British Journal of Radiology (2004) 77, 236-242
© 2004 British Institute of Radiology
doi: 10.1259/bjr/13096559
Two- and three-dimensional ultrasound in the development of a needle-free injection system
C J C Cash, MCRP, FRCR
1
L H Berman, MRCP, FRCR
1
G M Treece, BA, PhD
2
A H Gee, BA, PhD
2 and
R W Prager, BA, PhD
2
1 University Department of Radiology, Addenbrooke's Hospital, Hills Road, Cambridge CB2 2QQ and the 2 Cambridge University Department of Engineering, Trumpington Street, Cambridge, UK
 |
Abstract
|
|---|
Ultrasound was used to assess a needle-free injection device for both intradermal and subcutaneous injections. The aim of this study was, first, to differentiate intradermal from subcutaneous injections, both in vivo and in vitro using 2D ultrasound, and second, to quantify the amount of injectate that actually arrives within the dermis or subcutaneous tissues using volume measurements derived from high-resolution 3D ultrasound data sets, using a freehand system (Stradx), developed by the Cambridge University Departments of Engineering and Radiology. For the in vitro study the devices were filled with dye and injected into a pig preparation. The injection site was examined with high-resolution ultrasound and subsequently dissected to locate the injected dye with respect to the dermis. For the in vivo study, 8 volunteers received needle-free injections of normal saline. High-resolution 2D images and 3D data sets were obtained of the injected sites. Proprioceptive information for the 3D data sets was produced using an optically tracked freehand system. Segmentation of the 3D data sets gave an estimation of the volume of injected material (injectate) within the dermis. The results demonstrated that 2D ultrasound could identify the location of the injectate in the in vitro experiments and successfully distinguished an intradermal from a subcutaneous injection. In the in vivo study, 2D ultrasound clearly demonstrated the injectate location within the volunteers' dermis but was less able to demonstrate the dispersion of injectate within the subcutaneous tissues.
|
Introduction
|
|---|
Needle-free injectors operate by producing a high-pressure jet of injectate applied directly to the skin. The pressure is sufficient to cause a small puncture in the skin through which the injectate is delivered. The eventual location of the injectate, whether intradermal or deeper, is determined by several factors including those relating to the configuration of the device itself, such as injectate pressure and those relating to the physical properties of the skin such as dermal thickness. The advantages of needle-free systems over a needle and syringe are predominantly patient acceptance and secondly, a significant reduction in the biohazard risk to patients and health professionals posed by needlestick injuries. Standardized delivery to a specific location, e.g. the dermis or subcutaneous tissues, has been claimed.
The skin is immunologically active containing antigen-presenting Langerhans cells [1] and therefore the dermis and epidermis are potential sites for the delivery of vaccines [2]. Studies have shown that if delivery of the vaccine can be targeted to the immunologically competent epidermal/dermal layer, then only a quarter of the injection dose is required compared with a standard intramuscular injection [3]. Dermal thickness varies from about 1.5 mm to 4 mm between individuals and within an individual [4]. Where the dermis is thin, it may be difficult to administer a dermal injection accurately using a conventional needle without a substantial portion of the injectate entering the subdermis. The development of a system that repeatedly delivered the majority of the injectate to the dermis or the subcutaneous tissues, depending on the design of the device, would have important implications for drug delivery.
The bioavailability and pharmacokinetics of insulin injected into the subcutaneous tissues of diabetic patients have been studied. The variation in amount of subcutaneous tissue within a body region [5] may result in accidental intramuscular injection [6], which predisposes to faster absorption than subcutaneous injection [7].
The needle-free injection device assessed in this study delivers the injectate using a high-pressure jet. As there is no needle tip to locate prior to the injection, it is impossible to identify accurately the eventual location of the injectate. Deliberate variations in the pressure, orifice size and volume of injectate are among the variables that configure the device into either an intradermal or subcutaneous injection. This study was performed in collaboration with Weston Medical Limited (Cambridge, UK, http://www.weston-medical.com), the designer of the needle-free injection devices.
|
Method
|
|---|
In vitro study
The model consisted of fresh pig loin, with retained skin, obtained from a local abattoir. This was injected with coloured dye using intradermal and subcutaneous configurations of the needle-free devices. The area injected was scanned with 16 MHz and 22 MHz transducers and a Diasus ultrasound machine (Dynamic Imaging, Livingston, Scotland, UK, http://www.dynamicimaging.co.uk/). The thickness of the dermal/epidermal layer was measured. The injectate was demonstrated within the target site and its diameter measured with electronic calipers. The specimen was dissected along the plane of the injection to confirm the actual location of the injectate, identified as dye staining of the tissue. Photographs of the dissected specimen were taken alongside a ruler for later analysis.
In vivo study
The device development team at Weston Medical Limited (Cambridge, UK) regularly assesses the in vivo performance of the needle-free injector on its own volunteers using normal saline. The data required for this study were obtained during one of these assessments. Injections were administered into the skin of the abdominal wall by a qualified technician from Weston Medical using a sterile technique. The technician graded each injection according to the amount of saline that is expelled by the device onto the skin surface, i.e. the amount of injectate that fails to penetrate the skin. This is a subjective assessment of the success of the injection in delivering the "drug" into the cutaneous tissues. The accuracy of the device in delivering intradermal and subcutaneous injections was assessed using ultrasound images.
The 3D data sets were obtained from freehand 2D grey scale images from the high resolution Diasus ultrasound machine (Dynamic Imaging). The transducer position was continually tracked using optical sensors (Polaris, Northern Digital Inc., Waterloo, ON, Canada, http://www.ndigital.com/). The software used to process the data sets was written by Cambridge University Department of Engineering (Stradx, http://svr-www.eng.cam.ac.uk/
rwp/stradx) [8, 9].
The injection device has been developed so that a known volume is expelled on triggering. For the purposes of this study, the devices delivered 0.15 ml and 0.5 ml for an intradermal and subcutaneous injection, respectively. Some devices employed a mechanical spacer, i.e. the device had been designed with a gap between the injecting orifice and the surface of the skin; in the remaining devices the injecting orifice was applied directly onto the skin surface. In order to assess how much of the injectate failed to penetrate the skin of the volunteers following each injection, carefully pre-weighed dry filter paper was applied to the skin surface of the injection site to absorb any free injectate. The "wet" filter paper was then re-weighed in order to establish the volume of injectate left on the skin surface following an injection. 3D data sets of the injected site were then obtained and recorded to hard disc for later evaluation. The position of the injectate in relation to the dermal/epidermal layer was identified. If the injectate could be identified as a discrete area on the ultrasound image, manual segmentation of this area was performed and a volume of the injected material estimated.
|
Results
|
|---|
In vitro study
The dermis of the pig preparation was demonstrated as a well-defined, 1 mm layer of low reflectivity with respect to the more reflective subcutaneous layer. Typically, the injectate from an intradermal injection could be identified as a slightly reflective swelling within the dermis, but within this swelling there was a highly reflective arc casting a strong acoustic shadow (Figure 1
). Dissection of this specimen showed the dye to be mainly within the dermis. The dermal thickness was found to be 1 mm in depth and the diameter of the injectate was 9 mm. Some discoloration of the subcutaneous tissues was also present suggesting that some of the intradermal injection also entered this layer (Figure 2). The position of the echogenic arc on ultrasound appeared to correspond closely to the actual position of the injectate on dissection and this was used as the ultrasound marker of the location of the injectate. In the example shown, the diameter of the arc (injectate) was 8 mm.
View larger version (112K):
[in this window]
[in a new window]
|
Figure 1. In vitro specimen-intradermal injection: white arrowheads outline the acoustic arc; black arrow points to the dermal–subdermal junction.
|
|
Typically, following a subcutaneous injection, the reflective arc could be identified just beneath the level of the dermis at the dermal/subdermal interface (Figure 3). Again the dermal thickness measured approximately 1 mm in depth. The diameter of the arc (injectate) was 6 mm. Dissection of this specimen showed dye to be mainly just deep to the dermis. The diameter of the injectate measured 7 mm (Figure 4). The dispersion of dye within the subcutaneous fatty layer was more widespread and less clearly defined than the rather discrete swelling of the dermis produced by the intradermal device.
View larger version (121K):
[in this window]
[in a new window]
|
Figure 3. In vitro specimen-subcutaneous injection. White arrowheads outline the acoustic arc; black arrow points to the dermal–subdermal junction.
|
|
In vivo study
In the initial in vivo assessment of the needle-free device, ultrasound demonstrated strong acoustic shadows within the injected skin (Figure 5) in some but not all of the device configurations. The combination of reflective arc with acoustic shadow occurred more frequently in devices that employed the mechanical spacer, i.e. in devices designed with a gap between the injecting orifice and the skin surface. This acoustic shadow was usually not observed when the mechanical spacer was not used, i.e. when the injecting orifice was directly applied to the skin surface. The inference was that a small quantity of air was being injected into the skin when the spacer was employed. The spacer component of the device was subsequently found to be unnecessary and was abandoned in all later studies.
The high-resolution ultrasound images of living human skin showed the dermis as a clear relatively reflective layer measuring between 2 mm and 3 mm, varying between individuals. Intradermal injections produced a slightly reflective well-defined swelling within the dermis (Figure 6). In some cases there appeared to be small, poorly reflective, pockets seen just beneath the dermal swelling, suggesting that some of the injectate was reaching the subcutaneous structures (Figure 7).
View larger version (130K):
[in this window]
[in a new window]
|
Figure 6. Swelling of the dermis from an intradermal injection. Arrow points to dermal–subdermal junction.
|
|
View larger version (121K):
[in this window]
[in a new window]
|
Figure 7. Intradermal injection. Small poorly reflective pockets (arrows) suggest some injectate is reaching the subcutaneous tissues.
|
|
Subcutaneous injections were much harder to define with ultrasound. There was no obvious focal collection of injectate within the dermal or subcutaneous tissues. The injected material appeared to disperse diffusely into the subcutaneous adipose tissue (Figure 8), in a similar fashion to the in vitro studies. The absence of a focal collection prevented further volume analysis of these injections.
View larger version (126K):
[in this window]
[in a new window]
|
Figure 8. A subcutaneous injection illustrating some of the injectate in the dermis, seen as a moderately reflective swelling (black arrow), and some of the injectate in the subcutaneous tissues, seen as poorly reflective pockets (white arrowheads). White arrow points to the dermal–subdermal junction.
|
|
Using the 3D software, the swellings produced in the dermis by the intradermal injections could be analysed further. Manual segmentation around the dermal swelling, from a series of representative B-scans (Figure 9a,c) produced a volume measurement and a surface rendered outline of the dermal swelling itself (Figure 9e). From identical B-scans used to segment the bulging dermis, the volume of the same section of dermis was calculated, this time estimating the contour of unswollen dermis in order to correspond to the contour of the pre-injection dermis (Figure 9b,d,f). By subtracting this volume from the post-injection dermal contour a volume of the injectate within the skin could be estimated.
View larger version (87K):
[in this window]
[in a new window]
|
Figure 9. Manual segmentation of (a) the dermal swelling and (b) dermis alone from the same B-scan. A series of segmented B-scans is shown in 3D space (c) dermal swelling, (d) dermis alone. (e) Surface rendered dermal swelling and (f) surface rendered dermis alone are produced from the segmented B-scans, illustrating the change in dermal volume that occurs following an injection.
|
|
A total of 27 intradermal injections were delivered to the 8 volunteers using 5 differently configured experimental devices. Eight of these injections were graded as poor by the administering technician; a large amount of injectate dispersed over the volunteer's skin surface and little or no injectate could be identified within the cutaneous tissues on ultrasound. Further analysis of these technical failures was not performed. Of the remaining 19 intradermal injections, a clear dermal swelling could be identified allowing estimation of injectate volume deposited in the target. Table 1 outlines the results and illustrates the variation in performance of the devices between individuals.
View this table:
[in this window]
[in a new window]
|
Table 1. Illustrates a variation in performance of five differently configured dermal devices between 8 volunteers
|
|
|
Discussion
|
|---|
The needle-free injection is claimed to be less painful than standard needle injections [10] and is frequently reported by volunteers to be pain free. The reduction in biohazard is an obvious advantage to health professionals given that the reported incidence of needle-stick injuries is as high as 50 per 1000 medical staff per annum [11]. A further advantage would be the ability to target accurately the dermis or the subcutaneous tissues depending on the device configuration. In developing these devices many variables have been assessed including volume of injectate, gas pressure and size of delivery orifice. A subjective assessment is made by the administering technician following each injection, as previously described. Without imaging, the proportion of the injection penetrating the skin can be assessed but no assessment of location of the injectate can be made. Using a high-resolution transducer and software specifically designed to interpret and display high-resolution data sets, we were able to assess the performance of the needle-free injectors non-invasively. Whilst it is accepted that there will be errors in measuring the small volumes described, the 3D ultrasound system used has been thoroughly tested and claims a point location accuracy of 0.5 mm [8].
The in vitro studies demonstrated that high-resolution 2D images can differentiate between an intradermal and subcutaneous injection. The 3D technique employed in the in vivo study also enabled the volume of injectate reaching the target site to be assessed. As the data sets were acquired within seconds of the injection it was assumed that the detectable bleb was due to injectate deposition rather than a local tissue damage response.
The conspicuous reflective arc and acoustic shadow observed in the in vitro study and with injections given by some of the earlier devices may have been due to air entering the skin as a Venturi effect from the jet of injectate. Although the mechanical spacer was not used in the in vitro experiments the lax texture of the pig preparation may have resulted in poor contact between device and skin, allowing some air to enter with the jet of dye.
Using three-dimensional ultrasound data to estimate the percentage of the injected volume within the desired target, a range in performance was demonstrated for the intradermal devices. At best we were able to account for 95% of the injected volume and this was clearly demonstrated as a defined hyper-reflective swelling of the dermis (Figure 6
). At the other extreme we were only able to account for 39% of the injected volume. In this particular case 20% of the injected material was sprayed onto the skin surface. This still leaves 41% (0.06 ml) of injectate unaccounted for. In all the in vivo intradermal injections, even allowing for losses to the skin surface, a proportion of the injected volume could not be accounted for. The dermis is a much denser layer of tissue than the subdermis, which is composed mainly of fat. It is likely that if the injectate reaches the subdermis it will disperse within the fat more easily than it is able to within the tougher fibrous dermis. The in vitro study supports this suggestion, illustrating dispersal of the dye in and around the fatty lobules of the subcutaneous tissue (Figure 4). In some cases poorly reflective ill-defined pockets of an unquantifiable volume were identified within the subcutaneous tissues (Figure 7).
The performance of the differently configured devices varied between individuals. However, even the least accurate delivered 0.06 ml (39%) of the injected volume into the target site and this may well be sufficient for vaccination purposes. Given that this study was carried out on experimental devices that are yet to be optimized, performance of these devices is likely to improve. A potential criticism is that this study was carried out on a very small number of healthy young individuals. Once the optimal configuration of the intradermal injections has been discovered, 3D data sets could be used in a larger study to examine the range of performance on different individuals' skins. In such a study, volunteers of varying ages and ethnic groups ideally should be tested. Ideally, patients with medical conditions, which may alter the properties of the dermis, such as steroid therapy and connective tissue disorders, should also be included. Using segmentation of the 3D data to estimate the percentage volume of injected material located in the dermis, it would be possible to build a profile of how these devices perform in a larger sample. If the delivery site and volume are demonstrated to be reproducible, a smaller vaccine dose may be used to achieve a satisfactory immunological response.
|
Acknowledgments
|
|---|
We would like to thank Jonathan Wilkins and his team at Weston Medical for their help and co-operation with this study.
Received for publication January 2, 2003.
Revision received August 5, 2003.
Accepted for publication December 3, 2003.
|
References
|
|---|
- Bannister LH. Section 5: integumental system: skin and breasts. In: Williams PL, editor. Grays anatomy. London: Churchill Livingstone, 1995;376.
- Babiuk S, Baca-Estrada M, Babiuk LA, Ewen C, Foldvari M. Cutaneous vaccination: the skin as an immunologically active tissue and the challenge of antigen delivery. J Control Release 2000;66:199–214.[CrossRef][Medline]
- Zuberi BF, Rajput MR, Muzaffar L, Jatoi N, Shaikh WM. Efficacy of low dose intra-dermal hepatitis B vaccination schedule. J Pak Med Assoc 1998;48:377–8.[Medline]
- Bannister LH. Section 5: integumental system: skin and breasts. In: Williams PL, editor. Grays anatomy. London: Churchill Livingstone, 1995;392.
- Frid A, Linde B. Intraregional differences in the absorption of unmodified insulin from the abdominal wall. Diabet Med 1992;9:236–9.[Medline]
- Frid A, Linde B. Where do lean diabetics inject their insulin? A study using computed tomography. Br Med J 1986;292:1638.
- Frid A, Gunnarsson R, Guntner P, Linde B. Effects of accidental intramuscular injection on insulin absorption in IDDM. Diabetes Care 1988;11:41–5.[Abstract]
- Treece GM, Gee AH, Prager RW, Cash CJC, Berman L. High definition freehand 3D ultrasound. Ultrasound Med Biol 2003;29:529–46.[CrossRef][Medline]
- Treece GM, Gee AH, Prager RW, Berman L. Fast surface and volume estimation from non-parallel cross-sections, for freehand 3-D ultrasound. Med Image Anal 1999;3:141–73.[CrossRef][Medline]
- Hollingsworth SJ, Hoque K, Linnard D, Corry DG, Barker SG. Delivery of low molecular weight heparin for prophylaxis against deep vein thrombosis using a novel, needle-less injection device (J-Tip). Ann R Coll Surg Engl 2000;82:428–31.[Medline]
- Neuberger JS. Incidence of needlestick injuries in hospital personnel: implications for prevention. Am J Infect Control 1984;12:171–6.[CrossRef][Medline]
This article has been cited by other articles:
|
|
|
|
 
C J C Cash, A M Sardesai, L H Berman, M J Herrick, G M Treece, R W Prager, and A H Gee
Spatial mapping of the brachial plexus using three-dimensional ultrasound
Br. J. Radiol.,
December 1, 2005;
78(936):
1086 - 1094.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
C J C Cash, G M Treece, L H Berman, A H Gee, and R W Prager
3D reconstruction of the skeletal anatomy of the normal neonatal foot using 3D ultrasound
Br. J. Radiol.,
July 1, 2005;
78(931):
587 - 595.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
