We studied 7 patients aged 15 months-13 years, including a previously published case. The study design is shown in . All subjects presented with neonatal hyperammonemia and were documented to have a diagnosis of propionic acidemia by urine organic acid analysis. Molecular testing results were available for 3 patients. The study was approved by the Institutional Review Boards at the Children’s National Medical Center and the Children’s Hospital of Philadelphia. Informed consent was obtained for each individual before enrollment. This trial has been registered at Clinicaltrials.gov.
Figure 1 Overall trial timeline. Identical studies were performed before (‘day 1’) and immediately following (‘day 4’) a 3-day course of N-carbamylglutamate at a dose of 100 mg/kg/d or 2.2g/m2/d if ≥25 kg. Plasma samples (more ...)
An indwelling venous catheter was used for blood drawing in subjects who did not have a central venous line. For subjects without ready venous access, a peripherally inserted central catheter was introduced under conscious sedation (midazolam, fentanyl) by an experienced interventional radiologist. The PICC remained in place for the duration of the study, and was removed upon obtaining the last blood sample.
Following an overnight fast, a baseline sample (5 mL) of heparinized blood was obtained. At time 0, each subject received (either by ingestion or nasogastric tube) 0.33 mmol/kg, (27.5 mg/kg) of [1-13
C] sodium acetate (98 atom % excess) dissolved in 60 mL of water. Blood samples (2-4 mL) were subsequently obtained at 15, 30, 45, 60, 75, 90, 120, 180, and 240 min. For subjects under 5 years, the times were restricted to 15, 30 50 and 120 minutes in order to limit total blood taken to less than 3 ml/kg body weight. Each specimen was placed in a pre-cooled, heparinized tube and immediately centrifuged at 4°C to separate the plasma, which was immediately frozen on dry ice and kept frozen at -70°C until analysis performed within 2 days after colelction. Each sample was analyzed for plasma ammonia and urea (RXL Dade Behring, Siemens Healthcare Diagnostics, Deerfield, IL), and quantitative amino acids (Biochrom, Cambridge, UK) and [13
C]urea as previously described [8
All subjects underwent an identical study procedure before, and immediately after treatment for 3 days with NCG (Carbaglu®, Orphan Europe, Paris, 100 mg/kg/d, or 2.2g/m2/d if ≥ 25 kg body weight in four divided doses).
Subject 1 was an 11 year-old female born at term following an uncomplicated pregnancy, labor and delivery. She presented at 2 days of life with hyperammonemia and acidosis. She was treated for presumptive organic acidemia until laboratory results confirmed the diagnosis of propionic acidemia. She developed seizures at two weeks of life and required anti-epileptic drugs until one year of age. She required multiple admissions for acute illness leading to metabolic decompensation. She had a mild developmental delay, and was fed by a gastrostomy tube. Her historic ammonia concentrations while stable ranged between 81 and 127 μM.
Subject 2 was an 8 year-old female born at term following uncomplicated pregnancy labor and delivery. She presented on the 3rd day of life with hypothermia, metabolic acidosis and hyperammonemia (808 μM). She was diagnosed with propionic acidemia by urinary organic acid and plasma acylcarnitine analyses. She required hemodialysis and ventilatory and vasopressor support. She also had a dilated cardiomyopathy in the newborn period, requiring medications, but this subsequently resolved. Her ammonia levels ranged chronically between 55 to 122 μM. She required over 30 admissions or emergency department visits for acute exacerbations of her condition. She had developmental delay, including compromised speech, and received nutrition via a gastrostomy tube.
Subject 3 was a 5 year-old male born at term following uncomplicated pregnancy, labor and delivery. At 2 weeks of age, he presented with vomiting and lethargy, and was found to have a mild metabolic acidosis and a plasma ammonia level of 382 μM. A diagnosis of propionic acidemia was subsequently made by plasma acylcarnitine and urine organic acid analyses. After the initial episode, he was admitted 14 times for acute illness, including 5 admissions for pancreatitis. He had a mild developmental delay, more accentuated in language. He required continuous feeding via a gastrostomy tube. He was found to be heterozygous for a c.183+3 G>C mutation in the beta-subunit gene of propionyl-CoA carboxylase (PCCB). He was also noted to be heterozygous for two undocumented sequence variants, both potentially pathogenic: c.734G>A and c.967-14 A>G. His chronic plasma ammonia concentration ranged from 36 to 110 μM.
Subject 4 was a 15 month female born at 34 weeks gestation after uncomplicated labor and delivery. Apgar scores were 8 at 1 minute and 9 at 5 minutes. At 10 days of age, following findings of lethargy and hypotonia, she was noted to have acidosis, ketosis and a plasma ammonia level of 459 μM. Peak ammonia in the neonatal period was 1349 μM.
The diagnosis of propionic acidemia was confirmed via urine organic acid analysis. She required hemodialysis during the first episode. She has since had three episodes of hyperammonemia and acidosis requiring admission to the hospital. Her chronic ammonia concentrations ranged between 77 to 94 μM.
Subject 5 was a 13 year-old female born at term by a caesarean section. At 4 days of age, she presented with lethargy, acidosis, and a plasma ammonia level of 735 μM. The diagnosis of propionic acidemia was confirmed via urine organic acid analysis. She has had over 100 admissions to the hospital, primarily for lethargy and acidosis. One random plasma ammonia concentration was 47 μM. Molecular genetic tests are still pending, but her sister, subject 6, was found to be a homozygote for a mutation in the PCCB gene (see below).
Subject 6 was a 9 year-old sister of subject 5. She was born at 38 weeks gestation by a planned caesarean section. She was prospectively diagnosed in the newborn period due to her affected sister. She has been admitted to the hospital nearly 200 times, primarily for lethargy, vomiting and hyperammonemia. She was found to be homozygous for a known deleterious mutation in the PCCB gene, c.1218_1231delins12 (deletion of 14 bp, and insertion of 12 bp)[20
]. Her chronic ammonia levels ranged between 94 and 171 μM.
Subject 7 was a 6 year-old male, born at term, who presented at 4 days of age with severe hyperammonemia (1200 μM) and metabolic acidosis requiring hemodialysis. The diagnosis of propionic acidemia was established by urine organic acid analysis. He was admitted to the hospital on multiple occasions during the first 3 years of life, typically for acute exacerbations of his disorder, but once for acute pancreatitis. During these episodes, his plasma ammonia rose to 3 to 6 times the upper limit of normal. His development was significantly delayed: he was ambulatory but with no speech. He was fed exclusively by a gastrostomy tube. Mutation analysis revealed that he was homozygous for the frame-shift G216fs mutation in the alpha-subunit of the propionyl-CoA carboxylase A gene. A single random plasma ammonia concentration was 74 μM. Results from his study have been previously reported by our group[8
To evaluate the biochemical effects of NCG treatment, we used longitudinal mixed effects linear regression in Stata 10 (xtmixed). We used it first to compare grand mean levels of metabolites in the same participants when treated and not treated with NCG. The method allows us to perform a paired (pre vs post NCG) analysis and also to adjust variance estimates for the correlation of multiple measurements on the same subject within treatment condition. Thereafter, we used the same model to evaluate the detailed change over time within treatment condition. The model treated each participant as a random effect allowing him/her to have a different starting point prior to each treatment condition. In addition, the model allowed for nonlinear effects over time and different relationships (interactions) in time by treatment condition when estimating metabolic effects. Before we implemented the above parametric analyses, we evaluated the normality and variance homogeneity assumptions and when necessary implemented data transformations to ensure the data met these assumptions. Following analyses involving data transformations, the results were back transformed to return the estimates to their original scale and units. We set 95% confidence intervals (CI) around each estimate, as well as around group differences, and considered those differences to be statistically significant if the 95% CI failed to include zero.