Diagnostic criteria for polyarteritis nodosa in childhood Seza Ozen, MD, Nesrin Besbas, MD, Umit Saatci, MD, a n d Aysin Bakkaloglu, MD From the Department of Pediatric Nephrology and Rheumatology, Hacettepe Children's Hospital, Ankara , Turkey Clinical and laboratory features of 31 children with a diagnosis of polyarteritis nodosa were e v a l u a t e d retrospectively. All the patients had musculoskeletal involvement, renal involvement, or both during the course of the disease. We have d e f i n e d involvement of these two systems as the major diagnostic criteria in polyarteritis nodosa. Ten additional minor criteria were defined: (1) cutaneous findings, (2) gastrointestinal involvement, (3) peripheral neuropathy, (4) central nervous system involvement, (5) hypertension, (6) c a r d i a c involvement, (7) lung involvement, (8) constitutional symptoms, (9) presence of acute-phase reactants, and (4(]) presence of hepatitis B surface antigen. We propose that the presence of five of these criteria, including at least one major criterion, is highly suggestive of polyarteritis nodosa; such a c o m b i n a t i o n was present in 97% of our patients. Fourteen of the patients were treated with corticosteroids alone and 14 were treated with a c o m b i n a t i o n of steroids plus c y c l o p h o s p h a m i d e or azathioprine. At the last follow-up examination six patients in the steroid group and nine in the c o m b i n a t i o n group were considered to have c o m p l e t e remission of disease or inactive disease with persisting symptoms in an organ system. The overall mortality rate was 16%; renal involvement had the greatest adverse effect on outcome. We suggest that in patients with five of the 12 diagnostic criteria, especially those with renal involvement, therapy should be initiated promptly while diagnostic procedures are being carried out. (J PEDIATR1992; 120:206-9)
Polyarteritis nodosa is one of the vasculitis syndromes involving small and medium-sized arteries. 1 The clinical evidence is diverse, and differentiation of P A N from other rheumatic diseases, particularly from other vasculitis syndromes, may be difficult. 1' 2 Although the definitive diagnosis is made by histopathologic studies, such studies may be made difficult because of the segmental nature of vessel involvement t, 3 Thus the diagnosis of P A N requires knowledge of its clinical features. In this study we present the clinical features of 31 children with PAN that was diagnosed clinically and pathologically. The clinical data, therapy, and prognosis of these
patients have been studied retrospectively. To our knowledge this represents the largest series of patients with childhood P A N in the English-language literature. We have also attempted to describe diagnostic clinical criteria for PAN.
Submitted for publication May 24, 1991; accepted Aug. 13, 1991. Reprint requests: Seza Ozen, MD, Kuleli sok. 9/2, Gazi Osman Pasa, 06700 Ankara, Turkey. 9/20/33177
was made by renal biopsy in eight patients and by skin biopsy in 22. In one of the patients the diagnosis of P A N was established by the presence of characteristic findings on renal angiography. Patients with cutaneous PAN were e x -
METHODS Thirty-one patients with P A N who were seen at our department of pediatric rheumatology from 1979 to 1991 were included in this study. The histologic diagnosis of vasculitis FMF PAN
Familial Mediterranean fever Polyarteritis nodosa
Volume 120 Number 2, Part 1
cluded from this study. Retrospective analysis of the clinical and laboratory findings in these patients enabled us to designate 12 features as diagnostic criteria for P A N (Table I). Statistical analysis was carried out with the chi-square test. RESULTS The age range of the patients was 1 to 14 years, with an average age of 9.3 years. There were 22 boys and 9 girls, a ratio of 2.5:1. The most frequent clinical features are shown in the Figure. Constitutional symptoms consisting of fever, anorexia, and fatigue were present in 20 patients. Cutaneous changes and musculoskeletal involvement were the most common clinical symptoms. Cutaneous findings were in the form of livedo reticularis or maculopapular purpuric lesions, with or without peripheral gangrene. Nodules were present in two patients. Digital infarction occurred in another two. Complaints of myalgia, arthralgia, or both often involved the lower extremities and correlated well with disease activity. Severe myalgia unresponsive to analgesic therapy was the most frequent complaint at the time patients were first seen (24 patients); arthritis or arthralgia was present in 17. Gastrointestinal involvement was characterized by a range of symptoms, from abdominal pain to severe bleeding. Peripheral nerve involvement and central nervous system involvement were present in four and nine patients, respectively. Peripheral nerve involvement consisted of numbness, paresthesias, polyneuropathy in one patient. In four of the patients with central nervous system involvement, encephalitic symptoms were present and computed tomographic scans showed abnormalities. One patient had persistent hemiparesis and ptosis. Another had hemiparesis and palsy of the seventh nerve. Three patients had recurrent seizures and abnormal electroencephalographic findings. Evidence of renal involvement included proteinuria in 11 patients, hematuria and proteinuria in two, nephrotic syndrome with hematuria in two, and rapidly progressive glomerulonephritis in five. Patients with a diagnosis of rapidly progressive glomerulonephritis had elevated serum creatinine and blood urea nitrogen levels. Three of these patients had a rapidly fatal course and died as a result of complications of uremia, although therapy had been started in two cases. Five of the patients had prominent cardiac signs and symptoms, including pericarditis, arrhythmias, or cardiac failure. One of these patients also had myocardial infarction. Three patients had pulmonary involvement, as indicated by pulmonary infiltrates, pleural effusions, or hemoptysis. Two of the patients had associated familial Mediterra-
Childhood polyarteritis nodosa
T a b l e I. Proposed diagnostic criteria for P A N
Major findings Renal involvement Musculoskeletal findings Minor findings Cutaneous findings Gastrointestinal system involvement Peripheral neuropathy Central nervous system disease Hyperten sion Cardiac involvement Lung involvement Constitutional symptoms Acute-phase reactants Presence of HBsAg HBsAg, HepatitisB surface antigen.
nean fever diagnosed by family history and characteristic periodic attacks; these two patients had been treated with colchicine. Two brothers with P A N are included in this series; they had similar clinical findings, with skin and nervous system involvement. Ten patients had mild anemia with hemoglobin levels less than 10 gm/dl. Leukocytosis, with a leukocyte count >15,000 cells/#l, was detected in 14 of the patients, although the counts usually were not >22,000 cells/tzl. The erythrocyte sedimentation rate was elevated in 27 patients. In four patients with a normal erythrocyte sedimentation rate, constitutional symptoms were not present. Altogether acute-phase reactants, including elevated erythrocyte sedimentation rate, leukocytosis, and C-reactive protein, were present in 28 patients. Hepatitis B antigen was present in three. We have developed diagnostic criteria as a result of the findings in our patients. An essential diagnostic criterion was the lack of antinuclear and anti-DNA antibodies. Renal involvement, musculoskeletal symptoms, or both were present in all of our patients; these involvements were specified as the major features. Among the patients without renal involvement (n = 11), musculoskeletal findings were present in all of them. No other minor criterion was present in any of these 11 patients, and no other pair of criteria could be defined. The presence of 5 of these 12 criteria, including at least one major criterion, was used to indicate the diagnosis of PAN. Of 31 patients, 30 had 5 of the 12 diagnostic criteria for PAN, including at least one major criterion. Thus in 97% of the patients the proposed set of criteria was correlated with the histopathologic diagnosis of PAN. Fourteen patients were randomly assigned to treatment with steroids alone and 14 were treated with corticosteroids (2 mg/kg per day) plus a cytotoxic agent. In the steroidtreated group, follow-up data were not available for six patients. The mean duration of follow-up in the remaining
Ozen et al.
The Journal of Pediatrics February 1992
Constitutional symp Cutaneous findings Musculoskeletal symp Gut involvement Nervous system syrup Renal involvement Hypertension
Elevated ESR Anemia Leukocytosis
Number of patients Figure. Clinical and laboratory findings in 31 patients with PAN.
Table II. Comparison of important features Percentage of patients
Fink3 (n = 25)
Present series (n = 31)
Fever Skin lesions Myalgia, arthralgia Gut involvement Neurologic disease Hypertension Renal involvement Sedimentation
92 76 56 80 56 96 80 100
65 81 81 6l 48 65 65 87
eight patients in this group was 47 months. One died of renal failure and one died of cardiac failure and hypertension. Six patients are alive at present; two of these have complete remission of their disease. Three are continuing to receive low-dose ( 10 rag), alternate-day, oral corticosteroid therapy for their persisting mild symptoms with inactive vasculitic disease. The last patient in the steroid-treated group has persistent cardiac abnormalities. The patients treated with cortieosteroids plus cytotoiic drugs were treated with either cyclophosphamide or azathioprine, each at a dose of 2 mg/kg per day by mouth. Two patients were lost tO follow-up; the mean duration of follow-up of the remaining 12 patients was 42 months. One
patient in this group died of renal failure and another died of neurologic complications. Nephrotic syndrome has recently been diagnosed in one of these patients. Complete remission was achieved in four patients, and the remaining five have inactive vasculitic disease and are receiving maintenance corticosteroid treatment. In three of the patients no immunosuppressive treatment was started because of the rapid, fatal course of the disease in two and the noncompliance of the other with the treatment regimen. Among these three patients, one died of renal failure and one died of gastrointestinal bleeding. DISCUSSION It has been stated that "the diagnosis of PAN in children is probably most frequently made at autopsy. ''4 On the other hand, appropriate therapy markedly improves the chances of survival once the diagnosis is established J, 4, 5 Because of the segmental nature of the disease and the fact that diagnosis in reported childhood cases is sometimes uncertain, the crucial diagnosis may be delayedJ On the basis of our experience with 31 patients with a histologic diagnosis of PAN and on the basis of previously reported eases, l, 3, 6 we have proposed 12 features for the diagnosis of PAN in children. The presence of five of these criteria, including at least one major finding, correlated with the histopathologic diagnosis of PAN in almost all of our patients. If these criteria can be validated in a prospective study, they may be applied to establish an early diagnosis
Volume 120 Number 2, Part 1
and therapy while time-consuming procedures for a pathologic diagnosis are being carried out. It has been estimated that fewer than 150 cases of PAN in children were reported before 1976, and only small series have been reported since then.1 Our results are similar to the data from three childhood series reviewed by Fink3 (Table II). Although the data are similar, the most common clinical findings in our patients were skin involvement and myalgia, whereas in the other series the findings were constitutional symptoms and hypertension. The most striking difference was the prominence of myalgia in our group, with a 78% prevalence; myalgia was present in only 53% of the patients in the other childhood PAN series. Polyarteritis nodosa in children differs somewhat from that seen in adults.3,6 An elevated erythrocyte sedimentation rate is prominent in active cases of childhood PAN but correlates poorly with disease activity in adults. 1, 3, 5 Two of our patients with PAN were especially interesting because of the association with FMF. Thirteen cases of PAN have been reported in patients with FMF. 7 Glikson et al. 7 speculated that the development of PAN in patients with F M F may be related to streptococcal infections. We believe that it is difficult to substantiate such a conclusion because streptococcal infections are endemic in most of the countries where FMF is fairly common. We also described two brothers with PAN. Cases of familial P A N are rare but have been reported previously; both HLA-linked and non-HLA-linked genetic influences have been suggested in the pathogenesis of this condition,s Eight patients were treated with steroids alone and 12 patients were treated with steroids plus cytotoxic drugs. Although 25 % and 16% died in the group receiving steroids alone and the group receiving combination treatment, respectively, the difference in mortality rate between the
Childhood polyarteritis nodosa
two groups was not statistically significant. In our patients and in those reviewed by Meadow, 6 renal failure was the main cause of death. In our series 14% of the patients with renal involvement died; the mortality rate in those with gastrointestinal involvement was only 5.3%, and in patients with nervous system involvement the rate was 6.6%. Because PAN carries a grave prognosis and pathologic findings may be difficult to demonstrate in biopsy specimens, clinicians should be aware of the clinical features of the disease. We believe that in cases fulfilling the suggested criteria therapy should be initiated promptly while the diagnostic procedures are continued, especially in patients with renal involvement. REFERENCES
1. Cassidy JT, Petty RE.Vasculitis. In: Cassidy JT, Petty RE, eds. Textbook of pediatric rheumatology. 2nd ed. New York: Churchill Livingstone, 1990:377-424. 2. Cohen RD, Corm DL, Ilstrup DM. Clinical features, prognosis and response to treatment in polyarteritis. Mayo Clin Proc 1980;55:146-55. 3. Fink CW. Vasculitis. Pediatr Clin North Am 1986;33:120320. 4. Schaller JG, Wedgwood RJ. Vasculitis sydrome. In: Behrman RE, Vaughan VC, Nelson WE, eds. Nelson textbook of pediatrics. Philadelphia: WB Saunders 1987:527-30. 5. GuillevinL, Le Thi Houng Du, Godeau P, et al. Clinical findings and prognosis of polyarteritis nodosa and Churg-Strauss angiitis: a study in 165 patients. Br J Rheumatol 1988;27:25864. 6. Meadow R. Polyarteritis nodosa. In: Edelman CM Jr, ed. Pediatric kidneydisease, vol 2. Boston:Little, Brown, 1978:711-8. 7. Glikson M, Galun E, Schlsinger M, et al. Polyarteritis nodosa and familial mediterranean fever: a report of 2 cases and review of the literature. J Rheumatol 1989;16:536-9. 8. Reveille JD, Goodman RE, Barger BO, Acton RT. Familial polyarteritis nodosa: a serologic and immunogenetic analysis. J Rheumatol 1989;16:181-5.