| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Division of Metabolism, Endocrinology, and Nutrition, University of Washington School of Medicine (D.L.T., I.B.H.), and University of Washington Medical Center (J.L.K.), Seattle, Washington 98195-6176
Address all correspondence and requests for reprints to: Irl B. Hirsch, M.D., University of Washington Medical Center, 1959 NE Pacific Street, Box 356176, Seattle, Washington 98195-6176. E-mail: ihirsch{at}u.washington.edu.
 |
Abstract |
|---|
 Top Abstract References |
|---|
WHILE TRENDS IN the general population support decreasing mortality from cardiovascular disease, the patient with diabetes mellitus continues to have significant morbidity and mortality from cardiovascular disease (1). Death from cardiovascular disease accounts for 60% of all deaths associated with diabetes, and although very recent longitudinal data would suggest that indeed there has been a modest decline in the proportion due to heart disease, there has been no decline in cerebrovascular disease-associated death rate (2). It is also important to appreciate that the increasing population burden of diabetes prevalence actually results in an overall increase in cardiovascular event number (2). Hyperglycemia continues to be a marker of poor outcome in the hospitalized patient with acute myocardial infarction even in the era of thrombolytic therapy (3). Contributing to the burden of an acute event is the further correlation of hyperglycemia with poorer outcome after surgical intervention therapy for ischemic coronary disease (4). Treatment of the hospitalized patient with diabetes has recently focused on controlling glucose as a means of improving clinical outcomes.
The rationale for aggressive blood glucose control in hospitalized patients can be traced back 40 yr to Sodi-Pallares et al. (5). These investigators administered a polarizing solution that reduced the incidence of ventricular ectopy activity and improved early survival in an acute myocardial infarction (5). This polarizing solution was a glucose-insulin-potassium (GIK) infusion. The mechanism of the beneficial effect of the GIK infusion is still under debate.
In experimental low flow ischemia, the delivery of glucose and insulin improves ischemic and postischemic myocardial systolic and diastolic function as well as coronary vasodilatation. This, in turn, decreases the risk of reduced perfusion to the acutely ischemic muscle cell (6). The delivery of glucose and insulin increases glycolytic substrate, glycolytic flux, and ultimately ATP synthesis, thereby attenuating ischemia-induced decreases in ATP, phosphocreatine levels, and inorganic phosphate (7, 8). This lower production of inorganic phosphate results in preservation of a higher calculated free energy yield from hydrolysis of all ATP. Glucose and insulin, by increasing glycolytic flux, may also increase pyruvate generation, which, in turn, preserves the substrates of the citric acid cycle (9). Glycolytic ATP protects cell membranes (10), drives transport of calcium ion into sarcoplasmic reticulum (11), and in ischemic myocardium improves sodium homeostasis (12). In addition, myocardial stores of glycogen are preserved through provision of glucose and insulin. During ischemia, glycogen is rapidly depleted, which then has impact on calcium release and contractile function (13). Clinically, this is supported in studies of patients undergoing revascularization for coronary artery disease, as a positive correlation has been shown between enhanced glucose uptake, preservation of glycogen levels, and contractile function (14).
Additionally, increases in free fatty acid (FFA) levels, particularly nonesterified FFAs, are toxic to an ischemic myocardium. Elevated FFAs have long been associated with arrhythmias and have recently been found to be associated with cardiac sympathetic overactivity and a rise in oxidative stress in type 2 diabetes (15, 16). Elevated FFAs have been shown to cause endothelial dysfunction (17) and to increase blood pressure, perhaps by inhibition of endothelial nitric oxide synthase activity and endothelium-dependent vasodilation (18). It is important to note that circulating FFA levels and myocardial uptake of FFA are both decreased through GIK infusion (19).
Although FFA levels are increased in the setting of acute myocardial infarction as a result of increased catecholamine presence, FFA levels are also increased through the common clinical use of heparin. The antithrombotic effects of heparin are therapeutic; however, the activation of lipoprotein lipase by heparin and the resultant increase in FFA levels may not be beneficial. In patients receiving heparin, GIK infusion might be of particular benefit (19). Furthermore, in patients with type 2 diabetes undergoing elective coronary artery bypass grafting (CABG), FFA have been shown to be the main source of energy for the heart, with limited uptake seen of carbohydrates (20). Svedjeholm et al. (21) examined GIK infusion in 16 patients with severe cardiac failure after surgery for ischemic heart disease and reported rapid and near full recovery in 12 of the 16 subjects.
The beneficial effects of GIK infusion are not limited to the acute phase of myocardial infarction. In the postischemic state, there is imbalance between glycolysis and glucose oxidation. Impaired muscle contractility is thought to be the result either directly or indirectly of impaired glucose oxidation and accumulation of cytosolic protons (19). Agents that facilitate glucose oxidation have been shown to improve contractile dysfunction (22). It has been proposed that the depletion of glycogen stores and citric acid cycle intermediates in reperfused hearts is a major cause of impairment for the citrate synthetase reaction (acetyl coenzyme A and oxaloacetate). GIK could then potentially provide a "jump-start" for the impaired energy transfer of postischemic myocardial cells (19).
Glycolysis may be an important adaptive physiological change in ischemia. In isolated heart tissue made ischemic and then reperfused, glycolysis can prevent myocyte degeneration during forced ischemia and reperfusion transitions in the presence of excess oxidative substrate. Additionally, the same model shows the inotropic and metabolic effects of insulin to be additive, resulting in improved functional recovery in association with enhanced glucose uptake and utilization (23). Thus, different rationales for the use of metabolic support during ischemia and reperfusion of the myocardium complement each other.
Support of these metabolic changes in improving the physiology in ischemic conditions can be found in both animal and human models. Lazar (24), using pig models in which second and third diagonal vessels were occluded, followed by 30 min of cardioplegic arrest, then 180 min of reperfusion, simulated conditions of urgent CABG surgery. When GIK infusions were provided, there was significantly less acidosis, better recovery of ventricular wall motion, and smaller infarct size (24). In a prospective, randomized clinical study of patients undergoing CABG surgery for unstable angina, those patients that received GIK had a lower incidence of atrial arrhythmia, less weight gain, higher cardiac indexes, and overall shorter hospital stays from earlier extubation, as well as shorter intensive care unit (ICU) and post-ICU stays (24). Lindholm et al. (25) also showed that GIK infusion was associated with clinically important improvements during CABG. Less systemic vascular resistance, higher bypass pump flow, higher central mixed oxygen and saturation and oxygen tension, and higher hepatic venous oxygen saturation and oxygen tension were all reported.
For patients with diabetes, hyperglycemia at hospital admission for acute myocardial infarction correlates with increased in-hospital mortality (26, 27, 28, 29). Interestingly, measured systemic proinflammatory markers C3a and IL-6 are equally elevated in both those with and without GIK infusion (30). Diabetes is also an independent predictor of prolonged ICU stay, sternal instability and/or infection, sternal revision and respiratory insufficiency, postoperative delirium, perioperative stroke, renal dysfunction, and postoperative reintubation after cardiac surgical procedures (31). However, intensive insulin treatment, as shown in the Diabetes and Insulin-Glucose Infusion in Acute Myocardial Infarction study, produced improved survival at 1 yr postevent in the intensively treated patient group (P = 0.027) (32). Over an average subsequent follow-up of 3.4 yr, mortality continued to favor the intensively treated patient group (P = 0.011) (33).
Van den Berghe et al. (34) showed that intensive insulin therapy for patients admitted to a surgical ICU, treated to target glucose between 4.4–6.1 mM (80 and 110 mg/dl), resulted in dramatic clinical end points. This therapy reduced hospital mortality by 34%, sepsis by 46%, acute renal failure requiring hemodialysis or hemofiltration by 41%, and need for red cell transfusions by 50%. Compared with conventional therapy, there was also less critical illness neuropathy, and shorter durations of mechanical ventilation and ICU stays in these patients (34). The mechanism by which insulin produces theses beneficial effects has not been clearly established, but several contributing factors have been postulated. Normalization of blood glucose improves macrophage or neutrophil function (35, 36, 37) and glucose- and insulin-induced optimization of hemodynamic status in preventing renal failure (34) and reduced cholestasis (38, 39). It is also possible that the direct anabolic effect of insulin on respiratory muscle function may be responsible for the improved pulmonary outcomes (40, 41).
Others have postulated that decreasing insulin resistance, and thereby hyperglycemia, may enhance the responsiveness of leukocytes by inflammatory mediators (42). As noted above, C3a and IL-6 are not affected by the GIK infusion. However, insulin might by itself play a leading role through inhibiting other cytokines such as TNF
(43), with the combined effect of glucose and insulin infusion inhibiting macrophage inhibitory factor (44). The question of whether blood glucose levels should be a marker for other metabolic indicators, rather than the target for treatment itself, remains unclear (45). Given the effects of insulin on a variety of cytokines in addition to FFAs, insulin itself, rather than tight blood glucose control, may be the key to improving clinical outcomes for patients both with and without diabetes (46).
There are also a growing number of observational studies suggesting that hyperglycemia is related to stroke outcome (47, 48, 49, 50). To date, there is only one report of a treat to target intervention in stroke patients. The Glucose Insulin in Stroke Trial examined the safety of GIK infusion in treating to target blood glucose of 4–7 mM (72–126 mg/dl). Lowering plasma glucose levels was found to be without significant risk of hypoglycemia or excessive 4-wk mortality for patients with acute stroke and mild to moderate hyperglycemia (51). Due to the paucity of interventional clinical trial data, it seems premature to recommend GIK infusion to treat this population to normoglycemia as a standard of care.
The focus on in-hospital glycemic control must also be mitigated by attention to the prevention of hypoglycemia. In a tertiary care hospital, Fischer et al. (52) retrospectively reviewed 137 episodes of hypoglycemia that occurred in 94 hospitalized patients. A variety of circumstances were identified, the majority being related to decreased calorie intake. However, renal insufficiency, hepatic disease, infection, pregnancy, neoplasia, and burns were also associated factors. Although hypoglycemia was not believed to be the direct cause of death in any patient, overall mortality was 27% and was related to the degree of hypoglycemia and the number of risk factors for hypoglycemia.
These data, showing improvement in clinically important outcomes by aggressively treating hyperglycemia, has resulted in increased interest in diabetes management given the challenges of the in-patient setting. Strategies for the optimal treatment of diabetes are guided by two major principles. First, the type of diabetes must be considered, as treatments will obviously differ based on this factor. The second principle is the coordination of diabetes treatment, particularly the route of insulin administration, with the patient’s nutritional therapy. These issues are pertinent even for those patients admitted for reasons other than cardiovascular disease.
Medical patients admitted to the hospital today are more ill than patients of 20–30 yr ago, and many of these patients are too ill to eat. Couple this with the surgical patients who are frequently not allowed to eat (NPO), and the number of patients unable to eat in the hospital setting is high. The standard diabetes treatment for NPO patients is an iv insulin and glucose infusion. Formal studies examining the ideal quantity of glucose to administer (in those not receiving parenteral nutrition) has never been performed (53). However, in clinical practice patients should receive enough glucose and insulin to inhibit starvation ketosis, and thus a dose of 10 g/h of dextrose is usually recommended. For the patient with cardiovascular disease, the theoretical concerns of cardiotoxicity from elevated FFA (15, 16, 17) are yet another reason to ensure that starvation ketosis does not occur. If less glucose is provided for more than 24 h, urinary ketones should be measured to ensure that iatrogenic starvation is not occurring. Potassium chloride is also usually added to the infusion at a concentration of 20 mEq/liter for patients with normal renal function.
A variety of insulin infusion protocols have been published, and there are many more in clinical use that have not been published. Studies comparing different iv insulin infusion protocols are not available in the literature, and it is likely that different institutions will require different protocols based on their patient population as well as support personnel availability and physical plant. In general, the iv insulin infusion is appropriate for all type 1 patients hospitalized for an acute illness, where little or no food will eaten for at least 24 h, or for surgery during which general anesthesia will be administered for longer than 1–2 h. Similar recommendations can be made for patients with type 2 diabetes, as there is little role for oral antidiabetes agents for this population. Although there are obviously large differences in insulin requirements between different patients, iv insulin is significantly more predictable compared with sc insulin. The within-subject coefficient of variance (CV) for iv insulin is only 13%, whereas the between-subject CV is 26% (54). This is approximately half the CV seen with sc insulin (55).
Significant mention must be noted for the patient with diabetes undergoing CABG. The hypothermia that occurs during bypass may lead to extreme insulin resistance, resulting in the requirement of large doses of iv insulin. During warming, the insulin resistance subsides, and the rate of insulin administration needs to quickly be adjusted downward.
Historically, iv insulin infusion protocols were based solely on the patient’s blood glucose, and the infusion rate was changed by a fixed increment for all patients (53). In clinical practice this type of protocol rarely works, because it does not take into account differences in insulin sensitivity among patients. A young woman with type 1 diabetes who is NPO for a surgical procedure requires a smaller incremental change in insulin dose than a 60-yr-old man with type 2 diabetes who is undergoing cardiac bypass surgery. In addition, many nurses have noted that even if blood glucose is within target range, the infusion rate may need to be adjusted to prevent hypoglycemia. For example, if a patient’s blood glucose has decreased from 11.1 to 7.2 mM (200 to 130 mg/dl) in the preceding hour, the drip rate will need to be decreased to prevent hypoglycemia in the subsequent hour. An ideal insulin infusion protocol is based not only on the current blood glucose level, but also on the rate of change in blood glucose and insulin sensitivity.
One of the few published iv insulin algorithms, the Portland protocol, takes into account the rate of change in blood glucose by the percent change (56, 57). For example, for blood glucose levels between 5.6–8.3 mM (101–150 mg/dl), if the blood glucose is less than 10% lower than the last test, the rate of iv regular insulin is to be decreased by 0.5 U/h. If the blood glucose is more than 10% lower than the last test, the rate is to be decreased by 50% of the previous rate. There have been some changes in the Portland protocol, especially with regard to glycemic targets, but the use of changes in blood glucose by a percentage of the previous blood glucose and changes in insulin infusion rates based on percentages of the insulin infusion rates have not changed. At our institution, we have adapted the protocol initially published by Markovitz et al. (58) (Table 1
) to meet the needs of our patients. Our protocol is divided into four algorithms based on insulin sensitivity. The first algorithm is for the most insulin-sensitive patients, and the fourth algorithm is for the most insulin-resistant patients. The majority of patients start with algorithm 1. Insulin-resistant patients, such as those undergoing CABG or transplant, receiving glucocorticoids, or receiving more than 80 U insulin as outpatients, start with algorithm 2. The insulin infusion rate is determined by the patient’s blood glucose level and varies hourly until the patient is stable in the target range. If blood glucose targets are not achieved, or the blood glucose has not decreased by at least 3.3 mM (60 mg/dl) in the proceeding hour, the patient is moved up to the next algorithm. We occasionally have patients that are so severely insulin resistant that even the fourth algorithm is insufficient, and for these patients an endocrinology consult is encouraged.
|
) and the Portland protocol (57) initially require hourly testing, with a decrease in the frequency of testing as glycemia stabilizes. When we implemented our insulin drip protocol, we found that most patients required hourly blood glucose monitoring for only the first few hours, and then became stable, requiring less frequent blood glucose monitoring. Although nursing acuity is high for the first few hours of insulin drip therapy, patients quickly reach target blood glucose, resulting in lower nursing acuity and better blood glucose control. Those who are concerned that iv insulin infusions are too dangerous need to appreciate that with frequent capillary glucose monitoring, iv insulin infusions are safer than the traditional archaic sliding scale sc insulin strategies usually prescribed in the hospital (59). Hospitals differ on the need to admit patients to ICUs for iv insulin infusions, yet we see this practice as wasteful and expensive. Each institution will need to address where and when an iv insulin infusion can be administered based on their staffing and experience. We can suggest that an investment in nurse acuity up front will result in improved blood glucose control. Also, we can use the deficiencies in outdated sliding scale sc insulin as support for increase utilization of insulin drip therapy. If nurse staffing is not sufficient to maintain an iv insulin infusion on the general medical or surgical floor, then it probably is not sufficient to manage most patients receiving insulin, given the inherent problems of sliding scale insulin (53, 60, 61, 62, 63, 64). The thought that sc insulin is safe needs to be reconsidered when discussing all of the problems, and the reasons for them, with the routine use of insulin (63). The transition from an iv insulin/glucose infusion to sc insulin can be challenging. Subcutaneous insulin for in-patients follows the same strategy as out-patient insulin regimens (64). Insulin is classified as prandial, basal, and correction dose insulin for premeal hyperglycemia. For patients using insulin glargine, the iv infusion should serve as the basal insulin until the administration of insulin glargine. If dinner is planned, a prandial insulin, preferably insulin lispro or aspart, should be administered 0–15 min before the meal; this also applies if the first scheduled meal is lunch.
For those patients receiving insulin glargine who will be discharged from the hospital during the day and who did not receive insulin glargine the night before, there are two options for satisfying the basal insulin requirements until insulin glargine can be administered at home. The first option uses regular insulin as the usual basal/prandial insulin, supplementing with insulin lispro/aspart if premeal hyperglycemia is present. The lag time (duration between the injection and eating) is more critical with regular insulin than with the available prandial analogs, but the relatively long duration of regular insulin allows it also to function as a basal insulin (64) until insulin glargine is administered. Alternatively, a small dose of NPH (0.1–0.2 U/kg) may be administered at lunch and dinner, followed by the usual bedtime injection of insulin glargine. For patients using twice daily NPH, Lente, or Ultralente as their basal insulin, discontinuation of the infusion should coincide with the next scheduled injection of the basal and prandial insulins.
Table 2 shows a sample sc insulin order form that was developed and implemented to improve the safety and efficacy of sc insulin at our institution. Note that the physician is required to separate prandial, basal, and correction dose insulin. Lag times are often difficult to achieve, but if the patient can self-administer insulin, determining the timing between the injection and the meal becomes less of a challenge. Note that possible algorithms for correction dose insulin are provided, with an individualized algorithm also available to complete if required. Although formal studies have not been performed, we tend to only use either insulin lispro or aspart for the correction dose insulin, because theoretically it should result in less risk of hypoglycemia.
|
From physiological hypothesized benefit of the GIK infusion to the metabolic demonstrated impact on clinical outcomes in hospitalized patients with cardiovascular disease, glycemic control is important. Whether glucose is a marker of other physiological factors, such as cytokines, or other physiological changes, such as improved neutrophil function, remains unclear. However, the impact of glycemic control cannot be ignored, whether viewed from the individualized perspective of more prompt return to baseline health in an ill patient or from the socioeconomic perspective of health care dollars spent in hospital care. The time has come to treat hyperglycemia in the hospitalized patient as aggressively as we treat disorders of electrolytes or fluid balance.
Besides research focusing on mechanisms of how iv insulin and glucose alter these important outcomes, we also need to better understand which patient populations benefit the most from increased attention to glycemic control. We also need to better understand which strategies for diabetes treatment are most safe, efficient, and efficacious for different hospital settings.
Tempered by valid concerns about hypoglycemia, particularly iatrogenic hypoglycemia, tolerating hyperglycemia can no longer be considered a safer state for hospitalized patients with diabetes mellitus. The insulin treatment protocols presented in Tables 1 and 2 have been rapidly embraced by various specialty disciplines in our institution, as they are effective yet also very amenable to use on various in-patient wards, with varying support personnel. Additionally, patients who have been making sound decisions regarding their insulin dosing based on self-blood glucose management principles in an out-patient setting should be encouraged to continue these skills in the hospital setting. What endocrinologist has not received a phone call from a hospitalized patient with diabetes, rightly upset about having their glucose meter confiscated? There is also the problem of the prescribed insulin type or dose being changed despite the fact it had been quite effective in the past. How often have we noted a report of a blood glucose measurement grossly outside the target range with no additional insulin treatment? The added comment of "don’t worry, we see that a lot" or "that’s normal for that patient" only adds to our frustration.
Finally, there appears to be a momentum to improve upon our unacceptable standards of mediocre diabetes care for patients admitted to the hospital. Nonexperts in the field are finally questioning our time-honored tradition of sc sliding scale regular insulin for all patients. The use of iv insulin is becoming routine at many hospitals, especially for patients with acute myocardial infarction. The older practices of withholding insulin due to a procedure or due to euglycemia before eating in patients with type 1 diabetes will be buried with the fractional urine tests previously used to determine insulin dose. For the patient with type 2 diabetes, improved clinical outcomes and at less health care cost expenditure with insulin will outweigh continuing oral agent use in the hospital, a particularly appealing prospect, protecting the patient in congestive heart failure from continued use of metformin and thiazolidinediones. Although it will take time to change routine care for in-patients with diabetes, it is apparent that our old paradigm for this population is changing. It’s about time.
|
Footnotes |
|---|
Received February 28, 2003.
Accepted March 18, 2003.
|
References |
|---|
|
TopAbstractReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. A. Malesker, P. A. Foral, A. C. McPhillips, K. J. Christensen, J. A. Chang, and D. E. Hilleman An Efficiency Evaluation of Protocols for Tight Glycemic Control in Intensive Care Units Am. J. Crit. Care., November 1, 2007; 16(6): 589 - 598. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Wilson, J. Weinreb, and G. W. S. Hoo Intensive Insulin Therapy in Critical Care: A review of 12 protocols Diabetes Care, April 1, 2007; 30(4): 1005 - 1011. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. Inzucchi Management of Hyperglycemia in the Hospital Setting N. Engl. J. Med., November 2, 2006; 355(18): 1903 - 1911. [Full Text] [PDF]
|
|
|
|
 |
|
 A. D. Mooradian, M. Bernbaum, and S. G. Albert Narrative Review: A Rational Approach to Starting Insulin Therapy Ann Intern Med, July 18, 2006; 145(2): 125 - 134. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. X. Freire, L. Bridges, G. E. Umpierrez, D. Kuhl, and A. E. Kitabchi Admission Hyperglycemia and Other Risk Factors as Predictors of Hospital Mortality in a Medical ICU Population Chest, November 1, 2005; 128(5): 3109 - 3116. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J M. Schoeffler, D. A. Rice, and D. G Gresham 70/30 Insulin Algorithm Versus Sliding Scale Insulin Ann. Pharmacother., October 1, 2005; 39(10): 1606 - 1609. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. L. Thompson, K. C. Dunn, M. C. Menon, L. E. Kearns, and S. S. Braithwaite Hyperglycemia in the Hospital Diabetes Spectr, January 1, 2005; 18(1): 20 - 27. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. A. Goldberg and S. E. Inzucchi Selling Root Canals: Lessons Learned From Implementing a Hospital Insulin Infusion Protocol Diabetes Spectr, January 1, 2005; 18(1): 28 - 33. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. G. Pittas, R. D. Siegel, and J. Lau Insulin Therapy for Critically Ill Hospitalized Patients: A Meta-analysis of Randomized Controlled Trials Arch Intern Med, October 11, 2004; 164(18): 2005 - 2011. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. B. Campbell and S. S. Braithwaite Hospital Management of Hyperglycemia Clin. Diabetes, April 1, 2004; 22(2): 81 - 88. [Full Text] [PDF]
|
|
|
|
|
|
P. A. Goldberg, M. D. Siegel, R. S. Sherwin, J. I. Halickman, M. Lee, V. A. Bailey, S. L. Lee, J. D. Dziura, and S. E. Inzucchi Implementation of a Safe and Effective Insulin Infusion Protocol in a Medical Intensive Care Unit Diabetes Care, February 1, 2004; 27(2): 461 - 467. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. K. Plante and J. L. Nadler Diabetes and Vascular Disease Seminars in Cardiothoracic and Vascular Anesthesia, September 1, 2003; 7(3): 295 - 310. [Abstract] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
