Pediatric diabetic ketoacidosis, fluid therapy, and ... diabetic ketoacidosis, fluid therapy, ... to determine the optimal fluid treatment protocol for pediatric DKA. ... fluid

Pediatric Diabetes 2013: 14: 435–446doi: 10.1111/pedi.12027All rights reserved

© 2013 John Wiley & Sons A/S

Pediatric Diabetes

Original Article

Pediatric diabetic ketoacidosis, fluid therapy,and cerebral injury: the design of a factorialrandomized controlled trial

Glaser NS, Ghetti S, Casper TC, Dean JM, Kuppermann N, for the PediatricEmergency Care Applied Research Network (PECARN) DKA FLUID StudyGroup. Pediatric diabetic ketoacidosis, fluid therapy, and cerebral injury: thedesign of a factorial randomized controlled trial.Pediatric Diabetes 2013: 14: 435–446.

Treatment protocols for pediatric diabetic ketoacidosis (DKA) varyconsiderably among centers in the USA and worldwide. The optimal protocolfor intravenous (IV) fluid administration is an area of particular controversy,mainly in regard to possible associations between rates of IV fluid infusionand the development of cerebral edema (CE), the most common and the mostfeared complication of DKA in children. Theoretical concerns aboutassociations between osmotic fluid shifts and CE have promptedrecommendations for conservative fluid infusion during DKA. However,recent data suggest that cerebral hypoperfusion may play a role in cerebralinjury associated with DKA. Currently, there are no existing data fromprospective clinical trials to determine the optimal fluid treatment protocol forpediatric DKA. The Pediatric Emergency Care Applied Research NetworkFLUID (FLuid therapies Under Investigation in DKA) study is the firstprospective randomized trial to evaluate fluid regimens for pediatric DKA.This 13-center nationwide factorial design study will evaluate the effects ofrehydration rate and fluid sodium content on neurological status during DKAtreatment, the frequency of clinically overt CE and long-term neurocognitiveoutcomes following DKA.

Nicole S Glasera, SimonaGhettib, T Charles Casperc,J Michael Deanc, NathanKuppermanna,d and for thePediatric Emergency CareApplied Research Network(PECARN) DKA FLUID StudyGroup†

aDepartment of Pediatrics, Universityof California Davis, School of Medicine,Davis, CA, USA; bDepartment ofPsychology, University of CaliforniaDavis, Davis, CA, USA; cDepartment ofPediatrics, University of Utah School ofMedicine, Salt Lake City, UT, USA; anddDepartment of Emergency Medicine,University of California Davis, School ofMedicine, Davis, CA, USA†See Appendix for the members of thePECARN DKA FLUID Study Group.

Key words: brain injury – CerebralEdema – DKA

Corresponding author: NathanKuppermann, MD, MPH,Department of Emergency Medicine,University of California Davis,School of Medicine,2315 Stockton Blvd.,Davis, CA 95817,USA.Tel: (916) 734-1535;fax: (916) 734-7950;e-mail: [email protected]

Submitted 23 November 2012.Accepted for publication 29 January2013

The optimal treatment for pediatric diabetic ketoacido-sis (DKA) has been a topic of debate for decades. Mul-tiple working groups and consensus conferences havebeen convened to develop guidelines for pediatric DKA

treatment. These efforts, however, have been hamperedby a lack of high-quality data from randomized con-trolled trials to guide therapeutic recommendations(1–3). Intravenous (IV) fluid regimens for rehydration

435

Glaser et al.

of children with DKA have been the main topic of con-troversy. Consensus statements concerning IV fluidregimens for rehydration of children with DKA haveprovided broad, general guidelines because data areunavailable to support more precise recommendations.A recent informal poll of 20 hospitals participating inthe Pediatric Emergency Care Applied Research Net-work (PECARN) suggests that substantial variabilityin DKA management continues to exist (unpublisheddata), similar to that documented in older publishedliterature (4). According to currently used protocols inthe pediatric referral centers participating in PECARN,a 40-kg child with DKA could receive IV fluid at ratesas high as 215 mL/h or as low as 114 mL/h. Similarly,there is a disagreement about the optimal sodium con-tent of rehydration fluid with some using 0.45% saline,others 0.9% saline, and others using a combination.This substantial treatment variation reflects the lackof evidence to guide management and underscores theneed for a definitive randomized controlled trial.

At the center of the controversy surrounding DKAtreatment in children are physicians’ concerns aboutpossibly causing or exacerbating DKA-related cerebraledema (CE) or cerebral injury with inappropriateIV rehydration. Clinically overt and potentially life-threatening CE occurs in only 0.5–1% of DKAepisodes, making this entity difficult to study. (5,6) However, CE that is asymptomatic or associatedwith only minor mental status disturbances hasbeen documented to occur in most children withDKA (7–10). In addition, while it was previouslyassumed that children who did not develop clinicallyovert CE recovered fully, without lasting neurologicalinjury, recent data suggest that this is not the case.DKA episodes without clinically overt CE have beenassociated with permanent deficits in memory function(11). Evidence to guide clinical care of children withDKA is therefore essential not only for the goal ofdecreasing the rate of clinically overt, life-threateningCE, but also to reduce the incidence of subclinical CEresulting in neurocognitive dysfunction.

Some investigators hypothesized that CE may resultfrom osmotic shifts caused by rapid IV rehydration(12–14). As a consequence, many protocols manageDKA in children with conservative fluid therapy.Although this hypothesis is intuitively appealing,data showing clear associations between aggressivefluid therapy and CE are lacking. Instead, recentdata suggest that cerebral hypoperfusion and theeffects of reperfusion during DKA treatment mayplay a prominent role in the development of cerebralinjury and CE (6, 15–17). Conservative rehydrationprotocols could delay reestablishment of normalcerebral perfusion, and could be detrimental, ratherthan protective. Use of low sodium content fluidsmay exacerbate this problem by decreasing the volume

of fluid retained in the vascular space, while useof isotonic saline may slow repair of intracellulardehydration. Conversely, more rapid infusion offluids might increase vasogenic edema associated withcerebral reperfusion, particularly if breakdown of theblood–brain barrier has occurred from ischemia.

The PECARN Fluid Therapies Under Investigationin DKA (‘FLUID’) study is the first prospectiverandomized controlled clinical trial to investigate theimpact of fluid rehydration regimens on neurologicaland neurocognitive outcomes in children with DKA.The study will determine the effects of rehydration rateand fluid sodium content on neurological status duringDKA treatment, the frequency of clinically overt CE,and long-term neurocognitive outcomes.

Methods

Overview

The PECARN FLUID study is a factorial designrandomized controlled trial comparing four fluidtreatment protocols for children with DKA. Tworates of rehydration will be compared; a more rapidrate, designed to promote faster reperfusion of braintissue and a slower rate, geared toward more gradualreperfusion. Within each of these two rehydration rateschemes, we will compare two sodium concentrations(0.9% saline or 0.45% saline), although all initialfluid boluses will be with 0.9% saline. The studytreatment arms were based on the high and low endsof the range of treatment protocols in current usein PECARN hospitals. We will compare treatmentarms using comprehensive assessments for neurologicalinjury including measurements of subtle neurologicaldysfunction during DKA treatment (in addition torecording the frequency of acute, clinically overt CE)and measures of long-term neurocognitive functionseveral months after hospital discharge.

Inclusion/exclusion criteria

Children meeting the following criteria will beconsidered for enrollment:

(i) age <18-yr-old at enrollment;(ii) diagnosis of DKA (serum glucose or fingerstick

glucose concentration >300 mg/dL, and venouspH < 7.25 or serum bicarbonate concentration<15 mmol/L.

The following patients will be excluded from thestudy:

(i) Patients with underlying neurological disordersthat would affect either mental status testing

436 Pediatric Diabetes 2013: 14: 435–446

Pediatric DKA fluid therapy randomized trial

during DKA treatment or neurocognitive testingafter recovery;

(ii) Patients who present with DKA concomitantwith alcohol or drug use, head trauma,meningitis, or other conditions that affectneurological function;

(iii) Patients with DKA transferred to one of the studysites after receiving an IV fluid bolus of more than10 mL/kg;

(iv) Patients who have begun DKA treatment priorto being approached for enrollment and havereceived (a) more than 2 h of IV fluid infusion atmaintenance rates or higher, OR (b) more than4 h of DKA treatment with insulin and/or fluidsregardless of rate of infusion;

(v) Patients who are pregnant;(vi) Patients for whom the treating physicians feel a

specific fluid or electrolyte regimen is necessarysuch that patient safety or well-being could becompromised by enrollment into the study.

Enrollment and DKA treatment protocols

Upon arrival, patients will begin standard fluid therapyusing an initial IV fluid bolus of 10 cc/kg of 0.9% saline.During this initial therapy, study personnel will obtaininformed consent. Randomization will ideally occurprior to initial bolus completion. If consent cannot beobtained within this time frame, patients will be treatedwith the usual DKA protocol for the study site untilconsent is completed, but not beyond the time frameoutlined in the exclusion criteria.

Children will be randomized to one of four treatmentprotocols (Table 1). In all protocols, initial bolusvolumes will be subtracted from the fluid deficit used tocalculate the rate of fluid replacement. Because studiesshow that clinical estimates of percent dehydrationin children with DKA are inaccurate (18, 19), weassigned an assumed fluid deficit to each protocol(Table 1) based on the upper or lower end of the rangedocumented in recent investigations (18–21).

All four protocols are identical in regard to otheraspects of DKA treatment. Fluid boluses may berepeated at the discretion of the treating physicianto restore peripheral perfusion and hemodynamicstability. Potassium replacement is provided usingan equal mixture of potassium chloride andpotassium phosphate. Insulin treatment begins afterthe initial IV fluid bolus(es) as a continuousIV infusion of 0.1 units/kg/h. When the serumglucose concentration declines below approximately200–300 mg/dL, dextrose is added to the IV fluids tomaintain serum glucose between 100 and 200 mg/dL.

Glasgow Coma Scale (GCS) scores are evaluated atenrollment. For patients presenting with GCS scoresof ≥14, randomization is stratified by clinical center.

A separate, balanced randomization will be used forthose patients presenting with GCS scores <14, as thesepatients will not be included in the primary analysis,and present too infrequently to stratify by clinicalcenter. The PECARN Data Coordinating Centerhas prepared randomization schedules using variable-length permuted blocks to reduce predictability, as thefluid therapy is non-blinded. Assignments are madethrough the use of an interactive telephone service.

Outcomes

The primary study outcome is the occurrence of abnor-mal GCS (GCS < 14) during DKA treatment, a mea-sure previously shown to correlate with DKA-relatedCE measured by magnetic resonance imaging (8, 15).We will evaluate this dichotomous outcome only inchildren who present to the emergency departmentwith normal GCS scores (GCS scores ≥ 14). Subjectswho present with GCS < 14 will not be included in theanalysis of the primary outcome, but will be includedin the analyses of all other study outcomes.

Safety outcomes include the frequency of adverseevents (e.g. thromboses, hyperchloremic acidosis, andother electrolyte imbalances, renal failure, cardiacarrhythmias resulting in hemodynamic abnormalities)and death.

Secondary study outcomes will include:

(i) Forward and backward digit span recall test scoresduring DKA treatment. These tests will be used toevaluate working memory and are assessed every4 h during waking hours;

(ii) Clinically overt CE during DKA treatment.Development of clinically overt CE is defined bysevere mental status change diagnosed clinically asCE and associated with endotracheal intubationor administration of either 3% saline or mannitol;

(iii) Tests of memory 3 months after recovery from DKA.Memory tests will be used to evaluate variousmemory functions including item recognition andrecollection of contextual detail.

Intelligence quotient (IQ) test scores 3 months afterrecovery from DKA will be evaluated as an exploratoryoutcome.

Data collection

Clinical and demographic data to be recorded aresummarized in Table 2. For 3 yr and older children,rapid assessment of working memory (digit span recall)is conducted at enrollment (22–24). Participants areasked to repeat a series of numbers heard aloud, in thesame order as spoken (forward digit span), and thenare asked to do the same backwards. The forward task

Pediatric Diabetes 2013: 14: 435–446 437

Glaser et al.

Table 1. Treatment protocol overview

Protocol A1 Protocol A2 Protocol B1 Protocol B2

Standard initialfluid bolus*

10 cc/kg bolus of 0.9%saline




Additionalintravenousfluid bolus

Additional 10 cc/kg of0.9% saline

Additional 10 cc/kg of0.9% saline

No additional bolus No additional bolus

Assumed fluiddeficit

10% of body weight 10% of body weight 5% of body weight 5% of body weight

Replacementof deficit

Replace half of fluiddeficit + maintenancefluids over initial 12 h,remainingdeficit + maintenancefluids over subsequent24 h

Replace half of fluiddeficit + maintenancefluids over initial 12 h,remainingdeficit + maintenancefluids over subsequent24 h

Replacedeficit + maintenancefluids evenly over 48 h

Replacedeficit + maintenancefluids evenly over 48 h

Fluid used fordeficitreplacement

0.45% Saline 0.9% Saline 0.45% Saline 0.9% Saline

*This is standard treatment at all participating centers and is not part of the study protocol. Consent will occur during thisinitial fluid bolus after which the study treatment will be randomized.

measures the ability to maintain information online,whereas the backward task measures the ability tomentally manipulate information (25). The examinerincreases the number of digits by one unit on eachsuccessive trial as long as the child repeats themcorrectly. The test ends when the child makes amistake in two sequences of the same span in a row.The digit span recall test is repeated every 4 h duringnormal waking hours (7 am to 10 pm) with new digitsequences presented each time. Digit span recall is notassessed during usual sleep hours because the patients’cooperation during these hours is limited.

If any of the hourly GCS scores fall below 14,repeat GCS assessment is performed 15 min laterfor reassessment and/or confirmation. If the repeatGCS assessment confirms a GCS score below 14,the patient is classified as having abnormal mentalstatus. Abnormal GCS scores are reported to theattending physicians per usual hospital protocol,and suspected CE is treated at the discretion of theattending physician. Patients with abnormal mentalstatus resulting from hypoglycemia during DKAtreatment will not be considered to have met the studyoutcome.

Assessments of GCS scores and digit span testingcontinue for 24 h or until resolution of DKA (transitionto subcutaneous insulin administration), whichevercomes first. Data from previous studies show thatnearly all neurological injuries caused by DKA occurwithin the first 24 h of treatment, and the large majoritywithin the first 12 h (6). Therefore extending themonitoring period beyond this time is unlikely to beuseful. Biochemical data and any adverse events arerecorded until hospital discharge.

Additional clinical and biochemical monitoringconforms with the guidelines at each site, whichtypically follow international guidelines for themanagement of DKA in children (Table 2) (1, 3).

For children whose parents/guardians decline to par-ticipate, as well as for those who were eligible but notapproached, demographic, clinical, and biochemicalvariables are recorded to assess for enrolment bias

Post-recovery neurocognitive assessment

Patients 3–17 yr of age will return 3 months ± 4 wkafter recovery from DKA for neurocognitiveassessment. Immediately prior to testing, a fingerstickglucose concentration is measured. Neurocognitivetesting is rescheduled if children have hypoglycemia(glucose < 70 mg/dL) or hyperglycemia with ketosis.

Patients older than 6 yr are evaluated using theWechsler Abbreviated Scale of Intelligence (WASI)(26). The WASI has been standardized nationallyand yields the three traditional IQ scores; verbal,performance, and full scale IQ. In addition, theworking memory (digit span) test is repeated. Memoryis further tested with a color task and a spatial-positiontask that evaluate item recognition and recollectionof contextual detail. These tasks were used to collectdata in preliminary studies (11) and in other previousresearch (27, 28) (Table 3).

Children 3–5 yr of age undergo a modified versionof IQ and memory testing to accommodate theneurocognitive capacities and shorter attention spanof younger children. A short form of the WechslerPreschool and Primary Scale of Intelligence [WPPSI-III (29)] is used for IQ assessment, including four



Table 2. Summary of data collection during DKA

Data to be collected Frequency of measurement

Demographic and historical data• Demographic data (age, gender,

race/ethnicity, parental income, andeducation level)

• Diabetes history for children with knownDM (age at diagnosis of DM, HbA1c overpast year, number of previous DKAepisodes, number of previous episodes ofsevere hypoglycemia)

• Clinical data (medical conditions other thandiabetes, presence/severity of headache,assessment of peripheral perfusion)

• At time of enrollment

Biochemical monitoring• Blood glucose concentration • At presentation and hourly• Electrolytes (Na, K, Cl, HCO3, BUN, and

Cr)• At presentation and approximately every 2–4 h (per usual

site protocols)• Venous pH and pCO2 • At presentation and approximately every 2–4 h (per usual

site protocols)• Serum Ca, Phos, and Mg • At presentation and approximately every 4–8 h (per usual

site protocols)Assessment of neurological function

• Mental status assessment (GCS scores) • At enrollment and hourly• Brief memory assessment (digit span

recall)• At enrollment and every 4 h during normal waking hours

BUN, blood urea nitrogen; DKA, diabetic ketoacidosis; DM, diabetes mellitus; GCS, Glasgow Coma Scale.

Table 3. Summary of neurocognitive testing 3 months afterDKA recovery

Age Testing procedures

Younger than 3 yr No neurocognitive evaluation3–5 yr IQ assessment – abbreviated

WPSSI-RMemory assessment – abbreviated

versions of color task and spatiallocation task

6–18 yr IQ assessment – WASIMemory assessment – color task

and spatial location task

DKA, diabetic ketoacidosis; WASI, Wechsler AbbreviatedScale of Intelligence.

subtests for 3-yr-olds (receptive vocabulary, blockdesign, information, and object assembly) and sevensubtests for 4- and 5-yr-olds (block design, information,matrix reasoning, vocabulary, picture concepts, wordreasoning, and coding) (30, 31). The working memory(digit span) test is repeated, but the color and spatial-position tasks for contextual memory testing areshortened and simplified to accommodate the typicalcapacities of this age group.

Multiple DKA episodes

Some children may present with DKA more than onceduring the study period. To avoid excessively restrictingthe population available, children previously enrolled

into the study who present with another episodeof DKA remain eligible. Short-term neurologicaloutcomes (i.e. in-hospital) for both DKA episodeswill be included in the main analysis. To avoid biasresulting from very frequent enrollment of specificindividuals, children are not enrolled in the study morethan twice.

Non-DKA comparison group

Neurocognitive data for children enrolled in the studywill be compared not only among the four treatmentgroups but also to data from 400 children with type1 diabetes who have never experienced DKA. Thesechildren are recruited from the pediatric diabetes clinicsat the participating centers and tested as previouslydescribed. These analyses will allow us to bettercharacterize the effects of DKA (regardless of DKAtreatment regimen) on neurocognitive function ofchildren with diabetes, and to determine which subsetsof children are at greatest risk for neurocognitivedysfunction. We will control statistically for differencesin variables potentially affecting neurocognitiveoutcomes, including age, gender, duration of diabetes,socioeconomic status, episodes of severe hypoglycemia,and HbA1c.

Statistical analysis

We will analyze the study data according to theintention-to-treat principle. Additionally, we will


Glaser et al.

perform per-protocol or fluid-received analyses foradditional insight. These latter analyses will not replacethe intention-to-treat analysis, and the results of theper-protocol analysis will be examined with caution.

For purposes of sample size calculation, the indicatorof abnormal GCS score (i.e. GCS score < 14) duringDKA treatment was considered as the primaryoutcome. The rationale behind this selection of primaryoutcome was as follows. Severe, clinically overt CEoccurs in less than 1% of DKA episodes and thereforea study large enough to have adequate power to detectdifferences in this outcome would not be feasible. GCSabnormalities, however, are more frequent in childrenwith greater DKA-related CE measured by magneticresonance imaging (15), and increasing evidencesuggests that DKA-related CE presents as a continuumof severity, from asymptomatic to clinically overt (seeDiscussion). It seems reasonable to assume, therefore,that fluid protocols that decrease the frequency of GCSabnormalities are also likely to decrease the likelihoodof clinically overt CE.

Only subjects with baseline GCS scores of 14or 15 will be included in the primary analysis.A Mantel–Haenszel test, stratified by center andcontrolling for the other treatment factor, will beperformed for each of the two factors (rate of fluidadministration and sodium content of fluids). Tocontrol the type I error rate, each factor will be testedat a 0.025 level.

The interaction between fluid rate and sodiumcontent will be tested in an exploratory analysis.Furthermore, two additional analyses of GCS scoreswill be used: drop in GCS score (difference betweenbaseline and lowest GCS score) and the durationof time for which the GCS score is below 14. Thesetwo GCS outcomes will be analyzed using a VanElteren test, controlling for strata. We will performexploratory analyses of the GCS score outcomes byassessing the treatment effect after adjustment forcovariates. The frequency of clinically overt CE willbe tested using the Mantel–Haenszel test, as describedfor the primary analysis.

For digit span scores, we will apply longitudinaldata analysis methods by assuming a linear mixed-effects model. Time zero will be randomization time.The time–treatment interactions are the quantities ofinterest, as they represent the change over time dueto each treatment. We will additionally evaluate otherpossible exploratory models, including interactions andnon-linear relationships.

The main memory task analysis will focus on theaverage of the item color association rate and the itemspace association rate. These scores will be comparedusing a Van Elteren test, controlling for strata. Wewill investigate the effects of treatment after adjustingfor important covariates in a linear model. The main

tests for the three outcomes of forward digit span,backward digit span, and memory score will be subjectto Holm’s procedure for multiple comparisons. Thesame analytical approach proposed for memory willbe used for the analyses of IQ measures. General linearmodels will be used to adjust for known age-relateddifferences in measurement error and variance inIQ (26, 32).

We also plan to determine whether treatmenteffects are consistent across prospectively definedsubgroups. Variables used to define subgroups willinclude age (younger than 6 yr vs. 6 yr and older),baseline GCS scores (14 vs. 15) and history of previousDKA (possibly resulting in pre-existing neurocognitivealterations). For analyses that include patients withbaseline GCS scores <14, we will analyze the subgroupsof baseline GCS <14 vs. baseline GCS of 14 or 15. Thesignificance level for all subgroup-based tests will beadjusted in order to keep the overall type I error rateless than 0.05. Results of subgroup analyses will beused primarily to confirm a consistent magnitude oftreatment effect.

Power analysis

The power of the primary analysis (frequency ofabnormal GCS scores during DKA treatment) dependson the proportion of patients with GCS scores decliningbelow 14 in each group. Previous data show thatmental status abnormalities (GCS scores < 14) occur inapproximately 15% of children treated for DKA, andare associated with evidence of CE on neuroimaging(8, 15). We assume that the treatment group withthe highest rate of developing abnormal GCS scoreswould have an approximately 20% overall frequencyof GCS scores below 14 and we desire to detect anabsolute beneficial treatment effect of 7.5 with 90%power. Using a two-sided type I error rate of 0.025 andthe hypothesized proportions yields a required totalsample size of approximately 1200 patients. Allowingfor non-adherence to assigned treatment of up to 5%raises the required number to 1200/0.95 (2), or about1330. In order to adjust for O’Brien–Fleming interimmonitoring, a 2% increase will be made, bringing thesample size to approximately 1360. This represents thenumber of patients that present with GCS scores ofat least 14. We estimate that approximately 10% ofeligible patients will present with GCS scores less than14. This means that in the period required to enroll 1360patients presenting with normal GCS scores, about 150with abnormal scores will be enrolled. Thus, the totalplanned enrollment is 1510.

Data and safety monitoring board

This study will be monitored by a Data andSafety Monitoring Board (DSMB) approved by the



funding agency (NICHD). The FLUID DSMB iscomposed of five doctorate-level members not affiliatedwith the study, including experts in the fields ofemergency medicine, pediatric critical care, pediatricendocrinology, neuropsychology, and biostatistics.The DSMB will meet yearly and the monitoring planfor this study will include three interim analyses afterapproximately one fourth, one half, and three fourthsof the target sample size has been achieved. The DSMBwill make recommendations regarding study conduct,data quality, participant safety, and continuing orstopping the trial.

Discussion

DKA-related cerebral injury is the major cause ofmortality and morbidity in children with type 1diabetes (33–35). DKA occurs frequently in childrenand is often present at the time of diagnosis ofdiabetes (36). DKA may also occur in children withestablished diabetes during illness, diabetes equipmentmalfunction, or diabetes mismanagement. Clinicallyovert CE occurs in 0.5–1% of DKA episodes (5, 6)and has a high mortality rate (21–24%). Survivorsfrequently are left with permanent neurologicaldeficits (5, 6). Furthermore, children without overtneurological symptoms during DKA treatment maynonetheless have subtle evidence of brain injury afterrecovery from DKA, particularly memory deficits (11).Some have postulated that excessive rates of fluidadministration are responsible for DKA-related CE(13, 37, 38). Recent investigations, however, suggestthat dehydration and cerebral hypoperfusion may beassociated with DKA-related cerebral injury (6, 7, 16,17). The association between fluid therapy and DKA-related cerebral injury has never been investigated ina randomized controlled trial. Therefore, this studyis timely and necessary. The PECARN FLUID studywill test the associations of fluid therapy not only withacute neurological events during DKA treatment butalso with long-term neurocognitive outcomes.

Clinically overt CE, resulting in marked neurolog-ical depression, is infrequent, however, more subtleCE occurs with much greater frequency and is presentin most children with DKA (8–10). Studies using CTor MR imaging in children with DKA and in animalmodels of DKA have shown that CE may be presentbefore treatment as well as during therapy (8–10, 39).Children with abnormal mental status during DKAtreatment (assessed by GCS scores) are more likely tohave subtle CE (measured as narrowing of the cerebralventricles) than those with normal mental statusthroughout treatment (8). Abnormal GCS scoresduring DKA treatment have also been associated withgreater magnetic resonance diffusion-weighted imag-ing changes indicative of CE (15). These data suggest

that CE is not a rare phenomenon in children withDKA, but rather that CE occurs frequently with vary-ing severity. Severe, clinically overt CE that occurs in0.5–1% of children with DKA likely represents only themost extreme presentation of a common phenomenon.

Data suggest that even subtle DKA-related CE maynot be without long-term consequences. One importantstudy demonstrated that children with diabetes whohad experienced DKA performed more poorly in testsof memory capacity compared to children with diabeteswho had never had DKA (11). Most children inthat study had experienced only one DKA episode,generally at the time of diagnosis of diabetes, andthe DKA group was nearly identical to the non-DKAgroup in regard to duration of diabetes, measuresof glycemic control, and frequency of hypoglycemicepisodes. Exposure to DKA has similarly been shownto decrease maze learning ability in a rodent diabetesmodel (40). These data suggest that complications ofDKA may extend beyond the acute period and mayaffect neurocognitive function and quality of life forsubstantial numbers of children with type 1 diabetes.

Because children with DKA frequently present withnormal or only modestly altered mental status butmay subsequently experience a decline in mental statusduring treatment, investigators questioned whethersome aspects of treatment might be responsible (12, 13,37, 38, 41–43). For decades, it has been hypothesizedthat rapid fluid infusion with rapid changes in serumosmolality might lead to brain cell swelling. (12,13, 37, 38) Although this hypothesis is intuitivelysimple and attractive, data to support this hypothesishave been lacking. Properly controlled retrospectivestudies have not detected associations between osmoticchanges during treatment and risk of CE (6, 44, 45).Furthermore, although neurological decline associatedwith CE often occurs during DKA treatment, severalreports document the occurrence of symptomatic CE inchildren with DKA prior to the initiation of therapy. (6,46, 47) These data suggest that CE may not be primarilycaused by osmotic shifts induced by therapeuticinterventions, but instead that factors intrinsic toDKA may cause brain injury and this injury could beworsened during treatment. We and other investigatorshave hypothesized that DKA-related CE may becaused by cerebral hypoperfusion during DKA andthe subsequent effects of reperfusion (6, 7, 45, 48, 49).

Several additional lines of evidence suggestthat DKA-related CE may not be caused byosmotic fluctuations. One important descriptive studydemonstrated that many children with clinicallydiagnosed DKA-related ‘cerebral edema’ (49) haveno evidence of edema on cerebral imaging studiesdone at the time of neurological decompensation,despite a profound degree of neurological depression.Repeated imaging studies performed hours or days


Glaser et al.

later demonstrated evidence of CE, along withhemorrhage or cerebral infarction in some cases. Thesedata suggest that CE may possibly be a consequenceof cerebral injury during DKA, rather than the causeof injury, similar to hypoxic/ischemic brain injurieswhere edema typically develops after initial injury, as aconsequence of cellular energy failure and blood–brainbarrier dysfunction.

Data from both animal and human studies of DKAsupport the hypothesis that cerebral hypoperfusionand reperfusion may be involved in DKA-relatedbrain injury. Cerebral blood flow (CBF) is low duringuntreated DKA and rises to levels above normalduring treatment with insulin and saline (7, 17).Brain cell swelling (cytotoxic edema) is present duringuntreated DKA (39). During treatment with insulinand saline, vasogenic edema develops as CBF rises (7).Furthermore, brain levels of high-energy phosphatesare low during untreated DKA and cerebral lactatelevels are elevated, suggesting inadequate CBF. (16)Early in the course of treatment with insulin and saline,brain high-energy phosphate levels decline further.Brain ratios of n-acetylaspartate (NAA) to creatine,an indicator of neuronal health, are also low duringuntreated DKA and initially decline during DKAtreatment (16). These data suggest that DKA per semay have adverse effects on CBF and metabolism, butmore importantly, these data also suggest that furtherbrain injury may occur initially during treatment withinsulin and saline. These data therefore underscorethe importance of determining whether variationsin DKA treatment, particularly fluid infusionprotocols, might affect the likelihood or severity ofbrain injury.

Whether rates of fluid infusion during DKA treat-ment in children influence the risk of clinically overt CEhas been a topic of much controversy. Several investiga-tors have attempted to evaluate the effects of differencesin fluid treatment through retrospective studies (6, 12,14, 44, 50, 51). The results of these studies have beenvariable, and consistent associations between rates offluid infusion and risk of CE have not been found. Oneretrospective observational study documented similarrates of CE before and after an institutional DKAprotocol change from more rapid rehydration with0.9% saline to slower rehydration with 0.45% saline.(52) All of these retrospective studies, however, are bydefinition hampered by the lack of randomization oftreatments with inherent tendency toward bias causedby physicians’ treatment decisions related to patientappearance and severity of presentation.

In a large retrospective study by our group, childrenwith DKA-related CE were compared to controls withuncomplicated DKA, using multivariable analyses tocontrol for variables related to DKA severity. In thisstudy, we did not detect an association between fluid

infusion rates and risk of clinically apparent CE (6).Although the analysis controlled for DKA severity tothe extent possible in a retrospective study, treatmentdecisions may nonetheless have been biased by unmea-sured factors related to patient appearance and clinicalpresentation. In addition, the study lacked the statisti-cal power to make definitive conclusions regarding theeffect of fluid infusion rates or sodium content.

Some studies have demonstrated that failure ofthe measured serum sodium concentration to rise asthe glucose concentration decreases during therapyis associated with CE (6, 14, 53, 54). On the basisof these data, one study evaluated a DKA protocoldesigned to avoid a decrease in sodium concentrationby administering fluids at slow rates and utilizingfluids with relatively high concentrations of sodium(55). Using this protocol, the investigators found thatthe serum sodium concentration rose as the glucoseconcentration fell in 90% of DKA episodes. No deathsor apparent permanent neurological injuries occurred;however, six patients (2.6%) required administrationof mannitol because of alterations in mental status,likely indicative of CE. In addition, all previous studiesof fluid treatment protocols in pediatric DKA havebeen hampered by retrospective data collection and/orlack of appropriate control or comparison groups. Itis therefore imperative that a prospective randomizedstudy be conducted to answer this question.

Taking into account recent data from cerebralimaging studies in children with DKA and in animalmodels of DKA, the optimal fluid treatment protocolis not obvious. If cerebral hypoperfusion during DKAplays a role in DKA-related cerebral injury and edemaformation, arguments in favor, or against all of thetreatment arms in the PECARN FLUID study couldbe made. Conservative (slower) fluid resuscitationmight serve to prolong the state of cerebralhypoperfusion, resulting in increased risk of cerebralinjury. Furthermore, intravascular volume may declineduring therapy as intravascular osmolality decreases.If fluid resuscitation is inadequate during this time,cerebral hypoperfusion may be worsened. Use of lowsodium content fluids could exacerbate this problem.Conversely, it could be argued that more conservativefluid therapy might help to decrease vasogenic CElater in the course of DKA treatment, particularly ifbreakdown of the blood–brain barrier occurs.

At present, the impact of fluid resuscitation proto-cols on DKA-related brain injury in children remainsunknown. The design of the PECARN FLUID studywill allow us to determine the effect not only of fluidinfusion rate but also the sodium content of fluids.This study will be the first prospective clinical trial toprovide evidence for definitive recommendations inregard to fluid rehydration therapy in the setting ofpediatric DKA.



Acknowledgements

This study was supported by grant 1R01HD062417-01 fromthe Eunice Kennedy Shriver National Institute of ChildHealth & Human Development. The Pediatric EmergencyCare Applied Research Network (PECARN) is supportedby cooperative agreements U03MC00001, U03MC00003,U03MC00006, U03MC00007, U03MC00008, U03MC22684,and U03MC22685 from the Emergency Medical Services forChildren (EMSC) program of the Maternal and Child HealthBureau, Health Resources and Services Administration, USDepartment of Health, and Human Services. We are grateful forthe support and advice of Dr. Carol Nicholson, MD, MS of theEunice Kennedy Shriver National Institute of Child Health &Human Development, Bethesda, MA, USA.

Conflict of interest

None of the authors have any financial arrangementsthat represent conflicts of interest related to the study.

References

1. Sperling M, Dunger D, Acerini C et al. ESPE/LWPESconsensus statement on diabetic ketoacidosis in childrenand adolescents. Pediatrics 2003: 113: e133–e140.

2. Wolfsdorf J, Craig M, Daneman D et al. ISPADclinical practice consensus guidleines 2009 compendium:diabetic ketoacidosis in children and adolescents withdiabetes. Pediatr Diabetes 2009: 10: 118–133.

3. Wolfsdorf J, Glaser N, Sperling M. Diabeticketoacidosis in infants, children and adolescents: aconsensus statement from the American DiabetesAssociation. Diabetes Care 2006: 29: 1150–1159.

4. Glaser N, Kuppermann N, Yee C, Schwartz D,Styne D. Variation in the management of pediatricdiabetic ketoacidosis by specialty training. Arch PediatrAdolesc Med 1997: 151: 1125–1132.

5. Edge J, Hawkins M, Winter D, Dunger D. The riskand outcome of cerebral oedema developing duringdiabetic ketoacidosis. Arch Dis Child 2001: 85: 16–22.

6. Glaser N, Barnett P, McCaslin I et al. Riskfactors for cerebral edema in children with diabeticketoacidosis. N Engl J Med 2001: 344: 264–269.

7. Glaser N, Gorges S, Marcin J et al. Mechanism ofcerebral edema in children with diabetic ketoacidosis. JPediatr 2004: 145: 164–171.

8. Glaser N, Wooton-Gorges S, Buonocore M et al.Frequency of sub-clinical cerebral edema in childrenwith diabetic ketoacidosis. Pediatr Diabetes 2006: 7:75–80.

9. Hoffman W, Steinhart C, El Gammal T, Steele S,Cuadrado A, Morse P. Cranial CT in children andadolescents with diabetic ketoacidosis. AJNR 1988: 9:733–739.

10. Krane E, Rockoff M, Wallman J, Wolfsdorf J.Subclinical brain swelling in children during treatmentof diabetic ketoacidosis. N Engl J Med 1985: 312:1147–1151.

11. Ghetti S, Lee J, Holtpatrick C, DeMaster D, Glaser

N. Diabetic ketoacidosis and memory impariment inchildren with type 1 diabetes. J Pediatr 2010: 156:109–114.

12. Duck S, Wyatt D. Factors associated with brainherniation in the treatment of diabetic ketoacidosis.J Pediatr 1988: 113: 10–14.

13. Harris G, Fiordalisi I, Finberg L. Safe managementof diabetic ketoacidemia. J Pediatr 1988: 113: 65–67.

14. Harris G, Fiordalisi I, Harris W, Mosovich

L, Finberg L. Minimizing the risk of brainherniation during treatment of diabetic ketoacidemia: aretrospective and prospective study. J Pediatr 1990: 117:22–31.

15. Glaser N, Marcin J, Wooton-Gorges S et al.Correlation of clinical and biochemical findings withdka-related cerebral edema in children using magneticresonance diffusion weighted imaging. J Pediatr 2008:153: 541–546.

16. Glaser N, Yuen N, Anderson S, Tancredi D,O’Donnell M. Cerebral metabolic alterations in ratswith diabetic ketoacidosis: effects of treatment withinsulin and intravenous fluids and effects of bumetanide.Diabetes 2010: 59: 702–709.

17. Yuen N, Anderson S, Glaser N, O’Donnell M.Cerebral blood flow and cerebral edema in rats withdiabetic ketoacidosis. Diabetes 2008: 57: 2588–2594.

18. Koves I, Neutze J, Donath S et al. The accuracyof clinical assessment of dehydration druing diabeticketoacidosis in childhood. Diabetes Care 2004: 27:2485–2487.

19. Ugale J, Mata A, Meert K, Samaik A. Measureddegree of dehydration in children and adolescents withtype 1 diabetic ketoacidosis. Pediatr Crit Care Med2012: 13: e103–e107.

20. Smith L, Rotta A. Accuracy of clinical estimatesof dehydration in pediatric patients with diabeticketoacidosis. Pediatr Emerg Care 2002: 18: 395–396.

21. Sottosanti M, Morrison G, Singh R et al.Dehydration in children with diabetic ketoacidosis: aprospective study. Arch Dis Child 2012: 97: 96–100.

22. Case R. Capacity-based explanations of workingmemory growth: a brief history and reevaluation.In: Schneider F, ed. Memory Performance andCompetencies - Issues in Growth and Development.Mahwah, NJ: Erlbaum & Associates, 1995: 23–44.

23. Case R, Kurland D, Goldberg J. Operationalefficiency and the growth of short-term memory span. JExp Child Psychol 1982: 33: 386–404.

24. Dempster F. Memory span: sources of individualand developmental differences. Psychol Bull 1981: 89:63–100.

25. Baddeley A. Working Memory. Oxford: ClarendonPress, 1986.

26. Wechsler D. Wechsler abbreviated scale of intelligence.San Antonio, TX: Psychological Corporation, 1999.

27. Ghetti S, Angelini L. The development of recollectionand familiarity in childhood and adolescence: evidencefrom dual process signal detection. Child Dev 2008: 79:339–358.

28. Ghetti S, DeMaster D, Yonelinas A, Bunge S.Developmental differences in medial temporal lobefunction during memory encoding. J Neurosci 2010:30: 9548–9556.

29. Wechsler D. Wechsler Preschool and Primary Scaleof Intelligence. revised edn. San Antonio, TX:Psychological Corporation, 1989.


Glaser et al.

30. Kaufman A. Comparison of the performance ofmatched groups of black children and white childrenon the Wechsler Preschool and Primary Scale ofIntelligence. J Consult Clin Psychol 1973: 41: 186–191.

31. Tsushima W. Short form of the WPPSI and WPPSI-R.J Clin Psychol 1994: 50: 877–880.

32. Ferrer E, McArdle J, Shaywitz B, Holahan J,Marchione K, Shaywitz S. Longitudinal models ofdevelopmental dynamics between reading and cognitionfrom childhood to adolescence. Dev Psychol 2007: 43:1460–1473.

33. Daneman D. Diabetes-related mortality. Diabetes Care2001: 24: 801–802.

34. Edge J, Ford-Adams M, Dunger D. Causes of death inchildren with insulin-dependent diabetes 1990–96. ArchDis Child 1999: 81: 318–323.

35. Scibilia J, Finegold D, Dorman J, Becker D, Drash

A. Why do children with diabetes die? Acta EndocrinolSuppl 1986: 279: 326–333.

36. Usher-Smith J, Thompson M, Ercole A, Walter F.Variation between countries in the frequency of diabeticketoacidosis at first presentation of type 1 diabetes inchildren: a systematic review. Diabetologia 2012[Epubahead of print].

37. Arieff A, Kleeman C. Cerebral edema in diabeticcomas. II. Effects of hyperosmolality, hyperglycemiaand insulin in diabetic rabbits. J Clin Endocrinol Metab1974: 38: 1057–1067.

38. Prockop L. Hyperglycemia, polyol accumulation, andincreased intracranial pressure. Arch Neurol 1971: 25:126–140.

39. Lam T, Anderson S, Glaser N, O’Donnell M.Bumetanide reduces cerebral edema formation in ratswith diabetic ketoacidosis. Diabetes 2005: 54: 510–516.

40. Glaser N, Anderson S, Leong W, Tancredi D,O’Donnell M. Cognitive dysfunction associated withdiabetic ketoacidosis in rats. Neurosci Lett 2012: 510:110–114.

41. Edge J. Cerebral oedema during treatment of diabeticketoacidosis: are we any nearer finding a cause? DiabetesMetab Res Rev 2000: 16: 316–324.

42. Hammond P, Wallis S. Cerebral oedema in diabeticketoacidosis; still puzzling – and often fatal. BMJ 1992:305: 203–204.

43. Silver S, Clark E, Schroeder B, Sterns R.Pathogenesis of cerebral edema after treatment ofdiabetic ketoacidosis. Kidney Int 1997: 51: 1237–1244.

44. Lawrence S, Cummings E, Gaboury I, Daneman D.Population-based study of incidence and risk factorsfor cerebral edema in pediatric diabetic ketoacidosis. JPediatr 2005: 146: 688–692.

45. Mahoney C, Vlcek B, Del Aguila M. Riskfactors for developing brain herniation during diabeticketoacidosis. Pediatr Neurol 1999: 21: 721–727.

46. Couch R, Acott P, Wong G. Early onset of fatalcerebral edema in diabetic ketoacidosis. Diabetes Care1991: 14: 78–79.

47. Glasgow A. Devastating cerebral edema in diabeticketoacidosis before therapy. Diabetes Care 1991: 14:77–78.

48. Muir A. Cerebral edema in diabetic ketoacidosis: a lookbeyond rehydration. J Clin Endocrinol Metab 2000: 85:509–513.

49. Muir A, Rosenbloom A, Williams L, Yang M,Quisling R. Cerebral edema in childhood diabeticketoacidosis: natural history, radiographic findings andearly identification. Diabetes Care 2004: 27: 1541–1546.

50. Edge J, Jakes R, Roy Y et al. The UK case–controlstudy of cerebral oedema complicating diabeticketoacidosis in children. Diabetologia 2006: 49:2002–2009.

51. Rosenbloom A. Intracerebral crises during treatment ofdiabetic ketoacidosis. Diabetes Care 1990: 13: 22–33.

52. Mel J, Werther G. Incidence and outcome of diabeticcerebral oedema in childhood: are there predictors? JPaediatr Child Health 1995: 31: 17–20.

53. Bello F, Sotos J. Cerebral oedema in diabeticketoacidosis in children. Lancet 1990: 336: 64.

54. Hale P, Rezvani I, Braunstein A, Lipman T,Martinez N, Garibaldi L. Factors predicting cerebraledema in young children with diabetic ketoacidosis andnew onset type I diabetes. Acta Paediatr 1997: 86:626–631.

55. Harris G, Fiordalisi I. Physiologic management ofdiabetic ketoacidemia: a 5-year prospective pediatricexperience in 231 episodes. Arch Pediatr Adolesc Med1994: 148: 1046–1052.

56. The Pediatric Emergency Care Applied ResearchNetwork. The Pediatric Emergency Care AppliedResearch Network (PECARN): rationale, development,and first steps. Acad Emerg Med 2003: 10: 661–668.

57. Dayan P, Chamberlain J, Dean J, Maio R,Kuppermann N. The Pediatric Emergency CareApplied Research Network: progress and update. ClinPediatr Emerg Med 2006: 7: 128–135.

Appendix 1

Participating co-investigators of the PECARN DKAFLUID Study Group at the time of the designand initiation of the study include the following(institutions in alphabetical order):

Children’s Hospital, Boston, MALise Nigrovic, MD, MPHJoseph Wolfsdorf, MDMichael Agus, MD

Children’s Hospital ColoradoArleta Rewers, MD, PhDMarian Rewers, MD, PhDPeter Mourani, MD

Children’s Hospital of New York/PresbyterianMaria Kwok, MD, MPHMary Pat Gallagher, MDJohn Scott Baird, MD

Children’s Hospital of Philadelphia, Philadelphia, PAMonika Goyal, MDSteven Willi, MDVijay Srinivasan, MD

Children’s National Medical Center, Washington, DC



Kathleen Brown, MDFran R. Cogen, MD, CDESonali Basu, MD

Children’s Memorial Hospital, Chicago, ILJennifer Trainor, MDDonald Zimmerman, MDDenise Goodman, MD, MS

Nationwide Children’s Hospital, Columbus, OHBema Bonsu, MDJustin Indyk, MD, PhDTensing Maa, MD

Primary Children’s Medical Center, Salt Lake Center,UTJeffrey Schunk, MDMary Murray, MDJared Henricksen, MD

UC Davis Medical Center, Sacramento, CANathan Kuppermann, MD, MPHLeah Tzimenatos, MDNicole Glaser, MDJames Marcin, MD, MPH

Children’s Hospital of St. Louis, St. Louis, MOKimberly Quayle, MDNeil H. White, MDNikoleta S. Kolovos, MD

We acknowledge the efforts of the followingindividuals participating in PECARN at the time thisstudy was initiated.

PECARN Steering Committee: N. Kuppermann,Chair; E. Alpern, D. Borgialli, K. Brown, J.Chamberlain, L. Cimpello, P. Dayan, J. M. Dean,M. Gorelick, J. Hoyle, D. Jaffe, M. Kwok, R.Lichenstein, K. Lillis, P. Mahajan, D. Monroe, L.Nigrovic, E. Powell, R. Ruddy, R. Stanley, D. M.Tunik. MCHB/EMSC liaisons: T. Weik, E. Edgerton.

Central Data Management and CoordinatingCenter: J. M. Dean, C. Olsen, J. Yearly, R. Enriquez,A. Davis, R. Kelly, S.J. Zuspan, C. Casper, M. Fjelstad

Protocol Review and Development Subcommittee:D. Jaffe, Chair; J. Chamberlain, P. Dayan, J. M.Dean, R. Holubkov, L. Nigrovic, E. Powell, K. Shaw,R. Stanley, M. Tunik

Quality Assurance Subcommittee: K. Lillis, Chair;E. Alessandrini, R. Enriquez, R. Lichenstein, P.Mahajan, R. McDuffie, G. O’Gara, R. Ruddy, B.Thomas, J. Wade

Safety and Regulatory Subcommittee: W. Schalick,J. Hoyle, Co-Chairs; S. Atabaki, K. Call, H. Hibler,M. Kwok, A. Rogers, D. Schnadower, L. Tzimenatos

Feasibility and Budget Subcommittee: K. Brown, S.Goldfarb, Co-Chairs; M. Berlyant, B. Bonsu, E. Crain,M. Hounchell, D. Monroe, D. Nelson, S.J. Zuspan

Grant Writing and Publication Subcommittee: M.Gorelick, Chair; L. Alpern, J. Anders, D. Borgialli,L. Cimpello, A. Donaldson, G. Foltin, F. Moler, K.Shreve

We would like to thank Marci Fjelstad at thePECARN Data Coordinating Center for her dedicatedand diligent work, the Research Coordinators inPECARN, without whose dedication and hard workthis study would not have been possible, and all theclinicians around the PECARN Network who areenrolling children into this study.

Appendix 2

Structure of PECARN and the research teamand details of study data management

This study is being conducted through the PECARN(56, 57) and currently involves 13 individual hospitalsites. The research team is organized with carefulattention to specification of roles and clear lines ofresponsibility.

The fluid study PI’s coordinate and oversees allaspects of the study, along with the study ProjectAdministrator and the Project Data CoordinatorCenter (DCC). The participating PECARN EDs aredivided into six research nodes. Each node in PECARNsupervises three PECARN EDs. Each study site hasa research coordinator who is responsible for theoverall study organization at that site, coordinationof a patient enrollment plan at that site, day-to-daydata collection at the respective site and transmissionof data to the DCC, overseen by the study site PI.

At each of the participating PECARN centers, thephysician site co-investigator (‘Site PI,’ an emergencyphysician) is responsible for the overall conductof the study at that site. These investigators workclosely with their Divisions/Departments of PediatricEndocrinology and Pediatric Critical Care Medicine.Once received at the DCC, study data will bemanaged by DCC data mangers and analyzed by DCCstatisticians.

Research team committees

The study Steering Committee, consisting of thePIs, the 12 other Site PIs, the Project Managerand Data Manager, and the PI of the DCC,oversee the performance of the study, and resolveand adjudicate study problems as they arise. ThePublication Committee, which includes the PIs,and all site PIs, is responsible for reviewing andapproving proposed study manuscripts, and approvingauthorship agreements. This committee works closely


Glaser et al.

with the PECARN Grant Writing and PublicationSubcommittee, which serves in a similar capacity forall of PECARN.

Project Manager and Data Manager

The Project Manager and Data Manager at the DCCwork with the PIs on the coordination and oversight ofthis study. The Project Manager helps the PIs organizeresearch study meetings and communications, arrangetravel accommodations, maintain study documents inthe virtual electronic website, and answer questionsregarding budget and subcontracts. S/he assists theData Manager based on the DCC. The ProjectManager organizes and monitors inter- and intra-nodal research activities including institutional reviewboard submissions, electronic transmission of data tothe DCC, and report generation. The Project Manageralso assists with the preparation of documentation suchas study progress reports, and interacts closely with theData Manager, both based on the DCC.

The primary responsibility of the Data Manager atthe DCC, supervised by the PI of the DCC, is to overseethe timely transmission of high-quality data as well asits subsequent management. The Data Manager helpsto design the electronic database along with other DCCpersonnel. In cooperation with the study PIs and theProject Manager, the Data Manager coordinates andprovides the instruction on the Manual of Operationsand oversees multiple data entry quality assurance.

Study communications and training plan

The communications infrastructure of PECARNis well-established. Before the initiation of datacollection, the members of the study SteeringCommittee and all site research coordinators metfor a 2-d conference to receive instruction in thestudy manual of operations. At that time, researchcoordinators and site PIs also received small groupinstruction on the conduct of the neurocognitive testingby Dr. Ghetti and several graduate level assistants.Each research coordinator conducted a minimumof four observed practice sessions including all ofthe neurocognitive testing procedures. Each sessionwas followed by feedback and additional practice.

Equivalent training is provided when new sites andresearch coordinators join the study. The SteeringCommittee and all site PIs attend two PECARNmeetings annually throughout the course of this studyto discuss study issues and instruct site PIs. Eachsite PI instructs the group of ED physicians at theirhome institutions about the study, and serves as alocal advocate for the study, and coordinates thecollaboration and communications between membersof the Divisions of Pediatric Emergency Medicine,Pediatric Critical Care, and Pediatric Endocrinology.Throughout the study, the Steering Committee hasmonthly telephone conference calls, or more frequentconference calls as necessary.

Data management and transmission to the DataCoordinating Center (DCC)

Study data will be entered at each site onto a secure,encrypted, password-protected virtual private networkconnecting to a secure, password-protected databaseat a computer server. Site research coordinators willperform double entry to insure accuracy. The onlinedata entry screens will contain range and logic checksto minimize data entry errors. The Data Manager willmonitor data accuracy, contact sites with repeated dataentry errors, and identify ways to resolve these errors.

Study monitoring

All PECARN studies undergo site monitoring,coordinated by the DCC in conjunction with the studyPIs. The site monitoring plan is designed to identifyproblems with sites and methods for handling problemsthat arise. Each site was visited by a PECARN monitorduring the first year of the study. Additional in-personand/or remote monitoring visits are occurring annually.

In addition to the site monitoring visits, neurocog-nitive testing procedures are periodically re-evaluatedto insure that each site continues to conduct theseprocedures according to standardized methods. Eachresearch coordinator responsible for neurocognitivetesting is videotaped conducting the tests and the videois reviewed by the collaborating neuropsychologist.Any deviations from the correct testing procedures arereviewed with the site research assistants and corrected.


Documents

Pediatric diabetic ketoacidosis, fluid therapy, and ... diabetic ketoacidosis, fluid therapy, ... to determine the optimal fluid treatment protocol for pediatric DKA. ... fluid