RESEARCH

Cu

97440

T ycS ipP ilPAT IZABE

ABOb(TPcliDesmepaberecSubdestehoMammduStaranstuResclitransplant. When standardized for time, TPN recipientsat both institutions experienced significantly more hyper-glyexppofer

andifConmidupathejecJ A

Htwfecpoinfoblevposerved in the growth hormone studies conducted in thecritically ill population due to higher glucose levelsamong growth hormone recipients (8,9). Moreover, there

P.istiosiscol

ARDHuRo

CAs

0d

135cemia (P.05) after TPN initiation. TPN patients alsoerienced 69% of all infections and 100% of repeat

sitive cultures. Additionally, significantly greater dif-ences for TPN recipients were found for length of stay

is evidence that controlling hyperglycemia during illnessis associated with improved outcomes (10,11) and tightlycontrolled blood sugar levels have been demonstrated tolead to profound decreases in morbidity and mortality inthe critically ill (12).

Hyperglycemia is considered a major side effect of totalparenteral nutrition (TPN) (13-18). Although TPN hasbeen lifesaving in patients with permanent gastrointes-tinal failure, the use of TPN in most hospitalized patientsis appropriately limited based on the increased risk ofinfectious complications and lack of clear benefit in pa-tients with temporary gut failure. The increased risk ofinfection has previously been hypothesized to result froma loss of gastrointestinal barrier (19-21), a hypothesisthat has been clearly demonstrated in animals but re-mains controversial in humans (22). It is reasonable tosuspect that hyperglycemia associated with TPN plays animportant role in the increased infection rates that have

M. Sheean is a PhD candidate and C. Braunschweigan associate professor, Department of Human Nutri-n, University of Illinois at Chicago. E. Rich is an as-tant professor of medicine, section of Hematology/On-ogy, University of Chicago Hospitals, Chicago, IL.ddress correspondence to: Patricia M. Sheean, MS,, University of Illinois at Chicago, Department ofman Nutrition, M/C 517 1919 West Taylor Street,om 650, Chicago, IL 60612. E-mail: psheea1@uic.eduopyright 2004 by the American Dietetic

sociation.002-8223/04/10409-0007$30.00/0oi: 10.1016/j.jada.2004.06.024

2 Journal of THE AMERICAN DIETETIC ASSOCIATION 2004 by the American Dietetic Associationrrent Research

Continuing Education Questionnaire, page 14Meets Learning Need Codes 5000, 5360, and 5

he Incidence of Hypergltem Cell Transplant Recarenteral Nutrition: A PRICIA M. SHEEAN, MS, RD; CAROL BRAUNSCHWEIG, PhD, RD; EL

STRACTjective To determine whether total parenteral nutrition

N)induced hyperglycemia is associated with adversenical outcomes.ign A retrospective cohort investigation comparing thedical records of hematopoietic stem cell transplanttients was conducted to determine clinical differencestween those who received TPN and those who did noteive TPN during transplant.jects/Setting Forty-eight adult patients (18 years) un-

rgoing initial autologous or allogeneic hematopoieticm cell transplant at two urban university-affiliatedspitals were eligible for inclusion.in Outcome Measures Hyperglycemia (glucose 6.1

ol/L or 110 mg/dL), presence of infection, infectionration, and in-hospital mortality.tistical Analyses Performed 2, Student t, and Wilcoxonk-sum tests were used to detect differences among thedy participants.ults Patients had similar baseline demographic and

nical characteristics, with 63% receiving TPN duringemia in Hematopoieticients Receiving Total

ot StudyTH RICH, MD, PhD

d daily charges than those who did not receive TPN. Noferences were found for in-hospital mortality.clusions TPN is strongly associated with hyperglyce-

a, which may be linked to increased infections of longerration in a profoundly immunocompromised group oftients who frequently receive TPN. The implications ofse findings are limited by the small number of sub-ts; a larger investigation is warranted.m Diet Assoc. 2004;104:1352-1360.

yperglycemia associated with insulin resistance hasbeen increasingly linked to untoward events. In pa-tients with diabetes mellitus, the relationships be-

een inadequate blood sugar control and increased in-tion and death rates have been documented in surgicalpulations (1-4) and in patients following myocardialarction (MI) (5). These associations have also beenserved in nondiabetic patients with higher blood sugarels following MI (6) and ischemic stroke (7), and could

tentially explain the increased risk of mortality ob-

bestatieagcanothdaInvTPare

TmitraprimitalpreTPvaredtimnupatracemmo

MEStuAthetoptwlecme(hymabecinvoccwepuforarefluFunopa

APrverDemarecthegoareqcrimaonwe

P

betrePastegenpribeatyadallTPexcortheRe

DaCohalinmelecatiphporesanneAdumtioprotemPhgawhbathewadaabtribumiDiclastaph

OuThfincos

*TTobuTbymgen reported. Prospective investigations that comparendard hospital care (nothing by mouth until the pa-nt initiates oral intake) to TPN are limited. Whengregated, these investigations demonstrate a signifi-tly lower risk of infection and a trend toward fewerer complications in those patients randomized to stan-

rd hospital care compared with those on TPN (23).estigations that compare standard hospital care toN patients with tightly controlled blood sugar controlyet to be done.he purpose of this pilot investigation was to deter-

ne if the medical records of hematopoietic stem cellnsplant patients could be accurately reviewed for ourmary variables of interest (TPN exposure, hyperglyce-a, infection, in-hospital mortality) at two urban hospi-s with stem cell transplant programs, and to provideliminary data on observed associations between theN exposure to the incidence of hyperglycemia and onrious proxies of morbidity (length of stay, number of

blood cell and platelet transfusions, engraftmente, medication use and duration, neutropenic fevers,

tritional status) and mortality. We hypothesized thattients who received TPN during hospitalization fornsplant would experience significantly more hypergly-ia, which would be associated with increased rates of

rbidity and mortality.

THODSdy Design and Populationretrospective cohort investigation was conducted using

medical records of patients who had undergone hema-oietic stem cell transplant over a 6-month time period ato urban university hospitals. This study design was se-ted because it allowed for efficient and precise ascertain-nt of both the exposure (TPN) and the major outcomesperglycemia, infections, in-hospital mortality). The he-topoietic stem cell transplant population was selectedause it manifests several characteristics optimal for ourestigation. Specifically, the exposure of interest (TPN)urs in 50% to 60% of these patients, overall they are all-nourished group with fairly similar baseline cardiac,lmonary, renal, and hepatic function due to qualificationhematopoietic stem cell transplant. Additionally, theyimmunocompromised and prone to infection; thus, in-

ences on occurrence rates could be readily observable.rthermore, all patients are required to have central ve-us access for transplant thus providing uniformity amongrenteral nutrition delivery.

ll patients 18 years of age and older admitted to Rush-esbyterian St Lukes Medical Center (Rush) or the Uni-sity of Illinois Medical Center (UIC) between July andcember 2001 for an initial autologous or allogeneic he-topoietic stem cell transplant, who had not previouslyeived home TPN, and were without an acute infection attime of admission, were eligible for inclusion. With a

l of reviewing a minimum of 45 records, 59 were initiallyuested. Of these, 48 patients (81%) met our inclusion

teria. Five did not qualify based on transplant type (eg,tched unrelated donor, syngeneic, and umbilical cord),

e had an infection on admission, and five medical recordsre unobtainable for review.atients younger than 18 years of age were excludedcause pediatric hematopoietic stem cell transplantatment and response differs from the adult treatment.tients receiving less common types of hematopoieticm cell transplants (eg, matched unrelated donor, syn-eic, and umbilical cord) or those who had received a

or hematopoietic stem cell transplant were not eligiblecause they potentially represent a more severely ill orpical transplant population compared with patients

mitted for an initial autologous or matched relatedogeneic transplant. Individuals with a history of homeN and those admitted with an existing infection areluded because they have received the exposure (TPN)an outcome (infection) of interest prior to the start of

study. This study was approved by the Institutionalview Boards at Rush and UIC.

ta Collectionmputerized medical records were reviewed at UIC andrd-copy medical records were reviewed at Rush. Base-e clinical data (demographics, primary diagnosis, pastdical or surgical history, transplant type) were col-ted from the physicians history and physical. Prepar-ve chemotherapy regimens were verified from thearmacy chemotherapy dosing records. Laboratory re-rts were used to collect information regarding cultureults (date of infection, source, and infectious agent)d daily blood draw values (white blood count, absoluteutrophil count, triglycerides, albumin, and glucose).mission height and weight, as well as TPN dose, vol-e, and duration, were collected from the initial nutri-

n evaluation or from pharmacy records. Nursinggress notes were used for daily weights, maximumperatures, and 24-hour intake and output volumes.

ysician progress notes were used to gather data re-rding red cell transfusion and/or platelet transfusion,ich were verified by nursing progress notes and bloodnk summaries. Transplant records were used to verify

number of CD34 cells transplanted. Engraftments defined as the number of days between transplanty 0 and the first day of 3 consecutive days that thesolute neutrophil count was more than 0.5109/L. Nu-tional status was assessed using admission serum al-min less than 35 g/L* based on clinical relevancy, ad-ssion body mass index (BMI), using the Centers forsease Control and Prevention overweight and obesityssifications (24) and percent ideal body weight. Vitaltus (alive/dead) at discharge was obtained from theysician progress note.

tcome Measurese primary outcome measure was hyperglycemia, de-ed as blood glucose more than 6.1 mmol/L (12). Glu-e was recorded once per day from the morning venous

o convert g/dL albumin to g/L, multiply g/dL by 10.convert g/L glucose to g/dL, multiply g/L by 0.1. Al-

min of 3.5 g/dL35g/L.o convert mmol/L glucose to mg/dL, multiply mmol/L18.0. To convert mg/dL glucose to mmol/L, multiply/dL by 0.05551. Glucose of 6.0 mmol/L108 mg/dL.Journal of THE AMERICAN DIETETIC ASSOCIATION 1353

blotiopocostheglybloglywiTP

TdustupaanpaAbaleaflucorfroeatoanweine

StaAllanweifnodeofdatogly(yeglycoxmeressta

RESATh((na hmytutofcalTP

BenCDsig

crubuditlen(PreganonphduUIstayouho(PsioTPad

Tsenme1,8difrieprohocei

ThoanrecTPtiotimglydidstanuTPduofglymorec

Rter(Tatheautalfora micatioTPbe0.4nu0.3tiv10gropla

135od draw to achieve uniformity among patients. Addi-nal blood glucose monitoring is not routine among thispulation, and patients with more frequent blood glu-e monitoring were likely more prone to hyperglycemia;refore, morning blood glucose levels were used. Hyper-cemic events were tallied as total number of days withod sugar of 6.1 mmol/L or more. Percentage of hyper-cemic days was also tallied as number of hospital days

th blood sugar more than 6.0 mmol/L/length of time onN.he secondary outcome measures were number and

ration of infections and in-hospital mortality. At bothdy institutions, blood cultures are routinely drawn for

tients with temperatures more than 38C (100.4F),d stool cultures are tested for Clostridium difficile fortients with complaints of or documentation of diarrhea.culture-positive infection was a culture in which acterial or fungal pathogen was isolated or identified atst once from blood or normally sterile tissue or bodyid. For coagulase-negative staphylococcus and nonynebacterium jeikeium group corynebacteria, isolationm at least two blood cultures (drawn within 24 hours ofch other) containing the same organism was requiredbe deemed an infection. Routine surveillance culturesd culture-negative infections (ie, sepsis, pneumonia)re not included. In-hospital mortality was also exam-d to assess differences among groups.

tistical Analysisvariables were examined for the presence of outliers

d distribution of continuous variables. Variables thatre not normally distributed were log transformed, andnormality could not be achieved were analyzed usingnparametric tests. Within each group, standardizedscriptive statistics were calculated, including measuresmeans, median, standard deviations, ranges, and stan-rd errors for continuous variables. 2 analysis was useddetect differences in dichotomous variables [eg, hyper-cemia (yes/no), infection (yes/no), survival at discharges/no)]. For continuous variables (eg, percent of hyper-cemic days, number of infections), Student t and Wil-on rank-sum tests were used to detect differences inans that were normally and non-normally distributed,pectively. Statistical analysis was conducted using thetistical program SAS (Version 8, 2000, Cary, NC).

ULTStotal of 48 patients were included in this pilot study.e mean [standard deviation (SD)] age was 48.2 years13.4), and participants were predominantly male

28, 58%), white (n23, 48%), and 19% (n9) reportedistory of diabetes mellitus prior to admission. Multipleeloma and non-Hodgkins lymphoma (NHL) consti-ed the primary underlying diseases, and 77% (n37)the transplants conducted were autologous. The clini-and demographic characteristics by institution and byN recipient status are listed in Tables 1 and 2.etween hospitals, there were no significant differ-

ces regarding patient demographics or number of34 cells transplanted, but patients at UIC did havenificantly better admission nutritional status, as4 September 2004 Volume 104 Number 9dely assessed by overall BMI (P.002), admission al-min levels (P.03) and percent ideal body weight. Ad-ionally, patients at Rush had significantly longergths of stay (P.03) and higher daily hospital charges.0001) compared with UIC patients. Conditioning

imen differed with regard to rituximab use at Rush,d with an unequal number of TPN patients receivingly melphalan and total body irradiation with cyclophos-amide. Overall, 63% (n30) of patients received TPNring transplant, 73% (n24) at Rush and 40% (n6) atC. When examining these same characteristics by TPNtus, TPN recipients were found to be significantlynger (P.05), with significantly more expensive daily

spital charges (P.0004), longer lengths of stay.0005), and lower percent ideal body weight at admis-

n (P.05). There were no differences in TPN vs nonN recipients with regard to BMI, admission albumin,mission fasting glucose, or mortality.he TPN characteristics for each institution are pre-ted in Table 3. In general, TPN was initiated after aan (SD) of 11.0 (3.0) days, providing a mean of24464 kcal for a mean of 10.27.0 days. Statisticalferences were noted in the mean provision of kilocalo-s per kilogram of body weight (P.01) and grams oftein per kilogram of body weight (P.04) between

spitals, and the majority of the TPN patients did re-ve lipid emulsion (93%).he number of hyperglycemic episodes and percent of

spital days that glucose exceeded 6.1 mmol/L befored after TPN initiation between TPN and nonTPNipients was examined. For comparison in the nonN group, the mean number of days until TPN initia-n was used as the preTPN timeframe (Table 4). In thee period prior to TPN (before), the number of hyper-cemic events and the proportion of hyperglycemic daysnot differ between patients at Rush and at UIC. To

ndardize the timeframes for each group, the meanmber of days of TPN provision was calculated for theN recipients and used as the mean after TPN timeration for the nonTPN recipients. The mean numberhyperglycemic events and median proportion of hyper-cemic days following TPN initiation were significantlyre in the TPN recipients compared with the nonTPNipients at both Rush (P.01) and UIC (P.04).elevant clinical characteristics were evaluated to de-

mine differences between hospitals and TPN patientsble 5). Patients differed by institution with regard tonumber of platelet and red blood cell transfusions,

tologous engraftment time, and percent of total hospi-days with a fever. Further, the overall infection rateboth institutions was examined and found to be low atean (SD) of 0.71 (0.8), which did not differ signif-

ntly by hospital. Subsequently, the number of infec-ns was analyzed according to the before and afterN timeframes. Before TPN, the number of infectionstween TPN and nonTPN recipients was similar,30.57 and 0.440.61, respectively. After TPN thember of mean infections between groups was00.57 and 0.220.43, respectively, which is rela-ely low and also not statistically significant. Finally,0% of repeat positive cultures occurred in the TPNup, and TPN patients were found to have increasedtelet transfusions requirements compared with the

no(2.mogro

DISRepitdaitycohmaaccrecicaanhyavloc

Twa60proTP

T

ICc (

G

R

D

A 54.1A 7.8Le 20.3N

34.73

15

C ,785A 15 (

aTbRcUdSeTfBgMhM*PnTPN group during the after TPN time period23.0 vs 0.80.9; P.08, respectively). In-hospitalrtality differences were not detected between TPNups.

CUSSIONtrospective investigations using medical records of hos-alized patients are often criticized for the quality ofta they provide. Unlike prospective designs, temporal-of events is often difficult to establish in retrospectiveorts, and data is frequently missing. One of the pri-ry objectives of this investigation was to assess theessibility and availability of data from the medicalords and evaluate exposure distribution. Patient med-l records were easily obtainable at both institutions,d the primary variables of interest (TPN exposure,perglycemia, infections, in-hospital mortality) wereailable for all patients and found in more than oneation within the record for verification purposes.he distribution of TPN exposure between institutions

s comparable, and overall TPN exposure occurred in% of patients. To date, there are no evidence-basedtocols in place to classify individuals as appropriateN candidates. TPN initiation in this patient popula-

able 1. Patient characteristics by institution and by TPNa status

Rushb (n33) U

n % n

enderMen 22 67 6Women 11 33 9

aceWhite 16 49 7African American 13 39 6Hispanic 4 12 2

iabetes history 4 12 5

4ge (y) 45.511.2dmission glucosee (mmol/L) 7.33.1ngth of stay (days) 25.88.6

utritional statusBMIf 27.46.9Albumin (g/L)g 31% Ideal body weighth 120%

4harges per day (US $) 6,1891,393 2live at discharge [n (%)] 30 (91)

PNtotal parenteral nutrition.ushRush-Presbyterian St Lukes Medical Center, Chicago, IL.ICUniversity of Illinois Medical Center, Chicago, IL.Dstandard deviation.o convert the values for glucose to mg/dL, divide by 0.05551.MIbody mass index; calculated as kg/m2.ean on admission; to convert g/L to g/dL divide by 10.ean; actual body weight/ideal body weight100..05, difference between institution or TPN.n is often based on subjective complaints of anorexia,usea, or mucositis-related pain from the patient, re-lting in an inability to consume nutrition orally. There-e, TPN initiation at both study institutions was ran-m, and based on the attending physicians discretion,ich broadly represents current practice.he second major purpose of this pilot study was to

ntify ways to account for illness acuity between TPNups and to assess if patients who received TPN expe-nced more hyperglycemia. TPN recipients are oftenemed sicker; thus, any difference in outcomes must bemonstrated to be independent of preTPN health sta-. To address this, similar time points of comparisonong the TPN recipients and nonTPN recipients wereated. Admission glucose level, BMI, and albumin con-trations, previously established correlates of adverse

spital outcomes (25-27), were examined and no differ-ces were found between TPN and nonTPN groups.he second point of comparison for differences in dis-

se status was the before TPN time interval. Patientsboth institutions and in the TPN vs nonTPN groupsd similar numbers of preTPN hyperglycemic days andportions of hyperglycemic events. Although there is nole to measure severity of illness in the hematopoietic

n15) TPN (n30) No TPN (n18)

% n % n %

40 16 53 12 6760 14 47 6 33

47 13 43 6 3340 16 53 7 3913 1 4 5 2833 4 13 5 28

meanSDd 316.1* 45.413.0 52.913.0*2.9 7.83.5 6.92.13.6* 26.68.6 19.83.0*

8.7* 28.78.2 31.28.24* 31 329% 127% 140%

meanSD 3474* 5,9961,777 4,0321,633*100) 28 (93) 17 (94)tionasufordowh

Tidegroriededetusamcrecenhoen

TeaathaproscaJournal of THE AMERICAN DIETETIC ASSOCIATION 1355

stewiistingfouof

T tics bc

15)

TrA (93)A (7)PMNHACACRCC

MMTo

OCA 53.A 3 (n

aTbRcUdOeM

T

DDTokcPrLi

aRbUcS*P

135m cell transplant population, these findings togetherth similar baseline demographic and clinical character-ics, including preTPN infection rates, lends increas-

evidence that the TPN patients were not more pro-ndly ill than nonTPN recipients prior to the initiationTPN. The incidence of hyperglycemia after TPN initi-

able 2. Hematopoietic stem cell transplant related patient characteris

Rushb

(n33)UIC(n

ansplant type [n (%)]utologous 23 (70) 14llogeneic 10 (30) 1rimary diseaseultiple myeloma 9 9on-Hodgkins lymphoma 15 3odgkins lymphoma 3 1cute myeloid leukemia 2 1hronic myeloid leukemia 0 1cute lymphoblastic leukemia 1 0hronic lymphoblastic leukemia 2 0enal cell carcinoma 1 0onditioning regimenarmustine, etopside, cytarabine,

melphalan/rituximab 6/7 4/0elphalan/fludarabine 8/4 8/0elphalan, fludarabine, rituximab 2 0tal body irradiation,cyclophosphamide 2 1

therd 4 2ells transplanted (106/kg)e

utologous 5.094.00 (n19) 5.8llogeneic 4.341.76 (n10) 3.9

PNtotal parenteral nutrition.ushRush-Presbyterian St Lukes Medical Center, Chicago, IL.ICUniversity of Illinois Medical Center, Chicago, IL.thercombinations of busulphan, carmustine, cyclophosphamide, etoposide, fludarabineeanstandard deviation.

able 3. Total parenteral nutrition characteristics by hospital

Rusha

(n24)UICb

(n6)

4 meanSDc 3ays of TPN 10.96.6 6.72.9ays until TPN 11.03.0 12.01.8tal kcal 1,903413 1,508562al (per kg body weight) 274 20*7otein (g per kg body weight) 1.40.3 1.2*0.3pid emulsion type 4 n 310% 0 120% 23 4None 1 1

ushRush-Presbyterian St Lukes Medical Center, Chicago, IL.ICUniversity of Illinois Medical Center, Chicago, IL.Dstandard deviation..05, difference between groups.6 September 2004 Volume 104 Number 9on was examined to help discern if TPN was the sourceelevated blood glucose levels. Not only did TPN pa-nts experience significantly more hyperglycemicnts than nonTPN recipients, but the dramatically

ferent median proportions of hyperglycemic days be-een TPN groups (comparing after time points) per-ps signifies that the vast majority of these patientsve a sustained hyperglycemia while on TPN, which is

exception in the nonTPN patients. Although we didt anticipate enough statistical power to detect signifi-t differences between hyperglycemic events and hy-

rglycemic days, this method did allow us to address theporality of hyperglycemia and eventually will allow

tection of a dose-response using various definitions ofperglycemia and different levels of TPN provision.

ecause hematopoietic stem cell transplant patientscome severely immunocompromised during transplant,y receive multiple prophylactic medications to preventection and are closely monitored for fluid, electrolyte,d blood product support. Although the number of meanections that did occur in the after time period be-een the TPN and non-TPN groups was low and appearsilar (0.300.57 vs 0.220.43, respectively), this dif-

ence is 27% and represents a greater trend in theidence of infection in actual TPN recipients. Addition-y, 69% (9 of 13) of all infections after TPN was startedurred in the group who received TPN. Moreover, all

y institution and by TPNa status

TPN(n30)

No TPN(n18)

22 (73) 15 (83)8 (27) 3 (17)

10 810 84 02 11 01 01 11 0

7/4 3/311/3 5/11 1

3 01 5

15 (n13) 5.673.90 (n18) 5.073.34 (n14)1) 4.111.50 (n8) 4.832.34 (n3)

alan, or rituximab.atioftieevediftwhahathenocanpetemdehy

Bbetheinfaninftwsimferincallocc

, melph

reppeameitieleaau6

T istrat

Yes)

NPrmra

NPrmra

aTbRcTdadatePfPgU*P

T

R 8* (nPlEnAANFeAInMMR

aTbRcUdReMfP*Peat positive cultures occurred in the TPN recipients,rhaps signifying some reduced clearance capabilities asresult of hyperglycemia. In vitro studies have docu-nted that hyperglycemia is associated with abnormal-s in white blood cell function (28-32). Cetin and col-gues (33) conducted a prospective cohort study oftologous hematopoietic stem cell transplant patients (n1) who received either TPN (30 kcal/kg) or partial

able 4. Hyperglycemic events and days before and after TPNa admin

TPN (

Rushb beforec

(n24)

umber of hyperglycemic eventse 6.23.4oportion hyperglycemic days (%)ef 56.128.7edian 59.0nge 0-100

UICg before(n6)

umber of hyperglycemic eventse 6.24.1oportion hyperglycemic days (%)ef 51.434.3edian 41.7nge 17-92

PNtotal parenteral nutrition.ushRush-Presbyterian St Lukes Medical Center, Chicago, IL.he mean number of days prior to TPN initiation was 11 days at Rush and 12 days at Uniys from hospital admission until TPN was started at Rush and UIC, respectively.After represents the standardized time period during which TPN was infused at each hoRush and UIC, respectively.lus-minus values are meansstandard deviation.roportion of hospital days that glucose 6.1 mmol/L during actual or simulated TPN timICUniversity of Illinois Medical Center, Chicago, IL..05, difference between groups.

able 5. Other clinic characteristics by institution and TPNa status

Rushb UICc

BCd transfusionse 4.74.7 (n33) 2.11.atelet transfusionse 5.14.7 (n33) 1.41.graftment time (days)

utologous 13.0 (n21) 10.0* (nllogeneic 12.0 (n9) 12.0 (neutropenic fever n (%) 31 (94) 13 (8brile days (%)ef 24.015.4 (n33) 15.013

ntibiotic usage (days)e 7.14.8 (n33) 6.35.fectionsean 0.790.86 0.530.edian 1.0 0epeat cultures (n) 3 0

PNtotal parenteral nutrition.ushRush-Presbyterian St Lukes Medical Center, Chicago, IL.ICUniversity of Illinois Medical Center, Chicago, IL.BCred blood cells.eanstandard deviation.

roportion of total hospital days with a temperature 38C (100.4F).05, difference between institution or TPN category.ripheral parenteral nutrition (340 kcal/day) and foundt those receiving TPN had more hyperglycemic events,

ections, and longer delays in platelet engraftment. Al-ugh their findings are limited by a nonrandom design,

lure to document similarities between groups for ad-ssion medical and nutritional status, and exposure ofparticipants to some parenteral nutrition, these re-

lts give strength to our findings. Additionally, when the

ion

TPN (No)

Rush afterd

(n24)Rush before(n8)

Rush after(n8)

6.23.7 6.13.0 2.2*3.356.133.7 55.627.4 20.2*30.072.7 54.5 9.10-100 18-100 0-91

UIC after(n6)

UIC before(n9)

UIC after(n9)

5.51.8 5.73.0 2.8*2.578.625.1 47.225.0 39.7*36.385.7 50.0 28.643-100 17-92 0-100

of Illinois Medical Center (UIC) Chicago, IL. Before represents the 11 days and 12

ased on the mean; or 12 to 22 days and 13 to 18 for TPN and nonTPN recipients

.

TPN No TPN

15) 5.04.5 (n30) 2.12.9* (n18)15) 5.14.9 (n30) 2.12.2 (n18)

12.0 (n21) 11.0 (n14)12.0 (n9) 11.0 (n3)

27 (56) 17 (35)n15) 20.216.1 (n30) 19.913.5 (n18)15) 7.44.4 (n30) 5.94.4 (n18)

0.770.90 0.610.700.5 1.03 0pethainfthofaimiallsu

6* (n

14)1)7).0* (

0 (n

74versity

spital b

eframeJournal of THE AMERICAN DIETETIC ASSOCIATION 1357

incferfereadiflognegounodifsio(Pticwipathe

Peticuuntrathegrainauspaensuetiingprepepla60groofrecintietietricandiftioanlutfouThvivchwimemerechecluforTPtie

Aareexhho

thamoof

COItplaofstetriplestaprowidisicacourecitotiotoc

Ref1

2

3

4

5

6

7

8

9

135idence of in-hospital mortality was examined, no dif-ence between TPN groups was apparent; however dif-ences between hospitals were observed. Because dis-se severity and transplant-related mortality areferent between transplant types (eg, autologous vs al-eneic) and Rush conducted 91% (n10) of the alloge-ic transplants, while UIC conducted primarily autolo-s transplants, these results are not unexpected and

t anticipated to be different in this pilot study. Finally,ferences were found in the number of platelet transfu-ns during the TPN infusion timeframe between groups.08). Although this did not achieve traditional statis-

al significance (P.05), these results are consistentth the findings of Cetin and collegues (33) that TPNtients experienced prolonged platelet engraftment,reby requiring greater platelet transfusion support.revious nutrition studies conducted in the hematopoi-

c stem cell transplant population have primarily fo-sed on manipulation of different parenteral regimensder the assumption that TPN was necessary duringnsplant (34-38). The routine use of growth factors in1990s has substantially reduced the time until en-

ftment (39) and consequently the length of time withdequate nutrient intake, thus making the frequent

e of TPN for all hematopoietic stem cell transplanttients questionable. In the early studies, TPN recipi-ts did exhibit earlier engraftment (40) and improvedrvival (41) when compared with nonTPN hematopoi-c stem cell transplant recipients. Although these find-s are of interest, several design flaws limit their inter-tation. Specifically, the investigators combined

diatric and adult patients who may respond to trans-nt very differently, and in the later study more than

% of the control group received TPN, making theups treatments homogeneous and the interpretationthe true effect of TPN on survival questionable. Aent meta-analysis evaluated the effectiveness of TPNvarious populations, including 19 trials of cancer pa-nts (4 with hematopoietic stem cell transplant pa-nts) receiving chemotherapy (42). When these cancerals were aggregated, TPN was associated with signifi-tly increased infectious complications (absolute risk

ference: 16%; 95% CI: 8% to 23%) and total complica-ns (absolute risk difference: 40%; 95% CI: 14% to 66%)d lower tumor-response rates to chemotherapy (abso-e risk difference: 7%; 95% CI: 12% to 1%) werend in TPN recipients compared with the control group.e authors concluded that TPN did not improve sur-al, was associated with impaired tumor response to

emotherapy, and in general, was clearly associatedth net harm in patients undergoing oncology treat-nt. Hyperglycemia, however, was not evaluated in thista-analysis. Finally, Muscaritoli and colleagues (43)ently conducted an extensive review of the use TPN inmatopoietic stem cell transplant populations and con-ded that despite its broad use in this population, in-mation is lacking regarding the overall usefulness ofN for improving or influencing outcomes for these pa-nts.lthough these preliminary results are intriguing, theyconsiderably limited by the small sample size and

ibit only simple basic statistical measures. They will,wever, serve as the basis for a large-scale investigation8 September 2004 Volume 104 Number 9t will allow detection of differences in morbidity andrtality, temporality and dose-response, and inclusiona cost-benefit analysis.

NCLUSIONSwas determined that hematopoietic stem cell trans-nt patients who received TPN had a greater incidencehyperglycemia when compared with hematopoieticm cell transplant patients who received standard nu-tion care. Although these findings were limited by sam-

size, this pilot study provided a fundamental under-nding of the data collection, methodology and analysiscess that will be used in a future investigation, which

ll include a larger patient population and be able tocern if TPN administration is adversely affecting clin-l outcomes in this population. For now, dietitians en-ntering hematopoietic stem cell transplant patientseiving TPN should focus on routine blood sugar mon-ring, overall better blood sugar control, and investiga-ns that could potentially support evidenced-based pro-ols to determine appropriate TPN candidates.

erences. Pomposelli J, Baxter J, Babineau T, Pomfret E,

Driscoll D, Forse R, Bistrian B. Early postoperativeglucose control predicts nosocomial infection rate indiabetic patients. JPEN J Parenter Enteral Nutr.1998;22:77-81.

. Zerr K, Furnary A, Grunkemeier G, Bookin S, Kan-here V, Starr A. Glucose control lowers the risk ofwound infection in diabetics after open-heart opera-tions. Ann Thorac Surg. 1997;63:356-361.

. Golden S, Peart-Vigilance C, Kao W, Brancati F. Peri-operative glycemic control and the risk of infectiouscomplications in a cohort of adults with diabetes.Diabetes Care. 1999;22:1408-1414.

. Latham R, Lancaster A, Covington J, Pirolo J,Thomas C. The association of diabetes and glucosecontrol with surgical-site infections among cardiotho-racic surgery patients. Infect Control Hosp Epide-miol. 2001;22:607-612.

. Malmberg K, Ryden L, Efencid S, Herlitz J, Nicol P,Waldenstrom A, Wedel H, Welin L. Randomized trial ofinsulin-glucose infusion followed by subcutaneous insu-lin treatment in diabetic patients with acute myocar-dial infarction (DIGAMI Study): Effects on mortality at1 year. J Am Coll Cardiol. 1995;26:57-65.

. Capes SE, Hunt D, Malmberg K, Gerstein H. Stresshyperglycemia and increased risk of death after myo-cardial infarction in patients with and without diabetes:A systematic review. Lancet. 2000;355:773-778.

. Capes SE, Hunt D, Malmberg K, Pathak P, GersteinH. Stress hyperglycemia and prognosis in stroke innondiabetic and diabetic patients: A systematic re-view. Stroke. 2001;32:2426-2432.

. Takala J, Ruokonen E, Webster NR, Nielsen M,Zandstra D, Vundelinckx G, Hinds C. Increased mor-tality associated with growth hormone treatment incritically ill adults. N Engl J Med. 1999;341:785-792.

. Montori VM, Bistrian BR, McMahon MM. Hypergly-cemia in acutely ill patients. JAMA. 2002;288:2167-2169.

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39. Furnary AP, Zerr KJ, Grunkemeier GL, Starr A.Continuous intravenous insulin infusion reduces theincidence of deep sternal wound infection in diabeticpatients after cardiac surgical procedures. Ann Tho-rac Surg. 1999;67:352-362.

. Malmberg K. Prospective randomized study of inten-sive insulin treatment on long term survival afteracute myocardial infarction in patients with diabetesmellitus. DIGAMI (Diabetes Mellitus, Insulin Glu-cose Infusion in Acute Myocardial Infarction) StudyGroup. BMJ. 1997;314:1512-1515.

. Van den Berghe G, Wouters P, Weekers F, VerwaestC, Bruyninckx F, Schetz M, Vlasselaers D, Ferdi-nande P, Lauwers P, Bouillon R. Intensive insulintherapy in critically ill patients. N Engl J Med. 2001;345:1359-1367.

. Kalfarentzos R, Kehadias J, Mead N, Kokkinis K,Gogos CA. Enteral nutrition is superior to parenteralnutrition in severe acute pancreatitis: Results of a ran-domized prospective trial. Br J Surg. 1997;84:1665-1669.

. McClave SA, Greene LM, Snider HL, Makk L, Chea-dle W, Owens N, Dukes L, Goldsmith L. Comparisonof the safety of early enteral vs. parenteral nutritionin mild acute pancreatitis. JPEN J Parenter EnteralNutr. 1997;21:14-20.

. Gonzalez-Huix F, Fernandez-Banares F, Esteve-Co-mas M, Abad-Lacruz A, Cabre E, Acero D, Figa M,Guilera M, Humbert P, de Leon R, Gassull M. En-teral vs. parenteral nutrition as adjunct therapy inacute ulcerative colitis. Am J Gastroenterol. 1993;88:227-232.

. Bower RH, Talamini MA, Sax HC, Hamilton F, Fi-scher J. Postoperative enteral vs. parenteral nutri-tion. Arch Surg. 1986;121:1040-1045.

. Hamaoui E, Lefkowitz R, Olender L, Krasnopolsky-Levine E, Favale M, Webb H, Hoover E. Enteralnutrition in the early postoperative period: A newsemi-elemental formula versus total parenteral nu-trition. JPEN J Parenter Enteral Nutr. 1990;14:501-507.

. Iovinelli G, Marsil I, Varrassi G. Nutrition supportafter total laryngectomy. JPEN J Parenter EnteralNutr. 1993;17:445-448.

. Levine G, Deren J, Steiger E, Zinno R. Role of oralintake and maintenance of gut mass and disaccha-ride activity. Gastroenterology. 1974;67:975-982.

. Purandare S, Offenbartl K, Westerom B, BengmarkS. Increased gut permeability of floresein isothiocya-nate-dextran after total parenteral nutrition. ScandJ Gastroenterol. 1989;24:678-682.

. Deitch E, Winterton J, Ma L, Berg R. The gut is aportal of entry for bacteremia: role of protein malnu-trition. Ann Surg. 1987;13:565-571.

. Lipman T. Bacterial translocation and enteral nutri-tion in humans: An outsider looks in. JPEN J Par-enter Enteral Nutr. 1995;19:156-165.

. Braunschweig C, Levy P, Sheean P, Wang X. Enteralcompared with parenteral nutrition: A meta-analy-sis. Am J Clin Nutr. 2001;74:534-542.

. Centers for Disease Control and Prevention. Avail-able at: http:www.cdc.gov. Accessed July 10, 2003.. Umpierrez GE, Isaacs SD, Bazargan N, You X, Tha-ler LM, Kitabchi AE. Hyperglycemia: An indepen-dent marker of in-hospital mortality in patients withundiagnosed diabetes. J Clin Endocrinol Metab.2000;87:978-982.

. Hopkins B. Assessment of Nutritional Status. In:Gottsclich MM, Matarese LE, Shronts EP, eds. Nu-trition Support Dietetics Core Curriculum. 2nd ed.Silver Spring, MD: American Society of Parenteraland Enteral Nutrition; 1993.

. Gabay C, Kushner I. Acute-phase proteins and othersystemic responses to inflammation. N Engl J Med.1999;340:448-454.

. Bagdade J, Koot R, Bulger R. Impaired leukocytefunction in patients with poorly controlled diabetes.Diabetes. 1974;23:9-15.

. Bagdade J, Stewart M, Walters E. Impaired granu-locyte adherence: a reversible defect in host defensein patients with poorly controlled diabetes. Diabetes.1978;27:677-681.

. Mowat A, Baum J. Chemotaxis of polymorphonuclearleukocytes from patients with diabetes mellitus.N Engl J Med. 1971;284:621-627.

. Bagdade J, Nielson K, Bulger R. Reversible abnor-malities in phagocytic function in poorly controlleddiabetic patients. Am J Med Sci. 1972;263:451-456.

. Nolan C, Beaty H, Bagdade J. Impaired granulocytebactericidal function in patients with poorly con-trolled diabetes. Diabetes. 1978;127:889-894.

. Cetin T, Arpaci F, Dere Y, Turan M, Ozturk B,Komurcu S, Ozet A, Beyzadeoglu M, Kaptan K,Beyan C, Yalcin A. Total parenteral nutrition delaysplatelet engraftment in patients who undergo autol-ogous hematopoietic stem cell transplantation. Nu-trition. 2002;18:599-603.

. Taveroff A, McArdle AH, Rybka WB. Reducing par-enteral energy and protein intake improves meta-bolic homeostasis after bone marrow transplant.Am J Clin Nutr. 1991;54:1087-1092.

. Ziegler TR, Young LS, Benfell K, Scheltinga M, Hor-tos K, Bye R, Morrow FD, Jacobs DO, Smith RJ,Antin JH, Wilmore DW. Clinical and metabolic effi-cacy of glutamine-supplemented parenteral nutritionafter bone marrow transplantation. Ann Intern Med.1992;116:821-828.

. MacBurney M, Young LS, Ziegler TR, Wilmore DW.A cost-evaluation of glutamine-supplemented paren-teral nutrition in adult bone marrow transplant pa-tients. J Am Diet Assoc. 1994;94:1263-1266.

. Muscaritoli M, Conversano L, Torelli G, Arcese W,Capria S, Cangiano C, Falcone C, Rossi Fanelli F.Clinical and metabolic effects of different parenteralnutrition regimens in patients undergoing allogeneicbone marrow transplantation. Transplantation.1998;66:610-616.

. Schloerb PR, Skikne BS. Oral and parenteral glu-tamine in bone marrow transplantation: A random-ized, double-blind study. JPEN J Parenter EnteralNutr. 1999;23:117-122.

. Jansen J, Thompson JM, Dugan MJ, Nolan P, Wi-emann M, Birhiray R, Henslee-Downey P, Akard L.Journal of THE AMERICAN DIETETIC ASSOCIATION 1359

Peripheral blood progenitor cell transplantation.Ther Apher. 2002;6:5-14.

40. Weisdorf S, Hofland C, Sharp HL, Teasley K, SchisselK, McGlave PB, Ramsay N, Kersey J. Total paren-teral nutrition in bone marrow transplantation: Aclinical evaluation. J Pediatr Gastroenterol Nutr.1984;3:95-100.

41. Weisdorf SA, Lysne J, Wind D, Haake RJ, Sharp HL,Goldman A, Schissel K, McGlave PB, Ramsay NK,Kersey JH. Positive effect of prophylactic total paren-

teral nutrition on long-term outcome of bone marrowtransplantation. Transplantation. 1987;43:833-838.

42. Koretz RL, Lipman TO, Klein S, American Gastro-enterological Association. AGA technical review onparenteral nutrition. Gastroenterol. 2001;121:970-1001.

43. Muscaritoli M, Grieco G, Capria S, Iori A, Rossi F.Nutritional and metabolic support in patients under-going bone marrow transplantation. Am J Clin Nutr.2002;75:183-190.

1360 September 2004 Volume 104 Number 9

Continuing Education Questionnaire, page 1497Meets Learning Need Codes 5000, 5360, and 5440The Incidence of Hyperglycemia in Hematopoietic StemMETHODSStudy Design and PopulationData CollectionOutcome MeasuresStatistical Analysis

RESULTSDISCUSSIONReferences