V

PwA

KPW

a

b

c

d

e

RA

f

0d

accine (2008) 26, 1344—1352

avai lab le at www.sc iencedi rec t .com

journa l homepage: www.e lsev ier .com/ locate /vacc ine

reventing contamination between injectionsith multiple-use nozzle needle-free injectors:safety trial

imberly Kellya,∗, Anatoly Loskutovb, Darin Zehrunga, Kapaakea Puaaa,aul LaBarrea, Nancy Mullera, Wang Guiqiangc, Hui-guo Dingd, Darong Hue,illiam C. Blackwelder1

PATH, 1455 NW Leary Way, Seattle, WA 98107, United StatesPulse Needle Free Systems, Lenexa, KS, United StatesPeking University First Hospital, Beijing, ChinaCapital Medical University Beijing You an Hospital, Beijing, ChinaInstitute of Liver Diseases, General Hospital of Beijing Military Region, Beijing, China

eceived 3 October 2007; received in revised form 8 December 2007; accepted 19 December 2007vailable online 18 January 2008

KEYWORDSMultiple-use nozzlejet injectors(MUNJIs);Needle-freeinjection;Cross-contamination;

Summary Multiple-use nozzle jet injectors (MUNJIs), a type of needle-free injector, use ahigh-pressure stream to penetrate skin and deliver medicament. Concerns for their potential totransmit blood borne pathogens led to development of a hybrid MUNJI for use in mass immuniza-tions. The HSI-500®, referred to here as a protector cap needle-free injector (PCNFI), utilizesa disposable cap as a shield between the reusable injector nozzle and the skin to reduce therisk of contamination. This study aimed to determine the presence of hepatitis B virus (HBV)contamination in post-injection (‘‘next person’’) samples immediately following injection in

Hepatitis B virus(HBV)

HBV-carrier adults. Tolerability and pain were also assessed. The study ended early because thePCNFI failed to prevent contamination in the first batch tested (8.2% failure rate). The injec-tions were very well tolerated, with most followed by no bleeding (81.2%) or mild bleeding(7.8%). 55.2% of participants experienced no pain while 42.3% experienced mild pain followinginjection.© 2008 Elsevier Ltd. All rights re

I

∗ Corresponding author. Tel.: +1 206 285 3500;ax: +1 206 285 6619.

E-mail address: kkelly@path.org (K. Kelly).1 Independent Consultant.

Bocc

264-410X/$ — see front matter © 2008 Elsevier Ltd. All rights reserved.oi:10.1016/j.vaccine.2007.12.041

served.

ntroduction

y 1999, roughly one-third of all immunization and one-halff non-immunization injections in developing countries wereonsidered unsafe [1]. This suggests that reused syringesause millions of blood borne infections each year [2],

1345

M

P

Tfc((tcai

nbstream of an injection must first penetrate a thin plasticfilm, forming the smallest possible orifice for the injectionstream to pass through. The injection stream must thenpass through three additional, sequential coaxial orifices

Preventing contamination between injections

resulting in 1.3 million early deaths and US$535 million indirect medical costs [3]. Reuse of contaminated syringes,needle-stick injuries among health workers, and threats tothe community from improperly disposed and contaminatedsharps present serious health risks. While some develop-ing countries are addressing injection safety concerns byintroducing single-use, autodisable (AD) syringes for vac-cinations, the risks of transmitting chronic infections byneedle-stick injury from sharps waste will continue to bea leading hazard for health workers and waste disposal per-sonnel [4].

An alternative to needle and syringe injections areneedle-free injections that penetrate the skin with a high-pressure stream and deliver medication into intradermal,subcutaneous, or intramuscular tissues [5]. Some multiple-use nozzle jet injectors (MUNJIs) can have a short injectioncycle time, allowing rapid and cost-effective delivery ofa medicament to large numbers of people. This is espe-cially advantageous in epidemic situations. Studies show thepain of needle-free injection is generally less than or equalto the pain from a needle and syringe injection, thoughresults vary by study and may depend on the vaccine [6—8].Bleeding at the injection site is reported as more commonwith needle-free injection than with needles and syringes[9].

Evidence that MUNJIs could transmit blood bornepathogens between humans was documented in 1986, whena MUNJI was implicated in the transmission of hepatitisB virus (HBV) in 31 cases in a California clinic [10,11].These cases brought serious concerns and increased scrutinyregarding the safety of these devices [12,13]. In November1995, a joint meeting of the Centers for Disease Controland Prevention (CDC) and the World Health Organization(WHO) concluded that MUNJIs presented an unacceptablerisk to vaccine recipients. By December 1996, amid safetyconcerns during a large meningitis outbreak in northernNigeria, WHO revised its policy noting that needle-freeinjectors designed for use with multidose vials should notbe used for immunizations until safe needle-free injec-tors are identified through independent safety testing[14,15].

PATH has partnered with Pulse Needle Free Systems,formerly Felton International, of Kansas since 1998 in thespecification development, design, testing, and evaluationof a new-generation MUNJI. The design of this injector,designated the HSI-500® or the protector cap needle-freeinjector (PCNFI), uses a disposable plastic cap that acts asa shield between the injector nozzle and the skin to reducethe risk of contamination between injections. The PCNFIis designed for use in mass immunizations where the per-injection cost is estimated to be lower than the alternativeAD needle and syringe cost [16].

A similar small-scale pilot study to that reported here wasconducted in 2004 at Huntington Medical Research InstituteLiver Center in Los Angeles, California. That study, in whichno contamination was found, informed the design and plansfor this larger study in which we test the hypothesis that

the PCNFI would prevent contamination between injectionsin human volunteers infected with HBV. The primary goal ofthis study was to determine the presence of HBV contamina-tion in post-injection samples of sterile saline immediatelyfollowing injection with the PCNFI.

Fp

Figure 1 Foot pump and hydraulic hose.

aterials

rotector cap needle-free injectors

he PCNFI is powered by a spring that uses hydraulics and aoot pedal to provide the pressure needed to deliver a sub-utaneous injection (Fig. 1). It utilizes a reusable fluid pathFig. 2) attached to a vial containing the fluid for injectionvaccine or other medicament). The fluid path is loaded onop of the injector hand piece which includes a trigger. Theomponents of the injector are shown below in Figs. 1—3,nd Fig. 4 shows a fully assembled device as it would be usedn an injection session.

A single-use, disposable plastic cap shield separating theozzle from the skin is intended to reduce contaminationetween injections. The cap is designed such that the jet

igure 2 Protector cap needle-free injector reusable fluidath.

1346

Fw

owctrt

Fn

ci

hhippthfTdaotfvta

mftsstdmf

at

igure 3 Protector cap needle-free injector fully assembledith vaccine vial attached.

f 0.80 mm (the nozzle cup orifice), 0.85 mm (the centralasher orifice), and 2.00 mm (the base ring orifice) before

ontinuing unimpeded, penetrating the skin during an injec-ion. The series of four coaxial orifices is designed to reduceetrograde passage of infectious material from the injec-ion site onto the nozzle. Fig. 5 shows the cap and the

igure 4 All components of the fully assembled protector capeedle-free injector.

Tttwah00i0

S

AtS

T

BesDi

h

K. Kelly et al.

entral washer in the ‘‘enabled’’ configuration before andnjection.

After the injection is given, an interlock system in theand piece requires that the used cap be ejected from theand piece and a new cap installed before a subsequentnjection. During the ejection of the used cap from the handiece, a sequence of events occurs that disables the cap andrevents it from further use. The ejection sequence is ini-iated by the user hand-operating an ejection lever on theand piece which initially drives the cap nozzle cup structureorward while the cap is retained in the nozzle nut mount.his causes the cap nozzle cup structure to flex forward,isplacing the central washer from its mounting interfacend allowing it to fall away from a coaxial orientation of thether three orifices. Once displaced from coaxial alignment,he central washer acts as a barrier, impeding the jet streamrom passing through the cap. The cap is now disabled, pre-enting reuse of the cap for a second injection. Fig. 6 showshe cap and central washer in the ‘‘disabled’’ configuration,fter the central washer is displaced from its coaxial mount.

The cap is then pushed completely from the nozzle nutount, separating it from the injector nozzle in preparation

or disposal. A new cap can then be installed, which resetshe interlock system of the hand piece and allows for a sub-equent injection. If a used cap is installed, the interlockystem of the hand piece will reset, but the jet stream ofhe injection will not pass through the used cap due to theisplaced central washer that causes a non-coaxial align-ent of the series of orifices and impedes the jet stream

rom passing through the cap.We assigned serial numbers to each of the jet injectors

nd identification numbers to each of the fluid paths in ordero track the performance and use of devices at each site.his would also allow us to trace any positive samples backo the specific jet injector and fluid paths used for those par-icular injections in order to determine if a positive sampleas potentially due to device malfunction. The jet injectorsnd fluid paths assigned to each site are as follows: site 001ad jet injectors #001, 002, 003 and fluid paths #001, 002,03, 013, 020, 021, 022; site 002 had jet injectors #004, 005,06 and fluid paths #004, 005, 006, 014, 016; site 003 usednjectors #007, 008, 009 and fluid paths #007, 009, 010, 011,12, 018.

aline

sterile sodium chloride solution of 0.9% was used forhe injections. It was manufactured in Sichuan, China, byichuan Kelun Pharmaceutical Co., Ltd.

ests

lood tests to determine HBV viral load and participant

ligibility were conducted at each of the participatingites. Sites 001 and 002 used a test kit manufactured byaAn Gene Company Limited at Sun Yat-Sen University

n Guangzhou City, Guangdong Province, China.2 Site 003

2 The kits manufactured in China by Chinese manufacturers do notave brand names.

Preventing contamination between injections 1347

ge sh

M

S

T2p

Figure 5 New cap cross-section (central washer enabled) imacap when central washer is not displaced.

used a test kit manufactured by Shenzhen Piji Biotech-nology Company Limited in Shenzhen City, GuangdongProvince, China. Analysis of blood samples to confirm HBVviral load levels following each of the injection sessionswas conducted at a central lab at MDS Pharma facilitiesin Beijing, China, using the COBAS HBV AMPLICOR moni-tor assay manufactured by Roche, in Basel, Switzerland.

All tests were conducted according to the manufacturer’sinstructions.

Injector ejectate samples were tested using theUltraQualTM 1000 Assay manufactured by National GeneticsInstitute (NGI) in Los Angeles, California.

gDsor

Figure 6 Used cap cross-section (central washer displaced) imageto displacement of central washer.

ows injection stream passing through central axis of protector

ethods

tudy population

he study was conducted from July 2006 through October006, in Beijing, China, among HBV-carrier adults. Partici-ants were eligible if their HBV viral load was equal to or

reater than one million copies/mL as determined by HBVNA polymerase chain reaction (PCR) tests. China was cho-en as the study site because it has the highest prevalencef chronic HBV infections globally [17]. Volunteers wereecruited from three infectious disease hospitals within the

shows injection stream blocked from passing through cap due

1

m(tRPi

ealtaott

cItstc1clHgges

ftclc

S

THiWpvomF

mvFctcd

S0G

leUt(

mmoSaaAtPecvsipanpiowwwi

iopo

iiatsbtss

348

etropolitan Beijing area—–Peking University First Hospitalsite 001), Capital Medical University Beijing You an Hospi-al (site 002), and the General Hospital of Beijing Militaryegion (site 003). A contract research organization, MDSharma-Beijing, was hired to assist in study coordination andmplementation.

Men and women between the ages of 18 and 60 wereligible to participate if they were otherwise healthy andctively infected with HBV.3 Participants with HBV DNA viraloads of 106 or greater copies/mL were included. In ordero determine eligibility based on this viral load level, weccepted the results of HBV DNA tests conducted at eachf the respective study sites which are routinely used byhe hospitals for patient diagnosis and determination ofreatment.4

An HBV DNA viral load of 106 or greater copies/mL indi-ates that an individual’s blood is highly infectious with HBV.n 1984, Feinman et al. showed that HBV can be transmit-ed in as little as 10 pL (picoliters) of blood [18]. While thetudy did not specify the amount of HBV viremia transmit-ed in the 10 pL of blood, we assumed that 10 pL of bloodorresponds to 10 viral DNA’s when the viral load is equal to06 HBV DNA/mL of blood. The Chinese study investigatorsonfirmed that patients often experience viremia with viraloads higher than 106 HBV DNA/mL. Therefore, we chose anBV DNA viral load of one million copies/mL as a cut offiven that we could reasonably meet our enrollment tar-et of 300 participants and maximize the possibility thatven the slightest amount of contamination in post-injectionamples would be detected.

Individuals were not admitted to the study if any of theollowing criteria were present: (1) history of allergic reac-ions or anaphylaxis to immunizations, (2) dermatologicalonditions affecting the injection site, or (3) blood coagu-ation disorders or history of taking drugs that affect bloodoagulation, including any heparin-based medication.

tudy design and procedures

his open-label, non-randomized study was approved by theuman Subjects Protection Committee at PATH and the Eth-

cal Review Committee at Peking University First Hospital.ritten informed consent was obtained from all volunteersrior to any study procedures, and confidentiality and pri-

acy were maintained at all times. Device certification wasbtained from the National Institute for the Control of Phar-aceutical and Biological Products (NICPBP), a China State

ood and Drug Administration (SFDA)-accredited testing

3 Active infections were confirmed by a recent HBV DNA poly-erase chain reaction (PCR) test, and eligibility was confirmed if

iral loads were equal to or greater than one million copies/mL.or this study, ‘‘recent’’ was defined as within 14 days for patientsurrently on treatment for chronic HBV infection, within 21 days forhose not currently on treatment, and within 60 days for those noturrently on treatment and who would submit to a screening bloodraw for confirmation of their HBV viral load.4 Site 001 and site 002: DaAn Gene Company Limited. Sun Yat-en University, Guangzhou City, Guangdong Province, China; site03: Shenzhen Piji Biotechnology Company, Limited. Shenzhen City,uangdong Province, China.

O

Tp

ttl

L

Sn(L

K. Kelly et al.

aboratory. NICPBP verified device performance and adher-nce to ISO 21649:2006, ‘‘Needle-Free Injectors for Medicalse.’’ Additionally, the device is cleared for marketing byhe United States Food and Drug Administration (USFDA)#K041280).

Volunteers were identified by reviewing current patients’edical records at the three participating hospitals. Stricteasures were put in place in order to minimize the chances

f contamination of study samples due to study staff error.tudy staff was required to wear sterile latex gloves forll study procedures. During injections, the injector oper-tor installed a sterile cap onto the nozzle of the injector.newly sterilized fluid path was used for each volunteer

o minimize contamination risk between study volunteers.rior to volunteer injection, a clean sample of 0.5 mL salinejectate was obtained. The cap used in this step was dis-arded and a new cap was installed onto the injector. Theolunteer was then injected with 0.5 mL of 0.9% sterilealine in the right deltoid region. The cap used in thisnjection was discarded. A new cap was installed and aost-injection sample of saline ejectate was delivered intosample collection tube. The operator then installed a

ew sterile cap and repeated the same injection and sam-ling process in the left deltoid. Following each humannjection, a picture was taken of the injection site inrder to record any local reactions. 5.0 mL of venous bloodere collected in a separate room by a staff member whoas not involved in the injection procedures. The bloodas tested to confirm HBV viral load on the day of the

njections.Clean control samples taken immediately prior to the first

njection were used for confirmation that the PCNFI was freef HBV contamination prior to injection. An air control sam-le was taken during each session to evaluate the presencef any environmental HBV contamination.

Following the two injections in each volunteer, thenjector fluid path was fully disassembled and placedn an autoclavable pouch with a sterilization indicatornd sterilized at 121 ◦C for a minimum of 30 min. Afterhis cycle, the fluid path was reassembled and againterilized in an autoclavable pouch. The fully assem-led fluid paths were removed from the pouches at theime of the next injection to ensure sterility. Use andterilization logs were documented for each injectionession.

utcome measures

he primary safety endpoint was the presence of HBV in theost-injection samples.

Secondary safety endpoints were analyzed by pho-ographing and inspecting each injection site, interviewinghe volunteer to assess the pain of the injections, and ana-yzing any adverse event reports.

aboratory assessments

amples were shipped from China to Los Angeles, Califor-ia, for testing. A PCR-based assay for detection of HBVUltraQualTM 1000 Assay—–National Genetics Institute (NGI),os Angeles, California) was used to determine the presence

ap9

obworIai

F

Ps

OI1Tie

p4etS((pp

pw(bbbmf

Tiw

T

Ti‘

Preventing contamination between injections

of HBV contamination in post-injection samples. Althoughthis qualitative assay does not allow measuring the exactamount of viral DNA, it is more sensitive by an order of mag-nitude and more accurate than other qualitative methods.

In order to use this test in our study, validation testingof the method was performed at NGI. The purpose of thetesting was to determine the analytical sensitivity of theassay to detect trace amounts of HBV in saline samples andvalidate the suitability of the assay for use in this study.The testing was also designed to demonstrate whether anydestruction of the virus occurred (due to the force of theinjector device).

In order to determine the analytical sensitivity, anidentical set of HBV samples in saline were aliquotedusing either a pipette or the PCNFI. Eight concentrations(0.094—12.000 IU/mL) in each panel were assayed eighttimes using the NGI HBV PCR UltraQualTM 1000 Assay (0.5 mLsample extraction, one reaction per primer pair, for a totalof four). The same lot of baculovirus containing plasma wasadded to the samples aliquoted to serve as a control.

Results showed that NGI’s PCR-based assay is highlysensitive to detection of HBV in samples. The mean sen-sitivity and 95% detection limit of the assay was 1.589(1.227—2.058) IU/mL and 6.316 IU/mL, respectively.5

The estimated mean sensitivities for the dilution seriesaliquoted using a pipette and injector were not statisticallydifferent at a 95% confidence level. This suggests nosignificant viral destruction from passing HBV through theinjector. Finally, the testing also showed that the assay hadacceptable precision when analyzed on different days bydifferent analyst groups.

Serum specimens from volunteers following injectionwere analyzed to verify that viral load levels were ≥106

copies/mL on the day of injections. HBV DNA was detectedusing the COBAS HBV AMPLICOR monitor assay (lower limitof detection, 200 HBV copies/mL). Testing on the blood sam-ples collected on the day of injection was conducted at MDSPharma in Beijing, China.

Statistical considerations

We planned to enroll 300 volunteers and to collect two post-injection samples from each volunteer (600 data points). Weplanned for a 0% threshold for contamination, meaning thatwe expected no viral, lab, or environmental contaminationin any of the samples. If there were no contaminated sam-ples with this sample size, the upper limit of a two-sided95% confidence interval for the probability of contamina-tion would be 0.6%, assuming statistical independence of allresults.

On recommendation of the ethical review committees,batch sampling and iterative data analysis were planned toidentify the presence of HBV in the post-injection samplesamong early volunteers. If presence of HBV contamination

was found, the study team would thoroughly review the sit-uation and determine if the study should be stopped early.Interim testing on batches of samples from 100 volunteerswas planned.

5 Conversion equation: 1 International Unit (IU) = 3.44 copies.

iiMtaia

1349

The primary analysis consisted of estimation of the prob-bility of contamination after injection by the observedroportion of contaminated samples and an exact two-sided5% confidence interval.

Tolerability of injections was evaluated using categoriesf local reactions and pain. Local reactions were classifiedy presence and severity of bleeding and/or wheal. Whealsere reported as mild (>0—5 mm), moderate (>5—10 mm),r severe (>10 mm). Staff was trained to classify theseeactions using pictures and a wheal measurement guide.mmediately following each injection, volunteers weresked to rate any pain they experienced as a result of thenjections as none, mild, moderate, or severe.

indings

revention of contamination in post-injectionamples

f 294 volunteers screened for the study, 285 were enrolled.n accordance with the protocol, batch sampling on the first04 enrolled volunteers across all three sites was conducted.hree samples (one clean control, one right deltoid post-

njection sample, and one left post-injection sample) fromach volunteer were tested (N = 312).

Of the 208 post-injection samples, 8.2% (17/208) wereositive for HBV with an exact 95% confidence interval of.8—12.8%, assuming independence of the two results forach volunteer. The 17 positive samples came from 12 volun-eers; 5 volunteers had positive samples from both deltoids.tudy site 001 had 4.7% positive post-injection samples4/86), site 002 had 12.5% positive post-injection samples8/64), and site 003 had 8.6% positive post-injection sam-les (5/58). Table 1 provides detailed information for eachositive sample.

Of the 104 volunteers, the numbers of positive sam-les for contamination for neither, one, or both samplesere 92 (88.5%), 7 (6.7%), and 5 (4.8%), respectively

see Table 2). There was significantly more agreementetween the two samples for an individual, — i.e., eitheroth negative or, especially, both positive — than woulde expected by chance alone (kappa measure of agree-ent = 0.55, p < 0.001). Thus, the results for the two samples

rom an individual were not independent.None of the clean injector controls were positive (0/104).

hese results warranted testing of the air controls from allnjection sessions for the first 104 volunteers. These samplesere all negative for HBV contamination (0/27).

olerability of injections

olerability for all 285 volunteers was assessed. The major-ty of reactions (507/570, 89%), were classified as either‘none’’ or ‘‘mild bleeding’’ (Table 3). Moderate bleed-ng occurred after only 0.5% (3/570) of the injections. Noncidents of severe bleeding or bruising were reported.

ild intradermal wheals, an indication that the injec-

ion was partially delivered intradermally, were reportedfter 7.4% (42/570) of the injections. Additionally, 3/570njections (0.5%) were followed by both mild bleedingnd a mild wheal. Moderate and severe wheals were

1350K.

Kellyet

al.

Table 1 Positive post-injection samples by site and volunteer number

Studysite-volunteernumber

Post-injectionsamplepositivea

Serial numberof fluid pathusedb

ID number ofjet injectorusedb

Indications from injection site photographc HBV oninjection day(copies/mL)d

001—006First

#003 #003Right: no visible blood, slight reflux

8.00E + 07Second Left: inconclusive

001—020 Second #003 #002 Left: no visible blood, moderate reflux 1.60E + 09001—029 First #003 #002 Right: 8-10 mm intradermal wheal, no

visible blood, no reflux9.36E + 08

002—002First

#004 #004Right: single droplet of blood at theinjection site, no reflux 6.31E + 08

Second Left: 6—8 mm intradermal wheal, no visibleblood, significant reflux

002—005 Second #014 #004 Left: reflux with capillary bleeding 1.31E + 09002—020 First #006 #004 Right: 6-8 mm intradermal wheal, no visible

blood, no bleed-back1.63E + 08

002—027 First #005 #004 Right: no visible blood, very light reflux 3.76E + 09

002—028First

#004 #004Right: 3—4 mm intradermal wheal, lightreflux with light capillary bleeding 1.33E + 09

Second Left: no visible blood, no bleed-back

002—300 First #014 #004 Right: capillary bleeding, no bleed-back 6.89E + 08003—011 Second #007 #007 Left: no visible blood, very light reflux 2.10E + 08

003—013First

#012 #007Picture not available 2.21

E + 08Second Picture not available

003—017First

#012 #007Large 8—10 mm intradermal wheals andsignificant reflux observed. 4.25E + 08

Second Picture not availablea The first injection was delivered into the right deltoid, the second injection was delivered into the left deltoid.b Each fluid path and jet injector were individually numbered for identification.c Splash-back = injectate fluid on skin surface that did not penetrate the tissue. Injectate reflux = injectate fluid that flows back out of the tissue at the injection site. Bleed-back = capillary

bleeding at the injection site after injection.d Quantitites of HBV in participants’ blood were confirmed to be above the eligible level of a ≥106 copies/mL by PCR DNA assay after the study injections.

Preventing contamination between injections

Table 2 Results of the HBV PCR analysis by sample pairs

First injection(right arm)

Second injection(left arm)

Number ofobserved pairs

Negative Negative 92Negative Positive 3Positive Negative 4

dtiwp

o((opprc

pbftiiuist

iifobwbr

p

Positive Positive 5

All All 104

reported after 2.3% (13/570) and 0.4% (2/570) of injections,respectively.

Nearly all of the injections (556/570, 97.5%) were asso-ciated with either no pain or only mild pain (Table 4).Moderate pain was experienced following 2.1% (12/570) ofthe injections and severe pain was experienced with only0.4% (2/570) of injections.

Safety

No serious adverse events were reported.

Discussion

The investigators decided to end the study early based uponthe discovery of 17 positive samples in the first 208 (8%)post-injection samples, which demonstrated a significantlyhigher contamination rate than the 0% threshold. Additionalfactors contributing to this decision include the fact thatpositive samples were found across all study sites and, withthe exception of site 001, positive samples were traced tomore than one fluid path. Furthermore, the absence of anypositive clean control samples (0/104) validated the testing

mechanism at NGI labs and suggested that false positiveswere not likely.

The relatively frequent occurrence of two contaminatedsamples for a volunteer is most likely due to the studydesign. The fluid path was sterilized between injections for

vtvts

Table 3 Local reaction within 15 s following injection

None Mild bleeding Moderatebleeding

Mild

Right arm 228/285 80.0% 28/285 9.8% 1/285 0.4% 21/Left arm 235/285 82.5% 16/285 5.6% 2/285 0.7% 21/

Total 463/570 44/570 3/570 42/81.2% 7.7% 0.5% 7.4%

Table 4 Level of injection pain

None Mild

Right arm 166/285 58.2% 111/285 3Left arm 149/285 52.2% 130/285 4

Total 315/570 55.2% 241/570 4

1351

ifferent individuals, but no sterilizing was done betweenhe two samples from one individual. Thus, the probabil-ty of a positive second sample given a positive first sampleas increased due to residual contamination left on the fluidath.

For seven individuals, one sample was positive and thether negative (Table 2). Three cases were a negative firstright) post-injection sample followed by a positive secondleft) post-injection sample. A negative first/negative sec-nd sample pairing indicates that the device was functioningroperly. Conversely, a negative first/positive second sampleairing indicates device malfunction in the second injection,esulting in contamination of the nozzle or cap and thus aontaminated second post-injection sample.

The causality of positive first/negative second sampleairings, which occurred four times, is less clear. One possi-ility is that the contamination on the injector fluid pathrom the first deltoid injection was flushed away duringhe post-injection flow from the second deltoid injectionnto the second (left) sample tube, and possibly even dur-ng the second deltoid injection itself. This could leavendetectable levels of contamination in the second post-njection sample. The UltraQualTM1000 Assay, while highlyensitive for the detection of virons, provides only a quali-ative result.

Potential causes for positive first/negative second post-njection samples not associated with injector failurenclude inadvertent sample contamination from hospital sur-aces or staff and providing the injection sequence in thepposite order (injection in the left deltoid first followedy injection in the right). All air control samples testedere found to be negative, and it is unlikely that air-orne contamination contributed to any of the positive testesults.

Localized tolerability and volunteers’ perception of theain results show that injection with the PCNFI device is

ery well tolerated and does not lead to undue physicalrauma or pain. This is similar to what was found in pre-ious studies involving needle-free injectors compared toraditional needles and syringes [6—8]. Results in this studyhowed bleeding to be infrequent and mild.

wheal Moderatewheal

Severewheal

Mild bleedingand wheal

285 7.4% 6/285 2.1% 1/285 0.4% 0/285 0.0%285 7.4% 7/285 2.5% 1/285 0.4% 3/285 1.1%

570 13/570 2/570 3/5702.3% 0.4% 0.5%

Moderate Severe

9.0% 7/285 2.5% 1/285 0.4%5.6% 5/285 1.8% 1/285 0.4%

2.3% 12/570 2.1% 2/570 0.4%

1

ttifmmmdatbd(itbsetimt

A

PMavwistMTAtPG

R

[

[

[

[

[

[

[

[

352

Pulse Needle Free Systems is currently working to iden-ify the root cause of the PCNFI contamination identified inhis safety study. Several design features are being stud-ed to ascertain how contamination may occur. Areas ofocus include the cap ejection and disabling process (whichay have contributed to nozzle contamination due to theechanical forces involved with central washer displace-ent during cap ejection from the nozzle nut mount), theesign of the internal portion of the protector cap (toccommodate an increased volume of fluid during the injec-ion process), and manufacturing/assembly processes foroth the injector and projector cap. This work involves (1)imensional analysis of the injection system components,2) performance evaluations of the injection systems usedn the PATH safety study and other injection systems fromhe same lot of manufacture, (3) comparisons of laboratory-ased contamination models to actual field contaminationettings, and (4) laboratory-based contamination risk mod-ling using fluorescein and other markers. Data from thisesting may lead to additional contamination models. Test-ng results will guide the corrective actions and/or designodifications necessary to eliminate the risk of contamina-

ion.

cknowledgments

ATH was generously supported in this study by the Bill &elinda Gates Foundation. The opinions expressed hereinre those of the authors and do not necessarily reflect theiews of the Bill & Melinda Gates Foundation. The authorsould like to thank the talented staff at all three partic-

pating study sites for their assistance in conducting thistudy. The authors would also like to recognize the con-ributions made to the study and this manuscript by Dr.artin Friede, Dr. Mark Kane, Mr. John Lloyd, Dr. Myronong, Ms. Jennifer Handschin, Dr. Kathy Neuzil, Dr. Jefflbrecht, Ms. Shawna Pine, Dr. Lixia Wang, Dr. Junfeng Yang,he staff at the PATH office in Beijing, the staff at MDS-harma, Beijing, Sarah McGray, Louise Downing, and Emilyriswold.

eferences

[1] Simonsen L, Kane A, Lloyd J, Zaffran M, Kane M. Unsafeinjections in the developing world and transmission ofblood borne pathogens: a review. Bull World Health Organ1999;77:789—800.

[2] Kane A, Lloyd J, Zaffran M, Simonsen L, Kane M. Transmis-sion of hepatitis B, hepatitis C and human immunodeficiencyviruses through unsafe injections in the developing world:model-based regional estimates. Bull World Health Organ1999;77:801—7.

[

K. Kelly et al.

[3] Miller MA, Pisani E. The cost of unsafe injections. Bull WorldHealth Organ 1999;77:808—11.

[4] Steinglass R, Boyd D, Grabowsky M, Laghari AG, Khan MA, QaviA, et al. Safety, effectiveness and ease of use of a non-reusablesyringe in a developing country immunization programme. BullWorld Health Organ 1995;73:57—63.

[5] Mitragotri S. Immunization without needles. Nat Rev Immunol2005;5(12):905—16.

[6] Epstein JE, Gorak EJ, Charoenvit Y, Wang R, Freydberg N,Osinowo O, et al. Safety, tolerability, and lack of antibodyresponses after administration of a PfCSP DNA malaria vac-cine via needle or needle-free jet injection, and comparisonof intramuscular and combination intramuscular/intradermalroutes. Hum Gene Ther 2002;13(13):1551—60.

[7] Sarno MJ, Blase E, Galindo N, Ramirez R, Schirmer CL, Trujillo-Juarez DF. Clinical immunogenicity of measles, mumps andrubella vaccine delivered by the Injex jet injector: com-parison with standard syringe injection. Pediatr Infect Dis J2000;19(9):839—42.

[8] Greenberg RS, Maxwell LG, Zahurak M, Yaster M. Preanestheticmedication of children with midazolam using the Biojector jetinjector. Anesthesiology 1995;83(2):264—9.

[9] Weniger, BG. Jet injection with vaccines: overview andchallenges of mass vaccination with jet injectors, conferencepresentation, Innovative Administration Systems for vaccines,Rockville, MD, 18—19 December, 2003, http://www.hhs.gov/nvpo/meetings/dec2003/Contents/ThursdayPM/Weniger.pdf.Accessed April 23, 2007.

10] Centers for Disease Control and Prevention. Hepatitis B asso-ciated with jet gun injection-California. Morb Mortal Wkly Rep1986;35:373—6.

11] Canter J, Mackey K, Good LS, Roberto RR, Chin J, Bond WW, etal. An outbreak of Hepatitis B associated with jet injections ina weight reduction clinic. Arch Intern Med 1990;150:1923—7.

12] Robertson JS. Jet injectors and infection. Public Health1987;101:147—9.

13] Zachoval R, Deinhardt F, Gurtler L, Eisenburg J, Korger G. Riskof virus transmission by jet injection. Lancet 1988;1:189.

14] World Health Organization. Safety of Injections in Immu-nization Services: WHO Recommended Policy. Geneva: WorldHealth Organization; 1996. WHO/EPI/LHIS/96.05.

15] Jet Injectors for Immunization. Current Practice and Safety;improving designs for the future. Statement by Partic-ipants and Agenda, Atlanta, Georgia, October 2—4, 1996.http://www.cdc.gov/od/science/iso/pdf files/JetInjectImmu-niz-CurrentPracticeSafety-ParticipantStatement-1996Oct04.pdf Accessed November 12, 2007.

16] Ekwueme DU, Weniger BG, Chen RT. Model-based estimates ofrisks of disease transmission and economic costs of seven injec-tion devices in sub-Saharan Africa. Bull World Health Organ2002;80(11):859—70. ISSN 0042-9686.

17] Merican J, Guan R, Amarapuka D, Alexander MJ, Chutaputti

A, Chien RN, et al. Chronic hepatitis B virus infection in Asiancountries. J Gastroenterol Hepatol 2000;15:1356—61.

18] Feinman SV, Berris B, Guha A, Sooknanan R, Bradley DW, BondWW, et al. DNA: DNA hybridization method for the diagnosis ofhepatitis B infection. J Virol Methods 1984;8:199—206.