Pediatr Blood Cancer 2008;51:387–392
Hyperglycemia During Induction Therapy Is Associated With Increased Infectious
Complications in Childhood Acute Lymphocytic Leukemia
Rona Y. Sonabend, MD,1* Siripoom V. McKay, MD,1 M. Fatih Okcu, MD, MPh,2,3 Jinrong Yan,2,3
Morey W. Haymond, MD,1 and Judith F. Margolin, MD2,3
Background. Children with acute lymphocytic leukemia (ALL)
are at high risk for developing hyperglycemia. Hyperglycemic adult
ALL patients have shorter remissions, more infections, and increased
mortality. No corresponding data are available in children. We
hypothesized that children with ALL who become hyperglycemic
during induction chemotherapy have an increased risk for infection
during their first year of treatment. Procedure. We conducted a
retrospective chart review of 135 patients diagnosed with ALL during
1999–2001 at Texas Children’s Hospital. Infectious outcomes during
the first year of therapy were compared in three groups patients
based on blood glucose concentrations during induction therapy:
euglycemic (<140 mg/dl), mild hyperglycemic (MH) (140–199 mg/
dl) and overt hyperglycemic (OH) (blood glucose >200 mg/dl).
Results. Seventy-five (56%) patients met criteria for either MH (21%)
Key words: acute lymphoblastic leukemia; cancer; child; complications; hyperglycemia; infections; steroids
INTRODUCTION
Acute lymphocytic leukemia (ALL) is the most common
malignancy of childhood accounting for 25% of all childhood
cancers. Despite significant advancements in survival with an
overall cure rate of 80%, ALL therapy continues to pose significant
morbidity. Serious infections requiring hospitalization during active
treatment are common consequences of ALL therapy [1].
In recent years, much attention has been paid to the role of
hyperglycemia during acute illness and whether insulin therapy
leads to reduced mortality and morbidity [2–9]. Children with ALL
are at high risk to develop hyperglycemia. Patients undergoing a
critical illness and/or receiving corticosteroids have wide fluctua-
tions in glucose homeostasis [10,11]. Additionally, L-asparaginase,
a chemotherapeutic agent used in most ALL protocols, has been
linked to hyperglycemia and ketosis [12]. The risk of developing
hyperglycemia may not be transient as survivors of childhood ALL
have an increased prevalence of obesity, insulin resistance,
metabolic syndrome, and diabetes mellitus [13–16].
Hyperglycemia is seen in adult ALL patients during the
first month of therapy and is associated with shorter remission,
shorter median survival, and increased risk of sepsis, but its impact
on children with ALL remain unexplored [17]. In this report we
describe the prevalence of hyperglycemia and its relationship to
complications over the first year of therapy in children with ALL.
We hypothesized that children with ALL who experience hyper-
glycemia during their first month of treatment (induction therapy)
will experience more serious infections and have more frequent
hospitalizations during their first year of therapy.
PATIENTS AND METHODS
Study Population
Following IRB approval, the charts of all patients newly
diagnosed with ALL between January 1999 and December 2001
at Texas Children’s Hospital were reviewed. Patients were treated
ß 2008 Wiley-Liss, Inc.
DOI 10.1002/pbc.21624
or OH (35%). Hyperglycemia was more prevalent in older children
(P < 0.001) and those at risk for being overweight (BMI% >85%) at
diagnosis (P < 0.01). Patients with MH and OH were 2.5 times (95%
CI 1.0–6.2) and 2.1 times (95% CI 1.0–4.6) more likely to have
documented infections, respectively. Patients with OH were
4.2 times (95% CI 1.5–12) more likely to have bacteremia/fungemia,
3.8 times (95% CI 1.2–11.6) more likely to have cellulitis, and
4.0 times (95% CI 1.7–9.3) more likely to be admitted for fever
and neutropenia than the euglycemia group. Conclusion. Hyper-
glycemia, especially when overt, may be a previously unrecognized
risk factor of infectious complications in children with ALL during
the first year of treatment. Pediatr Blood Cancer 2008;51:387–
392. ß 2008 Wiley-Liss, Inc.
primarily on or according to Pediatric Oncology Group (POG)
legacy protocols [9,201 (n ¼ 13), 9,404 (n ¼ 12), 9,406 (n ¼ 9),
9,605 (n ¼ 8), 9,705 (n ¼ 8), 9,806 (n ¼ 9), 9,904 (n ¼ 20), 9,905
(n ¼ 48), 9,906 (n ¼ 6), and 9,917 (n ¼ 2)] which used the NCI
definition for standard or high risk of disease recurrence/cancer
treatment failure based on the patient’s age and white cell count at
diagnosis [18]. Patients received either 3- or 4- drug induction
regimen consisting of prednisone or dexamethasone, E. coli
asparaginase, vincristine, with or without daunorubicin. The
glucocorticoid dose schedule for these series includes either
prednisone (40 mg/m2) or dexamethasone (6 mg/m2) for 28 days.
Infants under one year of age were excluded from this study due to
high variability of treatment series and automatic hospitalization
while undergoing induction therapy. Hydration was administered
following standard protocol guidelines with 5% dextrose at a rate of
1.5–2 times maintenance. A glucose infusion rate (GIR) was
calculated on all patients at the beginning of induction therapy using
the following formula (dextrose% concentration
ml/kg/day)/144.
Any patient who had a pre-existing impaired glucose tolerance,
acanthosis nigricans, diabetes mellitus, or history of steroid use
within 1 month of diagnosis was excluded from the study.
Additionally, patients who failed to complete induction therapy at
our institution for any reason other than death were also excluded.
—
1
—————
Department of Pediatrics, Endocrinology and Metabolism, Baylor
College of Medicine, Houston, Texas; 2Department of Pediatrics,
Hematology and Oncology, Baylor College of Medicine, Houston,
Texas; 3Texas Children’s Cancer Center, Baylor College of Medicine,
Houston, Texas
Grant sponsor: Thrasher Research Fund.
*Correspondence to: Rona Y. Sonabend, Clinical Care Center, Suite
1020, 6621 Fannin St, CC 1020.05, Houston, TX 77030-2399.
E-mail: rysonabe@texaschildrenshospital.org
Received 19 November 2007; Accepted 14 April 2008
388 Sonabend et al.
Data Collection
Data collected at diagnosis included age, gender, height, weight,
BMI%, ethnicity, race, pre-existing medical conditions, family
history of diabetes, ALL risk category, CNS status of disease, and
record of insulin administration during induction therapy. Ethnicity
and race were assigned on the basis of parental report. Family
history of diabetes was defined as having a parent, sibling, or
grandparent with a diagnosis of either type 1 or type 2 diabetes
mellitus.
All inpatient and outpatient plasma blood glucose measurements
taken during induction chemotherapy were collected. Patients who
did not have at least two blood glucose measurements recorded
during induction were excluded from the study. For patients that
required insulin for hyperglycemia, blood glucose concentrations
obtained after initiation of insulin therapy were excluded from
analysis.
Patients, who had two or more plasma blood glucose concen-
tration >140 mg/dl without any measurements >200 mg/dl, were
classified into the mild hyperglycemia (MH) group. The overt
hyperglycemia (OH) group consisted of patients who had one or
more plasma blood glucose concentration of >200 mg/dl. These
cutoffs were chosen to correlate with the definitions of impaired
glucose tolerance and diabetes as established by the American
Diabetes Association and the World Health Organization, respec-
tively [19,20].
Clinical follow up data were collected for the year following
diagnosis. Information on any patient who had a relapse, bone
marrow transplant, or death prior to 1 year time point from diagnosis
was censored at the time of the event. Infections had to be
documented with positive cultures or radiologic findings. Only
infections that received a full course of antibiotic treatment were
included in our analysis. Common skin contaminants, that is, Staph
Epidermidis (CONS), were included in our analysis if the treating
physicians deemed them to be true infections requiring antibiotic
treatment. Hospitalization for fever and neutropenia (F/N) was
defined as an admission in which the patient had both a (1)
temperature >100.58F and (2) an absolute neutrophil count of less
than 1,000 cells/mm3. Hospitalization for suspected infection
included admissions to rule out infection in which the patient failed
to be both neutropenic and febrile.
Statistical Analyses
The prevalence of hyperglycemia was defined as the percentage
of patients who met our criteria for hyperglycemia during induction
chemotherapy. BMI percentiles (BMI%) were calculated using the
age and sex specific standardized BMI z-scores derived from the
Centers for Disease Control and Prevention (CDC) National Center
for Health Statistics (NCHS) growth curves. BMI% have not been
established for children under 2 years of age. The following
outcomes were compared: (1) Reason for admission after discharge
following initiation of induction therapy including F/N and
suspected infection and (2) Occurrence of infectious complications
including bacteremia/fungemia, urinary tract infections, pneumo-
nia, cellulitis, and meningitis.
Univariate differences between the hyperglycemic groups and
the euglycemic group were analyzed using ANOVA for continuous
variables and chi-square test for categorical variables. Proportion
comparisons were done by two-sample proportion test for
Pediatr Blood Cancer DOI 10.1002/pbc
comparisons of total occurrences of infection by glycemia group,
as each infection was analyzed as an independent variable. For
reason for admission and occurrence of infectious complication
comparisons, we calculated odds ratios (OR) for each outcome and
used logistic regression analysis including -age, ethnicity, risk of
disease, CNS status, gender, BMI, BMI%, steroid formulation, and
family history of diabetes as possible covariates. For these analyses,
adverse infectious complications were only counted once for each
patient regardless if the patient experienced multiple similar
complications. Final multivariable model included only statistically
significant variables. We used SPSS (Version 15.0; SPSS Inc.,
Chicago, IL) software package and considered a P-value of <0.05 as
significant. All tests of statistical significance were two sided.
RESULTS
One hundred forty-seven children with ALL were identified
during the study period and 135 met all inclusion/exclusion criteria.
Twelve patients failed to meet eligibility criteria for the following
reasons: age <1 year (3), discontinued induction therapy (4), less
than 2 blood glucose values (3), steroids use prior to induction (2).
Demographic Comparisons
Each eligible patient had an average of 14 (range, 2–52) blood
glucose measurements during induction therapy. Seventy-five
(56%) patients met previously defined criteria for hyperglycemia.
MH was seen in 28 (21%) and OH was seen in 47 (35%) patients.
Eleven patients (8%) received insulin during induction. The glucose
infusion rate at diagnosis was lowest in the OH group compared to
the MH and the euglycemic group (2.9mg/kg/min vs. 3.4mg/kg/min
vs. 3.5 mg/kg/min, respectively, P < 0.01).
Baseline demographics and mean blood glucose concentrations
according to glycemia status are shown in Table I. Mean blood
glucose concentrations were different among the OH, MH and
euglycemia groups (P < 0.001). Patients with hyperglycemia were
older and more likely to have high-risk disease at diagnosis
(P < 0.001). Additionally, they had a higher mean BMI% at the time
of diagnosis (P ¼ 0.04). Subsequent stratification of BMI% revealed
an increase in hyperglycemia in children who are at risk for
overweight, defined as BMI% of >85% (n ¼ 49; P < 0.01). There
were no differences in ethnicity, gender, or family history of
diabetes among the groups. CNS disease (8%) showed no significant
difference among groups (5 in euglycemia vs. 2 in MH vs. 4 in OH).
Data on steroid dose and formulation during induction therapy were
available on 127 patients. Patients who received prednisone (40 mg/
m2/day) were more likely to have OH, compared to patients who
received dexamethasone (6 mg/m2/day) (17/36 vs. 24/91). However
this difference did not reach statistical significance, P ¼ 0.07.
Reason for Hospital Admission
Table II shows comparisons of hospital admissions during year 1.
There was a trend toward increased hospitalization for infection
related complications as glycemia worsened (78% vs. 82% vs. 94%,
P ¼ 0.09). 77% of patients in the OH group were admitted at least
once for F/N compared to only 45% in the euglycemia group
(P ¼ 0.004). The median time to admission from the start of
induction for F/N was 43.5 days (range, 3–351 days). The
hyperglycemic (mild and overt combined) patients were 2.9 times
 Hyperglycemia in Children With ALL 389
TABLE I. Patient Demographics by Glycemia Group

Euglycemia Mild Overt P value

N 60 28 47
Mean blood glucose  SD 109  10 123  13 168  56 P < 0.001
Age (years) of dx  SD 53 75 10  5 P < 0.001
Gender P ¼ 0.98
Male 29 (48%) 13 (46%) 22 (47%)
Female 31 (52%) 15 (54%) 25 (53%)
Mean BMI%  SD 54  30 58  35 71  34 P ¼ 0.04
BMI% P < 0.01
85 48 (80%) 17 (61%) 21 (45%)
>85 12 (20%) 11 (39%) 26 (55%)
Ethnicity P ¼ 0.17
Caucasian 32 (53%) 11 (39%) 18 (38%)
Hispanic 22 (37%) 16 (57%) 21 (45%)
Other 6 (10%) 1 (4%) 8 (17%)
Risk group at diagnosis P < 0.001
High 11 (18%) 14 (50%) 33 (70%)
Low 49 (82%) 14 (50%) 14 (30%)
Family history of DM in 1st and 17 (28%) 10 (36%) 19 (40%) P ¼ 0.63
2nd relatives
Steroids P ¼ 0.07
Prednisone (40 mg/m2) 12 (20%) 7 (25%) 17 (36%)
Dexamethasone (6 mg/m2) 47 (78%) 20 (71%) 24 (51%)

Mild hyperglycemia: two or more blood glucose concentrations >140 and <200 mg/dl; overt
hyperglycemia: 1 blood glucose concentration >200 mg/dl; high risk defined as WBC >50,000 cells/
mm3 or age 10 years of age; comparisons between categorical data done by Chi Square analysis;
comparisons between continuous variables done by ANOVA.

(95% CI 1.4–6.0) more likely to be hospitalized for F/N (Table III,
Fig. 1). Patients with mild hyperglycemia were at 1.9-fold (95% CI
0.76–4.7) and overt patients at 4-fold (95% CI 1.7–9.3) greater risk
for hospitalizations due to F/N when compared to their euglycemic
counterparts. Hospital admission for suspected infection was
comparable in all groups.
Documented Infections
In total, 94 occurrences of infectious complications were seen in
76 patients during the first year of therapy. Due to the small numbers
of each infectious complication, MH and OH groups were
combined. There were a total of 43 bacteremia/fungemia occur-
rences, 36 of which occurred in the hyperglycemic patients
(P < 0.001, 95% CI 0.69–0.93). Fifteen different organisms were
found in our patients who experienced bacteremia/fungemia: CONS
(13), Staph Aureus (6), Alpha Strep (6), Pseudomonas (2), E. coli
(3), Klebsiella (3), Enterobacter (3), Enterococcus (2), Diptheroid
(1), Pasteurella (1), Bacillus (1), Acinetobacter (1), Salmonella (1),
Serretia (1), Rhodococcus Equi (1) and Aspergillosis (1). The
TABLE II. Proportion of Patients Admitted for Hospitalization During Year 1 of Treatment for
ALL by Glycemic Group
median time to occurrences of bacteremia/fungemia from day one
of induction therapy was 53 days (range, 2–268 days). Thirteen
patients experienced bactermia/fungemia in the first 29 days of
therapy. Of these, 2 were euglycemic, 4 had MH, and 7 had OH
(P ¼ 0.15). Cellulitis was also found to be more common in the
hyperglycemic patients as 76% of all cellulitis infections (16 out of
21) occurred in the hyperglycemic group (P ¼ 0.027, 95% CI 0.53–
0.92). The median time to the occurrence of cellulitis from the start
of induction therapy was 31 days (range, 9–336 days).
The hyperglycemic (mild and overt combined) patients were
2.2 times (95% CI 1.1–4.5) more likely to have any documented
infection (Table III, Fig. 1). Specifically, they were 3.7 times (95%
CI 1.4–9.9) more likely to develop bactermia/fungemia, and
3.0 times (95% CI 1.0–8.7) more likely to develop cellulitis than
their euglycemic counterparts. Patients with mild hyperglycemia
alone were 2.5 times (95% CI 1.0–6.2) more likely to have
documented infection than the euglycemic group. When compared
to their euglycemic counterparts, patients with OH were 2.1 times
(95% CI 1.0–4.6) more likely to have any documented infection,
4.2 times (95% CI 1.5–12.0) more likely to have bacteremia/

Euglycemia
(n ¼ 60) Mild (n ¼ 28) Overt (n ¼ 47) P value
Non-chemotherapy 47 (78%) 23 (82%) 44 (94%) 0.09
Suspected infection 23 (38%) 16 (57%) 25 (53%) 0.16
F/Na 27 (45%) 17 (61%) 36 (77%) 0.004
Other 14 (23%) 11 (39%) 17 (36%) 0.21
a
Fever and neutropenia.
Pediatr Blood Cancer DOI 10.1002/pbc
390 Sonabend et al.
TABLE III. Univariate Comparisons of Infectious Adverse Outcomes During Year 1 of Therapy by
Glycemia Group

OR 95% CI
a
Hospitalization for F/N
Euglycemia (n ¼ 27) 1
Hyperglycemia (mild þ overt) 2.9 1.4–6.0
Mild (n ¼ 16) 1.9 0.76–4.7
Overt (n ¼ 36) 4.0 1.7–9.3
Cumulative documented infections
Euglycemia (n ¼ 24) 1
Hyperglycemia (mild þ overt) 2.2 1.1–4.5
Mild (n ¼ 18) 2.5 1.0–6.2
Overt (n ¼ 28) 2.1 1.0–4.6
Bacteremia/fungemia
Euglycemia (n ¼ 7) 1
Hyperglycemia (mild þ overt) 3.7 1.4–9.9
Mild (n ¼ 9) 3.0 0.9–10.0
Overt (n ¼ 17) 4.2 1.5–12.0
Cellulitis
Euglycemia (n ¼ 6) 1
Hyperglycemia (mild þ overt) 3.0 1.0–8.7
Mild (n ¼ 6) 1.8 0.5–7.4
Overt (n ¼ 12) 3.8 1.2–11.6
a
Fever and neutropenia.

fungemia and 3.8 times (95% CI 1.2–11.6) more likely to have
cellulitis. Adverse outcomes for UTI and pneumonia were not
different among the three groups. Multivariable logistic regression
analysis showed that hyperglycemia is a statistically significant
independent risk factor for bactermia/fungemia (Table IV).
DISCUSSION
This study demonstrates that over half of children with ALL
experience some degree of hyperglycemia and that these patients
were more likely to develop adverse outcomes particularly
hospitalizations for F/N and infections, including cellulitis and
bacteremia/fungemia. Children undergoing induction chemother-
apy for ALL exhibit hyperglycemia at rates similar to their adult
counterparts [17]. Unique to our study is the observation of mild
hyperglycemia (21%) is higher than previously reported [15,21].
This may be due to the high percentage of overweight and obese
patients (36%) and/or our inclusion of non-fasting glucose values in
our study. We believe that the previous studies may have overlooked
the importance of postprandial hyperglycemia. The earliest
evidence of hyperglycemia is often seen postprandially; the 2-hr
glucose concentration in an oral glucose tolerance test is a more
sensitive test for diabetes mellitus than the fasting glucose
[20,22,23]. Additionally, postprandial hyperglycemia has been
shown to be a more powerful predictor of diabetic complications
than glycosylated hemoglobin (HbA1c) and fasting hyperglycemia
alone [24].
It is well known that corticosteroids and L-asparaginase, agents
used in induction protocols, have a direct effect on glucose
homeostasis. Corticosteroids produce a state of insulin resistance

Fig. 1. Univariate comparisons of adverse complications during year one of therapy. *Comparison between euglycemia (OR ¼ 1) and
hyperglycemia (overt þ mild).
Pediatr Blood Cancer DOI 10.1002/pbc
 Hyperglycemia in Children With ALL 391
TABLE IV. Multivariate Comparisons of Infectious Adverse Outcomes During Year 1 of Therapy

OR 95% CI
a
Hospitalization for F/N
Risk group
Standard risk 1
High risk 3.4 1.4–8.1
Family history of DMb
No 1
Yes 3.4 1.4–8.1
Blood glucose
Euglycemia 1
Mild hyperglycemia 1.2 0.5–3.4
Overt hyperglycemia 2.1 0.8–5.5
Cumulative documented infections
BMI%c
BMI% <85% 1
BMI% 85% 1.3 0.6–2.8
Blood glucose
Euglycemia 1
Mild hyperglycemia 2.7 1.0–6.9
Overt hyperglycemia 2.2 0.9–5.0
Bacteremia/fungemia
BMI%
BMI% <85% 1
BMI% 85% 0.8 0.3–2.1
Blood glucose
Euglycemia 1
Mild hyperglycemia 6.2 1.4–62.7
Overt hyperglycemia 8.9 2.3–34.9
a
Fever and neutropenia; bDiabetes mellitus; cBody mass index.

while L-asparaginase is postulated to damage to the pancreatic beta
cells [12,25]. Additionally, hyperglycemia can be secondary to
acute stress, which increases glucose production and peripheral
insulin resistance [10]. Higher BMI%, older age, and high-risk
disease contribute to hyperglycemia.
Although we do not have glucose measurements at the time of
the complications, there is sufficient data in adults to suggest that
mild hyperglycemia progresses and worsens over time [26].
Therefore hyperglycemia may, in fact, persist throughout induction
therapy and into consolidation and maintenance therapy. Hyper-
glycemia during induction may be a marker of patients who
remained hyperglycemic throughout treatment and may have had
unrecognized hyperglycemia at the times of infections. Whether a
predisposition to infections was secondary to the hyperglycemia
cannot be determined from our data.
From our comparison of the MH and OH groups, the frequency
of adverse infectious outcomes increases with a rise in blood
glucose concentrations. Whether therapeutic intervention aimed at
reducing hyperglycemia would improve the outcome of children
with ALL as has been reported in adults remains to be determined
[7–9].
There are several limitations to the current study that need to be
considered. This is a retrospective study and is therefore limited to
the data in the patients’ charts. Blood glucose concentrations were
not sampled in a standardized fashion with regard to timing or
frequency. It is likely that patients without significantly elevated
glucose levels were not monitored as closely for hyperglycemia and
may have had undetected subsequent hyperglycemia. This may have
Pediatr Blood Cancer DOI 10.1002/pbc
lead to an underestimation of the true incidence of mild hyper-
glycemia. Second, there were 16 patients that did not complete a
full year of treatment at our institution leading us to censor their
data. This could have lead to an omission error in the total number of
adverse outcomes. Third, family history of diabetes was likely
underreported, as we cannot be sure of how precise of a history was
obtained. Lastly, we do not know if hyperglycemia persisted beyond
the induction period. Therefore we are unable to say with any
certainty whether it is transient or chronic exposure to hyper-
glycemia that is associated with adverse outcome.
In our study, while hyperglycemia was common, only 11 patients
were treated with insulin. This study will raise awareness about
hyperglycemia and its associated adverse outcomes and lead to
routine monitoring of blood glucose concentrations in order to
detect hyperglycemia. Further investigation is underway to validate
our findings in a prospective fashion and also determine the
prognostic impact of early hyperglycemia on complications beyond
the first year, rates of remission and relapse of disease, and late-
effects seen in childhood cancer survivors.
ACKNOWLEDGMENT
The authors thank Dr. Angeta Sunehag for her support and
encouragement on this paper; Dr. Rebecca Kim and Rahul Suresh
for their help with chart review and data collection; and the medical
records staff for their dedication and hard work. The Thrasher
Research Fund, Salt Lake City, UT, sponsors this work. There are no
competing financial interests.
392 Sonabend et al.
REFERENCES
1. Carroll WL, Bhojwani D, Min DJ, et al. Pediatric acute
lymphoblastic leukemia. Hematology Am Soc Hematol Educ
Program 2003;102–131.
2. Devos P, Preiser JC. Tight blood glucose control: A recommen-
dation applicable to any critically ill patient? Crit Care
2004;8:427–429.
3. Faustino EV, Apkon M. Persistent hyperglycemia in critically ill
children. J Pediatr 2005;146:30–34.
4. Srinivasan V, Spinella PC, Drott HR, et al. Association of timing,
duration, and intensity of hyperglycemia with intensive care unit
mortality in critically ill children. Pediatr Crit Care Med
2004;5:329–336.
5. Ali NA, O’Brien JM, Jr., Blum W, et al. Hyperglycemia in patients
with acute myeloid leukemia is associated with increased hospital
mortality. Cancer 2007;110:96–102.
6. Pittas AG, Siegel RD, Lau J. Insulin therapy for critically ill
hospitalized patients: A meta-analysis of randomized controlled
trials. Arch Intern Med 2004;164:2005–2011.
7. van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin
therapy in the critically ill patients. N Engl J Med 2001;345:1359–
1367.
8. Vanhorebeek I, Langouche L, Van den Berghe G. Glycemic and
nonglycemic effects of insulin: How do they contribute to a better
outcome of critical illness? Curr Opin Crit Care 2005;11:304–
311.
9. Egi M, Bellomo R, Stachowski E, et al. Variability of blood glucose
concentration and short-term mortality in critically ill patients.
Anesthesiology 2006;105:244–252.
10. Gearhart MM, Parbhoo SK. Hyperglycemia in the critically ill
patient. AACN Clin Issues 2006;17:50–55.
11. Marik PE, Raghavan M. Stress-hyperglycemia, insulin and
immunomodulation in sepsis. Intensive Care Med 2004;30:748–
756.
12. Cetin M, Yetgin S, Kara A, et al. Hyperglycemia, ketoacidosis and
other complications of L-asparaginase in children with acute
lymphoblastic leukemia. J Med 1994;25:219–229.
13. Baillargeon J, Langevin AM, Lewis M, et al. Demographic
correlates of body size changes in children undergoing treatment
for acute lymphoblastic leukemia. Pediatr Blood Cancer 2007;49:
793–796.
14. Baillargeon J, Langevin AM, Lewis M, et al. Obesity and survival
in a cohort of predominantly hispanic children with acute
Pediatr Blood Cancer DOI 10.1002/pbc
lymphoblastic leukemia. J Pediatr Hematol Oncol 2006;28:575–
578.
15. Baillargeon J, Langevin AM, Mullins J, et al. Transient hyper-
glycemia in hispanic children with acute lymphoblastic leukemia.
Pediatr Blood Cancer 2005;45:960–963.
16. Neville KA, Cohn RJ, Steinbeck KS, et al. Hyperinsulinemia,
impaired glucose tolerance, and diabetes mellitus in survivors of
childhood cancer: Prevalence and risk factors. J Clin Endocrinol
Metab 2006;91:4401–4407.
17. Weiser MA, Cabanillas ME, Konopleva M, et al. Relation between
the duration of remission and hyperglycemia during induction
chemotherapy for acute lymphocytic leukemia with a hyper-
fractionated cyclophosphamide, vincristine, doxorubicin, and
dexamethasone/methotrexate-cytarabine regimen. Cancer 2004;
100:1179–1185.
18. Smith M, Arthur D, Camitta B, et al. Uniform approach to risk
classification and treatment assignment for children with acute
lymphoblastic leukemia. J Clin Oncol 1996;14:18–24.
19. Report of the expert committee on the diagnosis and classification
of diabetes mellitus. Diab Care 2003;26:S5–S20.
20. Gabir MM, Hanson RL, Dabelea D, et al. The 1997 American
Diabetes Association and 1999 World Health Organization criteria
for hyperglycemia in the diagnosis and prediction of diabetes. Diab
Care 2000;23:1108–1112.
21. Pui CH, Burghen GA, Bowman WP, et al. Risk factors for
hyperglycemia in children with leukemia receiving L-asparaginase
and prednisone. J Pediatr 1981;99:46–50.
22. Harris MI, Flegal KM, Cowie CC, et al. Prevalence of diabetes,
impaired fasting glucose, and impaired glucose tolerance in U.S.
Adults. The third national health and nutrition examination survey,
1988–1994. Diab Care 1998;21:518–524.
23. de Vegt F, Dekker JM, Jager A, et al. Relation of impaired
fasting and postload glucose with incident type 2 diabetes in a
Dutch population: The hoorn study. JAMA 2001;285:2109–2113.
24. Ceriello A, Hanefeld M, Leiter L, et al. Postprandial glucose
regulation and diabetic complications. Arch Intern Med 2004;
164:2090–2095.
25. Asensio C, Muzzin P, Rohner-Jeanrenaud F. Role of glucocorti-
coids in the physiopathology of excessive fat deposition and insulin
resistance. Int J Obes Relat Metab Disord 2004;28:S45–S52.
26. Meigs JB, Muller DC, Nathan DM, et al. The natural history of
progression from normal glucose tolerance to type 2 diabetes in
the Baltimore longitudinal study of aging. Diabetes 2003;52:
1475–1484.