Journal of the New Zealand Medical Association, 04-April-2003, Vol 116 No 1171
Minocycline-induced lupus: a case series
David Porter and Andrew Harrison
The potential for medications to cause lupus-like illness has been recognised since 1945, when sulphadiazine was reported to cause a lupus-like syndrome.1 More than 70 drugs have since been reported as causing a lupus-like syndrome.2 The more notorious of these drugs, such as procainamide and hydralazine, are no longer in common use, but others such as anticonvulsants, isoniazid, and chlorpromazine are still widely prescribed.
Minocycline is a semisynthetic tetracycline commonly prescribed for the treatment of acne vulgaris, and occasionally prescribed as a disease-modifying anti-rheumatic drug (DMARD) for rheumatoid arthritis. Since 1992, minocycline has increasingly been recognised as a cause of several autoimmune conditions, the most commonly reported of which are a lupus-like syndrome and hepatitis.3 Serum sickness and ANCA-positive vasculitis have also been described.3
Differentiating minocycline-induced lupus (MIL) from idiopathic lupus, or systemic lupus erythematosus (SLE), is important as it has greatly different prognostic and therapeutic implications. We describe six patients with clinical features consistent with minocycline-induced lupus, who presented to rheumatology clinics in Wellington in 1998 and 1999, and review those features in relation to the published literature.
An 18-year-old woman was referred with a three-month history of arthralgia and swelling of the wrists, proximal interphalangeal (PIP) joints and feet, widespread myalgia and early-morning stiffness of 30 minutes. She had had a transient rash over the shins, which had resolved prior to her clinic appointment. Laboratory testing showed erythrocyte sedimentation rate (ESR) 49, C-reactive protein (CRP) 10, and a positive antinuclear antibody (ANA) test with a titre of 1:1280. She had evidence of hepatitis with alanine transaminase (ALT) 804 and aspartate transaminase (AST) 696. Tests for hepatitis A, B and C, cytomegalovirus (CMV), and Epstein-Barr virus (EBV) were negative, as were tests for rheumatoid factor (RF), antibodies to extractable nuclear antigens (ENA), and histones.
The patient had taken three courses of minocycline for acne in the year prior to symptom onset, each of three months’ duration. Within two weeks of stopping minocycline, symptoms had almost fully resolved, ALT had fallen to 95 and AST to 77. Three months after stopping minocycline, all symptoms had resolved, ESR had fallen to 8, ALT to 22 and ANA to 1:320.
A 19-year-old woman presented with a three-month history of intermittent fever, and progressive arthralgia and morning stiffness involving the metacarpophalangeal (MCP) and PIP joints, wrists, elbows, shoulders, neck, ankles and feet. She had been taking minocycline for nine months for acne vulgaris. Three years previously, she had taken a 12-month course of minocycline. Laboratory testing showed ESR 38, CRP 24, ANA 1:320, with negative tests for RF, anti-dsDNA, and ENA. She was advised to discontinue minocycline. Within four days of stopping minocycline, all inflammatory symptoms had fully resolved.
A 31-year-old woman developed pain and morning stiffness involving the neck, shoulders, elbows, wrists, and the joints of the fingers and thumbs, 12 months after commencing minocycline. She had additional symptoms of profound lethargy, malaise, intermittent fever and night sweats, and numbness and tingling in a stocking distribution. She had a previous history of autoimmune thyroiditis, and a brother had SLE with predominant skin and renal involvement.
On examination, there was a right knee effusion, left ankle synovitis, and a malar rash. Laboratory testing revealed an ESR of 58, CRP 62, positive ANA 1:80, negative dsDNA and ENA. Immediately after stopping minocycline, her fever, rash, lethargy, malaise and arthralgia fully resolved. ESR fell to 13 and CRP to 0. She remained asymptomatic with normal inflammatory markers six months later.
A 32-year-old woman had been under the care of a dermatologist for severe pustular eczema, and been successfully treated with minocycline and azathioprine for two years. After stopping treatment for four months, the eczema returned and treatment was restarted. Three months later she became unwell with fever, sweating, headache, arthritis of the feet and hands, and arthralgia of the knees and shoulders with prominent morning stiffness.
ESR was 114, ANA 1:2560, dsDNA and RF negative. After a month of symptoms, minocycline was stopped and doxycycline substituted. Within a week, fever, joint symptoms, sweating and headache had all resolved. Four months later, she remained asymptomatic, although ANA was further elevated at 1:5120.
A 16-year-old male had previously taken minocycline for 15 months, and after a short break had been taking it for a further six months. He presented with intense arthritic pain of two weeks’ duration, involving the wrists, hands and feet. On examination, there was diffuse swelling of the fingers and wrists, and bilateral ankle arthritis. ESR was not elevated. ANA was positive 1:5120; dsDNA antibodies, ENA and RF negative. Tests for antineutrophil cytoplasmic antibody (ANCA) were positive in a perinuclear (p-ANCA) distribution. Antihistone antibodies were strongly positive. There was a mild hepatitis with ALT 257. Tests for hepatitis A, B and C, CMV, and EBV were negative. On stopping minocycline, his ALT rapidly returned to normal but he continued to have significant joint pain.
Four months later, the patient had a flare of joint symptoms, and his ALT rose to 177. He was treated with pulsed methylprednisolone (1000 mg x 3), followed by weekly methotrexate (7.5 mg), daily prednisone (10 mg) and hydroxychloroquine (200 mg). He gradually weaned himself off all medications and had stopped all treatment within six months. At review a further six months later, he remained asymptomatic. His ANA titre was negative on repeated testing, and transaminases remained consistently normal.
A 16-year-old female had taken minocycline for five months. After a break of three months she had taken a further two-week course, shortly after which she developed severe arthralgia in the elbows, wrists and knees, with early-morning stiffness of 3–4 hours. She had a photosensitive rash on the cheeks and an effusion in the left knee. Initial blood tests showed ESR 27, CRP 2, ANA 1:160, p-ANCA positive, RF negative, and ALT 139. She was initially diagnosed as having MIL.
Over the next six months, after stopping minocycline, the patient complained of increasing lethargy, morning stiffness and arthralgia, and she developed synovitis of the PIP joints. ESR was persistently elevated and peaked at 101. She was given pulsed methylprednisolone on three consecutive days, which settled her symptoms. Three months later, she had developed hepatosplenomegaly, anti-dsDNA antibodies and anti-ribonuclear protein (RNP) antibodies, and the ANA titre had risen to 1:640. She had also developed anaemia, lymphopenia, and neutropenia. She was commenced on prednisone 20 mg daily, and methotrexate 10 mg weekly. Methotrexate made her feel unwell, so she stopped it after 12 weeks. Prednisone was tapered to 0 mg over 14 months. Six months after stopping prednisone, symptoms remained well controlled by nonsteroidal anti-inflammatory drugs alone.
A summary of all six cases is shown in Table 1.
Table 1. Summary of six cases
Since 1945, many drugs have been reported as having the potential to cause lupus-like illness. Several features of drug-induced illness that differ from idiopathic SLE have been recognised.4 Males and females are equally affected in drug-induced lupus, whereas idiopathic SLE affects nine times more females than males. Drug-induced disease occurs up to six times more frequently in Caucasians than blacks, whereas idiopathic SLE is common in blacks. The age of onset is older in drug-induced disease, reflecting the age at which people are exposed to causative drugs. Although fever, arthralgia, serositis and ANA occur at least as frequently in drug-induced disease as idiopathic SLE, haematological, renal and CNS involvement, and dsDNA are rare. Antihistone antibodies are found in up to 95% of drug-induced disease, but are not specific, being found also in 50–80% of idiopathic cases.4
MIL is similar to lupus-like disease caused by other drugs. The preponderance of young and female patients among those presenting with the condition, which contrasts with those presenting with other drug-induced causes of lupus-like syndromes, may simply reflect the population to which minocycline is prescribed. Hepatic involvement is more common in minocycline-induced disease. The first report of MIL appeared in 1992,5 and at least 100 cases have now been reported in the literature. A systematic review of reported cases,6 and two subsequently published case series,7,8 have highlighted a number of features as outlined below and in Table 2.
Affected patients have usually been exposed to minocycline for long periods (mean 19–25 months), although length of exposure before symptom onset has varied from three days to six years. Patients are usually young (mean age 21–24 years) and female (at least 75% of cases). Acne vulgaris is the usual indication for minocycline, although the condition has been reported in a patient receiving minocycline for rheumatoid arthritis.8
Inflammatory arthralgia or frank arthritis is invariably present. Other features include lethargy, malaise, and fever. Hepatic involvement is common, as is dermatological involvement, with livedo reticularis, vasculitis, malar rash, erythema, nodular rash, mouth ulceration and alopecia all having been described. Lung disease,9 pleurisy,10 and peripheral neuropathy8 have occurred. There have been no reports of renal or CNS involvement.
Laboratory findings usually include elevated inflammatory markers, a positive ANA, negative antibodies to dsDNA, and negative tests for antihistone antibodies, although exceptions to all of these findings have been described. Hepatic transaminases are elevated in approximately one third of cases.8 P-ANCA, when measured, is frequently positive,8,11 and this may have pathogenetic implications.11 Anaemia, lymphopenia and hypergammaglobulinemia have less frequently been reported. The persistence of ANA well after resolution of symptoms and other laboratory abnormalities is common.
Table 2. Features of idiopathic SLE, drug-induced lupus, and minocycline-induced lupus
The time to resolution of symptoms on stopping minocycline has been highly variable, from a few days to two years. A series of 20 patients quoted a mean time to resolution of 15.7 weeks.7 No treatment other than withdrawal of minocycline is usually necessary, but corticosteroid or DMARD therapy has occasionally been required to aid resolution of symptoms.
The first four patients in our series are consistent with the majority of cases reported in the literature. They were young women who had been exposed to minocycline for between 9 and 26 months. They presented with abrupt onset of inflammatory polyarthralgia, malaise, elevated inflammatory markers, and positive ANA. Two had fever, one had sensory neurological symptoms, and one had markedly elevated hepatic transaminases. Symptoms resolved on cessation of minocycline, with resolution taking from a few days to a few weeks. Laboratory abnormalities also resolved, although ANA persisted in at least the two patients for whom follow-up tests were performed.
Diagnostic doubt may arise when there is a long delay between withdrawal of minocycline and resolution of symptoms, or if corticosteroids are required to treat symptoms. Both of these issues arose in the 16-year-old male in our series (Case 5). Rechallenge with minocycline may help in such difficult cases. Patients with MIL who are rechallenged with a single dose of minocycline develop recurrent symptoms associated with a rise in CRP within 12–72 hours of taking minocycline.8
As the number of reported cases has grown, the distinction between idiopathic SLE and MIL has become increasingly blurred and differentiation between the two conditions may be difficult. All of the features of idiopathic SLE (with the exception of renal and CNS involvement) have been described in minocycline-induced disease. Adding to this confusion, at least two cases in which idiopathic SLE has apparently been precipitated by minocycline have also been mentioned in the literature.12 Idiopathic SLE often has a marked elevation of ESR in the presence of a normal CRP (unless serositis or coincident infection is present).8 Most cases of MIL have elevated CRP, so elevated CRP may help differentiate the two conditions.11
Case 6 in our series highlights these issues, with our patient initially being diagnosed with MIL, with the clinical features, p-ANCA and hepatic involvement supporting the diagnosis. However, she had stopped taking minocycline before symptoms became troublesome and, with the subsequent development of highly elevated ESR in the presence of normal CRP, hepatosplenomegaly, anaemia, lymphopenia, neutropenia, and positive dsDNA and RNP antibodies, it is probable that she in fact had idiopathic SLE. It may be that minocycline was the precipitant of SLE that would have occurred anyway. The fact that one of our cases had a brother with SLE also raises the possibility of minocycline acting as a precipitant in a genetically predisposed individual.
In the three years during which the above cases came to light there were only two cases of lupus caused by drugs other than minocycline seen by the Wellington Regional Rheumatology Unit. Although MIL is very rare in patients who are taking minocycline, the frequency with which the drug is prescribed may well mean that it is now the most common cause of drug-induced lupus. Scanty information exists in the literature about the risk to the individual of developing MIL. One study found an 8.5-fold risk of developing a lupus-like syndrome in current users of minocycline.13 The same study stressed the low absolute risk of developing lupus-like symptoms, with a rate of 52.8 cases per 100 000 prescriptions. Other tetracycline antibiotics were associated with a much lower risk of developing a lupus-like syndrome, with current use of doxycycline, oxytetracycline, or tetracycline combined being associated with a 1.7-fold increased risk.13
It is important that clinicians are aware of the risk of patients on prolonged courses of minocycline developing a lupus-like syndrome, so that inappropriate investigation and treatment can be avoided.
Author information: David Porter, Registrar; Andrew Harrison, Rheumatologist, Rheumatology Unit, Hutt Hospital
Correspondence: Dr Andrew Harrison, Rheumatology Unit, Hutt Hospital, Private Bag 31907, Lower Hutt. Fax (04) 570 9231; email: Andrew.Harrison@hvh.co.nz
issue | Search journal |
Archived issues | Classifieds
| Hotline (free ads)
Subscribe | Contribute | Advertise | Contact Us | Copyright | Other Journals