Needle-Free Jet Injection of Rapid-Acting Insulin Improves Early Postprandial Glucose Control in Patients with Diabetes

Clinical Care/Education/Nutrition/Psychosocial Research ORIGINAL ARTICLE

Needle-Free Jet Injection of Rapid-Acting Insulin Improves Early Postprandial Glucose Control in Patients With Diabetes

ELSEMIEK E.C. ENGWERDA, BSC to at least occasionally skipping insulin in- CEES J. TACK, MD, PHD jections or restricting the number of daily BASTIAAN E. DE GALAN, MD, PHD injections (4). Although true needle phobia is rare, many patients with diabetes perceive insulin injections as painful or expe- d OBJECTIVE Clamp studies have shown that the absorption and action of rapid-acting in- rience some form of anxiety with injections sulin are faster with injection by a jet injector than with administration by conventional pen. To fi (5,6), the presence of which is strongly as- determine whether these pharmacokinetic changes also exist in patients with diabetes and bene t sociated with nonadherence and poorer postprandial glucose control, we compared the pharmacologic profiles of insulin administration by jet injection versus conventional insulin pen after a standardized meal in patients with type 1 glycemic control (7). or type 2 diabetes. Jet injectors for insulin administration provide a needle-free alternative to the use RESEARCH DESIGN AND METHODSdIn a randomized, double-blind, double- of pens or syringes and were originally dummy crossover study, 12 patients with type 1 diabetes and 12 patients with type 2 diabetes developed for patients with needle phobia. received insulin aspart either by jet injection or by conventional pen, in both cases followed by a Administration by jet injection significantly standardized meal. Blood was sampled for 6 h for determination of glucose and insulin levels to fi accelerates absorption of rapid-acting in- calculate pharmacologic pro les. sulin from the subcutaneous area into the RESULTSdInsulin administration by jet injection resulted in shorter time until peak plasma systemic circulation (8). Jet injectors de- insulin level (51.3 6 6.4 vs. 91.9 6 10.2 min; P = 0.003) and reduced hyperglycemic burden liver insulin at a high velocity (typically 2 during the first hour (154.3 6 20.8 vs. 196.3 6 18.4 mmol z min z L 1; P = 0.041) compared .100 m/s) directly across the skin in the with conventional administration. Jet injection did not, however, significantly reduce the hy- subcutaneous tissue and dispense the insu- perglycemic burden during the 5-h period thereafter. There was no indication that the jet injector lin over a larger area than does injection by performed differently in patients with type 1 and type 2 diabetes. syringe (9). With the euglycemic clamp CONCLUSIONSdThe considerably more rapid insulin absorption after administration by jet technique, we recently showed in healthy injector translated to a significant if modest decrease in postprandial hyperglycemia in patients volunteers that administration of insulin with type 1 and type 2 diabetes. The improved early postprandial glucose control may specifically aspart by jet injection reduced both the benefit patients who have difficulty in limiting postprandial glucose excursions. time until peak plasma insulin levels and the time to maximal glucose-lowering ef- Diabetes Care 36:3436–3441, 2013 fect by approximately 50% when compared with insulin administered by he pharmacologic profile of rapid- explain why the introduction of rapid- conventional insulin pen (10). acting insulin analogs, although con- acting analogs has had little effect on Although the euglycemic clamp tech- T nique is a reliable method to investigate siderably faster than regular insulin, is HbA1c in people with diabetes (2). Some still relatively slow compared with the have therefore suggested that these analogs the pharmacodynamics of therapeutic in- profile of endogenous insulin release. As a should be injected at least 15 min before sulin, it cannot be used to predict the consequence, patients with type 1 diabetes meals (3); however, this seems impractical glucose-lowering effect of insulin when or insulin-requiring type 2 diabetes who to implement in daily practice. injected before a meal, particularly in use these analogs still face the risk of im- Poor adherence to insulin therapy be- patients with diabetes. The aim of the mediate postprandial hyperglycemia and cause of injection-related anxiety may be current study was therefore to investigate late postprandial hypoglycemia. In partic- another, often neglected, reason for fail- the pharmacology of insulin injected by ular, postprandial hyperglycemia has been ure to reach glycemic targets with current jet injector before a standardized meal in recognized as an important contributor to rapid- and long-acting insulin analogs (4). patients with type 1 diabetes and insulin- suboptimal glucose control (1), which may A sizable proportion of insulin users admit requiring type 2 diabetes. We also wanted to investigate whether patients would ccccccccccccccccccccccccccccccccccccccccccccccccc perceive insulin administration with the current jet injector device as more or less From the Department of Internal Medicine, Radboud University Nijmegen Medical Centre, Nijmegen, the painful as insulin injection by pen. Netherlands. Corresponding author: Bastiaan E. de Galan, [email protected]. Received 27 February 2013 and accepted 20 May 2013. RESEARCH DESIGN AND DOI: 10.2337/dc13-0492. Clinical trial reg. no. NCT01438632, clinicaltrials.gov. d © 2013 by the American Diabetes Association. Readers may use this article as long as the work is properly METHODS Written, informed con- cited, the use is educational and not for profit, and the work is not altered. See http://creativecommons.org/ sent was obtained from 12 patients with licenses/by-nc-nd/3.0/ for details. type 1 diabetes and 12 patients with

3436 DIABETES CARE, VOLUME 36, NOVEMBER 2013 care.diabetesjournals.org Engwerda, Tack, and de Galan insulin-treated type 2 diabetes. Partici- insulin pen (NovoPen III; Novo Nordisk) times until reaching 50% of the C-INSmax pants were recruited from the outpatient and a comparable volume of placebo as insulin levels rose and until reaching the diabetes clinic of the Radboud University solution (Test Medium Penfill; Novo same value as insulin levels declined. The Nijmegen Medical Centre and by adver- Nordisk) by the alternate device in a pharmacodynamic parameters consisted of tisement in a local newspaper. All patients double-blind fashion, both administered the area under the baseline-subtracted were at least 18 years of age and had a BMI subcutaneously in the abdomen. On the plasma glucose concentration–time curve 2 ,32 kg/m and HbA1c #9%. Patients other occasion, the devices containing during the first hour (BG-AUC1h)andfirst were excluded if they had experienced a the insulin and placebo solution were 2 h (BG-AUC2h) after insulin injection, rep- major vascular event (e.g., myocardial in- reversed. Thus all participants received resenting the initial glycemic load, the max- farction, stroke, symptomatic peripheral both insulin and placebo on the two imal glucose excursion (BGmax), the area artery disease, coronary artery bypass experimental days (“double dummy”). under the total baseline-subtracted surgery, or percutaneous coronary or pe- The dose of insulin was individualized plasma glucose concentration time-curve ripheral artery angioplasty) in the previ- to the patient’s usual prandial insulin re- (BG-AUC6h), and the time until plasma ous 6 months; used immunosuppressant quirements and averaged 17.6 6 3.9 glucose had returned to baseline. Safety pa- agents, nonsteroidal anti-inflammatory units (8–40 units, 16.3 6 3.9 units for rameters included the number of patients drugs, anticoagulant therapy, or oral antidi- patients with type 1 diabetes and 18.9 6 requiring exogenous glucose infusion to abetic drugs other than metformin for type 2 11.1 units for patients with type 2 diabe- prevent postprandial hypoglycemia, the diabetic patients; or had symptomatic dia- tes; P =0.44).Two-by-twoblockran- amount of exogenous glucose required, betic neuropathy. Pregnancy was excluded domization was used to randomize the and the duration that exogenous glucose where appropriate. The study was approved sequence by which the two devices were was required. by the institutional review board of the used for insulin and placebo injections. Plasma glucose was measured with the Radboud University Nijmegen Medical To ensure blinding, both pen devices glucose enzymatic-amperometric method Centre. were prepared by a nurse who was not (Biosen C-line GP+; EKF-diagnostic All participants underwent two stan- otherwise involved in conducting the ex- GmbH, Barleben, Germany) during the dardized meal tests, separated by at least 2 periments. Insulin administration with experiments. Blood sampled for determi- weeks. Patients were requested to reduce both devices was performed by trained nation of plasma insulin levels was collected the evening dose of insulin or the basal staff only, as described in detail previ- in lithium-heparin tubes. After centrifuga- rate of insulin pump administration by ously (10). tion, the supernatant was stored at 2208C. 10–20% to avoid nocturnal hypoglyce- One minute after insulin administra- Plasma insulin was measured by radioim- mia and were instructed to consume a tion, the participant consumed a standard- munoassay (12). low–glycemic index meal on the evenings ized meal consisting of three white-bread before the experimental days. On each ex- sandwiches with marmalade and honey Power calculation and statistical perimental day, participants were ad- and a glass of orange juice (total energy analyses mitted to the research unit at 0730 h in 538 kcal, 108 g carbohydrates, 7 g fat, All data are expressed as mean 6 SEM un- fasting condition and having abstained 11 g protein) in 10–15 min. Plasma less otherwise indicated. Mean outcomes from smoking, alcohol use, and caffeine glucose measurements were measured at for all study end points and most safety use for at least 24 h. Patients with subcu- 5-min intervals during the first 3 h of the end points were tested with paired t tests. taneous insulin pumps were asked to stop study and at 10-min intervals for another A x2 test was used to compare the numbers the pump. The experiments were per- 3 h. Blood for plasma insulin levels was of patients requiring exogenous glucose formed with the patients in supine posi- sampled every 5 min during the first with the two injection devices. Continuous tion in a temperature-controlled room hour, every 15 min for the second hour, data were tested for normal distribution (22–248C). Two catheters were inserted and every 30 min thereafter. When (Shapiro-Wilk test and Kolmogorov- intravenously. One catheter was inserted plasma glucose values dropped below Smirnov test) and subsequently analyzed in retrograde fashion in a dorsal hand vein 4.8 mmol/L, 20% dextrose in water was with repeated measures ANOVA (Wilks for blood sampling, with the vein kept pat- administered intravenously to maintain L test), with the device as between-subjects ent by placing the hand in a heated box at normoglycemia. On the second experi- factor. P , 0.05 was considered statistically 558C (11). The other catheter was placed mental day, patients were asked to rate significant. All statistical analyses were per- in an antecubital vein of the contralateral the amounts of discomfort or pain experi- formed with SPSS software (Statistical arm for insulin and glucose administra- enced with the two administration meth- Package for the Social Sciences, version tion. After baseline variables were ob- ods on a visual analog scale from 0 to 18.0; IBM Corporation, Armonk, NY). tained, low-dose insulin was infused to 10 cm and to indicate which device they achieve normoglycemia, after which insu- would prefer should they have a choice. RESULTSdTable 1 summarizes the lin infusion was either terminated or, for All pharmacologic parameters were baseline characteristics of the partici- patients treated with insulin pumps, con- derived from the plasma glucose and in- pants. All 24 subjects completed the tinued at a rate corresponding to the basal sulin levels. With respect to the pharma- meal tests. The first test had to be resched- rate of the patient’s insulin pump. cokinetics, we calculated the time to uled in two cases. In one instance, the in- Thirty minutes after achievement of maximal insulin concentration (T-INSmax), sulin dose administered was erroneously stable normoglycemia, patients received the maximal insulin concentration calculated too low, and in the other, the insulin (aspart; Novo Nordisk, Bagsvaerd, (C-INSmax), the area under the insulin spring of the jet injector released prema- Denmark) either by jet injection (Insujet; concentration curve (C-INSAUC), and turely,sothatitcouldnotbeassessed European Pharma Group bv, Schiphol-Rijk, the time until 50% of insulin absorption how much insulin (or placebo), if any, the Netherlands) or by conventional (T-INSAUC50%). We also calculated the had actually crossed the skin. That jet care.diabetesjournals.org DIABETES CARE, VOLUME 36, NOVEMBER 2013 3437 Jet injection of insulin in diabetes

Table 1dBaseline characteristics (10.5 6 0.6 vs. 10.8 6 0.7 mmol/L; P = 0.65) or directly before the experiments 6 6 Whole group Type 1 diabetes Type 2 diabetes (5.61 0.13 vs. 5.45 0.18 mmol/L; P = 0.49). The time-action curves for the Male-to-female ratio 20:4 9:3 11:1 plasma glucose level after meal ingestion Age* (years) 49.9 6 16.0 39.2 6 16.2 60.6 6 7.2 were significantly different between the BMI (kg/m2)26.46 2.1 25.7 6 31.9 27.0 6 2.1 two devices (P = 0.018 by ANOVA). In 6 6 6 HbA1c (%) 7.5 0.8 7.3 1.0 7.7 0.6 line with the faster insulin pharmacoki- Daily insulin dose (units) 57.2 6 32.5 48.6 6 15.2 65.1 6 42.0 netics, the hyperglycemic burden during Duration of diabetes (years) 18.3 6 11.9 21.5 6 15.0 15.2 6 7.0 the first hour, as reflected by the area un- Systolic blood pressure (mmHg) 135.1 6 12.3 127.5 6 8.7 142.7 6 10.7 der the glucose concentration curve 6 6 6 Diastolic blood pressure (mmHg) 76.9 10.2 75.2 9.7 78.6 10.8 (BG-AUC1h)wassignificantly reduced Insulin regimen when insulin was administered with the Insulin pump 5 5 0 jet injector rather than by conventional Basal-bolus regimen 17 7 10 pen (Fig. 1B and Table 2). This benefit Premixed insulin 2 0 2 favoring the jet injector relative to the Use of metformin 8 0 8 conventional insulin pen was no longer Data n or mean 6 SD. *P , 0.05 for comparison between patients with type 1 and 2 diabetes. apparent after 2 h (Table 2). Although glucose values tended to be lower in the late postprandial phase after conventional pen administration, there were no injector was subsequently returned to the the jet injector (P = 0.003), a difference of fi 6 A signi cant differences between the two manufacturer and replaced. 40.6 12.3 min (Fig. 1 ,Table2).In devices with regard to the maximal glu- addition, T-INSAUC50% and the time for cose value, maximal glucose increment, Pharmacokinetic end points insulin levels to decline after reaching fi or area under the 6-h glucose concentra- Plasma insulin levels at initiation of in- peak values were signi cantly shorter tioncurve(Table2). sulin injections were slightly higher for when insulin was administered by jet in- the experiments where the conventional jection rather than by conventional pen pen was used than those for the jet in- (Table 2). In contrast, the total amount Safety jector (22.8 6 2.5 vs. 19.4 6 2.4 mU/L; of insulin absorbed during the entire 6-h There were no differences between the jet P = 0.037). Thereafter, plasma insulin levels period, as reflected by the C-INSAUC, did injector and conventional pen with re- increased faster when insulin was injected not differ between the two devices. spect to number of patients requiring with the jet injector than when injected by exogenous glucose to prevent hypoglyce- conventional pen (Table 2). The time until Pharmacodynamic end points mia (17 vs. 18; P = 0.75), the timing of peak plasma insulin concentrations was ad- There were no differences in plasma initiation of glucose administration (180 vanced from 91.9 6 10.2 min with the glucose values between the two experi- vs. 194 min; P = 0.79), or the amount of conventional pen to 51.3 6 6.4 min with mental conditions, either at baseline glucose administered (21.0 6 5.5 g vs. 23.7 6 5.7 g; P = 0.61). Both injection methods were well tolerated and elicited similar experiences of pain or discomfort Table 2dPharmacokinetic and pharmacodynamic parameters for insulin administration (visual analog scale 1.96 vs. 1.40; P = with the jet injector and the conventional insulin pen 0.14). Of the total of 24 patients, 13 preferred the conventional pen to the jet in- Jet injector Conventional pen P value jector, 4 preferred the jet injector, and 6 Pharmacokinetic parameters remained indifferent. One patient did not complete the questionnaire. T-INSmax50% (min) 13.9 6 1.3 31.8 6 3.2 ,0.001 6 6 T-INSmax (min) 51.3 6.4 91.9 10.2 0.003 6 6 T-INSAUC50% (min) 107.8 5.5 123.1 5.7 0.011 Subgroup analysis 6 6 C-INSmax (pmol/L) 98.8 8.9 103.3 8.4 0.53 The pharmacodynamic benefits of insulin 6 6 T-INSdec50% (min) 143.3 9.3 165.6 10.0 0.0497 administration by jet injection tended to z z 21 6 6 C-INSAUC (pmol min L ) 11,282.0 966.5 11,791.4 1,063.8 0.57 be numerically higher in patients with Pharmacodynamic parameters type 2 diabetes; however, the difference z z 21 6 6 BG-AUC1h (mmol min L ) 154.3 20.8 196.3 18.4 0.04 was not statistically different between z z 21 6 6 BG-AUC2h (mmol min L ) 475.5 68.0 554.4 65.7 0.18 patients with type 1 and type 2 diabetes. z z 21 6 6 BG-AUC6h (mmol min L ) 1,283.8 134.6 1,360.5 147.1 0.64 For type 1 diabetes, BG-AUC2h after jet 6 6 BGmax (mmol/L) 12.6 0.84 13.3 0.82 0.36 injection was 478 6 117 vs. 500 6 98 2 6 6 21 BGmax baseline (mmol/L) 7.0 0.8 7.9 0.8 0.27 mmol z min z L after conventional ad- 6 6 T-BGBL (min) 216.1 17.54 223.3 17.79 0.58 ministration (P = 0.87); values for type 2 Exogenous glucose (g) 22.2 6 5.9 23.3 6 5.9 0.77 diabetes were 471 6 82 vs. 662 6 96 z z 21 P 6 mmol min L ( = 0.09) (Fig. 2). There Data are mean SEM. T-INSdec50%, time to reach 50% of the C-INSmax as insulin levels declined; fi T-INSmax50%, time to reach 50% of the C-INSmax as insulin levels rose; T-BGBL, time until plasma glucose had were no differences in subgroups de ned returned to baseline. by age, sex, or BMI (data not shown).

3438 DIABETES CARE, VOLUME 36, NOVEMBER 2013 care.diabetesjournals.org Engwerda, Tack, and de Galan

Figure 2dThe 2-h BG-AUC values for the jet injector (white bars) and conventional pen (black bars) in subgroups according to type of diabetes.

between 5.9 and 22.9 mmol/L among subjects and between 0.1 and 12.9 mmol/L within subjects. Glucose levels at admission at the research unit were similarly variable, although all patients had consumed a low–glycemic index sup- per on the evening before the experiments. Factors contributing to this large variation may include the heterogeneity of the participants, the suboptimal glycemic control, and use of insulin by all participants, including those with type 2 diabetes. A second explanation concerns the finding that insulin absorption was much slower in the patients than we had previously observed in healthy volunteers (10). Although jet injection advanced insulin absorption to roughly similar ex- tents in both groups, insulin levels Figure 1dChanges from baseline during the standardized meal test. A: Changes in plasma in- peaked substantially later in diabetic sub- ○ - jects with than in those without diabetes sulin levels after insulin administration by jet injector ( ) and conventional pen ( ). B: Changes P in plasma glucose levels after insulin administration by jet injector (○) and conventional pen (51.3vs.30.6min; = 0.012). Because (-). the standardized meal consisted mainly of high–glycemic index carbohydrates, in- sulinabsorptionmayhavebeentoo slow to sufficiently counteract the fast CONCLUSIONSdPreviously, we Considering the substantial enhance- glucose load absorption. Whether jet in- showed in young, healthy volunteers ment of insulin action by jet injection in jection would have performed better for a that absorption and action of aspart in- our clamp study (10) and the current im- more usual mixed meal with slower food sulin occurred twice as fast when adminis- provement of insulin absorption, we absorption cannot be determined from tered by jet injection as by conventional anticipated a sizeable pharmacodynamic our data. Parenthetically, most patients pen (10). The current study confirms a benefit of insulin administration by jet in- in our study claimed never to consume more rapid absorption of insulin aspart jection in the current study. Although the high–glycemic index carbohydrates in when administered by the current jet injec- lower early postprandial glucose levels such large quantities, to avoid extreme tor compared with a conventional pen in and tendency toward lower glucose levels glucose excursions. patients with type 1 and type 2 diabetes. in the later postprandial phase confirmed Why rapid-acting insulin absorption These pharmacokinetic properties trans- the results of the clamp study under con- is slower in patients with diabetes than in lated into a significant, albeit modest, de- ditions more like an actual clinical situa- subjects without diabetes is unknown. crease in early (first 60 min) postprandial tion, the effects were rather modest. This We previously showed a strong associa- hyperglycemia after a standardized meal partial discrepancy may be explained tion between BMI and rate of insulin rich in carbohydrates. Beyond 1 h, the first by the relatively large variation, absorption in healthy subjects, arguing a benefit of jet injection on postprandial both inter- and intraindividually, in glu- role for greater subcutaneous tissue thick- glycemic burden was no longer statistically cose excursions after the meal test. In- ness (13). In the current patient group, significant. deed, maximal glucose excursions varied however, insulin absorption was unrelated care.diabetesjournals.org DIABETES CARE, VOLUME 36, NOVEMBER 2013 3439 Jet injection of insulin in diabetes to the BMI or any other measure of body has been reported to result in bruises E.E.C.E. performed the experiments, ana- composition. The presence of insulin anti- and “wet” injections, leading to discom- lyzed the data, interpreted the data, drafted the bodies, commonly found in patients on fort and unpredictable insulin absorption first version of the manuscript, edited the fi long-term insulin therapy, has been sug- profiles (21). Importantly, the current de- manuscript, and approved the nal version of gested to attenuate the absorption of sub- vice has a built-in lock-release system that the manuscript. C.J.T. analyzed the data, interpreted the data, edited the manuscript, and cutaneous insulin (14); however, insulin only releases insulin when proper skin fi fi approved the nal version of the manuscript. antibodies were not measured in our study. contact has been made and suf cient B.E.d.G. designed the study, wrote the study Alternatively, subcutaneous adipose tissue pressure has been applied to the nozzle protocol, performed the experiments, ana- blood flow (ATBF) may also affect the ab- of the injector. Other factors that deter- lyzed the data, interpreted the data, edited the sorption of insulin (15). Reduced subcu- mine optimal insulin delivery relate to jet manuscript, and approved the final version of taneous ATBF has been reported in velocity and nozzle diameter (22–24). the manuscript. B.E.d.G. is the guarantor of overweight nondiabetic subjects (16), as Our data are therefore in part specificto this work and, as such, had full access to all the well as in patients with type 1 (17,18) and the jet injector used and cannot be extrap- data in the study and takes responsibility for type 2 diabetes (19). In addition, ATBF olated to other jet injectors. the integrity of the data and the accuracy of the may fail to increase in response to dietary Various strategies are currently under data analysis. The authors thank Karin Saini, Anja Rasing, stimuli in overweight subjects (16) and in development to enhance absorption and and Mariëlle Verstegen (research nurses, Radboud patients with type 2 diabetes (19). action of subcutaneous insulin. These University Nijmegen Medical Centre) for their Although the pharmacodynamic ben- strategies include local skin heating to assistance during the clamps, Petra van de Ven efit of jet injection appears small, it is of stimulate tissue perfusion (25), coadmin- and Sandra Hendriks (diabetes nurses, Radboud potential clinical relevance for patients istration with hyaluronidase to break the University Nijmegen Medical Centre) for pre- with early postprandial glucose excur- solidity of the extracellular matrix (26), paring the insulin pens, and the patients for sions not sufficiently covered by conven- and destabilizing insulin hexamer forma- their participation in this study. The authors tional insulin injections. Administration tion by addition of EDTA and citric acid also thank Tim Heise, Profil Institut fur€ of rapid-acting insulin by this jet injector (27). The pharmacokinetic and pharma- Stoffwechselforschung GmbH, Neuss, Germany, may represent an especially valuable al- codynamic benefits of these interventions for his advice on study methodology and careful reading of the manuscript. ternative for patients with type 2 diabetes, are more or less comparable to those of jet in whom the postprandial glycemic benefit injection and range from ;10 to ;60% of jet injection tended to be more pro- advancement of peak insulin levels and – References nounced than in those with type 1 diabetes. maximal glucose-lowering effect (25 27). 1. Monnier L, Colette C, Owens D. Post- Parenthetically, a device that performs at A difference between jet injection and other prandial and basal glucose in type 2 di- least as well as, and potentially better than, developments to enhance insulin absorp- abetes: assessment and respective impacts. conventional pens may be a good option tion is that the former is already available Diabetes Technol Ther 2011;13(Suppl. 1): for any patient who does not tolerate in- for clinical application, whereas the latter S25–S32 sulin injections by needle or regards these are in still in various stages of development 2. Siebenhofer A, Plank J, Berghold A, et al. as otherwise uncomfortable. A recent and have not yet been marketed. Short acting insulin analogues versus reg- survey showed that 28.6% of patients In the current study, we found that jet ular human insulin in patients with diabetes with type 2 diabetes perceive insulin injection accelerated the absorption of in- mellitus. Cochrane Database Syst Rev 2006; 19:CD003287 injections as painful (20), putting such sulin aspart in patients with type 1 and 3. Rassam AG, Zeise TM, Burge MR, Schade patients at high risk of skipping at least insulin-treated type 2 diabetes. This better DS. Optimal administration of lispro in- occasional injections (7). When ques- pharmacokinetic profile was initially fol- sulin in hyperglycemic type 1 diabetes. tioned, most patients still preferred the lowed by a congruent reduction in glucose Diabetes Care 1999;22:133–136 conventional pen to the jet injector; how- excursions after a high-glycemic index meal, 4. Peyrot M, Rubin RR, Kruger DF, Travis LB. ever, these patients were unselected, gen- in particular during but not beyond the first Correlates of insulin injection omission. erally tolerated conventional injections hour. Jet injection may therefore be a good Diabetes Care 2010;33:240–245 5. Gale EA. Two cheers for inhaled insulin. without discomfort, consequently feeling needle-free alternative to conventional in- – no need to change the mode of adminis- sulin pens of at least equivalent pharmaco- Lancet 2001;357:324 325 tration, and did not handle the jet injector logical efficacy for the administration of 6. Peyrot M, Rubin RR. Levels and risks of depression and anxiety symptomatology (or the conventional pen) themselves un- insulin in patients with diabetes. Future among diabetic adults. Diabetes Care 1997; der study conditions. Future research research is needed to determine whether the 20:585–590 will need to reveal tolerability of the jet better pharmacologic properties of insulin 7. Aronson R. The role of comfort and dis- injector after personal experience with jet injection translate into beneficial long- comfort in insulin therapy. Diabetes Tech- the device in daily practice for a longer term effects on glycemic control and risk of nol Ther 2012;14:741–747 period. hypoglycemia in patients with diabetes. 8. Malone JI, Lowitt S, Grove NP, Shah SC. Handling a jet injector may be more Comparison of insulin levels after injection cumbersome than handling a conven- by jet stream and disposable insulin sy- fi ringe. Diabetes Care 1986;9:637–640 tional insulin pen. It requires suf cient d training in air-free filling of the chamber Acknowledgments European Pharma Group 9. Mitragotri S. Current status and future bv, Schiphol-Rijk, the Netherlands, funded the prospects of needle-free liquid jet injectors. with insulin and correct placement of the study. No other potential conflicts of interest Nat Rev Drug Discov 2006;5:543–548 injector on the skin to ensure that the relevant to this article were reported. 10. Engwerda EE, Abbink EJ, Tack CJ, de entire volume of insulin reaches the sub- The funder, European Pharma Group bv, Galan BE. Improved pharmacokinetic and cutaneous compartment. Inadequate was not involved in the design or execution of pharmacodynamic profile of rapid-acting contact between injector and the skin the study or in the writing of the manuscript. insulin using needle-free jet injection

3440 DIABETES CARE, VOLUME 36, NOVEMBER 2013 care.diabetesjournals.org Engwerda, Tack, and de Galan

technology. Diabetes Care 2011;34:1804– adipose tissue may accumulate the fat that 22. Stachowiak JC, Li TH, Arora A, Mitragotri 1808 has been recycled as VLDL and non- S, Fletcher DA. Dynamic control of needle- 11. Abumrad NN, Rabin D, Diamond MP, esterified fatty acids. Diabetes 2010;59: free jet injection. J Control Release 2009; Lacy WW. Use of a heated superficial hand 2465–2473 135:104–112 vein as an alternative site for the measure- 17. Hildebrandt P. Skinfold thickness, local 23. Schramm-Baxter J, Mitragotri S. Needle- ment of amino acid concentrations and for subcutaneous blood flow and insulin ab- free jet injections: dependence of jet pene- the study of glucose and alanine kinetics in sorption in diabetic patients. Acta Physiol tration and dispersion in the skin on jet man. Metabolism 1981;30:936–940 Scand Suppl 1991;603:41–45 power. J Control Release 2004;97:527–535 12. de Galan BE, Tack CJ, Lenders JW, et al. 18. Vora JP, Burch A, Peters JR, Owens DR. 24. Linn L, Boyd B, Iontchev H, King T, Farr Theophylline improves hypoglycemia Absorption of radiolabelled soluble insulin SJ. The effects of system parameters on in unawareness in type 1 diabetes. Diabetes in type 1 (insulin-dependent) diabetes: in- vivo injection performance of a needle- 2002;51:790–796 fluence of subcutaneous blood flow and free injector in human volunteers. Pharm 13. de Galan BE, Engwerda EE, Abbink EJ, anthropometry. Diabet Med 1993;10:736– Res 2007;24:1501–1507 Tack CJ. Body mass index and the efficacy 743 25. Raz I, Weiss R, Yegorchikov Y, Bitton G, of needle-free jet injection for the admin- 19. Tobin L, Simonsen L, Bulow€ J. The dy- Nagar R, Pesach B. Effect of a local heating istration of rapid-acting insulin analogs, a namics of the microcirculation in the device on insulin and glucose pharmaco- post hoc analysis. Diabetes Obes Metab subcutaneous adipose tissue is impaired kinetic profiles in an open-label, ran- 2013;15:84–86 in the postprandial state in type 2 diabetes. domized, two-period, one-way crossover 14. Radermecker RP, Renard E, Scheen AJ. Clin Physiol Funct Imaging 2011;31:458– study in patients with type 1 diabetes us- Circulating insulin antibodies: influence 463 ing continuous subcutaneous insulin in- of continuous subcutaneous or intraperi- 20. Peyrot M, Barnett AH, Meneghini LF, fusion. Clin Ther 2009;31:980–987 toneal insulin infusion, and impact on Schumm-Draeger PM. Factors associated 26. Muchmore DB, Vaughn DE. Review of the glucose control. Diabetes Metab Res Rev with injection omission/non-adherence mechanism of action and clinical efficacy 2009;25:491–501 in the Global Attitudes of Patients and of recombinant human hyaluronidase 15. Vora JP, Burch A, Peters JR, Owens DR. Physicians in Insulin Therapy study. Di- coadministration with current prandial Relationship between absorption of ra- abetes Obes Metab 2012;9999:1081– insulin formulations. J Diabetes Sci Tech diolabeled soluble insulin, subcutaneous 1087 2010;4:419–428 blood flow, and anthropometry. Diabetes 21. Schramm J, Mitragotri S. Transdermal 27. Steiner S, Hompesch M, Pohl R, et al. A Care 1992;15:1484–1493 drug delivery by jet injectors: energetics of novel insulin formulation with a more 16. McQuaid SE, Humphreys SM, Hodson L, jet formation and penetration. Pharm Res rapid onset of action. Diabetologia 2008; Fielding BA, Karpe F, Frayn KN. Femoral 2002;19:1673–1679 51:1602–1606

care.diabetesjournals.org DIABETES CARE, VOLUME 36, NOVEMBER 2013 3441