A comprehensive, bed-side scoring system to predict difficult lumbar puncture

Abstract

Background and Aims:

Spinal anesthesia (SA) is the most widely practiced neuraxial anesthesia. Lumbar puncture (LP) at multiple levels and multiple attempts due to any reason may cause discomfort and even serious complications. Hence the study was conducted to evaluate the patient variables that can predict difficult LP thus allowing for the use of alternate techniques.

Material and Methods:

We included 200 patients of ASA physical status I-II, scheduled to undergo elective infra-umbilical surgical procedures under spinal anesthesia. During preanesthetic evaluation, difficulty score was assessed using the 5 variables: Age, abdominal circumference, spinal deformity - assessed as axial trunk rotation (ATR) value, anatomical spine assessed by spinous process landmark grading system (SLGS) and patient position, by assigning a score of 0- 3 for each variable, with a total score of 0 – 15. The difficulty of LP was graded as easy, moderate or difficult based on total number of attempts and spinal levels by independent experienced investigator. The scores obtained during preanesthetic evaluation and the data collected after performing LP were analyzed using multivariate analysis and P value noted.

Results:

Our study showed that above patient variables correlated well with difficult LP scoring (P < 0.001). SLGS was noted to be a strong predictor, while ATR value a weak predictor. The correlation between the total score and grades of SA had a positive association (R = 0.6832, P < 0.00001) and was statistically significant. A median difficulty score of 2, 5 and 8 predicted easy, moderate and difficult LP respectively.

Conclusion:

The scoring system provides for a useful tool to predict difficult LP and helps both patient and anesthesiologist to choose an alternative technique.

Introduction

Spinal anesthesia (SA) has a wide range of clinical applications.^[1] It is widely used for patients undergoing infra-umbilical surgeries of known duration. It is a relatively safe technique with numerous advantages but when a difficult lumbar puncture (LP) is encountered it leads to patient discomfort and increases the incidence of postdural puncture headache, spinal hematoma, and nerve damage.^[2] Although a commonly practiced procedure, discomfort experienced by patients due to multiple attempts during difficult LP for needle placement is inevitable.^[3]

Hence, a scoring system that evaluates the probability of the difficulty encountered while performing LP is crucial. Although few studies^[2,4,5] have been done to assess the variables that predict the difficulty during LP, they have shown inconsistent results. Only Gvalani et al.^[5] showed a significant relationship between age and difficulty encountered during LP. Some of the studies^[2,4] had used radiological characteristics as a predictor which restricts its applicability in a broader population due to radiation exposure and cost.

We hypothesize that there is a positive correlation between five patient variables based on scoring and the level of difficulty in performing LP. We conducted this study with a research question that; whether the patient variable-based bedside scoring will help in predicting the difficulty in LP in patients above 20 years of age coming for infraumbilical elective surgeries under SA. The primary outcome of the study was to assess grading of difficult LP into easy, moderate, and difficult.

Hence, this study was designed to develop a comprehensive, bedside scoring system to predict the difficult LP with patient variables that are more safe, objective, validated, and cost-effective.

Material and Methods

After obtaining the institutional ethics committee approval (reference EC - 474/dated 14/11/17), a prospective, observational study was conducted at our tertiary care hospital, over a period of one year from January 2018 to January 2019. Two hundred patients who were posted for elective surgeries under SA involving surgical, urology, gynecological and orthopedic surgeries, aged 20 years and above, belonging to American Society of Anesthesiologist (ASA) physical status I and II were included in the study [Figure 1]. Any patient who did not give written informed consent, pregnant patients and those who were unable to perform Adam’s forward bend test were excluded from the study.

Figure 1.

STROBE diagram

Based on the previous study,^[4] prevalence of difficult spinal was 14.9%; taking level of confidence of 1.96 and 5% absolute precision, sample size was estimated to be 196 patients. We included 200 consecutive eligible patients coming for elective surgeries based on inclusion criteria.

During the preanesthetic evaluation, following five patient variables were assessed and each scored 0 to 3, as shown in Table 1:

Table 1.

Pre - operative assessment tool

Patient variable	Score				Assigned Score
	0	1	2	3
Age (years)	20-39	40-59	60-69	>70
Abdominal circumference (cm)	<80	80-99	100-120	>120
Anatomical spine assessment	Spinous process visible	Spinous process palpable and not visible	Spinous process not visible and not palpable, but interval between them is palpable as a low landmark under the thumb	None of the previous case
Spine deformity (ATR value)	No (<50)	Mild (5-7.50)	Moderate (7.5-100)	Severe (>100)
Patient position/Attitude	Can sit with flexed back (“C” shaped AP curvature)	Lateral position with flexed back	Sitting with stiff back	Right or left lateral position with stiff back
Difficulty score

ATR: Axial Trunk Rotation

1. Age (years),

2. Abdominal circumference (cm)^[6] was measured using a non-stretchable measuring tape at the level of umbilicus at the end of expiration with the patient in standing position.

3. Anatomical spine assessment was done by making the patient to sit comfortably with adequate exposure of spine and assessed using ‘spinal landmark grading system’^[7] (SLGS) and scored accordingly.

4. Spine deformity was assessed in terms of “Axial trunk rotation” (ATR) value and measured using the smartphone application ‘Scoliometer’.^[8-15] Patients were asked to perform Adam’s forward bending test first and then smartphone was placed on patient’s back with the centre mark on the summit of the lumbar spine in midline, subsequently moving it down over the vertebral column [Figure 2]. The value shown at each level was noted [Figure 3]. At least two values were recorded at each level and maximum degree of the ATR obtained was noted and a score was assigned.

5. Patient position was scored based on the attitude assumed by the patient that was comfortable for LP.

Figure 2.

(Original) Measuring ATR value using scoliometer in midline

Figure 3.

(Original) Display of ATR value in scoliometer smart phone app

By summating the scores of these five patient variables, a difficulty score was obtained (total score of 0 – 15).

All patients in the study followed standard hospital protocol and ASA fasting guidelines before taking up for surgery. An 18G intravenous line was secured and patients were preloaded with ringer lactate at 10 ml/kg. Standard monitoring devices like an electrocardiogram, noninvasive blood pressure and pulse oximeter were used for monitoring. An anesthesiologist with more than 5 years’ experience who was not involved in the preanesthetic assessment of patient variables performed the LP. Under aseptic precautions, the spinal space was identified by landmark palpation technique. Best palpable space was identified, lidocaine 1% infiltrated and LP performed using 25G Quinke’s needle and bupivacaine heavy 0.5% was injected. The visibility of cerebrospinal fluid (CSF) was considered as an end point. Using 2 outcome variables (total number of attempts and number of spinal levels attempted) the difficulty of LP was graded into easy, moderate and difficult. Each new skin puncture was considered as a new attempt. The first attempt done in a spinal space identified was considered first spinal level. Number of spinal levels required for a successful LP were noted. Not more than two attempts were allowed in one spinal level. Based on two variables i.e., number of spinal level and number of attempts the difficulty of LP was graded as follows:^[4]

• Easy – Free flow of CSF was obtained in first attempt.

• Moderate – Free flow of CSF was obtained after trying for more than one attempt in one spinal level or shifting to next spinal level. (2-4 attempts)

• Difficult – Free flow of CSF obtained after trying 5 or more attempts or shifting to third spinal level. (5 – 7 attempts)

The patients were then put to supine position and assessed for achievement of adequate sensory blockade by pin prink test and motor blockade based on modified bromage scale before the start of surgery.

Statistical analysis

Data were analyzed using Statistical Package for the Social Sciences (SPSS) version 22 software. Categorical data was represented in the form of frequencies and proportions. Continuous data were represented as mean and standard deviation. Chi-square test was used as test of significance for qualitative data. Independent t test or Mann Whitney U test was used as test of significance to identify the mean difference between two quantitative variables and qualitative variables respectively. ANOVA (analysis of variance) or Kruskal–Wallis test was the test of significance to identify the mean difference between more than two groups for quantitative and qualitative data respectively. Pearson correlation or Spearman’s correlation was done to find the correlation between two quantitative variables and qualitative variables respectively. P value of <0.05 was considered as statistically significant after assuming all the rules of statistical tests.

Results

A total of 200 patients participated in this study. Demographic details of these patients are provided in Table 2. The patients participated were from various surgical departments; General surgery and urology (n = 123, 61.5%), gynecology surgery (n = 57, 28.5%) and orthopedics (n = 20, 10%). Correlation of five patient variables scoring with different grades of LP is displayed in Table 3.

Table 2.

Demographic details of the study populationmographic details of the study population

Variables	Results
Gender
Male	97 (48.5%)
Female	103 (51.5%)
Age (years)
20-39 years	89 (44.5%)
40-59 years	75 (37.5%)
60-69 years	20 (10%)
>70 years	16 (8%)
Range	20-95 years
Abdominal circumference (cm): (Range)	68 cm-155 cm

Table 3.

Comparison of patient variables with different grades of LP

Patient variables	Gradation of variables	Score	Grades of LP			Chi Square
			Easy n	Moderate n	Difficult n
Age (years)	20-39	0	70	15	2	χ2=26.693,
	40-59	1	48	20	9	df=6,
	60-69	2	9	7	4	P<0.001
	>70	3	4	9	3
Abdominal circumference (cm)	<80 cm	0	8	0	0	χ2=63.195,
	80-99	1	109	31	4	df=6,
	100-120	2	14	20	12	P<0.001
	>120 cm	3	0	0	2
Spinal Deformity (ATR value)	No Deformity	0	116	33	10	χ2=26.420,
	Mild	1	13	13	4	df=6,
	Moderate	2	1	3	3	P<0.001
	Severe	3	1	2	0
Anatomical spine assessment (SLGS)	Spinous process visible	0	57	8	0	χ2=72.925,
	Spinous process palpable and not visible	1	53	14	3	df=6,
	Not visible and not palpable	2	19	28	10	P<0.001
	None of the previous	3	2	1	5
Patient position/Attitude	Sitting position with flexed back	0	118	38	4	χ2=67.041,
	Lateral position with flexed back	1	9	1	1	df=6,
	Sitting with stiff back	2	4	10	14	P<0.001
	Lateral position with stiff back	3	0	0	3

Age: Analysis showed a positive correlation (R-value = 0.6825) of age with the difficulty score, indicating that as age increases the difficulty score also increases and vice versa. It was noted that the LP was easy in patients aged <60 years, while the difficulty increased in the patients who were aged >60 years. Age had a weakly positive correlation with different grades of LP (R-value = 0.3479).

Abdominal circumference: Majority of the patients with abdominal circumference <100 cm had easy and moderate LP, while the patients with >100 cm had difficult LP. The correlation between the abdominal circumference, difficulty score and different grades of LP were statistically significant ((P < 0.001). A positive correlation was noted between number of attempts (R = 0.450, P < 0.001) and number of spinal level (R = 0.421, P < 0.001).

Spinal deformity: Correlation of spinal deformity with the difficulty score (z-score 15.81294, P < 0.00001) and grading of LP (z-score 14.12212, P < 0.00001) had normal distribution and was statistically significant indicating that severity of spinal deformity has an impact on the outcome. Pearson’s correlation coefficient for spinal deformity with number of attempts (R = 0.330, P < 0.001) and number of spinal levels (R = 0.311, P < 0.001) shows a weakly positive correlation indicating that with an increase in ATR values there was an increase in number of attempts and number of spinal levels approached for LP.

Anatomical spine assessment: Observation of the anatomical characteristics of the spine using SLGS revealed that patients with visible and palpable spinous process had an easy and moderate LP, while the rest with spinous process not visible or palpable had difficult LP. A good positive correlation was noted between SLGS with difficulty score (R = 0.7004, P < 0.00001) and grading of LP (R = 0.5116, P < 0.00001). By normal standards of spearman’s rho correlation, the association between SLGS and difficulty score was considered statistically significant (r_s = 0.73618, P (2-tailed) =0) and z-score is 12.29 with P value < 0.00001, showing that the higher the values of SLGS, greater the difficulty of LP.

Patient position: With regards to patient position, majority of the patient with flexed back either in sitting or lateral position had easy LP, whereas the patients with straight/stiff back in sitting or lateral position had difficult LP. A moderate positive correlation was noted between patient position and difficulty score (R = 0.652, P < 0.0001) and grades of LP (R = 0.5226, P < 0.00001). By spearman’s rho calculation, the association between the two variables considered was statistically significant. The decrease in flexibility increases the difficulty of LP.

Difficulty score: The summation of the scores for each variable gave an assigned total score for every patient. In our study, the overall score ranged from 0 to 15, but the analysis showed the lowest score was 0 and the highest score observed was 10. The correlation between the difficulty score and grades of LP had a positive association (R = 0.6832, P < 0.00001) and statistically significant; showing that the increase in the difficulty score, increases the difficulty of LP and vice versa [Table 4]. Amongst all the variables, SLGS had a strong correlation whilst ATR had a weak correlation with the grades of LP. The median score for each variable was obtained for the different grades of LP [Table 5] and we found that the difficulty score of 2, 5 and 8 predicted easy, moderate and difficult LP. This also shows that a difficulty score of more than 4 increases the difficulty of LP [Table 5]. The median score for the different variables was validated and analysis showed it was statistically significant.

Table 4.

Pearson’s correlation between patient variables with total scores and grades of LP

Variables	Total score	Grades of LP	P
Age	0.6992	0.3479	<0.001
Abdominal circumference	0.5859	0.5012	<0.0001
Spinal deformity (ATR value)	0.4682	0.3015	<0.001
Anatomical spine assessment (SLGS)	0.7004	0.5116	<0.00001
Patient position	0.6546	0.5226	<0.0001

SLGS: Spinal Landmark Grading system, ATR: Axial Trunk Rotation

Table 5.

Comparison of median scores of patient variables with grades of LP

Variables	Grades of LP			P
	Easy	Moderate	Difficult
Age	0	1	1	<0.001
Abdominal circumference	1	1	2	<0.001
Spinal Deformity (ATR)	0	0	0	<0.001
SLGS	1	2	2	<0.001
Patient position	0	0	2	<0.001
Total	2	5	8	<0.001

LP in all patients receiving SA was successful and did not require additional analgesia or change of technique for completion of surgery.

Discussion

Our study has demonstrated a positive correlation between age, abdominal circumference, spinal deformity, anatomical spine assessment, patient position and the difficulty score during univariate analysis. The difficulty score had significant correlation with the different grades of LP and can predict the difficulty of LP.

The predictors used for the scoring in the previously published studies^[2,4] include: Age, BMI, spinal bony landmark, spinal bone deformity, radiological characteristics and past history of spinal surgery or difficult spinal encountered. But the lacunae in the existing literature were the predictability of the variables used as they were subjective, with different ethnic background, included only certain population and had inconsistent results.

Age that has been used as a difficulty predictor showed a positive correlation with the final score and was statistically significant, a finding that was similar to Gvalani SK et al.^[5] and Ružman et al.^[16] indicating that age increases difficulty of LP, as the prevalence of osteoporosis, degenerative and other pathological processes of spine increase with age.

Various patient factors are involved in making the location of correct lumbar space difficult, one such factor being abdominal circumference. We observed the association between abdominal circumference and difficulty of LP while previous studies^[2,5] had evaluated weight and BMI. Snider et al.^[6] stated that the weight and BMI do not affect the radiographic tuffier’s line but found that the palpated bony anatomical projections (tuffier’s line) were affected by fat and muscle distribution. However, Lin et al.^[17] demonstrated a correlation between abdominal circumference and lumbar flexion. According to some studies,^[18,19] palpation of bony landmarks was difficult with abundant intervening tissue, and the distance from the skin to subarachanoid space had a correlation in performing a neuraxial blockade.

Atallah et al.^[2] and Gvalani et al.^[5] had reported that the quality of the spinous process and interspinous space was associated with technical difficulties of the neuraxial blocks and suggested that it was a sensitive and strong independent predictor of technical difficulty. The present study also had a strong correlation between the anatomical characteristics of bony landmarks and the difficulty score. Our study used ‘spinous process landmark grading system’ an objective, validated, anatomical characteristic scoring system which was used previously in studies conducted by Chein et al.^[7].

Spinal deformity was also an important factor for prediction of difficulty in performing neuraxial anesthesia. In this study, we used “Axial Trunk Rotation” (ATR) value to assess the spinal deformity using smart phone app ‘Scoliometer’^[8-15]. ATR value had a positive correlation with the difficulty score. Coelho DM et al.^[8] had demonstrated excellent intra-rater reliability and very good inter-rater values when he observed the correlation between the ATR scoliometer measurements and Cobb’s angle measurements. The method proposed in their study measured the reliability of all vertebral levels of the thoracic and lumbar spine, and proved that regardless of the vertebral level and magnitude of the patient’s ATR, the scoliometer measurement is reliable. Atallah et al.^[2] utilized X-rays to demonstrate difficult LP with radiological characteristics (like osteophytes, ligament calcification or narrow intervertebral spaces) and concluded it was a strong independent predictor of difficult LP.

Ružman et al.^[16] revealed that adequacy of patient position was a predictor of successful neuraxial block and found that sitting position (87.3%) was associated with better first puncture success. The findings in our study were comparable, with 90.1% first puncture success seen in patients who assumed sitting position with flexed back while performing neuraxial blockade and demonstrated that patient position had a correlation with predicting difficulty. The probable explanation for this finding would be due to better opening of the interspinous space. The decrease in flexibility of the spine due to degenerative changes would be probable reason for straight back.^[17,20] We also observed that sitting or lateral position with straight back was associated with increasing difficulty.

The summation of the score for all the patient variables in each patient gave the final score or difficulty score. Atallah et al.^[2] and Gvalani SK et al.^[5] stratified the difficulty score into nine grades (0–8), a score ≥ grade 4 was indicative of difficulty of LP in terms of increase of number of attempts and or number of spinal levels. In our study, we could develop and validate the scores and stratify the final difficulty score from 0 to 15; a score <3 indicated easy LP, 4 – 6 indicated moderate LP and ≥7 indicated difficult LP. This implies that a score ≥4 was associated with increase in difficulty in LP, a value that is comparable to the other studies.

Our study had a few limitations, we require large multicentric trials involving large population to further validate and confirm our study results; these predictors may not be applicable to pregnant patients and patients below 20 years of age; though the anesthesiologist performing the LP was unaware of patient variables, it is difficult to avoid observer bias completely; the trial was not registered in the Clinical trial registry.

Conclusion

We conclude that the clinically measurable, variable based comprehensive bedside scoring system will be beneficial in predicting difficult LP and thus help in reducing the incidence of multiple attempts. Further, it is confirmed that SLGS is a strong predictor while ATR value is a weak predictor of difficult LP.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.