients: in 16 cases, histological examination showed diagnostic PAN lesions with destruction of the mediumsized artery walls and fibrinoid necrosis, and in 13 patients the biopsies were non-diagnostic. Angiography was performed in two
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
G Model BONSOI-4129; No. of Pages 7
ARTICLE IN PRESS E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
3
Table 1 Patient characteristics at the time of diagnosis. n = 29 Age at onset (years) Sex ratio (M/F) Fever Cutaneous involvement Nodules Livedo reticularis Necrosis Purpura Muscle involvement Joints (arthralgia or arthritis) Hypertension Renal involvement Nervous system Peripheral Central Abdominal pain Testicular involvement ENT involvement Laboratory Leukocytes (x109 .L−1 ) Neutrophils (x109 .L−1 ) Platelets (x109 .L−1 ) ESR (mm.h−1 ) CRP (mg.L−1 ) 1992 diag. criteria fulfilled Ankara criteria fulfilled
6.9 ± 3.7 (2–14) 9/20 29 (100) 24 (82) 14 (48) 9 (31) 3 (10) 7 (24) 12 (41) 21 (72) 1 (3) 2 (7) 6 (21) 1 (3) 10 (34) 2 (22) 9 (31) 17.9 ± 8.7 (6.1–38) 14 ± 8.5 (3.8–31) 510 ± 239 (199–999) 75 ± 30 (10–160) 139 ± 73 (6–320) 4 (14) 16 (55)
ENT: ear-nose-throat; ESR: erythrocyte sedimentation rate; M: male; F: female; CRP: C-reactive protein. Quantitative data are expressed as means ± SD (range), and qualitative data are expressed as n (%).
patients (one cerebral angiography, one mesenteric angiography) due to seizures and severe abdominal pain, and neither showed any arterial abnormalities. Sixteen patients met the Ankara criteria. The 13 patients not fulfilling the Ankara criteria met clinical criteria but missed the histological diagnostic criterion and had not been angiographied. Clinical and biological presentation did not differ significantly according to whether or not patients met the Ankara criteria.
3.2. Comparison of children diagnosed as cutaneous PAN (group 1), cutaneous PAN with significant extra-cutaneous features (group 2), and visceral PAN (group 3) Eleven children initially presented as cutaneous PAN and 13 as cutaneous PAN with significant extra-cutaneous features. Two patients in each group subsequently developed visceral involvement at 42, 59, 66 and 398 days post-diagnosis, respectively, and were ultimately considered as visceral childhood PAN. Finally, 9 patients qualified as group 1, 11 as group 2, and 9 as group 3. The characteristics of the 3 groups are reported in Table 2. CRP and ESR differed significantly between the three groups (Table 2).
3.3. Treatment As first-line therapy, 18 patients were given glucocorticoids, 8 were given NSAIDs, and three were given IVIG. At last follow-up, 7 patients had received no glucocorticoids but only NSAIDs ± IVIG (Table 3), including two that qualified as group 3, 16 children had received glucocorticoids ± IVIG, and 6 had received glucocorticoids + cyclophosphamide with or without other therapies (adalimumab, rituximab, plasma exchange, cyclosporine, azathioprine, colchicine).
Figure 1. Cumulative incidence of glucocorticoid-free complete remission. Group 1: myalgia ± arthralgia, but absence of visceral involvement apart from mild abdominal pain and headache; Group 2: myalgia ± arthralgia + marked weight loss, abdominal pain and/or headache without proven visceral involvement; Group 3: proven visceral involvement.
3.4. Outcome Median follow-up was 70 months (range: 7–178) and was 62, 59 and 72 months for group 1, 2 and 3 respectively (P = 0.5). Median number of relapses was 1 (range: 0–10). One patient (from group 3) died of colonic perforation 10 years after the onset of severe relapsing PAN. 3.4.1. Clinical remission off-therapy At last follow-up, 15 children were in clinical remission offtherapy (no glucocorticoid and no other therapies): 4 from group 1 (44%), 4 from group 2 (36%) and 7 from group 3 (78%), P = 0.4 (Table 3). 3.4.2. Glucocorticoid therapy Glucocorticoid therapy was required for 8 (89%), 7 (64%) and 7 (78%) patients from groups 1, 2 and 3, respectively (P = 0.5). Seven patients did not require any corticoid therapy. In the other 22 cases, median duration of glucocorticoid therapy was 169, 83 and 132 months, respectively (P = 0.3). Maximal dose of oral glucocorticoids was 2, 1.6 and 1.5 mg/kg/day, respectively (P = 0.5), whereas dose 9 months after diagnosis was 0.16 mg/kg/d (range: 0.05–2), 0.4 mg/kg/d (range: 0.1–2) and 0.3 mg/kg/d (range: 0.1–1), respectively (P = 0.6). Cumulated dose of oral glucocorticoids was 777 mg/kg (range: 212–3270), 651 mg/kg (range: 315–2635) and 551 mg/kg (range: 60–2837), respectively (P = 0.7). Time-dependent probability of achieving glucocorticoidfree clinical remission was similar between the 3 groups (Figure 1). 3.4.3. Sequelae Three patients (one from each group) presented digital ischemia leading to amputation 1, 5 and 8 years after onset of the disease. All three children presented a histological hallmark of arterial necrosis. Overall, there were no significant differences between the three groups on the three selected outcome measures (Table 3). 3.5. Prognostic factors None of age, sex, PAN subgroup, Ankara criteria fulfillment, visceral involvement or leukocyte count pointed to a higher probability of glucocorticoid-free clinical remission.
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
G Model BONSOI-4129; No. of Pages 7
ARTICLE IN PRESS E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
4 Table 2 Characteristics of the three groups.
N Sex (M/F) Age (years) Typical histological findings Previous ENT infection Muscle involvement HTA Renal involvement Peripheral neuropathy Cutaneous features Livedo Nodules Necrosis Purpura Abdominal pain ENT involvement Mucosal involvement Joint Chills Headache Laboratory exams Leukocytes (x10e9/L) PNN (x10e9/L) Platelets (x10e9/L) ESR (mm/h) CRP (mg/l) ASLO > 600 UI/L 1992 diag. criteria fulfilled Ankara criteria fulfilled
Group 1 9
Group 2 11
Group 3 9
1/8 6 (3–13) 5 (56) 4 (44) 2 (22) 0 0 0
3/8 8 (4–13) 6 (54) 3 (28) 7 (63) 0 0 3 (27)
5/4 5 (2–14) 5 (56) 1 (11) 3 (33) 1 (11) 2 (22) 3 (33)
3 (33) 8 (89) 1 (11) 3 (33) 0 5 (56) 1 (11) 8 (89) 0 0
3 (27) 7 (64) 0 3 (27) 7 (64) 2 (18) 0 9 (82) 3 (27) 4 (36)
3 (33) 5 (56) 2 (22) 1 (11) 3 (33) 2 (22) 3 (33) 4 (44) 0 1 (11)
17.2 (7.5–33.1) 10.8 (4.5–31.0) 512 (309–990) 57 (10–78) 63 (6–150) 4 (44) 0 5 (56)
17.0 (6.1–31.5) 16.0 (4.3–30.0) 409 (199–905) 84 (59–101) 132 (59–320) 2 (18) 2 (18) 6 (54)
13.5 (7.3–38.0) 7.6 (3.8–25.0) 439 (307–999) 90 (32–160) 152 (113–270) 0 2 (22) 5 (56)
P
0.03 0.05
ASLO: anti-streptolysin antibodies; ENT: eye-nose-throat; ESR: erythrocyte sedimentation rate; M: male; F: female; CRP: C-reactive protein; PNN: polynuclear neutrophils. Group 1: myalgia ± arthralgia but absence of visceral involvement apart from abdominal pain and headache; Group 2: marked weight loss, abdominal pain and/or headache; Group 3: proven visceral involvement. P-values are shown only if ≤ 0.05. Quantitative data are expressed as median (range), and qualitative data are expressed as n (%).
Patients requiring glucocorticoids had higher leukocyte counts at diagnosis (19.5 ± 8 × 109 /L−1 [range: 7–38] vs 13.2 ± 6 × 109 /L−1 [range: 6–20], respectively, P = 0.06) and higher neutrophil counts (15.6 ± 8 × 109 /L−1 vs 9.3 ± 5 × 109 /L−1 , respectively, P = 0.06) than other patients, and fulfilled Ankara criteria in 14 out of 22 cases versus 2 out of 7 patients who remained glucocorticoid-free (odds ratio 4.38 at 95% CI [range: 0.53–53], P = 0.1). There were no differences on other variables. None of the studied parameters correlated with duration of glucocorticoid therapy or number of relapses. All three patients with ischemic disease-related sequelae fulfilled the Ankara criteria.
4. Discussion In this retrospective survey of 29 patients diagnosed with childhood-onset PAN from a single tertiary care pediatric rheumatology center, patient outcome could not be predicted by a histological necrotizing feature nor by initial disease presentation as none of the initial clinical or biological parameters correlated with patient status at last follow-up. Patients were fairly evenly spread across three groups at diagnosis – a group with visceral involvement, another group with no visceral involvement but marked constitutional symptoms (weight loss, abdominal pain or
Table 3 Treatments and outcome measures.
N Treatments Oral glucocorticoids Bolus methylprednisolone IVIG Cyclophosphamide Other Outcome Follow-up (months) Status Dead Active disease CR on therapy Treatment-free CR Sequelaea Glucocorticoid therapy duration (months) Number of relapses (mean, [median], range)
Group 1 9
Group 2 11
Group 3 9
8 (89) 0 (0) 6 (67) 2 (22) 0
7 (64) 1 (9) 4 (36) 2 (18) 0
7 (78) 2 (22) 4 (44) 2 (22) 2 (22)
62 (25–178)
50 (19–119)
72 (7–157)
0 2 (22) 3 (33) 4 (44) 1 (11) 69 (0–178) 2.4 [1] (0–5)
0 3 (27) 4 (36) 4 (36) 1 (9) 26 (0–119) 1.5 [0] (0–5)
1 (11) 0 1 (11) 7 (78) 1 (11) 39 (0–130) 2.2 [1] (0–10)
CR: clinical remission. Group 1: myalgia ± arthralgia, but absence of visceral involvement apart from mild abdominal pain and headache; Group 2: myalgia ± arthralgia + marked weight loss, abdominal pain and/or headache without proven visceral involvement; Group 3: proven visceral involvement. a Digital necrosis leading to amputation in the 3 cases, as well as diffuse lesions of the brain and testis in the patient from group 3
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
G Model
BONSOI-4129; No. of Pages 7
Organ involvement Author
No pts
PAN denomination
Definition used
Fever
Skin
HTA
Kidney
Joint
GIT
Muscle
Sore throat
Nervous system
Testis
CRP
ESR
ASLO
Ste
Cy
Relapsing pts
Evolution
1993
David
12
Associated with streptococcus
12
12
0
2
12
10
8
12
0
0
55
NA
–
10
2
7
2 proven visceral
1995
Kumar
10
Benign cutaneous
NA
NA
2
–
7
–
NA
NA
NA
–
NA
NA
NA
10
2
NA
7 autoamputations
1997 1998
Till Mocan
2 1
Cutaneous Cutaneous
Vasculitis associated with streptococcus infection Absence of systemic involvement No Absence of major organ system involvement
1 1
2 1
0 0
– –
2 1
1 0
1 0
2 0
0 0
0 0
NA NA
50 NA
2 1
1 1
0 0
2 0
0 0
2002 2004
Bauza Ozen
1 33
Cutaneous Cutaneous
1 14
1 33
0 0
– –
13
1 0
0 5
1 12
0 1
0 0
NA 37
52 60
1 NA
1 25
1 5
0 4
0 0
2004
Ozen
63
Systemic
36
58
27
NA
25
15
45
NA
14
4
NA
NA
NA
63
42
5
0 deaths
2005
Fathalla
4
Cutaneous
4
4
0
–
3
0
NA
2
0
–
NA
NA
3
4
1
3
0
2013
Elfetheriou 69
60
61
11
13
52
28
57
NA
7
4
56
64
21
69
57
24
3 deaths
Systemic
Confined to the skin with no organ involvement Organ involvement, regardless of vessel size. Constitutional symptoms and elevated acute-phase reactants Small and middle-sized vessels of the skin, muscles and joints EULAR PReS, cutaneous PAN excluded
ARTICLE IN PRESS
Year
E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
Table 4 Published series of childhood PAN.
5
G Model BONSOI-4129; No. of Pages 7 6
ARTICLE IN PRESS E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
headache) and a third group without any of these features – yet digital necrosis leading to amputation occurred in each group and there was no striking difference between the groups in terms of glucocorticoid usage or rate of glucocorticoid-free clinical remission. The rationale behind analyzing three groups of patients according to initial presentation was that the distinction between “cutaneous” and “visceral” PAN remains hazy, and intermediate forms could have been hard to classify. The clinical heterogeneity of the same disease is illustrated by the recently described monogenic form of PAN-related to an impaired function of adenosine deaminase 2, where within a same family, the homozygous carriers of the mutation have either cutaneous or visceral PAN [9]. Since cutaneous and visceral PAN share the same histopathological pattern, differentiating them is a challenge that can only be resolved based on clinical and imaging assessment at presentation and over the course of the disease [10]. This distinction is challenging for at least three reasons. First, the spectrum of cutaneous PAN, which includes fever, myalgia, arthralgia, non-erosive arthritis and, in some cases, peripheral neuropathy, is included within the spectrum of visceral PAN [11]. In adults, cutaneous PAN likely appears as an indolent disease, and some authors limit the definition of cutaneous PAN to cutaneous involvement without fever, myalgia, arthralgia or peripheral neuropathy (except if present at the same localization as skin lesions) [12]. In contrast, a review of 119 children with cutaneous PAN found that most children had fever and elevated acute-phase reactants, and that musculoskeletal features, joint involvement and peripheral neuritis were frequent (43%, 58% and 18%, respectively) [13]. Second, cutaneous PAN is frequently though to be associated with streptococcus. An infectious trigger is observed in many cases of either cutaneous or visceral PAN, but in a setting of systemic inflammation the diagnostic value of elevated ASLO is not straightforward for streptococcus infection. A high titer of ASLO or a suspected streptococcus infection can be associated with systemic PAN too [14,15]. Third, cutaneous PAN is thought to have a benign course. However, some cases of cutaneous PAN can prove severe due to digital necrosis and auto-amputations and will subsequently require glucocorticoids and/or other immunosuppressants (Table 4) [16,17]. In our series, four out of 20 cases without visceral involvement needed cyclophosphamide and two ultimately had finger amputations, whereas two patients with biopsy-proven visceral PAN did not require any glucocorticoid therapy. One major conceptual difference between cutaneous PAN and visceral PAN is that only visceral PAN carries a risk of vital organ involvement [4,15,18]. However, the heterogeneity of prognosis beyond organ involvement-based classifications prompts us to think that, within the PAN framework defined by the Ankara classification, prognosis should be evaluated based on specifically-validated tools, such as a “pediatric five-factor score” [19]. It has recently been reported that severe gastro-intestinal involvement correlates to a higher relapse rate [15]. Gastro-intestinal involvement is known to be a poor prognosis factor in adults with PAN [20]. Our sample population is too small to confirm these results, but note that the only patient who died had severe GI ischemia. Among the patients selected here with a diagnosis of pediatric PAN, several did not fulfill the recent Ankara criteria [11], yet showed no significant difference in clinical presentation and outcome to the other patients. The histological hallmark of PAN is the presence of fibrinoid necrosis with complete destruction of medium- and sometimes small-sized artery walls, but in some cases this typical pattern is missing or incomplete [3]. Angioimaging may have allowed some of the patients who missed the histological criterion to fulfill the Ankara criteria, but few patients
were angiographied. Even so, Ankara criteria were more sensitive in our series than the ACR criteria which were first designed in adults [21] then re-adapted to pediatric PAN [22]. All our patients had a clinical presentation that was compatible with Ankara criteria. Histological assessment was performed in all cases but was not always typical, likely related to the fact that PAN is a patchy disease. Consequently, and to better assess disease severity, there is a strong rationale for making wider use of angiographic imaging procedures in children with a suspicion of PAN. Arteriography is usually considered the most powerful imaging exam [23] but is not frequently performed, especially in moderate forms. Other imaging approaches considered for the Ankara classification include CT scan [24], angio-CT scan or angio-MRI. Although less sensitive than arteriography [23], these non-invasive approaches are key to the assessment of a systemic vasculitis and are now gaining in popularity. The treatment of childhood PAN is not well codified. Note that 7 of our patients received neither glucocorticoids nor immunosuppressants. Several patients first received IVIG due to a suspicion of Kawasaki disease and went on to show dramatic but transient improvement. It is unclear whether IVIG usage had a glucocorticoid-sparing effect in some patients [25]. Cyclophosphamide has been shown to reduce the number of relapses [15], but in our series it was only administered in a few patients with a severe disease course. We posit that the significant effect of immunosuppressive drugs on disease activity in several patients here can explain the absence of difference in outcome between patients with visceral forms and patients with non-severe relapsing disease. Given that most patients had been exposed to glucocorticoids for relatively long periods of time, a more modern approach with more frequent early introduction of immunosuppressive drugs might be worth considering. Our study has several limitations. First, the number of patients was too small to allow a fine-grained comparison between different groups of patients, and there may have been a recruitment bias towards more severe forms in our center. Second, the assessment of long-term outcome warrants caution in this retrospective study, as patients were included over a 20-year period in which therapeutic strategies may have differed over time. Nevertheless, this study still represents one of the largest series of pediatric PAN, and median follow-up was relatively long. In conclusion, this study failed to find evidence that initial presentation has predictive value for outcome in patients with pediatric-onset PAN. In particular, when considering a looser definition of “cutaneous PAN”, as also proposed in many other publications [11,13,16,26], we found a few cases where children with “cutaneous PAN” and no visceral involvement went on to develop severe complications, such as digital necrosis, ending up with extended periods of glucocorticoid therapy and immunosuppressive drugs and having long-lasting disease activity. We conclude that the organ distribution-based distinction between cutaneous and visceral childhood PAN had little prognostic power in this series.
Contributors Etienne Merlin, Richard Mouy and Pierre Quartier conceptualized and designed the study, drafted the initial manuscript, and approved the final manuscript as submitted. Bruno Pereira and Aurélie Bourmaud carried out the initial analyses, reviewed and revised the manuscript, and approved the final manuscript as submitted. Luc Mouthon, Jean-Charles Piette, Judith Landman-Parker, Patricia Chellun, Mustapha Layadi, Caroline Thomas, Loïc Guillevin and Anne-Marie Prieur coordinated and supervised the data
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
G Model BONSOI-4129; No. of Pages 7
ARTICLE IN PRESS E. Merlin et al. / Joint Bone Spine xxx (2015) xxx–xxx
collection, critically reviewed the manuscript, and approved the final manuscript as submitted. Disclosure of interest The authors declare that they have no conflicts of interest concerning this article. References [1] Kussmaul ARM. Uber eine bisher nicht beschreibene eigenthümliche Arterienerkankung (Periarteriitis nodosa), die mit Morbus Brightii und rapid fortschreitender allgemeiner Muskellähmung einhergeht. Deutsches Arch F Klin Med 1866;1:84–517. [2] Chen KR. Cutaneous polyarteritis nodosa: a clinical and histopathological study of 20 cases. J Dermatol 1989;16:429–42. [3] Dillon MJ, Eleftheriou D, Brogan PA. Medium-size-vessel vasculitis. Pediatr Nephrol 2010;25:1641–52. [4] Ozen S, Anton J, Arisoy N, et al. Juvenile polyarteritis: results of a multicenter survey of 110 children. J Pediatr 2004;145:517–22. [5] Ozen S, Pistorio A, Iusan SM, et al. EULAR/PRINTO/PRES criteria for HenochSchönlein purpura, childhood polyarteritis nodosa, childhood Wegener granulomatosis and childhood Takayasu arteritis: Ankara 2008. Part II: Final classification criteria. Ann Rheum Dis 2010;69:798–806. [6] Ruperto N, Ozen S, Pistorio A, et al. EULAR/PRINTO/PRES criteria for HenochSchönlein purpura, childhood polyarteritis nodosa, childhood Wegener granulomatosis and childhood Takayasu arteritis: Ankara 2008. Part I: Overall methodology and clinical characterisation. Ann Rheum Dis 2010;69:790–7. [7] Hellmich B, Flossmann O, Gross WL, et al. EULAR recommendations for conducting clinical studies and/or clinical trials in systemic vasculitis: focus on anti-neutrophil cytoplasm antibody-associated vasculitis. Ann Rheum Dis 2007;66:605–17. [8] Rothman KJ. No adjustments are needed for multiple comparisons. Epidemiology 1990;1:43–6. [9] Navon Elkan P, Pierce SB, Segel R, et al. Mutant adenosine deaminase 2 in a polyarteritis nodosa vasculopathy. N Engl J Med 2014;370:921–31. [10] Eleftheriou D, Brogan PA. Vasculitis in children. Best Pract Res Clin Rheumatol 2009;23:309–23.
7
[11] Ozen S, Ruperto N, Dillon MJ, et al. EULAR/PReS endorsed consensus criteria for the classification of childhood vasculitides. Ann Rheum Dis 2006;65: 936–41. [12] Nakamura T, Kanazawa N, Ikeda T, et al. Cutaneous polyarteritis nodosa: revisiting its definition and diagnostic criteria. Arch Dermatol Res 2009;301: 117–21. [13] Bansal N-K, Houghton KM. Cutaneous polyarteritis nodosa in childhood: a case report and review of the literature. Arthritis 2010;2010:687547. [14] David J, Ansell BM, Woo P. Polyarteritis nodosa associated with streptococcus. Arch Dis Child 1993;69:685–8. [15] Eleftheriou D, Dillon M, Tullus K, et al. Systemic polyarteritis nodosa in the young: a single centre experience over 32 years. Arthritis Rheum 2013;65:2476–85. [16] Mocan H, Mocan M, Peru H, et al. Cutaneous polyarteritis nodosa in a child and a review of the literature. Acta Pædiatrica 1998;87:351–3. [17] Williams VL, Guirola R, Flemming K, et al. Distal extremity necrosis as a manifestation of cutaneous polyarteritis nodosa: case report and review of the acute management of a pediatric patient. Pediatr Dermatol 2012;29:473–8. [18] Morgan AJ, Schwartz RA. Cutaneous polyarteritis nodosa: a comprehensive review. Int J Dermatol 2010;49:750–6. [19] Guillevin L, Pagnoux C, Seror R, et al. The Five-Factor Score revisited: assessment of prognoses of systemic necrotizing vasculitides based on the French Vasculitis Study Group (FVSG) cohort. Medicine (Baltimore) 2011;90:19–27. [20] Guillevin L, Lhote F, Gayraud M, et al. Prognostic factors in polyarteritis nodosa and Churg-Strauss syndrome. A prospective study in 342 patients. Medicine (Baltimore) 1996;75:17–28. [21] Lightfoot Jr RW, Michel BA, Bloch DA, et al. The American College of Rheumatology 1990 criteria for the classification of polyarteritis nodosa. Arthritis Rheum 1990;33:1088–93. [22] Ozen S, Besbas N, Saatci U, et al. Diagnostic criteria for polyarteritis nodosa in childhood. J Pediatr 1992;120:206–9. [23] Schmidt WA. Use of imaging studies in the diagnosis of vasculitis. Curr Rheumatol Rep 2004;6:203–11. [24] Adaletli I, Ozpeynirci Y, Kurugoglu S, et al. Abdominal manifestations of polyarteritis nodosa demonstrated with CT. Pediatr Radiol 2010;40:766–9, http://dx.doi.org/10.1007/s00247-009-1466-4. [25] Aries PM, Hellmich B, Gross WL. Intravenous immunoglobulin therapy in vasculitis: speculation or evidence? Clin Rev Allergy Immunol 2005;29:237–45. ˜ A, Idoate M. Cutaneous polyarteritis nodosa. Br J Dermatol [26] Bauzá A, Espana 2002;146:694–9.
Please cite this article in press as: Merlin E, et al. Long-term outcome of children with pediatric-onset cutaneous and visceral polyarteritis nodosa. Joint Bone Spine (2015), http://dx.doi.org/10.1016/j.jbspin.2015.01.007
