Can epidural kill?

Can epidural kill?

The probability of epidural causing death is extremely low. Theoretically it may occur as the result of either unexpectedly high block or local anesthetic toxicity. In addition, cases of cardiac arrest during spinal anesthesia for cesarean section have been reported.

Any invasive medical procedure may be fatal. However, during my medical career I have never observed or heard of epidural directly causing death of a patient. More objectively, Pubmed search of “fatal outcomes of labor epidural”, “death as the result of epidural” and similar key phrases did not return any reports or articles on the subject. Serious complications are often subject of medico-legal enquiries and don’t get reported in the literature, therefore the absence of publications does not necessarily mean fatalities don’t happen. On the other hand, the fact that anesthetic textbooks don’t even mention statistics of fatalities related to epidurals probably means that such an outcome is exceptionally rare.

Hypothetically there are several possible complications of epidural that may lead to the fatal outcome. Injecting local anesthetics directly into the blood vessel may lead to local anesthetic toxicity. Inadvertent dural puncture may result in high block leading to respiratory distress. Finally, in rare instances spinal block widely used for cesarean section may result in cardiac arrest. In the longer term, epidural abscess may also lead to a fatal outcome.

All these complications are discussed in separate chapters.

Toxic reactions to local anaesthetics

Every drug used in medicine has therapeutic index. Let’s say we take drug XYZ and start testing the effects of its various doses on the body. Let’s take something familiar, like caffeine. At very small doses caffeine will not have any effect on any parameter of the body. As the dose increases mild stimulatory effects become obvious, and you start feeling feel more awake. As the dose increases the feeling of being more awake becomes more pronounced. This is therapeutic effect, the reason many of us take this drug – in the form of coffee – in the morning. If we start increasing the dose of caffeine – take another cup of espresso – the wake up effect will become even more pronounced, but in addition to it we will also get other effects, such as increase in heart rate and blood pressure. We will also feel the need to go to the toilet more frequently, as caffeine is a diuretic at higher dose. Take another cup and you will start getting unpleasant side-effects: palpitations, anxiety, dizziness, shaking hands, insomnia and so on. Taking more and more caffeine may eventually lead to impaired consciousness, coma and death. Those of us who drink cofee in the morning subconsciously select the optimal therapeutic dose of caffeine which gives us the most awake state without the unpleasant effects. The range of dose or, more precisely, concentration of drug in the blood from which we get therapeutic effects to the concentration at which side-effects start appearing is called therapeutic window. The safety of every drug is measured by this parameter. Different drugs have different therapeutic windows. Paracetamol and vitamin C have wide, whereas drugs such as aminophylline have narrow therapeutic windows.

Local anaesthetics are interesting drugs in a sense that they are not given for their systemic, or general effects on the body, but for the effects at the site of injection. That’s why they are called “local”. As I explained earlier nerve cell is activated there is the flow of sodium ions into it. This action, in turn, starts from activating so called sodium receptor, sort of a lock that allows the flow of sodium. Molecules of local anaesthetics temporarily bind to these channels and thus make the cell unable to get stimulated. If nerve cell does not work, no pain impulse transferred from the skin and internal organs to the brain – the patient is happy!

And as discussed earlier, sodium channels are present not only in the nerve cells, but in other organs as well, including the heart muscle and the brain and are crucial for proper functioning of these organs. If local anaesthetics reach certain level in the blood two things happen. First, the patient loses consciousness and convulsions start. This results from local anaesthetics blocking certain areas in the brain and by itself does not present serious danger to the patient. If the concentration of these drugs keeps increasing, however, sodium receptors in the heart get blocked and the heart stops. Cardiac arrest arising from local anaesthetic toxicity is difficult to treat, and mortality is high.

When local anaesthetics are injected into the epidural space, they get absorbed into the blood stream slowly. While they are absorbed they are also inactivated by the liver and their concentration never reaches dangerous levels. However, in two situations we may encounter problems. One, rare, is when a patient is overly sensitive to the systemic effects of local anaesthetics and display symptoms and signs of toxicity at levels lower than usual. Two, when local anaesthetic is injected into a blood vessel by mistake. Every anaesthetist takes precautions to avoid this by aspirating, or pulling on the plunger of the syringe before injecting local anaesthetic, and when injecting, doing it slowly, so that if intravascular injection does happen, it is detected early, before the concentration of the drug reaches critical levels. Unfortunately, nothing is infallible, and intravascular injections of local anaesthetics happen, though very rarely.

Before the opponents of epidural start pointing fingers and say “I told you so”, I will state that in case of epidurals this problem is really rare, much less rare than even a decade ago. In fact, this complication is at its lowest. There are several reasons for it.

Different local anaesthetics display different potential for toxicity. Bupivacaine, for instance, is toxic and cardiac arrest that happened because of its administration is very difficult to treat. This drug is not widely used for epidurals anymore and is largely replaced by Ropivacaine, local anaesthetic with much better safety profile. The dose of Ropivacaine required to stop the heart is much higher than the dose that causes seizures, and intravascular injection of this drug is detected before it causes cardiac arrest. The treatment of cardiac arrest resulting from Ropivacaine is also easier to treat, probably of lower lipid solubility of this drug. Another recently introduced local anaesthetic is Levobupivacaine. It is also considered as having good safety profile. However, to my knowledge it is not widely used in labor epidurals, probably because of its slow onset and long duration of action, and probably because the safety of Ropivacaine is difficult to improve.

The main reason for low incidence of local anaesthetic toxicity though is that the doses that are used for epidural analgesia are low, and even if the whole amount intended for the epidural is given intravenously it is unlikely to cause serious problems. The highest safe dose of Ropivacaine is 2 mg per kilogram of body weight, which for average patient of, say, 70 kg is 210 mg. For epidurals we use it in concentration of 0.2%, which translates into 2 mg per ml. If we inject 20 ml of 0.2% Ropivacaine IV the patient will get 40 mg, the dose far, far below its toxic limit. More than that, the dose of 20 ml is rather on the upper border of that range, and usually 8 to 12 ml of this solution is sufficient for adequate epidural block. Because of this local anaesthetic toxicity is rare in obstetric practice.

The statement above is supported by currently available evidence. While there are no recent publications on local anaesthetic toxicity specifically in obstetrics, there are studies addressing this question in other areas of anaesthesiology. The review published in 2002 shows some interesting data1. The incidence of local anaesthetic toxicity seems to decline significantly after 1981, when the awareness regarding this complication increased among clinicians. Series published in the 50s reported the incidence of toxic reactions between 1 and 3%, the latter being observed in obstetric epidural population. It is important to note that resuscitation of obstetric patients in case of toxicity is especially difficult, due to changes in cardiovascular system characteristic for pregnancy. Luckily, in the review it is mentioned that “the decline in systemic toxicity is most dramatic in lumbar epidural anesthesia”, exactly the type of block used for pain relief in labor. Recent published series report the incidence of this complication of 1.2 to 1.3 per 10,000 patients or 0.12-0.13%. Moreover, as ropivacaine is safer from cardiovascular point of view, it is possible that most of the toxic reactions seen in recent years are neurological in nature and do not progress to cardiac arrest.

For the sake of objectivity I must mention at least one scenario where this is different. If often happens that a woman having epidural block has to undergo emergency cesarean section. When this surgery is not extremely urgent, anaesthetists often use epidural catheter placed earlier in order to provide anaesthesia for the cesarean. The amount of local anaesthetics needed to ensure pain free surgery with epidural is much higher than that required for pain relief in labor, and toxic reactions in this situation are possible. In case of Ropivacaine, anything between 10 and 20 ml of 0.75% solution is needed for successful epidural block, and it translates into 75 to 150 mg of total dose of the drug. If this is administered intravenously quickly, toxic reactions are possible.

Still, the complications of this sort are rare with epidurals for the reasons listed above. Every anaesthetist is trained to prevent local anaesthetic toxicity. We use test doses, inject the drug slowly, use low concentrations of local anaesthetics and are alert to the signs and symptoms of toxicity.

The verdict – dying from local anaesthetic toxic reaction is possible but very unlikely.

Unexpectedly high block, or “total spinal”

Nerves that supply the uterus and come from the thoracic segments of the spinal cord, T5 to T12. This is the level of epidural block that is necessary for the efficient pain relief in labor. Anaesthetists generally place the epidural catheter at the level of T3-4 or T2-3, below the level where the spinal cord ends. When local anaesthetic drug is administered through the catheter, it spreads upwards, and the dose of 10 to 20 ml usually achieves the desired effect.

If the block extends higher than the level indicated above, intercostal muscles – small muscles between the ribs – become affected. Normally, these muscles are not essential for breathing and are used for forced breaths and coughing. If these muscles become subjectively it feels to the patient that it is difficult to breath. However, there is no danger, and the respiratory system provides enough oxygen to the body while adequately eliminating carbon dioxide. This height of the block is often seen in cesarean patients receiving spinal anaesthetic, and simple reassurance is sufficient in the majority of patients. Besides, the level of the block decreases relatively fast, and unpleasant symptoms usually disappear within 15 – 20 minutes.

The problem arises when the block extends really high, to the cervical, or neck level. As you might remember from high school, the main muscle necessary for breathing is the diaphragm, a muscular layer that separates the chest cavity from the abdomen. When the diaphragm contracts, the air is pulled into the lungs – you inhale. Exhalation is mostly passive process normally not requiring any effort. The nerve that guides the movement of the diaphragm – phrenic nerve – originates from the spinal cord at the levels C3, 4 and 5 at the neck and passes through the same dural sack filled with cerebro-spinal fluid. If local anaesthetics spread that far this nerve becomes blocked and respiration becomes impossible.

At the doses and concentrations of local anaesthetics given for epidurals for the block to spread that high something unusual should happen. Several scenarios come to mind. The patient may have abnormally small epidural space where normal volume of drugs will spread abnormally high. This is very unlikely. Epidural catheter may also be placed incorrectly or even migrate through the dura and the block convert from the epidural to spinal. For spinal anaesthetic only two to three milliliters of local anaesthetic is enough to create fairly high block used for anaesthesia for cesarean section, and if big volumes used for epidurals – ten to fifteen milliliters – are administered, the block spreads abnormally high. But despite of the theoretical possibility this, again, is not very likely. First, every anaesthetist is trained to detect incorrect placement of the catheter. Second, migration of the catheter is very rare. Third, even if the catheter becomes intrathecal, the concentration of local anaesthetic used for epidural is so low that it should not cause severe compromise. Phrenic nerve is a large neural structure and due to its physiological properties fairly resistant to local anaesthetics. Ropivacaine, relatively sensory selective local anaesthetic, is even more unlikely to cause paralysis of the phrenic nerve.

Interestingly, one of the major operations on the carotid artery, so called carotid endarterectomy, in some centers is performed under cervical epidural anaesthesia with the patient awake and breathing spontaneously. Yes, it is correct, the epidural is intentionally placed at the neck level and local anaesthetics at reasonably high concentration and amount are given in order to achieve adequate block, yet there are no problems with the phrenic nerve.

In my practice I had three instances where phrenic nerve was affected to the degree that some intervention was necessary. All of them happened in theatre during cesarean section. In all three instances spinal anaesthetics were given, one of them probably too soon after large dose of local anaesthetic was given through the epidural catheter placed earlier for labor. All three cases were resolved quickly with no consequences. And despite of the distress caused by problems breathing and the need to intervene, all patients were satisfied with their childbirth.

The statements above are in agreement with the published evidence. The scenario where a spinal anaesthetic for caesarean section is given to a woman who had epidural for her labor has been discussed in the literature. Three such cases were published in 1994 from Sweden. All three women had to be intubated and ventilated and recovered without consequences. In the opinions of authors of the report spinal anaesthesia on the background of ongoing epidural was relatively contraindicated2. I would say caution should be exercised. Each patient is different and so is the risk-benefit ratio in every clinical situation. Besides, when laboring patient is booked for a caesarean, in most instances the epidural is stopped and enough time elapses for the epidural block to subside. Even if it does not, the only alternative to spinal is general anaesthesia, which is what we do if unexpectedly high block does occur.

Overall, the incidence of this complication is very low. A prospective analysis of more than 10,000 epidurals observed unexpectedly high blocks in eight patients or 0.07%, two of which required induction of general anaesthesia, intubation and ventilation3.

Cardiac arrest during spinal anaesthesia

In the early to mid-nineties the transition occurred from general anaesthesia to spinal for caesarean section. And as spinals became more and more popular and solved many problems related to general anaesthesia, reports of strange cardiac arrests started appearing in the medical literature.

No event in clinical medicine is as dramatic as cardiac arrest. Thanks to the medical TV dramas, it has become a household term, and most of readers probably know that we treat it with adrenaline!

In the vast majority of cases it happens in patients with serious pre-existing conditions or in those presenting with serious medical or surgical problems. There is not much surprise when a victim of a motor vehicle accident arrives to the emergency department with severe injuries and arrests on the table. Or a patient with the massive heart attack. Despite the fact that absolutely everything is always done in order to save a life, often the odds are against survival. However, when a young woman comes to the hospital for childbirth and arrests during caesarean section, the situation is far from ordinary.

A typical publication reporting cardiac arrest during spinal anaesthesia goes like this. A young female patient is brought to the operating theatre for an emergency caesarean section for non-reassuring fetal heart trace. The patient is given spinal anaesthetic which goes uneventful for 10 minutes or so. Then, suddenly, the heart rate of the patient slows down and the heart stops. The reported outcome of the event is variable, from quick successful resuscitation and full recovery to various degrees of brain damage to death. No joke, as you can see.

As a little sideline, it often amuses me how people build their arguments without knowing all the facts. If I wanted to discourage women from having elective caesarean sections, this is the first complication I would mention to them. Yet I have never heard the proponents of the natural birth even mention it.

Back to the science. The reported incidence of this complication varies. According to the authors of the review article in 2001 from Stanford University, cardiac arrests during spinal anesthesia are described as “very rare,” “unusual,” and “unexpected,” but are actually relatively common. The two largest prospective studies designed to evaluate the incidence of complications during spinal anesthesia reported two arrests in 1881 patients and 26 arrests in 40,640 patients for an overall incidence of seven arrests for every 10,000 (0.07%) spinal anesthetics. Another review of approximately 4000 regional anesthetics revealed six cases of severe bradycardia (pulse of 20 to 40 beats per minute) and six others (0.15%) with cardiac arrest after spinal anesthesia (3). These rates are high when compared with an incidence of three arrests from any cause for every 10,000 cases (0.03%) reported for patients undergoing other surgery4. The actual incidence of this complication in obstetric practice is difficult to estimate, as the numbers quoted above are based on all cases performed under spinal, not only caesarean sections, and include patients with serious pre-existing medical conditions. I am not going to downplay this statistic, however, as though advanced age and various medical conditions contribute to cardiac arrest during spinal, often these factors are absent, and according to the mentioned review, many of these arrests happened in young healthy patients.

Why do these arrests happen? There are a few theories about it. When spinal anaesthetic is given for procedures other than caesarean, sedative drugs are often administered in order to reduce the patient’s stress, and excessive oversedation has been blamed for depressed respiration, decrease in oxygen levels in blood and consequent cardiac arrest. Since the introduction of pulse oxymeters, the devices that continuously monitor oxygen levels in blood it was observed that these arrests often occur with saturation reading of 95-100% which is within perfect physiological range.

If respiratory system is not the one to blame, then one must search for the cause within the cardiovascular system. The activity of the heart is regulated by two main mechanisms: endocrine and vegetative nervous system. Endocrine system is represented by hormones circulating in blood, the most known being adrenaline and noradrenaline (also called epinephrine and norepinephrine). Vegetative nervous system influences the heart directly through the nerves and is divided into sympathetic and para-sympathetic systems that exert opposing effects on the heart. Increase in the sympathetic activity leads to the increase in the heart rate, blood pressure, the volume of blood expelled by the heart and strength of cardiac contraction, whereas para-sympathetic system leads to the opposite effects, slowing of the heart rate being the most commonly observed in practice.

The nerves representing sympathetic nervous system in the heart originate in the thoracic region of the spine. If spinal anaesthetic spreads high enough, these nerves may get blocked, with the consequent loss of the stimulatory influence of this part of the vegetative nervous system. Possible result is bradycardia and, if the opposing, parasympathetic system is activated, full arrest. Slowing of the heart rate is common during spinal anaesthetic, but by itself it rarely causes problems, even in blocks high enough to cause problems with respiration (see above). Even though the nerves supplying the heart are blocked, there are physiological levels of adrenaline in the blood that keep the heart going. Besides, the heart has the intrinsic ability to contract even if all nerves supplying it are blocked and there is no adrenaline around, and removing the influence of the sympathetic nervous system is just not enough for the heart to stop.

Sympathetic nerve fibers supply not only the heart muscle but blood vessels as well, without exception. When spinal anesthetic is given the nerves supplying the abdomen and the lower limbs are blocked – that’s why anesthesia is achieved – and so are sympathetic fibers that govern the blood vessels in these areas. That causes blood vessels to dilate. The effect follows simple logic: if the volume of pipes (blood vessels) in which certain volume of blood circulates increases, the pressure in these pipes decreases and less blood returns to the heart. This has been shown in various studies on cardiac physiology, and the pressure has been shown to go down by 30 – 50%, depending on the level of the block. If the blood is lost during the operation the pressures may go down even more. This, in turn, possibly triggers three regulatory reflexes, and as the result the heart slows down and, with sufficient decrease in pressures, stops completely.

It is thought that so called Bezold-Jarisch reflex is responsible for cardiac arrests under spinal anaesthesia. This reflex is activated when the heart is underfilled and its strong contractions cause the distension of the ventricular wall. As the result, sympathetic nervous system is inhibited and para-sympathetic is activated. As the result, the heart slows down and, if the activation of the para-sympathetic system is strong enough, stops completely.

Many of us have observed the effects of para-sympathetic system in working. Medical name for fainting – loss of consciousness caused by various factors – is vaso-vagal attack, or episode. Vagal nerve is the main representative of the para-sympathetic nervous system and is present in virtually every internal part of the body. Its activation universally leads to bradycardia of varying degrees and has been observed during many surgical actions, such as manipulating the bowel, the uterus, muscles surrounding the eyes, pleura and many others. When it gets hot and you get dehydrated it is the Bezold-Jarisch reflex that kicks in. But in this example it is short lived and self-correcting: you fall down, your body assumes horizontal position and the blood is re-distributed to the heart from the lower parts of the body. The heart fills and the influence of the vagal nerve is inhibited. In case of spinal anaesthesia this is not the case, and only the interference of the anaesthetist can avert the disaster.

Various measures have been taken to prevent this complication. The awareness and knowledge about its mechanisms is the first. If the problem is caused by the underfilling of the heart, simple logic dictates that patients must be given fluids, and intravenous administration of normal saline or a colloid fluid before establishing spinal block has long been the standard of practice.

Some patients have higher vagal tone before surgery, and several clinical signs have been identified that may point out at these patients. Taking prophylactic measures in order to decrease the activity of the para-sympathetic activity may prevent the development of severe bradycardia and cardiac arrest during caesarean section. Alternatively, in such patients general anaesthesia may be chosen for their surgery.

If the slowing of the heart rate does occur, it is treated aggressively. The patient is given drugs that mimic the action of the sympathetic (epinephrine) and inhibit para-sympathetic nervous system (atropine). Treating progressive bradycardia with atropine and ephedrine has been shown to improve the outcome.

Just for the sake of argument, I think that patients coming to theatre for the emergency caesarean section from labor ward have somewhat higher risk of cardiac arrest. Among several factors leading to this complication, dehydration is the major step in its physiological mechanism. Patients in labor are more likely to be dehydrated as the pain of childbirth often inhibits the feeling of hunger and thirst. They also strain and sweat and many birthing units have the policy of nil per mouth, which also limits the amount of water consumed by the patient.

The patients with increased vagal tone who from the start have heart rates below 50-60 per minute may be at increased risk. These patients might be at higher risk during spinal anaesthesia.


1. Mulroy MF. Systemic Toxicity and Cardiotoxicity From Local Anesthetics: Incidence and Preventive Measures. Regional Anesthesia and Pain Medicine, Vol 27, No 6, 2002. p. 556–561

2. Gupta A, Enlund G, Bengtsson M, Sjöberg F. Spinal anaesthesia for caesarean section following epidural analgesia in labour: a relative contraindication.Int J Obstet Anesth. 1994 Jul;3(3):153-6

3. Paech MJ, Godkin R, Webster S. Complications of obstetric epidural analgesia and anaesthesia: a prospective analysis of 10,995 cases. Int J Obstet Anesth. 1998 Jan;7(1):5-11

4. Pollard JB. Cardiac Arrest During Spinal Anesthesia: Common Mechanisms and Strategies for Prevention Anesth Analg. 2001;92:252–6



Dr. Eugene Smetannikov is a practicing anesthesiologist with the interest in obstetric anesthesia. He is the author of the most comprehensive book on the subject, The Truth About Labor Epidural