The Chinese ITQ study, involving 3853 patients from 20 centers in 8 cities (Table 7), was part of a worldwide survey conducted in 2014–2015 involved 13,289 subjects from 42 countries [11, 12]. That survey showed that LH was present in nearly 1/3 of insulin injectors worldwide. We consider this an underestimate since most HCPs conducting the ITQ were not trained in LH detection before the study. By step-wise regression analysis, incorrect rotation and years on insulin were the most important factors associated with LH (p < 0.001), while needle reuse is significantly but slightly less strongly associated (p = 0.023).
There was also a strong association in the worldwide ITQ between the presence of LH and the TDD of insulin, with over 10 IU more of insulin on average being consumed in those with LH (LH+) than those without LH (LH-). All currently used types of insulins are associated with the risk of LH, even newer analogs, as are all lengths of needles. The presence of LH is associated with 0.5% higher HbA1c values on average.
Ji et al. recently reported LH prevalence in China of 53.1% . LH+ patients in that study had higher HbA1c values than their LH- counterparts [8.2% vs. 7.7% (66 vs. 61 mmol/mol), respectively]. LH+ patients also consumed 11 IU (0.13 IU/kg or 31.7%) more insulin per day (38 vs. 27 IU). Daily consumption of insulin by LH + cost RMB 9.5 vs. 6.8 in LH- subjects. LH+ patients also reused needles more often and rotated sites less frequently. The authors conservatively extrapolated their results to the estimated 9 million insulin-injecting patients in China and concluded that LH-related excess insulin costs were approximately 2 billion RMB (290 million USD) annually. Famulla  has clearly shown that the increased consumption of insulin by those with LH is due to impaired insulin absorption (and action). Theirs was the only reported crossover euglycemic clamp study comparing insulin pharmacokinetics (PK) and pharmacodynamics (PD) when injected into areas of LH vs. normal tissue.
To prevent (or cure) LH, patients must carefully rotate their injection sites, a change that often requires that patients expand the zones into which they inject [18,19,20,21]. The vast majority of Chinese patients (92.0%) in the ITQ claim to rotate their injection sites (the best defense against LH), and even Chinese HCPs report most (77.3%) patients correctly rotate, i.e., leaving a space at least 1 cm between new injections and any old ones. However, when the rotation method was specifically queried it became clear that most Chinese patients use an inadequate rotational method (Table 5, italic options). Trends are similar in ROW, but even fewer Chinese use an adequate method (Table 5, bold options) than in ROW (p < 0.001).
Hence, it is clear that more effort must go into teaching Chinese patients what correct rotation is and how to perform it. When injecting over wider areas, however, HCPs may introduce an additional risk, that of giving injections IM. IM injections change variably (i.e., accelerate) the PK of insulin, especially if the muscle is actively working . To reduce the risk of IM injections, but still provide for wider injection zones, it is imperative that patients use the shortest needles possible. In addition, when a patient changes from injecting into areas of LH to sites without LH, it is advisable to simultaneously reduce the dose of insulin by 10–20% and to then titrate based on BGM [12, 23] to minimize the risk of hypoglycemia.
Our data show that most Chinese patients already use the 5-mm needle and that use of the 4-mm needle is growing. The 4-mm pen needle was introduced in 2010. The impetus for this ‘shift to short’ was the publication of a series of studies in persons with diabetes, which defined the mean skin thickness as 2.0–2.5 mm [24, 25]. One of these studies came from China . It agreed with other studies done in India , the Philippines , Korea , the USA , South Africa  and Italy , all showing essentially similar results but in different patient and ethnic populations.
The span of SC fat is the critical factor informing both the optimal needle length and the preferred technique for injection (e.g., angling the needle or lifting a skin fold). This distance also determines the risk of IM injections [32, 31]. The higher the BMI, the thicker is the SC [24, 32]. Chinese patients in our study had significantly lower BMI than those in ROW (Table 1). Wang  has published the dimensions of SC thickness in 508 adult Chinese patients with DM. It is clear that the risk for IM injections is considerable in Chinese patients, significantly higher than in western patients with much higher BMI values.
Although fewer patients in China still use the 8-mm needle than in ROW (10.3% vs. 28.5%), its use is still worrisome, given the lower BMI in China (Table 2). There seems to be a downward trend in the use of this needle, and Chinese HCPs should encourage this. Studies in various countries comparing many needles have clearly shown that the needle length has no effect on glucose control, including in overweight and obese patients [33,34,35,36,37,38,39,40,41,42]. Therefore, Chinese HCPs should not hesitate to move their patients from the 8-mm needle to shorter lengths. If the 4-mm length is not available, 5-mm pen needles can be used with a lifted skin fold for additional safety, especially in slim Chinese patients or those injecting into the thigh. Injecting the 6-mm needles at a 45° angle converts the injection depth to approximately 4 mm, another option whenever 4-mm needles are not available.
Chinese patients have a notably high needle reuse rate—nearly 78% of patients (Table 3). Reusing needles is a common practice also in ROW—slightly under one-half. Patients often cite convenience and cost concerns as the reasons. We found the latter to be the main factor in China, related to inconsistent reimbursement policies for pen needles across cities and provinces. Quite a few studies have now shown an association of needle reuse with LH [16, 43,44,45,46,47], especially when the reuse is excessive (≥ 5 times/needle)  as it is in China (44%). In the China LH study by Ji et al. , median reuse frequency was 13 uses/single needle in LH+ vs. 7.5 in LH- patients. Injection pain, which was present in a majority of Chinese patients (Table 4), can also be associated with needle reuse . Of note, only 4% of Chinese ITQ respondents indicated pain with each injection or several times per week—indicating that today’s pens with shorter and narrower diameters are acceptable to the patients using them in terms of injection comfort.
A subset of Chinese respondents to the ITQ reported an episode of DKA in the last 6 months in 16.2% compared with 17.5% in ROW (Table 6). This frequency appears somewhat high, given that < 6% of respondents in China had type 1 diabetes. However, the reported rate was similar in ROW patients, ~ 35% of whom have type 1 diabetes. These findings warrant additional evaluation.
The Chinese Diabetes Society has taken the lead on insulin injection education over the last decade. For a number of parameters in our study, China is ahead of ROW, and this may be a consequence of their efforts. Education in this field appears to work. A UK study  showed that education focused on systematic site rotation, use of the 4-mm pen needle and non-reuse of needles led to the disappearance or reduction in size of a significant percentage of LH lesions after only 6 months. Mean patient HbA1c values decreased by over 4 mmol/l and the prevalence of glucose variability and unexplained hypoglycemia fell as well. Mean TDD values decreased by 5.6 IU. Italian , French  and Russian  studies showed similar results. Currently, there is a randomized controlled interventional study ongoing in Tianjin, China, testing the efficacy of injection training and 4-mm needle use among Chinese injectors with LH. Results should be available soon.
Limitations of the Study
An important limitation of our data is that the results are based on surveys completed by patients/families and may be subject to recall bias, especially as it is related to recall of injection education that may have occurred months or years earlier and on a physical examination of injection sites at clinical visits. We have tried to mitigate this bias by including a nurse questionnaire that asked many of the same questions and attempted to verify, or not, the patients’ answers.