Nephrol Dial Transplant (1998) 13: 455–458
Nephrology
Dialysis
Transplantation
Case Report
Case of propylthiouracil-induced ANCA associated small vessel
vasculitis
Lorraine Harper, Paul Cockwell and Caroline O. S. Savage
Department of Nephrology, Queen Elizabeth Hospital, Birmingham, UK
Introduction
The aetiology of antineutrophil cytoplasmic autoantibody (ANCA)-associated vasculitis is unknown,
although environmental triggers have been proposed.
There are a few published reports of an association
between treatment with the antithyroid drug propylthiouracil and the occurrence of ANCA positivity and
ANCA-associated vasculitis. We discuss a patient
treated with propylthiouracil who developed an
ANCA-positive vasculitis with glomerulonephritis and
pulmonary haemorrhage.
Case report
In 1993 a previously healthy 30-year-old woman was
diagnosed as having autoimmune thyrotoxicosis. She
was treated with carbimazole for 5 months, then converted to propylthiouracil as she wished to conceive.
In November 1996 she was admitted with a 4-week
history of deafness, tinnitus, and general malaise. In
the week prior to admission she had developed progressive dyspnoea and frank haemoptysis. On clinical
examination she was tachypnoeic, pale and apyrexial.
Heart rate was 92/min and blood pressure
110/80 mmHg. On chest auscultation there were bilateral coarse basal crackles to both mid-zones. There
were no other abnormal clinical findings.
Investigation results included haemoglobin 4.8 g/dl,
white cell count 4.5×109/l with a normal differential
and platelets 349×106/l. Serum urea was 10.7 mmol/l
and creatinine 68 mmol/l. Creatinine clearance was
reduced at 51 ml/min. Urine analysis showed blood
(+++) and protein (+++) and albuminuria was
quantified as 0.008 g/l. Erythrocyte sedimentation rate
( ESR) was 51 mm/h and C-reactive protein (CRP)
156 mg/l (normal range <10). ANCA was positive
with a perinuclear pattern at a titre of 15400 with an
anti-myeloperoxidase (anti-MPO) specificity (antiproteinase 3 [anti-PR3] was not detectable), antiCorrespondence and offprint requests to: C. O. S. Savage, Senior
Lecturer, Department of Nephrology, Queen Elizabeth Hospital,
Edgbaston, Birmingham B15 2TH, UK.
nuclear antibodies (ANA) were positive at a titre of
15100. Anti double-stranded DNA (anti-dsDNA) and
anti-glomerular basement membrane antibodies were
negative. Complement levels were normal. Chest X-ray
showed bilateral alveolar shadowing consistent with
pulmonary haemorrhage ( Figure 1). Corrected gas
transfer ( KCO) was elevated at 3.67 (173% greater
than predicted). An ultrasound scan of the kidneys
was normal. Renal biopsy showed focal segmental
necrotizing lesions associated with extracapillary proliferation typical of a vasculitic glomerulonephritis
( Figure 2). Staining for immunoglobulin and complement components was negative.
She was treated with 6×2.5 l plasma exchanges,
pulse methyl prednisolone (500 mg/day for 3 days)
followed by oral prednisolone (60 mg/day), and oral
cyclophosphamide (2 mg/kg/day). Propylthiouracil
was discontinued. She made an uncomplicated clinical
recovery. Chest X-ray, urinalysis, CRP, and ESR
returned to normal. Her KCO is now 55% of the
predicted normal. She remains strongly pANCA positive with a titre of 15400. ANA has become negative.
Fig. 1. Chest X-ray on admission showing bilateral alveolar
shadowing consistent with pulmonary haemorrhage.
© 1998 European Renal Association–European Dialysis and Transplant Association
456
Fig. 2. Glomerulous showing a small active segmental lesion with
tuft disruption and extracapillary proliferation. Periodic acid-silver,
magnification ×250.
Treatment with cyclophosphamide was stopped after
5 months and changed to azathioprine 100 mg, she
continues on 10 mg of prednisolone as maintenance
therapy. She remains biochemically euthyroid off
treatment.
Discussion
Vasculitis associated with anti-thyroid medication
The overall incidence of important side-effects of propylthiouracil varies from 1–5%. Rare but serious complications include agranulocytosis, hepatotoxicity, and
drug induced hypersensitivity.
A clinical association between ANCA positivity and
clinical symptoms of drug induced hypersensitivity
associated with propylthiouracil was initially reported
by Dolman et al. in six female patients [1]. These
patients developed ANCA against human neutrophil
elastase and either proteinase 3 or myeloperoxidase
(three patients had all three autoantibodies). All had
normal renal function though three had microscopic
haematuria but no proteinuria. Two patients had skin
biopsies suggesting vasculitis and one had a nasal
biopsy suggesting Wegener’s granulomatosis but there
was no granuloma. Only one patient was treated with
immunosuppressive therapy (cyclophosphamide and
prednisolone), the others responded to withdrawal of
the propylthiouracil.
There are 14 reports of ANCA associated pauciimmune glomerulonephritis associated with propylthiouracil, of which two occurred in children [2–9] (see
Table 1). Three of these patients were treated with
prednisolone and cyclophosphamide, not all improved
with one of the treated children progressing to endstage renal failure [4,10]. One was treated with cyclophosphamide alone [2]. Seven patients were treated
with prednisolone alone, one deteriorated with a further decline in renal function [6–8]. Two patients
showed improvement following discontinuation of
their propylthiouracil, neither received immunosuppressive treatment [6 ].
L. Harper, P. Cockwell and C. O. S. Savage
Hypersensitivity vasculitis associated with propylthiouracil may affect any organ with skin lesions being
the most common [11]. Our patient is extremely
unusual in that she presented with pulmonary haemorrhage and renal vasculitis. To our knowledge there
have only been three other reports of pulmonary
haemorrhage associated with propylthiouracil-induced
ANCA positive vasculitis [10,12,13], of which only
one had renal involvement [10]. Treatment included
withdrawal of propylthiouracil in all the patients, one
patient also received treatment with corticosteroids
[13], both our patients and one reported by D’Cruz
[10] received steroids and cyclophosphamide. The
patient reported by Romas [12] received no immunosuppressive treatment. All patients (including our own)
improved following treatment.
Vasculitis has also been reported with the use of
carbimazole and methimazole. D’Cruz [10] reported a
patient who, a month after discontinuation of carbimazole, developed a Wegener’s type illness. Treatment
included plasma exchange, cyclophosphamide, and
prednisolone. Of the three patients reported by
Tanemoto, one was receiving both propylthiouracil
and methimazole, the causative drug could not be
identified and both were discontinued, with resolution
of symptoms [6 ].
At the molecular level, the three different antithyroid drugs are all based on a heterocyclic compound
containing a thiomide group, yet despite these structural similarities, there does not appear to be any cross
sensitivity. In two of the patients reported by
Tanemoto, methimazole was substituted for propylthiouracil and vice versa to control the hyperthyroidism
without recurrence of the symptoms induced by the
other drug [6 ]. Dolman also substituted methimazole
in three of his patients with no untoward effects [14].
However in two reports of vasculitis induced by propylthiouracil, one of which was non-renal, the patients
had previously demonstrated carbimazole intolerance
(developing a carbimazole rash) [10,15]. This suggests
that substitution of these drugs should be undertaken
with caution.
Withdrawal of the propylthiouracil does not always
result in a fall in titre or disappearance of circulating
ANCA. Our patient has remained ANCA positive
seven months after withdrawal of the propylthiouracil
at high titre (pANCA 1:400). It is interesting to note
that patients with the idiopathic disease who remain
ANCA positive are at increased risk of relapse [16,17].
Despite propylthiouracil withdrawal these patients may
still be at risk of disease recrudescence and should
remain under long term follow-up.
Idiosyncratic reactions associated with propylthiouracil such as agranulocytosis or vasculitis usually occurs
within weeks of the start of treatment although reports
occuring several years after commencement of treatment are reported. In the patients reported duration
of propylthiouracil treatment ranged from 1 week to
84 months.
Propylthiouracil-induced ANCA associated with small vessel vasculitis
457
Table 1. Reported cases of ANCA associated small vessel vasculitis and anti-thyrotoxcosis treatment
Reference
Sex
Age
Symptoms
Drug
Duration
on drug
Kitahari
2
Vogt
4
F
39
PTU
5 years
M
11
PTU
2 months
F
15
PTU
1 month
M
49
Myalgia, fever, scleritis,
proteinuria, nasal inflammation
Arthralgia, malaise,
ARF, rash
Arthralgia, fever,
Splenomegaly,
Impaired renal function
Impaired renal function
PTU
3 years
F
F
36
60
Impaired renal function
Impaired renal function
PTU
PTU
2 years
3 years
M
22
PTU, M
3 years
F
M
82
62
M
40
F
82
F
52
Fever, rash arthralgia
MH, proteinuria
Impaired renal function
Dyspnoea, haemoptysis,
Arthralgia, malaise,
Proteinuria
Epistaxis, arthralgia,
Fever, ARF
Scleritis, rash
Impaired renal function
Rash, impaired renal function
F
54
F
34
Rash, fever, arthralgia
Impaired renal function
Fever, impaired renal function
Yuasa
5
Tanemoto
6
D’Cruz
10
Ito
7
Kudo
8
Toda
9
PTU
PTU
4 years
11 months
carbim
12 weeks
PTU
?
PTU
?
PTU
84 months
PTU
4 years
Treatment
Outcome
C
Stopped PTU
C, P
Stopped PTU
C, P
Stopped PTU
Resolution on
stopping PTU
ESRF
P
Reduced dose
P
P
Stopped PTU
P
Stopped PTU, M
Stopped PTU
C,P
Stopped PTU
Resolution
C, P, PEX
Stopped carbim
P
PTU stopped
P
Stopped PTU
Stopped PTU
P
Stopped PTU
Resolution
Resolution
Resolution
Resolution
Resolution
Resolution
Resolution
Resolution
Decline renal function
Resolution
Resolution
F, female; M, male; ARF, acute renal failure; PTU, propylthiouracil; carbim, carbimazole; P, prednisolone; C, cyclophosphamide;
resolution, resolution of presenting symptoms and improvement in renal function.
Possible pathogenic mechanisms
It has been suggested that myeloperoxidase, which is
involved in the formation of reactive metabolites from
propylthiouracil, may bind covalently to one of these
metabolites [18]. The myeloperoxidase may be modified, forming an immunogenic conjugate with the metabolite acting as a hapten. When studied the ANCA
binding site did not involve the propylthiouracil metabolites, though this does not preclude the drug metabolites acting as a hapten. It is possible that the
mechanism suggested for the development of ANCA
may also occur in idiopathic ANCA associated vasculitis. However many of the patients reported had several
ANCA specificities detected (anti-PR3, anti-MPO and
anti-leucocyte elastase) which is uncommon in the
idiopathic disease and may therefore represent a
different pathogenic mechanism for the development
of autoantibody [19]. The mechanism behind ANCA
development is further complicated by the fact that
anti-neutrophil antibodies are not uncommonly seen
in patients with Graves’ disease, this may be a result
of the similarities between thyroid peroxidase and
myeloperoxidase. Development of ANCA may occur
with other drugs including hydralazine and penicillamine though their occurrence with these drugs is strongly
HLA-DR3 linked [20]. The HLA status of patients
developing ANCA associated with propylthiouracil
has not been studied.
It is unclear whether the altered state of self
tolerance present in these patients contributes to the
development of ANCA or whether there is an interaction between myeloperoxidase and propylthiouracil,
resulting in an immunogenic compound. At present no
good animal model of ANCA associated systemic
vasculitis exists. Thyrotoxic cats were treated with
propylthiouracil, two of five of these cats developed a
lupus like syndrome with ANCA [21]. The development of these antibodies supports the concept that
metabolites of propylthiouracil binding to myeloperoxidase are capable of inducing antibody production
and that these antibodies were involved in the pathogenesis of the syndrome. Unfortunately, when the
experiments were repeated the cats did not develop
vasculitis, though this may have been due to a change
in diet rendering these cats less susceptible to
immune disease.
The association between propylthiouracil and
ANCA associated vasculitis is an important, although
rare side-effect of this widely used drug. As in our
patient, it can cause life-threatening pulmonary-renal
syndrome which has a high morbidity and mortality.
Occurrence of symptoms of vasculitis should necessitate immediate withdrawal of the drug. The need for
treatment with immunosuppressive therapy is less clear,
although it is probable that aggressive disease should
be treated with immunosuppressive regimes.
458
References
1. Dolman K, Gans R, Vervaat T et al. Vasculitis and antineutrophil cytoplasmic autoantibodies associated with propylthiouracil
therapy. Lancet 1993; 342: 651–652
2. Kitahara T, Hiromura K, Maezawa A et al. Case of propylthiouracil-induced vasculitis associated with antineutrophil cytoplasmic antibody; review of literature. Clinical Nephrol 1997;
47: 336–340
3. Vasily D, Tyler W. Propylthiouracil induced cutaneous vasculitis.
Case presentation and review of the literature. J Am Med Assoc
243: 458–461
4. Vogt B, Kim Y, Jennette J, Falk R, Burke B, Sinaiko A.
Antineutrophil cytoplasmic autoantibody-positive glomerulonephritis as a complication of treatment with propylthiouracil
in children. J Paediatrics 1994; 124: 986–988
5. Yuasa S, Hashimoto M, Yura T et al. Antineutrophil cytoplasmic
antibodies associated crescentic glomerulonephritis and propylthiouracil therapy. Nephron 1996; 73: 701–703
6. Tanemoto M, Miyakawa H, Hanai J, Yago M, Kitaoka M,
Uchida S. Myeloperoxidase-antineutrophil cytoplasmic antibody-positive crescentic glomerulonephritis complicating the
course of Grave’s disease; report of three adult cases. Am J Kid
Dis 1995; 26: 774–780
7. Ito K. Propythiouracil induced vasculitis. Jpn J Nephrol 1994;
36: 1354
8. Kudo Y. Propylthiouracil induced vasculitis. Jp J Nephrol 1995;
37: 701
9. Toda A. PTU induced vasculitis. Jpn J Nephrol 1995; 37: 701
10. D’Cruz D, Chesser A, Lightowler C et al. Antineutrophil
cytoplasmic antibody-positive crescentic glomerulonephritis
associated with anti-thyroid drug treatment. Br J Rheumatol
1995; 34: 1090–1091
11. Fauci a, Haynes B, Katz P. The spectrum of vasculitis. Ann
Internal Med 1978; 89: 660–667
L. Harper, P. Cockwell and C. O. S. Savage
12. Romas E, Henderson D, Kirkham B. Propylthiouracil therapy:
An unusual cause of antineutrophil cytoplasmic antibody associated alveolar haemorrhage. J Rheumatol 1995; 22: 803
13. Stankus S, Johnson N. Propylthiouracil-induced hypersensitivity
vasculitis presenting as respiratory failure. Chest 1992; 102:
1595–1596
14. Dolman KM, Goldschmeding R, Sonnenberg A, Von dem Borne
AEGK. ANCA related antigens. Acta Pathol Microbiol Immunol
Scand 1990; 98 [Suppl. 19]: 28
15. Fong P, Pun K, Tai Y, Wang C, Yeung R. Propylthiouracil
induced hypersensitivity with circumstantial evidence for drug
induced reversible sensorineural deafness: A case report. Horm
Res 1991; 35: 132–136
16. van der Woude FJ, Rasmussen N, Lobatto S et al.
Autoantibodies against neutrophils and monocytes: tool for
diagnosis and marker of disease activity in Wegener’s granulomatosis. Lancet 1985; i: 425–429
17. Cohen Tervaert JW, van der Woude FJ, Fauci AS et al.
Association between active Wegener’s granulomatosis and anticytoplasmic antibodies. Arch Intern Med 1989; 149: 2461–2465
18. Waldhauser L, Uetrecht J. Oxidation of propylthiouracil to
reactive metabolites by activated neutrophils. Implications for
agranulocytosis. Drug Metab Disposition 1991; 19: 354–359
19. Kallenberg CGM, Mulder AHL, Cohen Tervaert JW.
Antineutrophil cytoplasmic antibodies: a still-growing class of
autoantibodies in inflammatory disorders. Am J Med 1992;
93: 675–682
20. Wooley P, Griffen J, Panayi G. HLA-DR antigens and toxic
reactions to sodium aurothiomalate and D-penicillamine in
patients with rheumatoid arthritis. New Engl J Med 1980;
303: 300
21. Waldhauser L, Uetrecht J. Antibodies to myeloperoxidase in
propylthiouracil-induced autoimmune disease in the cat.
Toxicology 1996; 114: 155–162
Received for publication: 5.9.97
Accepted: 12.9.97