Severe Hypoglycemia and Diabetic Ketoacidosis in Adults With

Type 1 Diabetes: Results From the T1D Exchange Clinic Registry

Ruth S. Weinstock, Dongyuan Xing, David M. Maahs, Aaron Michels, Michael R.
Rickels, Anne L. Peters, Richard M. Bergenstal, Breanne Harris, Stephanie N.
DuBose, Kellee M. Miller, Roy W. Beck, and for the T1D Exchange Clinic Network
DOI: http://dx.doi.org/10.1210/jc.2013-1589
Received: March 07, 2013
Accepted: June 04, 2013
Published Online: June 12, 2013
Abstract

Context:
Few studies have assessed factors associated with severe hypoglycemia (SH) and diabetic ketoacidosis
(DKA) in adults with type 1 diabetes (T1D).

Objective:
Our objective was to determine frequency of and factors associated with the occurrence of SH and DKA
in adults with T1D.

Design and Setting:

We conducted a cross-sectional analysis from the T1D Exchange clinic registry at 70 U.S. endocrinology
centers.

Patients:
Analysis included 7012 participants in the T1D Exchange clinic registry aged 26 to 93 years old with T1D
for 2 years.

Results:
Higher frequencies of SH and DKA were associated with lower socioeconomic status (P < .001). SH was
strongly associated with diabetes duration (P < .001), with 18.6% of those with diabetes 40 years having
an event in the past 12 months. SH frequency was lowest in those with hemoglobin A1c (HbA1c) levels of
7.0% (53 mmol/mol) to 7.5% (58 mmol/mol), being higher in those with HbA1c levels <7.0% (<53
mmol/mol) or >7.5% (>58 mmol/mol). DKA frequency increased with higher HbA1c levels (P < .001), with
21.0% of those with HbA1c 10.0% (86 mmol/mol) having an event in the past 12 months.

Conclusions:
SH and DKA are more common in those with lower socioeconomic status. DKA, most common in those
with HbA1c 10.0% (86 mmol/mol), should be largely preventable. In contrast, SH, most frequent with
diabetes 40 years duration, cannot be abolished given the limitation of current therapies. To reduce SH
in adults with longstanding diabetes, consideration should be given to modifying HbA1c goals, particularly
in patients with very low HbA1c levels.
Severe hypoglycemia (SH) and diabetic ketoacidosis (DKA) are major acute complications of type
1 diabetes (T1D). Many studies are available concerning the prevalence of these complications in
children and adolescents with T1D, but there is a paucity of data in adults. Although recent data are
available from European registries (1, 2), recent U.S. data are largely from clinical trials, which may not be
representative (3, 4). To provide current data in clinical practice settings in the United States, we used the

large database of the T1D Exchange clinic registry to assess the frequencies of and identify factors
associated with SH and DKA in adults with T1D. We hypothesized that rates of DKA would be positively
associated with hemoglobin A1c (HbA1c) and that SH would be inversely associated with HbA1c.
Moreover, we hypothesized that rates of DKA and SH would be associated with factors such as
socioeconomic status, age, pump use, and diabetes duration.

Subject and methods

The T1D Exchange Clinic Network includes 70 U.S.-based pediatric and adult endocrinology practices. A
registry of individuals with T1D commenced enrollment in September 2010 (5). Each clinic received
approval from an institutional review board, and informed consent was obtained according to institutional
review board requirements. Data were collected for the registry's central database from the
participant's medical record and by having the participant complete a comprehensive questionnaire, as
previously described (5). This report includes data on 7012 participants enrolled at 44 of the sites through
August 1, 2012, who were at least 26 years old and had duration of T1D of at least 2 years.
Information on the occurrence of SH and DKA in the previous 12 months was obtained from the
participant. Initially, SH was defined as an episode in which the assistance of another individual was
needed or glucagon was given. However, it became apparent that the interpretation of needing
assistance varied, and in some cases, glucagon was being given for mild hypoglycemia to avoid SH. As
a result, the questionnaire was modified to change the participant-reported SH definition to be an episode
in which hypoglycemia resulted in seizure or loss of consciousness (LOC); glucose data confirming an
event were not collected. Participant-reported DKA was defined as the occurrence of ketoacidosis that
resulted in overnight hospitalization. SH data were analyzed for the 4973 participants (71%) who
completed the modified questionnaire with the seizure/LOC definition, whereas DKA data were available
for 6796 participants (97%). In addition to participant reporting, information on the occurrence of DKA and
SH also was collected from the clinics' medical records. Clinic-documented DKA was defined as having
hyperglycemia and meeting all of the following criteria: 1) symptoms such as polyuria, polydipsia, nausea,
or vomiting; 2) serum ketones or large/moderate urine ketones; 3) arterial blood pH <7.30, or venous pH
<7.30, or serum bicarbonate <15 mmol/L; and 4) treatment provided in a healthcare facility.
Primary analyses used the participant-reported events because the data obtained frommedical records
were considered to be frequently incomplete. Additional analyses were performed using the clinicreported cases.
Statistical methods

The occurrence of at least 1 event was considered to be more reliable than the number of events
reported by a participant when multiple events in the past 12 months were reported. Therefore,
frequencies of SH and DKA were reported as the occurrence of 1 or more events in the previous 12
months. Separate logistic regression models evaluated the association of baseline demographic and
clinical factors with the occurrence of a participant-reported SH or DKA event in the previous 12 months.
Factors with a P value < .10 from individual factor models were included in an initial multivariate model,
and then a backward elimination procedure was used to remove variables with a P value > .01. A forward
selection process resulted in a similar model. Tests of significance were reported from models using
continuous or ordinal variables and odds ratios were reported from models using the categorical
variables.
Data analyses used SAS software, version 9.3 (2011; SAS Institute Inc, Cary, North Carolina).
All P values are 2-sided. In view of the multiple comparisons and large sample size, only P values < .01
were considered significant.

Results
The cohort included 7012 participants 26 to 93 years old; 54% were female, and 91% were non-Hispanic
white. Mean HbA1c (SD) was 7.7% (61 mmol/mol) 1.2% (13.1 mmol/mol). Median diabetes duration
was 24 years (interquartile range 15 to 34 years). Characteristics were similar in the cohort with available
DKA data (n = 6797) and those with available SH data (n = 4973) (Supplemental Table 1, published on
The Endocrine Society's Journals Online web site at http://jcem.endojournals.org).
Severe hypoglycemia

One or more SH events (seizure or LOC) within 12 months was reported by 11.8% of the 4973
participants with available data. SH frequency increased with longer diabetes duration, with the 12-month
frequency of 1 or more events being 18.6% in the 758 participants with diabetesfor 40 years (Figure 1A
and Table 1). Although the frequency of SH increased with older age (P < .001), as can be seen in Figure
1A, the age effect was largely explained by diabetesduration, and in the multivariate model, age was not
a significant factor (Table 1). With respect to HbA1c levels, SH frequency was lowest with HbA1c levels
7.0% (53 mmol/mol) to 7.5% (58 mmol/mol), with frequency being higher at levels above or below this
range (Figure 2A andTable 1). No evidence of a difference in SH frequency was found between adults
with HbA1c levels of 8.0% to 9.0%, 9.0% to 10.0%, and >10.0% (P > .05). There was a trend (P = .01) for
individuals with low (<0.43 U/kg) insulin requirements to have a higher SH frequency independent of body
mass index (BMI) (which was not associated with SH frequency).

Figure 1. A, The 12-month frequency of severe hypoglycemia according to age and diabetes duration.
Solid white bars represent <20 yearsdiabetes duration, horizontal black line bars represent 20 to <40
years diabetes duration, and solid black bars represent 40 years diabetesduration. B, The 12-month
frequency of diabetic ketoacidosis according to age and diabetes duration. Solid white bars represent <20
yearsdiabetes duration, horizontal black line bars represent 20 to <40 years diabetes duration, and solid
black bars represent 40 years diabetesduration.

(Tabel 1 & 2 di document lain, ini tempatnya tabel 1 disini....)

Figure 2. A, The 12-month frequency of severe hypoglycemia according to HbA1c level. * Mean HbA1c
averaged over 12 months before enrollment.P value obtained by categorizing HbA1c with 3 levels: <7.0%
(<53 mmol/mol), 7.0-<7.5% (53-<58 mmol/mol) and > 7.5% (>58 mmol/mol); P value adjusted for duration
of T1D, household annual income, insurance status, and education level. B, The 12-month frequency
of diabetic ketoacidosis according to HbA1c level. ** Mean HbA1c averaged over 12 months before
enrollment. P-value obtained by treating as continuous variable; P value adjusted for gender, household
annual income, education level, and insurance status.

In addition to diabetes duration and glycemic control, SH was more likely to occur in participants who had
lower education levels (P = .003), had lower household income (P < .001), and did not have private
insurance (P < .001 in multivariate model). In univariate models, non-Hispanic black and Hispanic
participants had a higher frequency of SH compared with non-Hispanic whites, injection users had a
higher frequency of SH than pump users, and current smokers had a higher frequency of SH than
nonsmokers, but none of these factors were significant (P < .01) in the multivariate model.
Diabetic ketoacidosis

One or more DKA events within 12 months were reported by 4.8% of the 6796 participants. Frequency of
DKA was lower with increasing age. However, the age effect was largely explained by HbA1c level, which
was strongly associated with the occurrence of a DKA event, ranging from 1.6% in the 864 participants
with HbA1c level <6.5% (<48 mmol/mol) to 21.0% in the 334 participants with HbA1c 10.0% (86
mmol/mol) (Figure 2B and Table 2). Unlike with SH, frequency of DKA was not associated
with diabetes duration (Figure 1B). In addition to HbA1c level, a higher frequency of DKA was associated
with lower socioeconomic status based on education level, income, and insurance status (P < .001 for
each in multivariate model) and female gender (P = .008). In univariate models, non-Hispanic black and
Hispanic participants had higher frequencies of DKA than non-Hispanic whites, and current smokers had
higher frequency of DKA than nonsmokers, but after adjusting for socioeconomic status, neither factor
was significant in the multivariate model. Frequency of DKA was not significantly different between pump
and injection users.

(tabel 2 ditaronya di bagian ini..)

Participant-reported versus clinic-reported events

The frequency of at least 1 DKA or SH event in the past 12 months was lower from clinic-reported data
from medical record review compared with participant-reported data (3.6% versus 11.8% for SH and 3.2%
versus 4.8% for DKA). Logistic models using clinic-reported SH and DKA events to assess associated
factors produced results similar to the models using participant-reported events (Supplemental Tables 2
and 3).

Discussion
Using the T1D Exchange clinic registry data, we determined the frequencies of SH and DKA over the past
12 months in adults with T1D who were being treated at endocrinology practices in the United States.
Compared with the pediatric and young adult cohort in the T1D Exchange clinic registry, the 12-month
frequency of hypoglycemia-induced seizure/LOC in adults was higher (11.8% versus 6.1%) and DKA
frequency was lower (4.8% versus 9.6%) (5). Higher frequencies of both SH and DKA were associated
with lower socioeconomic status, consistent with previous reports (610).
Our 12-month frequency of at least 1 SH episode with seizure/LOC (11.8%) is substantially higher than
the 2.2% 6-month frequency reported for control group participants with comparable characteristics in the
Juvenile Diabetes Research Foundation Continuous Glucose Monitoring (JDRF CGM) randomized
clinical trial (4) and the 4.8% 12-month frequency in the Sensor-Augmented Pump Therapy for A1c
Reduction (STAR3) trial (3) (R. Bergenstal, MD, personal communication, November 2012). This
suggests that clinical trial data may not be an appropriate source for estimating the real-world frequency
of SH. Most other registry and observational studies have reported SH frequency based on a definition of
an event in which the assistance of another person was needed. If it is assumed that about one-third of
SH events using this definition involved seizure/LOC, then the data on the 12-month SH frequency are
similar to our SH frequency: 31.5% in the Eurodiab Prospective Complications Study (2), 36.7% in a
Danish-English survey (7), 40.5% in a Dutch study (11), and 27% in a Swedish survey (12). The UK
Hypoglycemia Study Group reported an SH frequency of 22% over a 12-month period in adults with T1D
for <5 years and 46% in those with duration >15 years (13). It is difficult to compare our SH frequency
with that of certain other studies such as the Diabetes Control and Complications Trial (DCCT) due to
differences in populations, SH definition, and reporting metrics (eg, frequency of 1 or more events versus
rate per 100 000 person-years).
Individuals with longstanding diabetes were at greatest risk for SH irrespective of age, with almost 1 in 5
of those with duration of 40 or more years reporting hypoglycemia-induced seizure/LOC within the
previous 12 months. Based on the work of Cryer (14) and others, it is likely that the increased risk with
longer duration is related to the loss of endogenous insulin secretion (C-peptide negative) removing any
autoregulatory capability to reduce excessive insulin, a greater prevalence of defective glucose
counterregulation with hypoglycemic unawareness including a reduced plasma glucagon response to
hypoglycemia, and loss of the epinephrine response to hypoglycemia (hypoglycemia-associated
autonomic failure). Impairments related to chronic diabetes-related complications and other comorbidities
also could be contributing factors. We did not find an association between SH and gender in contrast with
the JDRF CGM randomized trial and the Diabetes Control and Complications Trial (DCCT), which found a
higher SH rate in females (15).
The lowest SH frequency was seen in those with HbA1c levels in the range of 7.0% (53 mmol/mol) to
7.5% (58 mmol/mol), with the frequency being higher in those with HbA1c levels <7.0% (<53 mmol/mol)
or >7.5% (>58 mmol/mol). These data support the contention that hypoglycemia is an important barrier
for the achievement of near-normal glycemic control in T1D. There was a suggestion in the data that SH
frequency might be higher in those who are the most insulin sensitive (based upon lower insulin

requirements), but there was not an association between SH and BMI. This suggests that the ability to
achieve glycemic targets varies in people with T1D. This may be related to differences in C-peptide
status, insulin sensitivity (independent of obesity), and/or varying capacities to secrete glucagon and
other factors. These data suggest that the avoidance of SH requires setting individualized flexible
glycemic targets, with caution in recommending a goal HbA1c of <7.0% (<53 mmol/mol), especially in
older patients with longstanding diabetes who experience frequent SH. On the other end of the spectrum,
the high rates of SH in poorly controlled patients demonstrate that elevated HbA1c levels, per se, do not
protect against SH. We did not find an association of smoking and SH in contrast to the findings of other
studies (8, 16).
With respect to DKA, the frequency not surprisingly was highest in those with higher HbA1c levels,
occurring within the previous year in 21.0% with HbA1c 10.0% (86 mmol/mol). Notably, the frequency
was no higher in pump users than injection users, an important finding because DKA is a potential risk
with pump infusion site failure. As with SH, DKA rates in clinical trials are likely not good estimates of realworld rates. In the JDRF CGM trial and Sensor-Augmented Pump Therapy for A1c Reduction (STAR3)
trial, DKA occurred in none of the adults (3, 4). Current DKA data in adults from registries and
observational studies are limited. The Eurodiab study reported an annual frequency of 8.6% from 1989 to
1991 (17).
In evaluating the data generated from this large registry, it is important to recognize that the cohort is not
population-based and participation in the cohort is predicated on being followed by an endocrinologist at a
participating center. This could affect the representativeness of the cohort, particularly because many
adults with T1D may not be regularly cared for by an endocrinologist. As a result, estimated frequencies
of SH and DKA could be biased, although we cannot surmise whether the frequencies would be more
likely to be over- or underestimates. Nevertheless, the T1D Exchange data likely are the best currently
available data on the frequency of SH and DKA in the United States. A lack of representativeness, even if
present, is not likely to affect the interpretation of the analyses of factors associated with SH or DKA.
A registry such as the T1D Exchange provides meaningful real-world data that may better represent what
occurs to patients than does data collected in controlled clinical trials. The tradeoff is that the data cannot
be as readily confirmed as would be the case in a prospective trial with frequent patient contact and
prospective recording of events. We found differences in the SH and DKA event data collected directly
from the patient participant and from data extraction from the medical record. We believe that the
discrepancies between the 2 more likely reflect underreporting in the medical records than overreporting
by the participants. It is likely that not all events are recorded in the medical record particularly for SH,
which may not require the involvement of the diabetes management team. In some respects, the use of
electronic health records has made data extraction of SH and DKA events from existing data more
problematic rather than less, in that insufficient information may have been recorded to fully classify the
severity of the events, assuming that the events were even captured. Although the frequencies of SH and
DKA events differed between self-report and medicalrecord review, it is notable that the results with
respect to associated factors were similar using either data source.
In comparing our results with those of other studies, it is important to distinguish between SH requiring
assistance and SH resulting in seizure or LOC. Initially in the registry, we queried the participants about
SH events that required the assistance of another individual due to an impaired cognitive state or the
administration of glucagon. After initial testing with the questionnaire, we concluded that the interpretation
of what constituted an SH event varied and that glucagon was sometimes given for mild hypoglycemia.
Although it is possible that other conditions could cause seizures and LOC, we changed the data
collection form to record only events resulting in seizure or LOC, believing that such events would more
dependably reflect SH and be more reliably reported by the participant or parent.
Our DKA results largely parallel previous studies, providing similar insights. With respect to SH, 2 findings
have important implications for patient management: the strong association of SH with
longstanding diabetes and the lower SH frequency in those with HbA1c levels near goal (7.0% [53
mmol/mol] to 7.5% [58 mmol/mol]) compared with those who achieve the goal of <7.0% (<53 mmol/mol)
or have higher HbA1c levels. Not only will SH not be abolished with current diabetes therapies, but it will
also remain an obstacle to optimizing glycemic control. To reduce SH in adults with

longstanding diabetes, given the limitation of current therapies, consideration should be given to
modifying HbA1c goals, particularly in patients with very low HbA1c levels and SH. In addition to
hypoglycemic risk, the decision to raise the target HbA1c should take into consideration age, self-care
capacities, presence of comorbidities including vascular complications, and life expectancy.

Abbreviations:
body mass index
BMI
diabetic ketoacidosis
DKA
hemoglobin A1c
HbA1c
JDRF CGM Juvenile Diabetes Research Foundation Continuous Glucose Monitoring
loss of consciousness
LOC
severe hypoglycemia
SH
type 1 diabetes.
T1D

Acknowledgments
We acknowledge the patients who chose to participate in this registry and the T1D Exchange Clinic
Network investigators and coordinators for their hard work and dedication. A listing of clinical sites,
investigators, and coordinators participating in the clinic registry is located in the Supplemental Appendix.
The T1D Exchange Clinic Network is funded through a grant provided by the Leona M. and Harry B.
Helmsley Charitable Trust. The Helmsley Trust had no role in the study design; collection, analysis, and
interpretation of data; writing of this report; or the decision to submit the report for publication.
R.S.W. initiated the idea for the manuscript, researched data, contributed to discussion, wrote the
manuscript, and reviewed/edited the manuscript. D.X. researched data, contributed to discussion, and
wrote the manuscript. A.M. contributed to discussion and reviewed/edited the manuscript. M.R.
contributed to discussion and reviewed/edited the manuscript. A.P. contributed to discussion and
reviewed/edited the manuscript. R.M.B. contributed to discussion and reviewed/edited the manuscript.
B.H. contributed to discussion and reviewed/edited the manuscript. S.D. researched data, contributed to
discussion, and wrote the manuscript. K.M. researched data, contributed to discussion, and
reviewed/edited the manuscript. R.W.B. contributed to discussion and reviewed/edited the manuscript.
R.W.B. is the guarantor of this work and, as such, had full access to all the data in the study and takes
responsibility for the integrity of the data and accuracy of the data analysis.
Disclosure Summary: D.X., B.H., S.D., and K.M. have no disclosures. R.S.W.'s nonprofit employer has
received grant money as the site for multicenter clinical trials sponsored by Eli Lilly, Medtronic, AstraZeneca, GlaxoSmithKline, and Johnson & Johnson. D.M.'s nonprofit employer has received research
grants from Eli Lilly, Abbott, and Diabetes Care. A.M.'s nonprofit employer has received a research grant
from Novartis. M.R. has received consultancy fees from Longevity Biotech. A.P.'s nonprofit employer has
received research grants from Sanofi; consultancy fees from Eli Lilly, Roche, Janssen, Amylim, and
Sanofi; and payment for lectures from Amylin, Abbott, Eli Lilly, NovoNordisk, Takeda, and Dexcom.
R.M.B. has served on a scientific advisory board, consulted or performed clinical research with
Abbott Diabetes Care, Amylin, Bayer, Becton Dickinson, Boehringer Ingelheim, Intuity, Calibra, DexCom,
Eli Lilly, Halozyme, Helmsley Trust, Hygieia, Johnson & Johnson, Medtronic, Merck, National Institutes of
Health, Novo Nordisk, ResMed, Roche, Sanofi, and Takeda. His employer, Park Nicollet, has contracts
with the listed companies for his services, and no personal income goes to R.M.B. He has inherited Merck
stock and has been a volunteer officer of the American Diabetes Association. R.W.B.'s nonprofit employer
has received consultant payments on his behalf from Sanofi and Animas and a research grant from
NovoNordisk with no personal compensation to R.W.B.

References
1. M Gimnez, Lpez JJ, Castell C, Conget I. Hypoglycaemia and cardiovascular disease in type
1 diabetes. Results from the Catalan National Public Health registry on insulin pump
therapy. Diabetes Res Clin Pract. 2012;96:e23e25. [CrossRef] [Medline]
2. Gruden G, Barutta F, Chaturvedi N, . Severe hypoglycemia and cardiovascular disease incidence in
type 1 diabetes: the EURODIAB Prospective Complications Study. DiabetesCare. 2012;35:1598
1604. [CrossRef] [Medline]
3. Bergenstal RM, Tamborlane WV, Ahmann A, . 2010 Effectiveness of sensor-augmented insulin-pump
therapy in type 1 diabetes. N Engl J Med. 2010;363:311320. [CrossRef][Medline]
4. Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study
Group;Tamborlane WV, Beck RW, Bode BW, . Continuous glucose monitoring and intensive treatment of
type 1 diabetes. N Engl J Med 2008;359:14641476. [CrossRef] [Medline]
5. Beck RW, Tamborlane WV, Bergenstal RM, Miller KM, Dubose SN, Hall CA; T1D Exchange Clinic
Network. The T1D Exchange Clinic Registry. J Clin Endocrinol Metab.2012;97:4383
4389. [Abstract] [Medline]
6. Mhlhauser I, Overmann H, Bender R, Bott U, Berger M. Risk factors of severe hypoglycaemia in adult
patients with type I diabetes: a prospective population based study. Diabetologia. 1998;41:1274
1282. [CrossRef] [Medline]
7. Musey VC, Lee JK, Crawford R, Klatka MA, McAdams D, Phillips LS. Diabetes in urban AfricanAmericans. I. Cessation of insulin therapy is the major precipitating cause
of diabeticketoacidosis. Diabetes Care. 1995;18:483489. [CrossRef] [Medline]
8. Pedersen-Bjergaard U, Pramming S, Heller SR, . Severe hypoglycaemia in 1076 adult patients with
type 1 diabetes: influence of risk markers and selection. Diabetes Metab Res Rev. 2004;20:479
486. [CrossRef] [Medline]
9. Randall L, Begovic J, Hudson M, . Recurrent diabetic ketoacidosis in inner-city minority patients:
behavioral, socioeconomic, and psychosocial factors. Diabetes Care.2011;34:1891
1896. [CrossRef] [Medline]
10. Umpierrez GE, Kelly JP, Navarrete JE, Casals MM, Kitabchi AE. Hyperglycemic crises in urban
blacks. Arch Intern Med. 1997;157:669675. [CrossRef] [Medline]
11. ter Braak EW, Appelman AM, van de Laak M, Stolk RP, van Haeften TW, Erkelens DW.Clinical
characteristics of type 1 diabetic patients with and without severe
hypoglycemia.Diabetes Care. 2000;23:14671471. [CrossRef] [Medline]
12. Bragd J, Adamson U, Lins PE, Wredling R, Oskarsson P. A repeated cross-sectional survey of severe
hypoglycaemia in 178 type 1 diabetes mellitus patients performed in 1984 and
1998. Diabet Med. 2003;20:216219. [CrossRef] [Medline]
13. UK Hypoglycaemia Study Group. Risk of hypoglycaemia in types 1 and 2 diabetes: effects of
treatment modalities and their duration. Diabetologia. 2007;50:11401147.[CrossRef] [Medline]
14. Cryer PE. Hypoglycemia in type 1 diabetes mellitus. Endocrinol Metab Clin North Am.2010;39:641
654. [CrossRef] [Medline]
15. Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group;FialloScharer R, Cheng J, Beck RW, . Factors predictive of severe hypoglycemia in type 1diabetes: analysis
from the Juvenile Diabetes Research Foundation continuous glucose monitoring randomized control trial
dataset. Diabetes Care. 2011;34:586590. [CrossRef][Medline]
16. Hirai FE, Moss SE, Klein BE, Klein R. Severe hypoglycemia and smoking in a long-term type
1 diabetic population: Wisconsin Epidemiologic Study
of Diabetic Retinopathy.Diabetes Care. 2007;30:14371441. [CrossRef] [Medline]
17. Stephenson J, Fuller JH. Microvascular and acute complications in IDDM patients: the EURODIAB
IDDM Complications Study. Diabetologia. 1994;37:278285. [CrossRef] [Medline]