The Scottish Government

« Previous | Contents | Next »

Listen

CHAPTER TWO: OVERVIEW OF THE LABORATORY STUDIES

2.1 This chapter is concerned with the first aim of the study:

To test paraphernalia items and injection preparation methods in the laboratory to quantify the theoretical benefits and/or risks that they present to health, thereby identifying from those tested the items of paraphernalia and preparation methods that present the least theoretical risks to individual health.

2.2 Several objectives were set in order to fulfil this aim. These were to:

a) Develop experimental methods for use in the laboratory that replicate the injection preparation practices of IDUs, based on the ethnographic work of Taylor et al (2004) and previous work.

b) Identify the key equipment variables and method variables in the preparation process to be investigated in the laboratory experiments.

c) Prepare injections using the developed method. Control all the variables in the preparation process (equipment and method) to allow the study of the impact of each variable.

d) Study the impact of each variable by performing scientific experiments on injections prepared in different ways using different equipment.

e) Where possible, benchmark the prepared injection results against standards used within the pharmaceutical industry for small volume injections and other relevant aspects of aseptic ('sterile') manufacturing. This will allow comparison of the preparation method and paraphernalia against theoretical standards that present minimal risk.

f) Establish the contents of a safer injection 'kit' and preparation method which presents the lowest theoretical risks to health based on the laboratory results.

Development of the standardised injection preparation process

2.2 Objectives covered in this section:

a) Develop experimental methods for use in the laboratory that replicate the injection preparation practices of IDUs.

b) Identify the key equipment variables and method variables in the preparation process to be investigated in the laboratory experiments.

Background and summary of method

2.3 Scientific experimentation requires all the variables that impact on a process to be controllable. This allows for the effects of a change in one to be studied. In the study of pharmaceutical injections this is simple as the manufacturing process is standardised and can therefore be easily controlled. One of the first challenges for this study was the need to develop a method of preparing injections that copies what IDUs do as closely as possible, but is controllable so it can be used in the laboratory.

2.4 Prior to this study, Professor Avril Taylor, University of Paisley, was commissioned by the Scottish Executive to undertake an ethnographic study of injecting practices. Her study (Taylor et al, 2004), focused on a sample of injectors from Glasgow and data was recorded using video. This provided a rich source of information on blood borne virus risk taking behaviours and also clear illustration of how those involved prepared and administered their injections. This information was systematically gathered for the study here by completion of a questionnaire by Taylor's researchers, while replaying each filmed injecting episode. Sixty questionnaires were completed. The data was analysed to establish a common preparation method used for a single person injection and practices that deviated from this. The results informed the development of a standardised injection preparation process that was copied in the laboratory. This process reflected as closely as was scientifically possible the 'typical' preparation method used by IDUs, balancing the scientific need for accuracy and reproducibility. Where a choice in the process had to be made, advice given in training information for needle exchange workers such as the Safer Injecting Briefing (Derricott et al, 1999) was used. It was considered better practice to investigate the potentially safer methods of preparation which are likely to be advocated by needle exchange staff. Full detail of the method used to develop the standardised injection preparation process is available from the author j.a.scott@bath.ac.uk.

The findings and how they informed the laboratory work

Drugs used by injection

2.5 As expected, heroin was the most common drug injected in the work of Taylor et al. It is also the most common drug injected in Scotland (Parry et al, 2004). The laboratory work investigated the preparation of heroin injections. Future work could examine other common drugs such as crack cocaine and amphetamines.

2.6 It was chosen to investigate preparation of injections for use by one person as sharing should be discouraged. However, it should be noted that 42% (n=25) of the preparation episodes documented were for batch preparation (i.e. for injections that were subsequently divided amongst injectors).

Drug quantities

2.7 Heroin quantities were described by those in the video study by their value in pounds sterling. A 'ten pound bag' was the most common quantity used when preparing injections for one person. This was therefore chosen for the laboratory method, but first, the corresponding average weight had to be estimated. This was done based on several sources of information: (i) advice from Avon and Somerset Constabulary (ii) advice from the Independent Drug Monitoring Unit ( IDMU) in Edinburgh, (iii) previous work with IDUs that measured fake powders (Ponton & Scott, 2004). A full description of what was done can be provided. The resulting weight of street heroin corresponding to a 'ten pound bag' used in the laboratory was 130mg.

Preparation of injections

Hand washing

2.8 Only one person was noted on Taylor's videos washing their hands prior to preparation, even though 83% (n=50) of injections were prepared in a building considered likely to have access to washing facilities. The Health Protection Agency 'Shooting Up' update report (2006) expresses concern about the increase in bacterial infections seen in IDUs. Unclean hands and skin could be a source of such bacteria. No research could be found examining the extent of microbial contamination on IDUs hands or the impact of hand washing in IDUs. Common sense and the application of knowledge from surgical and food hygiene suggests that hand washing prior to injection preparation should be encouraged. The lack of hand washing in the data from Taylor et al suggests it may not be seen as important by IDUs. The extent of contamination of IDUs hands, along with barriers to hand washing were further investigated in the 'Hand washing study' (section 2.2). A comparison of hand washing with a quicker and potentially more convenient source of hand cleansing using alcohol hand rub (alone without water) was also undertaken.

Order of preparation steps

2.9 Most IDUs in the videos used six steps in the preparation of heroin. The order of these steps varied with some, but the majority used the process summarised in figure one:

Figure 1: heroin injection preparation process most commonly used for single and multiple person injections

2.10 All heroin injectors added an acid and heated the injection. All but one used a filter. This was consistent with previous work with IDUs in the South West (Ponton & Scott, 2004). Detail of each step was required from the video data, in order to inform the laboratory standard method.

STEP ONE: Add drug to cooker:

2.11 The most common cooker in the videos was the teaspoon. The alternative used was the bottom of a drinks can. In Europe, and in some needle exchanges in the UK, the supply of commercially produced cookers is undertaken. When this work began, the Stericup" was the only one available. This is a thin aluminium bowl with a handle. A plastic cover placed over the handle protects the fingers from heat transfer. The Stericup is intended for disposal after use. It becomes more fragile after use due to the heat contorting the aluminium. The Stericup is sterile so should not contribute bacterial, fungal or viral contamination to the injection. For more information see: http://www.apothicom.org/index.php (English language version available).

2.12 There is a need to investigate the use of the Stericups as an alternative to injectors using their own spoons or drinks cans. Specific questions are:

• Do Stericups prevent IDUs sharing cookers?
• Do Stericups discourage reuse?
• Does the Stericup leak any aluminium into the injection and if so, are levels safe?

2.13 The first two questions are discussed in the qualitative section of the practice based study; the last question (3) was investigated in the laboratory, in the 'Stericup tests' (section 2.3).

STEP TWO: Add acid to cooker:

2.14 Acids are a necessary part of the injection preparation process. They have potential however to irritate veins, especially if too much is used. Most street heroin in the UK is in the chemical form called 'base' (King, 1997). Crack cocaine is also in base form. Bases are poorly soluble in small volumes of water, such as those used to prepare injections. They vaporise on heating so suitable for smoking. The acid converts the base drug into a soluble form by a process called 'protonation', which makes the drug dissolve in water. The heroin still exerts the same psychoactive effect.

2.15 Various amounts of acid were observed to be added on the videos. For the laboratory method a standard amount of 70mg, calculated as representing a 'pinch' of citric acid was chosen, with ascorbic acid and various amounts of both acids also investigated. The laboratory experiments were done to predict how much acid is needed, so the minimum amount can be advised, to reduce vein risks.

2.16 Citric acid sachets were supplied in Glasgow when the video work of Taylor et al was done. They were used in 41% (n=20) of heroin injection preparations recorded. Citric acid from other sources e.g. cooking grade citric and lemon juice were also observed. No ascorbic acid (vitamin C) was used. The use of various other household acids has been reported elsewhere (Ponton & Scott, 2004). The provision of acid of high quality (British Pharmacopoeial - BP grade) to IDUs, in sterile sachets has the potential to bring several benefits:

a) to ensure that the acid does not contribute contaminants or bacteria to the injections. This would allow needle exchanges to be reassured of the quality of the product supplied.

b) to perform favourably for clients so attract them into service, thus allowing the opportunity to deliver harm reduction messages and signpost into treatment services.

c) to minimise the risks to health, if the least risky BP grade acid and quantity can be established.

2.17 The BP grade and sterile nature of the acids assures that the first benefit is met. BP limits mean the acid has passed quality control tests e.g. for fungal contamination which can occur naturally in citric acid. The second benefit has been demonstrated by Garden et al (2003). However the third benefit has not been shown. Hence the question 'Which acid is least risky?' was studied in the laboratory. Both ascorbic acid BP (vitamin C) and citric acid BP are available in sterile sachets and both are permitted within the legislation. 'The acid tests' in laboratory section 2.4, address this.

STEP THREE: Add water to cooker:

2.18 For most of the injections videoed tap water was used. Bottled water tended to be used when preparation occurred outside. The most common volume used to prepare a £10 bag of heroin was 0.7ml, so this was taken as quantity used in the standard laboratory preparation method. This is similar to previous findings of 0.8ml (Scott, 2000, Ponton and Scott, 2004).

2.19 Sterile Water for Injection ( BP) ( WFI) is water that has been prepared to injectable pharmaceutical standards. It is sterilised, so is free of microbiological contamination and also has limited quantities of particles in it. This means there is minimal risk of this water contributing to any infection or damage to the vascular system. Since water can be a source of bacteria and shared contaminated water may transmit blood borne viruses ( BBVs), sterile water, in single use ampoules small enough to prevent sharing could have advantages. The supply of Sterile Water would also allow needle exchanges to be reassured that the water they were supplying was 'fit for purpose'. The advantages of single use sterile water ampoules prepared to pharmaceutical standards of sterility and particle content, over a cup of boiled and cooled tap water are evident, so laboratory experiments to further verify this were considered unnecessary.

STEP FOUR: Heating

2.20 In the videos all injections were heated as part of the preparation process. Therefore heat was used in the laboratory. The length of time of heating on the videos was variable. Several factors could influence this including volume and temperature of the water, thickness and type of spoon metal, the source of the heat and distance from the spoon. From previous work (Scott, 2000) and the videos, the established heating end point was when the injection 'bubbled and went clear'. Hence it was chosen to use this as the end point of heating in the laboratory. Figure 2 gives an illustration of a prepared injection after the endpoint.

Figure 2: clear heroin solution after heating with acid

STEP FIVE: Stirring

2.21 Stirring was used during the preparation of the majority of injections that were videoed, so it was used in the standard laboratory method. The end of the needle cap was used to do this. A new syringe was used each time and stirring undertaken until most powder was seen to dissolve.

STEP SIX: Filtering

2.22 Makeshift filters are used by IDUs to prevent the needle blocking during injection administration. It is believed that filters remove insoluble materials. The use of filters is advocated in harm reduction leaflets, as it is perceived that they reduce the risks of thrombus and granulomas. These can be caused by the injection of insoluble particles. However makeshift filters are not fit for purpose and needle exchange suppliers need to consider the performance and requirements of a filter when selecting one to supply.

2.23 A ¼ piece of cigarette filter was most commonly used as a filter in the videos. Cotton wool and hand rolling filter were also sometimes used. These are all 'makeshift' items and may not be fit for purpose. The laboratory experiments aimed to investigate whether these filters could potentially reduce risks and compare them with commercially available ones.

Commercial manufacturers are starting to produce filters for use by IDUs. At the start of this work the Sterifilt" had recently been produced by Association Apothicom in France. It is a polypropylene filter which fits over the needle and grips the sides of the syringe tip. It is designed to remove particles over 10 microns from solutions. It is illustrated in figure 3.

Figure 3: Sterifilt in use and close up of the filter. © Association Apothicom [used with permission]

2.24 The advantage of the Sterifilt is it is sterile and therefore new ones will not contribute microbiological contaminants into the injection. This has advantages in reassuring needle exchanges. The Sterifilt also has minimal absorption capacity so should not retain much drug. The Sterifilt has therefore less potential in theory of being kept for future reuse or sharing. Other commercially made filters were available at the start of this work. The Stericup filter was a small cellulose acetate filter supplied with Stericup cookers. 'Wheel filters' of the type used in medical and laboratory research were known to be supplied in Australia to IDUs. This raises the question of how the commercially available filters compare against each other and with makeshift filters.

2.25 The video data showed that filters were sometimes saved for later reuse or 'bashing down' to release trapped drug. Hence if filters were found to retain drug it was considered that they would likely be saved. Hence drug retention capacity was an important aspect of all filters to study. The filter experiments are summarised in section 2.5 'The filter tests'.

Development of the standardised injection preparation method

2.26 Based on the video data from Taylor et al and previous work, the standard method of injection preparation used in the laboratory was as shown in figure 4:

Figure 4: the basic preparation method for use in the laboratory.

2.27 In summary, the questions to be investigated were:

• How extensively microbiologically contaminated are IDUs hands?
• What barriers prevent IDUs from washing their hands prior to injecting?
• Is alcohol hand rub (a quicker and more convenient hand sanitiser) as good as soap and water?
• Do Stericup leak aluminium to the injection solutions and if so, are the levels safe?
• Which sterile acid presents the least theoretical risk to veins?
• What minimum quantities of acid dissolve heroin?
• How do commercially available filters (e.g. Sterifilt, Stericup filter and 'wheel' filters) compare against each other and with 'makeshift' filters used by IDUs (e.g. cigarette filters, hand rolling cigarette filters and cotton bud tips)
• To what extent do various filters retain drug and hence are likely to promote their retention for reuse or sharing?

Materials and methods

2.28 It is convention in scientific research to give full detail of the materials and experimental methods. This includes listing manufacturers of materials and describing equipment test conditions and settings. As this report is for a multidisciplinary audience, such detail has not been reproduced here. Instead, brief information has been given. Full detail can be obtained from the author j.a.scott@bath.ac.uk. The heroin samples were obtained from Avon & Somerset Constabulary Forensic Science Dept. A controlled drugs licence is held by the researcher and all storage, documentation and destruction have been verified by the police.

The Hand Washing Study

2.29 The hand washing study was designed to answer the questions:

• How extensively microbiologically contaminated are IDUs hands?
• What barriers prevent IDUs from washing their hands prior to injecting?
• Is alcohol hand rub as good a sanitiser as soap and water?

Method

Collection of the data

2.30 A study was designed to investigate the extent of contamination on IDUs hands and to compare the effectiveness of two different hand cleansing methods, which were:

(i) washing with Simple" soap and water and drying with a paper towel (a two step process that requires access to a sink)

(ii) rubbing with Spiri-Gel" 70% denatured ethanol hand rub until hands felt dry (a one step process that can be undertaken conveniently if gel carried e.g. belt clip packs as used by nurses)

2.31 The study was undertaken in association with Bristol Drugs Project ( BDP). Participants were recruited from IDUs attending three outreach mobile harm reduction services and the city centre needle exchange service. After consent was obtained, participants were asked to press all fingers of their dominant hand, defined as the hand they most commonly used to write with, into a tryptone soya agar plate. The dominant hand was chosen as this was the hand considered to be most involved in the preparation and administration of injections. This provided information on the contamination of IDUs hands. Participants were then randomised to one of the two cleansing methods. No instruction on how to cleanse the hands was given in order to gain insight into results from 'natural' practice. Participants then pressed their dominant hand fingers again onto another agar plate. Participants also completed a short questionnaire with the needle exchange worker, to describe their living arrangements and explore their views on hand cleansing.

Analysis of data

2.32 Plates were stored in a fridge until they were returned to the University where they were incubated at 37 degrees C for 48 hours. The number of colony forming units ( CFU), indicating microbiological contamination, were counted on each finger dab and an average for all visible dabs taken. Not all IDUs showed 5 finger dabs. 'Before' and 'after' plates were studied and compared. These and questionnaire data were analysed using Excel and SPSS. Data was compared using t-tests. The study was not designed to identify specific organisms.

Ethical approval

2.33 NHS Research Ethics Committee approval was sought as the needle exchange service received NHS funding. The study was given approval by Swindon Local Research Ethics Committee, ref: 04/Q2004/29.

Results

2.34 The target number of 50 participants completed the study. Twenty three from the city centre needle exchange and 27 from the mobile harm reduction service.

What is the extent of bacterial contamination on IDUs hands?

2.35 It should be noted that for some dabs the count was approximate as the close growth of the CFUs made the exact count impossible to determine. Table 1 shows the results prior to hand cleansing, grouped into bands. As can be seen, the majority of IDUs had between 20 and 50 CFUs per finger dab.

Table 1: Contamination of injecting drug users hands before cleansing, shown as band within which the average number of colony forming units per finger on the dominant hand fell.

Homeless vs. housed

2.36 Thirty four (68%) people said they lived in housing -either their own house or flat, with a parent or in shared accommodation with friends. Sixteen people (32%) termed themselves homeless, ten of these slept rough, for example in car parks, five slept in a hostel or shelter and one lived in a squat with other injectors. Homeless and housed IDUs showed no significant difference in the extent of the contamination on their dominant hand (p=0.104).

What is the impact of hand cleansing before preparing injections?

2.37 Figure 5 shows the 'Before' and 'After' results for the entire study group, not broken down according to hand cleansing method. Note that there are 48 'After' results because two plates could not be read due to 'spreaders'. This is where the finger must have been wet, resulting in extensive spread of the contamination across the plate. This gives a difficult to interpret, so unreliable result, so it was discounted. Hand cleansing significantly reduced microbial contamination (p=0.015).

Figure 5: Contamination before and after hand cleansing shown as the number of injecting drug users with CFU average count within the stated band.

Figure 6 shows one set of agar plates to illustrate the effect of hand cleansing. This has subsequently been used in an in-house BDP leaflet to highlight the importance of hand cleansing to IDUs.

Figure 6: 'Before' (top) and 'After' (bottom) plates for participant H3, illustrating the reduction in the number of colony forming units due to hand cleansing.

'BEFORE' illustrating heavy contamination (over50 CFU average per finger) 'AFTER' illustrating one colony forming unit per finger but with some 'spreader' effect.

Comparison between alcohol hand rub and soap and water

2.38 Comparison between alcohol hand rub and soap and water: which is better for IDUs?

Table 2 shows the impact of both types of hand cleansing, with the number of CFUs grouped into bands. Both methods of hand cleansing reduced contamination and when data was compared statistically, alcohol hand rub was not significantly better than soap and water (p=0.093).

Table 2: Comparison between average number of colony forming units before and after hand cleansing with the specified method, with counts in the stated bands.

How can hand cleansing be promoted to IDUs?

Participants' views of hand cleansing:

2.39 The questionnaire, which was administered by the needle exchange worker, asked participants whether they usually washed their hands before injection preparation. They were given a choice of 'YES -all the time or most of the time', 'YES -but only sometimes', 'NO -never or rarely' or 'Not Applicable, does not prepare own injections'. All participants prepared their own injections so the last option was not selected. Results are shown in figure 7:

Figure 7: Participants self-reported hand washing practices prior to injecting (n=50)

2.40 This shows that amongst the participants only 24% (n=12) said they always or mostly washed their hands before injecting. The majority either never or rarely washed their hands (42%, n=21), or reported doing this only sometimes (34%, n=17).

2.41 The questionnaire also asked 'What are the main reasons the client thinks prevent hand washing? A choice of the following options were given 'Lack of access to soap and water at the time', 'In too much of a hurry to get hit', 'No one else does it' or 'Client does not think there is a need to wash hands before injecting'. The option of ' other' and specifying a different reason was also given. Although the question was intended only for those who reported never or rarely hand washing, responses were recorded for 41 participants indicating that the needle exchange workers collecting the data asked this question to most. However since the question asked for a general opinion, data is considered meaningful from all who responded.

2.42 The most popular answer to this question was 'in too much of a hurry to get a hit', selected by 18 respondents (44%). Eleven participants (27%) said they thought there was no need to wash hands, highlighting a need for education around hand washing. Nine (22%) said ' lack of access to soap and water' was an issue and all of these nine were homeless. Two people gave other reasons: one said it depends on the circumstances around the injecting and the other said they had not thought about doing it. One person said 'no one else does it', which may mean it was not a learned part of the process or may mean they are deterred from doing it as no one else does.

Conclusions from the hand cleansing study

How extensively microbiologically contaminated are IDUs hands?

2.43 60% of IDUs produced average finger dab counts between 20 and 50 CFUs. There was no difference between homeless and housed IDUs contamination levels, which suggests that attempts to promote hand cleansing should be targeted at all IDUs not just homeless ones.

What barriers prevent IDUs from washing their hands prior to injecting?

2.44 Only 25% of participants reported cleansing their hands 'all or most of the time' prior to injecting. The main factors that prevented them from washing their hands were the hurry to get a 'hit' and the lack of perceived need. It should be noted that those who said lack of access to soap and water was the main barrier were all homeless.

Is alcohol hand rub as good as soap and water?

2.45 Hand cleansing by either soap and water or alcohol hand rub prior to injecting could be of benefit for IDUs, as both reduced the amount of contamination on their finger tips. This study showed alcohol hand rub reduced contamination on IDUs fingers in 23 out of 25 cases. The 70% alcohol hand rub gave a greater reduction overall in number of colony forming units than the soap and water, but this was not statistically different with the numbers in this study. It also did not on any occasion cause an increase in contamination. Hand rub may potentially be more convenient as it does not require access to a sink. This leads on to the question as to whether alcohol hand rub should be supplied to IDUs.

2.46 Alcohol hand rub within medical practice is promoted for use in addition to washing hands with soap and water. Hand rub use is endorsed by the US Centre for Disease Control (Dix, 2002) as its convenience has been demonstrated to promote compliance (Pittet, 2000). It is used in Scottish hospitals and in England, following a successful pilot study by the National Patient Safety Agency (see: www.npsa.nhs.uk/cleanyourhands). Alcohol hand rub has also been studied as an alternative to soap and water elsewhere and produced favourable results (Hernandes et al, 2004). Hand rub provides the convenience of a cleansing method that can be used without the need to access a sink or towel. It may be more likely to be used by IDUs as it is quicker to use, hence more possible when in a hurry to inject. However this would require supply in convenient containers, such as belt clipped small bottles or sachets and an accompanying education campaign to promote this practice. 27% of participants did not think hand cleansing was necessary. This belief was also identified in the field study reported later.

Aluminium cooker tests

2.47 The Stericup tests were designed to answer the question:

• Do Stericups add aluminium to the injection solutions and if so, are the levels safe?

Method

The injections that were prepared and tested

2.48 A series of injections were prepared, using the basic preparation method that was developed (figure 4). Both citric and ascorbic acids were tested, as both are supplied to IDUs. Various quantities of these were tested as it was thought that the acidity may make a difference to the results. 'Controls' prepared without acid or drug were tested too. Heat was used for approx 30 seconds. Controls with acid were also prepared in spoons instead of Stericups. The spoons were made of stainless steel, so did not contain aluminium. The individual injections tested are detailed in table 3 with the results.

The analysis technique

2.49 Aluminium levels were measured using a technique known as Atomic Absorption Spectrophotometry. This can detect levels of aluminium down to a concentration of 0.2 parts per million. A full description of the analysis method can be provided ( j.a.scott@bath.ac.uk).

Results

2.50 The results from the aluminium detection tests are shown in table 3:

Table 3: Aluminium detection test results.

What do the results show?

2.51 Water without acid or drug in the Stericup, did not yield any detectable aluminium levels. This was found both with and without the use of heat. When acid was used with water and heat in Stericups, aluminium was detected. This was found both with citric and ascorbic acids in various quantities. When injections were prepared using heroin, with citric or ascorbic acid and heat, aluminium was again detected. The concentration of aluminium detected when heroin was used was higher. All solutions with acid prepared on the stainless steel spoon did not yield any detectable aluminium levels. So some aluminium is deposited from the Stericup into the solution during preparation, although in all cases the quantity did not exceed 3.0ppm. Although not tested, other aluminium preparation cookers for IDUs, like Danicup, are likely to also exhibit this effect.

Why does this happen?

2.52 Ascorbic acid and citric acid are reported in the literature to increase aluminium absorption from the stomach (Fairweather-Tait et al, 1994). The authors suggest this may be due to the ascorbic and citric acids forming what is termed a 'soluble chemical complex' with the aluminium. This means the aluminium, which is not normally soluble, can dissolve in liquid because it reacts with the acid. The same mechanism may be occurring with the Stericup, since water and heat alone did not increase aluminium levels (i.e. water and heat did not dissolve any aluminium from the Stericup).

2.53 The injections that contained heroin had more aluminium in them. No control was performed with heroin on a spoon as drug supplies were limited. This means it cannot be established whether the higher level of aluminium is due to the heroin reacting with the acid and aluminium, or whether there is also aluminium in the street heroin as a contaminant. The former is considered most probable.

Are Stericups and other aluminium cookers safe for IDUs?

2.54 This leads to the question of whether the level of aluminium in injections contributed from the Stericup can be considered safe. Concerns in the literature around aluminium consumption relate to aluminium build up in bone, brain tissue and internal organs. However, much of the work is animal based (e.g. Fulton and Jeffery, 1990). Human studies are difficult as the effects take a long time to be seen, and it is impossible to control for all possible causes over a lifetime. Neither the British Pharmacopoeia, European Pharmacopoeia nor United States Pharmacopoeia specify a limit for aluminium in injectable medicinal preparations. This is likely to be because of the difficulty in establishing a safe limit. Aluminium is present in foods and drinking water sometimes in reasonably high quantities. Data from Wessex Water (Water Supply Zone Summary, Zone 59, ref: 44000059, 21/02/03) showed that analysis of 20 drinking water samples in Bath gave an average aluminium concentration of 21.6 micrograms per litre, which is equivalent to 0.02 micrograms per ml or 0.02 ppm. However aluminium is poorly absorbed from the gastrointestinal tract, therefore oral ingestion data does little to guide on safe limits for intravenous use.

2.55 Taking the highest experimental result which was 3 ppm (3 micrograms per ml), an injection of volume of 0.5ml would contain 1.5 micrograms. On this basis, an IDU who injected five times a day could be estimated to inject 7.5 micrograms of aluminium daily. Concern in the literature is largely around build up of aluminium in the body over time, so the real question is whether use of aluminium cookers by IDUs over several years could present any risks. This cannot be answered from this work. The potential advantages of single use sterile cookers in preventing sharing and BBV transmission need to be remembered when considering speculated longer term risks.

2.56 Future manufacturers of cookers may want to investigate other metals that do not react with acid or drug, but also do not withstand multiple use. One of the advantages of the Stericup is its fragile nature. It was found in the laboratory that the Stericup becomes flimsy when reheated, especially around the handle, which bent. This was also reported by some of the interviewees in the field study detailed later. This may prevent it being used by IDUs multiple times. Stronger metals and materials such as stainless steel are likely to be more durable, so potentially not be as 'single use' in nature. Their cost is likely to also make their supply prohibitive.

Conclusions from the aluminium cooker tests

Do Stericup add aluminium to the injection solutions and if so, are the levels safe?

2.57 Small amounts of aluminium were shown to be detectable in solutions made with acids in Stericups. The significance of this, if any, cannot be assessed at this point based on the limited literature on aluminium accumulation effects. Therefore it is unknown whether there is any cause for concern from aluminium deposits in injections. The Stericups were found to be 'single use' in nature. Reheating tended to make them bend at the handle and hence run the risk of spilling the contents. This was also noted by IDUs who were interviewed for the field study. The benefits of supplying single use cookers in potentially reducing the transmission of BBVs and in improve the cleanliness of injection preparation methods need to be remembered when considering supply.

The acid tests

The acid tests were designed to answer the questions:

a) Which sterile acid presents the least theoretical risk to veins?

b) What minimum quantities of acid dissolve heroin?

Method

2.58 Although a range of acidic substances have been noted to be used by IDUs e.g. lemon juice, vinegar, vitamin C tablets, it reasonable to propose that in order to reduce risks, the acid used should be in a 'pure' form, defined as of appropriate pharmaceutical grade. The British Pharmacopoeial ( BP) standard is recommended, as this means that the acid has passed all the necessary quality control tests stipulated, such as purity and limit tests for oxalic acid. Sterility is important as it can assure the needle exchange supplier that the acid they are distributing will not contribute towards microbiological risks in the injection preparation process. This laboratory work focused on citric acid BP grade sterile acid sachets and ascorbic acid (vitamin C) BP grade sterile acid sachets, as they present the less theoretical risks for IDUs.

2.59 The experiments performed investigated relative irritancy of prepared injections by measuring osmolality and pH. These terms are both explained below. The pH data can be used to predict the minimum amount of acid needed.

Osmolality measurements

2.60 Osmotic pressure (tonicity) is commonly described as osmolality. In the context of this study, this is the pressure that the injection may place on the blood and vein tissues when injected, due to a difference between the tonicity (amount of particles in the solution) of the injection and those in the blood/tissue cells. Hypotonic injections are ones were the tonicity of the solution is less than that of the blood/tissue cells. When this occurs, water from the injection may enter into the blood/tissue cells causing them to swell and possibly burst. Hypertonic injections are where the tonicity of the injection is greater than that of the blood/tissue and these cells may again attempt to redress this difference by leaking water, which may burst or dishevel them, possibly causing pain and reducing tissue function. Isotonic solutions are ones where the osmotic pressure of the injection is equal to the tissue or blood. Isotonic solutions cause no swelling or contraction of the tissues with which they come in contact. In the case of eye and nose drops isotonicity is essential to avoid pain on administration. However with injections, as the rapid moving blood quickly dilutes the injection, isotonicity is not essential.

2.61 Osmotic pressure is measured in units of milli-osmoles per litre (mOsmol/l). The osmotic pressure of plasma is 291 mOsmo/l. It is recommended that any fluid with an osmotic pressure above 550 mOsmo/l should not be injected rapidly as this would increase vein damage and ideally injections and infusions should be between 300-500 mOsmo/l (Florence and Attwood, 1998). Injections with a higher osmolality could be administered if this was done so very slowly, to allow adequate time for dilution in blood. Hence rate of administration is also an important factor in avoiding vein damage.

2.62 For this study, a series of heroin injections were prepared using the standard method (figure 4) and various quantities of either citric acid or ascorbic acid BP from sterile sachets were used, as described in table 4:

Table 4: samples prepared for osmolality measurement. *originally the equivalents of half a pinch were used but this was found not to be enough to dissolve all that was visible so these were increased to the stated amounts.

2.63 Injections were prepared in duplicate then combined as the minimum volume that the equipment could analyse was 1.0ml. Osmolality measurement was performed.

pH measurement

2.64 pH is an indicator of the concentration of hydrogen ions (charged molecules) (H+) in a solution. pH values are described as acidic (pH <7) or alkaline (pH >7) or neutral (pH = 7) on a scale of 0-14. Water has a pH of around 7. A high concentration of hydrogen ions gives a low pH (acid) and may be irritant to tissues and vein walls on injections. As said in 2.14, the acid dissolves the base heroin by a process called 'protonation'. The pH at which 99.9% of heroin molecules are 'protonated' is pH 4 (Florence & Attwood, 1998). In theory, adding additional acid will reduce the pH but have practically no effect on increasing the amount of dissolved heroin (i.e. be more irritant for no more drug gain). The pH of two of each injection described in table 4 was measured and the average calculated.

2.65 It was originally planned to also measure the amount of drug in resulting injections made with various quantities of acid to confirm concentration, but due to major equipment failure this was not possible.

Results

2.66 The pH and osmolality measurements are given in table 5.

Table 5: Average pH and osmolality measurement (n=2). Note the unit of measure mOsmo/Kg is approximately equivalent to mOsmo/l

2.67 The injections prepared with citric acid had a lower osmotic pressure, but also a lower pH. This indicates that although they may cause less cell bursting if injected rapidly, they may cause more vein lining irritation due to higher hydrogen ion content. The ascorbic acid injections had a higher osmolality, but higher pH, suggesting that if injected slowly to compensate for osmolality, they may be less irritant (Kuwahara et al. 1998). Citric acid is anecdotally reported by IDUs to cause more pain on injection than ascorbic acid, suggesting that irritation of the vein due to the hydrogen ion content is more significant than cell bursting, in causing pain. The results support the advice that small quantities of acid should be used and injections should be administered slowly.

2.68 The 330mg of ascorbic acid precipitated out of solution and would not freeze despite many attempts. This may possibly be due to the large quantity of ascorbic acid having an antifreeze type action or interaction with other materials present, causing precipitation. Dissolution occurred, then some minutes later the precipitate appeared.

2.69 When smaller quantities of citric and ascorbic acid (35mg and 70mg respectively) were added, not all the drug was seen to dissolve. Hence the minimum quantities used were increased to 50mg and 135mg, which were deemed approximately to be 'small pinches'. As said, in theory 99.99% of heroin base will dissolve at pH 4, at which the solubility is 120mg/ml (Florence and Attwood, 1998). However, this data relates to pure heroin base and the effects of other compounds present in street heroin cannot be accounted for. For example, other materials may 'use' some citric molecules (protons) to dissolve.

Conclusions from the acid tests

Which sterile acid presents the least theoretical risk to veins?

2.70 The results show advantages and disadvantages for both citric and ascorbic (vit C) acids and favour small quantities added stepwise. The results suggested ascorbic acid may be less irritant to vein lining due the injections being less acidic. However, they need to be administered slowly to avoid any osmotic effects. Citric acid produced theoretically more 'osmotically compatible' injections but it was shown that small amounts in excess of what is needed reduce pH a lot and hence it may be easier to cause burning, pain and irritation. Citric acid allows less 'margin for error' as smaller amounts compared to ascorbic acid produce similar pH changes. This concurs with anecdotal reports from IDUs.

What minimum quantities of acid dissolve heroin?

2.71 Less than half a sachet of either acid was sufficient for a £10 bag equivalent of the heroin used in these experiments. The results suggest that compared with the laboratory results, injectors may be using too large quantities of acid. This is based on data reported by Garden et al (2003) where whole sachets were common and the videos of Taylor et al (2004) where 8 out of 9 who used sachets to prepare £10 bags used one or more whole sachets. As well as using small quantities of acid, the results show the importance of administering injections slowly. As said, the diluting effects of blood are important to reduce irritation by the acid. Citric acid is cheaper than ascorbic acid, and for this reason services may be more able to fund citric acid supply.

2.72 Further work would be needed to measure the amount of resulting opiate in injections. Although all those prepared in the lab were observed visually to fully dissolve.

The filter tests

2.73 The filter tests were designed to answer the questions:

• How do commercially available filters (e.g. Sterifilt, Stericup filter and 'wheel' filters) compare against each other and with 'makeshift' filters used by IDUs (e.g. cigarette filters, hand rolling cigarette filters and cotton bud tips)
• To what extent do various filters retain drug and hence are likely to promote their retention for reuse or sharing?

Method

2.74 The first question was addressed by measuring particle content of prepared injections and comparing them when different filters were used. Injections were also studied under a powerful microscope technique called SEM to look for fibres from filters. The second question was answered by measuring the amount of drug retained in the filters after use.

Particle content analysis method

2.75 The particle content was measured using the British Pharmacopoeial standard method of particle size analysis for injections. The machine used was an LS-200 PMT particle sizer. It counts particles within specified size ranges, rather than reporting the exact size of each particle, which would be inaccurate. Each experiment was performed in triplicate and the average calculated. 'Controls' were performed. These were experiments that did not use drug to show the effects of filtering in itself. The particle content of various water sources was established. Then the amount of particles shed by the filters when used was also established. Finally injections prepared with heroin were prepared and filtered by different means and the results were compared. The SEM was used to examine and take pictures of prepared injections to look for shed fibres.

Analysis of used filters for retained drug

2.76 Capillary Zone Electrophoresis was used to measure opiate content, performed in conjunction with R Reid, The Robert Gordon University. The drug retained in the filters was 'bashed down' copying the methods of IDUs using 1.0ml of methanol, which was then diluted and analysed. The wheel filters were flushed with 1.0ml methanol, which was then diluted in the same way. Each one was tested twice and the average taken.

The filters that were used in the experiments

2.77 The filters analysed were cigarette filters (split into ¼), hand rolling cigarette filters (Rizla, split into ¼), cotton buds (removed from the plastic stalk and used whole), Sartoruis syringe ('wheel') filter (0.2 microns), Stericup filter and Sterifilt.

Microbiology of used filters

2.78 In the video work of Taylor et al, the practice of retaining filters for future use was observed. This could potentially present risks from infection if the filters harbour microbes. Demonstration of contamination of used filters could be used to convey harm reduction messages to IDUs to discourage reuse and sharing. An initial exploratory study would also be useful to indicate if further more detailed microbiological investigation is worthwhile. Participants in the hand washing study previously described (2.29-2.46) were invited to donate used filters for this study. Willing participants either donated filters they had on their person at the time of request or were issued with a sterile sample pot, marked with their study code, and asked to return it on their next visit to the needle exchange. As filters appear to have a value, as seen in the work of Taylor et al, it was anticipated that response may be quite low. LREC approval covered this aspect of the study. Returned filters were dropped using tweezers to avoid touching them, into 1ml of diluent nutrient broth and incubated for 48 hours. Where two or more filters were supplied by an IDU, two were chosen for analysis in duplicate. 100 microlitres of the resulting broth was plated onto a tryptone soya agar plate and growth counted after 24 hours as the number of colony forming units. The purpose was to conduct initial investigation into extent of contamination and speculate on likely species. No attempt was planned to identify the specific organisms. A microbiologist ( JC) assisted with classification.

Results and their significance

Particle content analysis & SEM: water only

Figure 8: Average number of particles per size range for each water sample (n=3)

2.79 The tap waters were relatively low in particles, with the kitchen cold tap giving slightly less particles than the lab cold tap. The sterile water BP in plastic ampoules performed favourably, giving a lower count than tap water. This would be expected as it is prepared to BP standard. The WFI was taken from glass ampoules and was more heavily contaminated. This may seem surprising at first, as it is prepared to the same BP standard the sterile water. However both waters were sampled straight from the packaging ampoule. The WFI ampoule was glass and it is likely that the increased particle load is due to glass particles that contaminate the water when the ampoule is snapped open. This is a known feature within the pharmaceutical industry. The tap water added to the Stericup was more contaminated than tap water alone. It is to be expected that some particles would be added from the atmosphere, the preparer and the Stericup. However the Stericup gave much less particle load than the spoon. This is likely to be because the spoon was cleaned prior to use and dried, whereas the new Stericup was removed directly from its packaging.

2.80 Particles detected were mainly less than 10 microns. The injection of small particles of this size is not without risk, for example they may cause granulomas. However, it is the injection of larger particles that may cause blood capillary blockage or the formation of clots that are of greatest concern. Overall the water control results support the use of the Stericup and support the use of sterile water BP, ideally taken from plastic not glass ampoules.

Filter controls

2.81 Figure 9 shows the results for control tests on the filters without drug. These were done to show the contribution to particle content from filters only. In all cases 0.7ml of sterile water from plastic ampoules was heated in a Stericup with 70mg citric acid, then filtered using one of the test filters.

Figure 9: Average number of particles per size range for each filter control (n=3)

2.82 The filter controls show that all filters add particles to the injection. The majority were below 10 microns. The Stericup filter added more large particles than the others, although overall few large particles were detected in any of the solutions. The Sterifilt added a lot of very small particles, but fewer larger ones. As particles would be generated in the manufacturing process of all filters (e.g. dust), the presence of particles was not a surprise.

Heroin injections

2.83 Figure 10 shows the results for heroin injections filtered by the stated means. The unfiltered injections ('no filter') contained the largest particles. All methods of filtration reduced the size distribution of particles towards the smaller end, to a greater or lesser extent. The makeshift filters performed less well than the commercially produced ones. The Sterifilt performed similarly to the wheel filter at the larger end of the particle size range. As may be expected, due to the difference in pore size, the wheel filter injections contained less of the very small particles.

Figure 10: Average number of particles per size range for each heroin injection filtered by the stated means (n=3).

2.84 At the small particle size range (up to 5 microns), the unfiltered heroin contained less particles than the filtered injections, with the exception of those filtered through the wheel filter. This shows that the filters or the act of filtration adds small particles to the injection. This was also shown by the filter controls, when compared to the sterile water ampoule water controls. However this does not mean that filtration should be discouraged, as when the larger particle data is studied the benefits of filtration, especially using the Sterifilt and wheel filter, are seen. It can be seen from fig 10 how the makeshift filters (cigarette, hand rolling and cotton bud) shed more particles especially in the 5 to 15 micron range. Since the smallest capillaries in the body are approximately 5 to 8 microns, these makeshift filters therefore have the potential to block these small vessels.

2.85 The Scanning Electron Microscope showed long thin fibres in the injections filtered with the cotton bud. Some of these images are given in figures 11 and 12.

Figure 11: SEM image of cotton bud fibre found in heroin injection

Figure 12: SEM image of cotton bud fibre found in heroin injection

Comparison with British Pharmacopoeial limits

2.86 The smallest capillaries in the body are approximately 5 to 8 microns, so particles smaller than this present less theoretical risk of capillary blockage but could however still form granulomas. The British Pharmacopoeia ( BP) (2004) states limits for the particle content of small volume injectables. These are defined as solutions of less than 100ml intended for injection. The BP limits require each container to have no more than 6000 particles equal to or greater than 10 microns. Of these, no more than 600 must be bigger than 25 microns. Table 6 summarises the filter heroin injection data according to BP limit sizes.

Table 6: Average total number of particles in injections up to 10 microns and up to 25 microns (n=3). *For definition of suggested limit see text.

2.87 Applying the BP limits finds all injections within limits except the unfiltered heroin. Clearly it would be preferable to promote use of the filters that gave the smallest number of particles, especially >25 microns (i.e.) the Sterifilt and the wheel filter. This is because the frequent and chronic use of street drugs by injection is in contrast to medicines, which are used by injection for the shortest, clinically appropriate time. Hence to minimise accumulative risks, minimal particle injection would be advocated.

2.88 Although the BP limit applies per container up to 100mls, clearly container volume can vary greatly. The BP limit allows small injections e.g. 0.5ml comparable to the size of an IDU injection to contain as many particles as a 100ml infusion. If the BP limit is taken to apply to 100ml containers only and a proportional calculation is made for 0.5ml volume (i.e. 1/20th) this would limit the IDU injection more strictly to 300 particles of 10 microns or greater with no more than 30 of these being greater than 25 microns. Referring to table 6, it can be seen that the Sterifilt would be within limits.

Drug retained in filters

The average amount of drug extracted from the filters is shown in figure 13.

Figure 13: Average amount of drug removed from used filters, shown in mg/ml of methanol which was used as the solvent for extraction.

2.89 The cotton bud and hand rolling filters had been stored for some time. As can be seen the heroin content had converted somewhat to 6-monoacetylmorphine (6- MAM) and morphine. The Sterifilt retained significantly less heroin (p<0.001), 6 MAM (p=0.001) and morphine (p=0.03). The incentive to retain makeshift filters for future used is illustrated by the opiates retained. The Stericup filter retained less than the makeshift filters, but as shown the qualitative interviews later, were still retained by some IDUs for future use. As the Sterifilt retained less drug it is less likely to be saved for reuse and future 'trading' or donation to IDUs. This was also reported in the qualitative interviews and was given as a reason by some for disliking the Sterifilt. The wheel filter retained much more drug because it retained a greater fluid volume. When the volume of water retained by filters was established, taking an average of 3 measures, the Sterifilt was found to retain 0.02ml. The wheel filter retained 0.3ml and the cigarette and hand rolling filters that both retained an average of 0.13ml. Therefore the wheel filter is unlikely to be acceptable to IDUs. It was found that the volume of retained injection can be reduced by pre-wetting the filters. However this would add another step and potential risk, if water was shared. The wheel filters are also expensive, being manufactured for hospital and laboratory use.

Microbiology of used filters

2.90 Table 7 gives the results of this work. JC judged the microbes to be largely a mixture of skin microflora, presumably transferred from the IDUs hands during preparation. No faecal contamination was evident. The filter itself may act as a filter for microbes in that the broth may not draw all the microbes off the filter. The results show that filters are contaminated with microbes, during reuse or sharing these could potentially cause infections. Dominant microbes can mask less dominant microbes in work such as this.

Table 7: microbiology of used filters donated by IDUs. * K4 and U2 only donated one filter.

2.91 The results support the need for a more detailed study to identify specific named organism (rather than type). Similarly, future work would attempt to collect detailed history of obtained filters, although reliability of such data would have to be considered carefully, as such information may not readily be known. Preliminary discussions with IDUs and drugs workers when designing this study suggested obtaining a detailed and reliable history on a specific filter could be difficult. It may be expected that those who willingly donated filters may have less of a tendency to store or reuse them as they were willing to donate them for no return. Only 12 of 50 IDUs donated filters for this work, despite a voucher incentive being offered.

2.92 The information from this study supports the harm reduction message that filters should not be stored for future use. It emphasises the need to address the culture of filter retention and to supply a filter that does not retain drug. The qualitative interviews showed there are several challenges in addressing such behaviour.

Conclusions from the filter tests

How do commercially available filters (e.g. Sterifilt, Stericup filter and 'wheel' filters) compare against each other and with 'makeshift' filters used by IDUs (e.g. cigarette filters, hand rolling cigarette filters and cotton bud tips)

2.93 The Sterifilt and wheel filter gave better performance in reducing the amount of particles in the heroin injections, compared to the makeshift filters. The Sterifilt passed the suggested stricter limit for 0.5ml injections that was based on the British Pharmacopoeial standard. The Sterifilt retained very little drug, giving evidence as to why it is unlikely to be retained for 'bashing down' (drug removal). Although the wheel filter performed well in the particle tests, it retained a lot of drug and is also expensive, which does not support its widespread use.

To what extent do various filters retain drug and hence are likely to promote their retention for reuse or sharing?

2.94 The Sterifilt retained significantly less opiates. Hence may be less likely to be stored for later use or shared. As discussed later, strategies should aim to discourage filter reuse, although this is difficult due to the perceived benefits of saving makeshift filters to save drug 'for a rainy day'.

2.95 Used filters were shown to be contaminated with a range of microbes, which may be capable of causing infections. This exploratory data supported the need for further work to perform a more detailed study to identify specific named organism.

« Previous | Contents | Next »