1University Department of Anaesthesia, Ninewells Hospital and Medical School, Dundee DD1 9SY, UK
Concerns about neurological sequelae of spinal (intrathecal) anaesthesia are almost as old as the technique, but are considered to be preventable by an understanding of all the possible mechanisms of harm and a careful technique.1 Spinal anaesthesia is now used frequently for a wide range of surgery on the lower half of the body, but it is difficult to know just how safe it is as few prospective data have been collected in recent years. A French survey did note the total incidence of neurological sequelae to be six in 10 000 spinal anaesthetics, with permanent deficit occurring in less than one in 10 000,2 although the prospective nature of this study has been questioned.3 However, there have been rumours emanating for some time from UK medico-legal circles of a disturbing number of cases of spinal cord damage associated with spinal anaesthesia. These rumours have been given substance by the publication by Reynolds4 of a series of seven patients, but with one recent estimate indicating that there may be at least 22 such.5 This may be a small number compared with the total number of spinal anaesthetics given, but the patients all have permanent and distressing symptoms. The nervous system has little capacity for repair, so it is vital that the causes are sought so that preventative strategies can be put in place.
The cases presented by Reynolds were collected over an 8-yr period, primarily through medico-legal work, and came from different parts of the UK. Six of the seven patients were obstetric (although that perhaps reflects her interest in that branch of anaesthesia), all received a pencil point needle (2527 swg) and four involved the combined spinal/epidural (CSE), needle through needle technique. Difficulty in siting the block was documented by the anaesthetist in only two of the cases, and one patient was noted to be obese. All seven patients felt pain on insertion of the needle (one on subsequent injection of local anaesthetic solution), neurological symptoms appeared as the block regressed, and six of the seven patients had magnetic resonance imaging (MRI) evidence of a lesion in the spinal cord at a level consistent with the clinical features. However, lesions were not consistent with the level of needle insertion documented in the case notes by the anaesthetists involved, and Reynolds has drawn attention to the frequency with which the level of insertion is identified inaccurately.6 Thus, she draws the conclusion that the spinal needle was inserted at an inappropriately high level and that it caused direct damage to the spinal cords of these unfortunate patients. Association does not equal causation, but the evidence of a causative link in this instance is compelling, if not absolute.
That accurate identification of vertebral level is difficult clinically is undeniable. Broadbent and colleagues7 studied the ability of four experienced anaesthetists to do this in the lumbar region. Two independent assessments were made in 100 patients adopting a flexed position, either sitting or left lateral. Skin markers were applied and then an MRI scan was obtained. The results were sobering. There was reasonable agreement between anaesthetists, but only 29% of the assessments were accurate, the true level being higher than that identified in a substantial majority: one segment higher in 51% of estimates, two in 15.5%, three in 1%, and four in 0.5%. Accuracy was not affected by the position of the patient or the experience of the anaesthetist, but it was impaired by patient obesity. However, certain factors may have exaggerated the results. The patients were assessed in the flexed position, but were scanned in a neutral position. This difference may, in itself, be important and the change in posture is likely to have changed the relative positions of the skin markers and the underlying bony structures. Previous studies, using spinal catheters or myelography needles to mark the level, have produced slightly better results, but still with inaccurate identification rates of 50 and 59%.8 9 The potential consequences of this are greater if the spinal cord ends lower than usual. Broadbent and colleagues7 noted that the conus medullaris was below the first lumbar vertebra in 19% of patients, and anatomical studies suggest that the figure may be higher.10 Indeed in some normal individuals the conus may be as low as the second lumbar interspace.
This means that the combination of a higher than intended level of needle insertion in a patient with a spinal cord terminating at a lower level than normal could well lead to problems. Broadbent and colleagues made the entirely reasonable recommendation that If a choice of suitable interspaces exists, the lower one should be selected for intrathecal injection, to reduce the risk of neurological damage from either misidentification of the vertebral level or an unexpectedly low conus. Reynolds has taken this a step further and advocated an approach for the prevention of these problems which requires that clinicians take greater care to ensure that lumbar puncture is performed below the termination of the spinal cord by using an interspace no higher than the third lumbar. The level of needle insertion is very important but is it the only issue? Limiting spinal anaesthesia to the third lumbar interspace or below would deny the technique to many older patients in whom those spaces are ankylosed, and, as the evidence shows, a clinician can never be certain which level is actually being used anyway. In addition, and because of the uncertainty regarding level, a significant proportion of spinal anaesthetics must be performed above the level of the conus, but there are only a few recorded cases of actual spinal cord damage. Indeed, Reynolds estimates the incidence of the problem to be about one in 20 000.11 A single error rarely leads to an accident, and it is more usual for there to be several contributory factors, a combination of active and latent errors.12 The active error here is the needle making contact with the spinal cord, but what are the relevant latent errors and safety mechanisms? To answer this we have to consider the factors that are important in the insertion of spinal needles: the patient, the equipment, and the technique.
The patient
First, there are factors that affect the assessment of level. Broadbent and colleagues identified obesity as an important factor, and they suggested that vertebral collapse and age might also be important. Additionally, if a patient is tense and the back musculature contracted, it can be very difficult to feel any bony landmarks and thus establish the true vertebral level. Secondly, there are factors that affect the level of termination of the spinal cord. As noted, there is some normal variation in spinal cord length, and it may be abnormally long if tethered as in conditions such as diastomyelia and some forms of spina bifida. Further, it seems reasonable to assume that the conus may alter its level slightly with changing position, and it has been claimed that flexion of the neck and trunk causes the cord to move cranially within the vertebral canal.13 Thirdly, the depth of the CSF-filled space between the arachnoid mater and the spinal cord varies. The lumbar section of the cord is enlarged by an increased amount of grey matter relating to the innervation of the lower limbs, but below that it tapers to almost nothing. Where the space is wider there is some safety margin, even if the needle has been inserted level with the spinal cord, but only millimetres are involved and the lumbar enlargement is very close to the arachnoid.
The equipment
All the patients described by Reynolds had a pencil point spinal needle inserted. These are now widely used to minimize the incidence of post-dural puncture headache, which is much higher with the traditional cutting type spinal needles. The fact that all the patients received such a needle may just be a reflection of their widespread current use, but certain issues must be considered. The increase in the use of spinal anaesthesia in the UK pre-dated the ready availability of pencil point needles by many years, but reports of spinal cord problems with cutting needles are very rare, although not unheard of.14 Without good comparative data it is impossible to say whether one type of needle is more likely to be associated with spinal cord problems, but there is some supporting information. Concern about the safety of pencil point needles was voiced as early as 1993 when Turner and Shaw15 described a high incidence of paraesthesiae. Their frequency has been noted to be as high as 12% with standard spinal anaesthesia,16 and 26% with the CSE technique,17 figures which are higher than had been documented previously with cutting needles. Paraesthesiae experienced on needle insertion indicate contact with either the spinal cord or the nerve roots of the cauda equina, and have been shown to increase significantly the likelihood of subsequent neurological deficit.18
Pencil point needles are often referred to as atraumatic, so why might they be more likely to cause such problems? A number of factors may be relevant:
1. A greater length of needle has to be inserted into the subarachnoid space to allow CSF to flow through the orifice because it is situated approximately a millimetre from the actual tip.
2. It is a misnomer to call these needles atraumatic, electron microscopy evidence having shown that a pencil point needle actually causes more damage to a membrane than a cutting one.19 The ragged hole left by a pencil point needle probably leaks less CSF, but the damage may be greater if the tip contacts the spinal cord. There is a parallel with work on peripheral nerve injury which showed that a short bevel needle is less likely to lead to nerve contact than a long bevel one, but that the damage will be greater if there is any contact.20 A clean cut may be less harmful than a contusion.
3. The blunt nature of the tip may require that slightly greater force is used to advance the needle, and this may increase the risk of precipitant entry into the subarachnoid space, with the risk of overshoot.
4. The blunt tip has also been shown to be more likely to tent the dura mater before penetrating it, and thus narrow the space between the arachnoid mater and the nervous tissue.21
The association of these spinal cord problems with CSE is even more tenuous, but the high incidence of paraesthesiae in some reports of its general use is a cause for concern. It is a relatively complicated technique, which requires significant instrumentation of a very confined space and is associated with a number of problems.22 Two components may make needle contact with nervous tissue more likely. First, the initial placement of an epidural needle may mean that the dura mater is tented inwards even before the spinal needle is inserted. Secondly, the insertion of the spinal needle through that epidural needle means that there is little or no tissue resistance to slow its insertion. Precipitant needle entry, with risk of overshoot, may be more likely particularly when the technique is being learnt.
The technique
Clearly every attempt must be made to identify a level of needle insertion that is below the termination of the spinal cord. However, this cannot be guaranteed so every effort must be made to minimize the risk of contact with the spinal cord should it be at the level of needle insertion.
It might be thought that use of the sitting position might confer some protection against a needle that has been inserted at too high a level making contact with the spinal cord. The increased CSF pressure will speed the flow of CSF through the needle and allow easier identification of correct placement, and it will distend the dural sac to increase the depth of the space between arachnoid and cord. However, four of Reynolds1 seven patients were managed in the sitting position, so this does not appear to provide any significant protection.
Cranial angulation of the spinal needle will result in the tip entering the subarachnoid space at an even higher level than it is inserted, so increasing the risk of contact with the spinal cord. However, if the patients back is properly flexed, the needle does not need such cranial angulation and can be inserted perpendicular to the skin, so avoiding this component of risk.
Gentle, slow insertion of the needle must be an essential component of every spinal anaesthetic, with the needle being inserted no further than is necessary to ensure free flow of CSF. Care must be taken not to advance the needle further when attaching the syringe containing the local anaesthetic. Although spinal cord damage has occurred after careful needle insertion, it is clearly more likely to occur if a heavy handed approach is adopted. The safe performance of a spinal anaesthetic requires an unhurried approach, free of distractions, with a heightened awareness of specific situations where errors are more likely to occur.
Conclusions
It must be emphasized that much of the above discussion is based on first principle analysis and circumstantial evidence rather than hard fact. However, it is important to make every effort to understand what might have contributed to the problems affecting these patients, because we must work to prevent others from being harmed for life. It may be that, on occasion, certain factors come together to make neurological damage that little bit more likely occur:
1. difficulty in identification of spinal level;
2. inadequate flexion of the spine;
3. use of pencil point needles; and
4. an incautious insertion technique.
In regard to this latter component, one final point is worth considering. All of Reynoldss patients experienced pain at the moment of needle insertion. Does this imply direct contact with fibres, which enter the cord through the dorsal roots and relay in the posterior horn, or with their secondary neurones, which cross the midline to ascend in the anterolateral spinothalamic tract? A spinal needle would have to traverse the posterior columns before reaching these structures, really implying an unnecessarily vigorous technique. No one should perform a spinal anaesthetic without an acute awareness that, on the far end of the needle, there may be somebody elses nervous system.
References
1 Wildsmith JAW, Lee JA. Neurological sequelae of spinal anaesthesia. Br J Anaesth 1989; 63: 5057[ISI][Medline]
2 Auroy Y, Narchi P, Messiah A, Litt L, Rouvier B, Samii K. Serious complications related to regional anesthesia. Anesthesiology 1997; 87: 47986[ISI][Medline]
3 Price JM, Carpenter RL. Anesthesia safety, outcomes, and prospective study design. Anesthesiology 1998; 89: 7901[ISI][Medline]
4 Reynolds F. Damage to the conus medullaris following spinal anaesthesia. Anaesthesia 2001; 56: 23547[ISI][Medline]
5 Reynolds F. Spinal anaesthesia and conus damage. Anaesthesia 2001; 56: 690713
6 Reynolds F. Logic in the safe practice of anaesthesia. Anaesthesia 2000; 55: 10456[ISI][Medline]
7 Broadbent CR, Maxwell WB, Ferrie R, Wilson DJ, Gawne-Cain M, Russell R. Ability of anaesthetists to identify a marked lumbar interspace. Anaesthesia 2000; 55: 110626[ISI][Medline]
8 Moore DC. Effects of speed and duration of injection on the level and duration of spinal anesthesia. Anesth Analg 1970; 49: 9128[Medline]
9 VanGessel EF, Forster A, Gamulin Z. Continuous spinal anesthesia: where do catheters go? Anesth Analg 1993; 76: 10047[Abstract]
10 Reimann AF, Anson BJ. Vertebral level of termination of the spinal cord with report of a case of sacral cord. Anatomical Rec 1944; 88: 12738
11 Reynolds F. Damage to the conus medullaris following spinal anaesthesia. Anaesthesia 2001; 56: 814
12
Arnstein F. Catalogue of human error. Br J Anaesth 1997; 79: 64556
13 Romanes GJ. The central nervous system: the spinal medulla. In: Romanes GJ, ed. Cunninghams Textbook of Anatomy, 12th Edn. Oxford: Oxford University Press, 1981; 61927
14 Greaves JD. Serious spinal cord injury due to haematomyelia caused by spinal anaesthesia in a patient treated with low-dose heparin. Anaesthesia 1997; 52: 15068[ISI][Medline]
15 Turner MA, Shaw M. Atraumatic spinal needles. Anaesthesia 1993; 48: 452
16 Hopkinson JM, Samaan AK, Russell IF, Birks RJS, Patrick MR. A comparative multicentre trial of spinal needles for Caesarean section. Anaesthesia 1997; 52: 9981014[ISI][Medline]
17 Turner MA, Reifenberg NA. Combined spinal epidural anaesthesia. The single space double-barrel technique. Int J Obstet Anesth 1995; 4: 15860
18 Horlocker TT, McGregor DG, Matsushige DK, Schroeder DR, Besse JA, the Perioperative Outcomes Group. A retrospective review of 4767 consecutive spinal anesthetics: central nervous system complications. Anesth Analg 1997; 84: 57884[Abstract]
19 Reina MA, de Leon-Casasola OA, Lopez A, De Andres J, Martin S, Mora M. An in vitro study of dural lesions produced by 25-guage Quincke and Whitacre needles evaluated by scanning electron microscopy. Reg Anesth Pain Med 2000; 25: 393402[ISI]
20 Rice ASC, McMahon SB. Peripheral nerve injury caused by injection needles used in regional anaesthesia: influence of bevel configuration, studied in a rat model. Br J Anaesth 1992; 69: 4338[Abstract]
21 Dittman M, Lehmann K. Physical properties (diameter, flowrate and tip geometry) of the latest spinal needles and their effect on tenting. In: Van Zundert A, ed. Highlights in Regional and Pain Therapy. IV. Barcelona: Permanyer Publications, 1995; 1124
22 Wildsmith JAW. Problems with combined spinal and epidural anesthesia. Reg Anesth Pain Med 1998; 23: 3889[ISI]