Insulin Injection Technique is Negatively Correlate to Short- and Long-run Glycemic Control in Type 2 Diabetic Patients with Long-Acting Insulin Analogue

25 Objective: To observe the effects of insulin injection technique (IT) on short- or long- 26 run glycemic control in type 2 diabetic patients (T2D) with long-acting insulin analogue. 27 Methods: This was a single-center, cross-over, observational and open-labled study. 28 Patients with T2D receiving long-acting insulin analogue insulin were enrolled as 29 inpatients. The study period lasting for 5 days including a 1-day screen period and 4- 30 day continuous glucose monitoring (CGM) period. During CGM period, patients 31 injected insulin themselves from day 1 to day 2, and patient’s insulin IT was given by 32 two independent specialist nurses, with insulin injected by nurses from day 3 to day 4. 33 The primary endpoint was the correlation between the insulin IT and the mean 34 amplitude of glycemic excursion (MAGE). 35 Results: A total of 60 diabetic inpatients were recruited and completed the study. The 36 mean score of patients’ insulin IT of patients was lower than that of nurses (p<0.05). 37 We observed that the MAGE value was significant different between the two injections 38 period (P<0.05), and needle reuse and rotation of injection site were negatively 39 correlated to MAGE and HbA1c values, respectively. 40 Conclusion: Insulin IT was negatively correlation to short- or long-run glycemic 41 control in T2D patients with long-acting insulin analogue therapy. 42


47
Insulin resistance and hyperglycemia are two features of type 2 diabetes (T2D) 1 . If oral 48 antidiabetic agents (OADs) are no longer sufficient to maintain glycemic control, 49 insulin therapy will be a strategy of interest 2 . Available evidences indicates that more 50 than half of T2D Asia population are on insulin therapy 3 , with most of them are 51 prescribed premixed insulin (77.3%) for maintaining their glycemic control 4-6 , 52 following by basal (11.8%) and prandial insulin (10.9%) 7 . 53 Although study demonstrated premixed insulin analogue therapy has beneficial for 54 glycemic variation in Chinese T2D patients 8 . Recently, a pilot study performed in China 55 reporting that only 27% of type 2 diabetic patients receiving premixed insulin analogue 56 therapy had adequate glycemic control (HbA1c 7%) 9 , which was much lower than 57 those the whole population in China 10 and the USA population 11 . Incorrect insulin 58 injection technique (IT) is a common phenomenon worldwide, which may be the  More than one fifth Chinese T2D received basal and prandial insulin 7 . However, there 68 are rare data regarding the impact of IT on GV in T2D patients receiving basal and 69 prandial insulin. Continuous glucose monitoring (CGM) provides a potential 70 opportunity to assess the 24-hrs GV in patients with T2D 16 . Therefore, we conducted 71 the single-center, cross-over, observational and open-labled study to observe the 72 relationship between GV and insulin IT in patients with T2D who received long-acting 73 insulin analogue therapy.   The study period lasting for 5 days including a 1-day screen period and 4-day CGM 93 period. Patients demography data were recorded at day 0, and fasting serum fasting 94 plasma serum was collected for measurement of HbA1c value on day 1 morning. All 95 recruited patients injected insulin themselves from day 1 to day 2, and patient's insulin 96 IT was given by two independent specialist nurses using a scale containing 15 skill-97 related items, with minor modification (Table 1), as previously described 9 . From days 98 3 to days 4, the same type (unopened) and dose of insulin was delivered by nurses.

99
During the insulin injection period (from day 1 to day 5), all recruited patients were 100 subjected to a 96-hrs retrospective CGM (Medtronic Incorporated, Northridge, USA), 101 as previously described 17, 18 . All patients were instructed to perform self-monitored 102 blood glucose measurements, using OneTouch Ultra Vue blood glucose monitor 103 (LifeScan, Milpitas, CA), before breakfast, lunch, and dinner and before going to bed, The analyses were performed using the SPSS 18.0 (SPSS, Science, Chicago, USA) 116 statistical package. All normal distribution of the data were presented as the means ± 117 SD, and non-normally distributed data were presented as median (25 th , 75 th percentile).

118
Pearson's analysis or Spearman's analysis in nonparametric variable were performed to 119 analyze correlation relationships between two variables. A two-way ANOVA for 120 repeated measurements was used in the comparison of indices between two groups.

121
Multiple linear stepwise regression analysis was performed to identify factors which 122 correlation with MAGE or HbA1c. All comparisons were 2-sided at the 5% significance 123 level. P value less than 0.05 was considered to be statistical significance.  The duration for using insulin was 7.2±7.6 years, with 2.2±1.4 injection daily, and the 136 mean insulin dose was 0.4±0.4 U/kg/day.

137
In this study, we assumed scores of insulin IT gained by specialist nurses were 28 points, 138 because they were all trained for diabetic patients care for at least 5 years. We observed 139 that the mean score of patients' insulin IT was significantly lower than those mean score 140 of specialist nurses (22.3±3.6 vs. 28, p<0.05). In addition, there was no differences in 141 score of insulin IT between male and female patients (21.6±1.7 vs. 23.8±4.1, p>0.05).

142
In the present pilot study, we expected to see a significant improvement in GV during 143 diabetic patients receiving insulin delivered by specialist nurses compared to those who 144 injection themselves period. Our CGM data showed that the GV, in terms of MAGE, 145 SD, CV%, and AUC >10 mmol/L in patient injection period were significantly higher 146 than those of nurse injection period (P<0.05, respectively). Although we did not observe 147 significant differences in the MG and the AOC <3.9 mmol/L between the two injection 148 periods (Table2).

149
To identify whether score of insulin IT was the independent risk factor for short-run In this pilot study, we observed that nearly half patients with T2D receiving long-acting 161 insulin analogue therapy had their glycemic control. We also found that patients mean 162 IT scores were lower than that of nurses, and the lower IT scores were significantly 163 negatively correlated to the GV, in terms of MAGE, SDBG, and CV%.

164
Inject site choice, needle use, insulin suspension, and other profile regarding insulin 165 injection, are common problems affect insulin efficacy in outpatient diabetic patients 166 worldwide 12, 13 . In this study, each patient scores of insulin IT were the sum of each 167 subitem score given by two special diabetic nurses according to them performance. we 168 observed that T2D patients with insulin therapy gained dramatically lower IT in the real 169 world, with the scores lost mainly focused on the rotation of injection site and needle 170 reuse.

171
In this observational study, we observed that only 20% of T2D patients had HbA1c less 172 than 7%. HbA1c is an established index employed for judging long-term glycemic 173 control 19-21 . However, HbA1c value itself can't describe well the glycemic variation, 174 especially acute glucose fluctuation [22][23][24] . CGM employee is one strength of this study, 175 our data showed that insulin IT not only negatively correlated to HbA1c value, but also 176 negatively correlated to GV. Our findings were in accordance with previous study 177 reporting that patients receiving pre-mixed insulin analogue with lower insulin IT had 178 significantly increase in GV and HbA1c value 9 .

179
Unlike premixed insulin analogue, insulin did not need resuspension before injection 25, 180 26 . However, other injection problems remain to be resolved. Study revealed that reuse 181 of needles may lead to bad short-run and long-run glycemic control, such as increased 182 in GV and HbA1c levels 13 , and lipohypertrophy (LH) 27 . Using self-monitoring of blood 183 glucose data, researchers observed that LH was an independent risk factor for the 184 increase in GV 27, 28 . Unfortunately, nearly half of patients with insulin therapy reuse 185 their needles more than once throughout the world 12 . In this study, only 4.2% patients 186 used their needles once, with most of them had their needles use more than 5 times. Of 187 importance, we found that the needle reuse score was significantly negative related to 188 MAGE. Much indices delivered from CGM were used to assess GV, such as MAGE, 189 SD, and CV%, in pre-or onset diabetic patients 29-31 , of which, there exist a high degree 190 of correlation between SD and MAGE 32, 33 . Interestingly, our data also revealed that 191 the score of needle reuse was negatively correlated to SD. 192 Rotate of injection site is one of the independent risk factor for induction of HbA1c level, 193 the underline mechanism may partially be the reason of the LH 12,13,34,35 . In this study,194 we observed that patients with insulin with lower rotation injection site score had 195 significantly increase in MAGE, and HbA1c value, which was slightly different to 196 previous study reporting that score of rotation of injection site was negative correlation 197 with HbA1c value in patients with premixed insulin analogue, rather than MAGE 9 .

198
Future studies identifying the relationship between score of rotation of injection site 199 and HbA1c or MAGE are warranted. Lifted a skinfold is most important in the past 200 decade during insulin injection. However, the score of lifted a skinfold of insulin IT 201 may be not importance as before, because 4 mm needle was used predominately 202 nowadays in China. In this study, all recruited patients lifted their skinfold well.

203
Several limitations should be addressed in this study. First, the sample size was modest; 204 Second, it was an observational study, not a perspective study, Third, the study 205 population may not be the same as other geography.

206
In conclusion, our data indicates that insulin IT was negatively correlation to short-or 207 long-run glycemic control in T2D patients with long-acting insulin analogue therapy in 208 Chinese population.    Table 2 The 24-hrs glucose profiles between the two different injection groups