STUDIES & RESEARCH

Closed-Loop Insulin Delivery for Glycemic Control in Noncritical Care

(New England Journal of Medicine)

List of authors.

Lia Bally, Ph.D., Hood Thabit, Ph.D., Sara Hartnell, B.Sc., Eveline Andereggen, R.N., Yue Ruan, Ph.D., Malgorzata E.
Wilinska, Ph.D., Mark L. Evans, M.D., Maria M. Wertli, Ph.D., Anthony P. Coll, M.B., B.S., Christoph Stettler, M.D., and
Roman Hovorka, Ph.D.

BACKGROUND

In patients with diabetes, hospitalization can complicate the achievement of recommended glycemic targets. There is
increasing evidence that a closed-loop delivery system (artificial pancreas) can improve glucose control in patients with
type 1 diabetes. We wanted to investigate whether a closed-loop system could also improve glycemic control in patients
with type 2 diabetes who were receiving noncritical care.

METHODS

In this randomized, open-label trial conducted on general wards in two tertiary hospitals located in the United Kingdom
and Switzerland, we assigned 136 adults with type 2 diabetes who required subcutaneous insulin therapy to receive
either closed-loop insulin delivery (70 patients) or conventional subcutaneous insulin therapy, according to local clinical
practice (66 patients). The primary end point was the percentage of time that the sensor glucose measurement was
within the target range of 100 to 180 mg per deciliter (5.6 to 10.0 mmol per liter) for up to 15 days or until hospital
discharge.

RESULTS

The mean (±SD) percentage of time that the sensor glucose measurement was in the target range was 65.8±16.8% in the
closed-loop group and 41.5±16.9% in the control group, a difference of 24.3±2.9 percentage points (95% confidence
interval [CI], 18.6 to 30.0; P<0.001); values above the target range were found in 23.6±16.6% and 49.5±22.8% of the
patients, respectively, a difference of 25.9±3.4 percentage points (95% CI, 19.2 to 32.7; P<0.001). The mean glucose level
was 154 mg per deciliter (8.5 mmol per liter) in the closed-loop group and 188 mg per deciliter (10.4 mmol per liter) in
the control group (P<0.001). There was no significant between-group difference in the duration of hypoglycemia (as
defined by a sensor glucose measurement of <54 mg per deciliter; P=0.80) or in the amount of insulin that was delivered
(median dose, 44.4 U and 40.2 U, respectively; P=0.50). No episode of severe hypoglycemia or clinically significant
hyperglycemia with ketonemia occurred in either trial group.

CONCLUSIONS

Among inpatients with type 2 diabetes receiving noncritical care, the use of an automated, closed-loop insulin-delivery
system resulted in significantly better glycemic control than conventional subcutaneous insulin therapy, without a higher
risk of hypoglycemia. (Funded by Diabetes UK and others; ClinicalTrials.gov number, NCT01774565. opens in new tab.)

The burden of diabetes is increasing worldwide,1 as is the proportion of patients with diabetes in hospitals. More than
one quarter of hospitalized patients in the United States and other developed countries have diabetes.2-4 In such
patients, the achievement of recommended glycemic targets5,6 is complicated by variable metabolic responses to acute

illness, changes in the amounts and timing of dietary intake, nutritional support, and drug-induced temporally rapid
alterations in insulin sensitivity from medications such as glucocorticoids.7-9

Strong associations have been reported between the rate of hyperglycemia among inpatients and an increased length of
hospital stay and increased rates of complications and death.10,11 Although the correction of hyperglycemia diminishes
the risk of adverse clinical outcomes,12 conventional insulin therapy increases the risk of hypoglycemia, which is
associated with increased morbidity and length of hospital stay.13 The implementation of current guidelines for
inpatient glycemic management is hindered by the need for vigilant and constant blood glucose monitoring and the
administration of insulin with meals, which increases the workload of hospital staff members and reduces staff
adherence.5,6 Consequently, glycemic control in hospitalized patients is often inadequate,2,14 which has spurred the
development of more effective and safe management strategies.15

An automated system that delivers insulin in response to glucose levels can address this need. Closed-loop glucose
control (also known as the artificial pancreas) consists of a continuous glucose monitor and an insulin pump, coupled
with a control algorithm that directs insulin delivery on the basis of real-time sensor glucose measurements.16 Such
autonomous glucose control obviates the need for the input of hospital staff members. There is increasing evidence that
closed-loop technology improves glucose control in patients with type 1 diabetes.17,18 In the critical care setting,
closed-loop technology has been evaluated for intravenous insulin delivery.19,20 However, for staffing and safety
reasons, subcutaneous insulin delivery has been feasible and pragmatic in patients receiving noncritical care.21 Here, we
report the results of a two-center, randomized, open-label trial of closed-loop insulin delivery without meal-associated
bolus administration in a diverse cohort of patients receiving noncritical care. We hypothesized that closed-loop insulin
delivery would be safe and improve glycemic control without increasing the risk of hypoglycemia.

Methods

PATIENTS

From August 2, 2016, to December 11, 2017, we recruited patients on general wards at the University Hospital in Bern,
Switzerland, and at Addenbrooke’s Hospital in Cambridge, United Kingdom. Inclusion criteria included an age of 18 years
or older and inpatient hyperglycemia requiring subcutaneous insulin therapy. Exclusion criteria were type 1 diabetes,
pregnancy or breast-feeding, and any physical or psychological disease or the use of medication that was likely to
interfere with the conduct of the trial or the interpretation of the results. Inpatients were identified through hospital
electronic records. All the patients provided written informed consent before the initiation of trial procedures.

TRIAL DESIGN

We randomly assigned the patients to receive insulin by means of a fully automated, closed-loop system (closed-loop
group) or conventional subcutaneous therapy (control group). Patients were followed for a maximum of 15 days or until
hospital discharge. Randomization was performed by means of the minimization method with the use of Minim
randomization software,22 which is a biased-coin approach with a probability of 0.7 to 0.8 for allocation of the “best
fitting” treatment. Randomization was stratified according to glycated hemoglobin level, body-mass index (the weight in
kilograms divided by the square of the height in meters), and pretrial total daily insulin dose to balance the two groups.
Investigators who analyzed the trial data were aware of the trial-group assignments.

TRIAL PROCEDURES

The body weight, height, and total daily insulin dose were recorded for each patient after enrollment. Throughout the
trial, the patients chose standard hospital meals at usual mealtimes, according to local practice. The patients were free
to consume other meals and snacks and were unrestricted in their usual activity on the general ward. In the two groups,
glucose levels were measured with the use of a continuous glucose monitor (Freestyle Navigator II, Abbott Diabetes
Care). A glucose sensor was inserted subcutaneously into the abdomen or upper arm by the investigator and calibrated
according to the manufacturer’s instructions. Point-of-care capillary glucose measurements (StatStrip Glucose Hospital
Meter System, Nova Biomedical, or Accu-Chek Inform II, Roche Diagnostics) were performed by nursing staff members
according to local clinical practice in the two trial groups.

CLOSED-LOOP INSULIN DELIVERY

Investigators discontinued each patient’s usual insulin therapy and sulfonylurea medication, if prescribed, on the day of
closed-loop initiation. All other medications were continued. The investigator inserted a subcutaneous cannula into the
abdomen for delivery of a rapid-acting insulin analogue (Humalog, Eli Lilly, or NovoRapid, Novo Nordisk) by means of a
trial pump (Dana Diabecare R, Sooil). The investigator initialized the control algorithm by using the patient’s weight and
pretrial total daily insulin dose. When sensor readings became available, the investigator initiated automated closed-
loop glucose control, which continued for up to 15 days. A low-glucose sensor alarm on the continuous glucose-
monitoring receiver was initialized at a threshold of 63 mg per deciliter (3.5 mmol per liter).

The automated closed-loop system consisted of a model predictive control algorithm (version 0.3.70) residing on a
control algorithm device (Dell Latitude 10 Tablet, Dell) linked by a USB cable to the continuous glucose-monitoring
receiver (Fig. S1 in the Supplementary Appendix, available with the full text of this article at NEJM.org). The tablet device
communicated with the pump by means of a Bluetooth wireless communication protocol. No prandial insulin boluses
were delivered, and the timing or carbohydrate content of meals was not included in the control algorithm. (Additional
details regarding the closed-loop system are provided in the Supplementary Appendix.)

At the end of the closed-loop period, patients completed a brief questionnaire to evaluate their satisfaction and trust of
automated glucose control with the closed-loop system, their acceptance of wearing trial devices, and their views as to
whether they would recommend the technology to other patients. Conventional insulin therapy and sulfonylurea
medication were resumed at the end of closed-loop use as appropriate.

CONVENTIONAL INSULIN THERAPY

For each patient, the usual insulin and other antihyperglycemic therapies were continued throughout the trial period. To
reflect usual care, the continuous glucose monitor was masked to the patient, investigators, and hospital staff members.
Each patient’s glucose control was managed by the clinical team, according to local clinical practice on the basis of
capillary glucose measurements. The clinical team was allowed to modify and adjust each patient’s insulin and other
antihyperglycemic therapies and to initiate additional point-of-care capillary glucose measurements as appropriate.

TRIAL OVERSIGHT

The protocol (available at NEJM.org) was approved by the local research ethics committee at each center and by
regulatory authorities in Switzerland (Swissmedic) and in the United Kingdom (Medicines and Healthcare Products
Regulatory Agency). The safety aspects of the trial were overseen by an independent data and safety monitoring board.
The trial was performed in accordance with the principles of the Declaration of Helsinki.

Abbott Diabetes Care supplied discounted continuous glucose-monitoring devices, sensors, and details regarding the
communication protocol to facilitate real-time connectivity; company representatives reviewed the manuscript before
submission but otherwise had no role in the trial conduct. All the authors participated in the design of the trial or
provided patient care and obtained samples. The first, second, and last author wrote the first draft of the manuscript.
The last author designed and implemented the control algorithm, and all the authors critically reviewed the manuscript.
The first and last authors vouch for the completeness and accuracy of the data and analyses and for the adherence of
the trial to the protocol. All the authors made the decision to submit the manuscript for publication.

PRIMARY AND SECONDARY OUTCOMES

The primary outcome was the percentage of time that the sensor glucose measurement was in the target glucose range
of 100 to 180 mg per deciliter (5.6 to 10.0 mmol per liter) for up to 15 days or until hospital discharge. Secondary
outcomes were the percentage of time that the sensor glucose measurement was either above or below the target
range; the percentage of time spent above 360 mg per deciliter (20.0 mmol per liter), below 70 mg per deciliter (3.9
mmol per liter), below 54 mg per deciliter (3.0 mmol per liter), and below 50 mg per deciliter (2.8 mmol per liter); the
area under the curve below 63 mg per deciliter (3.5 mmol per liter) and below 54 mg per deciliter; the mean daily sensor
glucose measurement; and the total daily insulin dose. We used data collected throughout the trial period to evaluate
glucose variability according to the standard deviation and the coefficient of variation in the sensor glucose
measurement. We calculated the between-day coefficient of variation in the sensor glucose measurement from daily
mean glucose values (midnight to midnight). Additional secondary outcomes and exploratory analyses are described in
the Supplementary Appendix.

Safety end points included clinically significant hyperglycemia (>360 mg per deciliter) with ketonemia and severe
hypoglycemia (<40 mg per deciliter), as determined by point-of-care capillary measurements, along with other adverse
events and serious adverse events.

STATISTICAL ANALYSIS

The trial was designed to have a power of 80% to detect a clinically significant between-group difference in the primary
outcome of 20 percentage points with the use of a two-sided t-test and an alpha level of 0.05. To reflect heterogeneity
among the patients, a standard deviation of ±39 for the primary outcome was used for the power calculations. We
planned that 150 patients would undergo randomization in order to permit the analysis of at least 48 hours of data from
120 patients.

The intention-to-treat analysis was performed on data collected during subcutaneous insulin delivery. Data from
patients who participated in a separate feasibility study21 were not included in the present analysis. Outcomes were
calculated with the use of GStat software, version 2.2 (University of Cambridge), and statistical analyses were performed
with the use of SPSS software, version 21.0 (IBM). We used the unpaired t-test to compare normally distributed
variables and the Mann–Whitney U test for highly skewed variables. The numbers of events that were related to a
capillary glucose measurement of less than 63 mg per deciliter and 40 mg per deciliter and more than 360 mg per
deciliter were tabulated in each trial group and compared with the use of Fisher’s exact test. Values are reported as
means (±SD) or medians (interquartile range), unless stated otherwise. All P values are two-tailed, and P values of less
than 0.05 were considered to indicate statistical significance.

Results

PATIENTS

Of the 165 patients who were invited to enroll in the trial, 138 consented to participate (Fig. S2 in the Supplementary
Appendix). One patient was withdrawn before randomization because of imminent hospital discharge. Of the remaining
137 patients, 70 were assigned to the closed-loop group and 67 to the control group. One patient in the control group
was excluded from the analysis because the transition from intravenous insulin to subcutaneous insulin did not occur as
originally planned.

Table 1.

Characteristics of the Patients at Baseline.
The demographic and clinical characteristics of the patients were similar with respect to sex, age, body-mass index,
glycated hemoglobin level, duration of diabetes, receipt of insulin, and insulin requirements (Table 1). Sepsis was the
predominant reason for admission (in 43% of the patients); approximately two thirds of the patients were being treated
with basal bolus insulin therapy. Additional data regarding the patients, including reasons for admission and antidiabetic
treatment before randomization, are provided in Tables S1 and S2 in the Supplementary Appendix. The burden of
coexisting illnesses was significantly higher in the closed-loop group than in the control group, according to the mean
score on the Charlson Comorbidity Index (9.4±3.4 vs. 7.0±2.8, P<0.001). Scores on this index range from 0 to 33, with a
score of ≥5 indicating a severe burden of illness. Additional details are provided in Table S3 and Fig. S3 in the
Supplementary Appendix.

OVERALL GLUCOSE CONTROL

Table 2.

Primary and Secondary Outcomes.
Figure 1.

Sensor Glucose Measurements and Insulin Delivery.
The mean (±SD) percentage of time that the sensor glucose measurement was in the target glucose range (primary
outcome) was 65.8±16.8% in the closed-loop group and 41.5±16.9% in the control group, for a difference of 24.3±2.9
percentage points (95% confidence interval [CI], 18.6 to 30.0; P<0.001) (Table 2). The mean sensor glucose
measurement was significantly lower in the closed-loop group than in the control group (154±29 mg per deciliter vs.
188±43 mg per deciliter; difference, 35±6 mg per deciliter; 95% CI, 23 to 47; P<0.001). Values above the target range
(>180 mg per deciliter) were found in 23.6±16.6% of the patients in the closed-loop group and in 49.5±22.8% of those in
the control group, a difference of 25.9±3.4 percentage points (95% CI, 19.2 to 32.7; P<0.001); there was no significant
between-group difference in the time spent at levels lower than the target range (<100 mg per deciliter, P=0.37) or
lower than 70 mg per deciliter (P=0.13). The burden of hypoglycemia was similar in the two groups, as measured by the
area under the curve of values below 63 mg per deciliter and below 54 mg per deciliter (Table 2). There was no
significant between-group difference in the median total daily insulin dose that was delivered (44.4 U in the closed-loop
group and 40.2 in the control group, P=0.50). The mean glucose variability in individual patients, as measured by the

standard deviation of the sensor glucose value, was significantly lower in the closed-loop group than in the control group
(46 vs. 59, P<0.001) (Table 2). The mean coefficient of variation in the sensor glucose measurement between 24-hour
periods was significantly lower in the closed-loop group than in the control group (15.6±8.0% vs. 21.7±12.2%, P=0.001).
The 24-hour sensor glucose measurements and insulin-delivery profiles are shown in Figure 1. End points for the first 48
hours and for the period thereafter until the end of the trial are provided in Table S4 in the Supplementary Appendix.

Capillary glucose measurements that were obtained before meals and before bedtime were significantly lower in the
closed-loop group than in the control group (P<0.01 for all comparisons) (Table 2). Hypoglycemic episodes with a
capillary glucose measurement of less than 63 mg per deciliter, as confirmed by point-of-care measurements, occurred
three times in the closed-loop group (in 3 patients) and nine times in the control group (in 8 patients). Hospital staff
members treated these episodes with oral carbohydrates according to local guidelines, without the need for intravenous
dextrose. As per protocol, two patients in the closed-loop group received supplemental insulin when their sensor
glucose measurements were greater than 434 mg per deciliter for more than 1 hour. Supplemental insulin was not
administered in the control group, since the sensor glucose measurement was masked in those patients.

OVERNIGHT AND DAYTIME GLUCOSE CONTROL

Table 3.

Daytime and Overnight Secondary Outcomes.
The mean percentage of time that sensor glucose measurements were in the target range was higher in the closed-loop
group than in the control group, both overnight (midnight to 8 a.m.) and during the daytime (8 a.m. to midnight).
Overnight, the percentage of time was 74.0±19.3% in the closed-loop group and 54.2±25.1% in the control group, for a
difference of 19.8±3.8 percentage points (95% CI, 12.2 to 27.4; P<0.001); during the daytime, the percentage of time
was 61.9±18.9% and 34.9±18.6%, respectively, for a difference of 26.9±3.2 percentage points (95% CI, 20.6 to 33.3;
P<0.001) (Table 3). The sensor glucose measurements were significantly lower in the closed-loop group than in the
control group, both overnight and during the daytime (P<0.001 for both comparisons), as were the standard deviations
of sensor glucose measurements during overnight periods (P<0.001) and daytime periods (P=0.001). In addition, the
between-night and between-day coefficients of variation in the sensor glucose measurement were significantly lower in
the closed-loop group than in the control group (P=0.004 for both comparisons). There was no significant between-
group difference in the nocturnal and daytime burden of hypoglycemia, as measured by the area under the curve for
values below 63 mg per deciliter (P=0.86 and P=0.24, respectively).

FOLLOW-UP PERIOD AND PATIENT FEEDBACK

The mean trial follow-up period, which was defined as the period from the first sensor reading until the last sensor
reading, was 7.9±3.9 days in the closed-loop group and 6.4±4.0 days in the control group (P=0.03). This time frame
included suspension of the trial period in 8 patients in the closed-loop group and 3 patients in the control group because
of surgery or other procedures that required transient removal of trial devices. Sensor glucose measurements were
available during 96% of the follow-up period in the closed-loop group and 92% of the follow-up period in the control
group (P=0.01). The closed-loop system was operational during 99% of the time when sensor glucose measurements
were available.

Overall, 54 of 62 patients (87%) in the closed-loop group reported that they were happy with their glucose levels during
the trial, and 61 of 62 (98%) reported that they were happy to have their glucose levels controlled automatically by the
closed-loop system (Fig. S4 in the Supplementary Appendix). All 62 patients reported that they would recommend the
system to a friend or family member during hospitalization.

ADVERSE EVENTS AND DEVICE DEFICIENCIES

Table 4.

Adverse Events and Safety Analyses.
No episodes of severe hypoglycemia or clinically significant hyperglycemia with ketonemia occurred in either group.
Adverse events that were related to trial devices occurred in three patients in the closed-loop group and three patients
in the control group. These events included skin irritation from sensor adhesive and bruising at cannula insertion sites. In
the closed-loop group, device deficiencies included sensor failures in two patients and a pump-check error in one
patient. The results for the safety end points are summarized in Table 4.

Discussion
In this trial involving hospitalized patients with type 2 diabetes, those who received insulin with a fully automated,
closed-loop system had significantly better glucose control than those who received standard subcutaneous insulin
therapy. The percentage of time that the sensor glucose measurement was in the target range was significantly higher in
the closed-loop group than in the control group, whereas the duration of hyperglycemia, the mean glucose level, and
glucose variability were significantly lower. These values were achieved without changing the total daily insulin dose and
without increasing the risk of hypoglycemia.

The advantage of a closed-loop system is the finely tuned, instantaneous glucose-responsive modulation of insulin
delivery, with its continual adaptation to changing insulin needs during the day and between days. In contrast,
conventional treatment approaches are less responsive to glucose changes and insulin needs; with tighter glycemic
control, such treatments are associated with an increased risk of hypoglycemia12,23 and adverse medical outcomes.13
The latter is a primary concern for many health care professionals and, we speculate, may explain why many
practitioners are reluctant to encourage tight glucose control.

Other techniques that address inpatient glycemic control include remote monitoring and consultation by a dedicated
specialist team,24 as well as algorithm-driven, computerized, tablet-based insulin-dosing support systems for hospital
staff members.25 Although glycemic benefits have been shown with the use of such systems, input by staff members is
still required, thereby decreasing usability, given the time constraints of daily practice.

Our findings expand the results of a single-center, randomized feasibility trial that evaluated a fully automated, closed-
loop system during a 72-hour period.21 Our trial was conducted at two centers in two countries, had a longer follow-up
period, and had a larger sample size in a considerably more diverse and complex inpatient population (including 19
patients who were receiving hemodialysis). In addition, patients in the closed-loop group in our trial did not receive
long-acting basal insulin, as was the case in the feasibility study. In spite of enrolling a more challenging and more
diverse inpatient population, we found that patients in the closed-loop group spent a higher percentage of time within

the glycemic target range than those in the control group (a between-group difference of 24 percentage points in our
trial vs. 21 percentage points in the feasibility trial) and a lower percentage of time above the glycemic target range (a
between-group difference of 26 percentage points vs. 19 percentage points). The observed differences may be
attributable to enhanced adaptive aspects of the control algorithm that we used and the longer trial duration, a
hypothesis that is supported by the greater benefit accrued beyond 48 hours of closed-loop operation (Fig. S5 in the
Supplementary Appendix).

A strength of our trial is that it addressed the unmet need for better glycemic control among hospitalized patients with
diabetes, an issue that affects nearly all areas of in-hospital care, patient outcomes, and health care costs. The two-
country design and sample size allowed for the evaluation of the safety and efficacy of closed-loop glycemic control over
a wide range of disease conditions, demographic characteristics, and different health care systems.

Our trial also has some limitations. Sensor glucose measurements were more available and the trial duration was longer
in the closed-loop group than in the control group. The observed imbalance may be attributable to between-group
differences in the burden of coexisting illnesses (which was higher in the closed-loop group than in the control group),
since the presence of such illnesses often increases the need for acute hospital care and may prolong
hospitalization.26,27 In addition, because the sensor glucose measurements were clinically unavailable in the control
group, any loss of connectivity between the sensor and the receiver device was not detected and may have contributed
to the collection of fewer sensor glucose data in the control group.

As part of the translation of research regarding the closed-loop system into clinical practice, further work is required to
determine practical considerations, facilitate ease of use, and assess costs. Standardized procedures will be needed to
ensure the most effective transition from acute care to outpatient care.28,29 Before closed-loop systems can have
widespread use, they may need to be integrated with electronic-record systems in hospitals and with training for health
care professionals.

In conclusion, in patients with type 2 diabetes who were receiving noncritical care, we found that the use of a fully
automated, closed-loop insulin-delivery system resulted in better glycemic control than standard insulin therapy. In
addition, the improved glucose control was achieved without increasing the risk of hypoglycemia in these patients.

Supported by a grant (14/0004878) from Diabetes UK, a grant (P1BEP3-165297) from the Swiss National Science
Foundation, the European Foundation for the Study of Diabetes, the JDRF, the National Institute for Health Research
Cambridge Biomedical Research Centre, and a Wellcome Strategic Award (100574/Z/12/Z). Abbott Diabetes Care
supplied discounted continuous glucose-monitoring devices, sensors, and details regarding communication protocol to
facilitate real-time connectivity. Dr. Bally received financial support from the University Hospital Bern, University of
Bern, and the Swiss Diabetes Foundation.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

Drs. Bally and Thabit contributed equally to this article.

This article was published on June 25, 2018, at NEJM.org.

We thank the trial volunteers for their participation; staff members at the Addenbrooke’s Hospital, Cambridge
University Hospitals NHS Foundation Trust, Cambridge, and at Inselspital, Bern University Hospital, Bern, for their
support; Josephine Hayes of the University of Cambridge for administrative support; Michèle Monnard of University
Hospital Bern for data-management support; and Jasdip Mangat for assisting in the development and validation of the
closed-loop system.

Author Affiliations

From the Departments of Diabetes, Endocrinology, Clinical Nutrition, and Metabolism (L.B., E.A., C.S.) and General
Internal Medicine (L.B., M.M.W.), Inselspital, Bern University Hospital, University of Bern, Bern, Switzerland; and the
Wellcome Trust–MRC Institute of Metabolic Science (L.B., H.T., Y.R., M.E.W., M.L.E., A.P.C., R.H.) and the Department of
Pediatrics (M.E.W., R.H.), University of Cambridge, and the Wolfson Diabetes and Endocrine Clinic, Cambridge University
Hospitals NHS Foundation Trust (S.H., M.L.E., A.P.C.), Cambridge, and the Manchester University Hospitals NHS
Foundation, Manchester Academic Health Science Centre (H.T.), and the Division of Diabetes, Endocrinology, and
Gastroenterology, Faculty of Biology, Medicine and Health, University of Manchester (H.T.), Manchester — all in the
United Kingdom.

Address reprint requests to Dr. Hovorka at the University of Cambridge Metabolic Research Laboratories, Institute of
Metabolic Science, Box 289, Addenbrooke’s Hospital, Cambridge CB2 0QQ, United Kingdom, or at rh347@cam.ac.uk.