DIABETES TECHNOLOGY & THERAPEUTICS Volume 16, Supplement 1, 2014 ª Mary Ann Liebert, Inc. DOI: 10.1089/dia.2014.1506

ORIGINAL ARTICLE

Insulin Pens and New Ways of Insulin Delivery Lutz Heinemann

Introduction

USAGE OF PENS IN DAILY LIFE

O

Dose accuracy of new versus used Novopen 4 insulin pens

ne has to acknowledge that not too much was published during the last year in this area of research that is worth being reported—at least when it comes to the results of clinical trials. However, the number of ‘‘reviews’’ and commentaries about noninvasive insulin delivery appears to be as high as in the previous years. So, interest in the topic is indeed there, but the progress made is simply not overwhelming. Interestingly enough, the number of publications on insulin pens was also lower than that in the years before. Nevertheless, a number of small biotechnology companies are still active in developing insulin products that can be administered by other routes than the subcutaneous (SC) route. Even the big players in the insulin arena are keeping their eyes open for promising developments, for example, intradermal (ID) administration or oral insulin. Insulin Pens A year ago, in the book chapter summarizing the data published in 2011/2012, I lamented the lack of studies on the usage of insulin pens in daily life. Happily enough, some studies looking into this were published during the last year. The assumption is that patients use pens according to the instructions of the manufacturer; however, when you talk with experienced diabetes nurses, they can tell you numerous stories—not all of them positive—about the reality of pen usage (the same holds true for all medical devices). Some patients use pens for years that are designed for multiple usages but appear to never change the needle during this period of time. Some interesting publications about needles, ID insulin delivery, and novel approaches for microneedles have been published. It remains to be seen if such approaches will make their way into clinical practice. Another question is: what do we actually know about what patients want when it comes to the design and handling of pens? It might very well be that companies know the answers to such questions but do not publish them, because they regard them as proprietary information.

Yucel H 1, Taks M 2, Menheere P 3, Grouls R 2, Bravenboer B 1 Departments of 1Internal Medicine and 2Clinical Pharmacy, Catharina Hospital, Eindhoven; and 3Maastricht University, The Netherlands Diabetes Technol Ther 2012; 14: 810–12

Background The number of studies evaluating dose accuracy in both new and used insulin injection pens is limited. We hypothesized that the dose accuracy of used insulin pens ( > 1 year) is less accurate than that of new insulin pens and studied if such differences influence the treatment. This study compared the dosing accuracy of 11 new and 11 used Novopen 4 pens. Methods Dosing accuracy differences between new and used pens were studied by weighing the volume of the dosage of 8 and 32 IU; each measurement was repeated 15 times. It was tested whether the pens complied with the ISO limits of 10% for 8 IU and 5% for 32 IU. Results For the 8 IU dose, the mean delivered dose was 8.04 IU in new pens and 7.91 IU in used pens; for the 32 IU dose, the mean delivered dose was 31.90 IU and 31.68 IU. The difference in the median values between the pens was significant ( p < 0.001). Three individual doses in the 32 IU dose exceeded the International Organization for Standardization (ISO) range in the lower range. The difference in mean variation coefficient between the two groups was also significant ( p < 0.001). Conclusions There was a significant difference between the accuracy of new versus used insulin pens. More studies with larger

Science & Co., Du¨sseldorf, Germany.

S-44

INSULIN PENS AND NEW WAYS OF INSULIN DELIVERY

S-45

sample sizes are necessary to confirm our findings and further elucidate the relationship between age of insulin pens and dose accuracy.

storing an in-use pen. For three-quarters of the insulin pens being used, users did not follow the manufacturer’s instructions for proper administration and storage of insulin pens. Correct usage scores were higher if initial education was performed by a pharmacist or nurse.

Comment Allow me to applaud the authors of this publication for addressing this important topic. This same significant difference in dosing accuracy between new pens and pens used in daily life over a period of time was also observed in a recent study in Germany. Patients use such devices for several years, and the dosing accuracy is usually never evaluated in daily practice by their treating diabetologist/diabetes nurse. What can we expect: such pens have to be ‘‘inexpensive’’ in terms of production costs and are probably used several times a day under a wide range of environmental conditions but are supposed to maintain the same precision in performance over time? Is this honestly realistic? It is of interest to note that the manufacturer guarantees dose accuracy for 3 years. Clearly this study should be repeated with both a larger sample size and other pens. Patients that show up at an outpatient clinic might make for good candidates for such studies. Another question that merits attention is: how well do such pens work when handled by the patients themselves and not by trained personnel?

Conclusions These data show that a majority of patients may be ignoring or unaware of key components for consistent insulin dosing using disposable insulin pens. A better education and reeducation on correct use of disposable insulin pens are needed. Comment This study also addressed the practice of daily pen usage, and it shows quite convincingly that many, if not most, of the patients do not use the pens according to the instructions of the manufacturer under daily circumstances. However, some of the instructions are either not practical or are cost intensive; for example, using a new PN for each insulin injection is most often simply not reimbursed. It remains to be studied if a group of patients who adhere carefully to instructions differs in any clinically relevant manner from a group that uses its pens without paying attention to the instructions and as it deems correctly. One wonders how well patients are trained in pen usage and who is performing this training step, which should take place when patients are switched to insulin or from using a syringe/vial to a pen. To my knowledge, this was never evaluated in daily practice. Also, when patients are switched from one pen to the other, the differences between these pens need to be adequately trained. It would be of interest to see data from many more patients from different settings and countries. It might very well be that certain differences exist in pen usage and in quality of this usage according to patient education strategies in various countries.

Administration technique and storage of disposable insulin pens reported by patients with diabetes Mitchell VD 1,2, Porter K 3, Beatty SJ 1 1

Division of Pharmacy Practice and Administration, Ohio State University College of Pharmacy, Columbus, OH; 2The Department of Pharmacy, Wexner Medical Center of the Ohio State University, Columbus, OH; and 3The Center of Biostatistics, The Ohio State University, Columbus, OH Diabetes Educ 2012; 38: 651–58

Background The aim of this study was to evaluate insulin injection technique and storage of insulin pens as reported by patients with diabetes. In addition, usage of pens was evaluated, as well as HbA1c, duration of insulin therapy, and duration of insulin pen usage. Methods A questionnaire was administered to patients at a university-affiliated primary care practice. Patients were 18 years or older and used a disposable insulin pen for at least 4 weeks. A correct usage score was calculated for each patient based on manufacturer recommendations for disposable insulin pen use. Results Sixty-seven patients completed the questionnaire, reporting total use of 94 insulin pens. The three components most often neglected were priming pen needle (PN), holding for specific count time before withdrawal of PN from skin, and

NEEDLES Insulin pen needles: effects of extra-thin wall needle technology on preference, confidence, and other patient ratings Aronson R 1, Gibney MA 2, Oza K2, Be´rube´ J 2, Kassler-Taub K 2, Hirsch L 2 1

LMC Diabetes & Endocrinology, Toronto, Ontario, Canada; and Beckton, Dickinson and Company, Franklin Lakes, NJ

2

Clin Ther 2013; 35: 923–33

Background PNs are essential for insulin injections using pens. The properties of these needs affect patients’ injection experience. The goal of this study was to evaluate the impact of a new extra-thin wall (XTW) PN versus standard PNs on patient preference, ease of injection, perceived time to complete the full dose, thumb button force to deliver the injection, and dose delivery confidence in patients with diabetes. The patients

S-46 injected insulin with the KwikPen (Eli Lilly), SoloSTAR (sanofi-aventis), and FlexPen (Novo Nordisk) insulin pens. Methods Quantitative testing of the XTW and comparable PNs with the three insulin pens was performed to evaluate thumb force, flow rate, and time to deliver medication. Subsequently, a prospective, randomized, 2-period, open-label, crossover trial was then conducted in patients with type 1 or type 2 diabetes mellitus who injected insulin by pen for ‡ 2 months, with at least one daily dose ‡ 10 U. Patients who used 4–8-mm-long PNs with 31–32G diameter were randomly assigned to use their current PN or the same/similar size XTW PN at home for *1 week and the other PN the second week. Results XTW PNs had better performance for each studied PN characteristic (thumb force, flow, and time to deliver medication) for all pens combined and each individual pen brand (all, p £ 0.05). Data sets of 198 patients randomized to study groups (80, SoloSTAR; 77, FlexPen; 59, KwikPen) were evaluable. Nearly all of these patients used a single PN. Patients rated the XTW PNs (mean [95% CI]) as preferable and it required less thumb force and less time to inject the dose, and was rated as providing greater confidence in full dose delivery. Results were similar for each of the three pens, those with impaired hand dexterity, and for all users of 4 mm PNs. Skin leakage and insulin dripping from the needle tip were rated as less frequent with the XTW PNs ( p < 0.05). Conclusions PNs with thin walls were preferred overall, rated as requiring less time and less thumb force to inject, and providing greater confidence in completing a full dose compared with usual PNs in this group of patients with type 1 or type 2 diabetes mellitus. Comment The advantage of shorter and thinner PNs is that the pain associated with insulin administration is reduced; this is also influenced by the sharpness of the needle tip. However, if the inner diameter of the needle is reduced too much, the pressure and time required to press the insulin formulation through increases too much. By reducing the thickness of the walls of the needle the inner diameter can be kept high. The development of thinner and shorter needles is ongoing; however, it appears that some technical limits have been reached. At least with the materials currently available—that is, steel—the walls cannot be made thinner without losing the necessary stability.

Comparison of the effects of a new 32-gauge · 4-mm pen needle and a 32-gauge · 6-mm pen needle on glycemic control, safety, and patient ratings in Japanese adults with diabetes Miwa T, Itoh R, Kobayashi T, Tanabe T, Shikuma J, Takahashi T, Odawara M

HEINEMANN Department of Diabetes, Endocrinology and Metabolism, Tokyo Medical University, Tokyo, Japan Diabetes Technol Ther 2012; 14: 1084–90

Background A prospective, open-label, controlled crossover study was performed to evaluate the potential effects of two PNs with the same diameter but different lengths (4 and 6 mm) and different needle tip shapes (straight and tapered) on glycemic control, perceived pain, safety, patients’ ease of use and preferences, and visual impression. Methods Forty-one insulin-treated patients with type 1 or type 2 diabetes were randomized into either Group 1 (4 mm PN in Study Period 1; 6 mm PN in Study Period 2) or Group 2 (the order for using the PNs was reversed). Results The 4 mm PN provided a comparable glycemic control as the 6 mm PN, with an equivalent occurrence rate of adverse events. The 4 mm PN was perceived as less painful and rated as more favorable than the 6 mm PN according to the survey results on patients’ ease of use and preferences and on their visual impressions. Conclusions The shorter PN was not only as safe and efficacious as the 6 mm PN, but also perceived as less painful, easier to use, and more favorable to Japanese adult patients with diabetes. Comment In an editorial accompanying this publication, an experienced U.S. clinician commented on the important progress made in PNs over the last decades and how much this has improved the convenience of pen usage (clearly also the case when insulin is administered by syringe). The term ‘‘convenience’’ is probably not strong enough to clarify how much the barrier of administering insulin several times per day can be reduced if administering insulin does not induce significant pain. That no difference in metabolic control could be observed in this study with Japanese patients is of no surprise; this study repeats results obtained in a larger U.S. study with more obese subjects (mean body mass index [BMI] 31.0 vs. 23.2 kg/m2). The idea that patients prefer the shorter needle has also been reported before. It is of interest to note that this study—and the one presented before— were supported by Becton Dickinson. One important question is how often patients change the PN in daily practice. This—beside its diameter and length—strongly determines how much pain the insulin injection causes. Clearly, PNs contribute to the cost of pen insulin therapy, which is more expensive than using a syringe and a vial. The question is, if the quality of life is improved (i.e., reduced injection pain) with better needles/insulin pen usage, how much does this contribute to

INSULIN PENS AND NEW WAYS OF INSULIN DELIVERY

S-47

an improved metabolic control and thereby reduce costs in the long run by avoiding the development of diabetesrelated late complications? Studies evaluating such questions would be of great interest.

studied for a number of years now, and it appears as if the slow but steady progress toward a practically usable product is taking significant steps in the right direction. Different time intervals between insulin administration and meal ingestion as well as different insulin doses (under- and overdosing) were tested, showing the expected differences in meal-related increases in glycemia. Of interest to note (and worth being studied in more detail) is the reduction in intra- and intersubject variability with ID insulin administration. It remains to be shown in larger clinical studies to what degree the benefits observed in such clinical–experimental trials translate into improvements in metabolic control under daily life conditions.

Pharmacokinetics and postprandial glycemic excursions following insulin lispro delivered by intradermal microneedle or subcutaneous infusion McVey E 1, Hirsch L2, Sutter DE 1, Kapitza C 3, Dellweg S3, Clair J 3, Rebrin K 4, Judge K1, Pettis RJ 1 1

BD Technologies, Research Triangle Park, NC; 2BD Medical– Diabetes Care, Franklin Lakes, NJ; 3Profil Institut fu¨r Stoffwechselforschung GmbH, Neuss, Germany; and 4BD Medical–Diabetes Care, Billerica, MA J Diabetes Sci Technol 2012; 6: 743–54

Background ID delivery by means of microneedles has been shown to accelerate insulin absorption. In this study, the pharmacokinetic (PK) and pharmacodynamic (PD) effects of insulin lispro administered before two daily standardized solid mixed meals (breakfast and lunch) were studied, using ID or traditional SC delivery. Methods Twenty-two patients with type 1 diabetes participated in an eight-arm full crossover block design. One arm established each subject’s optimal meal dose. In six additional arms, the optimal, higher, and lower doses (+ 30%, - 30%) were each given ID and SC delivery, in random order. The final arm assessed earlier timing for the ID optimal dose (-12 versus - 2 min).

Faster pharmacokinetics and increased patient acceptance of intradermal insulin delivery using a single hollow microneedle in children and adolescents with type 1 diabetes Norman JJ 1,2, Brown MR 1, Raviele NA 1, Prausnitz MR 2, Felner EI 1,2 1

Division of Endocrinology, Department of Pediatrics, Emory University School of Medicine, Atlanta, GA; and 2School of Chemical and Biomolecular Engineering, Georgia Institute of Technology, Atlanta, GA Pediatr Diabetes 2013; 14: 459–65

Background To improve compliance with insulin therapy and to accelerate insulin absorption, we tested if ID insulin delivery using a hollow microneedle causes less pain and leads to faster onset and offset of insulin pharmacokinetics in children and adolescents with type 1 diabetes compared with an SC insulin infusion. Methods

Results The primary end point, postprandial time in range (70– 180 mg/dL), showed no route-based differences with a high level of overall BG control for both SC and ID delivery. PK end points showed more rapid ID availability versus SC across doses and meals (DTmax - 16 min, DT50rising - 7 min, DT50falling - 30 min, all p < 0.05). Both intra- and intersubject variability for ID Tmax were lower. ID showed modest, secondary PD differences across doses and meals, generally within 90–120 min postprandially (D12 mg/dL BG at 90 min, D7 mg/ dL BGmax, D7 mg/dL mean BG 0–2 hours, all p < 0.05).

Sixteen patients received insulin lispro by microneedle and SC administration on separate days. Subjects rated the pain of insertion and infusion using a visual analog scale. Results Microneedle insertion pain was lower compared with insertion of the SC infusion catheter ( p = 0.005). Insulin onset time was 22 min faster ( p = 0.0004) and offset time was 34 min faster ( p = 0.017) after hollow microneedle delivery compared with SC infusion. Conclusions

Conclusions This study suggests that ID insulin delivery is superior to SC delivery in speed of systemic availability and PK consistency and helps to reduce postprandial glycemic excursions.

ID insulin delivery using a single hollow microneedle device resulted in less insertion pain and faster insulin onset and offset in patients with type 1 diabetes. A reduction in pain might improve compliance with insulin delivery.

Comment

Comment

This article demonstrated what can be achieved by using a different compartment when it comes to improving insulin absorption, thereby reducing postprandial glycemic excursions. ID insulin administration has been

As indicated above, ID insulin administration is a hot topic as it allows improving insulin pharmacokinetic and pharmacodynamic properties by simply administrating it via a different compartment.

S-48 Transdermal insulin application system with dissolving microneedles Ito Y, Nakahigashi T, Yoshimoto N, Ueda Y, Hamasaki N, Takada K Department of Pharmacokinetics, Kyoto Pharmaceutical University, Yamashina-ku, Kyoto, Japan Diabetes Technol Ther 2012; 14: 891–99

Background The aim of this study was to test a dissolving microneedle (DM) application system, where 225–300 insulin-loaded DMs were formed on a chip. After the heat-sealed sheet is removed, the system covered with the press-through package layer is put on the skin. By pressing with the hand, insulin DMs were inserted into the skin.

HEINEMANN The development and characteristics of novel microneedle arrays fabricated from hyaluronic acid, and their application in the transdermal delivery of insulin Liu S1, Jin MN 1, Quan YS 1,2, Kamiyama F2, Katsumi H 1, Sakane T 1, Yamamoto A1 1

Department of Biopharmaceutics, Kyoto Pharmaceutical University, Kyoto, Japan; and 2cOSmed Pharmaceutical Co. Ltd, Kyoto, Japan J Control Release 2012; 10: 933–41

Background The aim of this study was to test novel insulin-loaded microneedle arrays (MNs) fabricated from hyaluronic acid (HA) and characterize their applicability in the transdermal delivery of insulin.

Methods

Methods

Factors affecting the penetration depth of DM were studied using applicator in vitro and in vivo experiments. The penetration depth was determined for rat and human skin. Twolayered DM array chips were prepared to obtain complete absorption of insulin and administered to the rat abdominal skin. Plasma glucose levels were measured for 6 hours. By comparing the hypoglycemic effect with that obtained after SC injection, relative pharmacological availability (RPA) was determined.

The shape of MNs was observed via scanning electron microscopy. The characteristics of these MNs, including hygroscopy, stability, drug release profiles, and dissolution properties, were evaluated from a clinical application point of view. Transepidermal water loss was measured to investigate the piercing properties of MNs and the recovery of the skin barrier after the removal of MNs to confirm their safety. The transdermal absorption of insulin from MNs was examined via an in vivo absorption study in diabetic rats.

Results The penetration depth increased from 21 – 3 to 63 – 2 lm in proportion to application speed to isolated rat skin, at 0.8–2.2 m/s. Human skin showed similar results in the penetration depth. The in vivo penetration depth was dependent on the force (0.5–2.5 N) and duration (1–10 min), as the secondary application force. The penetration depth was 211 – 3 lm with a duration of 3 min in the in vivo rat experiment. DM array chips having an insulin-loaded space of 181 – 4 and 209 – 4 lm were evaluated in the rat. RPA values of insulin from DMs were 98.1 – 0.8% and 98.1 – 3.1%, respectively. Conclusions These results suggest the usefulness of DM for transdermal delivery of insulin.

Results The length of MNs was 800 lm with a base diameter of 160 lm and a tip diameter of 40 lm. MNs were found to maintain their skin piercing abilities for at least 1 hour, even at a relative humidity of 75%. After storing MNs for a month at - 40C, 4C, 20C, and 40C, more than 90% of insulin remained in MNs at all temperatures. It was also found that insulin is rapidly released from MNs via an in vitro release study. These findings were consistent with the complete dissolution of MNs within 1 hour of application to rat skin in vivo. MNs possess self-dissolving properties after their dermal application, and insulin appears to be rapidly released from these MNs. A continuous hypoglycemic effect was observed after 0.25 IU of insulin was administered to skin via MNs. Lower peak plasma insulin levels, but higher plasma insulin concentrations after 2 hour, were achieved compared with the SC administration of insulin of the same dose.

Comment This article does not present clinical data; however, it did a nice job of showing which interesting options are being developed to further reduce the burden of insulin therapy (i.e., reducing pain) while using a compartment that enables more rapid insulin absorption. To become a consumer product, such interesting developments have to fulfill a long list of requirements, not the least of which is production cost. So, we will have to wait and see which of the attempts (see following publications) to develop true microneedles will be successful (the ones used in the study presented before are large compared with the idea of ‘‘true’’ microneedles).

Conclusions These findings indicate that the MNs fabricated from HA are a useful alternative method of delivering insulin via the skin into the systemic circulation without inducing serious skin damage. It remains to be studied if HA MNs are an effective and safe method of transdermal insulin delivery in the clinic. Comment This interesting approach is only in the beginning stages and, although no human data were reported in this

INSULIN PENS AND NEW WAYS OF INSULIN DELIVERY

S-49

publication, the introduction of this article provided a useful review of the pros and cons of such MNs. The data presented are also impressive, that is, the decreases in blood glucose levels of rats when these MNs were administered with different insulin loads.

allow for direct clinical experience with the devices. On a positive note, it is good to see how pens improve from one generation to the next. This constant evolution of pens over the last decades has surely helped pave the way for their usage by most patients in most countries.

EVALUATION OF PENS

Study on the dosing accuracy of commonly used disposable insulin pens

A randomized, cross-over comparison of preference between two reusable insulin pen devices in pen-naı¨ve adults with diabetes

Krzywon M 1, van der Burg T 2, Fuhr U 3, Schubert-Zsilavecz M 1,4, Abdel-Tawab M 1

Wong M 1, Abdulnabib R2, Careyc MA 3, Fu H 1

2

1

Eli Lilly and Company, Indianapolis, IN; 2Pharmanet i3, Ann Arbor, MI; and 3Pharmanet i3, Blue Bell, PA Curr Med Res Opin 2013; 29: 465–73

Background The ease-of-use attributes of two reusable pen injectors, HPS (HumaPen Savvio) and HPL (HumaPen Luxura), and the final preference were evaluated in patients with type 1 or type 2 diabetes. Methods This was a 1-day, randomized, two-period crossover, openlabel, simulated-injection study in 203 pen-naı¨ve patients (mean age 58 years). For evaluation, a 16-item survey (7-point scale) was used: higher scores reflect greater preference and equal scores reflect no preference. The primary objective was final pen preference, with statistical gate-keeping to the ease of detecting an insufficient remaining dose (IRD) of insulin upon dose selection. Results HPS was chosen by 150 of 203 subjects (74%, 95% confidence interval [CI] = 67%–80%) for final overall pen preference. For the IRD item, HPS was preferred by 94 of 107 subjects with a preference (88%, 80%–93%). In 14 of the remaining 15 survey items, 64% to 88% of subjects with a preference preferred HPS over HPL. To confirm the results, subjects with no preference for either pen, which ranged between 95 and 148, were included in a Bayesian analysis. Conclusions The majority of pen-naı¨ve subjects preferred HPS over HPL. Some attributes of both pens were equally acceptable, as many subjects had no preference. Comment This is another one of those publications that give you a hint as to why such ‘‘studies’’ are not often performed by an independent institution in a head-to-head manner against other comparable pens. The authors themselves also acknowledged some other limitations: this was an injection simulation, an invalidated survey was used, and an office setting was used for the evaluation, which did not

1

Central Laboratory of German Pharmacists, Eschborn, Germany; Medical Devices/Quality, Regulatory & Risk Management, Sanofi-Aventis Deutschland GmbH, Frankfurt-am-Main, Germany; 3 Clinical Pharmacology Unit, Department of Pharmacology, University of Cologne, Cologne, Germany; and 4Institute of the Pharmaceutical Chemistry, J.W. Goethe-University, ZAFES, Frankfurt am Main, Germany Diabetes Technol Ther 2012; 14: 804–9

Background The advantages of insulin pens have led to the widespread use of insulin pens, particularly in Europe. In most of the former studies on the dose accuracy of insulin pens, only a small number of doses and pens were included. The present study was directed at the dose accuracy of one specific dose dispensed repeatedly with the same pen. This is the first study providing detailed comparative data on the accuracy of repeated dose delivery with prefilled disposable insulin pens at low, middle, and high doses dispensed over the entire pen volume. Methods Fifteen previously unused insulin pens from two lots of each pen type (SoloSTAR, FlexPen, Next Generation FlexPen, and KwikPen) were used to deliver 5 IU (low), 30 IU (middle), and 60 IU (high) doses, respectively, dispensed four times from each pen. Actual doses were determined gravimetrically taking the density of the respective insulin into account and were evaluated according to the ISO guidelines. Results All tested pens met the requirements for accuracy with none of the single values being outside the range of the ISO recommendations (1 – 1 IU, 30 – 2 IU, and 60 – 3 IU, respectively). Conclusion These data demonstrated a consistent and accurate dose delivery at all dosage levels for all tested pens, with no clinically relevant differences among the products. Comment It is good to see that this study (which was sponsored by one of the large manufacturers) showed no relevant differences in dose delivery between widely used disposable insulin pens in a head-to-head study over a wide range of insulin doses. In a sense, this publication ends a

S-50

dispute between manufacturers presenting studies that their own pens are more accurate than those of their competitors.

Perceptions of usability and design for prefilled insulin delivery devices for patients with type 2 diabetes Heron L 1, Reaney M 2, Hermanns N 3, Abetz L1, Gregg L 4 1

Adelphi Values, Bollington, Cheshire, United Kingdom; 2Eli Lilly, Windlesham, Surrey, United Kingdom; 3Diabetes Research Institute, Mergentheim, Germany; and 4Adelphi International Research, Bollington, Cheshire, United Kingdom

HEINEMANN

simulation study, three of the most widely used pens were studied. A relatively small number of insulin-naı¨ve patients and more experienced users were asked to answer a number of questions. An impressive number of tables and analyses showed that no pronounced differences exist between the pens studied.

Usability and preference assessment of a new prefilled insulin pen versus vial and syringe in people with diabetes, physicians and nurses Campos C, Lajara R, Deluzio T

Diabetes Spectrum 2013; 26: 16–28

The Institute for Public Health and Education Research, New Braunfels, TX

Background

Expert Opin Pharmacother 2012; 13: 1837–46

Insulin initiation in patients with type 2 diabetes is often delayed because of factors such as patients’ resistance to insulin therapy and worries about injections. Such delays affect glycemic control, have a direct effect on patient encounters, and may affect medication adherence. Usage of insulin pens addresses some of these concerns.

Background In this multicenter, crossover study, the preference and usability of a prefilled insulin pen (FlexTouch, FT) versus vial and syringe (V&S) was assessed. Methods

Methods Initially, semi-structured qualitative interviews were conducted to identify the most important features of insulin delivery devices for prandial use from the perspective of patients (n = 8) and healthcare professionals (HCPs; n = 10). In the second step, a 26-item questionnaire was developed. Patients (n = 33 insulin naive, n = 78 pen users) and HCPs (n = 151) were asked to indicate the most important features to them in insulin pens. Patients then simulated injection using three different pens (SoloSTAR, KwikPen, and FlexPen) and ranked them based on the same features. Results The most important features included knowing that the entire dose has been injected, ease of reading the dose correctly, and ease of correcting if the dose is over-dialed. KwikPen and SoloSTAR scored higher (paired t test, p < 0.05) in the simulation study than FlexPen on ‘‘knowing if you have injected the entire dose’’ (mean score out of 10: KwikPen, 8.9; SoloSTAR, 8.6; and FlexPen, 8.4). No other significant differences among the pens were noted in usability or design, and the mean ranking (from 1 to 3) of the pens was similar (KwikPen, 2.0; FlexPen, 2.1; and SoloSTAR, 1.9).

Patients with type 1 or type 2 diabetes (n = 60), and physicians (n = 30) and nurses (n = 30) with experience of diabetes management, performed test injections with FT and V&S. They answered written questions on ease of use and preference. The primary end point was preference for FT versus V&S. Results More respondents preferred using FT (88%) to V&S (5%; p < 0.001; the remainder chose ‘‘no preference’’), found FT (91%) easier to use than V&S (6%; p < 0.001), and would recommend FT (91%) over V&S (3%; p < 0.001). FT received better ratings than V&S for ease of use, holding the device stable when injecting, depressing the push-button/plunger, and reading the dose scale (all p < 0.001). Ratings for confidence in correct insulin delivery were also better with FT (both p < 0.001). Conclusions FT was preferred to V&S for insulin delivery in this comparative analysis. Usage of this pen may improve the experience of insulin injection compared with V&S for a wide range of patients. Comment

Conclusions HCPs can choose the most appropriate pen for patients by knowing which pens offer features that might help with earlier insulin initiation, greater patient adherence, and better clinical outcomes. Comment This publication—in an unusual journal for this topic— focused on relevant features of prefilled disposable pens from the perspective of patients and physicians. In this

This study was sponsored by the manufacturer of the FlexTouch. It is important to keep in mind that the pen uses a 32G extra-thin-wall needle with a length of 5 mm, whereas standard 30G needles were used with the syringes. It is interesting to note that many patients regard it as an advantage that using a syringe makes it easier to see that the full insulin dose is truly administered. One wonders how the outcome of this evaluation might have changed if another modern pen would have been used as comparator or as a third study arm. Without it, it appears as if the comparison was somewhat unfair.

INSULIN PENS AND NEW WAYS OF INSULIN DELIVERY

S-51

Accuracy and preference assessment of prefilled insulin pen versus vial and syringe with diabetes patients, caregivers, and healthcare professionals

Comment Just as with the previous article, the results of this study are not surprising. Thus, in view of the lack of a good comparator, the last sentence of the abstract can be interpreted as marketing.

Pfu¨tzner A1,2, Bailey T 3, Campos C 4, Kahn D 1, Ambers E1, Niemeyer M 5, Guerrero G6, Klonoff D 7, Nayberg I 7 1

ICRD Institute, San Jose, CA; 2IKFE, Mainz, Germany; 3AMCR Institute Inc., Escondido, CA; 4The Institute for Public Health and Education Research, New Braunfels, TX; 5Novo Nordisk NS, Søborg, Denmark; 6Novo Nordisk Inc., Princeton, NJ; and 7Mills Peninsula Health Services, San Mateo, CA Curr Med Res Opin 2013; 29: 475–81

Comparison of patient preference for two insulin injection pen devices in relation to patient dexterity skills Pfu¨tzner A1, Schipper C1, Niemeyer M 2, Qvist M 2, Lo¨ffler A1, Forst T 1, Musholt PB 1 1

Background The aim of this study was to investigate the dosing accuracy of the prefilled FlexTouch insulin pen (FT) in comparison to conventional vial and syringe (V&S) when used by patients (Pts), caregivers (CG), and healthcare professionals (HCPs). Methods One hundred twenty subjects participated in this study (40 patients with diabetes, age 61 – 11 years, 20 caregivers [parents and other relatives], 20 physicians, and 40 nurses/ Certified Diabetes Educators [CDE]). The participants were introduced to the devices in randomized order and were asked to perform injections of 5, 25, 43, and 79 IU doses into laboratory tubes. Dosing accuracy was analyzed by weighing the tubes on a pharmaceutical balance and calculating the mean absolute deviation ( MAD) from the intended doses. After completing a device assessment questionnaire, patient perception questionnaire (PPQ), with questions regarding device design and performance, the procedure was repeated with the other pen, and the patients were finally asked to complete a device preference questionnaire (DPQ).

IKFE–Institute for Clinical Research and Development, Mainz, Germany; and 2Novo Nordisk AB, Søborg, Denmark J Diabetes Sci Technol 2012; 6: 910–16

Background A number of patients with diabetes have impaired dexterity independent from the existence of diabetic neuropathy. In this study the impact of dexterity impairment on patient preference for two insulin pens (InnoLet and FlexTouch) was evaluated. Methods Ninety patients [54 male/36 female; age 62 – 8 years; disease duration 18 – 11 years; HbA1c 7.2 – 1.0%] were included in this study. They were stratified into four different groups based on the results of a dexterity test ( Jebsen–Taylor Hand Function Test) and assessment of visual impairment: 15 patients with type 1 (group A) and 30 with type 2 (group B) patients with impaired dexterity, 30 type 1/type 2 patients with visual impairment (group C), and 15 type 1/type 2 patients without any impairment (group D). The patients performed a cognitive function test (number connection test), were introduced to the devices in random order, and were asked to perform some mock injections before completing a six-item standardized preference questionnaire.

Results Dosing accuracy was better for FT when used by any of the cohorts at all doses; however, dosing accuracy with FT for all three subgroups was comparable (patients: 0.35–0.59 IU; HCP & CG: 0.29–0.54 IU; n.s.). Dosing accuracy with V&S for the three subgroups was not comparable: HCP and CG performed better with V&S than patients and delivered the doses with higher accuracy (range of mean MAD; patients: 0.81– 2.54 IU; HCP & CG: 0.51–1.30 IU, p < 0.005 at all doses). FT was ranked superior to V&S for all aspects of the PPQ. In the DPQ, 93% of the patients voted for FT.

Results

Conclusion

Patient dexterity skills may have an influence on device preference, especially if the impairment is more pronounced.

FT was more accurate at all tested doses and was used with similar accuracy by patients, HCPs, and CGs compared to V&S. Using questionnaires only, and without dexterity assessment, study participants rated FT higher than V&S in every component of the PPQ, and the vast majority of them preferred FT. These findings suggest a better alternative for dosing accuracy and improved adherence when using the prefilled insulin pen compared to V&S for insulin delivery in patients with diabetes.

Patients in all groups showed a strong preference for FlexTouch. All unimpaired patients (100%, group D) preferred FlexTouch, as did the vast majority in all other groups. Only 11% of the patients with impaired cognitive function preferred InnoLet, as did a few patients with more severely impaired dexterity or with visual impairment (group A 13%; group B 3%; group C 14%). Conclusions

Comment A number of factors determine why we prefer a given product (like a car), and the same holds true when it comes to insulin pens. Therefore, preference studies are a bit tricky to perform, and the reported outcome has to be taken with a grain of salt. In view of a large group of

S-52

HEINEMANN Comment

patients who use insulin pens and who have limited dexterity and/or who are visually impaired, it is important to be careful with which group of patients such studies are performed. This study focuses on a considerable number of patients with type 1 or type 2 and stratifies them according to the factors mentioned. The strong patient preference toward the newer insulin pen is impressive and somewhat counterintuitive as the InnoLet was developed with this specific patient group in mind. However, this device is not a pen by design. It would have been better to use another modern insulin pen as comparator.

This study confirmed that adding features to insulin pens does not necessarily result in a measurable benefit such as glycemic control. One would assume that adding a reminder or ‘‘memory’’ function to an insulin pen is helpful in detecting missed insulin injections, etc., by registering a number of parameters each time insulin is administered. However, this study with a considerable number of patients with suboptimal metabolic control performed across 32 sites in Germany showed no additional benefit of having such a memory function integrated into the pen; the observed improvement in HbA1c was comparable in both study groups. The authors discussed potential explanations for this negative study outcome: facilitation of corrective insulin injections after a meal when the preprandial insulin injection was forgotten, the number of corrective actions taken was too small, and also the handling of the basal insulin dose might have minimized any effects of changing prandial insulin delivery. Despite the negative outcome of this study, one can imagine many patients for whom monitoring insulin therapy details with such a pen might be advantageous.

No effect of insulin pen with memory function on glycemic control in a patient cohort with poorly controlled type 1 diabetes: a randomized open-label study Danne T1, Forst T 2, Deinhard J 3, Rose L 4, Moennig E 5, Haupt A 5 1

Bult Diabetes Centre for Children and Adolescents, Hannover, Germany; 2IKFE Institute for Clinical Research and Development, Mainz, Germany; 3Accovion GmbH, Eschborn, Germany; 4Diabetologische Schwerpunktpraxis, Mu¨nster, Germany; and 5Lilly Deutschland GmbH, Diabetes, Ban Homburg, Germany J Diabetes Sci Technol 2012; 6: 1392–97

Background Injection compliance is a major problem in patients with type 1 diabetes. Using an insulin pen with memory function might facilitate corrective dosing to reduce postprandial glycemic excursions and therefore might improve overall glycemic control.

Is the indicator magnifying window for insulin pens helpful for elderly diabetic patients? Lee JH 1, Hong ES 2, Ohn JH 2, Cho YM2 1

Department of Internal Medicine, Chungnam National University, Daejeon, Korea; and 2Department of Internal Medicine, Seoul National University, Seoul, Korea Diabetes Metab J 2013; 37: 149–51

Background Methods In this randomized, open-label, 24-week multicenter study, patients with inadequately controlled type 1 diabetes (HbA1c) ‡ 8% were randomized to use the HumaPen Memoir, an insulin pen device with memory function, for their mealtime insulin injections or the conventional device HumaPen Luxura. HbA1c, hypoglycemia, and pen acceptance were assessed at baseline and after 12 and 24 weeks. Results Of 263 patients randomized, 257 were eligible for analysis (baseline HbA1c 9.09 – 0.99%; age 40 – 17 years; diabetes duration 16 – 11 years): HumaPen Memoir 129, HumaPen Luxura 128. Changes of HbA1c up to week 24 were not different between the HumaPen Memoir [0.43% (- 0.59%, - 0.28%)] and the HumaPen Luxura group [0.48% (0.64%, 0.32%); p = 0.669]. Also, the overall incidence of hypoglycemic episodes did not differ between groups ( p = 0.982). Conclusions In this study, usage of a memory function pen was not associated with superior glycemic control, suggesting that adherence to mealtime injection schedules was not improved in a relevant manner. However, the memory function might be helpful for specific patient populations, for example, children or forgetful patients.

Patients with type 2 diabetes who require insulin therapy are commonly elderly and have poor visual acuity. In this study, we examined the clinical usefulness of the indicator magnifying window (IMW) for elderly patients. Methods We recruited 50 patients aged > 60 years who use insulin pens for insulin application. They were asked to set randomly selected doses at the insulin pen with or without an IMW. Dosing accuracy, convenience, self-confidence, need for eyeglasses, preference, and willingness to recommend the IMW to other patients was assessed. Results Although the IMW did not improve the dosing accuracy or convenience, it decreased the need for eyeglasses. Conclusions Overall, the clinical usefulness of the IMW is quite limited in elderly patients with type 2 diabetes. Comment One might be tempted to smile at such a study; however, for elderly patients/patients with impaired visual

INSULIN PENS AND NEW WAYS OF INSULIN DELIVERY

S-53 Methods

acuity—and this is a good portion of the patients treated in daily practice—such aspects are of high relevance. One might ask, why do IMWs exist if they apparently are of no great help for the patients, and how can we do better? Electronic gadgets exist or are in development that read the dose selected on the pen (see previous article). Not only do these gadgets allow the selected dose and the time of administration to be stored, but also they could be of help for patients in seeing the actual selected dose if this information were displayed directly on a TV screen or on their smart phone by means of a special app.

Inhaled Insulin When it comes to inhaled insulin, we are playing a sort of waiting game. The phase 3 studies with MannKind’s ultra-rapid inhalable prandial insulin (Afrezza) have been performed, and a new application for an Investigational New Drug was submitted to the U.S. Food and Drug Administration (FDA). A good review of this pulmonary insulin was published in 2012 (1). The tone of the press releases by MannKind suggests that FDA approval is probable. The question remains whether the FDA will ask for regular lung function tests, etc. The answer to this question will have an impact on the patient’s interest in using pulmonary insulin. The next question is whether diabetologists would be willing to prescribe inhaled insulin. This will depend to a certain extent on how smart the marketing activities of MannKind will be. It will be necessary to handle the burden of the Exubera failure and some safety concerns adequately. Nevertheless, if this inhaled insulin becomes available on the market, this will be the first time in a number of years that an insulin product not administered subcutaneously but via an alternative route makes its way to the market. If this finds its way into the armamentarium of diabetes therapy, this will be encouraging for alternative routes of insulin delivery in general, as people will hopefully regain trust in this area of research and more venture capital will be invested again. It is worth mentioning that some research and development in the area of inhaled insulin is still going on (www.dancepharma.com/). In view of the involvement of one of the godfathers of pulmonary application, John Patton, this development and story are interesting and worth following closely. Technosphere insulin effectively controls postprandial glycemia in patients with type 2 diabetes mellitus Zisser H 1, Jovanovic L 1, Markova K 1, Petrucci R 2, Boss A 2, Richardson P 2, Mann A 2 1 2

Sansum Diabetes Research Institute, Santa Barbara, CA; and MannKind Corporation, Valencia, CA

Diabetes Technol Ther 2012; 14: 997–1001

Eight subjects (7 men, 1 woman) with type 2 diabetes underwent dose optimization meal challenge (MC) visits (100% CHO) and MCs with varied CHO meal contents (50%, 200%, and 0% calculated CHOs). The observed change in postprandial glucose (PPG) excursions was the primary end point. Results Maximum PPG excursions with the 100% CHO meals were - 13 – 15 mg/dL for breakfast (B) and - 14 – 15 mg/dL for lunch (L). These increases were similar to those observed after 50% CHO meals (B, - 17 – 16 mg/dL; L, + 14 – 10 mg/ dL). As anticipated the largest PPG excursions occurred with the 200% CHO meals; however, they remained below the targets of the American Diabetes Association (B, + 19 – 16 mg/dL; L, + 32 – 29 mg/dL). During 15 of the MCs, subjects took their usual dose of the inhaled insulin and then had no meal (0% CHO). In this case, the largest PPG excursions were - 33 – 9 mg/dL at 60 min (B) and - 31 – 10 mg/dL at 60 and 90 min (L). Glycemic control improved from an HbA1c of 7.82 – 1.04% at Week 1 to 6.18 – 0.46% ( p = 0.00091) at Week 19. Conclusions These results obtained with a small sample size suggest that once an optimal dose of TI is determined, patients with type 2 diabetes can ingest meals with considerable differences in their CHO content—or even skip meals—without having the risk of severe hypoglycemia. In this uncontrolled pilot study, glycemic control improved significantly by - 1.63%. Comment For many patients with type 2 diabetes, it would be sufficient to have a good dose of insulin administered with each meal, even if it were not the ideal dose, as some reduction in postprandial glycemic excursions would be better than no reduction at all. The risk for postprandial hypoglycemic events is quite low. To get such a treatment accepted, it has to be quite easy and convenient; it has to be an inhalation that can be done easily and within seconds. This meal study with a small sample size shows how such a treatment could look. In this study, patients came into the center for numerous visits to determine the optimal insulin dose while eating meals with a considerable range of CHO content. Preprandial glycemia in these patients appeared not to have been well controlled; unfortunately, only changes in postprandial glycemic excursions are shown, not absolute changes. The improvement in HbA1c in such an uncontrolled study has to be interpreted with care. Nevertheless, in line with the above-mentioned thoughts, this study discusses an interesting new way of treating many patients.

Background In this pilot trial, it was studied if an optimal dose of Technosphere insulin (TI) inhalation powder (MannKind Corp., Valencia, CA) could be used to control postprandial glycemic excursions despite considerable variation in meal carbohydrate (CHO) content.

Oral Insulin This year, the literature search for publications on oral insulin (OI) resulted in less hits than in the years before. Ignoring the ‘‘usual’’ publications about promising new novel

S-54 OI formulations in pharmaceutical journals, the total number of publications about clinical studies with human subjects amounted to one (see below)! Despite this lack of published data, a number of review articles about OI were published [e.g., Fonte et al. (2) and Patel (3)]. Not a single publication could be found about other routes of insulin administration such as transdermal or nasal insulin. According to a recent press release, Novo Nordisk is focusing its development work on a long-acting OI formulation using the so-called GIPET platform, which was developed by the Ireland-based partner company Merrion. All other companies are focusing on covering prandial insulin requirements by OI. It is not easy to venture a guess about the progress of these companies or how much further the respective developments are: Oramed has continued with its insulin development (ORMD 0801) from phase 2 to phase 3; Biocon was able to find a partner (Bristol-Myers Squibb) for its OI IN-105. However, it is not clear whether other companies such as Tamarisk Technologies and Aphios have made it from the preclinical to the clinical phase or not; at least, no clinical data were published over the last year. It appears that buccal insulin formulation by Generex is ongoing (http://investor.generex.com/releases.cfm); positive data from a phase 3 trial in India were reported recently. However, one would be extremely interested in seeing convincing data from recent clinical–experimental studies and double-blind clinical studies showing a good metabolic effect with reasonable insulin doses. Glucose-reducing effect of the ORMD-0801 oral insulin preparation in patients with uncontrolled type 1 diabetes: a pilot study Eldor R 1, Arbit E2, Corcos A 3, Kidron M 2 1

Diabetes Unit, Internal Medicine, Hadassah University Hospital, Jerusalem, Israel; 2Oramed Pharmaceuticals, Jerusalem, Israel; and 3 Diabetes and Endocrine Clinic, General Health Services Affiliated to Hadassah Medical School, Jerusalem, Israel PLoS One 2013; 8: e59524

Background In efforts to provide patients with a more compliable treatment method, Oramed Pharmaceuticals tested the capacity of its OI capsule (ORMD-0801, 8 mg insulin). Methods Eight patients with type 1 diabetes in bad glycemic control (HbA1c 7.5–10%) were monitored throughout the 15-day study period by means of a blind continuous glucose monitoring (CGM) device. Baseline patient glucose profiles were monitored over a 5-day pretreatment screening period. During the ensuing 10-day treatment phase, patients were asked to treat themselves as usual and to self-administer an OI capsule three times daily, just prior to meal intake. CGM data sufficient for glucose analyses were obtained from 6 of the 8 subjects.

HEINEMANN crease in glucose area under the curve (66055 – 5547 vs. 55060 – 3068 mg/dL/24 hours, p = 0.023), with a greater decrease during the early evening hours. Conclusions The OI capsules in conjunction with SC insulin injections were well tolerated and effectively reduced glycemia throughout the day. Comment The results of this pilot study with a small sample size and a nonblinded study design (uncontrolled!) have to be interpreted with care, or as the authors stated it: ‘‘can only be addressed as hypothesis forming for future larger and longer trials planned.’’ The list of planned studies mentioned subsequently in the publication is impressive.

Miscellaneous Each year some studies are published that do not fit into one of the above categories but that are clearly interesting for insulin treatment in general. No human data have been published thus far with a truly innovative approach in which the insulin is activated by light (4). Body mass index and the efficacy of needle-free jet injection for the administration of rapid-acting insulin analogs, a post hoc analysis de Galan BE, Engwerda EEC, Abbink EJ, Tack CJ Department of Internal Medicine, Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands Diabetes Obesity Metab 2013; 15: 84–86

Background In a recent euglycemic glucose clamp study, it was shown that using jet injectors rather than insulin pens improved the time-action profiles of rapid-acting insulin analogs in healthy volunteers. Methods In this subsequent analysis, the authors analyzed whether the time-action profiles were modified by BMI and related weight parameters. They defined subgroups by BMI, waistto-hip ratio, waist circumference, and insulin dose. Results After insulin pen administration, times to peak insulin levels (T-INSmax) occurred 31.1 [95% confidence interval (CI) 13.7– 48.5] min later and time to maximum glucose requirement (TGIRmax) 56.9 (26.6–87.3) min later in more obese (BMI > 23.6 kg/ m2) than in lean subjects (BMI < 23.6 kg/m2). In contrast, T-INSmax and T-GIRmax were similar in subjects with high and low BMI, when insulin was administered by jet injection.

Results Treatment with OI was associated with a 24% reduction in the frequencies of glucose readings > 200 mg/dL (60 – 8% pretreatment vs. 45 – 5% with OI; p = 0.023) and a 17% de-

Conclusions These data suggest that insulin administration by means of a jet injector may be especially beneficial for obese subjects.

INSULIN PENS AND NEW WAYS OF INSULIN DELIVERY Comment This quite active group from the Netherlands drives this interesting development in a straightforward manner; that is, they have published a number of articles about administering insulin via an injector over the last years. However, in view of the numerous attempts presented previously (some injectors have been on the market for a while), one cannot help but ask oneself if this will finally become a success story this time. The number of studies in which the impact of body weight on insulin absorption has been studied is relatively small. This is somewhat surprising as in these studies the BMI/body weight showed the considerable impact of this parameter on insulin absorption/insulin action. The data presented here confirm these data when it comes to administering insulin by pen. It is clear that the BMI has no impact when the insulin is administered via the pulmonary route, as the absorptive surface in the lung can be assumed to be similar in slim and overweight subjects, but that the BMI has no impact on the absorption properties from the SC insulin depot when the insulin is administered with an injector is of notable interest. This most probably has to do with the ‘‘form’’ of the SC insulin depot: with a needle, the administered liquid sits in a circumscribed depot as compared with when the insulin is administered in a distinct spraylike dispersion pattern, ensuring a larger absorptive area.

S-55

These data suggest a benefit to administering prandial insulin in obese subjects using such a device. It would be of high practical relevance to repeat such studies with a larger sample size and larger range of BMI; optimizing insulin therapy in massively obese patients is difficult with the devices currently available to administer insulin.

Author Disclosure Statement L.H. is a consultant for a number of companies developing innovative diagnostic and therapeutic options for diabetes therapy. He is active, for example, with Sanofi and Roche Diagnostics. He is a partner and consultant for Profil Institut fu¨r Stoffwechselforschung, Neuss, Germany and Profil Institut for Clinical Research, San Diego, CA. References 1. Boss AH, Petrucci R, Lorber D. Coverage of prandial insulin requirements by means of an ultra-rapid-acting inhaled insulin. J Diabetes Sci Technol 2012; 6: 773–79. 2. Fonte P, Arau´jo F, Reis S, Sarmento B. Oral insulin delivery: How far are we? J Diabetes Sci Technol 2013; 7: 520–31. 3. Patel MM. Colon targeting: an emerging frontier for oral insulin delivery. Expert Opin Drug Deliv 2013; 10: 731–39. 4. Jain PK, Karunakaran D, Friedman SH. Construction of a photoactivated insulin depot. Angewandte Chemie 2013; 52: 1404–9.