A 14-month-old boy on zidovudine/lamivudine/nelfinavir and trimethoprim-sulfamethoxazole developed hepatitis and cholestasis with AST in the 5000 range, ALT in the 600 range, AlkP 800, and Bili 21. There was mild elevation of CPK, but no acidosis, and no other systemic findings. Work-up for HAV, HBV, HCV, EBV, and CMV was negative, and liver biopsy was compatible with drug-induced hepatitis. Currently he is off medications, recovering very slowly. His viral load is greater than 750,000 copies/mL and CD4+ cell count is 340 cells/mm, which corresponds to severe immunosuppression for his age. Any suggestions as to which medication is the most likely cause of the hepatitis, and what to use, both for PCP prophylaxis and for antiretroviral regimen?
Robert Posada, MD
Elevation of transaminases has been reported as a possible adverse event of all antiretroviral agents. Some agents directly cause liver toxicity and hepatitis while others exacerbate chronic liver disease. Zidovudine, the first nucleoside reverse transcriptase inhibitor (NRTI) to be developed, was also the first NRTI reported to cause hepatic toxicities including elevations in transaminases, lactic acidosis, and mitochondrial dysfunction. Although rare, lamivudine has also been associated with elevation of liver enzymes, as has nelfinavir. All 3 drugs can exacerbate an underlying liver problem. Similarly, trimethoprim-sulfamethoxazole (TMP-SMZ) has been reported to cause elevated transaminases. An additional fact to keep in mind is that there can be significant drug-drug interactions; for example, TMP-SMZ increases lamivudine levels and zidovudine-associated toxicity. Nevertheless, all of these agents have in general been well tolerated both individually and in combination.
Recently there have been reports of lactic acidosis and severe hepatomegaly with steatosis, including fatal cases, with the combined use of NRTIs. When evidence of severe liver toxicity occurs, all medications should be discontinued, as you did in this child. The task of determining which drug is the offending agent is often difficult. In addition to drug toxicity, one must search for other etiologies of the hepatitis and cholestasis, such as HIV-induced hepatitis or hepatitis caused by coinfections, eg, Mycobacterium avium complex(MAC), hepatitis B or C virus, or the herpesviruses (especially Epstein-Barr virus or cytomegalovirus [CMV]). Potential noninfectious causes such as autoimmune disorders must also be evaluated. I assume that your patient does not have an underlying hematologic disorder (eg, SS disease or hereditary spherocytosis) or inherited disorder (eg, galactosemia) that could have been exacerbated.
You present a very unusual situation that is fortunately very rare. Clearly, the levels of the bilirubin and liver enzymes are extremely worrisome and any intervention must be done with extreme caution and with very close follow-up of the patient. Your patient has apparently had a thorough work-up, hopefully using the most up-to-date technologies, to rule out other etiologies including any autoimmune disorders, and I assume there was no associated pancreatitis or liver steatosis, and no evidence for mitochondrial dysfunction given the lack of acidosis. It is certainly unusual that there were no symptoms with the laboratory values you report. However, since you do not report earlier liver function test results or clinical symptomatology, I assume that your patient had normal hepatic enzymes prior to this episode. With regard to the extremely high bilirubin, the patient must have been jaundiced and it would be interesting to know how rapidly this developed; whether the bilirubin was direct or indirect; and whether there was any evidence of hemolysis or anemia (or a positive Coombs' test) associated with this episode. This information would help in determining the extent of the hepatocellular damage and whether there was an underlying autoimmune disorder, hemolytic process, or obstructive phenomenon that may have been exacerbated by the drugs.
Because your patient has very low CD4+ cell counts and extremely high viral loads, he is at risk for MAC or CMV infection, as well as other AIDS-related opportunistic infections which may contribute to the findings you report. He is also at increased risk for progression of HIV disease or the development of opportunistic infections. Thus, after 2 weeks to 1 month of therapy discontinuation, you should re-evaluate the situation, working with a hepatologist or gastroenterologist who can help in assessing when the liver has recovered from the current episode.
Once the patient has recovered, therapy should be resumed with a different antiretroviral combination, taking into consideration previous medications, HIV viral load, immune function, and medical history. It is difficult to suggest any specific therapy without knowing the past history, since the choice of regimen should take into account:
previous medications, including whether all the drugs in last regimen were started simultaneously or sequentially
duration of therapy with previous regimens
tolerance and response to previous regimens
plasma HIV-1 RNA levels
immune function
other medical history
If the patient had a significant elevation of HIV-1 RNA and a low CD4+ cell count even before the last regimen was discontinued, it would be prudent to consider resistance testing before choosing the new regimen, although you must keep in mind that wild-type virus may have outgrown and replaced resistant virus during the period off therapy. If the patient had undetectable plasma HIV-1 RNA before discontinuation, you could feel more comfortable selecting a combination using the same classes of drugs. Since all antiretroviral agents have at one time or another been reported to cause elevation of transaminases or hepatitis, you may need to try several regimens and monitor your patient extremely closely with frequent evaluations of hepatic enzymes.
In terms of prophylaxis against Pneumocystis carinii pneumonia, the alternatives to TMP-SMX are dapsone, aerosolized pentamidine for children aged 5 years and over, or atovaquone.