50mL 0.9% saline

ICU length of stay reflects the number of ICU admission days.

Hospital length of stay reflects how long it takes a diabetic ketoacidosis patient to recover to the point where he/she can be released from the hospital.

The duration of insulin therapy is calculated by examining the hospital clinical information systems for all records of IV insulin infusion beginning at the time of enrollment. The start and the stop time-stamps of medication infusion are used to calculate a duration of infusion.

The SOFA score is a validated measure of organ dysfunction commonly used in critically ill patients, particularly those with sepsis. It evaluates the function of six organ systems-respiratory, cardiovascular, neurologic, hepatic, renal, and coagulation-based on routinely collected clinical and laboratory data. Each organ system is assigned a score from 0 (normal function) to 4 (most severe dysfunction), resulting in a total score ranging from 0 to 24. Higher SOFA scores indicate greater severity of organ failure and a higher risk of adverse outcomes. SOFA score in this study uses a modification in which the arterial oxygen saturation/fraction of inspired oxygen (SaO2 /FiO2) ratio is substituted for the partial pressure of arterial oxygen/fraction of inspired oxygen (PaO2 /FiO2 ratio).

Oxygen consumption rate of circulating peripheral blood mononuclear cells measured using Seahorse. Basal respiration represents oxygen consumption under baseline conditions. ATP-linked respiration represents oxygen consumption coupled to ATP production, while proton leak represents oxygen consumption not linked to ATP synthesis. Maximal respiration represents the maximal capacity of the electron transport chain after uncoupling. Spare respiratory capacity represents the difference between maximal and basal respiration and indicates the ability to respond to increased energy demand. Non-mitochondrial respiration represents oxygen consumption independent of mitochondrial activity.

Secondary outcome of change in PDH specific activity over 72 hours or at discharge if patient's hospital length of stay was less than 72 hours. PDH activity and total PDH protein quantity were measured in isolated peripheral blood mononuclear cells (PBMCs) following selective disruption of the mitochondrial membrane, using a previously validated immunocapture and microplate-based enzymatic assay protocol. PDH-specific activity was calculated as the ratio of measured PDH enzymatic activity to the natural logarithm of PDH protein quantity (PDH activity / ln[PDH quantity]), providing a normalized metric that reflects the functional efficiency of the enzyme independent of its expression level. Higher values suggest higher enzymatic efficiency.

The Hopkins Verbal Learning Test-Revised (HVLT-R) Total Recall score is a measure of verbal learning and immediate memory. Participants are read a list of 12 words (from three semantic categories) and asked to recall as many words as possible across three consecutive learning trials. The Total Recall score is calculated as the sum of correctly recalled words across the three trials, yielding a possible range of 0 to 36, with higher scores indicating better verbal learning and memory performance. This score reflects both initial acquisition and short-term retention of verbal information and is commonly used to assess cognitive function in clinical research.

The Hopkins Verbal Learning Test-Revised (HVLT-R) Delayed Recall score assesses retention of verbal information after a delay. Following completion of the three learning trials, participants are asked to recall the previously presented word list after a delay period (typically 20-25 minutes) without re-exposure to the words. The Delayed Recall score is calculated as the number of correctly recalled words, with a possible range of 0 to 12, where higher scores indicate better memory retention. This measure reflects the ability to consolidate and retrieve learned information over time and is commonly used to evaluate memory function in clinical and research settings.

The Hopkins Verbal Learning Test-Revised (HVLT-R) Retention (%) score reflects the proportion of learned information that is retained over a delay. It is calculated as the ratio of the Delayed Recall score to the highest number of words recalled on any of the three immediate recall trials, multiplied by 100, yielding a percentage value. Scores typically range from 0% to 100%, with higher values indicating better retention of previously learned material. This measure provides an index of memory retention independent of initial learning performance.

The Hopkins Verbal Learning Test-Revised (HVLT-R) Recognition Discrimination Index assesses recognition memory by evaluating the ability to distinguish previously learned words from novel distractors. During the recognition phase, participants are presented with a list of target words and distractor words and asked to identify those that were previously learned. The Recognition Discrimination Index is calculated as the number of true positives (correctly identified target words) minus the number of false positives (incorrectly identified distractor words), yielding a score typically ranging from -12 to 12. Higher scores indicate better recognition accuracy and discrimination ability, reflecting the integrity of memory retrieval processes.

The Brief Visuospatial Memory Test (BVMT) assesses visuospatial learning and memory. Participants are shown six geometric designs for a brief period and asked to reproduce them from memory across three learning trials. Performance is scored based on the accuracy and placement of each design, with a total recall score ranging from 0 to 36, where higher scores indicate better visuospatial memory. A delayed recall trial is administered after a delay to assess retention of visual information. This measure evaluates the ability to encode, store, and retrieve visuospatial material.

The Trail Making Test Part A (TMT-A) assesses visual attention, processing speed, and psychomotor function. Participants are instructed to connect numbered circles in sequential order as quickly as possible. The primary outcome is the time to completion, measured in seconds, with lower times indicating better performance. There is no fixed maximum score, although testing is typically discontinued at a predefined time limit (commonly 300 seconds).

The Trail Making Test Part B (TMT-B) assesses executive function, including cognitive flexibility and set-shifting, in addition to processing speed. Participants are required to alternate between numbers and letters in sequence (e.g., 1-A-2-B) as quickly as possible. The primary outcome is the time to completion, measured in seconds, with lower times indicating better performance. As with TMT-A, there is no fixed maximum score, and testing is typically discontinued at a predefined time limit (commonly 300 seconds).

The Wechsler Adult Intelligence Scale-Fourth Edition (WAIS-IV) Digit Span subtest assesses attention, working memory, and concentration. Participants are asked to repeat sequences of numbers in forward order (Digit Span Forward), reverse order (Digit Span Backward), and ascending order (Digit Span Sequencing). Scores are based on the total number of correctly recalled sequences, with higher scores indicating better attention and working memory capacity. The total score ranges from 0 to 48.

The Animal Naming test is a measure of semantic verbal fluency and executive function. Participants are asked to name as many animals as possible within a fixed time period. The score is the total number of unique, correct animal names generated, with higher scores indicating better semantic memory retrieval and executive functioning. Scores range from 0 with no fixed upper limit.

The Test of Premorbid Functioning (TOPF) estimates an individual's baseline cognitive ability prior to illness or injury. Participants are asked to read aloud a list of irregularly spelled words, and performance is scored based on correct pronunciation. Scores are used to estimate premorbid intellectual functioning. Scores range from 0 to 70, with higher scores indicating higher estimated premorbid functioning. This measure helps contextualize current cognitive performance relative to expected baseline levels.

Total of all reporting groups

Two randomized participants did not receive study drug and were excluded from the analysis population.

Two randomized participants did not receive study drug and were excluded from the analysis population.

Two randomized participants did not receive study drug and were excluded from the analysis population.

Two randomized participants did not receive study drug and were excluded from the analysis population.

Two randomized participants did not receive study drug and were excluded from the analysis population.

Two randomized participants did not receive study drug and were excluded from the analysis population.

Two randomized participants did not receive study drug and were excluded from the analysis population.

Two randomized participants did not receive study drug and were excluded from the analysis population.

Two randomized participants did not receive study drug and were excluded from the analysis population.

Two randomized participants did not receive study drug and were excluded from the analysis population.

Two randomized participants did not receive study drug and were excluded from the analysis population.

The SOFA score is a validated measure of organ dysfunction commonly used in critically ill patients, particularly those with sepsis. It evaluates the function of six organ systems-respiratory, cardiovascular, neurologic, hepatic, renal, and coagulation-based on routinely collected clinical and laboratory data. Each organ system is assigned a score from 0 (normal function) to 4 (most severe dysfunction), resulting in a total score ranging from 0 to 24. Higher SOFA scores indicate greater severity of organ failure and a higher risk of adverse outcomes.

Two randomized participants did not receive study drug and were excluded from the analysis population.

Secondary outcome of lactate over 24 hours (6, 12, 18, 24 hours) following enrollment

Secondary outcome of anion gap over 24 hours (6, 12, 18, 24 hours) following enrollment

ICU length of stay reflects the number of ICU admission days.

Hospital length of stay reflects how long it takes a diabetic ketoacidosis patient to recover to the point where he/she can be released from the hospital.

The duration of insulin therapy is calculated by examining the hospital clinical information systems for all records of IV insulin infusion beginning at the time of enrollment. The start and the stop time-stamps of medication infusion are used to calculate a duration of infusion.

The SOFA score is a validated measure of organ dysfunction commonly used in critically ill patients, particularly those with sepsis. It evaluates the function of six organ systems-respiratory, cardiovascular, neurologic, hepatic, renal, and coagulation-based on routinely collected clinical and laboratory data. Each organ system is assigned a score from 0 (normal function) to 4 (most severe dysfunction), resulting in a total score ranging from 0 to 24. Higher SOFA scores indicate greater severity of organ failure and a higher risk of adverse outcomes. SOFA score in this study uses a modification in which the arterial oxygen saturation/fraction of inspired oxygen (SaO2 /FiO2) ratio is substituted for the partial pressure of arterial oxygen/fraction of inspired oxygen (PaO2 /FiO2 ratio).

C-peptide levels measured at time of study drug administration, 24 hours, and at 72 hours (or at discharge if hospital length of stay was less than 72 hours).

C-peptide measurements were available for 98 patients because of incomplete sample collection.

Oxygen consumption rate of circulating peripheral blood mononuclear cells measured using Seahorse. Basal respiration represents oxygen consumption under baseline conditions. ATP-linked respiration represents oxygen consumption coupled to ATP production, while proton leak represents oxygen consumption not linked to ATP synthesis. Maximal respiration represents the maximal capacity of the electron transport chain after uncoupling. Spare respiratory capacity represents the difference between maximal and basal respiration and indicates the ability to respond to increased energy demand. Non-mitochondrial respiration represents oxygen consumption independent of mitochondrial activity.

Not all participants had analyzable data for each mitochondrial respiration parameter due to missing or incomplete measurements.

Secondary outcome of change in PDH specific activity over 72 hours or at discharge if patient's hospital length of stay was less than 72 hours. PDH activity and total PDH protein quantity were measured in isolated peripheral blood mononuclear cells (PBMCs) following selective disruption of the mitochondrial membrane, using a previously validated immunocapture and microplate-based enzymatic assay protocol. PDH-specific activity was calculated as the ratio of measured PDH enzymatic activity to the natural logarithm of PDH protein quantity (PDH activity / ln[PDH quantity]), providing a normalized metric that reflects the functional efficiency of the enzyme independent of its expression level. Higher values suggest higher enzymatic efficiency.

PDH activity measurements were available for 65 patients because of incomplete sample collection.

The Hopkins Verbal Learning Test-Revised (HVLT-R) Total Recall score is a measure of verbal learning and immediate memory. Participants are read a list of 12 words (from three semantic categories) and asked to recall as many words as possible across three consecutive learning trials. The Total Recall score is calculated as the sum of correctly recalled words across the three trials, yielding a possible range of 0 to 36, with higher scores indicating better verbal learning and memory performance. This score reflects both initial acquisition and short-term retention of verbal information and is commonly used to assess cognitive function in clinical research.

Data available for 37 patients because of incomplete sample collection.

The Hopkins Verbal Learning Test-Revised (HVLT-R) Delayed Recall score assesses retention of verbal information after a delay. Following completion of the three learning trials, participants are asked to recall the previously presented word list after a delay period (typically 20-25 minutes) without re-exposure to the words. The Delayed Recall score is calculated as the number of correctly recalled words, with a possible range of 0 to 12, where higher scores indicate better memory retention. This measure reflects the ability to consolidate and retrieve learned information over time and is commonly used to evaluate memory function in clinical and research settings.

Data available for 34 patients because of incomplete sample collection.

The Hopkins Verbal Learning Test-Revised (HVLT-R) Retention (%) score reflects the proportion of learned information that is retained over a delay. It is calculated as the ratio of the Delayed Recall score to the highest number of words recalled on any of the three immediate recall trials, multiplied by 100, yielding a percentage value. Scores typically range from 0% to 100%, with higher values indicating better retention of previously learned material. This measure provides an index of memory retention independent of initial learning performance.

Data available for 34 patients because of incomplete sample collection.

The Hopkins Verbal Learning Test-Revised (HVLT-R) Recognition Discrimination Index assesses recognition memory by evaluating the ability to distinguish previously learned words from novel distractors. During the recognition phase, participants are presented with a list of target words and distractor words and asked to identify those that were previously learned. The Recognition Discrimination Index is calculated as the number of true positives (correctly identified target words) minus the number of false positives (incorrectly identified distractor words), yielding a score typically ranging from -12 to 12. Higher scores indicate better recognition accuracy and discrimination ability, reflecting the integrity of memory retrieval processes.

Data available for 37 patients because of incomplete sample collection.

The Brief Visuospatial Memory Test (BVMT) assesses visuospatial learning and memory. Participants are shown six geometric designs for a brief period and asked to reproduce them from memory across three learning trials. Performance is scored based on the accuracy and placement of each design, with a total recall score ranging from 0 to 36, where higher scores indicate better visuospatial memory. A delayed recall trial is administered after a delay to assess retention of visual information. This measure evaluates the ability to encode, store, and retrieve visuospatial material.

Data available for 36 patients because of incomplete sample collection.

The Trail Making Test Part A (TMT-A) assesses visual attention, processing speed, and psychomotor function. Participants are instructed to connect numbered circles in sequential order as quickly as possible. The primary outcome is the time to completion, measured in seconds, with lower times indicating better performance. There is no fixed maximum score, although testing is typically discontinued at a predefined time limit (commonly 300 seconds).

Data available for 36 patients because of incomplete sample collection.

The Trail Making Test Part B (TMT-B) assesses executive function, including cognitive flexibility and set-shifting, in addition to processing speed. Participants are required to alternate between numbers and letters in sequence (e.g., 1-A-2-B) as quickly as possible. The primary outcome is the time to completion, measured in seconds, with lower times indicating better performance. As with TMT-A, there is no fixed maximum score, and testing is typically discontinued at a predefined time limit (commonly 300 seconds).

Data available for 36 patients because of incomplete sample collection.

The Wechsler Adult Intelligence Scale-Fourth Edition (WAIS-IV) Digit Span subtest assesses attention, working memory, and concentration. Participants are asked to repeat sequences of numbers in forward order (Digit Span Forward), reverse order (Digit Span Backward), and ascending order (Digit Span Sequencing). Scores are based on the total number of correctly recalled sequences, with higher scores indicating better attention and working memory capacity. The total score ranges from 0 to 48.

Data available for 37 patients because of incomplete sample collection.

The Animal Naming test is a measure of semantic verbal fluency and executive function. Participants are asked to name as many animals as possible within a fixed time period. The score is the total number of unique, correct animal names generated, with higher scores indicating better semantic memory retrieval and executive functioning. Scores range from 0 with no fixed upper limit.

Data available for 36 patients because of incomplete sample collection.

The Test of Premorbid Functioning (TOPF) estimates an individual's baseline cognitive ability prior to illness or injury. Participants are asked to read aloud a list of irregularly spelled words, and performance is scored based on correct pronunciation. Scores are used to estimate premorbid intellectual functioning. Scores range from 0 to 70, with higher scores indicating higher estimated premorbid functioning. This measure helps contextualize current cognitive performance relative to expected baseline levels.

Data available for 36 patients because of incomplete sample collection.

100mL 0.9% saline twice daily for two days

Placebo: 50mL 0.9% saline