In the present study, spinal anesthesia was scheduled for 69 babies. This series represents the first pediatric spinal anesthesia, performed since 2013, in Madagascar. In developed countries, such as in United States, SA was included since 1977, with 262 SA, on less than 1-year patients, in 15 years . Williams R.K. et al.  reported 95.4% pediatric spinal anesthesia. In Europe, 400 to 500 SA are performed annually, with 18% in preterms and 5% in newborns [3, 10]. In other countries, like India, in a one-year period study, 102 children (from 6 months to 14 years) received spinal anesthesia for subumbilical and lower limb orthopedic surgeries . In low-incomes countries, few studies on pediatric spinal anesthesia have been related. For instance, Ela A.A. et al. , in Cameroon, report a series of 55 children operated under spinal anesthesia. However, the use of spinal anesthesia especially in “precarious” or “difficult” situations is attractive because it requires fewer perioperative resources .
Spinal anesthesia in the present study was performed, even in very young age (27 were preterms) and low weight child, and 14 children had medical history of respiratory diseases. Spinal anesthesia is primarily indicated when general anesthesia is at high risk (= respiratory complications or postoperative apnea because of pulmonary disease or prematurity) [1, 7, 10, 13]. SA is the “gold standard” technique in preterms (gestational age ≤37 weeks) and high-risk patients (preterm infants with postconceptual age <60 CW) [2, 9]. Indeed, this population is at high risk of postoperative apnea, especially if general anesthesia is performed. Spinal anesthesia is a safe alternative when tracheal intubation should be avoided (due to bronchopulmonary dysplasia or respiratory diseases, …) [1, 4]. Indeed, spinal anesthesia can reduce or avoid apnea [9, 10]. Also, SA generates minimum respiratory complications [2, 10, 14, 15]. In this study, most of the patients had respiratory diseases and some were very young (20.3% rhino-bronchitis, and 39.1% prematurity). All these facts motivated spinal anesthesia in the present study. In addition, the characteristics of the patients were quite similar with a study by Hermanns H et al. : 34.5 (24–40) weeks at birth, 10 (5–24) weeks postnatal age at the time of the intervention, and 3.5 (2.2–5.2) kg in weight.
The surgeries (lasting 27.5 [17.5 - 40.0]), in the present study, were mostly hernia repairs. Spinal anesthesia is the gold standard for lower abdominal and lower limbs surgeries under 90 minutes duration [1, 2, 5, 7]. This was similar with a study of Ela A.A. et al.  (25 minutes to 78 minutes) and shorter than in a study of Frumiento C. et al.  (48 [15-130] minutes). The most concerned surgeries are inguinal hernia repair [1, 2, 5, 7]. But other surgeries (resection of ileostoma, sacral teratoma, …) can also be performed under SA [11, 13].
The spinal puncture (2 [1 - 2] attempts) was performed in sitting or lateral position, in the intersection between the line connecting the highest point of both iliac crests (Tuffier’s line) and the vertebral axis, with an 80mm – G25 Quincke spinal needle. The midline approach on Tuffier’s line is the most used in SA in small children, in lateral or seated position [6, 13, 16]. A 25G pencil-point needle such Whitacre (avoiding post lumbar puncture headache) or 25G neonatal spinal needle are recommended [11, 13]. These types of needles are not available in the CHU JRA, so 25G Quincke spinal needle was used for all patients.
Hyperbaric bupivacaine 0.5% was used with a dose of 4 [3.5 - 4] mg. The most commonly used local anesthetics are tetracaine 0.5% and bupivacaine 0.5% lasting 90 to 120 minutes . Hyperbaric bupivacaine (0.5%) is mostly used in a dose from 0.3 mg/kg to 1 mg/kg [1, 6, 13, 17].
In the present study, for all the patients, 2 anesthetists who had prior training in this technique performed SA for limiting the bias in performance. Even the spinal anesthesia can be performed by either an anesthetist-intensivist, or an anesthesia-intensivist trainee, or a state-certified nurse anesthetist, SA performer should be well trained for the technique . Trainees in anesthesia have a significant different success rate compared to anesthesiologists (83% versus 98.9%); the failure rate is 28% and the risk of total spinal anesthesia is approximately 0.63 to 0.8%, if the performer is not trained [1, 5].
The success of the lumbar puncture was 97.1% after 2 [1 - 2] attempts and SA success was 94.2%. Since the Bromage score is not assessable in this pediatric population, the success of the spinal anesthesia is estimated and based on sudden loss of leg movement while normal tonus in the arms and/or the relaxation of the anal sphincter and the possibility of performing the surgical procedure [14, 18]. Williams R.K. et al.  reported a success rate of 97.4%, quite similar with our results. In the present study, no complementary local anesthesia by the surgeon was required and the GA conversion was 5.8%. This failure rate was similar to literature, varying from 1.04% to 24.6% [7, 9, 17, 19]. Dohms K. et al.  find a failure rate of 7.5% and 16% required supplemental anesthesia, also more than two punctures were needed in 28%. In Kachko L.'s  study, conversion to general anesthesia was 1.04%. In inguinal hernia repair, Frumiento C. et al.  describe 91.4% of adequate spinal anesthesia, 78,6% no supplemental anesthesia, 4.5% complementary local anesthesia and 2.2% general anesthesia conversion.
The heart rates were stable throughout perioperative period. Spinal anesthesia allows remarkable cardiovascular stability and can avoid bradycardia with minimum cardiac complications [2, 8, 10, 14, 15, 18]. But in some cases, 1.5% patients experienced bradycardia in operating room, and 1.9% received vagolytics .
Spinal anesthesia offers a good balance between safety and perioperative risks and appears to be a safe technique, provided that the contraindications are respected; the frequency of complications is 30% [1, 5, 10, 15]. SA causes less bradycardia, apnea, desaturation, requiring postoperative respiratory assistance than GA; ventilation and oxygenation are not generally compromised, even in patients at high risk [8, 18, 20]. In the present study, no perioperative complications were observed.
In Antananarivo, this series is the first to have been reported. The strength in the present study is the characteristics of the study population (preterms, newborns, infants).