Lumbar Puncture Needle Depth


Recently, I prepared up to perform a lumbar puncture for the first time in a few months and a quiet voice at the back of my brain whispered ;

How deep would I need to go?
Which length needle would be the best?

I asked a handful of senior and junior colleagues, both at the time and in the writing of this post, and the response was almost universally “deep enough that the CSF comes out.” Certainly true, but not very pragmatic, and lacking the kind of detail I was hoping for…

I know there’s much discussion about the tip shape of a lumbar puncture needle, and in honesty, I’ve yet not read sufficiently to have strong opinions. However, in the fifteen minutes before the procedure, I had a look at the literature around needle length, and swiftly realized there was much more to this than I’d thought. Procedure finished, I was back to the drawing board.

Essentially, the balance is that a needle that is too short won’t reach the sub-arachnoid space, and a needle too long confers additional technical difficulty and increases the risk of going through.

So first, some basic anatomy; the aim of the exercise for lumbar puncture and CSF examination is to be in the sub-arachnoid space. To reach this space, the needle must pass through (in order) skin, superficial fascia, supraspinous ligament, interspinous ligament, ligamentum flavum, epidural space, dura mater and the arachnoid. I’m no neurosurgeon, but I’m pretty sure that it’s impossible to feel each of these layers on the end of the needle.

The anatomical target is either the L3/4 or L4/5 vertebral interspace, which respectively lie one vertebral body above & below the level of Tuffier’s line. Tuffier’s line is the imaginary line running between the superior iliac crests, and is used to demarcate the lower end of the spinal cord (which, in neonates, ends around L3 and moves superiorly with linear growth).

Finding a formula:
One of the more widely used formulas is from a 1997 paper where Craig et al. derived an elegant formula that;

 LP needle depth (cm) = 0.03 x height of child (cm).

Easily memorable and from a sample of 107 children receiving an LP with macroscopically clear CSF, the authors’ intention was a formula requiring only one variable that could be obtained in a critically unwell child – height being easily obtained with a measuring tape or Broselow tape.

In my department, the most common single measure recorded is weight; Bilic’s 2003 study of 195 Croatian children (over 3m of age) found the best correlate for LP depth was weight, using the formula

LP depth (cm) = 1.3 + (0.07 x Body weight (kg) )

The above formulae use a single variable and hence are probably more useful and pragmatic in the setting of an unwell child. Several other articles have discussed the most accurate formula for LP depth; all of which are reliant on at least two measured parameters. The following formulae may be more beneficial for elective CSF examination.

Several formulae were derived for LP depth from a cohort of 279 paediatric oncology patients in Malaysia; the best fit for their dataset was

y = 10 (weight (kg)/height (cm)) + 1

For this cohort, the LP depth was measured by perhaps a less reliable method than other datasets described, as the investigators measured the distance from their finger on the needle when pressed to the back at withdrawal. Irrespective, this paper summarizes many of the preceding papers in the discussion section.

Abe and foundation DFTB contributor Loren Yamamoto took a slightly different approach in a 2005 study; they reviewed 175 abdominal CTs to identify spinal canal depth at the iliac crest, deriving the formula of

LP depth (cm) = 1+ 17( weight/height).

Crucially, they went on to compare standard needle sizes to these depths to identify if the needle was too short or too long.


Defining the needle depth in this way has several benefits – firstly, it’s relatively prescriptive and secondly, it draws to attention the risks associated with using a needle that is too short (multiple punctures, anatomically impossible to reach the CSF), which amount to avoidable harm. In this context, it’s pertinent to know your tools. That is, identify which spinal needles are available in your department, their lengths and the type of tip.

LP needles are available in the following lengths (mm), depending on the brand, introducer, tip type: 25, 35, 38, 50, 64, 70, 75, 90, 103, 120, 150. Find the stock in your department  and see what’s there.



What about ultrasound?
The use of ultrasound to identify the depth of the spinal cord has been trialed in a number of papers; the two mentioned here were both produced from Addenbrooke’s Hospital in Cambridge, UK.

Firstly, in a neonatal population (105 neonates), weighing between 500g and 4500g, USS was used to measure median spinal cord depth (MSCD). They subsequently derived a formula of

LP depth (median spinal cord depth in mm) =  2(Weight) + 7 mm (R^2 0.76).


Subsequently, this nomogram was validated (albeit by the same author group and unit) in this study.

A later study by the same group undertook USS on 225 children aged 3m to 17 years presenting for echocardiography. The majority of patients were over 5 years of age. MSCD was identified as above, and a number of prediction models developed. The formula put forward by the group as satisfying the inherent tradeoff between accuracy (R^2 =0.72) and utility is

MSCD (mm)=0.4 W (kg)+20


So, does this change my practice? I will admit that I don’t have any of the above formulas fixed in my head, as yet.  Spinal needles in my hospital don’t have depth markings (it would be interesting to know if these exist). Instead, the above information serves to help in selecting a needle, particularly in those patients somewhere between neonate and adult sized. On this basis, I suspect I’m most likely to utilize formulae with weight as the single variable. I also went and re-read Ben Lawton’s post on champagne taps before the next one.


In summary;

  • Formulae are not yet in regular practice to identify needle depth for lumbar puncture.
  • We advocate increased awareness of the depth of the target structure, particularly when it comes to needle selection.
  • A needle can be too short, but it can’t be too long – it just becomes harder to use.


Print Friendly, PDF & Email


A Paediatric Trainee based in Queensland, Australia, Henry is passionate about Adolescent Medicine & General Paediatrics, with a strong interest in Medical Education & Clinical Teaching. An admitted nerd & ironman with a penchant for Rubik's Cubes & 'Dad jokes'.

@henrygoldstein | + Henry Goldstein | Henry's DFTB posts

Author: Henry Goldstein A Paediatric Trainee based in Queensland, Australia, Henry is passionate about Adolescent Medicine & General Paediatrics, with a strong interest in Medical Education & Clinical Teaching. An admitted nerd & ironman with a penchant for Rubik's Cubes & 'Dad jokes'. @henrygoldstein | + Henry Goldstein | Henry's DFTB posts

Leave a Reply