Injectable Ammonium Chloride Used Enterally for the Treatment of Persistent Metabolic Alkalosis in Three Pediatric Patients

Abstract

Enteral administration of injectable ammonium chloride may offer an effective method for the treatment of persistent metabolic alkalosis, without the adverse effects associated with the intravenous route. This case series describes 3 pediatric patients who received ammonium chloride enterally for the treatment of persistent metabolic alkalosis. The patients were a 2-month-old female infant, a 6-week-old male infant, and a 3-year-old male toddler. Four to 18 doses of ammonium chloride were administered enterally (range, 3-144 mEq/dose). Two of the 3 patients achieved resolution of metabolic alkalosis with ammonium chloride, while 1 patient's condition was refractory to treatment. Resolution of metabolic alkalosis occurred at 4 and 8 days, which required a total weight-based dose of 10.7 mEq/kg and 18 mEq/kg, respectively. No adverse effects were recorded. The use of ammonium chloride injection administered enterally was a safe and effective option in 2 of the 3 pediatric patients with persistent metabolic alkalosis.

INTRODUCTION

Metabolic alkalosis is routinely diagnosed by an elevated serum pH above 7.45 and an increased serum bicarbonate (HCO₃⁻) level above 26 mEq/L. This occurs as a result of proton loss from the gastrointestinal tract or kidneys or from a gain in bicarbonate.¹ Metabolic alkalosis persists if renal function is compromised. Common causes of metabolic alkalosis include vomiting, diarrhea, edema, diuretics, potassium depletion, and magnesium deficiency.^2,3 In addition to evaluating metabolic alkalosis based on pH and HCO₃⁻, base excess (BE), a measure of the in vitro amount of acid required to bring 1 L of blood to pH 7.4, which normally ranges from −2 to 2 mmol/L, is used.^4,5

Intravenous (IV) ammonium chloride is a treatment option for severe cases of metabolic alkalosis.¹ Ammonium chloride is a systemic and urinary acidifying agent that is converted to ammonia and hydrochloric acid through oxidation by the liver. Ammonium chloride is poorly tolerated due to adverse effects including encephalopathy, metabolic acidosis, and ammonia toxicity, which consists of diaphoresis, altered breathing, bradycardia, arrhythmias, and coma.⁶ Additionally, rapid IV infusion of ammonium chloride has been associated with central nervous system toxicity that includes mental confusion, irritability, seizures, and coma.

Enteral administration, including per os (PO), nasogastric, or orogastric administration of ammonium chloride injection, would prevent the associated infusion-related reactions and offers a potential alternative route with increased tolerability. Due to the lack of evidence to support this option, the purpose of this case series was to describe the effects of administering ammonium chloride injectable via the enteral route in 3 pediatric patients with metabolic alkalosis.

PATIENT A

A 2-month-old female, born at 32 weeks gestational age, weighing 3.1 kg, with a medical history of fetal cocaine exposure and poor prenatal care, was admitted to the pediatric intensive care unit (PICU). The patient was intubated, diagnosed with respiratory syncytial virus bronchiolitis complicated by ventilator-dependent respiratory failure (VDRF), and remained in the hospital for 42 days. Concomitant medications administered during the management of metabolic alkalosis included albuterol, ipratropium, aminophylline, cefotaxime, chloral hydrate, diazepam, methadone, and ketamine. Ventilator settings throughout the PICU course were respiratory rate (RR), 30 to 38 breaths/minute; tidal volume (TV), 24 to 26 mL; fraction of inspired oxygen (FiO₂), 40%; and positive end-expiratory pressure (PEEP), 5 cm H₂O.

Between hospital days 6 and 24, the patient received 7 doses of furosemide that ranged from 0.75 to 1.5 mg/dose (up to 0.5 mg/kg/dose) for diuresis. On day 15, IV fluids were changed from 0.9% NaCl at 2 mL/hr to 5% dextrose and 0.22% NaCl with potassium chloride, 20 mEq/L, at 12 mL/hr maintenance for 24 hours. To correct the patient's metabolic alkalosis, we initially administered 8 doses, from 3 to 15 mg/dose (1 to 4.8 mg/kg/dose), of acetazolamide. Laboratory data on day 18, 1 day prior to ammonium chloride administration, showed a normal arterial pH of 7.35. Abnormal values were partial pressure of carbon dioxide (PaCO₂), 57 mm Hg; HCO₃⁻, 27 mEq/L; and BE, 5 mEq/L (Figure 1). On hospital day 18, fluid input and output were 136.3 mL/kg/day and 4.97 mL/kg/hour, respectively. For persistent BE elevation, the patient received ammonium chloride at 3 mEq/dose (1 mEq/kg/dose) via nasogastric tube every 12 hours. The metabolic alkalosis correction equation for BE was used: 0.3 L/kg multiplied by weight (kg) and BE.⁶ After 9 doses, the frequency was increased to every 6 hours for an additional 9 doses and a total of 18 doses administered by hospital day 25. On day 23, IV fluids were returned to increased maintenance levels with 5% dextrose, 0.45 NaCl, and potassium chloride at 20 mEq/L. The rate was then decreased to 8 mL/hr on day 25. The average fluid input and output while ammonium chloride was administered were 130 mL/kg/day and 4.9 mL/kg/hour, respectively. Serum HCO₃⁻ was normalized by the third dose, but the PaCO₂ did not normalize until hospital day 25. Throughout the course of treatment, arterial pH was between 7.32 and 7.37 and remained within the defined range of normal.

Figure 1.

Patient A Change in Base Excess Associated with Each Dose of Ammonium Chloride

Predose (closed circles and solid line); post-dose (closed diamonds and dashed line)

PATIENT B

A 6-week-old male, born full term, weighing 5.6 kg, with no significant past medical history, initially presented to the emergency department, was diagnosed with respiratory syncytial virus infection and sent home. The patient returned to the hospital 20 days later after a cyanotic apneic episode and was subsequently intubated and admitted to the PICU with a subsequent 38-day hospital course complicated by VDRF. Concomitant hospital medications during ammonium chloride treatment were albuterol, methylprednisolone, chloral hydrate, acetaminophen, and ranitidine. During the patient's PICU stay, the course of ventilator settings was RR, 35 to 60; TV, 32 to 35 mL; FiO₂, 40%; and PEEP, 5 cm H₂O. On hospital day 3, IV fluids were at 85% maintenance at 20 mL/hour: 5% dextrose, 0.45% NaCl, and potassium chloride at 20 mEq/L. On day 4, IV fluids were discontinued when the enteral nutrition goal had been met. From days 7 to 13, the patient was edematous and was managed with 19 doses of oral furosemide, up to 6 mg/dose (1.07 mg/kg/dose). Metabolic alkalosis was initially managed with 5 intermittent doses of acetazolamide, 28 mg/dose (5 mg/kg/dose) IV, from days 5 to 14 of hospitalization. The patient also received metolazone, 1.1 mg/dose (0.2 mg/kg/dose) IV and/or by mouth twice daily from days 9 to 20. On the eighth day of hospitalization, 1 day prior to ammonium chloride treatment, fluid input was 122 mL/kg/day, fluid output was 6.6 mL/kg/hour, and laboratory results included pH 7.39; PaCO₂, 60 mm Hg; HCO₃⁻ , 30 mEq/L; and BE, 8 mEq/L (Figure 2). The patient received a total of 6 doses of ammonium chloride, 10 mEq/dose (1.8 mEq/kg/dose) administered via nasogastric tube intermittently from days 9 to 12. Dosage for the correction of alkalosis was calculated based on BE.⁶ During the administration of ammonium chloride, the average fluid input was 105 mL/kg/day, the fluid output was 5 mL/kg/hour, and the pH ranged from 7.35 to 7.41. By the 12th day of hospitalization, after the last dose was administered, the BE was 4 mEq/L. The HCO₃⁻ remained high throughout administration of ammonium chloride and was 27 mEq/L by the end of treatment.

Figure 2.

Patient B Change in Base Excess Associated with Each Dose of Ammonium Chloride

Pre-dose (solid triangles and dashed lines); Post-dose (Solid diamonds and solid line)

PATIENT C

A 3-year-old male weighing 17 kg with a pst medical history of asthma was admitted status-post a motor vehicle accident. The patient was sedated, intubated, and admitted to the PICU. The patient's 20-day hospital course was complicated due to spleen and liver laceration, hemoperitoneum, right adrenal hemorrhage, and coagulase-negative staphylococcus bacteremia. Concurrent medications at the time of the metabolic alkalosis were albuterol, ampicillin/sulbactam, midazolam, fentanyl, chloral hydrate, docusate, metoclopramide, and ranitidine. During the PICU stay, ventilator settings were RR, 14 to 18; TV, 120 mL; FiO₂ , 30% to 40%; and PEEP, 5 cm H₂O. The patient received a bolus of 1 L of 0.9% NaCl on hospital day 3 and continued on half-maintenance IV fluids (5% dextrose, 0.45% NaCl, and potassium chloride at 20 mEq/L infused at 30 mL/hr), which was subsequently discontinued on day 7. The patient concurrently received 5 intermittent doses of furosemide, 8.5 to 15 mg/dose (0.5 to 0.9 mg/kg/dose) IV, for fluid overload on days 6 and 7. On the seventh day of hospitalization, laboratory values were within range, except for PaO₂ , 70 mm Hg; PaCO₂ , 57 mm Hg; HCO₃⁻ , 39 mEq/L; BE, 18 mEq/L; and pH 7.47 (Figure 3); and fluid input and output were 96 mL/kg/day and 2.8 mL/kg/hour, respectively. The patient received ammonium chloride at 36 mEq/dose by mouth (2.1 mEq/kg/dose) on day 8. The dose was increased to 54 mEq/dose (3.2 mEq/kg/dose) on day 9 and to 144 mEq/dose (8.5 mEq/kg/dose) on day 10 due to a lack of effect. The initial dose to correct the metabolic alkalosis was calculated based on BE. IV fluids were restarted on day 9 with 5% dextrose, 0.45% NaCl, and 20 mEq/L potassium chloride at half maintenance (35 mL/hr) and were continued until hospital day 12. During ammonium chloride treatment from days 8 to 10, the average fluid input was 76 mL/kg/day, and the average output was 5.6 mL/kg/hr. HCO₃⁻, BE, and pH values normalized to 24 mEq/L, −1, and 7.4, respectively, on day 10. On the 11th day of hospitalization, the patient was successfully extubated.

Figure 3.

Patient C Change in Base Excess Associated with Each Dose of Ammonium Chloride

Pre-dose (solid triangles and solid line); Post-dose (Solid square and dashed)

DISCUSSION

Treatment algorithms for metabolic alkalosis are based on the patient's response to administration of IV sodium chloride solution.⁷ In the adult population, first-line treatment for patients with contraction metabolic alkalosis, characterized by a decreased intravascular volume, is sodium chloride infusions to replace volume and chloride stores. For patients who are volume overloaded or sodium intolerant, acetazolamide or potassium chloride is recommended to acidify the serum or provide chloride, respectively. Acetazolamide is an acidifying agent that inhibits carbonic anhydrase. This action in the kidneys causes a reduction in hydrogen proton secretion and increases excretion of sodium, HCO₃⁻, potassium, and water.⁶ In severe cases (serum pH >7.6) or persistent metabolic alkalosis, diluted IV hydrochloric acid has been used.^1,2 However, the use of hydrochloric acid requires frequent monitoring of the arterial blood gases (ABG) and serum electrolytes.³ Adverse effects, including severe transient respiratory acidosis and infusion site reactions resulting in tissue extravasation and skin necrosis, can occur.

IV ammonium chloride is an additional treatment option for refractory metabolic alkalosis. Currently, there is a lack of sufficient data demonstrating the effect of enteral administration of ammonium chloride, especially within the pediatric and neonate populations.^8,9 A previous study evaluated the effect of ammonium chloride-induced acute metabolic acidosis through an oral load of 2.8 mEq/kg/dose on the renin-angiotensin-aldosterone system and renal function in 7 neonates.¹⁰ Ammonium chloride significantly decreased blood pH, BE, and total CO₂ values, while significantly elevating plasma concentrations of potassium, sodium, and chloride. Additionally, urine flow rate and net acid excretion were increased. The limited evidence-based support forces clinicians to use their best judgment when deciding whether to use enteral ammonium chloride for the treatment of metabolic alkalosis.

All 3 patients initially received sodium chloride-based IV fluids appropriately, based on the treatment algorithm. In addition to evaluating the recorded input and output during treatment with ammonium chloride, we assessed fluid status by reviewing the IV fluid administered to each patient and comparing this to maintenance requirements.¹¹ Input and output alone do not adequately reflect a patient's fluid status because insensible fluid losses (i.e., respiration, perspiration) are not quantified. Patients A and B continued to follow the treatment algorithm and received acetazolamide prior to ammonium chloride administration.

The indication for ammonium chloride was the elevated BE, which was consistently outside the defined range of normal. The general use of BE for ABG analysis has been criticized for clinical utility as it is a calculated value that varies among laboratories.⁴ In addition to abnormal BE, the clinical symptoms, pH, and bicarbonate levels must be evaluated before treatment with ammonium chloride. After consideration of laboratory values and prior treatment, if the patient appears to be an appropriate candidate for ammonium chloride infusion, the enteral route may be preferred over that of IV. Patients A and C achieved resolution of metabolic alkalosis after enteral ammonium chloride therapy, while patient B's BE remained refractory. The increase in patient A's IV fluids from 0.9% NaCl at 2 mL/hr to maintenance fluids during ammonium chloride treatment may have contributed to successful management.

Because there is little evidence for conversion from an IV to enteral dose of ammonium chloride, the calculated enteral dose was based upon the recommended IV dose (mEq/dose); this assumes a bioavailability of 100%. Because the patients were given the IV product enterally and the bioavailability of ammonium chloride is unknown, equations for metabolic alkalosis correction were used as a guide to determine the dose (mEq).⁶ Of the 3 known equations, selection and final calculated correctional amount were decided by the primary physician with assistance from a clinical pharmacist. According to the correction equation for IV ammonium chloride, to prevent overcorrection, the dose should be divided to provide one-third to one-half of the calculated amount, and patients should be reevaluated before receiving additional dosages. For the purpose of this case series, dosages were calculated and compared to what the physicians administered to each patient. The calculated dose of ammonium chloride based on the correction for BE equation did not compare with the administered dose in patients A and B (Table 2). Patient C's dose initially correlated but required escalation to continually decrease BE to normalization (Table 2). This leads to the hypothesis that using the corrective equations for IV ammonium chloride will not correlate with the required dosage for enteral administration, possibly due to bioavailability of less than 100%.

Table 2.

Indication, Dose Description, and Concomitant Medications for the 3 Patients

Figures 1 through 3 show each patient's BE in relation to dose. Patients A and C both demonstrated an unexpected increase in BE after the first dose of ammonium chloride. The effects of ammonium chloride on patient C subsequently resulted in a decrease from pre- to post-BE values after each dose.

A potential cause of metabolic alkalosis in all three patients is excessive volume depletion from multiple doses of diuretics. According to the treatment algorithm for metabolic alkalosis, removal of the causative agent is recommended to improve alkalosis.¹ Continuation of diuretic therapy in Patient B may have prevented the resolution of metabolic alkalosis. Management with ammonium chloride was potentially futile and would have required repeated doses to continually reduce the BE.

For all three patients, no adverse effects were documented in the patient charts to suspect ammonium chloride-induced encephalopathy or toxicity. This is important to note due to the case report of a 54-year-old Caucasian woman with metabolic alkalosis who experienced adverse effects associated with ammonium chloride IV infusion (280 mEq over 14 hours).⁸ Several hours into the infusion, the patient became increasingly confused, combative, lethargic, and ultimately comatose. Approximately 16 hours after ammonium chloride was discontinued, the patient was arousable with a normal mental status. Evaluation of ammonium chloride-induced encephalopathy and toxicity in these patients may have been difficult for providers, as all three patients were intubated and were 3 years of age or less. This may indicate why there was a lack of documented adverse effects associated with the enteral ammonium chloride.

Further investigations are required to analyze the potential efficacy and adverse effects of this alternative route of ammonium chloride in a larger pediatric population size. Ammonia levels could be an indicator to correlate with efficacy or toxicity and determine bioavailability. In this case series, the use of enteral ammonium chloride injection in 2 of 3 pediatric patients achieved resolution of metabolic alkalosis. Administration of ammonium chloride enterally for the treatment of persistent metabolic alkalosis may be a feasible and well-tolerated option.

ABBREVIATIONS

ABG

arterial blood gases

BE

base excess

FiO₂

fraction of inspired oxygen

KVO

keep vein open

PMH

past medical history

PICU

pediatric intensive care unit

PEEP

positive end-expiratory pressure

RR

respiratory rate

TV

tidal volume

VDRF

ventilator dependent respiratory failure

Table 1.

Endpoints for Patients Who Achieved Resolution of Metabolic Alkalosis