Clinical Care/Education/Nutrition/Psychosocial Research

O R I G I N A L A R T I C L E

Effects of Carbohydrate Counting on

Glucose Control and Quality of Life Over
24 Weeks in Adult Patients With Type 1
Diabetes on Continuous Subcutaneous
Insulin Infusion
A randomized, prospective clinical trial (GIOCAR)
ANDREA LAURENZI, MD1 ELENA PERETTI, MD2 the type of daily carbohydrate intake
ANDREA M. BOLLA, MD1 ALESSANDRO SAIBENE, MD2 (4–6).
GABRIELLA PANIGONI, RD2 GABRIELLA GALIMBERTI, MD2 Over the years, a number of methods
VALENTINA DORIA, MD2 EMANUELE BOSI, MD1 have been proposed to help patients with
2
ANNACHIARA UCCELLATORE, MD MARINA SCAVINI, MD, PHD2
diabetes to quantify the carbohydrate
content of a meal in real life, for example,
exchange lists, portion/servings, grams,
OBJECTIVE—Few studies have assessed the efficacy of carbohydrate counting in type 1 diabetes,
and none have validated its efficacy in patients who are treated with continuous subcutaneous
glycemic index, and insulin:carbohydrate
insulin infusion (CSII). The aim of our study was to test the effect of carbohydrate counting on ratio (I:CHO) (7–9). Among these, the
glycemic control and quality of life in adult patients with type 1 diabetes who are receiving CSII. I:CHO is considered the most advanced
counting technique, consisting of esti-
RESEARCH DESIGN AND METHODS—Sixty-one adult patients with type 1 diabetes mating the grams of carbohydrates in a
treated with CSII were randomly assigned to either learning carbohydrate counting (interven- meal and then calculating the pre-meal
tion) or estimating pre-meal insulin dose in the usual empirical way (control). At baseline and 12 insulin dose based on this estimation
and 24 weeks, we measured HbA1c, fasting plasma glucose, BMI, waist circumference, recorded and an insulin sensitivity measure (8).
daily insulin dose, and capillary glucose data, and administered the Diabetes-Specific Quality-of-
Life Scale (DSQOLS) questionnaire.
Carbohydrate counting was devised in
the 1960s, but it has become widely
RESULTS—Intention-to-treat analysis showed improvement of the DSQOLS score related to used as part of intensive diabetes manage-
diet restrictions (week 24 – baseline difference, P = 0.008) and reduction of BMI (P = 0.003) and ment after the Diabetes Control and Com-
waist circumference (P = 0.002) in the intervention group compared with control subjects. No plications Trial (DCCT) (10,11).
changes in HbA1c, fasting plasma glucose, daily insulin dose, and hypoglycemic episodes (,2.8 Although carbohydrate counting is
mmol/L) were observed. Per-protocol analysis, including only patients who continuously used widely used by patients worldwide, few
carbohydrate counting and CSII during the study, confirmed improvement of the DSQOLS score studies have validated its efficacy in type
and reduction of BMI and waist circumference, and showed a significant reduction of HbA1c
(20.35% vs. control subjects, P = 0.05).
1 diabetes (12,13), and none have vali-
dated its efficacy in adult patients receiv-
CONCLUSIONS—Among adult patients with type 1 diabetes treated with CSII, carbohydrate ing continuous subcutaneous insulin
counting is safe and improves quality of life, reduces BMI and waist circumference, and, in per- infusion (CSII). We designed the current
protocol analysis, reduces HbA1c. study with the aim of testing the effect of
carbohydrate counting on glucose con-
Diabetes Care 34:823–827, 2011 trol and quality of life over 24 weeks in
adult patients with type 1 diabetes treated

N
utritional management is a corner- Some studies have examined the con- with CSII.
stone in the management of diabe- tribution of quantity and type (i.e., simple
tes, and monitoring of carbohydrate vs. complex) of carbohydrates in patients RESEARCH DESIGN AND
intake, a major determinant of postprandial with type 1 diabetes and showed that METHODS—The GIOCAR (contegGIO-
blood glucose, is a key strategy for achiev- the daily insulin requirement is indeed CARboidrati) was designed as a prospective,
ing good glucose control (1–4). associated with the amount rather than randomized, controlled, open-label clinical
trial with a duration of 24 weeks. The study
c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c was approved by the Ethics Committee of
From the 1San Raffaele Vita-Salute University, Milan, Italy; and the 2Diabetes and Endocrinology Unit, De-
the San Raffaele Scientific Institute in Milan
partment of Internal Medicine, San Raffaele Scientific Institute, Milan, Italy. and was registered at ClinicalTrials.gov
Corresponding author: Emanuele Bosi, bosi.emanuele@hsr.it. (no. NCT01173991). After having received
Received 2 August 2010 and accepted 25 January 2011. detailed information about the study and
DOI: 10.2337/dc10-1490. Clinical trial reg. no. NCT01173991, clinicaltrials.gov. before any study procedure, participants
A.L. and A.M.B. contributed equally to this work.
© 2011 by the American Diabetes Association. Readers may use this article as long as the work is properly signed a written informed consent.
cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ We recruited adult patients with type
licenses/by-nc-nd/3.0/ for details. 1 diabetes treated with CSII and followed

care.diabetesjournals.org DIABETES CARE, VOLUME 34, APRIL 2011 823

Carbohydrate counting in patients on CSII

at the CSII Outpatient Clinic of the San (group 1 intervention, group 2 control counting (,75% of the meals) (six partic-
Raffaele Scientific Institute in Milan. We subjects) with a 1:1 ratio. An investigator ipants) or shift from CSII to multiple in-
included patients with type 1 diabetes, without contact with study participants sulin injections for .7 consecutive days
aged 18–65 years, who had been treated generated the treatment allocation se- (two participants in the carbohydrate
with CSII for .3 months. Exclusion cri- quence using a computerized random counting group, one participant in the
teria were serum creatinine .124 mmol/L number generator (Stata, version 10.0; control group).
in women and .150 mmol/L in men, pre- Stata Corp, College Station, TX). Because Baseline characteristics of study
vious training in carbohydrate counting, of the type of intervention, blinding was participants in the two groups were
celiac disease, pregnancy, severe comor- not possible. Patients were given the compared using the x2 test, unpaired,
bidities, and any disability preventing same glucose meter (OneTouch Ultra2; two-tailed t test, or Mann–Whitney two-
compliance with study procedures. Three LifeScan Inc., Milpitas, CA) for self- sample statistic as appropriate. Changes
diabetologists with experience in manag- monitoring of blood glucose during the from baseline of DSQOLS scores, BMI
ing patients on CSII and trained in carbo- study period and were asked to measure and waist circumference, total daily insu-
hydrate counting and one dietitian capillary glucose six times per day, accord- lin dose, fasting plasma glucose, LBGI,
certified in carbohydrate counting con- ing to American Diabetes Association and HBGI in the two groups were com-
ducted the study. We used the Complete Standards of Medical Care (4). Before ran- pared using the unpaired, two-tailed t test
Guide to Carb Counting (2nd ed.) (14) as a domization, all participants attended a or the Mann-Whitney two-sample statis-
reference for carbohydrate counting. group lesson with the dietitian about the tic, as appropriate. HbA1c levels and hy-
Learning carbohydrate counting involves recommended diet for patients with dia- poglycemic events during the study in the
several steps. The first step is keeping a betes. After randomization, patients in two groups of participants were analyzed
food diary: complete food records include group 1 (intervention) were trained on using mixed-effects models.
day of the week, meal time, amounts of carbohydrate counting and bolus calcula-
food, carbohydrate grams for each food, tion in the first 12 weeks using the I:CHO RESULTS—The clinical trial was car-
total carbohydrate grams for the meal or and sensitivity factor during four to five in- ried out between October 2008 and July
snack, preprandial and postprandial (2 h dividual sessions with the dietitian and a 2009. The flow diagram of the study is
after the start of the meal) blood glucose, diabetologist, whereas patients in group 2 shown in Fig. 1. Of 67 patients assessed
short-acting insulin dose, and physical ac- (control subjects) continued estimating for eligibility, 61 were randomized and 56
tivity. The I:CHO tells how much insulin their pre-meal insulin dose in an empirical concluded the study (28 in each group),
is needed to “cover” the amount of carbo- way. HbA1c and fasting plasma glucose with a dropout rate of 8.2%. Patients as-
hydrates eaten and bring blood glucose were measured at baseline and after 12 signed to group 1 attended on average 4.4
level back to pre-meal target (14,15). and 24 weeks. At baseline and after 24 (SD 1.13) individual training sessions on
This ratio is calculated on the basis of in- weeks, we measured BMI and waist cir- carbohydrate counting.
dividual recorded diary data by dividing cumference, recorded total daily insulin The baseline characteristics of study
the total grams of carbohydrates of a meal dose, and asked patients to complete a val- participants are shown in Table 1. The
by the number of units of short-acting in- idated instrument for assessing diabetes- two groups were similar in age, sex, years
sulin that were able to hold post-meal glu- specific quality of life (DSQOLS) (16). of school completed, duration of diabetes,
cose excursions within 1.6 mmol/L. The Capillary glucose measurements were duration of CSII, type of insulin used,
sensitivity factor or correction factor is cal- downloaded from the memory of glucose daily insulin requirement, and HbA 1c
culated by dividing 1,800 by the total daily meters at 12 and 24 weeks at the time of levels.
insulin requirement (14,15) and corre- the outpatient visits, and LBGI and HBGI
sponds to the glucose lowering obtained were calculated as reported (17). Study Metabolic control
with one unit of short-acting insulin. By data were recorded on a paper Case Report In the ITT analysis, HbA1c levels during
combining the I:CHO and sensitivity fac- Form and then entered in a dedicated the 24 weeks of the study were similar in
tor, patients are instructed to estimate the database maintained in Microsoft Office the two groups (P = 0.252). However, the
preprandial insulin dose, taking into con- Access (Microsoft Corp., Redmond, WA), PP analysis showed significantly lower
sideration preprandial blood glucose and and de-identified datasets were extracted HbA1c levels in the carbohydrate count-
the amount of carbohydrates they plan for statistical analyses. HbA1c was measured ing group than in control subjects (in-
to eat. using ion-exchange high-performance tervention group 20.4 vs. 20.05% in
The primary outcome of the study liquid chromatography (DCCT-certified control subjects; Δ 20.35%, P = 0.05)
was the change in HbA1c at week 24. Sec- method) (18), with a normal range of (Fig. 2). BMI change was not a signifi-
ondary outcomes were the changes of the 3.5–6.0%. cant predictor of HbA1c. No differences
following variables at week 24: quality of between groups were observed in total
life, assessed with the Diabetes-Specific Statistical analysis daily insulin dose, LBGI, HBGI, and
Quality-of-Life Scale (DSQOLS) question- The intention-to-treat (ITT) analysis in- fasting plasma glucose levels (data not
naire (16), BMI and waist circumference, cluded all randomized patients who con- shown).
hypoglycemic events (capillary glucose cluded the trial, i.e., 56 patients (28
2.8 mmol/L), hypoglycemia and hyper- patients per group). The per-protocol Anthropometrics
glycemia risk indexes (low blood glucose (PP) analysis included 20 patients in the The median changes in BMI and waist
index [LBGI] and high blood glucose in- carbohydrate counting group and 27 circumference for the two groups are
dex [HBGI]) (17), total daily insulin dose, patients in the control group. For this shown in Table 2. Among patients in the
and fasting plasma glucose. Participants analysis, we excluded nine patients because carbohydrate counting group, we ob-
were randomly assigned to two groups of discontinuous use of carbohydrate served a significant reduction in BMI

824 DIABETES CARE, VOLUME 34, APRIL 2011 care.diabetesjournals.org

 Laurenzi and Associates

CONCLUSIONS—This study reports

the results of the first randomized clinical
trial testing the effects of carbohydrate
counting in adult patients with type 1
diabetes treated with CSII. Carbohydrate
counting improved the DSQOLS score
related to diet restrictions and was associ-
ated with a modest, although significant,
decrease in BMI and waist circumference.
When patients who did not continuously
use carbohydrate counting or CSII during
the study were excluded from the analyses,
carbohydrate counting was also associated
with a significant reduction of HbA1c not
accompanied by an increase of hypoglyce-
mic events.
Patients with type 1 diabetes must con-
tinuously match capillary glucose levels,
food intake, and physical activity with
the administration of the appropriate dose
of exogenous insulin to maintain glucose
levels within the recommended target (4).
Dietary restrictions and the stress associ-
ated with the daily management of diabe-
tes have a negative impact on the quality
of life of these patients (19). Carbohy-
drate counting is a tool that helps patients
to estimate in a systematic way the
amount of the pre-meal insulin bolus to
minimize the glucose increase after a
meal, and if necessary, to correct an either
inappropriately high or low pre-meal glu-
cose level.
The improvement of the DSQOLS
score related to diet restrictions that we
observed in our study extends previous
findings by Trento et al. (12) to patients
with type 1 diabetes treated with CSII,
although we did not observe an improve-
ment in HbA1c with carbohydrate count-
ing in our ITT analysis, possibly because
our study was relatively short (6 months)
Figure 1—Flow diagram for the GIOCAR trial. or not all participants trained in carbohy-
drate counting used it. However, when
only patients who continuously used car-
bohydrate counting for the daily manage-
and waist circumference compared with group compared with control subjects. ment of their diabetes were included in
control subjects. The significant differ- The significant difference in the diet re- the analyses, we observed a significant
ence in BMI and waist circumference per- strictions score between the two groups of improvement in HbA1c (20.35%) com-
sisted in the PP analysis (BMI: P = 0.020; participants persisted in the PP analysis pared with control subjects. The fact
waist circumference: P = 0.007). (P = 0.004). that one of five participants who learned
carbohydrate counting did not use it to
Quality of life Adverse events estimate their meal boluses suggests
The median changes in the score for the The frequency of hypoglycemic events that, although relatively simple, this
seven sections of the DSQOLS instrument (i.e., capillary glucose 2.8 mmol/L) was method may be difficult to implement
between the two groups are shown in similar in the two groups in both the ITT for a non-negligible proportion of pa-
Table 2. Baseline scores were similar in and the PP analyses. No episodes of severe tients with type 1 diabetes.
the two groups of participants. At week hypoglycemia requiring assistance from a In our study, the observed improve-
24, we observed a significant increase third party were observed during the ment in metabolic control among patients
(meaning a better quality of life) in the study. One patient in the control group using carbohydrate counting in the daily
scores related to diet restrictions among was diagnosed with painful diabetic neu- management of their diabetes was achieved
participants in the carbohydrate counting ropathy during the study. without the increase of body weight, waist

care.diabetesjournals.org DIABETES CARE, VOLUME 34, APRIL 2011 825

Carbohydrate counting in patients on CSII

Table 1—Baseline characteristics of study participants by the allocated treatment group with the dietitian about the recommended
(ITT analysis) diet for patients with diabetes, with the only
difference between the two groups being
Carbohydrate counting Control subjects learning and using carbohydrate counting.
(n = 28) (n = 28) P value A decrease in BMI with carbohydrate count-
ing, although not significant, also was ob-
Female participants 13 (46.4%) 19 (67.9%) 0.105 served in the study by Trento et al. (21), in
Age (years)* 41.2 6 10.0 39.8 6 9.8 0.601 which the control group had similar expo-
Years of school completed† 14 (10–18) 13 (13–15.5) 0.840 sure to the diabetes care team. We suggest
Duration of diabetes (years)* 21.9 6 11.0 19.8 6 11.7 0.490 that carbohydrate counting may provide
Duration of pump therapy (years)† 2 (1–3) 2 (0–3.5) 0.796 users with some additional benefits that fa-
Type of insulin cilitate weight loss, most likely through im-
Glulisine 14 (50.0%) 17 (60.7%) 0.340 proved nutrition or increased physical
Lispro 12 (42.9%) 7 (25.0%) activity.
Aspart 2 (7.1%) 4 (14.3%) Our study has several strengths. First
Insulin requirement (IU/day)† the sample size is at least as large as that of
Total 36 (24.5–49) 33 (28.5–39.5) 0.282 previously published research involving
Basal 22.5 (15–26) 18.5 (14–22) 0.268 patients with type 1 diabetes who receive
Boluses 15 (10.5–21.5) 12.5 (10–20.5) 0.522 multiple daily injections. Second, patients
BMI (kg/m2)† 23.7 (21–25.2) 23.8 (20.8–26.8) 0.670 were provided training in carbohydrate
Waist circumference (cm)† 83 (78.5–91) 78 (74–85.5) 0.194 counting that is feasible in the setting of a
Glycated hemoglobin (%)*^ 7.9 6 0.9 8.1 6 1.5 0.526 diabetes clinic. On the other hand, our
Categorical variables are presented as frequency with percent in parentheses. *Continuous variables with a study has some limitations. First, patients
normal distribution are presented as mean with SD in parentheses. †Continuous variables that do not have in the intervention group had more con-
a normal distribution are presented as median with the interquartile range in parentheses. ^Normal range
3.5–6.0%.
tact with the diabetes care team during the
teaching of carbohydrate counting, thus
preventing us from ruling out that the
intervention group lost weight and im-
circumference, or frequency of hypogly- unexpectedly observed a small, although proved metabolic control secondary to
cemic events usually reported with inten- significant, weight loss, for which we have the extra attention, rather than the use of
sive diabetes management (10,20). Indeed, no obvious explanation. At baseline, all ran- carbohydrate counting. However, several
in the carbohydrate counting group we domized patients attended a group lesson evidences support a direct effect of car-
bohydrate counting on the improvement
of glucose control: 1) in our study a sig-
nificant improvement in HbA1c was ob-
served only in the PP analysis, which
included only those participants who in-
deed use carbohydrate counting to esti-
mate their meal boluses; 2) a similar
decrease in HbA1c after learning carbohy-
drate counting was observed in a study
with a control group with similar expo-
sure to the diabetes care team (21); and 3)
the study patients in the control group
were long-term attendants of our CSII
outpatient clinic, making it unlikely
that a transient increase in contact with
the diabetes care team could significantly
affect their diet and diabetes manage-
ment. Moreover, our study has a relatively
short duration, not allowing the assess-
ment of the effects of carbohydrate count-
ing in the long-term, as it would be
desirable for a lifetime intervention. Fi-
nally, we did not measure physical activ-
ity and food intake during the study, not
allowing us to assess their contribution
to weight loss and improved metabolic
control.
Figure 2—PP analysis: HbA1c levels (mean and 95% CI) in the two study groups during the In conclusion, our study shows that
GIOCAR trial. The carbohydrate counting group (◆) had significantly lower HbA1c levels than offering carbohydrate counting to pa-
the control group (◇) (P = 0.050). tients with type 1 diabetes treated with

826 DIABETES CARE, VOLUME 34, APRIL 2011 care.diabetesjournals.org

 Laurenzi and Associates

Table 2—Changes from baseline at week 24 in DSQOLS scores, BMI, and waist 10. The Diabetes Control and Complications
circumference by the allocated treatment group (ITT analysis) Trial Research Group. The effect of intensive
treatment of diabetes on the development
and progression of long-term complications
Carbohydrate counting Control subjects in insulin-dependent diabetes mellitus.
(n = 28) (n = 28) P value N Engl J Med 1993;329:977–986
11. Anderson EJ, Richardson M, Castle G,
DSQOLS scores et al.; The DCCT Research Group. Nutri-
Social relations 2 (22.5 to 3.5) 0 (21.5 to 5) 0.993 tion interventions for intensive therapy in
Leisure-time flexibility 20.5 (22 to 1) 0 (22 to 3) 0.413 the Diabetes Control and Complications
Physical complaints 2 (0–4.5) 2 (20.5 to 5) 0.483 Trial. J Am Diet Assoc 1993;93:768–772
Worries about future 1 (21 to 4) 0 (21.5 to 3) 0.466 12. Trento M, Borgo E, Kucich C, et al. Quality
Diet restrictions 5.5 (0.5–8.5) 0 (22 to 3.5) 0.008 of life, coping ability, and metabolic control
Daily hassles 1.5 (22.5 to 6) 2 (21.5 to 3.5) 0.488 in patients with type 1 diabetes managed by
Fears about hypoglycemia 0.5 (22 to 7.5) 1 (25.5 to 5.5) 0.643 group care and a carbohydrate counting
BMI (kg/m2) 20.32 (20.65 to 0) 0.15 (0–0.40) 0.003 program (Letter). Diabetes Care 2009;32:
e134
Waist circumference (cm) 21 (22 to 0) 0 (0–2) 0.002
13. Mehta SN, Quinn N, Volkening LK, Laffel
Data are presented as median with the interquartile range in parentheses. LM. Impact of carbohydrate counting on
glycemic control in children with type 1 di-
CSII improves quality of life related to abetes. Diabetes Care 2009;32:1014–1016
diet restrictions and obtains a modest, References 14. Warshaw HS, Kulkarni K. Complete Guide
1. American Diabetes Association. Standards to Carb Counting. 2nd ed. Alexandria, VA,
although significant, decrease in BMI and of medical care in diabetes—2009. Di- American Diabetes Association, 2004
waist circumference. A reduction of HbA1c, abetes Care 2009;32(Suppl. 1):S13–S61 15. Davidson P. Bolus & supplemental in-
not accompanied by an increase in hypo- 2. Pastors JG, Warshaw H, Daly A, Franz M, sulin. In The Insulin Pump Therapy Book -
glycemic events, may be expected when Kulkarni K. The evidence for the effec- Insights From The Experts. Fredrickson L,
patients continuously use carbohydrate tiveness of medical nutrition therapy in Ed. Sylmar, CA, Minimed Technologies,
counting in the daily management of their diabetes management. Diabetes Care 1995, p. 58–72
diabetes. 2002;25:608–613 16. Bott U, Mühlhauser I, Overmann H,
3. DAFNE Study Group. Training in flexible, Berger M. Validation of a diabetes-specific
intensive insulin management to enable quality-of-life scale for patients with type 1
dietary freedom in people with type 1 diabetes. Diabetes Care 1998;21:757–769
Acknowledgments—This study was sup- diabetes: dose adjustment for normal 17. Kovatchev BP, Cox DJ, Gonder-Frederick
ported by an unrestricted educational grant eating (DAFNE) randomised controlled LA, Clarke W. Symmetrization of the blood
from GlaxoSmithKline. No other potential con- trial. BMJ 2002;325:746 glucose measurement scale and its applica-
flicts of interest relevant to this article were 4. American Diabetes Association. Executive tions. Diabetes Care 1997;20:1655–1658
reported. summary: standards of medical care in 18. Goodall I. HbA1c standardisation
A.L. conducted the study and had primary diabetes—2010. Diabetes Care 2010;33 destination—global IFCC Standardisation.
responsibility in patient care. A.M.B. contrib- (Suppl. 1):S4–S10 How, why, where and when—a tortuous
uted to patient care, data analysis, and article 5. Halfon P, Belkhadir J, Slama G. Correla- pathway from kit manufacturers, via inter-
editing. G.P. was responsible for carbohydrate tion between amount of carbohydrate in laboratory lyophilized and whole blood
counting teaching and dietary education. V.D., mixed meals and insulin delivery by arti- comparisons to designated national com-
A.C.U., and E.P. contributed to patient care and ficial pancreas in seven IDDM subjects. parison schemes. Clin Biochem Rev 2005;
education. A.S. contributed to carbohydrate Diabetes Care 1989;12:427–429 26:5–19
counting teaching and education. G.G. had 6. Franz MJ. Carbohydrate and diabetes: is 19. Rubin RR, Peyrot M. Quality of life and
primary responsibility in patient care. E.B. the source or the amount of more im- diabetes. Diabetes Metab Res Rev 1999;
conceived and coordinated the study, and wrote portance? Curr Diab Rep 2001;1:177–186 15:205–218
the article. M.S. had primary responsibility in 7. Chiesa G, Piscopo MA, Rigamonti A, et al. 20. UK Prospective Diabetes Study (UKPDS)
data analysis and wrote the article. Insulin therapy and carbohydrate counting. Group. Intensive blood-glucose control with
Parts of this study were presented in poster Acta Biomed 2005;76(Suppl. 3):44–48 sulphonylureas or insulin compared with
form at the 70th Scientific Sessions of the 8. Gillespie SJ, Kulkarni KD, Daly AE. Using conventional treatment and risk of compli-
American Diabetes Association, Orlando, Florida, carbohydrate counting in diabetes clinical cations in patients with type 2 diabetes
25–29 June 2010. practice. J Am Diet Assoc 1998;98:897–905 (UKPDS 33). Lancet 1998;352:837–853
The authors thank Dr. Stefania Del Rosso, 9. Pastors JG, Waslaski J, Gunderson H. Di- 21. Trento M, Trinetta A, Borgo E, et al. Car-
Laboraf, for support with clinical laboratory abetes meal-planning strategies. In Diabetes bohydrate counting improves coping
tests; Giorgio Antopulos, Laboraf, for support Medical Nutrition Therapy and Education. ability and metabolic control in patients
with data management of laboratory results; Ross TA, Boucher JL, O’Connell BS, Eds. with type 1 diabetes managed by Group
and Silvano Berni, Nesh.it, for graphic ex- Chicago, IL, American Dietetic Associa- Care. J Endocrinol Invest. 3 May 2010
pertise. tion, 2005, p. 201–217 [Epub ahead of print]

care.diabetesjournals.org DIABETES CARE, VOLUME 34, APRIL 2011 827