Case report

Abstract

Several cases of lactic acidosis, some fatal, have been reported in women receiving long-term antiretroviral therapy [1]. Nucleoside analogue antiretroviral drugs are implicated in lactic acidosis and steatosis, with or without pancreatitis, via mitochondrial toxicity [2]. Women are at greater risk than men, but it is unclear whether pregnancy increases susceptibility, possibly in relation with disorders of mitochondrial fatty acid oxidation or riboflavin deficiency [3]. Clinical symptoms, such as fatigue, dyspnoea, abdominal pain or vomiting, may be difficult to recognize in pregnant women, or may mislead the obstetrician towards a diagnosis of pre-eclampsia or preterm labour, and may appear at an advanced stage. Because consequences for the mother and fetus can be devastating, early diagnosis is crucial. We report two cases of severe lactic acidosis in the third trimester of pregnancy in women receiving stavudine and didanosine, both of which were preceded by a moderate elevation of liver enzymes. Case 1 A 36-year-old HIV-1-infected multiparous woman presented with epigastric pain and intense uterine contractions at 35 weeks of a previously uneventful pregnancy, and was delivered by emergency caesarean section, giving birth to a 3180 g male infant, whose postnatal course was normal. She had received stavudine, didanosine and efavirenz for 18 months, with efficacy (CD4 lymphocytes 363 × 106/l and plasma HIV RNA < 200 copies/ml). Efavirenz had been continued despite potential teratogenicity because the pregnancy was diagnosed late. Six days before presentation, serum aspartate aminotransferase (AST), performed as part of a systematic surveillance, was three times the upper limit of norm (ULN), and the serum bicarbonate level was below the norm. At entry, AST was five times the ULN and serum lipase was 14 times the ULN. After delivery, vomiting began and lactic acidosis appeared (pH 7.26, serum lactate 8.2 mm/l, 4 × ULN, bicarbonates 12 mm/l). There was no bile duct anomaly, and computerized tomography showed severe necrotic pancreatitis and hepatic steatosis. Antiretroviral drugs were discontinued, the patient was placed under intensive care with non-specific support and received thiamine, riboflavine and l-carnitine. She recovered normal lactate concentrations by day 12, had necrotic cellulitis of the abdominal wall requiring two further operations, and was discharged 6 weeks after delivery. Case 2 A 43-year-old HIV-1-infected multiparous woman was hospitalized at 30 weeks for a mild elevation of liver enzymes (1.5 × ULN) and hyperlactataemia (4.3 mm/l). She had been treated with stavudine, didanosine, ritonavir and saquinavir for 16 months, with good efficacy and tolerance. The pregnancy had been marked by an ultrasound diagnosis of a fetal atrioventicular canal, but the couple declined a genetic amniocentesis. Serum liver enzymes, tested monthly as part of the systematic follow-up, were already slightly above the ULN a month before entry. The patient had gained no weight over 3 months, but was otherwize asymptomatic. Upon admission, antiretroviral therapy was discontinued. Nonetheless, metabolic acidosis (pH 7.22) appeared within 24 h, serum lactate increased to 12.6 mm/l, concomitantly with abdominal pain, vomiting and dyspnoea. A caesarean section was performed. The child, a 1000 g female with trisomy 18, died at 2 weeks of age in neonatal intensive care. The mother was treated in intensive care with haemodialysis and infusions of amino acids, thiamine and riboflavine and l-carnitine, and lactic acidosis resolved within 2 days without further complications. Although the natural history is unpredictable, these cases suggest that the early detection and management of lactic acidosis is likely to be of benefit. After our first experience, we managed the second case more actively. In both, moderately elevated liver enzymes preceded the onset of acidosis and clinical symptoms. This is consistent with the previously published cases. We therefore suggest that biological surveillance should be performed monthly in women receiving nucleoside analogue combinations, including liver enzymes, lipase and bicarbonate (anion gap). The usefulness of routine lactate measurements remains controversial, because mild hyperlactataemia does not always lead to lactic acidosis [2]. Both of our cases, and all but one previously reported case, followed long-term treatment including stavudine and didanosine. Since January 2001, the product information [4] warns that this combination should be avoided during pregnancy unless ‘the potential benefit clearly outweighs the potential risk'. Another case of lactic acidosis occurred in a woman receiving stavudine and lamivudine [5], but none has been reported with zidovudine monotherapy among several thousand pregnancies. However, mitochondrial dysfunction has been reported in infants exposed to zidovudine or zidovudine–lamivudine [6]. Antiretroviral therapies are used with major benefits in pregnant women, for their health and to prevent mother–infant HIV transmission. Nonetheless, the mitochondrial toxicities of nucleoside analogues, for the mother and child, should be considered in choosing the most appropriate regimen.