J C E M

O N L I N E

B r i e f

R e p o r t — E n d o c r i n e

R e s e a r c h

Angiotensin-Converting Enzyme Insertion/Deletion Polymorphism and Severe Hypoglycemia Complicating Type 2 Diabetes: The Fremantle Diabetes Study Wendy A. Davis, Simon G. A. Brown, Ian G. Jacobs, Max Bulsara, John Beilby, David G. Bruce, and Timothy M. E. Davis School of Medicine and Pharmacology (W.A.D., D.G.B., T.M.E.D.), Centre for Medical Research (S.G.A.B.), School of Primary, Aboriginal, and Rural Health Care (I.G.J.), and School of Surgery and Pathology (J.B.), University of Western Australia, Crawley, Western Australia 6009, Australia; Institute of Health and Rehabilitation Research (M.B.), University of Notre Dame, Fremantle, Western Australia 6959, Australia; and Department of Clinical Biochemistry (J.B.), PathWest, Nedlands, Western Australia 6909, Australia

Aims/hypotheses: The aim of this study was to determine whether the angiotensin-converting enzyme (ACE) gene I/D polymorphisms independently predict severe hypoglycemia in communitydwelling type 2 patients. Methods: Six hundred and two patients who were ACE genotyped at baseline and assessed in 1998 were followed up to the end of June 2006. Severe hypoglycemia was defined as that requiring documented health service use as the primary diagnosis. Cox proportional hazards modeling was used to determine the predictors of first episode and zero-inflated negative binomial regression modeling identified predictors of frequency. Results: Forty-nine patients (8.1%) experienced 63 episodes of severe hypoglycemia. After adjusting for previously identified significant independent predictors of time to first episode, both ACE DD genotype and ACE inhibitor therapy, but not their interaction, added to the model [hazard ratio (95% confidence interval): 2.34 (1.29 – 4.26), P ⫽ 0.006, and 1.77 (0.99 –3.13), P ⫽ 0.052, respectively]. Similarly, after adjusting for previously identified risk factors for multiple episodes of severe hypoglycemia, ACE DD genotype was independently associated with increased risk [incidence relative risk (95% confidence interval): 1.80 (1.00 –3.24), P ⫽ 0.050]. Conclusions/interpretation: ACE DD genotype was associated with an approximately 2-fold increased risk of the first episode of severe hypoglycemia and its subsequent frequency in well-characterized patients with type 2 diabetes. Consistent with previous case-control studies, ACE inhibitor therapy was a weak predictor of severe hypoglycemia. ACE I/D genotyping might provide useful adjunctive prognostic information when intensive glycemic control measures are contemplated. (J Clin Endocrinol Metab 96: E696 –E700, 2011)

here is evidence that the renin-angiotensin system (RAS) is implicated in the etiology of severe hypoglycemia. The deletion (D) allele of the angiotensin-converting enzyme (ACE) gene is a strong risk factor in adults

T

with type 1 diabetes (1, 2), probably through its association with increased serum ACE activity relative to the insertion (I) allele and a consequently limited ability to maintain cognitive function during hypoglycemia (3, 4).

ISSN Print 0021-972X ISSN Online 1945-7197 Printed in U.S.A. Copyright © 2011 by The Endocrine Society doi: 10.1210/jc.2010-2087 Received September 7, 2010. Accepted December 29, 2010. First Published Online February 2, 2011

Abbreviations: ACE, Angiotensin-converting enzyme; ACE-I, ACE inhibitor; D, deletion; eGFR, estimated glomerular filtration rate; FDS, Fremantle Diabetes Study; FSG, fasting serum glucose; HbA1c, glycosylated hemoglobin; I, insertion; RAS, renin-angiotensin system; ZINB, zero-inflated negative binomial.

E696

jcem.endojournals.org

J Clin Endocrinol Metab, April 2011, 96(4):E696 –E700

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 January 2017. at 14:05 For personal use only. No other uses without permission. . All rights reserved.

J Clin Endocrinol Metab, April 2011, 96(4):E696 –E700

However, there is no such relationship in younger age groups (5, 6), and equivalent data in type 2 diabetes are restricted to a cross-sectional study of 308 insulin- or sulfonylurea-treated patients that also found no association between ACE genotype and severe hypoglycemia (7). The role of ACE inhibitor (ACE-I) therapy is also unclear. The associated suppressed serum ACE activity should reduce the risk of severe hypoglycemia as seen in a cross-sectional study of insulin-treated type 2 patients (8). However, large case-control studies suggest the opposite effect (9), perhaps due to a short-term post-dose increase in insulin sensitivity (10). Given this background and because previous studies have typically involved selected samples and not adjusted for all important confounding variables, we investigated the role of RAS-associated genotypes, ACE-I use and their interaction on the occurrence and frequency of severe hypoglycemia in representative well-characterized, community-dwelling, type 2 patients. The present analyses extend previously published predictors of severe hypoglycemia in this cohort (11).

Materials and Methods The Fremantle Diabetes Study (FDS) took place in a postcodedefined Australian community of 120,097 people. Descriptions of recruitment, sample characteristics, and details of nonrecruited patients have been published (12). Of 2258 diabetic patients identified between 1993 and 1996, 1426 (63%) were recruited and 1294 had type 2 diabetes. Eligible patients who declined participation were a mean 1.4 yr older than FDS subjects, but their sex distribution, the proportion with type 2 diabetes, and the distribution of treatment modalities were similar (12). The FDS protocol was approved by the Fremantle Hospital Human Rights Committee. All subjects gave informed consent before participation. The present study comprised follow-up of FDS patients from January 1999 when severe hypoglycemia could be ascertained accurately from relevant local linked databases (11). At that time, 1123 (86.8%) of the original FDS type 2 cohort were known to be alive. The present sample comprised the 602 of these patients (53.6%) who had undergone annual assessment during the preceding 12 months and were ACE genotyped. The assessment included demographic, socioeconomic, lifestyle and medical questionnaires, documentation of medications, and comprehensive physical examination. Fasting biochemical tests were performed using standard automated methods in a single laboratory. The estimated glomerular filtration rate (eGFR) was calculated from the serum creatinine, and complications were identified using standard definitions (11). Genomic DNA was isolated from peripheral blood and ACE I/D polymorphisms detected by validated PCR and, when appropriate, restriction typing of PCR products (13). Severe hypoglycemia was defined as an episode in which a patient with a subnormal blood/plasma/ serum glucose required documented health service use (ambulance attendance, emergency department attendance, or hospitalization) and hypoglycemia was the primary diagnosis (11).

jcem.endojournals.org

E697

The computer packages SPSS for Windows (version 17.0; SPSS Inc., Chicago, IL) and Intercooled Stata (version 10.0; StataCorp, College Station, TX) were used for statistical analyses as detailed previously (11). Data are presented as proportions, mean ⫾ SD or, for nonnormally distributed variables, median and (interquartile range). Genotype frequencies were tested for deviation from Hardy-Weinberg equilibrium by ␹2 test. Cox proportional hazards modeling [forward conditional variable entry (P ⬍ 0.05) and removal (P ⬎ 0.10)] and zero-inflated negative binomial (ZINB) regression modeling were used to determine independent predictors of first episode of severe hypoglycemia or frequency of severe hypoglycemia, respectively. The ZINB model is based on two latent groups; the first never has an episode (assessed in a logit model) and the second has a probability of severe hypoglycemia conforming to negative binomial distribution in which randomness is not assumed but certain subgroups have a higher chance of an event than others (count model). Plausible predictors in both models were selected from prior reports (11) and comprised age, gender, diabetes duration, educational attainment, English-speaking ability, body mass index, glutamic acid decarboxylase antibody status, fasting serum glucose (FSG), glycosylated hemoglobin (HbA1c), sulfonylurea treatment, insulin treatment and its duration, other treatment associated with hypoglycemia (aspirin, anticoagulants, nonsteroidal antiinflammatory drugs, fibrates, and allopurinol), self-monitoring of blood glucose, a history of hospitalization for severe hypoglycemia between FDS entry and January 1999, exercise, alcohol consumption, renal impairment (eGFR ⬍ 60 ml/ min per 1.73m2), peripheral neuropathy, polypharmacy (five or more medications), and any hospitalization in the previous 12 months. ACE genotypes (II, ID, DD) and carrier status (D or I) were entered into the models separately.

Results The 602 patients had a mean age of 67.1 ⫾ 9.8 yr, 52.0% were male, and they had been diagnosed a median of 7.7 (5.2–11.8) years previously. Their median HbA1c was 7.2% (6.5– 8.1%) and 13.2% were insulin treated. There were 194 patients (32.3%) taking an ACE-I and 12% (2.0%) on an angiotensin receptor blocker. No patient was on ACE-I/angiotensin receptor blocker dual therapy. These and other baseline characteristics categorized by ACE genotype are shown in Table 1. Forty-nine patients (8.1%) experienced a total of 63 episodes of severe hypoglycemia. After adjusting for the previously established predictors of time to first episode [history of severe hypoglycemia, renal impairment, peripheral neuropathy, and post-primary school education (11)], ACE DD genotype (which was in Hardy-Weinberg equilibrium at P ⫽ 0.92 with a D-allele frequency of 0.52) increased the risk more than 2-fold (Table 2). ACE-I use was of borderline significance, whereas ACE DD/ACE-I interaction was not an independent associate. For severe hypoglycemia frequency assessed using the ZINB model and after adjustment for the logit model for patients who were certain zeros [i.e. more likely to be non-insulin

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 January 2017. at 14:05 For personal use only. No other uses without permission. . All rights reserved.

E698

Davis et al.

ACE Polymorphism and Severe Hypoglycemia

J Clin Endocrinol Metab, April 2011, 96(4):E696 –E700

TABLE 1. Characteristics of 602 participants with type 2 diabetes at annual assessment in 1998, categorized by ACE genotype n Age on January 1, 1999 (yr) Male sex (%) Body mass index (kg/m2) Educational attainment greater than primary level (%) English ability (not fluent) (%) Any exercise in past 2 wk (%) Daily alcohol consumption (standard drinks) Glutamic acid decarboxylase antibody positive (%) Diabetes duration on January 1, 1999 (yr) FSG (mmol/liter) HbA1c (%) Sulfonylurea treatment (%) Insulin treatment (⫾oral agents) (%) Self-blood glucose monitoring (%) History of severe hypoglycemia (%) eGFR ⬍ 60 ml/min per 1.73m2 (%) Peripheral neuropathy (%) Orthostatic hypotension (%) Electrocardiographic QTc interval (msec0.5) Five or more prescribed medications (%) Anticoagulant therapy (%) Regular aspirin use (ⱖ75 mg/d) (%) Nonsteroidal antiinflammatory drug treatment (%) Allopurinol treatment (%) Fibrate treatment (%) ␤-Blocker treatment (%) Hospitalized during 1998 (%)

DD 166 66.5 ⫾ 9.9 51.8 29.6 ⫾ 5.1 70.5 15.1 70.5 0.1 (0 –1.2) 4.2 8.2 (5.7–12.5) 8.8 (7.4 –10.1) 7.3 (6.6 – 8.1) 45.2 17.0 84.9 3.0 28.9 53.6 27.7 429 ⫾ 30 43.4 2.4 28.3 15.1 5.4 6.6 16.9 37.3

ID 299 67.3 ⫾ 9.9 54.5 29.5 ⫾ 5.2 76.8 12.7 69.6 0.1 (0 – 0.8) 2.7 7.5 (5.0 –11.5) 8.5 (7.1–10.2) 7.2 (6.4 – 8.1) 49.5 12.8 86.9 2.0 34.1 45.5 30.1 428 ⫾ 27 33.6 3.7 30.4 9.4 7.7 5.7 19.1 36.5

II 137 67.4 ⫾ 9.5 46.7 30.4 ⫾ 6.6 86.0 7.3 67.2 0 (0 – 0.8) 2.2 7.7 (4.9 –11.8) 8.8 (7.4 –10.7) 7.3 (6.5– 8.3) 48.9 9.5 78.1 0.7 33.6 41.6 27.0 430 ⫾ 31 38.0 2.9 29.9 16.1 5.8 3.6 10.2 33.6

P value 0.66 0.33 0.27 0.005 0.10 0.81 0.10 0.65 0.54 0.47 0.34 0.67 0.16 0.07 0.38 0.49 0.09 0.67 0.81 0.11 0.83 0.91 0.06 0.64 0.55 0.06 0.79

Data are percentages, mean ⫾ SD or median and (interquartile range).

treated or to have short duration insulin treatment and unlikely to have impaired renal function or peripheral neuropathy or to have been educated beyond primary school level (11)], associates of severe hypoglycemia frequency among patients who were not certain zeros were higher HbA1c and lower fasting serum glucose [as found previously (11)] as well as ACE DD but not ACE-I use or ACE DD/ACE-I interaction (Table 2).

Discussion The present study is one of the largest to have examined the interaction between the RAS and hypoglycemia in diabetic patients and the largest in type 2 diabetes. We used longitudinal objective health service usage and other detailed sociodemographic, clinical, and laboratory data to examine whether, after adjustment for previously identified risk factors (11), aspects of the RAS are associated with the first occurrence and frequency of severe hypoglycemia. ACE DD genotype proved to be a significant independent predictor of both first and multiple episodes. ACE-I use weakly predicted first occurrence. These findings extend previous studies of severe hypoglycemia complicating type 2 diabetes and the knowledge of the role of the RAS be-

yond data obtained primarily from studies of Scandinavian type 1 patients. Our hazard ratio for first occurrence of severe hypoglycemia associated with the DD genotype (2.35) was similar to the relative risk of 3.2 described for adult type 1 patients who were not taking an ACE-I (1). These comparative data, the lack of an effect of ACE DD/ACE-I interaction in the present study, and the independent weakly positive association between ACE-I use and severe hypoglycemia in our patients all suggest that the effects of ACE-I on glucose metabolism are more complex than a protective role through reduction in serum ACE activity. Factors favoring increased glucose disposal such as increased muscle blood flow and elevated kinin levels may be involved (10, 14). In a study of type 1 adults (2), RASinhibiting treatment blunted ACE genotype-hypoglycemia associations, but the numbers of such patients were relatively small (28 of 108 cases and 35 of 262 unmatched controls) and the analysis was likely to be underpowered as a result. It has been suggested that previous reports of a positive association between ACE-I and severe hypoglycemia may reflect the unadjusted influence of powerful confounders such as diabetes duration and late complications (2). In-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 January 2017. at 14:05 For personal use only. No other uses without permission. . All rights reserved.

J Clin Endocrinol Metab, April 2011, 96(4):E696 –E700

jcem.endojournals.org

TABLE 2. Independent baseline predictors of time to first severe hypoglycemic event and frequency of severe hypoglycemia during follow-up Time to first event Time on insulin (increase of 1 yr) History of severe hypoglycemia eGFR ⬍ 60 ml/min per 1.73m2 Peripheral neuropathy Educational attainment beyond primary level ACE DD genotype ACE-I use Frequency Logit model Time on insulin (increase of 1 yr) eGFR ⬍ 60 ml/min per 1.73m2 Peripheral neuropathy Educational attainment beyond primary school level Count model HbA1c (increase of 1%) FSG (increase of 1 mmol/liter) ACE DD genotype

Hazard ratio (95% CI) 1.33 (1.15–1.53)

P value ⬍0.001

5.48 (2.05–14.64) 2.63 (1.46 – 4.73) 2.57 (1.36 – 4.84) 2.82 (1.25– 6.38)

0.001 0.001 0.004 0.013

2.35 (1.13–1.53) 1.77 (0.99 –3.13)

0.006 0.052

Incidence rate ratio (95% CI)

P value

0.34 (0.18 – 0.66)

0.001

0.18 (0.06 – 0.50) 0.18 (0.06 – 0.49) 0.17 (0.04 – 0.87)

0.001 0.001 0.033

1.36 (1.08 –1.71) 0.83 (0.73– 0.94) 1.80 (1.00 –3.24)

0.009 0.004 0.050

deed, patient selection, sample size, and availability of potential explanatory/confounding variables might explain why the only other previous study of ACE genotypes in type 2 diabetes did not show a significant association with severe hypoglycemia (7). The strengths of the present study include the range of available variables, the objective nature of severe hypoglycemia ascertainment, and the use of statistical techniques that minimize bias in identifying its predictors. Adherence with ACE-I therapy may have been suboptimal in some of our patients, but the positive association between ACE-I use and severe hypoglycemia suggests that this was not a major issue. Indeed, full adherence to ACE-I may have strengthened this association. The apparently discordant results in studies of the association between ACE genotype and severe hypoglycemia in type 1 diabetes (1, 2, 5, 6) may reflect the overriding influence of attempts to attain strict glycemic targets with imperfect insulin regimens relative to the variable lifestyle of most young patients. Our older type 2 patients are likely to have had a more regular lifestyle, whereas most severe hypoglycemia was not in insulin-treated patients (approximately 50% were on a sulfonylurea both at baseline and at the time of the event). It is possible that the modulating effects of ACE genotype become more evident in this setting. The mechanisms underlying the relationship between the DD genotype and severe hypoglycemia are unclear, but there is evidence that the associated high ACE activity

E699

increases susceptibility to cognitive dysfunction during hypoglycemia (3, 4). This may inhibit a patient’s ability to take appropriate corrective action, thus increasing potential severity. The lack of a significant association with the ID genotype suggests that the I-allele is dominant and that, in contrast to data from type 1 adults (2), D-allele homozygosity rather than simply its carriage is important. Inclusion of ACE genotypes in the two models in the present study did not displace previously reported predictors (11). Insulin treatment or its duration, prior severe hypoglycemia, renal impairment, peripheral neuropathy, and higher educational attainment proved independent predictors of first and multiple episodes. Frequency of severe hypoglycemia was also associated with a lower FSG but a higher HbA1c. In comparison with these other significant variables, the DD genotype represents a nonmodifiable risk factor for first and recurrent episodes. Other more easily accessible variables such as time on insulin therapy and renal impairment are stronger predictors of severe hypoglycemia, but it is possible that ACE genotyping could also be used to guide intensity of blood glucoselowering treatment in diabetic adults, especially when effects of hypoglycemia on cognitive function are of concern. Clinical and other features do not differentiate between ACE genotypes in the present and other (15, 16) studies, and the cost of such a test [US $37 based on current Australian Government rebates for comparable HFE genotyping (17)] has declined over recent years. Given current concerns relating to the effects of hypoglycemia on outcome in type 2 diabetes (18, 19), such additional prognostic information might justify this expense.

Acknowledgments We thank the patients for their participation, Fremantle Diabetes Study and Fremantle Hospital staff for help with recruiting patients, and collecting and recording clinical information. We also thank Paul Chubb and other members of the Biochemistry Department at Fremantle Hospital and Health Service for performing laboratory tests. Address all correspondence and requests for reprints to: Professor Timothy M. E. Davis, School of Medicine and Pharmacology, University of Western Australia, Fremantle Hospital, P. O. Box 480, Fremantle, Western Australia 6959, Australia. E-mail: [email protected]. This work was supported by an educational grant from Sanofi-Aventis. The Fremantle Diabetes Study was funded by the Raine Foundation, University of Western Australia. T.M.E.D. is supported by a National Health and Medical Research Council of Australia Practitioner Fellowship. Disclosure Summary: The authors declare that, apart from the stated funding sources, there is no duality of interest associated with this manuscript.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 January 2017. at 14:05 For personal use only. No other uses without permission. . All rights reserved.

E700

Davis et al.

ACE Polymorphism and Severe Hypoglycemia

J Clin Endocrinol Metab, April 2011, 96(4):E696 –E700

1. Pedersen-Bjergaard U, Agerholm-Larsen B, Pramming S, Hougaard P, Thorsteinsson B 2001 Activity of angiotensin-converting enzyme and risk of severe hypoglycaemia in type 1 diabetes mellitus. Lancet 357:1248 –1253 2. Pedersen-Bjergaard U, Nielsen SL, Akram K, Perrild H, Nordestgaard BG, Montgomery HE, Pramming S, Thorsteinsson B 2009 Angiotensin-converting enzyme and angiotensin II receptor subtype 2 genotypes in type 1 diabetes and severe hypoglycaemia requiring emergency treatment: a case cohort study. Pharmacogenet Genomics 19:864 – 868 3. Nordfeldt S, Samuelsson U 2003 Serum ACE predicts severe hypoglycemia in children and adolescents with type 1 diabetes. Diabetes Care 26:274 –278 4. Pedersen-Bjergaard U, Thomsen CE, Høgenhaven H, Smed A, Kjaer TW, Holst JJ, Dela F, Hilsted L, Frandsen E, Pramming S, Thorsteinsson B 2008 Angiotensin-converting enzyme activity and cognitive impairment during hypoglycaemia in healthy humans. J Renin Angiotensin Aldosterone Syst 9:37– 48 5. Johannesen J, Svensson J, Bergholdt R, Eising S, Gramstrup H, Frandsen E, Dick-Nielsen J, Hansen L, Pociot F, Mortensen HB 7 June 2010 Hypoglycemia, S-ACE and ACE genotypes in a Danish nationwide population of children and adolescents with type 1 diabetes. Pediatr Diabetes 10.1111/j.1399-5448.2010.00660.x 6. Bulsara MK, Holman CD, van Bockxmeer FM, Davis EA, Gallego PH, Beilby JP, Palmer LJ, Choong C, Jones TW 2007 The relationship between ACE genotype and risk of severe hypoglycaemia in a large population-based cohort of children and adolescents with type 1 diabetes. Diabetologia 50:965–971 7. Freathy RM, Lonnen KF, Steele AM, Minton JA, Frayling TM, Hattersley AT, Macleod KM 2006 The impact of the angiotensinconverting enzyme insertion/deletion polymorphism on severe hypoglycemia in type 2 diabetes. Rev Diabet Stud 3:76 – 81 8. Akram K, Pedersen-Bjergaard U, Carstensen B, Borch-Johnsen K, Thorsteinsson B 2006 Frequency and risk factors of severe hypoglycaemia in insulin-treated type 2 diabetes: a cross-sectional survey. Diabet Med 23:750 –756 9. Herings RM, de Boer A, Stricker BH, Leufkens HG, Porsius A 1995 Hypoglycaemia associated with use of inhibitors of angiotensin converting enzyme. Lancet 345:1195–1198

10. Rave K, Flesch S, Ku¨hn-Velten WN, Hompesch BC, Heinemann L, Heise T 2005 Enhancement of blood glucose lowering effect of a sulfonylurea when coadministered with an ACE inhibitor: results of a glucose-clamp study. Diabetes Metab Res Rev 21:459 – 464 11. Davis TM, Brown SG, Jacobs IG, Bulsara M, Bruce DG, Davis WA 2010 Determinants of severe hypoglycemia complicating type 2 diabetes: the Fremantle Diabetes Study. J Clin Endocrinol Metab 95: 2240 –2247 12. Davis TM, Zimmet P, Davis WA, Bruce DG, Fida S, Mackay IR 2000 Autoantibodies to glutamic acid decarboxylase in diabetic patients from a multi-ethnic Australian community: the Fremantle Diabetes Study. Diabet Med 17:667– 674 13. Hung J, McQuillan BM, Nidorf M, Thompson PL, Beilby JP 1999 Angiotensin-converting enzyme gene polymorphism and carotid wall thickening in a community population. Arterioscler Thromb Vasc Biol 19:1969 –1974 14. Uehara M, Kishikawa H, Isami S, Kisanuki K, Ohkubo Y, Miyamura N, Miyata T, Yano T, Shichiri M 1994 Effect on insulin sensitivity of angiotensin converting enzyme inhibitors with or without a sulphydryl group: bradykinin may improve insulin resistance in dogs and humans. Diabetologia 37:300 –307 15. Harrap SB, Tzourio C, Cambien F, Poirier O, Raoux S, Chalmers J, Chapman N, Colman S, Leguennec S, MacMahon S, Neal B, Ohkubo T, Woodward M 2003 The ACE gene I/D polymorphism is not associated with the blood pressure and cardiovascular benefits of ACE inhibition. Hypertension 42:297–303 16. Kajantie E, Rautanen A, Kere J, Andersson S, Yliha¨rsila¨ H, Osmond C, Barker DJ, Forse´n T, Eriksson J 2004 The effects of the ACE gene insertion/deletion polymorphism on glucose tolerance and insulin secretion in elderly people are modified by birth weight. J Clin Endocrinol Metab 89:5738 –5741 17. Commonwealth of Australia 2010 Medicare benefits schedule book. Pathology services. Canberra, Australia: Department of Health and Ageing 18. Kelly TN, Bazzano LA, Fonseca VA, Thethi TK, Reynolds K, He J 2009 Systematic review: glucose control and cardiovascular disease in type 2 diabetes. Ann Intern Med 151:394 – 403 19. Zoungas S, Patel A, Chalmers J, de Galan BE, Li Q, Billot L, Woodward M, Ninomiya T, Neal B, MacMahon S, Grobbee DE, Kengne AP, Marre M, Heller S 2010 Severe hypoglycemia and risks of vascular events and death. N Engl J Med 363:1410 –1418

References

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 15 January 2017. at 14:05 For personal use only. No other uses without permission. . All rights reserved.