You are working in a small, rural hospital staffed by one physician and one nurse. There are multiple sick patients, all of whom require your attention, but the sickest is probably the 62 year old female with pneumonia and a blood pressure of 75/40 despite 3 liters of normal saline. This is septic shock, and you need to start a vasopressor, but the hospital protocol is that norepinephrine should only be given through a central line. You just haven’t had time to get one started, and wonder if it wouldn’t be better to get the patient started on the vasopressor using their peripheral IV.
The myth: It is an absolute contraindication to administer vasopressors through a peripheral intravenous line.
Although this is probably a myth, it is important to be clear: vasopressors can cause tissue damage after extravasation. Loubani et al (2015) performed a systematic review of adult patients given vasopressors peripherally, and looked for cases of tissue necrosis or limb ischemia after extravasation. They found 85 articles, almost all of which were case reports or case series. In total, there were 270 patients with 325 local tissue injuries or extravasation events. We cannot determine the incidence of these events from this data, because it is unclear how many total patients received vasopressors (the denominator).
There were 114 extravasation events from peripheral IVs, of which 75% did not result in any injury to the patient, and most injuries were minor. However, there were 3 patients who suffered major disability and 1 in whom the extravasation was thought to contribute to death. Most events occured in IVs placed distal to the antecubital fossa, and the average length of infusion before extravasation occured was 35 hours.
There were 204 local tissue injuries without obvious extravasation after peripheral vasopressors (179 skin necrosis, 5 tissue necrosis, and 20 gangrene). Only 9 of these resulted in major disability. In 4 cases the injury was thought to be a major contributor to death. Again, most of the events (85%) occured in IVs placed distal to the antecubital fossa and the average time of infusion before tissue injury was 56 hours.
What does this data tell us? First, we need to note that this data is imperfect. It is primarily case reports, there is missing data (time to tissue injury was only available for 80% of cases), and a lot of data comes from chart review. Clearly, tissue damage after peripheral vasopressor use occurs, although it is probably rare. (This data does not actually prove cause and effect, and these are sick patients who may have other reasons to develop tissue necrosis.) When tissue necrosis occurs, it most often happens in distal veins. Many have interpreted this to mean that it is safer to avoid distal veins if administering peripheral vasopressor, but I think the more likely explanation for this observation is simply that IVs are more commonly placed at distal sites. The rate of injury is probably proportional to the rate of placement at different locations. The time to extravasation and injury was generally quite long, although there were a handful of reported cases in the first 4 hours. However, the authors’ conclusion that in emergency situations short term use of peripheral vasopressors is unlikely to cause significant tissue injury seems reasonable.
The primary problem with this review is the lack of a denominator. This study provides us with no information on the actual incidence of injury after peripheral vasopressor administration, and therefore it is impossible to perform an appropriate risk benefit analysis. What we need are prospective studies.
There are a couple observational studies that can give us a better idea of the rate of extravasation and tissue injury with peripheral vasopressors. Cardenas-Garcia et al (2015) performed a prospective observational study of adult patients receiving vasoactive medications through a peripheral IV over a 20 month period at a single center ICU. They used peripheral vasopressors for days, but had a strict protocol that included selection of appropriate IVs that may not be reproducible in most emergency departments (see table) and a plan for extravasation. They included 734 patients, and only 19 (2%) experienced an extravasation event. None of those patients had any local tissue injury, but all were treated with phentolamine and nitroglycerin paste.
There is another observational study looking at 55 emergency department patients who received vasopressors through a peripheral IV. (Medlej 2017) There were 3 complications: 2 extravasations and 1 thrombophlebitis. None of these 3 patients had any treatment for the complication and all did fine.
So in observational studies, the rate of extravasation is between 2 and 6%, but there were no serious adverse events among the 789 total patients. Patients in the Cardenas-Garcia paper were treated after extravasation, but it isn’t clear if that treatment was necessary.
There is one RCT to help guide our management. Ricard et al (2013) randomized 263 patients at 3 French ICUs to either receive a central line or a peripheral IV as their initial venous access. The primary outcome was the rate of catheter-related complications, and at first glance it appears that the peripheral IVs were much worse (133 complications in 128 patients with peripheral IV versus 87 complications among 135 patients with central lines). However, a closer look at the data is a lot more reassuring. 56 of the “complications” in the PIV group were simply patients where they had difficult placing an IV. Those aren’t complications I care about. Some of the data is hard to pick out, because they report the intention to treat data. Therefore, there are 3 pneumothoraces in the peripheral IV group – not a common complication of that procedure. For the complication we are most interested in, there were 22 extravasation events (17%) in the peripheral IV group and 2 (1%) in the central line group. They don’t actually report tissue damage in the paper, but Scott Weingart contacted the authors, and there were no significant tissue injuries among the patients with extravasations. So clearly extravasations will occur, but the patient oriented outcome of tissue injury is pretty rare.
So, how should we interpret this data? Clearly, short term use of peripheral vasopressors will be appropriate in some patients, but we shouldn’t take the potential for injury lightly. Injury can occur, and occasionally that injury will be severe or even life threatening.
I think it makes sense to ensure that you have a high quality IV, although it isn’t entirely clear what that should mean. I would want vein that was easily cannulated on the first attempt and that flushes very easily. Some people argue that you should use the largest most proximal vein possible, but I am not sure that is an absolute requirement. Location probably does matter though, because it is one thing to need a skin graft on the forearm, but another thing altogether to lose function to your hand. We need to keep a close eye on the site, with standardized checks by the nursing team. Also, if the blood pressure drops, your first move should be to check the IV site, as the patient might have extravasated, and therefore is no longer receiving their required vasopressor. (Another tip: don’t place the blood pressure cuff on the same arm, because intermittently cutting off the patient’s vasopressor seems like a bad idea).
However, actual harms from peripheral vasopressors are actually quite rare, and to really make this decision we have to consider the harms of requiring a central line. Central lines have their own risks, including infections, DVTs, arterial punctures, and pneumothoraces. Furthermore, requiring a central line for vasopressors often results in a delay to necessary therapy. When considering those harms, I think it is clear that the administration of peripheral vasopressors is not only acceptable, but will occasionally be the best option for our patients.
Cardenas-Garcia J, Schaub KF, Belchikov YG, Narasimhan M, Koenig SJ, Mayo PH. Safety of peripheral intravenous administration of vasoactive medication. Journal of hospital medicine. 2015; 10(9):581-5. PMID: 26014852
Fitzcharles-Bowe C, Denkler K, Lalonde D. Finger injection with high-dose (1:1,000) epinephrine: Does it cause finger necrosis and should it be treated? Hand (New York, N.Y.). 2007; 2(1):5-11. PMID: 18780041
Loubani OM, Green RS. A systematic review of extravasation and local tissue injury from administration of vasopressors through peripheral intravenous catheters and central venous catheters. Journal of critical care. 2015; 30(3):653.e9-17. PMID: 25669592
Medlej K, Kazzi AA, El Hajj Chehade A. Complications from Administration of Vasopressors Through Peripheral Venous Catheters: An Observational Study. The Journal of emergency medicine. 2018; 54(1):47-53. PMID: 29110979
Ricard JD, Salomon L, Boyer A. Central or peripheral catheters for initial venous access of ICU patients: a randomized controlled trial. Critical care medicine. 2013; 41(9):2108-15. PMID: 23782969
Central line myth #2: location, location, location
While on the topic of central line, I might as well mention the other common myth: that one central line location is clearly superior to others. The myth will change a little from location to location, because everyone has their own preference. Subclavians are the best because they have the lowest infection rate. Subclavians are the worst because you will cause a pneumothorax. Femoral lines are dirty. Didn’t I hear that IJs actually have the highest complication rate?
While everyone has their opinions, we actually have data to guide us. There is a systematic review by Marik et al (2012) looking at the rate of infection that includes 2 RCTs and 8 observational cohorts covering a total of 17,376 lines. The rate of bloodstream infections was the same between femoral lines and subclavian lines. The observational data demonstrated increased infection at the femoral site as compared to the IJ site, but the RCT data showed no difference. Unfortunately, this data is imperfect, as the majority of studies are observational, and RCT data is primarily a single study looking at emergent dialysis catheters.
By far the best data we have to guide management is by Parienti et al (2015). They performed a large, multi-center randomized trial in which patients were randomized to get the central line in either the femoral, subclavian, or jugular site. A total of 3471 catheters were placed in 3027 adult ICU patients. A contra-indication to 1 site was allowed. For the primary outcome, a composite of adverse events, the subclavian line was statistically superior to both the femoral and jugular lines, but the overall rates of complications in all 3 lines was quite low. The rates of complication are summarized below:
But unfortunately, even this trial isn’t perfect. The clinicians were unblinded and were allowed to exclude one of the 3 sites if they thought it was inappropriate. The subclavian line was excluded more often than the other sites, potentially biasing the data. Furthermore, ultrasound was not mandated in this trial, but is considered by many to be standard of care when placing central lines, making generalization somewhat difficult. Also, although all supervisors had to have placed more than 50 central lines, the actual procedure could be done by a trainee. It is also worth noting that there was a high rate of failure (15%) with the subclavian line, so although it has the lowest overall complication rate, you have to consider the possibility that your patient will require another attempt at another site, with the added risk of complications.
Although this paper is occasionally interpreted as indicating that the subclavian line should be prefered, I think it is pretty clear that all 3 options have low complication rates and all 3 options are reasonable. The choice should be based on clinician judgement and an assessment of individual patients.
Marik PE, Flemmer M, Harrison W. The risk of catheter-related bloodstream infection with femoral venous catheters as compared to subclavian and internal jugular venous catheters: a systematic review of the literature and meta-analysis. Critical care medicine. 2012; 40(8):2479-85. PMID: 22809915
Parienti JJ, Mongardon N, Mégarbane B. Intravascular Complications of Central Venous Catheterization by Insertion Site. The New England journal of medicine. 2015; 373(13):1220-9. PMID: 26398070 [free full text]