Metabolic acidosis

A 45-year-old man with type 2 diabetes, hypertension, and chronic renal failure was admitted to the hospital with a 3-day history of upper abdominal pain, anorexia, nausea, low-grade fever, and altered sensorium. He had had diabetes for 8 years, and hypertension and renal failure for 9 months but had not undergone dialysis. The patient was receiving therapy with glipizide, 20 mg qd, and phenformin, 100 mg qd, for his diabetes, and amlodipine, 10 mg qd, clonidine, 0.2 mg qd, metoprolol, 100 mg qd, minoxidil, 2.5 mg qd, and furosemide, 40 mg qd, for his hypertension. He stopped receiving his oral hypoglycemic agents 3 days before hospital admission because of reduced food intake but was still taking his antihypertensive medications. He procured all his medications from India.

Physical Examination

On hospital admission, the patient's temperature was 94.0[degrees]F, BP was 140/90 mm Hg, pulse was 85 beats/min, respiratory rate was 24 breaths/min, and arterial oxygen saturation was 92.0% while receiving 100% nonrebreathing mask. His oral mucosa was dry, his neck veins were flat, and his chest was clear to auscultation. Cardiovascular findings were within normal limits without rubs, murmurs, or gallops. The findings of the abdominal examination were benign. The patient's lower extremities had 3+ edema. He was drowsy but arousable and could communicate. There were no focal neurologic deficits.

Laboratory Findings

The results of the patient's laboratory tests were as follows: WBC count, 28,000 cells/[micro]L, with 77% neutrophils and 8% bands; hematocrit, 29.1%; hemoglobin level, 8.5 g/dL; sodium level, 145 mEq/L; potassium level, 5.4 mEq/L; bicarbonate level, 7.0 mEq/L; chloride level, 101 mEq/L; anion gap, 37; BUN, 48 mg/dL; creatinine level, 8.1 mg/dL; glucose level, 100 mg/dL; calcium level, 7.5 mg/dL; phosphorus, 9.2 mg/dL; total protein level, 6.4 g/dE; albumin level, 3.4 g/dL; lipase level, 189 IU/L; serum osmolality, 329 mOsm/kg; acetone, positive 1:8. The results of his urine chemistry tests were as follows: creatinine level, 160 mg/dL; sodium level, 16 mEq/L; chloride level, 18 mEq/L; potassium level, 16.2 mEq/L; and urea level, 299 mg/dL. Arterial blood gas measurements were as follows: pH, 6.92; PC[O.sub.2], 18.0 mm Hg; and P[O.sub.2], 85.7 mm Hg. A chest radiograph revealed basilar and perihilar infiltrates. ECG had tall, peaked T waves in the anterior leads. A CT scan of the brain was negative for acute events. Blood was drawn for cultures, and the results were pending.

What is the etiology of this patient's profound metabolic acidosis? What are the pathophysiologic mechanisms for the generation of the acidosis?

Diagnosis: Phenformin-induced lactic acidosis

This patient was receiving glipizide, 20 mg qd, and phenformin, 100 mg qd, for management of his diabetes. He had procured the phenformin in India. His initial lactate level was 18 mmol/L.

Phenformin is a biguanide drug that inhibits gluconeogenesis and increases the peripheral uptake of glucose. The association between phenformin and fatal lactic acidosis led to its withdrawal from the US market in 1977. This drug is still available in other countries and is often available without a prescription. As a result, sporadic reports of phenformin-induced lactic acidosis come from within the United States in patients who have obtained this medication from other countries. One in 4,000 patients taking phenformin develops lactic acidosis. The mortality rate associated with phenformin-induced lactic acidosis approaches 50%.

The risk factors for lactic acidosis with the use of biguanide agents include age (> 60 years), decreased cardiac, hepatic, and/or renal function, diabetic ketoacidosis, surgery, respiratory failure, ethanol intoxication, and fasting. However, not all patients have comorbid conditions. Prior episodes of lactic acidosis and hypoxia also increase the risk for recurrent lactic acidosis. Typical presenting symptoms include nausea, vomiting (approximately 50%), diarrhea, anorexia, epigastric pain (approximately 35%), Kussmaul respiration, and altered sensorium (approximately 80%). Other clinical findings include dehydration, hypothermia, and hypotension. Other causes of lactic acidosis, such as generalized hypoxia and sepsis, should be ruled out.

The metabolic consequences of phenformin toxicity in diabetic patients reflect both drug effects and inadequate insulin levels. Phenformin binds to mitochondrial membranes and inhibits oxidative phosphorylation and the formation of high-energy adenosine triphosphate (ATP). This slows or limits gluconeogenesis in diabetic patients by reducing the amount of ATP that is necessary to reverse the glycolytic pathway and by slowing or limiting pyruvate carboxylase, a key enzyme in this process that requires the use of ATP as a substrate. Therefore, pyruvate and its precursors cannot be used for glucose synthesis. In addition, low levels of ATP activate glycolysis through an increase in phosphofructokinase activity and an increase in the synthesis of pyruvate from phosphoenolpyruvate by pyruvate kinase. Since reutilization is decreased and formation is increased, intracellular concentrations of pyruvate increase.

The symptoms associated with phenformin toxicity strongly influence the evolution of the metabolic disorder. GI symptoms (ie, anorexia, nausea, and vomiting) limit oral intake, and serum insulin levels fall. Low levels of insulin allow protein and fat catabolism in diabetic patients. Protein catabolism mobilizes alanine and other glucogenic amino acids, which are converted into pyruvate and tricarboxylic acid cycle intermediates. The pyruvate produced during glycolysis and other metabolic processes is usually converted into acetyl coenzyme A (CoA), which then enters the citric acid cycle. However, the inhibition of oxidative phosphorylation by phenformin limits the regeneration of nicotinamide adenine dinucleotide (NAD+) from NADH. Therefore, intracellular concentrations of NADH increase, and high ratios of NADH/NAD+ stimulate covalent phosphorylation and the concomitant inhibition of the pyruvate dehydrogenase complex, which blocks pyruvate entry into the citric acid cycle. Low levels of insulin in fasting diabetic patients also allow fat catabolism. Free fatty acids are oxidized in mitochondria and form acetyl CoA, NADH, and flavin adenine dinucleotide (reduced form). The high levels of NADH inhibit the metabolism of acetyl CoA by the citric acid cycle. Therefore, intracellular concentrations of acetyl CoA increase, and high acetyl CoA/CoA ratios also stimulate covalent phosphorylation and the concomitant inhibition of the pyruvate dehydrogenase complex. This further limits pyruvate entry into the citric acid cycle. Since pyruvate cannot enter either the gluconeogenesis pathway or the citric acid cycle, it is metabolized to lactic acid by lactic dehydrogenase. This increases the concentrations of lactate and hydrogen ions, resulting in the metabolic acidosis. This reaction is strongly driven by the high levels of NADH. Finally, the high concentrations of acetyl CoA produced by fatty acid metabolism form acetoacetic acid and [beta]-hydroxybutyric acid. These ketone bodies contribute to the overall metabolic disorder. Consequently, diabetic patients with phenformin toxicity usually have a severe anion gap metabolic acidosis with increased concentrations of lactic acid, acetoacetic acid, [beta]-hydroxybutyric acid, and often other acids retained as a consequence of deteriorating renal function.

Medical therapy for phenformin-induced lactic acidosis includes supportive care and efforts to limit or reverse toxicity. Supportive care includes mechanical ventilation and hemodialysis. Mechanical ventilation prevents respiratory muscle fatigue and helps to maintain the respiratory compensation for metabolic acidosis. It also optimizes oxygen delivery to the tissues and thereby increases the disposal of NADH by maximizing oxidative phosphorylation. Hemodialysis corrects the acidosis through the removal of lactate, acetoacetate, [beta]-hydroxybutyrate, and associated H+ ions. Bicarbonate can potentially correct the pH. However, its use in the treatment of metabolic acidosis remains controversial. Bicarbonate administration frequently causes hypernatremia, hypokalemia, and volume overload. It may cause paradoxical intracellular and cerebrospinal fluid acidosis. Increases in pH shift the hemoglobin-[O.sub.2] disassociation curve to the left and reduce [O.sub.2] delivery to vulnerable tissues. Finally, bicarbonate increases membrane permeability to biguanides and could increase intracellular lactate formation. Consequently, bicarbonate administration is not recommended.

Several interventions for limiting or reversing the toxicity of phenformin have been evaluated. Phenformin is lipid-soluble and has a large volume of distribution (5 to 10 L/kg). Conventional hemodialysis does not facilitate removal of phenformin. Prolonged filtration using continuous venovenous hemofiltration may clear a sufficient amount of phenformin to partially reverse its toxicity. Sodium dichloracetate increases the activity of pyruvate dehydrogenase, which could facilitate the metabolism of pyruvate by the citric acid cycle. This would, in turn, decrease lactate production. However, available studies indicate that dichloracetate therapy in lactic acidosis does not significantly alter hemodynamics or improve survival rates. Exogenous insulin and glucose should be beneficial. Insulin reduces protein and fat catabolism and therefore should reduce the production of pyruvate and ketoacids. It also reverses the phosphorylation of pyruvate dehydrogenase, and this increases this enzyme's activity and the metabolism of pyruvate by the citric acid cycle. Dembo et al reported a detailed study of a patient with phenformin-induced lactic acidosis who was treated with insulin and reviewed 76 published cases. This review demonstrated that the survival rate was higher in patients who had been treated with insulin (65%) than in patients who had not received insulin (48%). Insulin infusions at a rate of 2 to 5 U/h with sufficient glucose to maintain euglycemia should be used. [K.sup.+] levels will need close monitoring.

The frequency of lactic acidosis associated with metformin therapy is much lower than the frequency associated with phenformin therapy and may not exceed the background rate in diabetic patients who have comorbid conditions. There are several potential explanations for this. Metformin does not bind to mitochondrial membranes as efficiently as phenformin. Metformin has a shorter half-life (1.5 to 4.9 h) and has no hepatic metabolism, whereas phenformin has a longer half-life (4 to 13 h) and undergoes hepatic oxidation. Hence, patients with hepatic dysfunction may be at increased risk for toxicity when receiving phenformin therapy. Toxicity from metformin can be avoided if contraindications are strictly observed. The most important contraindications are renal insufficiency (ie, creatinine level of [greater than or equal to] 1.5 in men and [greater than or equal to] 1.4 in women) and medical conditions with increased risk for renal hypoperfusion (eg, congestive heart failure) or lactic acidosis (eg, sepsis). Patients with metformin toxicity should benefit from conventional dialysis.

Our patient had most of the clinical features of phenformin-induced toxicity. Chronic renal failure secondary to diabetes was his most important risk factor for developing phenformin-induced lactic acidosis. He was admitted to the medical ICU, underwent hemodialysis, and received ventilatory support. He was also started on a regimen of empiric antibiotic coverage (ie, piperacillin/tazobactam) for possible sepsis, as suggested by the presence of leukocytosis and pulmonary infiltrates. His blood cultures were reported to be negative on the third hospital day. His pulmonary infiltrates subsequently cleared after the second hemodialysis run and were thought to be the result of volume overload. His other laboratory parameters improved, as shown in Table 1. He was extubated successfully after 48 h in the ICU. Presently, he is receiving hemodialysis while waiting for renal transplantation.

CLINICAL PEARLS

1. Biguanide-induced lactic acidosis should be considered in the differential diagnosis of high-anion-gap metabolic acidosis in diabetic patients.

2. The index of suspicion for lactic acidosis should be increased in a clinical setting where the patient is of foreign origin, especially if from those countries where phenformin is available.

3. The incidence of biguanide toxicity can be reduced if renal function is closely monitored while the patient is receiving these drugs.

4. Insulin administration can correct many of the metabolic derangements induced by phenformin toxicity.

SUGGESTED READINGS

Dembo AJ, Marliss EB, Halperin MD. Insulin therapy in phenformin-associated lactic acidosis: a case report, biochemical considerations and review of literature. Diabetes 1975; 24: 28-35

Gan SC, Barr J, Arieff A, et al. Biguanide-associated lactic acidosis. Arch Intern Med 1992; 152:2333-2336

Kwong SC, Brubacher J. Phenformin and lactic acidosis: a case report and review. J Emerg Med 1998; 16:881-886

Lalau J-D, Race J-M. Lactic acidosis in metformin therapy. Drugs 1999; 58:55-60

Luft D, Schmulling RM, Eggstein M. Lactic acidosis in biguanide-treated diabetics: a review of 330 cases. Diabetologia 1978; 14:75-87

Mariano F, Benzi L, Cecchetti P, et al. Efficacy of continuous venovenous haemofiltration (CVVH) in the treatment of severe phenformin-induced lactic acidosis. Nephrol Dial Transplant 1998; 13:1012-1015

Owen MR, Doran E, Halestrap AP. Evidence that metformin exerts its anti-diabetic effects through inhibition of complex 1 of the mitochondrial respiratory chain. Biochem J 2000; 348: 607-614

Stacpoole PW, Wright EC, Baumgartner TG, et al. A controlled clinical trial of dichloroacetate for treatment of lactic acidosis in adults: the Dichloroacetate-Lactic Acidosis Study Group. N Engl J Med 1992; 327:1564-1569

Voet D, Voet JF, Pratt CW. Fundamentals of biochemistry. New York, NY: Wiley and Sons, 1999; 382-528

Department of Internal Medicine, Texas Tech University Health Sciences Center, Lubbock, TX.

Manuscript received June 19, 2002; revision accepted August 25, 2002.

Correspondence to: Kenneth M Nugent, MD, FCCP, Department of Internal Medicine, Texas Tech University Health Science Center, Lubbock, TX 79430; e-mail: Kenneth. Nugent@ttuhsc.edu

COPYRIGHT 2003 American College of Chest Physicians
COPYRIGHT 2003 Gale Group