Tuberculosis remains an important public health problem, particularly in large, inner-city, urban populations such as exist in New York City. Adverse effects of isoniazid (INH), the mainstay of treatment and prevention of tuberculosis, are commonly seen following both overdose and therapeutic use. We report a case of an adolescent who experienced both a psychotic break after initiating INH and seizures after overdosing on INH in response to her psychotic delusions.
A 14-year-old girl was brought to the hospital by city Emergency Medical Services following several generalized tonic-clonic seizures at home. She had been in her usual state of health until several hours prior to arrival at the hospital, when she complained of abdominal and chest pain and “felt [she] was going to die.” In addition, she reported paresthesias of her fingers and toes and dazzling lights in her vision. One hour after onset of these symptoms, the patient suffered the 1st of 4 generalized tonic-clonic seizures, each reported to last a few minutes. Between seizures, she was reported to be agitated and unresponsive to commands.
The family denied any history of fever, trauma, or seizure disorder. Specifically, they denied any previous history of visual disturbances, paresthesias, psychiatric disease, suicide attempts, or substance abuse. The family had moved to the Bronx from Guyana 3 months prior to this illness. There was no history of psychiatric illness in the family. Both the patient and her sister had recently been prescribed isoniazid and pyridoxine for tuberculin skin test conversion. One day prior to admission, the sisters were started on a prophylactic regimen of isoniazid 300 mg and pyridoxine 50 mg per day. Examination of the patient’s medication bottles revealed 3 pills missing from each container. The family denied the presence of other medications or drugs in the home.
On arrival, the patient was intermittently combative, responding only to painful stimuli. She was afebrile with a pulse rate of 94 beats per minute and a blood pressure of 110/60 mmHg. She was breathing without difficulty and had a normal pulse oximetry reading in room air. Aside from her altered mental status, her neurologic examination was nonfocal. Pupils were midsize and normally reactive. Cardiac and respiratory examinations were unremarkable. Cardiac monitoring displayed a normal sinus rhythm.
Shortly after arrival, the patient had 2 additional brief generalized tonic-clonic seizures. She was treated with intravenous lorazepam, phenytoin, and pyridoxine. Chest radiography, computer tomography scan of the head, and analysis of spinal fluid were normal. Serum electrolyte analysis revealed an anion gap metabolic acidosis of 20; serum glucose was 217 mg/dL. Serum levels of acetaminophen, salicylate, and ethanol were undetectable.
As the patient’s mental status normalized gradually over the subsequent 3 days, she was able to supply additional historical details of her illness. She reported that 1 day following the initiation of INH and pyridoxine, she experienced auditory hallucinations. She recalled hearing a voice that commanded her to “take the pills.” In an attempt to obey this command, she obtained 20 INH pills from a friend and ingested them all. She remembered feeling ill and seeing “colors and rainbows” shortly afterward. She could not recall subsequent events until her return to consciousness at the hospital. Psychiatric evaluation during hospital recovery found the patient neither actively suicidal nor psychotic. The patient was discharged from the hospital with a diagnosis of INH-induced seizures and psychosis classified as either toxic, brief reactive, or not otherwise specified. At 1 week following the overdose, the patient’s mother reported no subsequent psychotic symptoms, hallucinations, suicidal ideation, or seizures. Follow-up psychiatric evaluation several months later revealed no recurrence of psychiatric symptoms.
Tuberculosis accounts for more morbidity and death than any other single infectious disease (1). In particular, urban environments have had very high local rates in this decade, with over 200 cases per 100,000 people per year (2). The highest rates of tuberculosis in the US occur in recent immigrants from high-risk countries, the indigent, and those with immunodeficiency states. With the rise in tuberculosis, the use of INH has likewise increased and remains the mainstay of prophylaxis and treatment (3). INH has been used for the treatment of tuberculosis since 1952 (4). It is highly effective in treating symptomatic patients with tuberculosis and preventing primary disease. The acute neurotoxic effects of INH in overdose are well known. Therapeutic at doses of 5 to 10 mg/kg/day (maximum 300 mg/day), seizures can occur with doses as low as 20 mg/kg, and high mortality has been reported following overdoses of 80 to 150 mg/kg (5). Liver injury results from conversion of hydrazine intermediates by P450 isoenzymes to a directly hepatotoxic compound. The peripheral neuropathy associated with chronic INH use is a distal sensory-motor axonopathy which is exacerbated by a pyridoxine-deficient diet. There are other, less frequently reported, adverse effects of INH, including optic neuritis, hematologic abnormalities, hypersensitivity reactions, a pellagra syndrome, a tyramine-like reaction, vasculitis, encephalopathy, and various psychologic complaints (6). Rarely, psychosis has been reported in patients taking INH. We report a case of an adolescent who attempted suicide by drug overdose during an apparent psychotic break that occurred while taking INH. In this case, the drug of overdose was also INH.
INH is well absorbed from the gastrointestinal tract. Peak levels are reached within 1 to 2 hours of ingestion of a therapeutic dose. It distributes to a volume of distribution of 0.6 L/kg. INH then undergoes a complex metabolic transformation, which is predominantly hepatic and characterized by multiple pathways and intermediates. Among other reactions, the drug undergoes 2 acetylation steps with an intervening hydroxylation reaction. The rate of acetylation is genetically determined. People of Asian and African descent are predominantly rapid acetylators, whereas whites tend to be slow acetylators. Plasma half-life of the parent drug is about 70 to 80 minutes in rapid acetylators and 4 to 6 hours in slow acetylators (7). It is uncertain what effect, if any, acetylator status has upon the incidence of other adverse reactions. Following biotransformation, INH undergoes renal excretion in the form of its several metabolites.
Psychosis associated with therapeutic INH occurs infrequently, but its actual incidence rate is not well established. As early as 1957, merely 5 years after INH came into general use for tuberculosis, 5 cases of psychiatric disturbances in patients taking INH were described (8). Of these, 2 actually exhibited psychotic symptoms, one with paranoid delusions and the other with gender identity delusions. Both patients were taking relatively low doses of INH (2.6–2.9 mg/kg) in addition to other antituberculous drugs (para-amino salicylic acid and streptomycin). Psychotic symptoms occurred 7 to 8 weeks following initiation of the drug in one patient, and 1 week after initiation in the other. Both experienced improvement within a few months of discontinuation of INH.
Noting that a history of psychiatric disorder seemed to be a risk factor for recurrence of psychiatric symptoms while taking INH, Duncan and Kerr (9) discussed 38 cases of INH-associated psychosis in patients who had no history of psychiatric illness. They described observations of a prodrome followed by a “psychotic phase” in most of these patients. For a period of days to weeks preceding the onset of frank psychosis, patients complained of paresthesias and anxiety. They exhibited emotional lability and facial twitching. These symptoms progressed to agitation, irrational fear, confusion, paranoid delusions, and hallucinations, both auditory and visual. Other general observations of this group included dosage regimens up to 12 mg/kg/day and a latency to onset of psychiatric symptoms ranging from 2 days to 10 months. Neither variable seems to contribute to severity of psychosis or ultimate prognosis. Duncan and Kerr postulated that concomitant administration of para-amino salicylic acid (PAS), known to be hepatotoxic, may have decreased INH clearance and contributed to the toxic psychosis but found no evidence for this claim within their case series. Twenty-five patients experienced resolution of psychiatric symptoms within 2 weeks to greater than 10 weeks after INH discontinuation. INH was restarted in 1 patient without adverse events, but reintroduction of the drug in 2 other patients led to recurrence of psychotic symptoms. The authors stressed that discontinuation of INH had the greatest impact on resolution of psychotic symptoms, and that the benefits of pyridoxine administration were unproved.
In a case series of 37 patients with INH-associated psychiatric symptoms, a prodromal phase consisting of headaches, dizziness, inability to concentrate, and irritability was also described. Lasting for a mean of 5 days, these symptoms usually occurred within 3 months of initiating INH therapy (4). Including patients with a wide variety of psychiatric complaints, some patients suffered from delirium with psychotic features, whereas others exhibited predominantly neurotic symptoms. One subset of eight patients exhibited more classically schizophreniform delusions uncomplicated by alteration of consciousness or dysautonomia. In these patients, symptoms resolved within 3 to 6 weeks after treatment with neuroleptics and vitamins despite continuation of INH.
A latency of 3 days to 8 weeks between initiation of INH and onset of psychiatric symptoms was observed in a case report and review of eight other patients from the literature in 1993 (10). INH had been used both alone and in combination with numerous other antituberculous agents in these patients. In general, symptoms remitted when INH was withdrawn. In 1 case, symptoms recurred when INH was reintroduced. However, a few other patients were able to continue INH with the addition of neuroleptics.
Since this report, cases of INH-associated psychosis have occurred sporadically in the literature. In a letter, one physician describes the compelling case of a 64-year-old man who developed the dramatic onset of visual hallucinations and disruptive behavior 12 days after initiation of a tuberculosis prophylaxis regimen of 300 mg INH and 50 mg pyridoxine per day (11). The man, although a recovering alcoholic, had no history of psychiatric illness. A comprehensive evaluation for organic etiology was unrevealing. The patient’s symptoms resolved completely after discontinuation of INH and 48 hours of treatment with haloperidol and lorazepam. He remained free of symptoms 1 year later. A 31-year-old woman with no psychiatric history developed paranoid delusions after 8 weeks of taking INH and pyridoxine for a positive tuberculin skin test. In her case, symptoms abated only minimally after INH was discontinued. She responded to 6 weeks of therapy with an antipsychotic medication and was symptom-free 11 months later (12).
The toxicity of INH is complex and not completely understood. Several of its actions may contribute to psychiatric adverse effects. INH metabolites inhibit the activation of pyridoxine to pyridoxal 5-phosphate. The latter is a cofactor of the enzyme glutamic acid decarboxylase which catalyzes the conversion of glutamic acid to gamma aminobutyric acid (GABA), the major inhibitory neurotransmitter in the central nervous system. The resulting GABA depletion leads to central nervous system disinhibition and, clinically, seizures following INH overdose. It is unclear whether the observed prodromal symptoms and frank psychosis in patients taking INH represent a lesser degree of GABA-mediated disinhibition or a different pathophysiology entirely. INH is closely related to iproniazid, another antituberculous agent which, because of its monoamine oxidase inhibitor (MAOI) activity, was used as an antidepressant in the 1950s (13). Decreased intracellular degradation and resultant accumulation of catecholamines within nerve terminals presents the risk for sympathetic crisis upon release of the catecholamines. INH therapy may be associated with a pellagra-like syndrome with dementia, diarrhea, and dermatitis. Niacin deficiency results from the INH-mediated depletion of pyridoxal phosphate which is also required as a cofactor for the synthesis of niacin from tryptophan.
At the root of the difficulty with isoniazid-associated psychosis lies the incomplete understanding of the pathogenesis of psychosis itself (14). Both overproduction and underproduction of catecholamines have been implicated, obviously oversimplified theories at best. The reported cases strongly suggest an association of toxic psychosis with isoniazid in therapeutic doses. There exists a wide variation both in latency from initiation of the drug to onset of psychiatric symptoms and interval from discontinuation of isoniazid to clinical improvement. In our patient, the period of less than 1 day from initiation of isoniazid to onset of auditory hallucinations is the briefest latent period of isoniazid-associated psychosis to our knowledge. The lack of preexisting psychiatric illness and complete resolution of hallucinations and suicidal ideation upon discontinuation of the drug strongly supports the diagnosis of toxic psychosis due to isoniazid in this patient. The role of pyridoxine supplementation in prevention or treatment of isoniazid associated psychiatric symptoms is unclear. Most cases of psychosis either resolved or dramatically improved upon withdrawal of isoniazid. One recommendation that emerges from the collective experience is that isoniazid discontinuation should be the primary intervention.
1. Kahn EA, Storke JR. Tuberculosis. In: Burg, et al., eds. Current Pediatric Therapy. Philadelphia: WB Saunders; 1999:91.
2. Frieden TR, Fujiwara PI, Washko RM, et al. Tuberculosis in New York City - Turning the tide. N Engl J Med 1995; 333:229–233.
3. Centers for Disease Control and Prevention. Guidelines for preventing the transmission of Mycobacterium tuberculosis in health-care facilities, 1994. MMWR Morb Mortal Wkly Rep 1994;43(RR-13):1–132.
4. Wasik A. Mental disorders caused by isonicotinic acid hydrazide (INH) in the course of treatment of pulmonary tuberculosis. Pol Med J 1970; 9:1498–1503.
5. Sievers ML, Herrier RN. Treatment of acute isoniazid toxicity
. Am J Hosp Pharm 1975; 32:202–206.
6. Pitts FW. Tuberculosis: Prevention and therapy. In: Hook EW, Mandell GL, Gwaltney JM, et al., eds. Current Concepts of Infectious Diseases. New York: Wiley; 1977:181–194.
7. Jeanes CW, Schaefer O, Eidus L. Inactivation of isoniazid
by Canadian Eskimos and Indians. Can Med Assoc J 1972; 106:331–335.
8. Jackson SLO. Psychosis
due to isoniazid
. BMJ 1957; 2:743–746.
9. Duncan H, Kerr D. Toxic psychosis
due to isoniazid
. Brit J Dis Chest 1962; 56:131–138.
10. Pallone KA, Goldman MP, Fuller MA. Isoniazid
: Case report and review of the literature. Ann Pharmacother 1993; 27:167–170.
11. Ibrahim ZY, Menke JJ. Comment: Isoniazid
[letter]. Ann Pharmacother 1994; 28:1311.
12. Alao AO, Yolles JC. Isoniazid
. Ann Pharmacother 1998; 32:889–891.
13. Jacobson E. The early history of psychotherapeutic drugs. Psychopharmacology 1986; 89:138–144.
14. Davis KL, Kahn RS, Ko G, et al. Dopamine in schizophrenia: A review and reconceptualization. Am J Psychiatry 1991; 148:1474–1486.