Central venous access device insertion

Like the thoracocentesis chapter, the purpose of this summary is to unite the information which is otherwise contained in numerous widely spread-out sources. The main audience would probably be the CICM supervisor of training who would need something akin to a reference manual in order to be able to assess the trainee who needs their WCA form filled out. Alternatively, the trainee doing the WCA might benefit from the bibliography at the end of this chapter (but probably not from the chapter itself). Past paper SAQs rarely touch on this, but Question 15 from the second paper of 2015 had asked about the potential complications of CVC insertion. 

What one might describe as "required reading" for this topic consists of the following resources, themselves concise literature reviews:

• NSW Health Policy Directive: Central Venous Access Device Insertion and Post Insertion Care (CEC, 2016)

• ANZICS Central Line Insertion and Maintenance Guideline 

These are 25 and 14 pages, respectively. Of the two documents, the ANZICS Guideline is probably better referenced, but the NSW Health policy is more detailed. Additional resources could include the paywalled treasure chest of UpToDate ("Overview of Central Venous Access"). A free (and excellent) resource which offers excellent detail about the actual technique of insertion is the Central Venous Access document by Rigby et al (undated),  from Queen's University in Canada. As far as peer reviewed literature goes, nobody describes the technique better than Bannon et al (2011). Good articles describing PICC insertion are more difficult to track down;  Dawson et al (2011) seems to be the most comprehensive.

What the hell is a "central" line

The ANZICS guideline defines a central line in terms of tip position; "the device must terminate in one of the great vessels ... or in or near the heart to qualify as a central line". 

Greater vessels, for the purpose of this definition, are also listed by ANZICS:

• pulmonary artery
• superior vena cava
• inferior vena cava
• brachiocephalic veins
• internal jugular veins
• subclavian veins
• external iliac veins
• common iliac veins
• femoral veins

Basically, the definition is based on rates of blood flow, which is a function of vessel caliber as determined by the Hagen-Poiseuille equation. Paul Marino's The ICU Book contains Table 2.1 (Ch. 2, 3rd edition, p.18) which lists the blood flow rates through various central veins, albeit without any reference as to where he got these numbers from (presumably, he measured it himself). This table's values are reproduced below:

Vein	Flow rate
Superior vena cava	1800-2000 ml/min
Inferior vena cava	1200-2000 ml/min
Femoral vein	700-1100 ml/min
Internal jugular	500-1400 ml/min
Subclavian vein	350-800 ml/min

Indications for central venous access 

There are only a few strong reasons for the insertion of a central venous access device:

• IV access
• Infusion of irritant substances
• CVP monitoring
• Advanced haemodynamic monitoring (PICCO, PA catheter)
• Central venous oxygenation monitoring

Extended indications also include:

• Inadequate peripheral access
• Extracorporeal therapies (ECMO, CRRT)
• IVC filter placement
• Venous stenting
• Transvenous pacing
• Catheter-guided thrombolysis
• Repeated blood sampling

 Contraindications to central venous access 

Generic contraindications to CVC insertion at any site include:

• Obstructed vein (eg. clot)
• Stenosis of the vein
• Raised ICP (IJ line)
• Severe coagulopathy
• Respiratory failure with high FiO₂
• Contaminated site
• Traumatised site (eg. clavicle fracture and subclavian line)
• Burned site
• Uncooperative awake patient
• No absolute contraindications

At various sites, other contraindications can be invented, but they would all fall into the categories of "some bone broken or deformed near the site" or "the site itself is infected". For example, a previously fractured clavicle or severe kyphosis with forward rotation of the shoulders us a relative contraindication to subclavian line insertion. The possibilities are too numerous to list for all possible sites.

Factors involved in site selection

The NSW Health policy lists a series of factors which should be "considered" before inserting a CVAD, without offering much guidance on how one should react to them. These factors included:

• Obesity
• Bleeding diathesis:
  • Platelets less than 50
  • INR over 1.5
  • APTT over 50
  • Antiplatelet drugs: clopidogrel and ticagrelor especially
• Hypotension
• Previous surgery at the proposed site of insertion
• Previous central line at the same site
• Infection at the site
• Presence of LBBB (relevant to PA catheters only)
• Lymph node dissection in the same area
• Previous DVT in the limb being drained by the central vein you're looking at.

The same policy also has some selection criteria for when to use one CVAD insertion site over another. To these, a few more suggestions have been added. The whole thing would benefit from a tabulated format where advantages and disadvantages of each site are listed.

Central Venous Access Sites

Site	Advantages	Disadvantages
Subclavian	Suitable for long term use (~14 days) Lowest rates of infection	Increased risk of subclavian stenosis if used for vascaths (which prevents a fistula from being formed on that arm) Impossible to control bleeding from arterial puncture or severe coagulopathy (non-compressible) Least suitable for the patient with severe lung pathology (risk of pneumothorax) Requires the patient to be supine and head down for insertion Insertion interferes with CPR Arm position can "pinch off" flow in a subclavian vascath
Internal Jugular	Convenient Easy to access during an operation Least acute complications Best site for a vascath Fairly unlikely to result in a haemothorax or pneumothorax No "pinch-off" phenomena   Compressible	Only short-term use is optimal May be better to save it for when a vascath, PA catheter or ECMO cannula are required May get in the way of a future tunneled central line Requires the patient to be supine and head down for insertion May cause increased ICP if the patient develops an IJ thrombosis Cannot be an option if the C-spine collar needs to remain in situ Not an ideal choice if tracheostomy is planned (will get in the way)
Femoral	Zero risk of pneumothorax or haemothorax Easily compressible site No need for supine/head down position; suitable for patients who are in respiratory distress No need for CXR confirmation - useable immediately	Risk of infection is greatest A poor choice for morbidly obese patients with extensive pannus Risk of retroperitoneal haematoma Unsuitable for CVP monitoring Circulation of drugs is comparatively delayed Impairs patient mobilisation
PICC	Low risk of serious complications Suitable for prolonged use, (up to 6 weeks in some cases) No need to position the patient supine	Higher risk of thrombosis More difficult to assure correct tip position Unsuitable for high-volume or highly viscous infusions (too much resistance to flow) Usually unsuitable for CVP monitoring or central venous blood sampling (high risk of lumen obstruction and poor waveform fidelity)

What's best? The college declares "a preference for the subclavian vein in most patients", quoting Chittick et al (2010) who suggested this site to be the least likely to get infected and cause a reportable CLABSI. This was on the basis of conflicting evidence; not all the studies agree that the choice of site matters and in fact Deshpande et al (2005) found no difference between sites, apparently because highly experienced operators were inserting the lines. Presumably their aseptic technique was more microbicidal and the number of passes through the skin was lower.

Right vs left sided lines

The decision to insert a line on the left or right depends on a few considerations:

• It is easier to insert a right subclavian (more likely to end up with the tip in the correct position)
• It is easier to insert a left PICC (less likely to go up into the neck)
• It is better to insert a vascath into a right IJ because the relatively straight course of the catheter will promote better flow and fewer access pressure problems
• It is better to save the right IJ for a vascath if the patient is likely to require ECMO or dialysis at some stage in the near future
• The right IJ vein is bigger than the left: Ishuzuka et al (2010) found a 3-4mm difference in their diameter, on average
• Femoral vessels are bilaterally symmetrical (i.e. it does not matter which vein you access in terms of venous access alone). However, the right femoral artery is the favoured site for arterial angiography access, and in general the femoral arteries are ideal for this; so if a coronary angiogram IABP or ECMO are planned, one may wish to leave the groins alone.
• In a patient with some sort of severe one-sided lung pathology the IJ or subclavian line should be placed on the affected side, so that you do not cause a pneumo/haemothorax of the "good" lung.
• Left-sided lines (IJ and subclavian) tend to have a higher chance of damaging the thoracic duct. Teichgraber et al (2003) report a case where the guidewire from a left subclavian insertion attempt accidentally penetrated the thoracic duct, perforating it and causing a chylothorax. You'd have to be extremely unlucky and the rarity of this complication approaches case report status, but it needs to be considered in situations where thoracic duct damage would be for some reason disastrous.

Central venous catheter selection

The general principles as outlined in the local policy:

1. The minimum number of lumens should be used
2. TPN (or any lipid-rich solution for that matter) requires a dedicated lumen all to itself
3. Antimicrobial (instead of antiseptic) catheters should be saved for those who need them most

Patients who should be considered for an expensive minocycline/rifampicin coated line are identified by the ANZICS guideline as follows:

• Immunosuppressed patients.
• Those expected to have long term need for CVC
• Those at a high risk of CLABSI (eg, burns patients)
• If CLABSI rates in the unit have remained high in spite of good insertion technique and attention to routine maintenance

All CVC packages should have some information regarding the number of lumens, priming volume and theoretical maximum flow rate:

Appropriate process of consent

An appropriate consent procedure would need to include the following important matters:

• Description of the procedure
• Risks and complications of insertion:
  • Pain during insertion
  • Haemothorax / pneumothorax, need for chest drain
  • Cardiac or greater vessel injury
  • Air embolism
  • Death
• Minor but common complications:
  • Need to reposition the device
  • Arrhythmia
• Possible later complications:
  • The line may get blocked
  • Wound or bloodstream infection may occur
  • Chronic pain from injured intercostal nerves

This information was collected from the QLD Department of Health website, where a comprehensive Central Vascular Access Device Insertion Consent Form is available.

Additionally, basic pre-conditions for consent must be demonstrated:

1. The patient must be legally capable of giving consent (competent)
2. Consent must be informed
3. Consent must be specific
4. Consent must be freely given
5. Consent must cover that which is actually done

In accepting consent from a patient, we operate under the presumption of competence: i.e. patient is over 18 and is “duly qualified: having sufficient, capacity, ability or authority” to give consent. This is tested by the following criteria:

• The patient can understand the relevant information
• The patient is able to appreciate the situation and its consequences
• The patient is able to manipulate this information in a rational fashion, and then make a reasoned internally consistent decision
• The patient is able to communicate this choice to you

The correct environment for CVC insertion

The NSW Health policy recommends a safe space which has specific characteristics.

You need to have access to:

• Adequate lighting
• Adequate space around the patient
• Some means of assuring aseptic technique (i.e. a trolley and drapes?)
• Immediate access to cardiac resuscitation equipment and drugs
• Skilled assistants
• "Electrical safety support" (i.e. you need to be in a cardiac-protected electrical area, where there are Residual Current Devices (RCDs), Line Isolation Monitors (LIMs) and equipotential earthing). 

These characteristics are shared by ED resuscitation areas, ICU rooms, operating theatres, as well as some anaesthetic bays, endoscopy suites, angiography and interventional radiology suites, and potentially other locations around the hospital. A reader has pointed out that whatever "electrical safety support" was supposed to be, it has disappeared from the NSW Health document and the ANZICS paperwork, suggesting that we either no longer care about these risks, or we are no longer equipped to discuss them.

Preparation and positioning of the patient

In summary:

• Flat and supine is best
• Trendelenburg position  is preferred for upper body lines

Basically, the patient should be supine for just about every CVC insertion, except PICC.  Femoral line insertion can be performed in a semirecumbent position, but also benefits from maximal flatness. In any case, the major guiding principle is to make the central vein in the region of interest the most dependent vein in the body, so that it dilates and makes puncture easier. In an intravascularly depleted patient, the veins will be collapsed unless it is possible to lay them head-down for CVC insertion. This is supposedly less likely in the subclavian vein, which is apparently "suspended" by soft tissue and is therefore less likely to collapse. 

The practice of putting people in a Trendelenburg position has a dual purpose, of helping prevent air embolus. If the spontaneously breathing patient takes a deep sudden breath during an IJ venous puncture, they could potentially suck enough air into their central veins to fill the right ventricle. Air flow though a 14G needle can be approximately 100ml per second, according to Teichgräber et al (2004) who presented a case of CVC-associated air embolism to NEJM. Only approximately 100ml is required to fill the RV outflow tract and kill you.

How far head-down do you tilt them? The correct answer is probably "enough to get the veins to dilate, but not so much that they fall off the bed". The ideal position is a compromise between respiratory function, comfort, intracranial pressure and the sinister upward creep of neck fat and massive breasts. Bazaral et al (1981) tilted patients by 14 and found that this increased the crossectional area of the IJ by about 50%, which they said was similar to what you can expect from a Valsalva manoeuvre.

In addition to these concerns, each site has its own anatomical peculiarities which can be exploited to dilate or bring to surface the central veins in relation to surrounding structures. 

Positioning for internal jugular (IJ) puncture

To make IJ cannulation easier, Bannon et al (2011) recommend a neutral head position, or only as much head rotation as is required to give access to the neck (i.e. don't have the patient trying to bite their own shoulder). Turning the head away from the puncture actually impedes the procedure because it puts the vein directly under the sternocleidomastoid (making an anterior approach especially difficult).  Sulek et al (1995) found CT evidence that a contralateral rotation beyond 40° also brings the carotid artery anteriorly, making it more likely to get in the way of the needle.

Positioning for subclavian puncture

Ipsilateral arm traction: Anecdotes and non-peer-reviewed sources recommend pulling on the ipsilateral arm to lower the shoulder and make it easier to get under the clavicle without having to angle the needle posteriorly. However, Raju et al (2016) found that this position actually increases the rate of catheter malposition.

Towel under the shoulders: A major decision in positioning for a subclavian line is whether or not to retract the shoulders (eg. by shoving a rolled up towel between the scapulae).  Bannon et al (2011) recommend a neutral position of the head and shoulders, quoting radiological evidence that in fact retraction of the shoulders tends to squish the vein between the first rib and clavicle, which will probably make cannulation more difficult.  Having said this, notable authors swear by this technique. Bova et al (1996) found a statistically significant improvement in the rates of successful (within three passes) subclavian cannulation with the shoulders retracted which suggests that the difficulty of  radiologically proven venous compression is overcome by some sort of other positive factors.

Which position is correct? Nobody can say. Bova et al used a single house officer to cannulate all of their patients, who may have had some considerable experience with one technique and not the other. Raju et al had a selection of experienced anaesthetists, and had so few complications in their small sample that it is difficult to discuss the superiority of one technique over the other.

How to manage the hairy patient

Hair at the insertion site will interfere with your line insertion. Hair will prevent proper asepsis, trap blood and clot chunks, and then act as a barrier to applying your adhesive dressing. This hair should be removed using clippers as opposed to a razor. Tanner et al (2010) performed a meta-analysis and discovered that shaving increases the risk of operative site infections (the risk approximately doubles).

Appropriate monitoring  for central line insertion

At minimum, you should have:

• An ECG monitor
• A pulse oximeter
• A pressure transducer

The NSW Health policy recommends all patients must have some ECG monitoring. Ones which have a reduced level of consciousness should have a blood pressure monitor and pulse oximetry at minimum. 

ECG monitoring is essential because most techniques will at some stage result in the guidewire entering the heart and causing some irritation of the endocardium. Usually, this manifests as ectopics - for example, if in the ventricle, then these will be ventricular ectopics. These arrhythmias are in fact usually consequence-free, but Stuart et al (1990) caution regarding "a distinct possibility of a malignant arrhythmiabeing precipitated by a guidewire".

In addition to the NSW health guideline recommendation, the author would like to add a pressure transducer, which - it goes without saying- belongs in the room when you are going to be inserting a central line. Even if you do not plan to actually use the CVP, the trace and pressure are important means of confirming that you are in the correct vessel.

Indications for ultrasound guidance in central line insertion

Local guidelines follow NICE and recommend that "ultrasound should always be used if available and the operator trained in the use" etc. This essentially says that if you are untrained in the use of ultrasound, it will not add any element of safety to your technique. Bannon et al (2011) add that "ultrasound does not obviate the need for anatomic knowledge" because you still need to know which way to safely aim your needle and which anatomical structures you are seeing on ultrasound.

Ultrasound guidance adds the following meaningful benefits to an already safe technique:

•  Eliminates the total reliance on "mind’s eye" visualization of deep structures
• Allows you to aim the needle directly at the deep arteries if they are underlying the vein
• Confirms that anatomy in the region is normal
• Confirms that the vessel is patent (i.e. there is no thrombus)
• Allows visualisation of the needle in the correct vessel (a means of confirming venous position)
• Allows visualisation of the guidewire in the vessel (another means of confirming venous position)

Sterile tray for central line insertion

An exhausting long list of suggested equipment is listed in the local guidelines. Ideally the equipment should consist of the following minimum:

• Sterile gown and gloves
• The CVAD
• Antiseptic solution for skin preparation (refer to Appendix 4 – Antiseptic solutions)
• 0.9% sodium chloride solution
• Sterile drapes
• Sterile procedure trays
• Gauze squares
• Local anaesthetic
• Needles and syringes
• Suture equipment
• Small sterile container to isolate used sharps
• Swabable capless valves or similar for each lumen
• Appropriate dressings

Here's one which was prepared with an almost ridiculous attention to neatness, as if for the purpose of some sort of equipment photoshoot.

The CVC packs may come with a Raulerson syringe. This thing is under patent held by Dr J.D. Raulerson (1989). In essence, the syringe has a hollow plunger which contains a valve preventing the aspiration of air into the patient. This hollow plunger can be used to introduce the guidewire without disconnecting the syringe from the needle. Though not everybody (anybody?) uses this technique, it could come in handy, for example in helping advance a guidewire by flushing ahead with saline (Pastewski et al, 2006). Most normal people, when faced with this proposition, would raise an eyebrow, as theoretically there is nothing stopping you from hydrodissecting a nice false passage for your guidewire using this method.

Aseptic scrub

The NSW Health policy recommends a two-minute hand scrub. All sorts of local hand-wash policies are available, but for those who want to read more the WHO probably supplies the definitive international document for the rationale behind surgical scrub technique (Widmer et al, 2010).

Sterile barrier attire for central line insertion

The operator should dress up surgically:

• Surgical mask
• Surgical hat (head and facial hair cover)
• Eye protection 
• Sterile gown
• Sterile gloves

The assistant should at least wear a surgical mask, a hat and eye protection.

Appropriate antiseptic for central line insertion

The NSW Health policy has an entire appendix dedicated to choice of antiseptic solution. Bottom line, they recommend some sort of alcohol-rich chlorhexidine wash.

Apparently they found some sort of advantage in ditching povidone-based solutions. The recommended agent ended up being 2% chlorhexidine in 70-80% alcohol (ANZICS would be happy with a mere 0.5%). The recommendation to wait until they dry is again made here, even though it seems to have more to do with the risk of a diathermy-associated alcohol fire.

Sterile barrier drapes for central line insertion

The NSW Health policy recommends "sterile drape/s fully covering the patient and their bed (unless this is impractical)". A lazy man might argue that it is always impractical to provide that much drape cover. 

Here are some drape options for different sites:

Local and systemic analgesia/anaesthesia

Basic things recommended by ANZICS and UpToDate:

• Usually lignocaine without adrenaline is enough - adrenaline is only useful if you are doing a tunnelled line and want to prevent hemorrhage into the tunnel.
• Infiltrate generously, but keeping in mind that it distorts anatomical landmarks
• Try not to inject any air into the subcutaneous tissues, as this tends to interfere with ultrasound imaging of the vessels.
• When performing a subclavian puncture, anaesthetise the clavicular periosteum if you can.

Preparation of the equipment for insertion

Generally, both experienced operators and inexperienced novices tend to benefit from some sort of order. This means lining the equipment up in a manner which represents the order of use. First the needle, guidewire, scalpel, dilator, CVC, flush, bungs and caps, then suture materials and instruments and finally the dressing. A trainee undergoing assessment who performs this ritualised "preparation of the altar" will score more highly by giving the impression of orderly calm.

On the other hand, the trainee whose equipment is strewn  haphazardly across their sterile field will give the impression of being disorganised, no matter how technically competent they are. Moreover, a random pile of equipment gives rise to rooting behaviour mid-procedure, which generates errors. While looking for the guidewire in your pig sty of a sterile trolley, you might lose the needle position and need to re-puncture, which would be some combination of unsafe and inelegant.

Anatomic landmarks at different puncture sites

The most detailed peer-reviewed (and freely available) publication regarding actual technique of central line insertion is probably this 2011 article by Bannon et al, the goal of which is made conspicuous by being published in Risk Management and Healthcare Policy.

Internal jugular vein landmarks for CVC insertion

The key landmarks are the borders of Sedillot's triangle, named after Charles Sédillot (1804–1883). Bannon et al recommend a somewhat lower approach than that which is commonly practiced by experienced ultrasound operators - 1/3rd of the way from the clavicle  to the mastoid. These days most people recommend a halfway point. The closer to the clavicle you get, the more the lateral safety distance separating the carotid and the jugular vein; but also the closer you get to the apex of the lung. Ergo, the high approach has become popular with the ultrasound-savvy crowd, where carotid puncture becomes less of an issue (i.e. you can see it).

Three possible approaches exist:

• Anterior approach: The skin is punctured at the medial border of the sternocleidomastoid at the level of the cricoid cartilage. The needle tip should pass medial to the sternocleidomastoid directed 30°–45° posteriorly from a coronal plane and 15°–45° laterally from a sagittal plane; i.e. you aim away from midline.
• Central approach: The apex of the triangle is punctured. The needle and syringe are angled 45° off a coronal plane, and the needle is advanced in a sagittal plane. 
• Posterior approach: The skin is punctured at the posterior border of the sternocleidomastoid muscle. The needle is advanced toward the ipsilateral sternoclavicular joint at an angle of 30°–40° off the skin. This approach reduces the chances of posterior wall puncture (i.e. "through and through" puncture of the internal jugular vein). 

Safety manoeuvres which are recommended include:

• Never angle the needle medially (or, you prang the carotid)
• Never puncture deeper than 2-3cm (or you hit the apex of lung)
• Stay "high" in the triangle (the lower the more anterior the carotid becomes).
• Use ultrasound (if competent)

Subclavian vein landmarks for CVC insertion

The infraclavicular approach is practiced by most sane people. It was first described by Dudrick et al (1968) in an effort to secure long-term access for TPN. The key landmarks are  the junction of the lateral two thirds and the medial third of the clavicle.

• The index finger of the non-dominant hand should be on the supraclavicular notch, and the thumb should be on the needle, pushing it down (parallel to the floor.)
• Puncture should occur about 1-2cm inferior and lateral to the junction (that usually is at the level of the deltoid tuberosity of the clavicle, if you can identify it)
• The needle should pass under the junction of the medial one-third and lateral two-thirds of the clavicle.
• The needle tip should be directed at the index finger (or just superior to its tip)

Safety manoeuvres include:

• Never angle the needle posteriorly (you will find artery or lung). The scalene muscle is about 1.5cm thick and provides some protection to the subclavian artery, but any properly dedicated person will eventually find it with a posteriorly-pointed needle.
• Never angle the needle inferiorly (you will mainly find lung)
• Puncture a reasonable distance from the clavicle (else you'll never get under it). Additionally, this limits how deep you can advance your needle. Which is good: the more medially you puncture, the more the subclavian vein becomes adirectly pposed to the pleura: i.e. a deep puncture is more likely to cause a pneumothorax here. If you are especially unlucky you may even cannulate the cuff of the endotracheal tube. 

Femoral vein landmarks for CVC insertion

Femoral vein CVCs canb be inserted using palpation alone, but ultrasound-guided methods are gaining popularity.  Tke key landmarks are the inguinal ligament and the midpoint of the femoral arterial pulse. This has significant relevance to the CICM Fellowship exam; Question 3 from the first paper of 2005 asked candidates to "outline the anatomical structures relevant to the insertion of a femoral venous catheter."

• The inguinal ligament runs from the pubic tubercle medially to the anterior superior iliac spine laterally.
• The femoral artery pulse is roughly at the midpoint of the inguinal ligament
• The femoral vein is medial to the femoral pulse
• The puncture should be approximately 1-1.5cm medially to the maximal femoral pulse, and approximately 1cm inferior to the inguinal ligament

Safety manoeuvres include:

• Ensure the puncture of the vein is below the inguinal ligament. In fact the skin puncture is usually below, but you've got to remember that the needle takes a long subcutaneous course, particularly in obese patients. Your actual venous puncture site might be deep and much more proximal than you think. Beyond the inguinal ligament, the femoral vein now becomes the external iliac vein, which courses deeper into the retroperitoneum where it is impossible to compress.

PICC line insertion landmarks

This is an area of greater diversity, and there are more possible insertion sites peripherally which will all end up with the tip in the central vessels. However varied, there are still some standards to follow. Dawson et al (2011) describes the ideal area for PICC puncture as the site where the best ultrasound image of the basilic vein can be seen, which is approximately midway up the arm (approximately 12cm medial to the medial epicondyle of the humerus). Any closer to the axilla and you enter an area with a dangerous amount of microbial life and moisture; any closer to the elbow and you expose the catheter to flexion and possible migration/dislodgement/fracture. Local guidelines for PICC insertion appear to concur with Dawson at least to some extent, as they recommend a site of puncture somewhere above the cubital fossa.

Safety manoeuvres:

• Always ensure you use the basilic vein: cephalic has a greater risk of thrombosis
• Always either use a post-precedure aseptic neck ultrasound to check for catheter malposition, or whatever alternative method you might have, eg. the ECG technique

Difficult central line insertion

A "pass" is described by the CEC as "each complete insertion of the needle that is intended to cannulate the central vein". Essentially, an otherwise suitably qualified but CVC-untrained person should only have a maximum of three of these passes. Why three? 

• Risk of insertion-related complications increases with the number of passes (Bova et al, 1996)
• Infection risk increases beyond three passes
• Multiple passes (beyond three) do not increase success rate (Bova et al, 1996) - i.e. if you suck at CVCs, perseverance will not rescue your poor technique 
• Three needle passes are somehow safe. Takeyama et al (2006) found that more than three passes resulted in a greater risk of complications from subclavian line insertion, but the first second and third passes did not differ in terms of complication rates (though admittedly all their operators were quite senior and the overall complication rate was fairly low).

Seldinger technique of catheter insertion

Sven Ivar Seldinger published on this in 1953, bringing order into the previously lawless Mad Max-like wasteland of vascular catheter insertion (they were just shoving totally random tubing into people, like animals). Seldinger used a "flexible round-ended metal leader" to safely exchange an artery-puncturing needle for a flexible polyethylene catheter. "This technique is simpler than appears on paper and after a little practice should present no difficulties", he wrote. It is now used to introduce just about any sort of hollow object into just about any other hollow object.

In honor of the fact that Seldinger appears to have done his own illustrations for publication, they are reproduced here with all respect (and no permission whatsoever).

To assist insertion, the lumen of a CVC needle hub is tapered (which allows easier insertion of the guidewire) whereas many other needles will have a blunt cylindrical crossection of their hub, all the way to the needle.

The J-tipped Seldinger guidewire

A promotional propaganda article from Diagnostic and Interventional Cardiology (Fornell, 2011) discusses angiography guidewire technology in some detail. Perhaps more detail than is expected from CICM trainees in the middle of their program (and let's face it, cardiologists require much more performance from their guidewires than we do).  The much more relevant 2015 article by Wolfram Schummer (2015) describes the mechanical properties of Seldinger guidewires we typically use for CVC insertion. Judging by the author's Figure 1, these properties were generally disappointing.

Basically, the following characteristics are expected of an ideal Seldinger guidewire:

• Flexible, conforming easily to the vessel
• Strong, resistant to lacerations
• Kink-resistant
• Soft-tipped, to prevent damage to vascular structures
• Sufficiently stiff to allow manipulation of its direction
• Approximately twice the length of the device it is intended to guide.

These characteristics are met by the usual CVC guidewire. Flexibility is conveyed by the spiral coil construction of the outer wire; stiffness is conveyed by the core wire which does not go all the way to the tip. The tip is J-shaped and composed entirely of the coil, which makes it flexible and atraumatic. The distal end of the guidewire contains the core wire and is stiff all the way to its tip; this makes it easy to threat the line over it. It also makes it easy to perforate the vessel wall if you reverse the guidewire and shove the non-J-tipped stiff end into the needle. This is the wrong way to do it. 

In summary, important things about handling the guidewire are:

• The J-tip can be straightened by putting traction on the wire distal to it (see this Weingart video, which is otherwise full of excellent tips)
• Rack the wire: As the dilator is advanced over the guidewire, the wire should be moved back and forth a little bit. This doesn't so much prevent kinking as alerts you to the fact that it has kinked. If it has kinked, it may be possible to withdraw the kink back into the dilator, and then proceed. One may be extremely unlucky and end up with a frayed dilator tip, which a) will not dilate anything, and b) will tear a hole in the vessel which is going to be much more difficult to fix.

The importance of not severing the guidewire

Most health services have some sort of dedicated policy which specifically demands that the whole guidewire be visualised after insertion. This is because of problems which happened. Bad things clearly followed. Patients have re-presented to the ED weeks following central line insertion with J-wires eroding out of their groins.

So: important not to do that.

Apart from carelessly losing control of the guidewire, it is also possible to sever it:

• When doing the initial "stab" prior to dilating
• When re-inserting the needle over the guidewire (never do this!)
• When the guidewire does not pass, and you keep trying to advance it in a variety of directions (NSW Health policy recommends you remove the guidewire and the needle together so as not to shear the wire).

Non-idiotic dilation of the CVC tract

Usually, the CVAD kit comes with a plastic taper-tipped dilator. 

A dilator is  required to create a tract which the (usually, soft) CVC cannot create by itself. This is supposed to be a subcutaneous tract. The dilator  should never enter more than a few millimetres into the actual vein. Thus, it is important to remember that is is not necessary to shove it into the patient to the hilt. For instance, when inserting a right subclavian or left IJ line, the dilator is often long enough and stiff enough to reach the opposite wall of the greater vessels, creating a tear in a deep mediastinal structure. Gupta et al (2011) present a case report of a (fortunately non-lifethreatening) haematoma from an overvigorous deep dilation. 

Methods to confirm venous placement

There are several simple bedside methods which you can use to determine whether you are in a vein or artery.

• Create a simple manometer: attach minimum volume extension tubing to the CVC, stretch the tubing up and allow the blood level in the tubing to rise. Given the convertion of mercury manometry to blood, at a systolic BP of 120 mm Hg your "blood manometer" would read 1.63m, i.e. you'd run out of tubing. This is why we used mercury in those things.
• Attach the transducer to the needle. It's sterile and already attached - may as well use it before dilating the vessel
• Visual ultrasound confirmation - you'd have to actually visualise the needle in the central vein.
• Blood gas analysis - though you'd be surprised how often arterial oxygenation is so disappointing that it looks venous.

The importance of flushing and capping all lumens

Needless to say, if you don't flush them they will clot.

All lumens should be capped, because:

• otherwise the patient will bleed to death,
• OR a massive air embolism will develop

The NSW Health policy recommends you use "swabable capless valves" to cap your lumens. These are colloquially known as "bungs". The manufacturer reports that they can withstand a backpressure in excess of 45 psi, which the patient is unlikely to ever generate using their thorax alone (as that would come to approximately 2327 mmHg, or approximately 31,600 cm H₂O). As the CEC implies, they can be swabbed, and this favours a relatively clean handling technique. The instruction is to swab them before and after they are accessed.

Interestingly, the volume of a syringle matters. It is significantly more likely for you to fracture the CVC when flushing a blocked lumen with a small syringe, even though this seems counterintuitive. Hayward et al (2011) were able to demonstrate this in the context of injecting various solid lesions in need of hydrodissection.

Appropriate fixation and dressing

The main issue to consider is that the CVAD should not be secured by tape alone. Suture or a sutureless "catheter clamp" should be used to secure the device. Re-positioning the device will require you to patientl undo all the fixation and reapply it all again after you have finished, in a way that prevents malposition.

An ideal dressing has the following characteristics:

• Sterile
• Transparent (allows surveillance of the site)
• Strong enough to stabilise the catheter
• Positioned with the insertion site at its centre
• Covers the CVC from insertion site to the hub

If two catheters are close together (eg. where an arterial line and central line have both been placed into the same groin) you may use the same dressing.

As far as deciding between sutures and "catheter clamps", there is no specific evidence to guide us. Frey et al (2008) were able to demonstrate that sutureless devices are probably better - at least for PICCs. ANZICS remain unconvinced whether CVCs would benefit.

Both NSW health policy and ANZICS recommend that the CVAD dressings should be changed every 7 days unless soiled, wet or loose. If there is a gauze piece under the dressing soaking up blood from the oozing insertion site, one needs to change it every 2 days.

Appropriate depth and radiographic confirmation of placement

Having inserted the central line, one should perhaps develop interest in where the tip is positioned. There are two competing needs here. For optimal dilution of infused substances the tip should be closest to the right ventricle; but for optimal safety it should be as far away from it as possible.  The NSW Health policy suggests that it is "reasonable" to expect your tip:

• above the cephalic limit of the pericardial reflection, which is
• at a level corresponding to the carina on a chest radiograph

The "pericardial reflection" mentioned here presumably refers to the serous pericardium, as opposed to the fibrous pericardium which is seamlessly continuous with the fibrous coats of the greater vessels. A line tip positioned below this reflection is generally viewed as dangerous because a perforation of the wall at this level will give rise to cardiac tamponade and death. Whereas a perforation of the wall above the reflection will merely give rise to a haemomediastinum, which is somehow viewed as a more benign complication.

Where is this pericardial reflection, and how can we determine its position? The gold standard is autopsy. Albrech et al (2004) performed many of these to delineate the limits of the pericardium, and found that it generally ended about 0.8cm below the carina. Unfortunately this technique is poorly suited for routine clinical use. Thankfully Stonelake & Bodenham (2006) were able to validate the carina as a convenient radiological landmark for the non-cadaveric setting. 

Radiological confirmation of insertion depth

Radiologically, the most useful landmark is the carina. Bannon et al (2011) mentions that the carina is (almost) always about 1.5cm inferior to the origin of the superior vena cava, and the right main bronchus is almost always about 2-3cm superior to the cavoatrial junction. Stonelake and Bodenham suggested that though some of the lower SVC lies within the border of the pericardial reflection, this "may represent a necessary compromise for left-sided CVCs to ensure they lie parallel to the vessel wall". Right sided lines are already reasonably straight and their tips should be in the upper SVC around the junction of the innominate veins, where blood flow is still excellent but where there is zero risk of cardiac tamponade. The innominate veins themselves are a gray zone. Most people would agree that the blood flow there is satisfactory to avoid thrombosis, but nobody would want to infuse anything particularly noxious into those vessels, nor for a prolonged period of time

Ergo, you should aim for the carina. If your catheter ends up within 2cm or so of the carina, it is unlikely to be in a disastrous place.

The importance of being parallel with the SVC

Most manuals and policy documents report that the tip of the catheter should not abut the SVC wall, and that the catheter should be parallel to the midline of the SVC wherever possible. This is because of the risk of erosion though the SVC wall, especially if horrible substances like TPN are being given through it. Some authors even go as far as to give a precise "angle of safety" to aim for.

That angle is 40°. This figure arose from the study by Gravenstein et al (1991), who eroded some central line tips  through a "vessel wall" simulated from polyethylene film of 12.7μm thickness, pulsated against the line tip at 80bpm. This might seem like a highly unrealistic model but in fact the very next year Duntley et al (1992) were able to  scrape up enough literature for a case series of 61 patients with catheter tip erosions, in whom the majority had catheter tips abutting the SVC wall at an angle greater than 45°. 

This is of course not a super-urgent problem, as the line is not going to erode instantly- you have some time to reposition it. However, while in this position the distal lumen cannot have anything corrosive hypertonic vasoactive or high-volume going though it, nor can the CVP be measured - which means the central line is unusuable for most of the conventional indications of central line insertion.

This is usually a problem associated with left-sided lines. As the line exits the left innominate vein, it has the opportunity to poke the SVC wall before continuing downward towards the lower SVC. Unless it is not long enough, in which it will never continue downwards, remaining up against the SVC wall, poking it vigorously each time the patient coughs. The consequences are predictably hideous. One may find oneself draining a massive pleural effusion composed almost entirely of propofol, for example. Duntley et al report a mortality rate of 12% (and that's just the ones they reported). 

Height-based estimation of insertion depth

It is possible to estimate the depth of insertion by the use of the Peres equations. Peres et al (1990) prospectively explored the relationship between patient height and catheter position, arriving at a series of empirically derived height-based formulae. Essentially, they all rely on dividing the patients height (in cm) by 10, and then applying some sort of modifier (for instance, for left external jugular catheters  the expected depth = height/10 + 4 cm). Thus, for a right IJ line the required depth (using the central approach) ends up approximately 16cm in the average 160cm adult female and perhaps 17-18cm in the tall male.  

Now, these equations have undergone extensive modification since 1990. For instance, a variant can be seen in the excellent resource by Phillipe Le Fevre which modifies the formulate with (-2cm) for right SCV, (-1cm) for right subclavian and (+2) for left-sided lines. This follows the results of a study by Czepizak et al (1995), who tested the predictive qualities of Peres' equations and found that in 95% of cases the equations accurately predicted the required line depth provided these minor modifications were made. Here they are, illustrated on the surface of this early Flavian marble.

The problems with developing and using these empirical equations are manifold, and mainly come down to three main facts:

• Technique varies (eg. central IJ approach, high cricoid etc; lateral or medial SCV punctures, and so on).
• Patients vary (eg. somebody with an unusually fat neck will have a central line with an unusually long subcutanous course).
• Goalposts vary (i.e. not everybody is in agreement as to what the ideal position actually is). For instance, Peres et al used the cavoatrial junction as their target, which not everybody would agree with (too close to the pericardial reflection, they might whinge).

There are other limitations to equation-based depth estimates. For instance, how do they perform in people with severe scoliosis? What is the guideline for their use in people who have unusually long necks? What of patients with acromegaly, pituitary gigantism, achondroplastic dwarves? Not to speak of the paediatric population. In summary, the equations are occasionally useful in the normal Homo sapiens group, insofar as their anatomy remains "classical", and are a useful way of saving yourself some time repositioning and re-suturing the line, not to mention the patient being disadvantaged by additional needles and a second Xray. 

Estimating the depth of CVC insertion using surface anatomy

It is also possible to measure along the surface of your patient to decide how deep the line should go. For example, the surface landmark corresponding to the carina is supposed to be the manubriosternal joint, otherwise known as the angle of Louis. One should therefore aim to have one's tip somewhere around there. For example, Kim et al (2011) used the junction of the second costal cartilage with the manubrium and sternum as the surface point to aim for.

They draped the patient and then laid the line out on the draped skin, with the tip at this surface landmark. The line was then inserted to the depth predicted by this method. The authors reported that the tip was within 2cm of the carina radiologically in 95% of the inserted lines. Vinay and Tejesh (2016) found this method to be superior in accuracy to height-based formulae.

Immediate complications and their management

Again something which might benefit from a big table:

Immediate Complications of CVAD insertion

Complication	Management strategy
Bleeding	Risk of this is quite low (van de Weerdt et al, 2017) Direct pressure is the first recourse Gauze or something prothrombotic like Kaltostat calcium alginate can be introduced under the transparent to act as a haemostatic agent Adrenaline-soaked gauze can be used, but may have little effect Administration of blood products or Vitamin K may be required
Arterial puncture (16G needle)	Remove the needle Local pressure for 5 minutes Post-procedure chest Xray if this was a thoracic line to check for haemothorax or pneumothorax  Beyond haemothorax, the other major risk is dissection of the artery  If this was a carotid arterial puncture, one needs to be performing neuro obs regularly. A carotid haematoma may also compromise the airway
Arterial dilation	Leave the catheter or dilator in situ Consult vascular surgery See the chapter on accidental arterial insertion of a large-bore venous vascular device
Guidewire lost	Interventional radiology or vascular surgery will need to go fish it out
Pneumothorax / haemothorax	Thoracocentesis will likely be required
Endotracheal cuff puncture	You're going to need to reintubate the patient
Nerve injury	Usually, by pressure from local haematoma after arterial puncture If clinically significant (eg. paraesthesia develops) a referral to vascular surgery may need to be made

Important matters surrounding the documentation of insertion

Essential elements:

• Aseptic technique
• Sedation given and total volume local anaesthetic instilled 
• Depth of insertion and any complications
• Type of CVAD inserted including serial number and bar code
• Guidewire independently observed to be removed intact

Correct technique of accessing a central line

The ANZICS Guideline has a whole section on routine central line maintenance which covers several important aspects:

• Accessing the CVC needs to be done in a manner which minimises the risk of infection. Whenever handling the line connectors or bungs, gauze soaked in chlorhexidine should be used. The specific goal is to reduce the colonisation of surfaces which might interface with the patient's bloodstream or the substance which is being infused into it. 
• Unused lumens should be flushed with saline 4th-hourly
• Change of administration sets needs to happen every 4-7 days, unless the solution contains lipid: in which case, it is every 24 hours. This mainly refers to propofol bottles and TPN.
• Mechanical valve connectors need to be watched carefully: Jarvis et al (2009) found that these were associated with an increased rate of CLABSI. For the record, the majority of connector "bungs" used in Australia are "split-septum", i.e. where the inserted line parts a pair of rubbery flaps (instead of displacing a spring-loaded piston, or something similar).  
• Blood sampling from central lines needs to be done in a sterile manner using full aseptic technique, 

Routine central venous catheter surveillance

Daily review of the CVAD insertion site involves looking for:

• erythema
• drainage
• tenderness
• pain
• redness
• swelling
• integrity of suture
• dressing integrity
• catheter position
• ongoing need for line

In summary, every day you need to answer the questions, "is it infected" and "does the patient still need it?"

For PICC lines, there is also the need to think seriously about thrombosis. The NSW Health policy recommends that we "compare the mid-upper limb circumference" with previous measurements whenever a concern regarding swelling gets raised.

Indications for CVC removal

At what stage should the CVC be removed? There are several possible reasons you might want to decannulate somebody. 

• Pointless: the indications for the line have resolved and there is no further reason for it to remain in situ.
• Infected: either a CLABSI is suspected or the site of insertion looks erythematous.
• Blocked lumen: ANZICS recommend that you remove a central line within 24 hours if the lumen has become blocked. 
• Inconvenient: the presence of the line in its current location is counterproductive, eg. the patient has an IJ CVC and needs a tracheostomy, or they need to mobilise by their femoral line makes it difficult. This is an indication for removal of the line; whether you insert another one elsewhere is a different issue.
• Old: line age tends to be related to infection risk (Moro et al, 1994)
  • Antiseptic-coated lines: 
    • 7 days for femoral lines
    • 7-10 days for IJ lines
    • 10-14 days for subclavian lines
  • Antibiotic coated lines: up to 21 days

Correct technique of CVC removal

The NSW Health policy recommends a prescriptive technique for CVAD removal. In summary:

• Aseptic technique
• Position the patient supine with head down
• The patient is to remain supine for 30-60minutes
• Airtight dressing on the site for 24 hours
• If a CLABSI is suspected, you send the tip for culture- but not otherwise (i.e. CVC tips should not be routinely cultured)
• Removal and especially the removal of an intact tip should be documented. 

Late complications of central line insertion

These can be grouped into the categories of "device dysfunction" and "vessel dysfunction"

Device dysfunction 

• Catheter fracture
• Catheter tip displacement
• Lumen blockage
• fibrin sheath formation

Patient-related complications

• Infection : on average 2.5 infections/ 1000 catheter days
• vascular erosion
• vessel stenosis
• thrombosis
• osteomyelitis of clavicle (subclavian access)
• Nerve injury
• Extravasation

Role and complications of "rewiring" or guidewire exchange

"Rewiring" refers to threading a guidewire though an old line, removing it (leaving the guidewire in situ)and then introducing a new central line over the guidewire. Generally speaking, this is frowned upon. ANZICS recommend that this desperate act be considered only in the following circumstances:

• The risks of using another site outweigh the risk of infection using the same site,
  OR
• The CVC has only been in situ for less than 72 hours and there is neither suspicion of CLABSI nor concern that the line was inserted in a non-sterile manner

The increased risk of infection is surprisingly low. Cook et al (1997) - the study used by ANZICS to justify their position - looked at 19 RCTs totaling 918 patients and 1913 catheters, and found that the relative risk of CLABSI is around 1.76, as compared to a new puncture. Having said this, the removed catheter tip should be cultured, and if it was colonised the new required catheter should also be removed.

A variation of this dodgy technique is the recycling of a PA catheter introducer sheath, where the removed PA catheter is replaced with a central line. The NSW Health policy doesn't even mention this, as it is an esoteric thing for most environments. Again, as in the case of guidewire exchange, this should only be done if the sheath has been in situ for less than 72 hours.

Audit process for safe and responsible central line practice

Accreditation of practitioners  

Locally, the CEC has published a document which outlines a training framework for CVC insertion. Their recommendations can be summarised in the form of a diagram (stolen from their own document):

In brief:

• There are minimum theoretical knowledge requirements, which should undergo some sort of theory assessment
• There are a minimum of practice insertions which are supervised, of which there should be three at each site
• There should be a minimum of five independent insertions, which do not need to be observed but which should have a supervisor available nearby.
•  If the practitioner can demonstrate "relevant prior experience",  one observed insertion and one supervised insertion at each site is satisfactory 

The CEC document linked above has a helpful logbook which can be printed and used as a record.

Audit of complications

Bundles of care should be used to manage quality assurance related to central line insertion. (Blot et al, 2014) found that these tend to create a significant improvement in the rates of complications. 

Infection rates are recorded in the ANZICS CORE CLABSI Registry. This is a central register of all inserted lines, and CLABSIes are reported to this central body for the purpose of data collection and audit. The CLABSI forms tend to also collect some peripheral data, such as the tip osition and any other complications. More ANZICS CLABSI stuff can be found on their main page for this.