A middle aged man with IgG kappa multiple myeloma previously treated with bortezomib and lenalidomide presented to the hospital with altered mental status. He had completed chemotherapy months prior to presentation. Shortly after being admitted, he progressed to obtundation associated with tachypnea and a profound respiratory alkalosis requiring intubation for airway protection. His initial arterial blood gas at the time of intubation revealed a pH of 7.35, an undetectably low pCO2, and a bicarbonate of 14 (on concurrent labs) with an anion gap of 14. Mechanical ventilation was initiated with a minimal amount of pressure support. All subsequent blood gases demonstrated a pH greater than 7.55 with ongoing respiratory alkalosis. The anion gap normalized. Intensive work-up of the altered mental status resulted in the sole finding of hyperammonemia with a serum ammonia level of 100 umol/L. There was no liver injury evident on labs. Lactulose was initiated to treat the elevation in ammonia with no improvement noted.
We were consulted regarding the possibility of dialysis to correct the hyperammonemia.
Do we need to correct high ammonia levels?
Hyperammonemia carries a significant morbidity and mortality, and patients frequently require ICU-level care for encephalopathy. With acute presentations of hyperammonemia, levels of ammonia greater than 200 umol/L are associated with cerebral edema and herniation due to cerebral dysautoregulation. In the presence of chronic hyperammonemia, compensatory increases in ammonia metabolism by the muscles and hepatic and splanchnic vascular beds may blunt symptoms. Interestingly, arterial, venous, and brain levels of ammonia typically do not correlate in patients with chronic hyperammonemia but have a better correlation in acute hyperammonemia. In our patient, we did not check arterial ammonia levels to look for a correlation.
Why is our patient hyperammonemic?
Ammonia is produced primarily in the gut as a product of protein breakdown and bacterial metabolism, and it is broken down primarily in the liver. Increased production of ammonia can occur in the presence of protein breakdown from GI bleeding. In patients with liver failure who are already predisposed to having high ammonia levels, GI bleeding is a known risk factor for the precipitation of hepatic hyperammonemic encephalopathy.
Our patient had no laboratory evidence of liver dysfunction. He had recently suffered an episode of GI bleeding from a bleeding mechanical abnormality seen on EGD that was corrected. Perhaps this could have been the inciting factor for the elevated ammonia, but the patient's ammonia levels remained elevated throughout his hospital stay despite termination of the bleeding.
Inborn errors of metabolism can be considered, but most of these present in childhood. However, urea cycle disorders can be unmasked in adulthood by medications, protein intake, and infections. The drugs most likely to be involved include salicylates, valproate, carbamazepine, sulfadiazine, pyrimethamine, glycine, and TPN. None of these were implicated in our patient. Urea-splitting organisms and herpes infection can also raise ammonia levels. Neither were present.
Alas, multiple myeloma can be a cause of hyperammonemia. Rising beta-2 microglobulin levels in the absence of acute kidney injury in our patient suggested worsening myeloma off of chemotherapy.
Hyperammonemic encephalopathy in multiple myeloma
In vitro, myeloma cell lines secrete ammonia into culture medium in greater amounts than other hematological malignant cells. This may be due to excess protein synthesis in myeloma cells. In vivo, the exact mechanism for hyperammonemia in multiple myeloma patients is unknown.
It is important to rule out hypercalcemia and hyperviscosity as causes of altered mental status in patients with multiple myeloma before encephalopathy is attributed to elevated ammonia levels. In our patient, calcium and viscosity levels were checked and found to be normal.
In a published review of 27 cases of hyperammonemic encephalopathy in multiple myeloma patients, depressed levels of consciousness were noted to occur at ammonia levels ranging from 35 to 39,342 umol/L. The degree of ammonia elevation did not appear to correlate with death (if one ignores the outlier of 39,342) but in all cases where the patient's ammonia level did not improve, death was the unfortunate outcome. Presenting symptoms included respiratory alkalosis, asterixis, myoclonus, hallucinations, and hyperdynamic heart failure. Of the 27 reported cases, 4 patients received dialysis (3 HD, 1 PD) and 9 received ammonia-lowering medications such as antibiotics, lactulose, carnitine, and flumazenil. All patients received chemotherapy.
Should we dialyze our patient?
In drug intoxications associated with hyperammonemic encephalopathy, dialysis serves a definitive role in clearing the toxin and correcting the hyperammonemia. However, in cases resulting from malignancy, the answer is less clear.
Of the 4 patients who received dialysis in the study described above, 3 survived. Of the remaining 23 patients who were not dialyzed, 12 survived. Ammonia levels decreased in all 4 patients who received dialysis and in 19/23 patients who did not undergo dialysis. This suggests that dialysis plays perhaps only a minor role in lowering ammonia in this patient population. Given the potential for harm associated with dialysis catheter placement, we felt that the risks outweighed the unclear (if any) benefits of dialysis.
Our patient received steroids and further chemotherapy. Interestingly, administration of steroids can raise ammonia levels in the short term because of increased catabolism. Ammonia levels fluctuated wildly but improved slightly. The patient was extubated succesfully. His mental status improved but not to baseline. He continued to have a respiratory alkalosis at the time of discharge.
Ammonia: Chicken, Egg, or Bystander?
In hepatic encephalopathy, it is well known that ammonia levels do not correlate with the degree of encephalopathy. For this reason, following ammonia levels after initiating treatment is discouraged. The same may be true here. Perhaps ammonia is the cause of the encephalopathy; perhaps it is a biomarker of disease; perhaps it is an innocent bystander. Further study is needed to elucidate this as it may clarify the role of dialysis as a treatment for hyperammonemia in myeloma.
2. Lora-Tamayo J, Palom X, Sarrá J, Gasch O, Isern V, Fernández de Sevilla A, Pujol R. Multiple myeloma and hyperammonemic encephalopathy: review of 27 cases. Clin Lymphoma Myeloma. 2008 Dec;8(6):363-9.
3. Furer V, Heyd J. Hyperammonemic Encephalopathy in Multiple Myeloma. Isr Med Assoc J. 2007 Jul;9(7):557-9.