In the early 20th century, Lord H. Tyrrell Gray advocated the use of spinal anesthesia for surgery in infants, declaring that “…it will occupy an important place in the surgery of children in the future” (1). By 1933, infant spinal anesthesia was even proposed for pediatric thoracic surgical procedures such as lobectomy and pneumonectomy (2). However, as the safety of general anesthesia improved, spinal anesthesia fell into disuse. Despite scattered reports (3), it was not until a 1984 study by Abajian et al. (4) that infant spinal anesthesia was successfully reintroduced to the modern era. Since that time, infant spinal anesthesia has been used alone or in combination with epidural anesthesia for a variety of surgical procedures, including inguinal hernia repair, exploratory laparotomy, repair of gastroschisis, orthopedic procedures, pyloromyotomy, meningomyelocele repair, and as an adjunct to general anesthesia in infants undergoing repair of complex congenital heart disease with cardiopulmonary bypass (5–18).
Epidemiological data suggest that infants are at an increased risk of complications associated with general anesthesia as compared with older children and adults (19–22). A technique that avoids the use of general anesthesia in this group of high-risk patients could therefore have important safety advantages. Infant spinal anesthesia exhibits a high degree of cardiovascular and respiratory stability (23,24). In several small comparative trials, infant spinal anesthesia has been associated with a decreased incidence of hypotension, hypoxemia, bradycardia, or postoperative apnea compared with general anesthesia (25–29). However, infant spinal anesthesia remains underutilized relative to general anesthesia.
A study involving a large patient population is needed both to confirm the safety of spinal anesthesia in infants and to determine the clinical feasibility of using this technique in the current operating room environment. We have prospectively studied all patients undergoing infant spinal anesthesia at the University of Vermont since 1978. Patient demographic, surgical, and anesthetic data are recorded in real time as the case progresses and are entered into a database, the Vermont Infant Spinal Registry. Data from small subsets of Vermont Infant Spinal Registry patients have been reported previously (4–10). We now report our comprehensive experience with every infant spinal anesthetic attempted at the University of Vermont.
At the University of Vermont, spinal anesthesia is used for most lower abdominal and lower extremity surgical procedures in infants. General anesthesia is used for infants with coagulopathies, infections, or congenital malformations in the area of proposed lumbar puncture (LP) or those infants already receiving mechanical ventilation at the time of surgery. With approval of the IRB, all patients younger than 1 yr undergoing attempted spinal anesthesia are included in the Vermont Infant Spinal Registry. Patient demographic data include chronological and post-concept-ual ages, birth weight, weight at time of surgery, and pertinent medical history. Technical data concerning the anesthetic include type and dose of local anesthetic, size of the LP needle, and whether cerebrospinal fluid was successfully obtained. The training level of the operator (resident, nurse anesthetist, or attending) and the background of the supervising attending (general anesthesiologist versus fellowship trained pediatric specialist) are recorded.
LP was performed with the infant in either the sitting or the lateral decubitus position. After cerebrospinal fluid was obtained, local anesthetic in a 1-mL tuberculin syringe was injected to produce subarachnoid block. If epinephrine was used, 20 μg of a 1:1000 solution was added to the syringe. The sensory level of spinal block was judged to be adequate by sequential observation of profound motor weakness in the lower extremities and a lack of response to firm skin pinch in the dermatomal level appropriate for the surgical procedure. If an insufficient sensory level was obtained, the case was either postponed or general anesthesia was induced. In either situation, the case was classified as a spinal anesthesia failure.
The need and indications for supplemental oxygen, assisted ventilation, or conversion to general anesthesia are reported. The use of medications including sedatives, anticholinergics, or vasopressors and their indications are recorded. Since 1991, time and efficiency data including elapsed time until spinal anesthesia is successfully induced, time until surgical incision is made, length of the surgical procedure, and time until the patient leaves the operating room have also been recorded.
Our review of Vermont Infant Spinal Registry data includes 1554 consecutive patients. Birth weights ranged from 540 g to 7800 g. The mean patient weight at the time of surgery was 4379 g (range, 650 g–13.6 kg). Surgeries performed include general, urologic, orthopedic, neurosurgical, and thoracic procedures (Table 1).
In most cases (79.1%), the primary anesthesia provider was either an anesthesia resident or a nurse anesthetist. An attending anesthesiologist supervised all cases. Although in the majority of cases the attending was a pediatric anesthesia specialist, 36.9% of cases were supervised or performed by an anesthesiologist without subspecialty training in pediatric anesthesia.
LP was performed using 25-gauge (53.0%), 22-gauge (43.0%), or 26-gauge (3.9%) needles. A successful LP, as defined by return of cerebrospinal fluid through the needle, was obtained in 97.4% of patients. Data on whether the successful operator was an attending or trainee anesthesiologist were available for 995 patients. Anesthesia trainees had an LP success rate of 83.0% (n = 826). This included residents having their first experiences with infant spinal anesthesia and more experienced trainees. When an attending anesthesiologist performed the case alone, the LP success of 98.9% (n = 194) was significantly more frequent than the success rate when the case was performed by a trainee (Fisher’s exact test, P < 0.001). Overall, spinal anesthesia appropriate to begin surgery was achieved in 1483 patients (95.4%).
Hyperbaric tetracaine was used in 99.6% of cases; the remaining cases used either hyperbaric lidocaine or bupivacaine. Excluding patients undergoing patent ductus arteriosus ligation (n = 20), the mean tetracaine dose was 0.54 ± 0.20 mg/kg. The total dose of tetracaine ranged between 0.375 mg and 4.5 mg. Epinephrine was used in 91.0% of cases. The majority of patients did not require any supplement to the spinal anesthesia. These patients were calmed if necessary by stroking, soothing, or giving a pacifier. IV sedation was administered to 24.1% of patients. The most commonly used sedative was midazolam, with typical doses of 0.20–0.25 mg. Supplemental local anesthesia from the surgeon was used in 2.7% of cases, primarily during the latter part of surgery as the subarachnoid block was regressing. In 18 patients (1.2%), the surgical procedure outlasted the duration of spinal anesthesia, requiring conversion to general anesthesia to complete the operation.
A mean time of 10 min elapsed from when the patient first entered the room until the spinal anesthetic was successfully administered (Table 2). The mean time from patient entry into the operating room until skin incision was 27 min. Mean duration of surgery was 47 min. Mean time from completion of the surgery until the patient left the operating room was 6 min.
Although most patients remained in room air, 55 of 1483 patients (3.7%) received supplemental oxygen either by nasal cannula or blow by. Oxygen hemoglobin desaturation to <90% was noted in 10 of 1483 patients (0.6%).
In 56 of 1483 patients (3.8%) the sensory level of spinal anesthesia obtained was higher than intended. The mean dose of tetracaine for these patients was 0.57 ± 0.16 mg/kg, which was similar to the tetracaine dose for the population as a whole (Wilcoxon’s rank sum test, P = 0.41). Twenty-three of these patients were given supplemental oxygen while maintaining adequate spontaneous ventilation. In 10 patients, the level of spinal anesthesia appeared to compromise spontaneous ventilation. These patients received either bag mask ventilation (5 patients) or had their trachea intubated (5 patients). Two of the intubated patients were undergoing exploratory laparotomy, one had a colostomy performed, one was undergoing gastroschisis repair, and one was undergoing inguinal hernia repair. In one patient, the onset of the high block appeared to correspond with dosing an epidural catheter, which had been placed subsequent to the induction of subarachnoid block. All but one of these patients were tracheally extubated at the conclusion of surgery. The remaining patient, undergoing gastroschisis repair, required postoperative mechanical ventilation. The mean dose of tetracaine for the 5 patients requiring intubation was 0.70 ± 0.14 mg/kg, which was significantly larger than the dose for the population as a whole when compared with Wilcoxon’s rank sum test (P = 0.03). However, using logistic regression, we found no significant association between the dose of tetracaine and the need for intubation (P = 0.09; odds ratio, 5.77; 95% confidence interval, 0.75–44.5).
Bradycardia, as defined by a heart rate <100 bpm, occurred in 24 of 1483 patients (1.6%). Fifteen patients received atropine or glycopyrrolate for the bradycardia. Three patients had bradycardia in association with the onset of a high spinal; of these patients, 2 received IV atropine. One received tracheal intubation and a brief period of cardiac chest compressions while atropine was circulating. The endotracheal tube was removed at the conclusion of surgery without sequelae. No patient received vasopressor treatment for hypotension. The dose of tetracaine for those patients with bradycardia was 0.60 ± 0.17 mg/kg, which was similar to the group as a whole (Wilcoxon’s rank sum test, P = 0.12).
The results of this series confirm the safety and efficacy of infant spinal anesthesia in a very large population and support the notion that it is clinically feasible. In previous studies, infant spinal anesthesia has been shown to have an infrequent incidence of complications and a high degree of cardiorespiratory stability (4–18,23,24). However, these studies involved only modest numbers of patients. In this large population, oxygen desaturation to <90% was rarely observed and <5% of patients required supplemental oxygen despite the fact that neonates and infants are notoriously prone to hypoxemia in response to external stressors. Moreover, the Vermont Infant Spinal Registry includes a wide variety of patients, including a number of smaller premature infants undergoing complicated surgical procedures. The infrequent rate of observed respiratory complications may have been a result of the inherent stability of spinal anesthesia in infants, our infrequent use of supplemental sedation, or a combination of both factors.
Anesthesiologists have long suspected that infants, particularly neonates, are at increased risk for adverse outcomes under anesthesia when compared with older children and adults (19–21). Recent data examining the incidence of cardiac arrest in children during the perioperative period have confirmed these suspicions (22). Consequently, there is strong interest in determining the safest method of administering anesthesia to this high-risk group. Studies with modest numbers of patients have suggested that when compared with general anesthesia, spinal anesthesia is associated with fewer cardiovascular and respiratory complications, less need for postoperative mechanical ventilation, and a shorter hospital stay (25–29). General anesthesia using short-acting IV or inhaled drugs could possibly provide a more rapid recovery profile with fewer complications. However, in a prospective study using sevoflurane in ex-premature infants, general anesthesia still had significantly more complications when compared with spinal anesthesia (28). Our study was not designed to compare the rate of complications between spinal anesthesia and general anesthesia. However, the very infrequent respiratory and cardiovascular complications we observed under spinal anesthesia is compelling.
Despite the overall stability of this technique, intraoperative complications did occur. A small number of patients required either bag mask ventilation or endotracheal intubation in association with a high spinal level. With the exception of an individual patient undergoing a complicated gastroschisis repair, the need for assisted ventilation in these patients was transient and no postoperative complications were observed. The incidence of bradycardia in the Vermont Infant Spinal Registry is more frequent than the 1.27% incidence reported by Keenan et al. (30) in 1994 among infants undergoing general anesthesia. However, these two groups are difficult to compare. The report by Keenan et al. (30) was a retrospective review with much more frequent morbidity and mortality associated with the bradycardia. Keenan et al. reported a rate of hypotension of 30%, asystole or ventricular fibrillation in 10%, and death in 8%. In contrast, the episodes of bradycardia seen within the Vermont Infant Spinal Registry were transient and without discernible sequelae. The occurrence of bradycardia was not predictable based on the dose of local anesthetic administered.
Concern has been expressed that spinal anesthesia in infants may be too technically difficult or time consuming to perform (28). However, successful LP and subarachnoid block were very frequent despite a wide variation in patient size and age, including small, premature infants undergoing major surgery early in the neonatal period and larger, healthy infants undergoing outpatient procedures. Furthermore, our experience demonstrates that anesthesiologists without fellowship training in pediatric anesthesia and even anesthesia trainees under supervision can perform infant spinal anesthesia with the expectation of frequent success. Likewise, infant spinal anesthesia can be performed efficiently. Our average time requirement of 10 minutes for induction of subarachnoid block and 27 minutes for total anesthesia and surgical preparation time are within accepted operating room guidelines and similar to the time required for induction of general anesthesia in neonates undergoing inguinal hernia repair (31,32). Because the University of Vermont is a teaching institution with surgical and anesthesia residents and medical students, these times are likely to be longer than in institutions without anesthesia and surgical trainees.
Infant spinal anesthesia is particularly time-efficient at the conclusion of surgery because there is no need for emergence, tracheal extubation, and stabilization before postanesthesia care unit (PACU) transport. The mean interval of 6 minutes between completion of surgery and PACU admission after infant spinal anesthesia compares favorably to infants emerging from either sevoflurane or propofol general anesthesia, who required an average time of 16 minutes between the end of surgery and eye opening during emergence (33). Furthermore, because previous studies have documented the very infrequent complications in the PACU with spinal anesthesia compared with general anesthesia, additional efficiency may be realized specifically in the PACU (25,26). Fewer complications related to postoperative hypoxemia, bradycardia, and apnea and the absence of emergence delirium could reduce delays in PACU discharge and the need for additional interventions from the anesthesiologist.
A potential limitation to the use of infant spinal anesthesia is a shorter duration of surgical anesthesia as compared with spinal anesthesia in adults. However, despite the wide variety of surgical procedures in the Vermont Infant Spinal Registry, only 1.2% of patients required conversion to general anesthesia as a result of insufficient duration of the subarachnoid block. Other modalities, such as combining the use of spinal and epidural anesthesia or the addition of intrathecal additives such as clonidine, show potential to significantly extend the duration of surgical anesthesia (5,14,34).
Our experience confirms that the incidence of serious complications associated with the use of spinal anesthesia in infants is very infrequent, even in small premature infants undergoing major surgical procedures. Anesthesiologists who care for infants should be aware of this option for anesthesia. Given the infrequent incidence of side effects of this technique, particularly in a population of infants at increased risk of complications, its use as an alternative to general anesthesia should be strongly considered for every infant undergoing lower abdominal or lower extremity surgery during the first 6 months of life.
The authors would like to gratefully acknowledge Taka Ashikaga, PhD and Mark Hoeft, BS for their assistance with statistical analysis and Mark Rockoff, MD for his insightful reviews of this manuscript.
1. Gray T. A study of spinal anesthesia in children and infants. Lancet 1909;3:913–7.
2. Junkin C. Spinal anesthesia in children. Can Med Assoc J 1933;28:51–3.
3. Melman E, Penuelas J, Marrufo J. Regional anesthesia in children. Anesth Analg 1975;54:387–9.
4. Abajian JC, Mellish R, Browne A, et al. Spinal anesthesia for surgery in the high risk infant. Anesth Analg 1984;63:359–62.
5. Williams R, McBride W, Abajian JC. Combined spinal and epidural anesthesia for major abdominal surgery in infants. Can J Anaesth 1997;44:511–4.
6. Williams R, Abajian JC. High spinal anesthesia for repair of patent ductus arteriosus in neonates. Pediatr Anaesth 1997;7:205–9.
7. Sartorelli K, Abajian JC, Kreutz J, et al. Improved outcome utilizing spinal anesthesia in high risk infants. J Pediatr Surg 1992;27:1022–5.
8. Vane D, Abajian JC, Hong A. Spinal anesthesia for primary repair of gastroschisis: A new and safe technique for selected patients. J Pediatr Surg 1994;29:1234–5.
9. Viscomi C, Abajian JC, Wald S. Spinal anesthesia for repair of meningomyelocele in neonates. Anesth Analg 1995;81:492–5.
10. Aronsson D, Gemery J, Abajian JC. Spinal anesthesia for spine and lower extremity surgery in infants. J Pediatr Orthop 1996;16:259–63.
11. Harnik E, Hoy G, Potolicchio S, et al. Spinal anesthesia in premature infants recovering from respiratory distress syndrome. Anesthesiology 1986;64:95–9.
12. Mahe V, Ecoffey C. Spinal anesthesia with isobaric bupivacaine in infants. Anesthesiology 1998;68:601–3.
13. Blaise G, Roy W. Spinal anesthesia in children. Anesth Analg 1984;63:1140–1.
14. Peutrell J, Hughes D. Combined spinal and epidural anesthesia for inguinal hernia repair in babies. Pediatr Anaesth 1994;4:221–7.
15. Parkinson S, Pecsok J, Little W. Use of hyperbaric bupivacaine with epinephrine for spinal anesthesia in infants. Reg Anesth 1990;15:86–8.
16. Hammer G. Regional anesthesia for pediatric cardiac surgery. J Cardiothorac Anesth 1999;13:210–3.
17. Somri M, Gaitini L, Vaida S, et al. The effectiveness and safety of spinal anaesthesia in the pyloromyotomy procedure. Paediatr Anaesth 2003;13:32–7.
18. Shenkman Z, Hoppenstein D, Litmanowitz I, et al. Spinal anesthesia in 62 premature, former premature or young infants-technical aspects and pitfalls. Can J Anaesth 2002;49:262–9.
19. Tiret L, Nivoche Y, Hatton F, et al. Complications related to anaesthesia in infants and children. Br J Anaesth 1988;61:263–9.
20. Cohen M, Cameron C, Duncan P. Pediatric anesthesia morbidity and mortality in the perioperative period. Anesthe Analg 1990;70:160–7.
21. Holzman R. Morbidity and mortality in pediatric anesthesia. Pediatr Clin North Am 1994;41:239–56.
22. Morray J, Geiduschek J, Ramamoorthy C, et al. Anesthesia related cardiac arrest in children. Anesthesiology 2000;93:6–14.
23. Dohi S, Naito H, Takahashi T. Age related changes in blood pressure and duration of motor block in spinal anesthesia. Anesthesiology 1979;50:319–23.
24. Oberlander T, Berde C, Lam K, et al. Infants tolerate spinal anesthesia with minimal overall changes: Analysis of heart rate variability in former premature infants undergoing hernia repair. Anesth Analg 1995;80:20–7.
25. Krane E, Haberkern C, Jacobson L. Postoperative apnea, bradycardia and oxygen desaturation in formerly premature infants: Prospective comparison of spinal and general anesthesia. Anesth Analg 1995;80:7–13.
26. Welborn L, Rice L, Hannallah R, et al. Postoperative apnea in former preterm infants: Prospective comparison of spinal and general anesthesia. Anesthesiology 1990;72:838–42.
27. Somri M, Gaitini L, Vaida S, et al. Postoperative outcome in high risk infants undergoing herniorrhaphy: Comparison between spinal and general anaesthesia. Anaesthesia 1998;53:762–6.
28. Williams JM, Stoddart PA, Williams S, Wolf A. Postoperative recovery after inguinal herniotomy in ex-premature infants: Comparison between sevoflurane and spinal anesthesia. Br J Anaesth 2001;86:366–71.
29. Kim G, Song J, Gwak M, Yang M. Postoperative outcome in formerly premature infants undergoing herniorrhaphy: Comparison of spinal and general anesthesia. J Korean Med Sci 2003;18:691–5.
30. Keenan R, Shapiro J, Kane F, Simpson P. Bradycardia during anesthesia in infants. Anesthesiology 1994;80:976–82.
31. Overdyk F, Harvey S, Fishman R, Shippey F. Successful strategies for improving operating room efficiency at academic institutions. Anesth Analg 1998;86:896–906.
32. Burge D. Operating and anesthetic time required for inguinal hernia repair in the neonate. Pediatr Surg Int 2003;19:646–8.
33. Cohen I, Finkel J, Hannallah R, et al. Rapid emergence does not explain agitation following sevoflurane anaesthesia in infants and children: A comparison with propofol. Ped Anesth 2003;13:63–7.
34. Rochette A, Raux O, Troncin R, et al. Clonidine prolongs spinal anesthesia in newborns: A prospective dose-ranging study. Anesth Analg 2004;98:56–9.