Chronic dialysis, NAT2 polymorphisms, and the risk of isoniazid-induced encephalopathy – case report and literature review

Despite recent advances and the fact that nearly all cases can be cured, tuberculosis remains one of the world’s biggest threats. In 2014, 9.6 million people developed tuberculosis and 1.5 million died from the disease [1]. Isoniazid is one of the oldest but also the most widely used anti-tuberculosis agent; it is easily administered, inexpensive, and exerts a high bactericidal activity against Mycobacterium tuberculosis.

However, isoniazid may lead to serious adverse events, including hepatotoxicity, peripheral neuropathy and encephalopathy, especially in populations at risk [2]. The following case illustrates the impact of end-stage renal disease and NAT2 polymorphisms on the risk of isoniazid toxicity, and discusses how these life-threatening complications may be prevented and managed.

A 71-year-old African woman on chronic dialysis was referred to our institution because of unexplained lethargy and altered state of consciousness. Hemodialysis – using high-efficiency/high-flux membranes - had been initiated 5 years earlier because of end-stage nephrosclerosis. No significant complication occurred since the onset of dialysis. Her neurological status was normal until 6 weeks before her transfer, when she developed confusion and, over a few days, unresponsiveness. An extensive workup had been performed at another university hospital and found no evidence for a metabolic or vascular etiology, nor meningo-encephalitis; the electroencephalogram showed severe generalized abnormalities in keeping with a diagnosis of encephalopathy. The patient was transferred to our hospital for further investigation, upon request of the family.

At admission, blood pressure was 120/70 mmHg and heart rate regular at 86 bpm. She was somnolent, opened her eyes to verbal stimuli, uttered only sounds and had an avoiding response to pain. She scored 9/15 on the Glasgow coma scale. Two months before her transfer, she had undergone a workup for a chronic dry cough, including chest computed tomography-scan, positron emission tomography, bronchoalveolar lavage and interferon-gamma release assay. The results suggested latent tuberculosis infection, and the patient was prescribed isoniazid 300 mg o.d. along with weekly pyridoxine 250 mg. Altered mental status with somnolence then unresponsiveness developed two weeks after the start of isoniazid. Three hours after intake of isoniazid, blood level of the drug was 2.89 mg/l (normal range 1–2 mg/l), and the ratio of inactive acetyl-isoniazid to isoniazid was 0.33, a result compatible with a slow acetylator phenotype (<0.48, slow acetylators; >0.77, fast acetylators) [3]. Slow acetylator phenotype was confirmed by sequencing of NAT2 – the gene encoding the human isoenzyme arylamine N-acetyltransferase 2 (NAT2) (NM_000015.2) – which showed a heterozygous status for the variants c.282C > T(rs1041983), c.341 T > C(rs1801280), c.481C > T(rs1799929), c.590G > A(rs1799930) and c.803G > A(rs1208), compatible with a NAT2*5B/*6A genotype (Fig. 1).

Isoniazid was discontinued soon after admission in our unit, and the patient was given high doses of oral pyridoxine (250 mg daily, equivalent to 1750 mg weekly). A remarkable improvement of the patient’s condition ensued, and Glasgow coma scale was 15/15 ten days later. She was discharged two weeks after admission and was prescribed physical therapy at home, in addition to continuing maintenance hemodialysis. Three months after discharge she had returned to her previous condition, being again able to have a full conversation and walk on her own. The life-threatening complication of isoniazid therapy, along with the debatable indication of treatment, prompted watchful waiting rather than reintroduction of any anti-tubercular treatment. Clinical follow-up was excellent, and a computed tomography of the chest performed one year later did not show any sign suggestive of tuberculosis infection.

More than 2 million people worldwide have end-stage renal disease and require dialysis as a life-saving therapy [4]. In the dialysis population, Mycobacterium tuberculosis infection has a high prevalence, ranging between 5 and 25%, and isoniazid is frequently used to prevent reactivation [5]. Latent tuberculosis is frequently diagnosed in renal transplant candidates, who all should be screened during pre-transplantation evaluation, as the risk of active tuberculosis in transplant recipients is estimated to be 20–74 times higher than in the general population [6]. Patients on dialysis are at risk for isoniazid toxicity, because of altered pyridoxine metabolism resulting in severe deficiency in pyridoxal phosphate (the active form of pyridoxine), and extensive removal of pyridoxal phosphate (a small molecule of 247 Da) by renal replacement therapies [7]. While the recommended dose of pyridoxine supplementation in case of isoniazid treatment is usually 10–25 mg per day, [8] it needs to be increased to >100 mg/day to prevent drug toxicity among patients on dialysis [7]. Patients on high-flux high-efficiency hemodialysis – as our patient – are particularly at risk of pyridoxal phosphate depletion because of higher clearances during dialysis sessions [9].

Isoniazid is primarily metabolized by the polymorphic arylamine NAT2. NAT2 activity is genetically determined and basically relies on the number of active alleles in NAT2 gene (NAT2*4 and *12), thereby providing a molecular mechanism for the large inter-individual variability in toxicity and efficacy of isoniazid [10]. Individuals are therefore classified as rapid metabolizers if they have one or more NAT2*4 alleles, and slow metabolizers only if they carry two slow metabolizer variants, as our patient (Fig. 1). During treatment with the standard regimen, slow acetylators carry a higher risk of drug toxicity, while rapid acetylators are prone to treatment failure, due to insufficient exposure to isoniazid. As a result, it has been suggested that NAT2 genotype may help determining the dose of isoniazid in individual patients [11]. The appropriateness of such a pharmacogenetics-based has been recently validated in a large randomized controlled trial. NAT2 genotype-guided regimen significantly reduced the risk of both toxicity and treatment failure, in slow and rapid acetylators, respectively [12]. In our patient, the combination of increased isoniazid and decreased acetyl-ionazid levels pointed towards a slow acetylator phenotype, an hypothesis that was further verified by the identification of NAT2*5B/*6A genotype. Intriguingly, Sub-Saharan populations show a high prevalence (>70%) of slow acetylator phenotype related to NAT2 polymorphisms, possibly as a consequence of selective pressure related to the chemical environment, climate, biome, and dietary habits [13].

In conclusion, the combination of end-stage renal disease, removal of pyridoxal phosphate by hemodialysis, and deleterious NAT2 polymorphisms, contributed to the severe toxicity of isoniazid in our patient. Drug withdrawal and administration of high doses of pyridoxine rapidly reversed drug toxicity. Physicians should be aware of the increased risk of isoniazid toxicity in patients on dialysis and in those with a slow acetylator phenotype conferred by NAT2 polymorphisms. Adaptation of prescription – either with higher doses of pyridoxine or decreased doses of isoniazid, respectively – has been suggested to effectively reduce the risk of potentially life-threatening toxicity of isoniazid, while avoiding treatment failure. Future studies will need to validate NAT2 genotype-guided drug dosage of isoniazid as a safe and appropriate alternative to standard dose regimen.