3684

Chin Med J 2010;123(24):3684-3688

Original article
Appropriate insulin initiation dosage for insulin-naive type 2 diabetes outpatients receiving insulin monotherapy or in combination with metformin and/or pioglitazone
LIAO Lin, YANG Ming, QIU Lu-lu, MOU Ya-ru, ZHAO Jia-jun and DONG Jian-jun Keywords: initiation dosage; insulin; metformin; pioglitazone; type 2 diabetes
Background Few studies have given suggestions on appropriate initiation insulin dosage when combined with oral antidiabetic drugs (OADs). This research was to investigate appropriate initiation insulin doses for insulin-naive type 2 diabetes patients with different combinations and the relationship between insulin dosage and relevant factors. Methods This was a randomized, open-label, treat to target study. The target was 20% decrease of both fasting plasma glucose (FPG) and 2 hours post-breakfast blood glucose (P2hBG). One hundred and forty-seven insulin-naive Chinese patients recruited were randomly assigned to 3 groups: group A, patients received insulin monotherapy; group B, received insulin plus metformin (0.5 g, tid) and group C, received insulin plus metformin (0.5 g, tid) and pioglitazone (15 mg, qd). Insulin doses were initiated with a dose of 0.3 Ukg-1d-1 and titrated according to FPG and P2hBG till reached the targets. Results Both the time of getting 20% reduction of FPG and P2hBG showed significant differences among the three groups. The time was shortest in Group C. The insulin doses needed to achieve glucose reduction of 20% in three treatment groups were (0.400.04) Ukg-1d-1 for Group A, (0.370.04) Ukg-1d-1 for Group B, and (0.350.03) Ukg-1d-1 for Group C, respectively. Multiple linear stepwise regression analysis showed that insulin doses correlated with body weight, FPG, diabetes duration, age and history of sulfonylurea treatment. The standardized regression coefficients were 0.871, 0.322, 0.089, 0.067 and 0.063 (with all P <0.05). Conclusions To achieve blood glucoses reduction of 20% within safety context, initial insulin doses were recommended as the following: 0.40 Ukg-1d-1 for insulin mono-therapy, 0.37 Ukg-1d-1 for insulin plus metformin treatment, and 0.35 Ukg-1d-1 for insulin plus metformin and pioglitazone treatment in Chinese type 2 diabetes outpatients. Body weight is found the most closely related factor to the insulin dosage. Chin Med J 2011;124(24):3684-3688

here is a rapid increase in the incidence of type 2 diabetes in China. Strict glycaemic control is essential to the prevention of acute and chronic complications that have serious impact on the physical and mental health of patients. As a very important therapeutic regimen in the treatment of diabetes, insulin administration has been recommended when fasting plasma glucose (FPG) >13.9 mmol/L, random blood glucose >16.7 mmol/L, glycosylated hemoglobin A1c (HbA1c) >10.0%, or there is occurrence of ketonuria, apparent polyuria, polydipsia and weight loss, according to the consensus statement for treatment of type 2 diabetes issued in 2008 by American Diabetes Association and European Association for the study of Diabetes Mellitus.1 However, the current clinical situation is that patients are treated with oral antidiabetic drugs (OADs) even when they meet the above criteria. One of the reasons that might be related to this situation is the lack of knowledge of insulin initiation.2-4 Physicians find it hard to prescribe proper initial doses of insulin to lower the blood glucose levels and at the same time avoid hypoglycemia, especially in combination therapy with other OADs. In addition, there is still not enough evidence to recommend the initial insulin dosage prescribed for insulin-naive patients with type 2 diabetes.

The aim of this study was to explore the appropriate insulin initiation doses for insulin-naive patients who received insulin monotherapy or in combination with metformin and/or thiazolidinedione (TZD, a kind of insulin sensitizer), as well as the relationship between insulin dosage and relevant factors. METHODS Study population This was a parallel-group, open-label, randomized, treat-to-target study. A total of 147 eligible insulin-naive patients were enrolled from daily diabetic clinics of Shandong Provincial Hospital and Qilu Hospital of
DOI: 10.3760/cma.j.issn.0366-6999.2010.24.031 Division of Endocrinology, Department of Medicine, Provincial Hospital affiliated to Shandong University, Jinan, Shandong 250021, China (Liao L, Yang M, Qiu LL, Mou YR and Zhao JJ) Department of Internal Medicine, Peking University Peoples Hospital, Beijing 100044, China (Yang M) Division of Endocrinology, Department of Medicine, Qilu Hospital of Shandong University, Jinan, Shandong 250012, China (Dong JJ) Correspondence to: DONG Jian-jun, Division of Endocrinology, Department of Medicine, Qilu Hospital of Shandong University, Jinan, Shandong 250012, China (Tel: 86-531-82169538. Fax: 86-531-86927544. Email: dongjianjun@medmail.com.cn)

Chinese Medical Journal 2010;123(24):3684-3688

3685

Shandong University, Jinan, China during January 1, 2006 to October 30, 2007. Inclusion criteria were: type 2 diabetes patients with unsatisfactory glycaemic control (HbA1c >7.0%) after diet control and OADs. Exclusion criteria included: various acute complications; hepatic transaminase >2.5 times normal reference value (glutamicpyruvic transminase >100 U/L, glutamic-oxalacetic transaminase >100 U/L), abnormal renal functions (serum cretinine >the normal reference value), cardiac insufficiency (America NYHA cardiac function >3); type 1 diabetes mellitus; ongoing hormone therapy; women in gestation and lactation; patients with other endocrine disorders. All subjects discontinued their previous OADs at randomization. Block randomization method was used in this study to minimize bias among the three treatment groups. The protocol was approved by local ethical committees. All participants gave written informed consent before inclusion in the study. The primary endpoint of the study was the insulin dosages that can decrease blood glucose of both FPG and postprandial blood glucoses (2 hours after breakfast, P2hBG) by 20% compared with baseline data. Group assignment and treatments Initial diabetes education was provided to all enrolled patients. After a 2-week diet and exercise treatment, blood glucose levels were re-examined and patients were randomly assigned to 3 groups using randomization number. Group A (49) received premixed insulin alone (insulin monotherapy group), twice-daily injections before breakfast and supper at an initial insulin dose of 0.30 Ukg-1d-1; Group B (49) received premixed insulin and metformin (insulin + metformin group) at an initial insulin dose of 0.30 Ukg-1d-1 and metformin dose of 0.5 g tid; Group C (49) received premixed insulin, metformin and pioglitazone hydrochloride (insulin + metformin + pioglitazone group) at initial doses of 0.30 Ukg-1d-1 for insulin, 0.5 g tid for metformin and 15 mg qd for pioglitazone, respectively. The premixed insulin used in this study was Novolin 30R (Novo Nordisk A/S, Denmark). The doses of pioglitazone and metformin remained constant throughout the trial. Efficacy assessments Designated staffs were to provide diabetes education and instruction of drug usage for patients during their treatment. FPG and P2hBG were monitored every day and right away if the patients felt any discomfort such as palpitations, sweating, dizziness, et al. Dose titration was based on blood glucose values every three days. If FPG decreased by less than 20% of the origin, pre-supper insulin dose was to be increased by 12 units (if the decrease of FPG was much less than 20% or even elevated, the pre-supper insulin dose was to be increased by 2 units; if it was slightly less than 20%, then, increased 1 unit) till reaching the target (20% decreases). If P2hBG after breakfast did not reach the target, pre-breakfast

insulin was to be increased by 12 units till P2hBG decreased to 20% of the original P2hBG. When both FPG and P2hBG reached the targets, the corresponding insulin dosage was to be calculated. All the patients reached the targets within 4 weeks of treatment. Safety assessments Safety was assessed by physical examination findings, laboratory evaluations, reporting of adverse events and hypoglycemic episodes. If patients experienced symptoms such as palpitation, sweating, tremor of hands and feeling of hungry, self-measured blood glucose (peripheral finger blood glucose) was monitored at once in order to find and correct hypoglycemia. Minor hypoglycemic episodes were defined as blood glucose values less than 3.9 mmol/L, and the patients were capable of self-treated (with or without symptoms). Major hypoglycemia was an episode with neurological symptoms consistent with hypoglycemia that required assistance and had either a plasma glucose value less than 3.9 mmol/L or reversal of symptoms after food intake, glucagon, or intravenous glucose. Patients who experienced hypoglycemic episodes were to be treated with reduced insulin doses and withdrawn from the study. Peripheral blood glucose was determined using One Touch Ultra Blood Glucose Meter (Life-Scan, UK). The original FPG and other biochemistry parameters were determined using X.100 automatic biochemistry analyzer (Backman, USA). Statistical analysis It was planned that 147 patients be randomized to three treatment groups in a 1:1:1 ratio, assuming a dropout rate of 10%, allowing a power of 80% to detect that the differences in insulin dosage among the three groups (=0.025, =0.2). Measurement data were presented as mean standard deviation (SD); enumeration data were presented as constituent ratios or rates, and 2 test was used for rate comparisons. Analysis of variance was used for comparisons of normal distribution data among the three groups. Analysis of covariance was used to correct effects of various factors on dependent variables. Correlations between factors were analyzed using multiple linear regression analysis. The regression analysis was used with the insulin doses at 20% reduction of blood glucose as dependent variable, and age, course of disease, body weight, triglyceride, FPG and history of sulfonylureas (SUs) therapy as independent variables. Statistical product and service solutions (SPSS) 11.5 software (SPSS Inc., USA) was used for statistical analysis. RESULTS Subject characteristics Totally, 147 Chinese patients with type 2 diabetes including

3686

Chin Med J 2010;123(24):3684-3688 Table 1. Baseline characteristics of the patients among three treatment groups

Groups A B C

n 48 46 45

Age (years) 53.010.2 52.610.6 51.410.2

Course of disease Body weight Triglyceride (months) (kg) (mmol/L) 37.5435.00 69.418.38 2.040.73 38.9644.00 70.9412.37 1.850.90 32.8836.40 72.0613.04 1.881.04 The differences among 3 groups were not significant, P >0.05.

FPG (mmol/L) 10.61.89 10.431.85 10.441.81

Proportion of patients using SUs before 21/48 18/46 20/45

73 men and 74 women were enrolled. Mean age was (52.210.4) years; mean body mass index (BMI) was (24.544.12) kg/m2, mean FPG was (10.581.90) mmol/L. Most (n=139, 94.6%) of the subjects completed the study. A total of 8 subjects discontinued the study. Seven patients withdrew because of the gastric discomfort or diarrhea, three from insulin + metformin group and four from insulin + metformin + pioglitazone group. One patient withdrew from insulin monotherapy group due to allergy of insulin. No patient was withdrawn from the study due to hypoglycemic episode. Finally, there were 48, 46 and 45 patients in group A, B and C, respectively. Baseline demographic characteristics including age, disease course, BMI, triglycerides and FPG levels and proportion of patients using SUs before were compared (Table 1) and the above baseline characteristics did not have any significant differences among the three groups (P >0.05). Glycemic control FPG and P2hBG levels were improved in all treatment groups after 4 weeks of treatment (Figures 1 and 2). The mean time to achieve FPG reduction of 20% was (21.24.8) days for Group A (insulin monotherapy group), (18.85.2) days for Group B (insulin + metformin group), and (16.65.0) days for Group C (insulin + metformin + pioglitazone group). Statistical analysis revealed a significant time difference to reach the target among the 3 groups. Analysis of variance showed that F=8.12, P <0.01, and Bonferroni method for multiple comparison indicated: t(A:B)=3.40, P <0.05, t(B:C)=2.85, P <0.05, t(A:C)=6.25, P <0.05. The mean time to achieve P2hBG reduction of 20% was (20.44.5) days for Group A, (17.64.8) days for Group B, and (15.24.2) days for Group C. Statistical analysis revealed a significant time difference to reach the target among the 3 groups. Analysis of variance showed that F=7.68, P <0.01, and Bonferroni method for multiple comparison indicated: t(A:B)=3.20, P <0.05, t(B:C)=2.90, P <0.05, t(A:C)=6.10, P <0.05. To achieve FPG and P2hBG reduction of 20%, insulin doses needed were (0.400.04) Ukg-1d-1 for Group A (insulin monotherapy group), (0.370.04) Ukg-1d-1 for Group B (insulin + metformin group), and (0.350.03) Ukg-1d-1 for Group C (insulin + metformin + pioglitazone group). Statistical analysis revealed significant differences among the 3 groups. Analysis of variance showed that F=6.43, P <0.01, and Bonferroni method for multiple comparison revealed that t(A:B)=4.20, P <0.05, t(B:C)=3.12, P <0.05, t(A:C)=7.32, P <0.05. The mean daily insulin dose was

Figure 1. FPG control in three treatment groups during the study. The mean time to achieve FPG reduction of 20% was (21.24.8) days for Group A (insulin monotherapy group), (18.85.2) days for Group B (insulin + metformin group), and (16.65.0) days for Group C (insulin + metformin + pioglitazone group). Statistical analysis revealed significant time differences to reach the target among the 3 groups.

Figure 2. P2hBG control in three treatment groups during the study. The mean time to achieve P2hBG reduction of 20% was (20.44.5) days for Group A (insulin monotherapy group), (17.64.8) days for Group B (insulin + metformin group), and (15.24.2) days for Group C (insulin + metformin + pioglitazone group). Statistical analysis revealed significant time differences to reach the target among the 3 groups.

0.05 Ukg-1d-1 lower in subjects in Group C taking metformin and pioglitazone compared with those injecting insulin alone in Group A. Allocation of insulin doses in the mornings and evenings when reaching the blood glucose target were 1.39:1 (58.20% in the morning and 41.80% in the evening). Regression analysis of insulin doses and relevant parameters Body weight, FPG, course of disease, age and history of

Chinese Medical Journal 2010;123(24):3684-3688

3687

SUs therapy were correlated with insulin doses. The standardized regression coefficients for the 5 variables were 0.871, 0.322, 0.089, 0.067 and 0.063, respectively; the corresponding P values were 0.000, 0.000, 0.006, 0.001 and 0.018. The coefficient of determination of variables on insulin doses at 20% reduction of blood glucose was 0.964. Tests on regression equation showed statistical significance as F=643.205 and P=0.000 (Table 2).
Table 2. Multiple linear stepwise regression analysis of insulin and relevant factors
Variables Standard error t values P values Constant 8.229 0.782 0.000 10.616 <0.001 Body weight 0.377 0.008 0.871 48.483 <0.001 FPG 0.861 0.078 0.322 10.979 <0.001 Course of disease 0.406 0.145 0.089 2.796 0.006 Age 0.032 0.010 0.067 3.312 0.001 History of using SUs 0.640 0.267 0.063 2.400 0.018 R2=0.964, F=643.205, P=0.000, the equation had statistical significance. It showed that insulin doses correlated with factors of body weight, FPG, diabetes duration, age and history of SUs treatment.

Why we chose 20% glucose reduction? Our purpose of the study was to find out the appropriate initiate insulin dose in insulin naive type 2 diabetes outpatients. Then, what was a satisfactory dose for a doctor and blood glucose decrease for the patients? A drop less than 10% was not satisfactory. Because if a patient begins insulin therapy at FPG of 15 mmol/L, a decline of 10% was 1.5 mmol/L, the blood glucose remains at 13.5 mmol/L, still quite high. However, a decline larger than 30% was easily to produce hypoglycemia for outpatients who usually regularly come to see the doctor weekly. So, we choose 20% decreases as our target. A decrease of 20% could lower blood glucose approvingly meanwhile to avoid hypoglycemia. Exploration of initial insulin doses and its influencing factors in Chinese population in this study has shown that body weight had the most impact on insulin doses, followed by fasting blood glucose, while triglyceride had no effect. Therefore, it is suggested to give primary consideration to body weight in determination of initial insulin dose for insulin-naive patients with type 2 diabetes, while other factors like FPG level, course of disease, age and history of SUs therapy are of less importance as compared to body weight, but still should be taken into consideration since they are also related to insulin doses. Significant glucose lowering effects can be obtained when patients of insulin monotherapy group received premixed insulin at dose of (0.400.04) Ukg-1d-1, while it was (0.370.04) Ukg-1d-1 in insulin plus metformin group and (0.350.03) Ukg-1d-1 in insulin plus metformin and pioglitazone group. Metformin can improve insulin sensitivity by activating the adenosine monophosphate (AMP)-activated protein kinase cascade,10,11 while TZD by activating its nuclear receptor peroxisome proliferator-activated receptor (PPAR) affecting gene regulation in target cells.12,13 Because of the different mechanisms, combination of the two agents results in good pharmacological effect, as each improves insulin resistance primarily in different tissue. Also, there may be some potential benefit of limiting weight gain when using TZD with metformin. Furthermore, with the data on preservation of -cell function,14,15 TZD may help preserve insulin secretory capacity and thus delay secondary failure of oral agent therapy.16 As well, this combination may allow for tighter glycemic control without the risk of severe hypoglycemia as SUs might bring. So, patients with metformin plus TZD will have a better insulin sensitivity and can reach the targets (blood glucose decrease by 20%) at the lowest insulin dose. Body weight was shown to have the largest impact on insulin dosage compared to other factors. However, since the objective of this study was to evaluate initial insulin dose, BMI was not used as an independent variable due to the following reasons: BMI is an index used in calculating standard body weight and includes the factor

Safety The overall rate of hypoglycemia including minor hypoglycemia was zero. Five subjects reported hypoglycemia-related symptoms such as palpitation, sweating and feeling of hungry, their blood glucose at that time were larger than 3.9 mmol/L. The insulin doses of those five patients were decreased by 2 units and titrated more slowly. DISCUSSION Insulin therapy is typically begun only after lifestyle modification and OADs fail to normalize HbA1c values. According to the Chinese guideline for prevention and treatment of type 2 diabetes mellitus (2007), in most circumstances, insulin is initiated in combination with biguanide or TZD. Since intensive glycemic control using insulin is associated with an increased risk of hypoglycemia,5 how to achieve effective hypoglycemic effects while avoiding hypoglycemia had become a hard task. Determination of initial insulin doses in combination therapies with other agents has therefore been a challenge for physicians. This study evaluated three approaches for initiating insulin therapy in patients with type 2 diabetes. One was the monotherapy of insulin, the second was combination of metformin and insulin, and the third was the combination of metformin, TZD and insulin. These three schemes coverd the most frequent therapeutic combinations in clinical practice.6,7 In order to get a better glucose control, many doctors prefered twice-daily premixed insulin as initial insulin therapy since basal insulin monotherapy is insufficient to control postprandial hyperglycemia and premixed insulin 70/30 provided signicantly improved overall glycemic control compared with once-daily insulin glargine.8 This is consistent with the fact that as cell function declines, HbA1c rises.9

3688

Chin Med J 2010;123(24):3684-3688 4. Rosenstock J, Riddle MC. Insulin therapy in type 2 diabetes. In: Cefalu WT, Gerich JE, LeRoith D, eds. The CADRE Handbook of Diabetes Management. New York, NY: Medical Information Press; 2004: 145-168. Davis S, Alonso MD. Hypoglycemia as a barrier to glycemic control. J Diabetes Complications 2004: 18: 60-68. The American Diabetes Association. Executive Summary: Standards of Medical Care in Diabetes 2009. Diabetes Care 2009; 32: s6-s12. The American Association of Clinical Endocrinologists Medical Guidelines for the Management of Diabetes Mellitus: the AACE system of intensive diabetes self-management-2000 update. Endocr Pract 2000; 6: 43-84. Raskin P, Allen E, Hollander P, Lewin A, Gabbay RA, Hu P, et al. Initiating insulin therapy in type 2 diabetes. Diabetes Care 2005; 28: 260-265. Maedler K, Donath MY. Beta-cells in type 2 diabetes: a loss of function and mass. Horm Res 2004; 62: 67-73. Zhou G, Myers R, Li Y, Chen Y, Shen X, Fenyk-Melody J, et al. Role of AMP-activated protein kinase in mechanism of metformin action. J Clin Invest 2001; 108: 1167-1174. Hundal RS, Inzucchi SE. Metformin: new understandings, new uses. Drugs 2003; 63: 1879-1894. Girard J. Mechanisms of action of thiazolidinediones. Diabetes Metab 2001; 27: 271-278. Lebovitz HE, Banerji MA. Insulin resistance and its treatment by thiazolidinediones. Recent Prog Horm Res 2001; 56: 265-294. Fonseca V, Rosenstock J, Patwardhan R, Salzman A. Effect of metformin and rosiglitazone combination therapy in patients with type 2 diabetes mellitus: a randomized controlled trial. JAMA 2000; 283: 1695-1702. Einhorn D, Rendell M, Rosenzweig J, Egan JW, Mathisen AL, Schneider RL. Pioglitazone hydrochloride in combination with metformin in the treatment of type 2 diabetes mellitus: a randomized, placebo-controlled study. The Pioglitazone 027 Study Group. Clin Ther 2000; 22: 1395-1409. Bell DS, Ovalle F. Long-term efficacy of triple oral therapy for type 2 diabetes mellitus. Endocr Pract 2002; 8: 271-275.

of body height. Our statistical analysis has shown that body height affects initial insulin dose calculation and causes irrelevance between body weight and initial insulin dose. In contrast, statistical analysis has showed that body weight becomes the most relevant factor when body mass index is substituted with body weight. The reason is that potency of insulin in the body is largely dependent upon insulin concentration in body fluids. Since total body fluids constitute approximately 60% of body weight, higher body weight with more body fluids required accordingly larger doses of insulin to obtain equivalent insulin concentration. Therefore, initial insulin dose is closely and proportionally correlated with body weight. Using a cohort of insulin-naive patients with type 2 diabetes, we investigated the relationship of several factors and the initial insulin doses at blood glucose reduction. Considerations should be given to body weight, fasting blood glucose, course of disease, age, history of SUs therapy, or combined hypoglycemic agents when determining initial insulin doses. It should also be noted that insulin doses required by various ethnic Chinese groups living at different regions may vary due to their distinct dietary habits or life styles. Corresponding adjustment are needed in clinical practice.
Acknowledgements: We thank our colleagues for their kindly help and support during the research progress and the follow-up. REFERENCES 1. Nathan DM, Buse JB, Davidson MB, Ferrannini E, Holman RR, Sherwin R, et al. Medical management of hyperglycemia in type 2 diabetes: a consensus algorithm for the initiation and adjustment of therapy. Diabetes Care 2009; 32: 193-203. Hirsch IB, Bergenstal RM, Parkin CG, Wright E Jr, Buse JB. A real-world approach to insulin therapy in primary care practice. Clin Diabetes 2005; 23: 78-86. Polonsky WH, Jackson RA. Whats so tough about taking insulin? Addressing the problem of psychological insulin resistance in type 2 diabetes. Clin Diabetes 2004; 22: 147-150.

5. 6.

7.

8.

9. 10.

11. 12. 13.

14.

15.

2.

16.

3.

(Received August 23, 2010) Edited by GUO Li-shao