Case

Several decades ago, a 2 6/12-year-old girl was transferred to the children’s hospital because of severe dehydration and shock following gastroenteritis. Numerous attempts to obtain venous access were unsuccessful, and the anaesthesia team was called for help. The anaesthetist inserted an intraosseous Cook cannula into the right proximal tibia. Fluid resuscitation was performed, and some hours later a peripheral venous cannula was successfully placed. The decision was taken to leave the intraosseous needle in place for the following 48 hours to have a back-up in case of loss of peripheral venous access. The girl recovered and left hospital.

Six weeks later, she was readmitted and diagnosed with osteomyelitis of the right proximal tibia, the exact same site where the intraosseous cannula had previously been placed. The site was surgically drained. From the pus enteritic salmonella were cultured, the same organism found in stool and blood cultures during the first hospitalization.

Discussion

The presented case of osteomyelitis after intraosseous cannulation reminds us that this clearly beneficial and often life-saving technique can have potentially significant complications (Neuhaus 2014). It is very likely that local tissue damage caused by the cannula in the spongiosa enhanced adherence of bacteria during bacteraemia. In retrospect, the intraosseous needle was unnecessarily left in place for a prolonged period. It seems advisable to remove an intraosseous needle as early as possible, i.e. immediately after venous access has been obtained, or after 24 hours at the latest, to minimize risks of infection. Following this case, it has been the author’s practice to administer a single dose of an antibiotic, e.g. cefuroxime, in the rare cases when intraosseous needles had been placed.

Complications do occur after intraosseous cannulation. In a large Scandinavian series, including 1802 cases, the following complication rates were reported: dislocation (8.5%), extravasation (3.7%), compartment syndrome (0.6%) and osteomyelitis (0.4%) (Hallas et al. 2013). In addition, bone fractures (La Fleche et al. 1989) and broken needles (Hallas et al. 2013) have been reported. The author has encountered a circular skin burn caused by the EZ-IO system: a needle was used that was too short because of marked tissue swelling secondary to the administration of local anaesthetic (Fig. 4.1). The most serious complication is extravasation leading to compartment syndrome; in some cases, this has even resulted in the loss of an extremity (Suominen et al. 2015). Therefore, it is of paramount importance to continuously observe the area of needle insertion in order to be able to stop fluid administration immediately when swelling occurs (Dasgupta & Playfor 2010). This can easily be done during resuscitation, but may be difficult during surgery, when the legs may not be accessible.

Figure 4.1 Circular scar caused by an intraosseous needle.

The EZ-IO system, an easy-to-use drill-assisted device, is the most popular product used today (Hafner et al. 2013). When dedicated material is not available, a normal 18G intravenous cannula can be used (Hamed et al. 2013). In children, the proximal tibia is the insertion site preferred by most practitioners, followed by the distal tibia or the distal femur.

An indication for intraosseous access is clearly given for resuscitation when peripheral venous cannulation is not immediately successful. Most clinicians will even directly proceed to intraosseous cannulation, when a rapid successful peripheral venous puncture seems unlikely. In contrast, there is an ongoing debate regarding the semi-elective use of intraosseous cannulation for anaesthesia in cases of difficult venous access (Neuhaus et al. 2010). The author has used this approach a few times in severely ill children for fluid administration and anaesthesia induction until central or peripheral venous catheterization was successful.

In cases of difficult venous access during elective anaesthesia, however, most clinicians will rely on an ultrasound-guided puncture, e.g. of the cephalic vein on the forearm, which is highly successful in experienced hands in the anaesthetized non-moving child. Alternatively, deep veins such as the femoral, internal jugular, brachiocephalic or subclavian vein can be cannulated. In exceptional cases, experienced anaesthetists occasionally proceed with anaesthesia for short (Jöhr & Can 1993) or non-invasive procedures (Wilson & Engelhardt 2012) without an intravenous access.

Summary and Recommendations

Intraosseous cannulation is an important technique in emergency medicine and, in a few exceptional cases, for the care of severely ill children in the operating theatre.

Intraosseous access should be removed as early as possible in order to minimize the risk of extravasation, compartment syndrome and infectious complications.

Case

A long time ago, a girl born at 38 weeks of gestation with a birth weight of 2500 g was scheduled for the repair of a left-sided diaphragmatic hernia on day of life 5. She was nasotracheally intubated and on conventional ventilation with an FiO₂ of 0.21. When she arrived in the operating theatre, an umbilical arterial catheter (UAC), a peripherally inserted 27G central venous catheter (PICC) and a closed 24G catheter (Insyte) on the left foot were in place. After careful evaluation, a decision was made to use these devices during the case. It was reported that the peripheral venous catheter was functioning well and only closed some minutes ago for transfer to the theatre. In addition, the line was rapidly flushed with 5 ml of normal saline without adverse events.

The anaesthetic management included a bolus of thiopental, sevoflurane and a continuous infusion of mivacurium; in addition, a caudal block with ropivacaine and morphine was performed. The PICC was used for the continuous administration of medications and glucose, whereas crystalloids and colloids were infused via the peripheral venous line. By the end of the 3-hour case, 30 ml hydroxyethyl starch, 40 ml of a balanced crystalloid solution (Ringerfundin) and a 6 ml bolus of an antibiotic had been given. The peripheral line was flushed with normal saline with the intention to close it again for transfer. When the drapes were removed, massive swelling of the foot with blister formation was seen (Fig. 4.2a). The blister was removed by the surgeon and the foot was aseptically dressed.

Figure 4.2a This extravasation was visible after the removal of the drapes. The infusion was administered through an initially well-running 24G cannula in the great saphenous vein.

The postoperative course was uneventful and the girl was extubated on postoperative day 3. Fortunately, the extravasation did not result in ischaemia or deep tissue necrosis. Nevertheless, the patient had to return repeatedly to the operating theatre for debridement and dressing changes.

Discussion

Infiltration of a venous line is an ever-present threat. Even when the catheter tip has a perfect intravascular position at the beginning of the case, fluid can leak along the catheter into the subcutaneous tissue, especially when high pressures are applied and the venous flow is impaired by thrombus formation. Over time, incremental swelling can cause the catheter to be pulled out. Even if the correct position is verified by flushing and by ultrasound (Gautam et al. 2017), there is no guarantee that late displacement, as described in this case, will not occur. When a line has just been inserted and a long segment of it is in an intravascular position, one can be confident that it will stay in place. However, many of these venous catheters are only 15–18 mm long, and only a few millimetres are placed within the vessel. Sometimes, replacing such a short catheter immediately after insertion by a longer one via a 0.018 inch guidewire is an option. Following this case, for major neonatal surgery the author mostly used venous catheters that had been newly inserted under his supervision, at least when intraoperative access to the access site was not possible. Alternatively, particularly when massive volume shifts were expected during neonatal surgery, he preferred to insert a central venous catheter into the internal jugular vein, e.g. a 4F double-lumen catheter.

When vasoactive drugs are infused over peripheral lines, massive local vasoconstriction impeding venous drainage seems to increase the risk of infiltration; nevertheless the administration of vasoactive compounds, especially dopamine, over a restricted time period using a peripheral line is common practice with a low morbidity (Patregnani et al. 2017).

On the ward, prevention of extravasation consists in a regular inspection of the infusion site (Tofani et al. 2012), as well as in setting the pressure alarm of the infusion pump at a very low level. It is noteworthy that even with an umbilical venous catheter that is not in the desired ideal position, extravasation can occur (Hollingsworth et al. 2015). Among all superficial veins, the external jugular vein is probably the one most prone to dislodgement, leading to extravasation into the loose neck tissue. For this reason, it was the author’s practice never to use an external jugular venous catheter for postoperative intravenous therapy. Even when the drip is running spontaneously, the tip of the cannula could be in the extravascular space.

Treatment of extravasation mainly consists in avoiding additional damage. Sometimes, especially in acute cases like the one described here, a proportion of the fluid can be squeezed out through the entry hole of the cannula, allowing rapid reduction of tissue pressure. Accidental subfascial infusion can lead to a compartment syndrome requiring fasciotomy (Fig. 4.2b) (Pasquesoone et al. 2016). When necrotic lesions are observed, it is wise to wait until they become clearly delineated. If only crystalloids, artificial colloids or blood products and no toxic compounds have been administered, prognosis is usually good. However, there are no evidence-based recommendations for the management of extravasations (Reynolds et al. 2014).

Figure 4.2b This extravasation caused ischaemia of the hand. Some of the fluid could be squeezed out through the entry hole of the cannula, and the hand subsequently turned pink.

Every undesired event should be communicated to the parents as early as possible. In the presented case, in line with his usual practice, the anaesthetist had spoken with the parents every 90 minutes over the phone to inform them of the progress of surgery and any potential problems during anaesthesia. They were understanding and forgiving, and even comforted the frustrated anaesthetist.

Summary and Recommendations

The case presented here shows that extravasation is an ever-present threat, even when the venous catheter was initially in a correct intravascular position.

The site of venous access should be monitored regularly; if this is not feasible, e.g. during surgery, it may be wise to insert a new device, and to select one which can be inserted into the vein over a long distance.

Case

A long time ago, a 9-year-old girl, weighing 30 kg, sustained a femoral fracture in a skiing accident. On the scene, paramedics established intravenous access in the right cubital fossa, administered fentanyl 50 µg and midazolam 2.5 mg, followed by a Ringer’s lactate drip. She was transported first by rescue sledge and then by ambulance over 60 km to the paediatric hospital. After an x-ray had been taken, the girl was brought directly to theatre for closed reduction and internal fixation of her femur. She was still on the initial emergency stretcher wearing her padded ski suit.

The anaesthetic plan included intravenous induction, endotracheal intubation and a femoral nerve block for postoperative analgesia. Proper functioning of the venous line was verified by the injection of fentanyl 50 µg flushed in with 10 ml of normal saline. Ringer’s lactate was dripping freely by gravity alone, leaving only a small remainder in the bag.

Anaesthesia was induced by injections of thiopental 2.5% 175 mg and succinylcholine 50 mg. Both were forcefully flushed in with 10 ml of normal saline. We wished her a good sleep and, well behaved, she closed her eyes. Cricoid pressure was applied. About 30 seconds later, she opened her eyes and asked ‘Why are you pressing on my throat?’

The suspected course of events soon became clear when the padded ski suit was retracted, exposing a massively swollen right elbow. Rapidly, venous access was established on the opposite hand and additional doses of thiopental and succinylcholine were injected intravenously. The further anaesthetic course was uneventful.

Discussion

This case of paravenous injection of thiopental and succinylcholine prompts consideration of two issues: first, the unrecognized malposition of a venous cannula, and second, the potential tissue toxicity of injected compounds.

Every effort should be made to ensure the correct intravenous position of a cannula before it is used for induction. This includes flushing without resistance and pain (done in the presented case), and inspection of the cannula site (not done in the presented case). Even an infusion dripping freely by gravity does not reliably prove correct cannula placement, especially in regions with loose subcutaneous tissue, e.g. at the neck or, as in this case, the cubital fossa. Undoubtedly, inspection of the cannula site during flushing would have avoided the complication. The author has to admit that in daily practice, especially when infants and young children are not cooperative upon arrival to the theatre, he often uses an existing intravenous access and does not always remove the bandages and the splint to inspect the cannula site. Ultrasound is increasingly used to verify that injected fluid can be seen in a venous vessel (Gautam et al. 2017).

Subcutaneously injected drugs will be absorbed and can cause clinical effects. While sedation by the hypnotic agent may be desirable, the onset of paralysis caused by subcutaneously injected succinylcholine would be a disaster for the child. In this case, 1.7 mg/kg of succinylcholine was given, less than the dose required for complete paralysis after intramuscular injection (Schuh 1982); nevertheless it is wise in such cases to induce general anaesthesia as rapidly as possible.

Paravenous injection can cause tissue damage (Le & Patel 2014). It has been known for a long time that subcutaneously injected thiopental can lead to tissue necrosis (Davies 1966). Solutions with lower concentrations, e.g. thiopental 1% or 2.5%, are less disastrous than the classical 5% solution. Indeed, the author has never encountered relevant tissue damage during his career (spanning 40 years in clinical anaesthesia), despite a number of accidental subcutaneous injections of 1% or 2.5% solutions of thiopental. Nevertheless, it is probably wise to use the less toxic propofol instead of thiopental if there is the slightest doubt about the proper positioning of a venous line. In the author’s experience, ketamine may be a good option in toddlers, because it will rapidly calm the situation to allow inspection of the venous access site. The most common cause of tissue necrosis following paravenous injection involves calcium-containing solutions; the author remembers several such cases, especially in neonates (Fig. 4.3).

Figure 4.3 Skin necrosis after extravasation of calcium gluconate in a neonate.

Another reason why thiopental should not be injected through a cannula lying in the cubital fossa is the possibility of unrecognized intra-arterial injection (cf. Case 4.4), which can lead to catastrophic complications including gangrene and limb amputation (Khan & Noorbaksh 2004, Kinmonth & Sheperd 1959). It seems wise to never inject thiopental into a catheter lying in the cubital fossa.

Summary and Recommendations

The described case of accidental paravenous injection of thiopental and succinylcholine in a 9-year-old girl shows a complication that appears completely avoidable (at least when judged in retrospect from the writing desk).

Both local toxicity and absorption, with ensuing clinical effects, have to be considered after paravenous injection.

The case also nicely illustrates that performing only some parts of a safety checklist is not enough. A venous line must be dripping freely and easily flushable without causing any discomfort; in addition, inspection of the cannula site should raise no concerns.

The delicious irony of the presented story is that the author did not follow the rule to ‘never inject thiopental into a vein in the cubital fossa’, a rule that he had himself established in his institution.

Case

Many years ago, a 2-month-old boy, weighing 4 kg, was scheduled for inguinal hernia repair. After inhalational induction, the airway was secured with a size 1 classic type laryngeal mask airway (LMA). Venous access proved to be difficult, leading to attempts on the scalp. Finally, a 26G Abbocath was successfully inserted into a vessel. There was only moderate backflow of blood; however, when the line was flushed with normal saline, a wide skin area immediately blanched (Fig. 4.4). Intra-arterial position was assumed and the catheter was removed before any active compounds had been injected.

Figure 4.4 Typical intra-arterial placement of a cannula with blanching of the skin following flushing

(from Jöhr, M. (2017). Complications in paediatric anaesthesia. Anästh Intensivmed, 58, 259–266, with permission).

Discussion

In the presented case, inadvertent intra-arterial placement of the cannula was recognized in time. Intra-arterial placement of peripheral catheters typically occurs on the scalp, in the antecubital fossa and at the radial aspect of the distal forearm. On the head, skin blanching can readily be recognized when the line is flushed. At all other sites, this is not the case and a high index of suspicion is required. Numerous cases of accidental intra-arterial injections have been reported in adults (Ghouri et al. 2002) and in children (Duggan & Braude 2004), most of them involving the use of thiopental and some of propofol. Whereas thiopental is directly toxic to the vascular endothelium (MacPherson et al. 1991), propofol is not (MacPherson et al. 1992). With propofol, except for one case (Ang 1998), only transient erythema but no long-term consequences have been reported. Intra-arterial injection of atracurium can lead to vasospasm (Kessell & Barker 1996) or no visible reaction (as described below).

In some children, usually with an extensive past medical history, venous access can be very difficult and only a very thin vein can be cannulated, e.g. on the back of the hand or on the palmar surface of the distal forearm. In such cases, there can be marked resistance to injection and blanching of the skin because venous drainage is limited and injected fluid is flooding back into the capillary system. It can thus be difficult to determine the correct position of such lines. Such cannulas should only be used very cautiously and replaced as early as possible.

A previously placed arterial line can erroneously be used to inject drugs. In his professional career, the author has made this painful error on two occasions in adults, once injecting midazolam and once injecting atracurium, fortunately with no long-term consequences. Colour-coding of the three-way stopcocks is helpful to avoid such errors, and it may be wise to additionally protect the injection port of an arterial line by applying a tape.

In some desperate cases, it might be tempting to use an existing arterial line to administer drugs. In neonatology, a large variety of compounds have been given via an umbilical artery catheter without obvious negative consequences. This can probably be explained by the fact that the injection is made into the aorta, with a very high blood flow and rapid drug dilution; when a peripheral artery is used, this is not the case. In anaesthesia, the only generally accepted drugs for intra-arterial use are heparin (to keep the line open) and local anaesthetics (Koscielniak-Nielsen & Horn 1993). Because of the paucity of available safety data, voluntary intra-arterial administration of any drug should probably be avoided.

Once intra-arterial injections have occurred, evidence-based treatments are largely lacking (Crawford & Terranova 1990). It seems reasonable to use heparin to prevent further thrombosis and to maintain sufficient perfusion pressure. The use of sympatholytic blocks has been reported (Kessell & Barker 1996).

Summary and Recommendations

Even in expert hands, intra-arterial cannula placement occasionally occurs when venous access is attempted. Timely recognition of cannula misplacement is of paramount importance.

Whenever there is the slightest doubt, a non-toxic compound should be chosen, e.g. propofol instead of thiopental. Remember: ‘Never inject thiopental into an elbow vein.’

Colour-coding of three-way stopcocks and even taping the injection port can help to prevent its inadvertent use for drug administration.