Regional anesthesia is both an old and a new technique. It is now a well-established technique of anesthesia and its use has increased very much during the last 20 years [1]. Providing estimates of the incidence of the various complications related to regional anesthesia is not a new concern. In two classic studies, each assessing a large number of spinal blocks, Dripps and Vandam assessed the risk associated with the use of procaine and tetracaine in 10,098 patients [2], whereas Phillips et al. monitored 10,440 patients after lidocaine spinal anesthesia [3]. The main message of these prospective studies was that complications related to spinal anesthesia are very rare. Such results and the numerous advantages associated with regional anesthesia have contributed to the perception that regional anesthesia is “safe” and this has translated into an increasing number of regional anesthesia procedures performed worldwide. However, one should be very careful before extrapolating these old results to our current practice.

The comparison cannot probably be made not only because of methodological concerns and but also because of tremendous quantitative and technical changes during this 30-year period. This factor also restricts our ability to conduct meta-analysis studies [4]. Unfortunately, the number of recent prospective studies assessing the incidence of severe complications related to regional anesthesia is low, and this is particularly true when peripheral nerve blocks are concerned. Severe complications are rare and this is the main factor explaining the low number of studies. Indeed, the number of monitored procedures has to be very large in order to estimate the level of risk with sufficient statistical power [5]. In the case of rare events, other approaches that have been developed in other fields of research need to be used to understand and to control the risk associated with regional anesthesia techniques [6, 7].

In 1996 a large epidemiologic study evaluating the incidence of serious complications associated with regional anesthesia and evaluating their characteristics, was published in Anesthesiology [8]. In 1998, a completely new service, entitled SOS Regional Anesthesia (RA) Service, was established [9]. This service first included a hot line and three experts (Pr Samii, Pr Ecoffey, and Pr Benhamou) rotated each week to respond to any question asked by participants on regional anesthesia at any time (even at night if necessary) and 7 days a week (even Sunday if necessary). SOS RA Service had four main goals: (1) to provide an online clinical help for the practitioner facing a severe complication, (2) to obtain immediately relevant clinical information for every complication reported (and obviate the loss of pertinent information related to late collection as this occurred in the first survey), (3) to provide advice on difficult clinical cases before any anesthesia is given (generally at the time of the preanesthetic visit), (4) to estimate the incidence of complications from a prospective declaration of all regional techniques performed by practitioners who had subscribed to the service. The SOS RA Service works currently according to the three first initial goals as the calculation of incident rates was not maintained after the first 10-month period because of the complexity related to exhaustive case collection. Even with this deficiency, this expert system remains highly demanded by practitioners (one phone call each day as a mean) and is very useful for detecting the emergence of “new” complications. From the voluntary participation of 487 anesthesiologists who performed 158,083 regional blocks in a 10-month period, 56 major complications (including four deaths) were reported in the SOS RA survey.

The incidence of cardiac arrest that occurred after spinal anesthesia was 2.7/10,000. Interestingly, the clinical situations associated with cardiac arrests were homogeneous because bradycardia was recorded before each cardiac arrest that occurred during spinal anesthesia, and cardiac arrest causing death occurred in the course of a central block performed during hip surgery in an elderly patient. Spinal anesthesia is mainly a vasoplegic process and crystalloid preloading has a limited protective effect [10]. The use of vasoactive drugs, mainly drugs with a potent alpha effect, are needed. This has been well demonstrated in obstetrics [11], a situation in which the dose of local anesthetic used has to spread over a large metameric distribution (up to T4–T5). As bradycardia is an important warning sign, surveillance is important. During cesarean delivery, heart rate is known to be a surrogate factor of cardiac output and bradycardia occurs more often with phenylephrine than with ephedrine [11]. Bradycardia should be treated immediately by reducing the infusion rate, adding ephedrine or atropine if blood pressure is low [11]. Since cardiac arrests may occur later, attention should be maintained throughout the procedure, especially when high block is used or when additional contributing factors can be encountered during the procedure. The factors involved in cardiac arrest occurring during central blocks are numerous and the risk probably increases from the beginning of the procedure until it ends. Factors causing hemodynamic instability superimpose on those previously present. In cardiac arrests that occur “later,” additional factors that add to an already unstable situation, include sympathetic blockade and hemorrhage. Special attention should therefore be given to correct each factor that might contribute to decompensation.

One case of cardiac arrest and two respiratory complications (not leading to cardiac arrest) occurred during a lumbar plexus block performed via the posterior approach and the incidence of 80/10,000 seen after posterior lumbar plexus block is obviously much higher than after spinal anesthesia . Complications are related to cephalad diffusion of the local anesthetic in the epidural or intrathecal space [12]. Although it was difficult to draw any definite conclusion regarding this block, French anesthesiologists were warned against the high rate of complications that was found with the posterior lumbar plexus block and advised to manage this block with at least the same vigilance as for a central block [13].

In the main SOS-RA survey [9], local anesthetic systemic toxicity consisted of seizures only, without cardiac toxicity. The results suggested a decreased rate of local anesthetic-induced systemic toxicity when compared with the first survey [8] although methodological differences between the two studies preclude any definitive conclusion. If this result proves to be true, the low incidence of toxic systemic complications may be related to better physician information, improved practice patterns (lower doses, slow injection, test dose, fractionated injection …), and the introduction of ropivacaine in clinical practice (at the time the first study was performed, ropivacaine was not available in France). In the face of these reassuring results, two important points were emphasized at that time: (a) the most important factor for increased safety is to maintain a high level of vigilance even if ropivacaine was introduced to prevent systemic toxicity [14], (b) the “good” prognosis of these complications (neither cardiac arrest nor death were reported) could become worse if such complications occur outside the operating theater (i.e., in case of postoperative analgesia on the wards). A few years after, however, case reports describing cardiac arrest were published [15]. Although the safety of ropivacaine can be questioned after the report of these cardiac arrests, it should be noted that both patients were easily resuscitated, a characteristic that is obviously different from bupivacaine . This also shows that the absence of adverse events in large surveys cannot lead to the conclusion that the incidence is zero. Calculation of the incidence of rare complications thus remains difficult and might be underestimated, again suggesting that epidemiologic surveys are not the only way to study rare events.

A recent prospective survey performed under the auspices of SOS-RA was aimed at gaining information related to complications associated with ultrasound guided axillary plexus blocks [16]. This was believed to be useful as initial large-scale studies failed to clearly demonstrate a reduced incidence of LAST [17]. Of 27,031 blocks performed, the incidence of systemic toxicity of local anesthetic was very low at 1.5 per 10,000 [16]. This incidence was in agreement with other recently performed studies and also confirms that ultrasound guidance probably reduces the risk of LAST although this technique does not nullify the risk [18–20]. Interestingly, there was no systemic toxicity from local anesthetic due to delayed absorption from the tissues [16]. Inadvertent vascular puncture may not be reduced with ultrasound guidance but may be associated with a reduced dose of local anesthetic administered thereby reducing the risk of delayed LAST [21].

Another issue which has recently attracted attention is the occurrence of seizures or cardiac arrest in patients who had undergone TAP block. In most of these reports, women of relatively small height and weight and bilateral TAP blocks using relatively large doses of local anesthetic were involved, reminding us that these muscle planes are highly vascularized [22, 23].

In the main SOS-RA survey, lidocaine spinal anesthesia was associated with more neurologic complications than bupivacaine spinal anesthesia (14.4 versus 2.2/10,000) [9]. Most neurologic complications were transient. These results about transient neurologic symptoms and neurologic toxicity of lidocaine contributed to the declining use of intrathecal lidocaine in France.

Among 12 complications that occurred after peripheral nerve blocks [9], 9 were observed in patients in whom a nerve stimulator had been used, demonstrating that nerve stimulation is not a definitive guarantee against neurologic complications. Moreover, the exact incidence of neurologic complications after nerve stimulation (versus other techniques) cannot be calculated from this study because of the low number of cases. In cases reported in our files since 1998, inadequate patient positioning and/or non-cooperative patients, insufficient physician experience, insufficient patient information on the procedure, excessive sedation, or a non-gentle technique are often critical factors that contribute to increased risk of neurologic complications. Obviously, these factors hold true also when a nerve stimulator is used. The use of nerve stimulation was already accepted in European institutions and a relatively new debate emerged related to the significance of a paresthesia occurring during puncture. This debate is far from being closed. Experts using ultrasound guidance have, for example, added to the discussion by reporting several cases in which the needle had made physical contact with a nerve, but no paresthesia was felt by the patient [24]. Others have also shown that intraneural injection can follow a puncture in which nerve stimulation has been used without any warning sign [25].