Valproic acid overdose and haemodialysis

The utilization of valproic acid (VA) has expanded since its introduction in 1978 for the treatment of epilepsy. It is now also prescribed for bipolar affective disorder, schizoaffective disorder, schizophrenia and migraine prophylaxis [1]. With increased usage, VA overdose will probably become more common. Unfortunately, the best approach to treating severe VA overdose remains unclear. Most cases of VA overdose are benign. Severe overdose, however, has been linked with coma, haemodynamic instability, bone marrow depression, cerebral oedema, multiorgan system failure and occasionally death [2]. Recommended treatments for severe VA overdose include supportive therapy alone, continuous activated charcoal infusion and naloxone [2]; however, evidence of the benefit of these therapies is lacking. Recently, a number of reports have examined extracorporeal techniques for VA‐induced coma (Table 1). While some reports have been promising, clinical experience in this area is limited and the role of haemodialysis remains unclear. We report a case of VA‐induced coma treated successfully with one session of haemodialysis, and summarize the literature on extracorporeal techniques in this setting. Published case reports of VA overdose treated with haemodialysis and/or haemoperfusion Abbreviations: VA, valproic acid; t½, half‐life; Cl, clearance; F, female; M, male; yo, year old; mo, month old; HD, haemodialysis; HP, haemoperfusion. *Calculated from data in paper. Published case reports of VA overdose treated with haemodialysis and/or haemoperfusion Abbreviations: VA, valproic acid; t½, half‐life; Cl, clearance; F, female; M, male; yo, year old; mo, month old; HD, haemodialysis; HP, haemoperfusion. *Calculated from data in paper. A 56‐year‐old woman with a history of bipolar affective disorder and three previous suicide attempts was found unconscious with divalproex sodium (Epival) pills scattered beside her. On arrival in the Emergency Department, she was comatose with pinpoint pupils and a Glasgow Coma Score (GCS) of six (the patient would withdraw from pain). She was haemodynamically stable with heart rate 76 bpm, respiratory rate 16 breaths per minute, and blood pressure 90/50 mmHg. Shortly after presentation, the patient's systolic blood pressure fell to 70 mmHg, which resolved quickly with a litre bolus of normal saline. Initial blood work revealed a VA level of 998 μg/ml (6930 μmol/l), as measured by a fluorescent polarization immunoassay (Abbott Laboratories Inc, Abbot Park, IL). Additional blood work demonstrated a respiratory acidosis (pH 7.14, pCO2=91 mmHg, pO2=64 mmHg, HCO3=30 mmol/l), an anion gap of 12 mmol/l, an albumin level of 25 g/l, and a lactate level of 2.7 mmol/l. Serum electrolytes, creatinine, complete blood count and liver enzymes were all normal. Urine toxicology was negative for benzodiazepines, cocaine, amphetamines, THC, phencyclidine, opiates, barbiturates and tricyclic acid. Blood toxicology detected no ethanol, acetaminophen or salicylates. Initial treatment included oxygen, iv fluids and two doses of 2 mg iv naloxone, which were administered with no effect. Five hundred mililitres of activated charcoal was given by nasogastric tube. The patient was intubated, ventilated and given 100 mEq of iv bicarbonate. Despite these treatments her status deteriorated. Three hours after presentation, her GCS was three (completely unresponsive) and she was transferred to a tertiary centre for haemodialysis. Haemodialysis was initiated with a Centrysystem 3 dialysis machine (Cobe Renal Care, Lakewood, CO) utilizing a polysulphone haemodialysis membrane (F80A, Fresenius Medical Care, Lexington, MA). The patient was dialysed for 7.7 h via a femoral venous catheter. Dialysis was performed with a blood flow rate of 400 ml/min and a dialysate flow rate of 500 ml/min. No ultrafiltration was performed. Immediately prior to haemodialysis, and on one occasion during dialysis, the patient became hypotensive with a systolic pressure between 70 and 80 mmHg. On both occasions her systolic pressure rose to>90 mmHg with transfusion of 250 cc of pentaspan. At no time was inotropic support necessary. Figure 1 illustrates the marked decrease in serum VA that occurred with haemodialysis. VA elimination follows first order kinetics [6], thus the elimination rate constant was calculated using the equation Ke=ln(Ci/Cf)/time and the half‐life of VA was calculated using t½=ln2/Ke. The elimination constant and half‐life of VA prior to haemodialysis was 0.022 and 31.3 h, respectively. With haemodialysis these values were improved to 0.308 and 2.25 h, respectively. Significant clinical improvement was also correlated with the use of haemodialysis. Haemodialysis was initiated at 04:00 when the patient had a GCS of three. At 05:30 the patient began to move her arms and legs. At 06:30 she opened her eyes spontaneously. By 11:40, when haemodialysis was completed, the patient was awake, alert and communicating with a note pad. She was extubated at 12:00 and transferred to the general medical floor later that day. As illustrated in Figure 1, after haemodialysis was discontinued the serum VA level rebounded to a peak of 98 μg/ml (within the therapeutic range of VA). This increase probably reflected redistribution of VA out of peripheral tissues into the blood. There was no change in the patient's clinical status and a second haemodialysis session was not necessary. The VA concentration slowly decreased from this rebound peak with a post‐dialysis half‐life of 22.7 h. Valproic acid concentration vs time. Note: time between arrows indicates duration of haemodialysis (HD). Pre‐HD half‐life=31.3 h, intra‐HD half‐life=2.25 h, post‐HD half‐life=22.7 h. The asterisk denotes the calculated data point. Haemodialysis clears toxins that are <500 g/mol, are water soluble, have small volumes of distribution, follow single compartment kinetics, and are poorly bound to plasma proteins. In comparison, haemoperfusion is able to remove water and lipid soluble toxins and is less affected by protein binding [3]. VA is a small (144 Da), water‐soluble molecule with a volume of distribution of only 0.1–0.4 l/kg. However, at therapeutic doses it is 90–95% protein bound [4]. Thus, it would appear to be a better candidate for haemoperfusion than haemodialysis. Indeed it has been demonstrated that at therapeutic doses, haemodialysis has negligible impact on serum VA concentrations [5]. In the case of overdose, however, the pharmacokinetics of VA change due to saturation of protein binding [4]. As a result, a much larger percentage of VA is unbound and accessible by haemodialysis. Albumin level and age also affect the proportion of unbound VA. Lower albumin levels result in more unbound drug, while older age is correlated with less protein binding independent of albumin level [4]. Our case was especially suited to haemodialysis in that the patient was middle‐aged, had a low serum albumin, and a very high plasma VA concentration. There are three published English language case reports on haemoperfusion for VA overdose; however they show conflicting results (Table 1) [6–8]. Graudins and Aaron [7] concluded that haemoperfusion benefited their patient, while Matsumoto [8] was uncertain, and Van der Merwe [6] reported that haemoperfusion was not helpful. Case reports of haemoperfusion combined with haemodialysis are more optimistic. Four authors conclude that the intervention benefited their patients [9–12]; one concludes that it did not [13] (Table 1). One group [10] collected data on the two techniques separately and concluded that haemodialysis is an effective method of removing VA, while haemoperfusion is limited by saturation of the columns and by thrombocytopenia. Prior to the presented case, there were two published case reports of VA overdose treated with haemodialysis alone. Johnson et al. [14] concluded that the treatment was beneficial. They reported a pre‐treatment half‐life of 7.2 h and an intra‐treatment half‐life of 2.4 h. Brubacher et al. [15] also reported a reduction in the half‐life of VA with haemodialysis; however, the VA concentration in their case had decreased significantly prior to haemodialysis, thus they could not conclude that dialysis had a significant impact. The presented case is consistent with the results of Johnson et al. In our case, the plasma VA concentration fell precipitously with the use of haemodialysis and was in the therapeutic range after only 5 h (Figure 1). These laboratory changes were mirrored by marked improvements in the patient's clinical status. After termination of haemodialysis, VA concentration did increase slightly (Figure 1). This rebound suggests that at overdose concentrations VA may not observe single compartment kinetics. The volume of distribution of VA may be higher than previously realized. Johnson et al. [14] observed the same phenomenon. Haemoperfusion has been recommended previously, either alone or in combination with haemodialysis, for the treatment of VA‐induced coma. From both a technical and logistical perspective, haemoperfusion is more difficult than haemodialysis. It is also associated with a higher incidence of harmful side effects such as thrombocytopenia. Moreover, it has not been demonstrated to be superior to haemodialysis alone. Based on our review of the literature and our own experience, we recommend haemodialysis for VA overdose causing coma. Correspondence and offprint requests to: Dr Philip A. McFarlane, Division of Nephrology, St Michael's Hospital, 61 Queen Street East, 9th Floor, Toronto, Ontario, Canada M5C 2T2.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.