Education and debate

Grand Rounds--Hammersmith Hospital: Persistent fever in pulmonary tuberculosis

Drug malabsorption should be considered

Malabsorption of rifampicin and other antituberculous drugs is becoming an increasing problem, particularly in HIV positive patients with tuberculosis. Its occurrence presents a formidable challenge to doctors as persistently low drug doses can result in multiple drug resistance. We describe the case of an HIV negative man with pulmonary tuberculosis and a persistent fever, despite two months treatment with rifampicin, isoniazid, and pyrazinamide. We discuss possible causes of this fever, including cytokine release, drug induced fever, drug resistance, and drug malabsorption. The case presented here highlights the difficulties in detecting the malabsorption.

Case history

A 47 year old white, unemployed factory worker presented with a four month history of cough, sputum, weight loss, and anorexia. His bowel habit was normal with no diarrhoea or steatorrhoea. He had a history of chronic schizophrenia, and his only other serious illness was a pneumococcal pneumonia in 1990. He was taking no drug treatment, smoked 20-30 cigarettes a day, and drank 40 units of alcohol a week.

On examination he was unwell and cachectic. He had a fever of 38.3°C and had finger clubbing. He was normotensive (pulse 90 beats/min). Auscultation of the chest revealed crepitations and bronchial breathing at the right base. Chest radiography showed multiple cavitating lung lesions, and sputum was positive for acid fast bacilli. Pulmonary tuberculosis was diagnosed, and combination therapy (rifampicin, isoniazid, and pyrazinamide) was started with three tablets a day (daily total: rifampicin 360 mg, isoniazid 150 mg, pyrazinamide 900 mg). His HIV status was negative.

After three weeks of supervised hospital treatment, the patient had orange urine, but his swinging fever persisted (up to 39°C). Repeated blood cultures were negative, and an abdominal ultrasound scan was normal. A two week course of prednisolone 50 mg daily made no difference to the fevers, and after an episode of hypotensive collapse the prednisolone was changed to intravenous hydrocortisone 100 mg four times a day. Intramuscular streptomycin 500 mg was added to cover the possibility of a resistant mycobacterial strain. Nasogastric feeding was also started for albumin support.

The fevers settled with the intravenous hydrocortisone, and a week later this was exchanged for oral prednisolone (60 mg daily), at which point the fevers returned. Sensitivities of the original sputum became available: the Mycobacterium tuberculosis was sensitive to rifampicin, isoniazid, and pyrazinamide. Drug induced fever was excluded as 48 hours without rifampicin, isoniazid, and pyrazinamide made no difference to the swings in fever. Daily streptomycin was continued during this period.

The patient failed to improve, and repeat sputum analysis two months into treatment showed persistence of acid fast bacilli. Serum rifampicin concentrations were 0.6 µg/ml one hour after a dose and 0.5 µg/ml three hours after (therapeutic range 6-10 µg/ml). It was not possible to measure isoniazid and pyrazinamide concentrations. The patient was prescribed daily intravenous rifampicin 400 mg and intravenous isoniazid 300 mg. He also continued taking daily oral pyrazinamide 1.5 g, intramuscular streptomycin 500 mg, oral pyridoxine (now increased from 10 mg to 20 mg empirically), and oral prednisolone 60 mg. His urine turned a deeper orange. On this regimen the fever settled and he recovered. Table 1 shows the blood test results, and figure 1 summarises the response of the patient's temperature during the different combinations of drugs.

Table 1--Patient's results of blood tests during course of treatment*
--------------------------------------------------------------------------------------------------------------
Test                                    31 Dec  2 Jan  25 Jan  30 Jan  8 Feb  13 Feb  26 Feb  7 Mar   11 Mar
--------------------------------------------------------------------------------------------------------------
C reactive protein (U/I)                        155       92    139    139       9     59       68     123
Erythrocyte sedimentation rate
  (mm in first hour)                                      70            50      12     38       34
White cell count (x109/l)          8.4     9.0      21     20    21.9    15.3   18.2     13.1    11.2
Haemoglobin (g/l)                         88    108      102    111     88     113   109       111    120
Platelet count (x109/l)            353    438      936    932    570     519   721       575    659
Mean cell volume (fl)                   83.8     78.5     82            82.7    87.8  89.4      88.6   91.3
Plasma sodium (mmol/l)                   127    132      128    135    128     137   136       133    140
Plasma potassium (mmol/l)                5.5     4.2     4.4    4.1     4.6     3.7   4.8       4.0    4.1
Plasma urea (mmol/l)                     3.8     3.1     2.4    2.9     3.7     4.2   4.8       6.8    3.3
Plasma creatinine (µmol/l)             49    43        49     45     48     40    46        47     41
Plasma glucose (mmol/l)                          5.8
Plasma albumin (g/l)                            19        17     21     21     20    29        30     33
Serum bilirubin (µmol/l)                     21         9     10     12      7     8         9      7
Serum alkaline phosphatase (U/I)                78       130    142    116     79    67        63     61
Serum aspartate aminotransferase (U/I)          26        12     13     16     12    16        21     17
Serum iron (µmol/l)                           1.7
--------------------------------------------------------------------------------------------------------------
*See figure 1 for the four phases of treatment.

View larger version (25K): [in this window] [in a new window]   Fig 1--Temperature chart of patient during four phases of treatment. Phase 1: oral rifampicin, isoniazid, and pyrazinamide and oral prednisolone (swinging fever); phase 2: oral rifampicin, isoniazid, and pyrazinamide, intramuscular streptomycin, and intravenous hydrocortisone (fever settles); phase 3: oral rifampicin, isoniazid, and pyrazinamide, intramuscular streptomycin, and oral prednisolone (fever returns); phase 4: intravenous rifampicin and isoniazid, oral pyrazinamide, intramuscular streptomycin, and oral prednisolone (fever settles). Daily streptomycin was continued between phases 3 and 4

The cause of the low rifampicin concentrations was not clear, but malabsorption was suspected. The patient, however, refused endoscopies and small bowel biopsies, barium studies, a pancrealauryl test, and a glucose hydrogen breath test. The patient's serum B12, serum and red cell folate, plasma calcium, and plasma magnesium concentrations were all normal, and the low plasma albumin concentration was corrected with nasogastric feeding. Stool microscopy showed no fat globules.

Comment

Possible causes of persistent fever in pulmonary tuberculosis (once non-compliance and supra-added infections have been excluded) include cytokine release, drug induced fever, drug resistance, and drug malabsorption.

Proinflammatory cytokine release is the host response to mycobacterial persistence.^{1 2} Mitchison reviewed the published triple or quadruple drug regimens between 1973 and 1982.³ Among those which included rifampicin, isoniazid, and pyrazinamide, only 48% of patients had culture negative sputum by one month, 88% by two months, and 97% by three months. Even after three months of treatment 3% of patients will therefore have viable M tuberculosis in their sputum.

The host response to the mycobacterial persistence can be modulated by steroids.^{4 5 6} Adjunctive steroid therapy is lifesaving in miliary and cardiac tuberculosis and in tuberculosis affecting the central nervous system and may be beneficial in cavitating pulmonary tuberculosis, endobronchial tuberculosis--in which large pleural effusions occur--and in severe hypoxaemia. Such treatment is unsuitable for mild to moderate disease without toxaemia, old fibrocavitary disease, and when there is resistance to one or more drugs.

Drugs are also a recognised cause of a fever, but these (particularly rifampicin and isoniazid) usually cause raised aminotransferases, a relative bradycardia, and an erythrocyte sedimentation rate that initially improves then rises again.⁵ A drug induced fever should resolve within 48 hours of stopping the drug(s).

A more worrying cause of a persistent fever is the presence of resistant strains of mycobacteria. This is particularly problematic as sensitivities take 6-8 weeks to obtain. In addition, initial sensitivities may change. It is therefore essential for all patients producing sputum to have analysis both for mycobacterial quantification and sensitivities during treatment.

Drug malabsorption as a cause of a persistent fever applies to this case. Rifampicin, a derivative of rifamycin (from Streptomyces mediterranei) is 60% protein bound, 60% excreted in bile, and 10-15% excreted in urine (hence the orange urine).⁷ Its half life is four hours but can be up to 14 hours in biliary obstruction.

In a small study from Hyderabad, India, there was a 50% fall in the plasma concentration time curve of rifampicin in undernourished patients.⁸ This was attributed to both malabsorption and increased renal clearance. This effect was in part offset by reduced plasma protein binding. As there is increased renal clearance, the urine is still orange despite low plasma concentrations.

It was initially recommended that rifampicin should be taken when fasting, but having breakfast was shown to have no significant effect on absorption.⁹ Later, however, when the dietary constituents of breakfast were analysed separately, 50 g of fat reduced rifampicin levels by 20-50%, whereas 100 g of glucose and two egg whites had no effect.¹⁰

Several cases of rifampicin malabsorption have been reported. In 1978 a 28 year old diabetic patient with pulmonary M tuberculosis resistant to isoniazid and malabsorption of rifampicin was successfully treated with intravenous rifampicin.¹¹ In France a diabetic child with coeliac disease had selective malabsorption of rifampicin and not isoniazid.¹²

More recently, there has been evidence of drug malabsorption in HIV positive patients,^{13 14 15} even in the absence of malabsorptive symptoms. In one of the reports two patients with HIV infection and tuberculosis became resistant to rifampicin, of whom one became resistant also to isoniazid after initially being sensitive.¹⁵ Serum concentrations of all antituberculous drugs except pyrazinamide were low in both patients. Therapeutic drug monitoring is therefore essential in such patients as persistently low drug concentrations select for multiple drug resistance.¹⁶

In the case that we have presented, the persistent fever seemed to be related to malabsorption of antituberculous drugs. Such malabsorption may be due to cytokine destruction of villi, undiagnosed gastrointestinal tuberculosis, chronic pancreatitis (due to high alcohol intake), bacterial overgrowth, pre-existing coeliac disease, or small bowel lymphoma. The malabsorption is mild (normal serum B12 and serum and red cell folate concentrations and a plasma albumin concentration that corrected with nasogastric feeding) and may be selective for rifampicin.

Discussion

JMBH: This is a very instructive case: pulmonary tuberculosis can kill patients even when they have been receiving treatment for a month with drugs to which the mycobacterium is sensitive. Particularly at risk are patients who are alcohol dependent, destitute, or infected with HIV. We normally think (perhaps with overconfidence) that we have sterilised patients after two weeks on the basis that the mycobacterial load has decreased by 2 log decades during this period. As Professor Mitchison's study showed, however, 50% of patients are still culture positive after one month of treatment.³ The only evidence we have for malabsorption in this case is the very low serum rifampicin concentration, although from the orange colour of his urine, we would not have expected such low plasma concentrations.

JS: Of course, the other evidence is that his fever went away when he was treated parenterally.

JMBH: Yes, that is indirect evidence that we have overcome the malabsorption. Direct evidence of malabsorption is unfortunately scanty, but we think that malabsorption is the most likely explanation.

MW: Could you have quantified the orange colour of the urine with a spectrophotometer?

JMBH: There have been colorimetric studies of urine versus rifampicin concentration, and it would have been interesting to have done this.

JC: We sometimes use radio labelled albumin or transferrin imaging for detecting protein losing enteropathies in patients like this. Had you consideredthis, or was the patient not keen on having further invasive investigations?

MTB: I am afraid that he refused even non-invasive investigations such as a pancrealauryl test to look at pancreatic exocrine function. He certainly refused endoscopy and biopsy to rule out coeliac disease, with jejunal aspiration to look for bacterial overgrowth.

JC: Drug malabsorption is not generally associated with coeliac disease, but it has been documented with rifampicin. It is a bit surprising that drug absorption is decreased by small intestinal mucosal disease, but I suppose that it is possible if a drug were protein bound in the gut. I can envisage, in a drug bound to proteins in the intestinal lumen, that impaired protein digestion might affect drug absorption. Patients who are ill and cachectic may have diminished acid secretion and pancreatic exocrine output. Furthermore, in protein losing enteropathies there is obviously further loss of protein in the gut, which may compound matters.

JS: Do you think the fact that you were filling up the lumen of his gut with albumin from his nasogastric feed may have been relevant?

JC: The nasogastric feed was overnight, and it is the lipids in food that have been shown to decrease rifampicin absorption.

PWI: We have become a little complacent about tuberculosis; only 2-3% of patients with tuberculosis are admitted to hospital. We assume that the drugs will work and know from large scale clinical trials that antituberculous drugs are highly effective. This sort of complication is really unusual. I wonder if the increase in inflammatory markers now reflects malabsorption of his prednisolone, because earlier on he had suppression of his cytokines while receiving intravenous hydrocortisone and now oral prednisolone has been restarted. This case emphasises that occasionally we have to rely on pharmacokinetic studies of antituberculous drugs. Usually we look at the colour of the urine, which we take as a measure of compliance.

CMcK: This man also has chronic schizophrenia, and a feature of his illness in the past has been a refusal to take treatment. His schizophrenia is well controlled currently, so I am sure that this was not a factor.

JSF: Some literature now shows that vitamin D deficiency may increase the severity of the condition and may affect response to treatment. I wonder whether you measured his vitamin D level?

MTB: We did not measure his vitamin D levels, but it would have been interesting to do so.

The BMJ welcomes grand rounds from other hospitals.

1. Friedland JS. Chemotactic cytokines and tuberculosis. Biochem Soc Trans 1994;22:310-2. [Medline]
2. Friedland JS. Hartley J, Hartley CGC, Shattock RJ, Griffin GE. Inhibition of ex vivo proinflammatory cytokine secretion in fatal Mycobacterium tuberculosis infection. Clin Exp Immunol 1995;100:233-8. [Medline]
3. Mitchison DA. Assessment of new sterilizing drugs for treating pulmonary tuberculosis by culture at 2 months. Am Rev Resp Dis 1993;147:1062-3. [Medline]
4. Cunha BA. Pulmonary tuberculosis and steroids. Chest 1995;107:1476-87. [Free Full Text]
5. Muthwuswamy P, Hu T-C, Carasso B, Antonio M, Dandamudi N. Prednisolone as adjunctive therapy in the management of pulmonary tuberculosis. Chest 1995;107:1621-30. [Abstract/Free Full Text]
6. Senderovitz T, Viskum K. Corticosteroids and tuberculosis. Resp Med 1994;88:561-5. [Medline]
7. Proust AJ. The Australian rifampicin trial. Med J Aust 1971;2:85-94. [Medline]
8. Polosa K, Murphy KJR, Krishnaswamy K. Rifampicin kinetics in undernutrition. Br J Clin Pharmacol 1984;17:481-4. [Medline]
9. Sieger DI, Bryant M, Burley DM, Citron KM. Effect of meals on rifampicin absorption. Lancet 1974;2:197-8. [Medline]
10. Purohit SD, Sarkar SK, Gupta ML, Jain DK, Gupta PR, Mehta YR. Dietary constituents and rifampicin absorption. Tubercle 1987;68:151-2.
11. Anker N, Da Cunha Bang F. Long-term intravenous rifampicin treatment: advantages and disadvantages. Eur J Resp Dis 1981;62:84-6. [Medline]
12. Dieterlen P, Cassereau H, Lestradet H. Permanent malabsorption of rifampicin in a diabetic with coeliac disease. Arch Fr Pediatr 1986;43:421-2. [Medline]
13. Berning SE, Huitt GA, Iseman MD, Peloquin CA. Malabsorption of antituberculosis medications by a patient with AIDS. N Engl J Med 1992;327:1817-8. [Medline]
14. Peloquin CA, MacPhee AA, Berning SE. Malabsorption of antimycobacterial medications. N Engl J Med 1993;329:1122-3. [Free Full Text]
15. Patel KB, Belmonte R, Crowe HM. Drug malabsorption and resistant tuberculosis in HIV-infected patients. N Engl J Med 1995;332:336-7. [Free Full Text]
16. Peloquin CA. Pharmacology of the antimycobacterial drugs. Med Clin North Am 1993;77:1253-62. [Medline]