Short-Term Outcome of N-Carbamylglutamate in the Treatment of Acute Hyperammonemia: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for Short-Term Outcome of N-Carbamylglutamate in the Treatment of Acute Hyperammonemia (NCT01599286), a randomized phase 2 study. The primary outcome is Time to the Primary Outcome (Earlier of Ammonia <50 µmol/L or Hospital Discharge).

2. Clinical Question

The trial asked whether N-carbamylglutamate (NCG), given with standard care, could shorten the time to resolution of acute hyperammonemia compared with placebo given with standard care. Its primary outcome was the earlier of reaching an ammonia level of 50 µmol/L or less and hospital discharge. The registry’s outcome title refers to an ammonia level below 50 µmol/L; its outcome description defines the threshold as 50 µmol/L or less.

The primary analysis was described as an intention-to-treat composite and as a survival analysis based on time to the earlier event. The record reports hazard ratios separately for PA/MMA episodes and CPS1D/OTCD episodes. It states that higher ratios reflect an NCG advantage. The clinical question therefore concerns the relative timing of the composite outcome within these diagnosis groupings, not a single pooled estimate across all conditions.

The investigators also described interest in neurologic function, plasma glutamine, and duration of hospitalization. Those aims were secondary to the ammonia-related primary question; secondary outcome results are not reported in the registry record. The posted primary results alone cannot establish whether changes in those other outcomes accompanied the primary endpoint.

3. Trial Design

This completed phase 2 study used randomized allocation, quadruple masking, and a parallel model. The two interventions were Carbaglu and placebo; the arm descriptions specify that both were given alongside standard care. The registry lists an enrollment of 35. It also lists primary outcome results for PA/MMA episodes, with 24 episodes, and CPS1D/OTCD episodes, with 11 episodes. These episode counts describe the posted outcome groupings and should not be read as a statement about the number of unique people represented in each grouping.

The listed conditions were propionic acidemia type I and/or type II, methylmalonic acidemia, carbamoyl-phosphate synthase I deficiency disease, and ornithine carbamoyltransferase deficiency. The outcome description says that patients were randomized to study drug or placebo at each hyperammonemia episode analyzed by diagnosis. It does not describe how repeated episodes from the same patient were handled in the statistical model, or whether the analysis accounted for any correlation between episodes from one person.

The study started on 2012-09-01, reached primary completion on 2019-04-30, and is listed as completed. The sponsor is Mendel Tuchman. A detailed description of the randomization procedure, masking roles, treatment duration, and allocation concealment is not reported in the registry record. The record also does not report a sample-size justification, a protocol-level analysis plan, or whether there were interim analyses. These details should not be inferred from the randomized, masked, parallel design labels.

Although the intervention names and design identify a two-group comparison, the posted primary results are separated by diagnosis grouping rather than presented as one overall estimate. The record describes the primary outcome as based on all measurements of hyperammonemia averaged over up to 7 days, while also noting that ammonia measurements occurred at uncontrolled times during episodes. This makes the timing and interpretation of repeated measurements important to understanding the result.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Before assignment: There were 65 eligible patients, 35 were enrolled with a randomization of hyperammonemia. Eligible patients were randomized at each episode of hyperammonemia. To be enrolled, a patient had to have an eligible episode of hyperammonemia and was then randomized to either placebo or NCG at each randomization. 13 patients had multiple randomizations, 22 only had one.

Overall Study

Analysis population for the primary outcome, as defined in the registry: Patients were randomized to either study drug or placebo at each episode of hyperammonemia analysed by diagnosis. Data presented as a hazard ratio based on the time to reach an ammonia level of ≤50 µmol/L. Higher ratios reflect an NCG advantage.

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

Population: 65 participants were eligible based on diagnosis. There were 35 individual participants enrolled due to a hyperammonemia episode. 24 individuals had episodes where they were randomized to the NCG arm, and 23 had episodes where they were randomized to the placebo arm. 13 participants had multiple randomizations to both arms throughout the study, while 22 participants only had a single randomization to either placebo of NCG. 35 unique participants were randomized providing 106 randomized episodes.

6. Primary Endpoint

The registry lists the primary outcome as Time to the Primary Outcome (Earlier of Ammonia <50 µmol/L or Hospital Discharge), measured over Average of all measurements of hyperammonemia, for up to 7 days.

The primary endpoint is a time-to-event composite: the time until the earlier of an ammonia level at or below 50 µmol/L or hospital discharge. The registry describes the analysis as a survival analysis and reports a hazard ratio based on time to reach that earlier event. Discharge is included because the investigators considered it a point at which the patient was no longer at risk of neurological injury from ammonia.

A composite endpoint can capture whichever of its specified events occurs first, rather than requiring every participant to experience the same event. Time-to-event methods are generally useful when both whether an event occurs and when it occurs are relevant. In general, they can use information about event timing rather than reducing follow-up to a simple event count at a fixed point. That general explanation does not establish which specific modeling choices were made in this trial beyond the survival analysis described in the registry.

The registry says the primary outcome was assessed over an average of all hyperammonemia measurements for up to 7 days. It also cautions that post-treatment ammonia measurements were taken at uncontrolled times. If testing frequency and timing differ across episodes, a simple average could be biased; the record therefore says it is difficult to compute a meaningful average. The time-to-event analysis is the posted primary approach, but further details about how event times were assigned between tests are not reported in the registry record.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

The registry reports a hazard ratio and a 95% confidence interval for a time-to-event primary outcome. The following concepts are general explanations of these methods; details of the trial’s exact statistical implementation are not reported in the registry record.

What does a time-to-event endpoint represent?

A time-to-event endpoint records the time until a defined event occurs. Here, the registry defines the primary event as the earlier of reaching an ammonia level of 50 µmol/L or less and hospital discharge. The analysis is described as a survival analysis, a family of methods used for outcomes where timing matters and where participants may have different observed follow-up times.

In general, time-to-event analyses can account for the fact that an event may not have occurred by the end of observation. Such observations are often called censored. The registry record does not state how censoring was handled in this trial, how event times were determined between ammonia tests, or whether any other events affected follow-up. Those are important implementation details and cannot be inferred just from the phrase “survival analysis.”

How should a hazard ratio be read?

A hazard ratio is a relative measure used in time-to-event analysis. It compares event rates over time under a model; it is not the same as a ratio of average times or an absolute difference in time. The trial record describes its reported ratio as reflecting the relative probabilities in the two groups of reaching the endpoint and states that higher ratios reflect an NCG advantage. The estimate should be read using that stated direction, while recognizing that the record does not give enough modeling detail to specify all assumptions behind the estimate.

A ratio of 1 is the reference value for no relative difference on the ratio scale. Values above or below 1 indicate opposite directions under a consistent definition of the ratio. The registry’s direction statement is especially useful because the intervention order and the way a time-to-event model encodes groups can otherwise be easy to misread. For PA/MMA, the point estimate is 1.37; for CPS1D/OTCD, it is 0.79. The uncertainty intervals around both estimates include 1.

What information does a confidence interval add?

A 95% confidence interval conveys uncertainty in an estimate. In general, if the same study and method were repeated many times, intervals constructed in this way would cover the underlying value at the stated frequency, subject to the method’s assumptions. It is not a probability statement that assigns a 95% chance to the true value being inside this particular interval.

The PA/MMA interval runs from 0.88 to 2.13. The CPS1D/OTCD interval runs from 0.22 to 2.81. Since each interval includes 1, the results remain compatible with no relative difference on the hazard-ratio scale, as well as with values in either direction. The width of an interval is a practical indication of precision: a broad interval leaves more uncertainty about the magnitude and direction than a narrow one. The CPS1D/OTCD interval is broad, and its result is correspondingly uncertain.

Why does the endpoint combine ammonia threshold and discharge?

In general, a composite endpoint can represent a clinically relevant event defined by more than one component, with the first component to occur determining the event time. The record explains that discharge was treated as a point at which a patient was no longer at risk of neurological injury from ammonia. This rationale links the discharge component to the concern motivating the ammonia threshold. Still, the registry does not report separate estimates for the threshold component and discharge component, so the reported hazard ratio does not show which component contributed more to the result.

The registry’s title describes the ammonia criterion as below 50 µmol/L, while the detailed outcome description uses 50 µmol/L or less. The latter is the explicit operational definition in the outcome description. The distinction is worth retaining when interpreting the endpoint; it should not be silently replaced with a different threshold.

Why can measurement timing matter?

The record says that multiple post-treatment ammonia levels were measured at uncontrolled times during an episode. In general, measurements made at different intervals can affect the apparent timing of threshold attainment: a threshold crossing can only be observed when a measurement is taken. More frequent testing may identify a crossing sooner than less frequent testing, even if underlying biological trajectories were similar. The registry explicitly notes that the timing and frequency of testing make a meaningful average difficult to compute without potential bias. It does not report a method that corrects for this measurement schedule, so no such adjustment should be assumed.

What does the diagnosis-specific presentation imply?

The posted primary results are separated into PA/MMA and CPS1D/OTCD episode groupings. This presentation allows the estimates to be read within those categories, but it does not establish that the treatment effect differs between them. In general, a formal interaction analysis or other prespecified comparison would be needed to assess variation by diagnosis. Such an analysis is not reported in the registry record. The episode counts listed for the groupings are 24 and 11, respectively, and the estimates have different levels of precision.

The record describes randomization at each episode and analysis by diagnosis. It does not say whether participants could contribute more than one episode or how any within-participant correlation was addressed. In general, repeated observations from one participant may not be statistically independent. A model that treats correlated observations as independent can misstate uncertainty; however, whether that issue arose or how it was handled in this trial is not reported in the registry record.

What can be concluded from the posted method description?

The registry establishes that the primary outcome was analyzed as a survival outcome and reported as hazard ratios with 95% confidence intervals. It does not report the model specification, assumption checks, missing-data approach, censoring rules, analysis software, or the detailed statistical plan. These omissions limit reproducibility and make it inappropriate to infer a particular model or diagnostic procedure. The results can still be described as posted, while distinguishing the reported estimates from details that are not available in the registry record.

10. Multiplicity, Missing Data and Analysis Populations

The registry lists one outcome measure and zero posted analyses. It describes the primary endpoint as an intention-to-treat composite, but the record does not report a full analysis plan, the exact set of planned comparisons, or whether any multiplicity adjustment was used. Because multiple diagnoses are presented separately and the investigators described additional secondary aims, the absence of a reported adjustment method matters when considering how many statistical questions may have been examined. It does not establish that an adjustment was or was not used outside the information in the registry record.

The brief summary identifies neurologic function, plasma glutamine, and duration of hospitalization as additional interests. Results for secondary outcomes are not reported in the registry record. No conclusion about those outcomes can be drawn from the posted primary hazard ratios. The record also does not report subgroup interaction tests, adjusted comparisons, or a single pooled estimate across diagnosis groups.

The outcome description refers to an intention-to-treat analysis, but further details about the analysis population are not reported in the registry record. In particular, the record does not describe exclusions after randomization, handling of protocol deviations, or the relationship between the number of patients enrolled and the number of episodes contributing to each result.

Missing outcome data, loss to follow-up, censoring rules, and imputation procedures are not reported in the registry record. These details can affect a time-to-event analysis, so their absence limits assessment of whether missingness could have influenced the estimates. The registry also notes that ammonia testing times were uncontrolled, but it does not explain how unobserved threshold crossings or uneven measurement schedules were handled analytically. No assumptions about missing data, multiplicity, or analysis populations should be attributed to the study beyond the stated intention-to-treat description.

11. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: gastrointestinal disorders (20.0%), infections and infestations (16.0%), nervous system (12.0%), injury (4.2%), metabolic and nutritional (4.2%).

The registry reports serious adverse events and other adverse events separately for placebo and NCG. For serious adverse events, 24 participants were at risk in the placebo group and 6 were affected; 25 were at risk in the NCG group and 8 were affected. For other adverse events, 24 were at risk in the placebo group and 9 were affected; 25 were at risk in the NCG group and 11 were affected. The registry also lists 10 deaths in the placebo group and 14 in the NCG group.

These are posted counts, not estimates of treatment-caused risk. The record does not provide enough detail to determine the timing, causes, relationship to study treatment, or overlap among serious events, other events, and deaths. It also does not state whether the death counts refer to the same observation period or analysis population as each adverse-event count. The at-risk figures and affected counts should therefore be read as the registry’s safety reporting, not converted into comparative rates or used to infer causation.

The registry lists serious-event categories and percentages, including gastrointestinal disorders at 20.0%, infections and infestations at 16.0%, and nervous system events at 12.0%. The denominator and group allocation for these category percentages are not reported in the registry record. They describe the listed categories but cannot be used to compare treatment groups without that context.

12. What the Result Does, and Does Not, Mean

The primary results provide diagnosis-specific estimates for the relative timing of a composite event: reaching an ammonia level of 50 µmol/L or less or hospital discharge, whichever occurred first. The point estimate is 1.37 for PA/MMA episodes and 0.79 for CPS1D/OTCD episodes. Under the registry’s stated direction, higher ratios reflect an NCG advantage. The confidence intervals are 0.88 to 2.13 and 0.22 to 2.81, respectively, and both include 1.

These estimates do not establish a clear direction for either diagnosis grouping. An interval including 1 is not proof that the groups are identical, and it is not evidence of equivalence. It indicates that the posted results remain compatible with no relative difference on the ratio scale, as well as with effects in either direction. The broad range for CPS1D/OTCD especially limits precision about the magnitude of the comparison.

The hazard ratios are not absolute time differences. They do not say how many hours or days earlier participants reached the threshold or were discharged, and the record does not post absolute event probabilities by a specified time. The composite also combines two possible events, and separate component results are not reported. Consequently, the estimate cannot identify whether threshold attainment or discharge was more influential in the observed comparison.

The primary estimates do not answer the secondary questions about neurologic function, plasma glutamine, or hospitalization duration as separate outcomes. Those results are not reported in the registry record. The safety counts likewise do not establish whether an adverse event was caused by treatment. The record’s description of uncontrolled ammonia measurement times is a further interpretive constraint, because testing schedules can affect when threshold attainment is observed.

In general, a randomized and masked comparison can reduce some sources of bias, but the registry’s design labels do not resolve every question about analysis conduct or precision. Details about repeated episodes, missing data, model assumptions, and the statistical plan are not reported in the registry record. The most defensible reading is therefore limited to the posted estimates, their confidence intervals, the outcome definition, and the stated direction of the ratio.

13. Why This Trial Matters Statistically

This trial raises several statistical issues that are important for studies of acute episodes. The primary outcome concerns both the occurrence and timing of resolution, so a time-to-event approach is a natural way to express the question. The composite combines an ammonia threshold with hospital discharge, making the endpoint relevant to the investigators’ stated concern about ongoing neurological risk while also requiring care in interpreting what event drove the result.

The study’s outcome description highlights a practical measurement problem: ammonia tests were taken at uncontrolled times. In general, the time a threshold is first observed is not necessarily the exact time it was crossed. If measurement schedules vary, apparent event times can be affected by observation patterns. The registry recognizes that a simple average could be biased by the frequency and timing of testing, making this an important issue rather than a purely technical detail.

The results are presented for two diagnosis groupings, with separate estimates and confidence intervals. This makes visible that the estimates are not identical, while their uncertainty intervals remain broad and include 1. It also illustrates why point estimates alone should not be used to claim that effects differ across diagnoses. A formal test of heterogeneity is not reported in the registry record.

The record describes randomization at each episode, which draws attention to the distinction between participants and episodes. If a participant contributes multiple episodes, those observations may be related; the registry does not report how that possibility was handled. Clear reporting of episode-level inclusion, dependence between repeated episodes, and event-time construction would help readers assess the precision and reproducibility of the analysis.

More broadly, this record demonstrates why estimates should be read together with their confidence intervals and analysis details. The posted ratios convey a relative comparison, but they do not supply absolute time differences, separate composite components, or secondary outcome results. Statistical interpretation depends not only on the numerical estimate but also on the endpoint definition, observation process, and limits of what has been reported.

14. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry describes the primary outcome as a survival analysis but does not report the model specification, the assumptions assessed, or the method used to determine event times between ammonia tests. It notes that testing occurred at uncontrolled times, which can complicate the timing of observed threshold attainment. A meaningful average could be biased by test frequency and timing, according to the outcome description, but no adjustment approach is described.

Episodes were analyzed by diagnosis, and randomization is described as occurring at each episode. The record does not report whether participants contributed repeated episodes or how within-participant dependence was handled. It also does not explain the relationship between the overall enrollment of 35 and the episode counts listed for the two diagnosis groupings. These gaps make the analysis population and the basis of the reported uncertainty difficult to assess fully.

The confidence intervals are broad, particularly for CPS1D/OTCD, and include 1 for both groupings. The registry does not report a pooled result or a formal test comparing the diagnosis-specific estimates. Separate point estimates should not be treated as evidence of a diagnosis-by-treatment difference.

Other limitations of interpretation include the composite endpoint and the lack of separate results for its ammonia-threshold and discharge components. The registry does not post results for the stated secondary interests in neurologic function, plasma glutamine, or hospitalization duration. Details about missing outcome data, censoring, multiplicity, analysis populations, protocol deviations, and interim analyses are not reported in the registry record. These limitations do not invalidate the posted values, but they restrict the conclusions that can be drawn from them.

15. Related Tutorials

16. Related Statistical Calculators

17. Sources

• ClinicalTrials.gov: Short-Term Outcome of N-Carbamylglutamate in the Treatment of Acute Hyperammonemia — NCT01599286.
• Registry record last updated 2021-02-15; status last verified 2021-01.
• Results first posted 2021-02-15.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

18. Record Summary

The registry lists a completed, randomized, quadruple-masked, parallel phase 2 trial comparing NCG plus standard care with placebo plus standard care for acute hyperammonemia in PA, MMA, late-onset CPS1 deficiency, and late-onset OTCD. Enrollment is listed as 35, with the primary outcome reported for PA/MMA and CPS1D/OTCD episode groupings.

The primary outcome was time to the earlier of an ammonia level of 50 µmol/L or less or hospital discharge, analyzed as a survival outcome. The posted hazard ratio is 1.37, with a 95% confidence interval of 0.88 to 2.13, for PA/MMA episodes; for CPS1D/OTCD episodes it is 0.79, with a 95% confidence interval of 0.22 to 2.81. The registry states that higher ratios reflect an NCG advantage. Both intervals include 1, and the record does not report secondary outcome results or further detail on key analysis procedures.