Content uploaded by Andreas Thomas
Author content
All content in this area was uploaded by Andreas Thomas on Sep 11, 2018
Content may be subject to copyright.
ORIGINAL ARTICLE
Glycaemic variability in paediatric patients with type 1 diabetes
on continuous subcutaneous insulin infusion (CSII) or multiple
daily injections (MDI): a cross-sectional cohort study
C. Schreiver*, U. Jacoby*, B. Watzer†, A. Thomas‡, D. Haffner*
,
§
,1
and D.-C. Fischer*
,1
*Department of Paediatrics, University Hospital Rostock, †Department of Paediatrics, Philipps University Marburg, ‡Medtronic
GmbH, Meerbusch and §Department of Paediatric Kidney, Liver and Metabolic Diseases, Hannover Medical School, Hannover,
Germany
Summary
Objective This cross-sectional observational cohort study was
designed to investigate i) whether glycaemic variability in paedi-
atric patients with type 1 diabetes is lower in those using an
insulin pump (CSII) compared with those using multiple daily
insulin injections (MDI) and ii) whether urinary F
2
-isoprostanes
and/or urinary prostaglandin F
2
excretion as surrogate marker
of oxidative stress and cyclooxygenase activity are associated
with glycaemic variability.
Methods 48 paediatric patients with type 1 diabetes (22 using
an insulin pump) underwent an ambulatory 3-day continuous
glucose monitoring. All patients continued with normal daily
activities and collected urine for two consecutive 24 h periods.
The glucose pentagon was used to calculate the glycaemic risk
parameter.
Results Insulin requirements, HDL-cholesterol, the mean of
glycaemic excursions (P<001) and the standard deviation of
mean glucose concentration (P<005) were significantly lower
in patients with CSII compared with those using MDI. By con-
trast, averaged HbA
1c
during the last twelve months as well as at
the time of sensor insertion did not differ significantly between
both groups. Summarizing characteristic parameter of acute and
long-term metabolic control into the glucose pentagon revealed
a significantly lower glycaemic risk parameter in CSI patients
compared with both, healthy subjects and patients using MDI
(P<005).
Conclusions Paediatric patients with type 1 diabetes using an
insulin pump presented with lower glycaemic variability and a
concomitantly lower glycaemic risk parameter compared with
those using MDII. Whether these findings translate into a lower
risk of diabetes associated cardiovascular complications remains
to be elucidated.
(Received 11 September 2012; returned for revision 16 October
2012; accepted 1 November 2012)
Introduction
Patients suffering from type 1 diabetes are at a seriously elevated
age-adjusted relative risk of developing cardiovascular disease
when compared with their healthy peers or patients with type 2
diabetes.
1,2
This association is only partly due to traditional risk
factors (e.g. hypertension, dyslipidaemia and smoking), but
rather reflects the causal role of diabetic micro- and macroangi-
opathy.
3
Hyperglycaemia has been accused to impair endothelial
function, either directly or secondary to its effect on cells signal-
ling e.g. via the secretion of growth factors, cytokines and/or
vasoactive substances.
3
Endothelial cells, like mesangial and
Schwann cells, can regulate glucose uptake only within narrow
limits and become rather vulnerable if extracellular glucose levels
rise, i.e. extracellular hyperglycaemia translates rather quickly
into an intracellular one with subsequent severe metabolic
aberrations and finally increased oxidative stress.
4,5
The produc-
tion of F2-isoprostanes (FISO) is thought to reflect the free radi-
cal-catalysed peroxidation of arachidonic acid. Consequently, the
urinary concentration of FISO is used as marker of oxidative
stress, i.e. lipid peroxidation. However, there is strong evidence
that the formation of prostaglandin F
2
(PGF
2
) and probably
also other isoprostanes depends on cyclooxygenase activity, too.
6
Continuous subcutaneous insulin infusion (CSII) and multiple
daily insulin injections (MDI) are established therapies for type
1 diabetes and both are thought to be equally well suited to pre-
vent hyperglycaemia together with the deleterious consequences
outlined above. It is still a matter of discussion, whether CSII
therapy is superior to MDI with respect to glycaemic control.
While elevated levels of HbA
1c
indicate previous hyperglyca-
emia and overall poor metabolic control during the last three
months, real-time continuous glucose monitoring (CGM) is the
Correspondence: Dagmar-Christiane Fischer, PhD, Department of Paedi-
atrics, University Hospital Rostock, Ernst-Heydemann-Str. 8, 18057
Rostock, Germany. Tel.: +49 381 494 7041; Fax: +49 381 494 7044;
E-mail: dagmar-christiane.fischer@med.uni-rostock.de
1
These two authors contributed equally to this work.
©2012 John Wiley & Sons Ltd 641
Clinical Endocrinology (2013) 79, 641–647 doi: 10.1111/cen.12093
only way to assess more subtle aspects of glucose homeostasis,
e.g. nocturnal and/or asymptomatic hypoglycaemia.
7,8
We there-
fore used CGM in paediatric patients with type 1 diabetes on
CSII and MDI treatment to investigate i) whether glycaemic var-
iability is lower on CSII compared with MDI treatment and ii)
whether urinary F2-isoprostanes and/or urinary PGF
2a
excretion
as surrogate marker of oxidative stress and cyclooxygenase
activity are associated with glycaemic variability in these
patients.
Patients and methods
All children suffering from type 1 diabetes and being treated at
our institution were eligible for this cross-sectional observational
cohort study. Inclusion criteria: age 6–18 years, duration of
disease >1 year, C-peptide below 03 nmol/l, intensive insulin
treatment with either MDI or CSII, constant mode of insulin
therapy for at least 6 months. Exclusion criteria: any febrile ill-
ness during the last three months, chronic inflammatory-/rheu-
matic disease (e.g. Crohn’s disease, rheumatoid arthritis),
hepatitis, HIV, glucocorticoid treatment, liver-, renal- or cardiac
failure, hereditary dyslipidaemia, skin conditions prohibiting
sensor needle insertion, inborn or acquired errors of metabolic
disease except type 1 diabetes. Forty-eight patients (22 boys)
consented to participate and were enrolled between January and
December 2010. Patient characteristics are given in Table 1. CSII
treatment was introduced in our centre in 2007 and rarely used
as first line treatment before 2010. Thus, in this study, the
majority of patients on CSII treatment were switched from MDI
mainly to prevent recurrent hypoglycaemia and patients (or par-
ents) wish. In patients with CSII therapy (n=22) solely short-
acting insulins were used, while patients on MDI therapy
(n=26) received a combination of either short- and long-acting
insulins (n=8), short- and intermediate-acting insulins (n=9),
normal and long-acting insulins (n=5), or short-, normal and
long-acting insulins (n=4). Patients applied 4 to 5 insulin boli
per day. The study received appropriate Ethics committee
approval from the institutional review board and was performed
in accordance with the Declaration of Helsinki. Subjects and/or
their parents gave assent and written informed consent prior to
participation, i.e. to undergo continuous glucose monitoring
(CGM) for a 72-h period.
Procedures
All children and adolescents were seen in our outpatient clinic
in the afternoon. Demographic and clinical data including yearly
assessments of retinopathy by a trained ophthalmologist as well
as results from routine laboratory analysis were gathered by
interview and chart review, respectively. Furthermore, blood was
taken for determination of actual HbA
1c
, lipids and creatinine to
estimate the glomerular filtration rate (eGFR) according to the
Schwartz formula.
9
Pubertal stage was assessed according to
Tanner and blood pressure (BP) was measured according to the
updated Task Force Report on high blood pressure by using an
oscillometric device (Dinamap 1846SX; Critikon, Tampa, USA).
Individual age- (height) and gender-related SD scores (SDS) for
height, weight, BMI and BP were calculated as described.
10,11
Patients were classified as hypertensive in case of BP values
above the height and gender-related 95
th
percentile.
11
The patients and/or parents were instructed to the use of the
CGM (Guardian REAL-Time, Medtronic, Northridge, USA) and
Table 1. Patient characteristics and biochemical findings at initiation of
CGM
All patients
(22 m/26 f)
MDI
(14 m/12 f)
CSII
(8 m/14 f) P-Value
Age [years] 1293313130126360702
Male 1283313630114360162
Female 1293312530133350504
Height [SDS] 043 089 036 105 050 067 0694
Weight [SDS] 025 082 011 091 042 068 0214
BMI [SDS]
‡
004 102 -009 111 018 090 0535
Blood pressure
Systolic
[mmHg]
1175141 1187164 11601070475
Diastolic
[mmHg]
71297704947181020634
Systolic [SDS] 084 141 090 162 077 115 0468
Diastolic [SDS] 034 135 029 124 039 150 0569
Duration of
Diabetes
[months]
6344425553667295110320
Mode of
therapy
[months]
4123605553662432 277<0001
Average HbA
1c
[mmol
/month]
†
70 12 73 12 66 11 0079
Average HbA
1c
[%]
†
852 111 880 114 819 099 0079
Current HbA
1c
[mmol/mol]
71 19 74 22 68 13 0419
Current
HbA
1c
[%]
857 206 903 042 828 025 0419
Insulin dose
[IE/kg]
095 035 105 039 084 026 0022
Cholesterol
[mmol/l]
479 128 525 143 426 081 0003
HDL-Cholesterol
[mmol/l]
164 041 177 044 149 031 0017
LDL-Cholesterol
[mmol/l]
274 120 309 143 232 065 0018
LDL/HDL-ratio 176 090 189 113 161 050 0526
Lipoprotein
(a) [g/l]
028 007 038 012 015 003 0,324
eGFR
§
[ml/min
/173m
2
]
1534364 1596269 15172980396
Albuminuria
[mg/day]
126232149300901090732
Data are given as mean SD. *P<005; **P<001 between MDI and
CSII patients;
†
mean HbA
1c
during the last 12 months before initiation
of CGM;
‡
BMI-SDS calculated according to
10
and
§
eGFR, estimated glo-
merular filtration rate calculated according to Schwartz et al.
9
, BP, blood
pressure
©2012 John Wiley & Sons Ltd
Clinical Endocrinology (2013), 79, 641–647
642 C. Schreiver et al.
to enter a minimum of two blood glucose values per day for cal-
ibration. It should be noted that none of the patients enrolled in
this study had used CGM before. The sensor was placed sub-
cutaneously and in CSII patients contralaterally to the catheter
by the same investigator (CS) in all subjects. Patients on MDI
therapy were asked to inject insulin contralaterally to the site of
the sensor. Patients/parents kept a study diary and recorded
meals, symptomatic hypo- and hyperglycaemic events and sport
activities. The patients were asked to collect urine for determina-
tion of albumin, F
2
-isoprostanes and PGF
2
during two consecu-
tive 24-h periods. At the end of the 72-h observation period, the
CGM sensor was removed and data transfer was achieved using
the carelink personal therapy management software (Medtronic,
Northridge, USA).
Consecutive quantification of urinary F
2
-isoprostanes
and PGF
2
by GC-MS/MS
Urine for determination of F
2
-isoprostane and PGF
2
was avail-
able from 45 patients (25 on MDI therapy) and aliquots were
initially stored at 80 °C until quantitative determination of
FISO and PGF
2
by gas chromatography/triple quadrupole mass
spectrometry (GC-MS/MS) essentially as described previously.
12
Per urine an aliquot was spiked with deuterated internal PGF
2
standard (IS), acidified with formic acid to pH 2.6, extracted
with ethyl acetate/hexane and subsequently derivatized to the
correspondent pentafluorobenzyl esters. Samples were purified
by means of thin-layer chromatography with ethyl acetate/hex-
ane as developing solvent. A broad zone with R
f
0.03–0.34 was
scraped off, eluted with pure ethyl acetate and subjected to prep-
aration of trimethysilyl ethers. A Finnigan MAT TSQ700 mass
spectrometer (Thermo Electron, Dreieich, Germany) equipped
with a Varian 3400 gas chromatograph (Palo Alto, CA) and a
CTC A200S autosampler (Zwingen, Switzerland) was used. The
first quadrupole was set to monitor the [M-PFB]
ions (m/z
569 for PGF
2a
and m/z 573 for tetradeuterated IS). Collision cell
pressure was 0.2 Pa and collision energy was 12 eV). For F
2
-iso-
prostane determination, product ions [M-3TMSOH]
(m/z 299
and m/z 303 for the internal standard) were monitored. In a
second run for quantification of PGF
2
product ions [C
12
H
17
]
(m/z 161) and [C
122
H
4
H
13
]
(m/z 165 for the tetradeuterated
internal standard) and a collision energy of 16 eV were used.
Assessment of glycaemic control
Data from the 3-day CGM were used to calculate i) the mean of
the daily differences (MODD), ii) the continuous overlapping
net glycaemic action for one, two and four hours (CONGA-1,
-2, -4), and iii) the mean of glycaemic excursions (MAGE) as
described previously.
13,14
In addition, the frequency of mild and
severe hypoglycaemias, i.e. glucose concentrations below
3.9 mmol/l and 28 mmol/l, was counted. The mean glucose
concentration, the standard deviation of the mean glucose con-
centration, the mean length of time with glucose concentration
>888 mmol/l, the intensity of hyperglycaemia (area under the
curve (AUC) for glucose concentration >888 mmol/l) as well
as the HbA
1c
were used to generate the glucose pentagon and to
calculate the Glycaemic Risk Parameter (GRP) as described
before.
15
Statistics
Data were analysed using the SPSS statistical package 15.0 (SPSS
Inc. Chicago, Illinois, USA). Normal distribution was evaluated
by the Kolmogorow-Smirnow test and comparison between
groups were done using Student’s t-Test or Mann–Whitney-U
test, if appropriate. All P-values are two sided and a P-value
below 005 was considered significant. Data are given as
mean SD or median and range, if appropriate.
Results
At the time of enrolment, patients on MDI and CSII therapy
were quite comparable with respect to age, gender distribution,
BP and anthropometric values (Table 1). In addition, the num-
ber of prepubertal and pubertal patients were similar in both
groups (8/18 vs 9/13 prepubertal/pubertal patients on MDI and
CSII therapy, respectively; P=0469). Apart from one patient
on MDI therapy with already known and adequately treated
hypertension, six and one patient on MDI and CSII therapy
showed an elevated BP, respectively (P<005). Patients on CSII
therapy tended to present with longer duration of disease and to
use CSII for a shorter period of time compared with patients on
MDII therapy, although these differences did not reach statistical
significance. Both, time averaged HbA
1c
during the last twelve
months as well as at the time of sensor insertion tended to be
higher in patients on MDI compared with CSII therapy. Insulin
requirement and (HDL)-cholesterol levels were significantly
lower in CSII compared with MDI patients (each P<005).
Urinary albumin-to-creatinine ratios in spot urine were normal
(i.e. <300 lg/g) in all except one patient on MDI therapy, who
presented with a ratio of 603 lg/g. Likewise, albuminuria was
similar in both groups (Table 1). None of the patients smoked
and/or revealed signs of retinopathy.
Continuous glucose monitoring and glycaemic variability
CGM was well tolerated in all patients, none of the sensors
failed and per patient, the pre-defined set of characteristics was
calculated. Patients on MDI therapy revealed significantly higher
values for MAGE and standard deviation of mean glucose con-
centration (Table 2). Mean glucose concentration, mean length
of time and intensity of hyperglycaemia tended to be higher in
patients on MDI compared with CSII. In contrast, mean values
for MODD, CONGA-1, -2, and 4 as well as the frequencies of
mild or severe hypoglycaemias did not differ significantly
between groups. Moreover, glycaemic variability was almost
comparable in pubertal and prepubertal patients and even sport-
ing activities did not differ (data not shown). We combined the
data derived from CGM with the HbA
1c
concentration at time
of sensor insertion to calculate the glycaemic risk parameter,
i.e. the area of the glucose pentagon (Fig. 1). This approach
©2012 John Wiley & Sons Ltd
Clinical Endocrinology (2013), 79, 641–647
Glycemic variability in children: CSII vs MDI 643
revealed a significantly enlarged area in patients on MDI therapy
compared with both healthy subjects and patients on CSII ther-
apy (each P<005).
Excretion of FISO and PGF
2
The urinary concentrations of FISO and PGF
2
were significantly
correlated ((r =0769; P<005) and the urinary albumin con-
centration was associated with either one of these metabolites as
well (FISO: r =0406, P<001; PGF
2
:r=0339, P<005).
However, patients on MDI and CSII therapy did not differ with
respect to creatinine normalized FISO and PGF
2
excretion
respectively (FISO: 48730 ng/mg vs 49831 ng/mg and
PGF
2
:287 018 ng/mg vs 253 024 ng/mg). There was no
association between parameters of glycaemic variability and
either absolute or creatinine normalized urinary FISO and PGF
2
(data not shown).
Discussion
High glycaemic variability is a common feature in patients with
diabetes and has been accused to be the main responsible fac-
tor for co-morbidities, e. g. diabetic nephropathy and retinopa-
thy.
4
In general, self-monitoring of glucose two-to-four times
per day with concomitant adjustment of insulin dosages and
carbohydrate intake is used to ensure glucose homeostasis,
whereas HbA
1c
is used as an established marker of long-term
metabolic control. It is still a matter of discussion whether or
not the additional costs related to CSII are justified from a
medical point of view, i.e. whether CSII is superior to MDI in
terms of metabolic control and prevention of glycaemic excur-
sions.
16–23
Instead of individual glucose measurements several
hours apart, CGM appears to be ideally suited to detect “auto-
matically” otherwise unheralded extremes and fluctuations of
glucose concentrations during a reasonable period of times, i.e.
Table 2. Results of CGM in pediatric patients with type 1 diabetes
receiving MDI and CSII therapy
MDI (n=26) CSII (n=22)
MODD [mmol/l] 350 142 337 134
CONGA-1 [mmol/l] 081 045 090 047
CONGA-2 [mmol/l] 083 043 082 045
CONGA-4 [mmol/l] 085 041 101 061
MAGE [mmol/l] 650 221 496 162**
Number of measured
Mild hypoglycemias
(<39 mmol/l)
292 225 245 255
Severe hypoglycemias
(<28 mmol/l)
077 108 059 113
Mean glucose concentration
within 24 h [mmol/l]
929 211 864 177
SD of mean glucose
concentration [mmol/l]
315 096 254 078*
Mean length of time of
hyperglycemia
(>888 mmol/l) [h]
1093 518 908 477
Intensity of hyperglycemia
(AUC >888 mmol/l)
[mmol/l 924 h]
188 131 127 108
Glycemic risk parameter 400 229 284 137*
Data are given as mean SD. *P<005 and **P<001. MODD, mean
of the daily differences; CONGAn, continuous overlapping net glycaemic
action for nhours
Fig .1 The recently suggested glucose pentagon
reflecting the risk for diabetes-specific complications
in healthy individuals ( ) and in patients on MDI
() and CSII-therapy ( ). The axes were scaled as
suggested.
15
©2012 John Wiley & Sons Ltd
Clinical Endocrinology (2013), 79, 641–647
644 C. Schreiver et al.
to record glucose concentrations every one or five minutes for
at least 72 h.
7,24
Despite being a very appealing approach and
necessary step towards the development of closed loop systems,
one ends with huge amounts of data. The latter requires addi-
tional tools and algorithms to separate out and highlight infor-
mation relevant to answer the question at hand. Several indices
and statistical tools have been suggested for this purpose.
13,25,26
Among these, the mean and standard deviation of the glucose
concentration during the entire observation period are fairly
direct characteristics, whereas MODD and CONGAnas marker
of inter- and intraday glycaemic variation are more sophisti-
cated ones. Both MODD and CONGAn were quite similar in
children on MDI and CSII therapy respectively. This may at
least in part simply be due to the fact that our patients used
CGM under conditions of everyday life rather than during an
artificial hospital setting with highly standardized meals and
activities. Furthermore, it highlights the inherent danger of data
reduction, as those parameters directly calculated from the
entire set of data indicate significant differences between groups
(see below).
While our data for MODD are well in line with those
reported earlier for smaller groups of adult (n=28) and paedi-
atric (n=10) patients with type 1 diabetes, the CONGA values
in our study were about one-third of those described earlier.
The latter may be due to the fact that activities and meals
during everyday life in children are quite homogeneous com-
pared with adults.
13,14
Furthermore, none of the aforementioned
studies was conducted to decipher differences related to the
mode of therapy, but rather to compare patients and healthy
controls. There is one recent report indicating that CONGA-4
values are lower only in adults with type 1 diabetes on CSII
therapy with already good metabolic control (HbA
1c
58
mmol/mol ( 75%)). In those with poor metabolic control
(HbA
1c
>77 mmol/mol (>92%)), the results for patients on
CSII therapy were even worse compared with those on MDI
therapy.
27
In our patients, averaged and actual HbA
1c
values
indicate neither good nor poor metabolic control and even did
not differ between groups. In this regard, our findings are well
in line with the data reported by Lepore et al.
27
In contrast to MODD and CONGAn, the mean amplitude of
glycaemic excursions (MAGE) as an individually standardized
index of glucose fluctuations differed significantly between
patients on MDI and CSII therapy respectively. The “glucose
pentagon” has been suggested as a model to integrate different
parameters related to acute glycaemic variability (CGM data)
and long-term metabolic control (HbA
1c
).
15
Although all values
entered into this model only tended to be higher in patients
on MDI therapy, the GRP was significantly worse in patients
on MDI therapy compared with both healthy controls and
patients using CSII therapy. This result is remarkable, in that i)
the duration of disease tended to be longer in patients on CSII
therapy and that ii) apparently all patients in our study group
were started on MDI and a subgroup was switched to CSII
later. Keeping this in mind, the glucose pentagons and the
derived GRP indicate that on average two years of CSII are
sufficient to significantly reduce, but not normalize, the GRP.
In other words, switching from MDI to CSII therapy several
years after onset of disease has only limited potential to nor-
malize the GRP and this might be especially valid for those
patients showing initially poor glycaemic control with MDI
therapy.
Another parameter to be considered when speculating on gly-
caemic risk and complications of diabetes is oxidative stress. A
strong linear relationship between urinary 8-iso PGF
2
has been
described in a recent study with adults suffering from type 2
diabetes and being treated with oral anti-diabetic agents rather
than insulin.
28
By contrast, we and others were unable to show
any association between glycaemic variability and urinary excre-
tion of PGF
2
or FISO in paediatric and adult patients with type
1 diabetes.
14
But perhaps even more impressive is the notion
that our patients had approximately 100-fold higher urinary cre-
atinine normalized PGF
2
and FISO excretion than adults with
either form of diabetes. First of all, methodological problems
related to the quantitative determination of individual isoforms
of the PGF
2
family of prostaglandins have to be considered.
While a combination of solid-phase extraction and enzyme-
immunoassay was used by Monnier et al., we and Wentholt
et al. applied more sophisticated, although still slightly different
technologies, i.e. GC-MS/MS with pre-column derivatization
and immunoaffinity chromatography followed by HPLC-MS/
MS.
14,28
Both the extremely different concentrations of prosta-
glandins in adult and paediatric patients with type 1 diabetes on
the one hand, and the apparently missing link between glycae-
mic variability and oxidative stress in patients with type 1 and
type 2 diabetes on the other hand indicate that i) oxidative
stress is higher in paediatric patients with type 1 diabetes and
probably reflects more complex metabolic aberrations rather
than glycaemic variability, and/or ii) that the pathological mech-
anisms behind either diabetes form differ.
The present study has several limitations and the results might
be biased due to the fact that i) patients on CSII treatment
tended to have longer duration of diabetes than those using
MDI therapy, ii) some patients were previously transferred from
MDI to CSII therapy because of unstable diabetes and recurrent
hypoglycaemia and iii) metabolic differences between both
groups of patients are likely to be underestimated by the parallel
design. However, due to ethical reasons, a (randomized) cross-
over study design was not feasible in this paediatric patient
cohort. Indeed, children and their parents usually fight for the
permission, to use an insulin pump instead of MDI therapy and
one can hardly imagine that these parents and patients accept
the withdrawal of the pump and go back to MDI therapy.
Patients had to adapt to this “new” therapy for several months
before reapplication of CGM will yield reliable data on glycae-
mic variability.
In conclusion, this observational cohort study carried out
under “real-life” conditions revealed lower glycaemic variability
and glycaemic risk profile in paediatric patients with type 1 dia-
betes on CSII compared with MDI therapy. However, whether
on the long-run CSII is superior to MDI to delay the occurrence
of diabetes associated complications remains to be elucidated in
carefully designed prospective trials.
©2012 John Wiley & Sons Ltd
Clinical Endocrinology (2013), 79, 641–647
Glycemic variability in children: CSII vs MDI 645
Acknowledgments
We thank all families, especially the children, who participated
in this study and gratefully acknowledge the support of our col-
leagues. The outstanding technical assistance of Susan Schilling,
Birgit Salewski and Anja Rahn (all Department of Paediatrics,
Rostock) is highly appreciated.
Funding
The sensors were made available free of charge by Medtronic,
Meerbusch Germany and this company funded our study by an
unrestricted grant.
Declaration of competing interests
Nothing to declare.
Authors Contribution
CS and UJ were responsible for patient care, collection and anal-
ysis of data and preparation of the manuscript. BW measured
the FISO and PGF
2
concentrations, AT participated in data
analysis and interpretation of the glucose pentagon. DH and
DCF (guarantors) were responsible for the study concept
and design, supervised the study, participated in data analysis
and interpretation, reviewed and edited the manuscript.
References
1 Libby, P., Nathan, D.M., Abraham, K. et al. (2005) Report of the
National Heart, Lung, and Blood Institute-National Institute of
Diabetes and Digestive and Kidney Diseases Working Group on
Cardiovascular Complications of Type 1 Diabetes Mellitus. Cir-
culation,111, 3489–3493.
2 Giannini, C., Mohn, A., Chiarelli, F. et al. (2011) Macrovascular
angiopathy in children and adolescents with type 1 diabetes. Dia-
betes Metab Res Rev,27, 436–460.
3 Schalkwijk, C.G. & Stehouwer, C.D. (2005) Vascular complica-
tions in diabetes mellitus: the role of endothelial dysfunction.
Clin Sci (Lond),109, 143–159.
4 Brownlee, M. (2005) The pathobiology of diabetic complications:
a unifying mechanism. Diabetes,54, 1615–1625.
5 Colette, C. & Monnier, L. (2007) Acute glucose fluctuations and
chronic sustained hyperglycemia as risk factors for cardiovascular
diseases in patients with type 2 diabetes. Hormone and Metabolic
Research,39, 683–686.
6Schweer,H.,Watzer,B.,Seyberth,H.W.et al. (1997) Improved
quantification of 8-epi-prostaglandin F2 alpha and F2-isoprostanes
by gas chromatography/triple-stage quadrupole mass spectrometry:
partial cyclooxygenase-dependent formation of 8-epi-prostaglandin
F2 alpha in humans. Journal of Mass Spectrometry,32,1362–1370.
7 Danne, T., Lange, K. & Kordonouri, O. (2008) Real-time glucose
sensors in children and adolescents with type-1 diabetes. Hor-
mone Research,70, 193–202.
8 Tamborlane, W.V., Beck, R.W., Bode, B.W. et al. (2008) Contin-
uous glucose monitoring and intensive treatment of type 1 dia-
betes. New England Journal of Medicine,359, 1464–1476.
9 Schwartz, G.J., Mu~
noz, A., Schneider, M.F. et al. (2009) New
equations to estimate GFR in children with CKD. Journal of the
American Society of Nephrology,20, 629–637.
10 Kromeyer-Hauschild, K., Wabitsch, M., Kunze, D. et al. (2001)
Percentiles of body mass index in children and adolescents eval-
uated from different regional German studies. Monatsschr Kin-
derh,149, 807–818.
11 National High Blood Pressure Education Program Working
Group on High Blood Pressure in Children and Adolescents.
(2004) The fourth report on the diagnosis, evaluation, and treat-
ment of high blood pressure in children and adolescents. Pediat-
rics 114, 555-576.
12 Schweer, H., Watzer, B. & Seyberth, H.W. (1994) Determination
of seven prostanoids in 1 ml of urine by gas chromatography-
negative ion chemical ionization triple stage quadrupole mass
spectrometry. Journal of Chromatography,652, 221–227.
13 McDonnell, C.M., Donath, S.M., Vidmar, S.I. et al. (2005) A
novel approach to continuous glucose analysis utilizing glycemic
variation. Diabetes Technol Ther,7, 253–263.
14 Wentholt, I.M., Kulik, W., Michels, R.P. et al. (2008) Glucose
fluctuations and activation of oxidative stress in patients with
type 1 diabetes. Diabetologia,51, 183–190.
15 Thomas, A., Schonauer, M., Achermann, F. et al. (2009) The
“glucose pentagon”: assessing glycemic control of patients with
diabetes mellitus by a model integrating different parameters
from glucose profiles. Diabetes Technol Ther,11, 399–409.
16 Hanas, R. & Adolfsson, P. (2006) Insulin pumps in pediatric
routine care improve long-term metabolic control without
increasing the risk of hypoglycemia. Pediatric diabetes,7,
25–31.
17 Jeitler, K., Horvath, K., Berghold, A. et al. (2008) Continuous
subcutaneous insulin infusion versus multiple daily insulin injec-
tions in patients with diabetes mellitus: systematic review and
meta-analysis. Diabetologia,51, 941–951.
18Derosa,G.,Maffioli,P.,D’Angelo,A.et al. (2009) Effects of
insulin therapy with continuous subcutaneous insulin infusion
(CSII) in diabetic patients: comparison with multi-daily
insulin injections therapy (MDI). Endocrine Journal,56,
571–578.
19 Fatourechi, M.M., Kudva, Y.C., Murad, M.H. et al. (2009) Clini-
cal review: Hypoglycemia with intensive insulin therapy: a sys-
tematic review and meta-analyses of randomized trials of
continuous subcutaneous insulin infusion versus multiple daily
injections. Journal of Clinical Endocrinology and Metabolism,94,
729–740.
20 Jakisch, B.I., Wagner, V.M., Heidtmann, B. et al. (2008) Com-
parison of continuous subcutaneous insulin infusion (CSII) and
multiple daily injections (MDI) in paediatric Type 1 diabetes: a
multicentre matched-pair cohort analysis over 3 years. Diabetic
Medicine,25,80–85.
21 Karagianni, P., Sampanis, C., Katsoulis, C. et al. (2009) Continu-
ous subcutaneous insulin infusion versus multiple daily injec-
tions. Hippokratia,13,93–96.
22 Monami, M., Lamanna, C., Marchionni, N. et al. (2010) Contin-
uous subcutaneous insulin infusion versus multiple daily insulin
injections in type 1 diabetes: a meta-analysis. Acta Diabetologica,
47,77–81.
23 Weissberg-Benchell, J., Antisdel-Lomaglio, J. & Seshadri, R.
(2003) Insulin pump therapy: a meta-analysis. Diabetes Care,26,
1079–1087.
©2012 John Wiley & Sons Ltd
Clinical Endocrinology (2013), 79, 641–647
646 C. Schreiver et al.
24 Mastrototaro, J., Shin, J., Marcus, A. et al. (2008) The accuracy
and efficacy of real-time continuous glucose monitoring sensor
in patients with type 1 diabetes. Diabetes Technol Ther,10,
385–390.
25 Clarke, W. & Kovatchev, B. (2009) Statistical tools to analyze
continuous glucose monitor data. Diabetes Technol Ther,11(Sup-
pl 1), S45–54.
26 Rodbard, D. (2009) Interpretation of continuous glucose moni-
toring data: glycemic variability and quality of glycemic control.
Diabetes Technol Ther,11(Suppl 1), S55–67.
27 Lepore, G., Corsi, A., Dodesini, A.R. et al. (2010) Continuous
subcutaneous insulin infusion is better than multiple daily insu-
lin injections in reducing glucose variability only in type 1 diabe-
tes with good metabolic control. Diabetes Care,33, e81.
28 Monnier, L., Mas, E., Ginet, C. et al. (2006) Activation of oxida-
tive stress by acute glucose fluctuations compared with sustained
chronic hyperglycemia in patients with type 2 diabetes. JAMA,
295, 1681–1687.
©2012 John Wiley & Sons Ltd
Clinical Endocrinology (2013), 79, 641–647
Glycemic variability in children: CSII vs MDI 647
