Rifampicin: not always an innocent drug

Abstract

Rifampicin has been widely used due to its broad antibacterial spectrum. Acute haemolysis is a rarely encountered complication of rifampicin. A 58-year-old woman was admitted to our department because of high-grade fever with rigors, accompanied by abdominal and lumbar pain and laboratory evidence of acute haemolysis. She had been treated for brucellosis initially with doxycycline and streptomycin. Due to subsequent appearance of myositis, ciprofloxacin and rifampicin were added for treatment of localised brucellosis. After intravenous administration of rifampicin, the patient deteriorated significantly. After exclusion of other causes of haemolysis, autoimmune haemolytic anaemia related to rifampicin was established by strongly positive direct Coombs test. Drug withdrawal in conjunction with intravenous immune globulin and prednisolone resulted in resolution of haemolysis and no relapse in the ensuing 1-year period. Our case highlights the importance of recognising commonly administrative drugs as cause of haemolytic anaemia, that can often be life threatening.

Background

Drug-induced immune haemolytic anaemia (DIIHA) is a rare condition with an estimated incidence of 1 case per 1 million population.¹ Since the first report of DIHA due to mephenytoin in early fifties, the range of implicated drugs has widened during the following decades as a result of significant innovations in medicine and drug development.² Nowadays, over 130 agents have been comprised in the list of drugs related to DIIHA and the three main categories include antimicrobials, non steroid anti-inflammatory drugs and chemotherapeutic agents.^{1 2}

The major suggested pathogenetic mechanism in DIIHA is the induction of drug-related antibodies, while non-immunological protein adsorption caused by some drugs is another implicated mechanism.^{3 4} Drug-related antibodies are divided to drug dependent, necessitating the drug to be in the test system or drug independent, being reactive without the drug presence in the test. The latter are serologically indistinguishable from idiopathic warm autoantibodies.¹ Haemolysis is often caused by activation of more than one mechanism. For the diagnosis of DIIHA, the following criteria are essential: existence of haemolytic anaemia, a temporal relationship with a specific drug at initiation of haemolysis, improvement of haematological parameters after drug withdrawal and a positive direct antiglobulin test (Coombs test).⁴

Herein, we describe a case of severe DIIHA due to rifampicin, a rare complication of a widely used drug and highlight the crucial steps in establishing the diagnosis in conjunction with approach to treatment.

Case presentation

A 58-year-old Caucasian woman with a recent history of well-established uncomplicated brucellosis based on positive blood cultures and positive Wright agglutination test was admitted to the department of medicine because of high- grade fever with rigors (38.5°C) accompanied by abdominal and lumbar pain and laboratory evidence of anaemia. She had been treated initially in the outpatient clinic with doxycycline (200 mg/day) for 6 weeks and streptomycin (1 g/day) intramuscularly for the first 2 weeks but she reported aggravated pain of the left hip despite treatment. This pain was reproduced by passive movement of the hip. Pelvic MRI revealed myositis of the left gluteus minimus muscle and 6 days before admission, oral ciprofloxacin and rifampicin were added to doxycycline, in order to treat localised brucellosis with a triple-drug regimen according to the Hellenic Society for Infectious Diseases and the Hellenic Centre for Disease Control and Prevention clinical practice guidelines for the diagnosis and treatment of infectious diseases.⁵ Her medical history was unremarkable apart from the recent history of brucellosis and she denied any travel history last year, the use of any drug in the past including rifampicin, as well as intravenous or nasal illicit drugs use. Physical examination on admission was unrevealing apart that she was pale and febrile.

On admission, the same three drugs were administered daily intravenously. During the second day of hospitalisation, the patient deteriorated with the presentation of rigors, dyspnoea, diarrhoea, nausea and vomiting. These symptoms were accompanied by abdominal and lumbar pain that progressively deteriorated in conjunction to the presence of dark urine. The patient appeared tired and icteric with altered mental status. The abdomen was soft, with diffuse tenderness that was greatest in the upper quadrants. No rash, dysuria or melena was observed.

Investigations

The laboratory workup on admission showed significant normochromic normocytic anaemia (haematocrit: 28.7%; haemoglobin: 10.1 g/dL) that had never existed before and raised serum inflammatory markers (table 1). On the second day of hospitalisation, blood tests showed an abrupt reduction of haemoglobin to 5.5 g/dL with normal white cells and platelets count, elevated reticulocytes and total bilirubin levels with the conjugated form within normal limits, increased ferritin and C reactive protein values and a sharp increase of lactate dehydrogenase while haptoglobin serum levels were almost undetectable (table 1).

Table 1.

Abnormal laboratory tests before and during hospitalisation

Variable	Before rifampicin	On admission	Hospitalisation day 2	Discharge day 14
Haematocrit (%)	37.1	28.7	10.0	32.7
Haemoglobin (g/dL)	12.4	10.1	5.5	11.0
Reticulocytes (%)			10.1	1.2
Haptoglobin (normal range: 30–200 mg/dL)			<7.3	75
Total bilirubin (mg/dL; ULN: 1.1)	0.18	1.39	4.70	0.6
Conjugated bilirubin (mg/dL; ULN: 0.4)		0.31	0.30
Lactate dehydrogenase (U/L; ULN: 480)	355	533	1165	338
Ferritin (ng/mL; ULN: 300)		2731	2780	1323
Creatine kinase (U/L; ULN: 145)	371	92	287	32
C reactive protein (mg/dL; ULN: 0.7)	0.68	5.7	7.1	0.4

ULN, upper limit of normal.

Peripheral blood smear revealed spherocytes, anisocytosis, nucleated red cells and polychromasia while haemoglobin electrophoresis was normal. The remaining haematological (including international normalised ratio and activated partial thromboplastine time), microbiological, virological and biochemical parameters including liver tests, blood, stools and urine cultures and investigation for hepatitis A, B and C and other viruses (Epstein-Barr virus, Cytomegalovirus, HIV, parvovirus B19), Mycoplasma, tuberculosis, malaria, leptospirosis, leishmaniasis, autoantibodies related to rheumatic diseases including antinuclear antibodies and antibodies against double-stranded DNA, rheumatoid factors, erythrocyte sedimentation rate and cryoglobulins were within normal limits. Chest X-ray, CT of the abdomen, investigation for Clostridium difficile infection, glomerular filtration rate and bone marrow biopsy and flow cytometry for potential underlying lymphoproliferative disorders were also unrevealing. However, the direct Coombs test was strongly positive (++++) for IgG and C3d, suggestive of an immune-mediated process of haemolysis.

Differential diagnosis

Taking into account the above findings and that the condition of the patient deteriorated significantly during the first 2 days of hospitalisation, after rifampicin administration, we considered a diagnosis of acute haemolytic anaemia due to DIIHA (rifampicin) highly likely as the patient had no specific family history of anaemia and her medical history was unrevealing (absence of lifelong anaemia) while there was no evidence to support a diagnosis of inherited haemolytic anaemia. The possibility of DIIHA due to ciprofloxacin administration was rather unlikely as the development of DIIHA because of the use of fluoroquinolones is not common. In addition, the patient had normal platelets count and kidney function, while no schistocytes were observed on the peripheral blood smear rather excluding causes of microangiopathic haemolytic anaemia.

Treatment

Immediate discontinuation of rifampicin (900 mg/day) was done on the third day of hospitalisation and considering the severity of intravascular haemolysis, the patient was treated with intravenous immune globulin (400 mg/kg/day for five consecutive days) together with prednisolone (initial dose 1.5 mg/kg/day) that was rapidly tapered off in 1 month. In addition, aggressive expansion of plasma volume was performed with intravenous infusion of normal saline, as a prophylaxis against acute kidney injury due to severe intravascular haemolysis. Regarding treatment for localised brucellosis, rifampicin was replaced by trimethoprim–sulfamethoxazole (800+160 mg two times a day).

Outcome and follow-up

The patient responded quickly with clinical and laboratory improvement (table 1) and was discharged in good health after 14 days of hospitalisation, with no evidence of relapse of haemolysis in a follow-up period of 1 year while the localised brucellosis was also resolved after a total of 3 months treatment duration taking into account as first day of treatment the initiation of doxycycline with streptomycin.

Discussion

Rifampicin is an old antimicrobial agent that belongs to the class of rifamycins. Despite its association with several adverse events, it is considered safe and easy to use drug. However, discontinuation of treatment as a result of adverse events is a rare phenomenon, particularly if patients are managed by experienced physicians.^{6 7} Review of the existing literature on rifampicin-induced haemolysis revealed a limited number of reports, including some originated back in the seventies and some not written in the English language.^8–17

First reports on rifampicin-related haemolysis were thought to occur only in patients under intermittent high-dose treatment, suggesting these conditions to be a prerequisite for the increased formation of rifampicin-dependent autoantibodies.¹⁸ It has been previously suggested that continuous formation and clearance of rifampicin-antibody complexes could prevent the increase of antibody titres in those on daily rifampicin treatment.¹⁹ In this context, daily rifampicin had been considered to carry a low risk for developing severe complications related to formation of antibodies after exposure to the drug, including haemolytic anaemia and acute kidney injury.¹⁹

Still, with the increasing use of low-dose rifampicin, few cases of intravascular haemolysis have been reported related to daily drug administration.²⁰ It is important for clinicians to keep in mind this rare complication, because late recognition could be fatal.^{4 16} Key points for early recognition of this process are signs of haemolysis (red or dark urine) and other more insidious symptoms like abdominal and/or back pain. However, other more common causes of haemolytic anaemia like microangiopathy, infections, hereditary causes, warm autoimmune haemolytic anaemia and transfusion-associated haemolysis should be quickly and appropriately excluded. In this context, we should keep in mind that autoimmune haemolytic anaemia has also been reported as a rare complication in the spectrum of brucellosis.²¹ However, in our case anaemia developed after the addition of rifampicin and was worsened dramatically after the intravenous administration of the drug. In addition, there was a temporal association between intravenous administration of rifampicin and the deterioration of symptoms making the diagnosis of brucellosis-related autoimmune haemolytic anaemia very unlikely.

The responsible mechanism for rifampicin-induced haemolysis is the existence of rifampicin-dependent antibodies that are directed at I antigens located at the surface of red blood cells (RBCs). As a consequence complement is activated and severe intravascular haemolysis can occur as noticed in our case.^{1 4} Drug-dependent antibodies are investigated by testing drug-treated RBCs or by testing RBCs in the presence of a drug solution.²² Considering that the pathophysiology has been well described, the fact that we have not performed specific evaluation for the detection of rifampicin-specific antibodies cannot debilitate our diagnosis as this test is not widely available for everyday clinical practice.

With regard to the treatment of DIIHA, withdrawal of the suspected drug is the cornerstone of management.⁴ Nevertheless, in some cases drug withdrawal is not sufficient as the only therapeutic measure as is the case of cefotetan.²³ The benefit of steroids is not clear and often physicians are not able to distinguish whether a favourable outcome can be attributed to drug withdrawal per se or to the administration of steroids.¹ Accordingly, experts advise management on an individual basis depending on the severity of haemolysis and strength of clinical suspicion of DIIHA.²⁴ Considering our patient to suffer from autoimmune-driven severe intravascular haemolytic process with a sharp decrease in haemoglobin level, we decided to administer a combination of immunoglobulin that blocks macrophage Fc receptors and steroids, which suppress macrophage activation, to encounter the autoimmune process.²⁵

In conclusion, our case highlights the importance of recognising commonly administrative drugs like rifampicin, as cause of haemolytic anaemia, that can often be life threatening. After excluding other cases of haemolysis, prompt drug withdrawal is the cornerstone of treatment, while combination of immunoglobulin and steroids can be useful tools to tackle the autoimmune process in cases of severe intravascular haemolysis.

Learning points.

• Rifampicin has been established as backbone treatment for tuberculosis, while it has been widely used in various antibiotic combinations for bacterial infections.

• Autoimmune-driven acute haemolysis is a rarely encountered complication of rifampicin.

• Prompt drug withdrawal is the cornerstone of treatment, while combination of immunoglobulin and steroids can be useful tools to tackle the autoimmune process in cases of severe intravascular haemolysis.

• Our case highlights the importance of recognising commonly administrative drugs as cause of haemolytic anaemia, that can often be life threatening.