The present study represents the first series of spinal anesthesia performed in pediatric patients, in a hospital center in Madagascar. Over a seven-year period, 69 children were scheduled for spinal anesthesia, with a GA conversion following failure rate of 5.8%.
1. Incidence, clinical features and indications of pediatric spinal anesthesia
The indication of spinal anesthesia was mainly limited to the situations where general anesthesia leaded a great risk for the child, especially respiratory risk [1, 4, 6]). After the decline of caudal anesthesia in the years 1990–2000, SA had an upsurge of 2.1 to 3.6% in regional anesthesia techniques, even if spinal anesthesia in newborns (or even in premature infants) is controversial [3, 7]. Plus, SA is the “gold standard” technique in the former preterm infant (< 60 weeks PCA) for lower abdominal and lower extremities surgeries under 90 minutes duration . In a work by Williams R.K. et al.  spinal anesthesia was performed in 95.4% of children. In France, this technique represents 18% of regional anesthesia in preterms and 5% in newborns . In Finland, 400 to 500 spinal anesthesias are performed annually . As related in the present study, first spinal anesthesias in Antananarivo were performed in 2013, with 69 cases in seven years, for surgeries ranging 10 to 65 minutes.
The success of the technique was estimated on the reflux of cerebrospinal fluid (CSF) which was 97.1% in the present study, almost similar to the rate of 97.4% reported by Williams R.K. et al. . One lumbar puncture was perfomed in 47.3% and general anesthesia conversion was 5.8%. Sedation is often necessary, during insufficient block and the conversion to general anesthesia is indicated in case of failure of the technique . Dohms K. et al.  find a failure rate of 7.5% and in 28% cases, more than two punctures were needed and 16% required supplemental anesthesia. In Kachko L.'s  study, general anesthesia conversion occurred in 1.04%.
The success of the spinal anesthesia was estimated and based on the motor skills of the lower limbs and the relaxation of the anal sphincter (when present), as well as the effectiveness of the surgical gesture. The use of general anesthesia after spinal puncture was 2.9% due to the quality of the product. Since the Bromage score is not assessable in this population category, the effectiveness of spinal anesthesia can be assessed by the possibility of performing the surgical procedure .
The dose of bupivacaine used in our series is 4[3.5-4] mg. The most commonly used local anesthetics are 0.5% tetracaine and 0.5% bupivacaine; the usual dose is 0.6–0.8 mg / kg to reach average levels, and 1 mg / kg for higher levels (thoracic) . These two molecules act during 90 to 120 minutes .
2. Spinal anesthesia, safe and real alternative to general anesthesia?
Spinal anesthesia offers an interesting and reassuring alternative if tracheal intubation should be avoided by the underlying pathologies such as bronchopulmonary dysplasia or respiratory diseases [1, 7]. Indeed, spinal anesthesia can avoid apnea or bradycardia and have minimum cardiorespiratory complications [2, 6, 10, 11]. In the cases presented, most of the children presented an anesthetic risk in case of general anesthesia, due in particular to respiratory diseases and very young age (20.3% of rhino-bronchitis, and 39.1% of prematurity). This situation motivated indication of spinal anesthesia in the CHU JRA.
Spinal anesthesia offers a good balance between safety and perioperative risks and seems to be a secured technique for the operated child, as long as compliance with contraindications is observed [5, 6, 11]. SA is more effective in blunting the neuroendocrine stress and adverse effects of surgery and provides additional effective intraoperative analgesia [2, 12].
Spinal anesthesia in children allows remarkable cardiovascular stability [13, 14]. The frequency of complications is 30% . Ventilation and oxygenation are not generally compromised, even in patients at high risk, in preterm and expremature [13, 14]. SA has a moderate risk of apnea and bradycardia (RR = 0.72), desaturation (RR = 0.82) and a low risk of needing postoperative respiratory assistance (RR = 0.09) .
In this work, no bradycardia of less than 100 bpm was observed, probably due to the very slow infiltration of the anesthetic and the 45° head up tilt position of the patient directly after infiltration. No other complications arose during these SA.
In addition, spinal anesthesia can provide other benefits, such as a shorter length of stay compared to the length of stay after general anesthesia as well as faster recovery of gastrointestinal function .
3. Spinal anesthesia in developing countries? And .... in Madagascar?
The use of spinal anesthesia especially in “precarious” or “difficult” situations is attractive because it requires fewer perioperative resources . However, a conversion to general anesthesia in case of failure could incur higher costs, hence the need for some experience in the gesture and a well knowledge of the technique. Very few studies on spinal anesthesia in children, much less in infants, newborns and preterms / ex-prematures, in low-incomes countries, have been found.
Ela A.A. et al.  report a series of 55 patients from one day to 16 years old, operated under spinal anesthesia with bupivacaine associated with fentanyl, with a Whitacre 25 G or 22 G needle. The spinal anesthesia performed in their study population was performed by (i) either an intensivist, (ii) or an anesthesia-intensivist resident, (iii) or a state-certified nurse anesthetist . The indications were predominantly inguinal and scrotal hernias (48.15%), and surgeries range from 25 minutes to 78 minutes . No complications were found in their study. Due to the many categories of age of the patients in the Ela A.A. et al.  study, the results of our study differ from theirs, particularly in terms of the puncture site and the products used.
In Antananarivo, this series is the first to have been mentioned. The strength in our results is the specificity of the study population (preterms, newborns, infants). These first results being apparently acceptable, this technique should be extended, nevertheless after training of anesthesiologists. Indeed, this technique requires a high level of skill, particularly if it is carried out in very little patients, who are "non-cooperating" with small anatomical dimensions [1, 13]. The failure rate is 28% and the risk of total spinal anesthesia is possible (around 0.63 to 0.8%), especially since anesthesia trainees have a significant different success rate compared to anesthesiologists (83% vs 98.9%) [1, 8].