Guest Skeptic: Dr. Simon Carley. Simon is Professor of Emergency Medicine in Manchester, England. He is a keen educator and loves sharing what he knows through the blog and podcast at St. Emlyn’s.
Case: You are working hard in the emergency department. It is a busy night and you are working flat out dealing with the stream of patients. In the middle of this busy night one of your nurses asks you to resite an IV on an 80-year-old lady with confusion and urosepsis. You had placed the IV yourself under ultrasound guidance earlier and it took some time and effort. The line has ‘fallen out’ and she needs re-siting before heading off to the ward. Your nursing colleague asks if you want to glue this one to stop it from getting pulled out?
Dr. Simon Carley
Background: Placement of an IV is arguably the most common invasive procedure in the emergency department. It is almost routine for our sicker patient and yet it is not without risk.
IVs can act as a source of infection, are restrictive to patient movement, require monitoring and can fall out requiring replacement. All of these factors take time away from other aspects of clinical care and put patients at risk.
It is also fair to say that every clinician will remember the IV that took them ages to site, only for it to fall out later.
Back in 2012, Cliff Reid on the Resus.me site reviewed two papers looking at the use of tissue glue to secure central venous catheters and the results looked good. I am also aware of the use of glue amongst our anaesthetists to secure arterial lines in young children. These are both examples of “high stakes” lines. Glue is used to secure the line as it will be difficult and potentially dangerous for them to be removed and it makes sense to me and others to use it for securing them.
The glue used is cyanoacrylate glue, or skin glue as we know it. For peripheral IVs the question is slightly different.
We still need to question whether we need to place as many IVs in the ED as we do. A paper reviewed on the St.Emlyn’s website back in 2013 showed that in an Australian ED fewer than 60% of peripheral IVs were used for anything more than taking blood. So clearly they are not as high stakes as our CVC and arterial lines.
However, some clearly are and in the case described we really do not want our IV to fall out and our patient may come to harm if they have delays in IV fluids or antibiotics.
The bottom line is that for many of our patients the securing of the IV is an important clinical intervention.
Clinical Question: Does the addition of skin glue decrease the failure rate of emergency department inserted peripheral intravenous catheters compared to standard peripheral intravenous catheter care?
Reference: Simon Bugden et al. Skin Glue Reduces the Failure Rate of Emergency Department- Inserted Peripheral Intravenous Catheters: A Randomized Controlled Trial. Ann Emerg Med 2015
Population: Adult emergency department patients requiring IV cannulation for therapeutic intervention.
Excluded: Allergy/irritation to skin glue or standard IV catheter securement material; presence of infection near the IV site, upper limb phlebitis, or venous thrombosis; high likelihood of intentional IV removal (ex: agitated patient); and non-English speaking patients without an interpreter.
Intervention: Fixation of the IV with one drop of glue at the peripheral IV skin site and one drop of glue under the peripheral IV catheter hub. There is a YouTube video demonstrating the technique.
Comparison: Fixation of the IV with standard dressing (details and figures are in the paper).
Primary Outcome: Peripheral IV failure at 48hrs
Defined failure as a composite of one or more:
Infection: Clinical impression of cellulitis or pus
Phlebitis: Two or more symptoms of pain, redness, swelling or palpable venous cord
Occlusion: Inability of flush 10ml of saline or history of IV being removed because “it was not working”
Dislodgement: Subcutaneous extravasation or history of “it fell out”
Secondary Outcomes: Modes of peripheral IV failure (infection, phlebitis, occlusion and dislodgement)
Author’s Conclusions: “This study supports the use of skin glue in addition to standard care to reduce peripheral intravenous catheter failure rates for adult emergency department patients admitted to the hospital.”
Quality Checklist for Randomized Clinical Trials:
The study population included or focused on those in the ED. Yes.
The patients were adequately randomized. Yes. Though they could have been more sophisticated about it.
The randomization process was concealed. Yes. They used an online system to do this so the people allocating the therapies would not have been able to interfere with the allocations.
The patients were analyzed in the groups to which they were randomized. Yes. This was an intention to treat study. No patients crossed over and so we can be satisfied with the analysis in this regard.
The study patients were recruited consecutively (i.e. no selection bias). No. Patients were screened by three trained ED research nurses 16hrs/day, seven days a week.
The patients in both groups were similar with respect to prognostic factors. Unsure. Probably but we can’t be sure.
All participants (patients, clinicians, outcome assessors) were unaware of group allocation. No.
All groups were treated equally except for the intervention. Yes. Apart from the intervention they were treated the same.
Follow-up was complete (i.e. at least 80% for both groups). Yes. They followed up 97.1% of the patients.
All patient-important outcomes were considered. Yes. I think so. I quite like the composite outcome in this paper. As the authors tell us, the patient is less bothered by why an IV fails; rather they are interested in the fact that it has failed. In addition the authors have broken the outcomes down for us too so we can get an idea about why.
The treatment effect was large enough and precise enough to be clinically significant. Yes/Unsure.It’s a large enough treatment effect, but I’m not so sure about the precision.
Key Results: A total of 369 patients were enrolled and had data to be analyzed (179 in the intervention group and 190 in the control group). The most common cannula was 20 gauge with the most common site being the antecubital fossa.
Failure rate at 48hrs: 17% (glue) vs. 27% (standard) NNT=10
Secondary Outcomes: Mode of peripheral IV failure was similar in both groups except for dislodgement. Peripheral IV dislodgement was 7% less frequent by 48hrs (95% CI -13% to 0)
We like this study. It is a simple question, a simple design and an important question so first up we want to say thanks to all the people who put a lot of hard work into getting us to where we are now and for giving us the opportunity to discuss it. Anyone and everyone who has ever conducted an RCT in an emergency department will know – it’s jolly hard work.
So on to the nerdy stuff. The bits that we need to discuss not to wreck a paper, but to understand how we can interpret the results in the paper and also for the applicability to our own practice.
External Validity: This is a paper from Australia, which has a similar EM practice to the UK and much of the developed world, including Canada. This means that the decisions about IVs and the reasons for their placement are likely to be similar to our practice.
Single Centre: This will always limit the generalizability of the findings. We do not know much about the patients in this study; they are likely to be similar to ours but we do not really know.
Most ED Patients: We need to think deeper about the patients. The patients in this study are those admitted for more than 48 hours. That’s a subset of my patients who get IVs and so I’m not sure that we can apply this to every patient who might need an IV at the door of the ED. The patients themselves looked similar in table 1 in terms of their characteristics but there are only 360 patients in that table yet they randomised 380. It is unclear where the missing 20 are.
Exclusion: They excluded agitated patients and yet suggested these patients may have benefited the most. However, agitated patients often have super human strength and I am not sure a little dab would do ya.
Randomization: Randomization process itself was good in terms of the process; they used a computer generated randomisation sequence but there is some work out there that tells us that certain sites and uses of IVs are likely predisposed to failure. For example some studies suggest that the use of antibiotics increases the failure rate. The authors could have stratified their randomisation for factors such as this though that may have led to the need for a larger study.
Consecutive Patients: This was not consecutive recruitment and that may bias the results. Admittedly they recruited 16 hours/day 7 days a week, which is better than most EM studies, but there are problems with this. Firstly, patients overnight (I am guessing that’s where those missing 8 hours were) are different. Secondly, the fact that it had to be when the research nurses were present may have influenced the cases. I can not see any data on the number of patients “screened” for inclusion in this trial and so it is possible that patients deemed “unsuitable” for whatever reason may have had different outcomes.
Blinding: As this is a therapeutic trial and an RCT a key question is blinding, or masking as we like to say in my house. Ideally, in a therapeutic trial everyone should be unaware of which group the patient has been assigned to until right at the end. Now clearly patients and staff could not be blinded here as it is obvious who gets the glue. However, they could have blinded those doing the data analysis, just giving them data but not telling them which group they were allocated. This is Nerd level – EXPERT but is increasingly seen in RCTs.
Follow-Up: Follow-up was fantastic at over 97%. In Virchester we wound struggle to find 97% of my patients 48 hours after seeing them in the ED! You could question the outcome measure of 48hrs as many admitted patients require an IV for more than two days. In addition, 209/369 (57%) of patients were discharged home before being reviewed by the research nurse. These patients did not have direct visualization of their IV but rather a telephone interview with a standardized questionnaire, chart review and discussion with ward staff.
Harm: Harm is often underreported in studies. They did not report any incidents of adverse skin events but patients did comment on a “pulling” feeling during removal. This seemed to occur when the glue was not wiped off properly or patients with very hairy arms.
Cost: It’s also worth mentioning that there is no assessment of cost here. Wound glue is expensive and the wipes to remove it similarly so. Having said that, the time, effort and equipment required to re-site IVs is also expensive and so there is a balance here. We simply don’t know from the data presented how that would pan out in different health economies. In the UK the ED would bear the increased cost, and the wards would reap the benefits with the patients stuck in the middle. Such insanities exist in the financial world of medicine and so this would need to be a carefully negotiated intervention if we were to take it forward.
Effect Size and Precision: Lastly we need to think about the size of this study and I think the final question on the checklist is telling. The effect size is great, really great. The failure rate in the glue group was 17% as compared to 27% in the standard care group. That is an absolute reduction of 10% and thus a number needed to treat of just 10. If this study is true then we would save one resiting of a cannula for every ten that we put in. That effect size is huge as compared to most of the interventions that we deliver in the ED. So it is a big effect, but sadly it is not a very precise one. If we look at the confidence intervals for the effect they are pretty broad, ranging from an absolute risk reduction of 18% (an NNT of just over 5) through to 2% (an NNT of 50). That range would have a real impact on whether this is a worthwhile, routine, ED intervention.
Comment on authors’ conclusion compared to SGEM Conclusion: We generally agree with the authors’ conclusions that skin glue reduces failure rate of peripheral IVs for adult ED patients admitted to the hospital.
SGEM Bottom Line: Skin glue does appear to decrease the failure rate of IVs in patients admitted to hospital from the ED at 48 hours. We do not know if this is a good idea for all ED patients and we do not know the true effect size, but for high stakes cannulas that we really want to stay in this intervention should be considered.
Case Resolution: On the basis of this trial, this patient would appear to be a good candidate for glue fixation of her cannula. You resite the line using a dab of skin glue to fix it in place. She leaves for the ward looking happier already.
Clinically Application: I think that the evidence here does not convince me to do this in every single patient, and of course if we are to start doing this we would need to consider resource and training implications. However, for some IVs, ones where I would consider them “high stakes” then yes I will consider and will probably use this technique in the ED.
What do I tell my patient? We know that about 1 in 5 IVs fail in the first 48 hours. It looks as though you will need one for at least a couple of days. Putting a little skin glue on the IV site can make it less likely to fail. The skin glue makes it a little trickier to remove but we think it is a good idea to use it in your case.
Keener Kontest: Last weeks’ winner was Amanda Johnston. Amanda knew caffeine can compete with adenosine, causing a higher likelihhod of faiure to convert SVT or a need for increased dosage (Acad Emerg Med 2010).
Listen to the podcast for this weeks’ question. If you think you know the answer then send an email to TheSGEM@gmail.com with keener in the subject line. The first correct answer will receive a cool skeptical prize.
Remember to be skeptical of anything you learn, even if you heard it on the Skeptics’ Guide to Emergency Medicine.