Retrospective analysis of predisposing factors for difficult lumbar punctures requiring image guidance in pediatric oncology patients

Abstract

Success rates of lumbar punctures (LPs) in children are reportedly as low as 50%. In addition to procedural complications and failure, difficult LPs are a risk factor for traumatic LPs (TLPs), which can potentially affect diagnostic utility and alter treatment plans for pediatric oncology patients. To identify the intrinsic factors associated with technically difficult LPs in the pediatric oncology population, we performed a retrospective review of patients who required diagnostic imaging modalities for LP procedures at a single pediatric oncology institution between September 2008 and November 2018. We evaluated data from 64 LPs performed in 33 patients who were referred for image-guided LPs after undergoing technically difficult LPs that were unsuccessful using anatomic landmarks. In these cases, 96.9% of patients had at least one of the following intrinsic factors: body mass index (BMI) ≥ 25, anatomic spinal abnormalities, history of ≥ 5 previous LPs, age < 12 months, and history of back surgery. Elevated BMI was the most common factor associated with difficult LP (81.8%), followed by spinal abnormalities (51.5%), and history of ≥ 5 previous LPs (33.3%). Age < 12 months and history of back surgery were also associated with difficult LPs, but at a lower frequency. On the basis of these findings, we propose clinical recommendations for preprocedural identification of patients at risk of difficult LPs to reduce complications, including TLP, failure, and exposure to general anesthesia.

1. Introduction

Lumbar puncture (LP) is a relatively safe and common procedure used for diagnostic and therapeutic purposes in pediatric oncology. Avoiding traumatic LPs (TLPs), defined as > 10 red blood cells per microliter of cerebral spinal fluid (CSF), is important because TLPs can potentially obscure the diagnostic utility of the procedure. This is particularly important for pediatric oncology patients, who receive frequent LPs for collection and examination of CSF for malignant cells and intrathecal chemotherapy administration. TLPs in children with acute lymphoblastic leukemia (ALL) are associated with worsened survival outcomes, and patients with two successive TLPs experience event-free survival outcomes similar to those of patients with overt central nervous system disease.¹ TLPs can not only obscure diagnostic utility but may also introduce leukemic cells into the CSF or cause inadequate administration of intrathecal chemotherapy. For these reasons, newer institutional ALL treatment protocols treat patients with a history of TLPs (with blasts evident in CSF) similarly to that for patients with overt CNS disease.¹ Because these patients are most likely to receive more LPs than are their counterparts without a history of TLPs, minimizing the incidence of TLPs is paramount, especially for patients experiencing their first LP.

The risk factors for TLPs include obesity, infancy, age > 10 years, African American race, history of TLPs within the prior 2 weeks, previous TLPs with platelet count less than 50,000/mL, difficulty visualizing/palpating spinous processes, advancement of needles with stylets in situ, lack of anesthesia/sedation, and platelet counts < 100,000/mL at the time of procedure.^2–4 The role that operator experience plays in TLPs is uncertain, but operator inexperience is also apparently a risk factor for TLPs.^3,5 Many of these risk factors are considered intrinsic or nonmodifiable, such as patient age, body-mass index (BMI), race, and spinous process visualization/palpation.^3,4 Glatstein et al. suggest that the “occurrence of TLP may be related to intrinsic (most likely anatomic) factors rather than systemic/extrinsic factors.”⁵

Optimizing the modifiable risk factors (i.e. quality of anesthesia/sedation or platelet counts) is important to minimize the incidence of TLPs. However, identifying and defining nonmodifiable risk factors is equally important. Clearly, age and race are two easily definable intrinsic risk factors. Less clear, or more subjective, however, is the concept of “difficulty with spinous process visualization/palpation.” Moreover, obesity is not well-defined in this setting. Such factors may contribute to what practitioners refer to as “difficult LPs.” Difficult LPs are distinct from TLPs. TLPs may occur in the setting of technically easy LPs; however, difficult LPs, which may require multiple needle redirects or attempts, are a risk factor for TLPs. Indeed, the rate of TLPs increases with increasing LP attempts.⁴

Although clinicians often successfully perform LPs by palpating anatomic landmarks, procedure difficulty is reported in 32% of LPs performed in adults.⁶ Several studies have analyzed the factors predisposing adults to unsuccessful LPs (Table 1). Performing LPs in children is often considered more challenging than in adults because the spinal anatomy of children, especially infants, may be more difficult to interpret, particularly by less experienced providers.⁷ Consequently, LP success rates in children are reportedly as low as 50% to 65%.^4,5 Table 2 summarizes the pediatric studies that have examined the risk factors associated with TLPs.

Table 1.

Review of studies reporting LP difficulty in adult patients.

Study design (n)	Patient population	Mean patient age	Patient age range	Study conclusion
Meta-analysis (957 patients)	Adults and children	24 days– 70.2 years	ND	“US-assisted LPs were associated with higher success rates, fewer traumatic LPs, shorter time to successful LP, fewer needle passes, and lower patient pain scores. US should be considered prior to performing all LPs, especially in patients with difficult anatomy.”
Review (NA)	Adults	NA	NA	“Our goal with this review was to describe our techniques for lumbar puncture in the difficult patient, with emphasis on using fluoroscopy in the obese patient and to suggest maneuvers that might make the procedure easier”
Meta-analysis (NA)	Adults	ND	ND	“US evaluation of the lumbar spine has been shown in randomized trials to improve LP success rates while reducing the number of attempts and the number of traumatic taps.”
Prospective RCT (80 patients)	Adults	62.6 years	>60 years	“US of the lumbar spine was feasible in patients with lumbar conditions even in obese and old ones and allowed the visualization of the epidural space. However, pre-procedure US examination did not reduce pain during the procedure.”
Retrospective (742 LPs)	All ages	43.5 years	0–90 years	“Fluoroscopy-guided LP at the L4-L5 level is associated with nearly twice the risk for traumatic puncture compared with the L2-L3 or L3-L4 level. Rates of traumatic result are twice as high in adults older than 80 years compared with younger patients. Failure rates for fluoroscopy-guided LP are low except in children younger than 1 year, in whom failure occurs in most cases.”
Prospective RCT (46 LPs, 46 patients)	Adults	36.9 years	ND	“The use of US for LP significantly reduced the number of failures in all patients and improved the ease of the procedure in obese patients.”
Prospective cohort (148 LPs, 148 patients)	Adults	42 ± 18 years	16–96 years	“It may be possible to predict which patients will have difficult or traumatic LPs before performing the procedure. Simple bedside assessments of spine visibility and palpability may assist in planning the approach to an LP in patients.”

Abbreviations: LP, lumbar puncture; NA, not applicable; ND, not described; RCT, randomized controlled trial; US, ultrasound.

Table 2.

Review of studies reporting LP difficulty in pediatric patients.

Ref. (year)	Study design (n)	Patient population	Mean patient age	Patient age range	Study conclusion
22(2019)	Meta-analysis (957 patients)	Adults and children	24 days–70.2 years	ND	“US-assisted LPs were associated with higher success rates, fewer traumatic LPs, shorter time to successful LP, fewer needle passes, and lower patient pain scores. US should be considered prior to performing all LPs, especially in patients with difficult anatomy.”
9(2018)	Retrospective (201 LPs, 161 patients)	Children	ND (median 125 days)	2 days–17.7 years	“This study demonstrates that real-time US-guided lumbar puncture is a safe and effective method for performing pediatric lumbar punctures.”
2(2017)	Review (NA)	Children	NA	NA	“We have discussed technical nuances and procedural tips that result in reduced radiation burden and improved success rates for difficult lumbar punctures in children.”
20(2015)	Retrospective (9088 LPs, 440 patients)	Children with cancer and thrombocytopenia	7.8 years	1 month–20.9 years	“Even in thrombocytopenic patient, an epidural hematoma would be a relatively rare complication following lumbar puncture. Despite the large number of punctures performed on patients with platelet counts below 100 000 mm3 (n = 1108), further studies are necessary in order to determine a lower safe platelet count threshold for the performance of lumbar punctures in healthy patients undergoing neuraxial anesthesia.”
13(2013)	Retrospective (4 patients)	Children	16 years	14–19 years	“Our preliminary data suggest that US may be an efficacious alternative to fluoroscopy. By using US to identify the landmarks in the lumbar region, the appropriate puncture point can be determined allowing access to the intrathecal space with relative ease.”
12(2013)	Prospective cohort (9 LPs, 9 patients)	Children	ND	7 weeks–16 years	“Sonography is a promising modality for image-guided lumbar punctures without radiation in children”
3(2011)	Prospective cohort (127 LPs)	Children	ND	0–16 years	“Incidence of traumatic LP was independent of physicians’ experience, sedation use or time of procedure, suggesting an intrinsic factor as the cause of traumatic LPs.”
4(2007)	Prospective cohort (1459 LPs)	Children	ND (median 2.8 months)	0–22 years	“Of the factors associated with traumatic or unsuccessful lumbar punctures in children, advancement of the spinal needle with the stylet in place and lack of local anesthetic use are the most modifiable”
5(2002)	Retrospective (5609 LPs, 956 patients)	Children with newly diagnosed ALL	ND (median 5.5 years)	0–18 years	“The unmodifiable risk factors for traumatic and bloody LP include black race, age younger than 1 year, a traumatic or bloody previous LP performed within the past 2 weeks, and a previous LP performed when the platelet count was 50 × 103 or less. Modifiable risk factors include procedural factors reflected in treatment era, platelet count of 100 × 103µL or less, an interval of 15 days or less between LPs, and a less experienced practitioner.”
11(2001)	Prospective RCT (47 LPs, 32 patients)	Neonates and infants	22 ± 18 days	3–86 days	“US can disclose the cause of failed LP, can help determine whether or not to intervene further, and can provide guidance for LP.”

Abbreviations: LP, lumbar puncture; NA, not applicable; ND, not described; RCT, randomized controlled trial; US, ultrasound.

To improve the chances of LP success and thereby reduce the risk of TLPs, any predisposing factors associated with failed LPs should be identified in pediatric patients before an initial LP attempt, and additional measures, such as ultrasound (US) or fluoroscopic guidance, should be implemented accordingly. US guidance in pediatric patients is a particularly safe and effective method for increasing the success rate of LPs, reducing procedure time and cost and improving patient satisfaction (Table 2).^7–12 However, no data are currently available to suggest which patients are optimal candidates for image-guided LPs. No a priori hypotheses were considered for this study. Therefore, to reduce the likelihood of TLPs in pediatric oncology patients, we analyzed the intrinsic patient factors associated with difficult LPs. On the basis of our findings, we advise preprocedural identification of patients at risk of difficult LPs to inform clinicians of when early use of image-guidance should be considered.

2. Materials and methods

This study was approved by the Institutional Review Board of St. Jude Children’s Research Hospital (St. Jude) as a retrospective review of patients who required diagnostic imaging support for identification of the subarachnoid space for an LP procedure at St. Jude between September 2008 and November 2018 (Study: XPD18-11 Pro00009393). The requirement for obtaining informed consent for this retrospective study was waived by the institutional IRB. These patients were referred for image-guided LPs after an unsuccessful, technically difficult LP using anatomic landmarks.

The data collected included demographic and clinical data, factors expected to be associated with technical difficulties during LPs, and procedural details (Table 3). Classification as overweight (BMI ≥ 25) or obese (BMI ≥ 30) was in concordance with the current Centers for Disease Control and Prevention guidelines.¹³ The current recommendations for platelet counts of > 50 × 10⁹ and international normalized ratio (INR) levels of < 1.5 for LP procedures were considered when evaluating the results.¹⁴ A single clinical research associate collected the data, and a second researcher (HM) verified the data accuracy in 20% of the dataset. Data were managed with the Research Electronic Data Capture (REDCap) web-based application hosted at St. Jude.^15,16 REDCap is a secure platform designed to support data capture for research studies, providing an interface for data capture, audit trails, automated data export, and data integration.

Table 3.

Data collected for analysis.^a

Category	Data acquired
Demographic and clinical details	Patient demographics (i.e., age, sex, race) Diagnosis (hematologic or oncologic) Coagulation status:  Platelet count (within 5 days of procedure)  INR level (within 14 days of procedure)
Factors associated with difficult LP	BMI (collected at the time of each LP procedure) Anatomic abnormalities of the lumbar spine (as identified by prior MRI studies) History of previous LPs Infant age (< 12 months) History of back surgery
Procedural details	Mode of imaging (fluoroscopy or US) Duration of anesthesia Incidence and description of anesthesia complications Duration of fluoroscopy Fluoroscopy radiation exposure LP procedure success Incidence and description of LP procedure complications

Abbreviations: BMI, body mass index; INR, international normalized ratio; LP, lumbar puncture; MRI, magnetic resonance imaging; US, ultrasound.

^aData obtained from primary clinic notes, interventional radiology notes, procedure notes, anesthesia records, and laboratory results.

3. Results

3.1. Patient characteristics

During the study period, 34 patients required diagnostic imaging to identify the subarachnoid space for 65 LP procedures. One procedure/patient was excluded from the analysis because the indication for the image guided procedure was placement of an intrathecal catheter, rather than a prior failed LP by landmark approach. We retrospectively analyzed 64 LPs performed in 33 patients. Twenty-two patients received a single LP, and 11 patients received multiple LPs with image guidance (n = 42 total LPs, median: 3 LPs/patient, range: 2–6 LPs/patient). The ages of the study participants at the time of the procedure ranged from 10 months to 23 years (median: 16.5 years), and the majority (72.7%) of patients were older than 14 years. The cohort consisted of 18 (54.5%) male and 15 (45.5%) female patients. Twenty (60.6%) patients were white, and 13 (39.4%) were African American. The most common diagnosis was leukemia (n = 20 patients, 60.6%), 6 (30%) of which were relapsed cases (Table 4).

Table 4.

Demographic and clinical details.

Category	Finding	Patients
		n (%)
Age	<12 months	3 (9.1)
	1–13 years	6 (18.2)
	14–23 years	24 (72.7)
Sex	Male	18 (54.5)
	Female	15 (45.5)
Race	White	20 (60.6)
	African American	13 (39.4)
Diagnosis	Leukemiaa	20 (60.6)
	Lymphomab	5 (15.1)
	Brain Tumorc	4 (12.1)
	Otherd	4 (12.1)

^aAcute lymphoblastic leukemia (n = 8), acute myeloid leukemia (n = 1), and biphenotypic leukemia (n = 1).

^bBurkitt lymphoma (n = 2), diffuse large B-cell lymphoma (n = 1), non- Hodgkin lymphoma (n = 1), and T-cell lymphoma (n = 1).

^cMedulloblastoma (n = 2), craniopharyngioma (n = 1), and glioblastoma/anaplastic astrocytoma (n = 1).

^dEvans syndrome, sickle cell disease, HIV⁺, and rhabdomyosarcoma (n = 1 each).

3.2. Identified factors associated with difficult lumbar puncture

Of the 64 cases referred for image-guided LPs during the study period, 62 (96.9%) had at least one of the following factors associated with technically difficult LPs (in decreasing order of frequency): BMI ≥ 25, anatomic abnormalities of the lumbar spine, history of ≥ 5 previous LPs, age < 12 months, and history of back surgery (Table 5).

Table 5.

Factors associated with difficult lumbar punctures and procedural details.

Factors associated with difficult lumbar puncture		Patients	Procedures
Factor	Findings	n=33 (%)	n=64 (%)
BMIa	BMI < 25 (normal)	8 (24.2)	19 (29.7)
	BMI ≥ 25 (overweight/obese)	27 (81.8)	45 (70.3)
Anatomic abnormalities of the lumbar spine	Known abnormalities	17 (51.5)	39 (60.9)
	Spinal stenosis	12 (36.4)	35 (54.7)
	Arachnoiditis	4 (12.1)	9 (14.1)
	Postsurgical changes	2 (6.1)	2 (3.1)
	Scoliosis	2 (6.1)	2 (3.1)
	Otherb	4 (12.1)	9 (14.1)
	No known abnormalities	16 (48.5)	25 (39.1)
History of previous LPsa	≥ 5 previous LPs	11 (33.3)	25 (39.1)
	< 5 previous LPs	14 (42.4)	32 (50)
	Unknown	8 (24.3)	7 (10.9)
Age <12 months	Age < 12 months	3 (9.1)	4 (6.25)
	Age > 12 months	30 (90.9)	60 (93.8)
History of back surgery	History of back surgery	3 (9.1)	3 (4.7)
	No history of back surgery	30 (90.9)	61 (95.3)
Procedural details		Patients	Procedures
Category	Details	N=33 (%)	n=64 (%)
Mode of imaging	Fluoroscopy	31 (93.9)	60 (93.8)
	Ultrasound	2 (6.1)	4 (6.3)
LP procedure success rate	Successful	29 (87.9)	59 (92.2)
	Unsuccessful	5 (15.2)	5 (7.8)
	At least one complication	24 (72.7)	37 (57.8)
LP procedure complication details	Bloody CSFc	19 (57.6)	27 (42.1)
	Multiple attempts	11 (33.3)	16 (25)
	Multiple levels	4 (12.1)	4 (6.25)
	Postprocedure headache	2 (6.1)	2 (3.1)

Abbreviations: BMI, body mass index; CSF, cerebrospinal fluid; LP, lumbar puncture.

^aAs recorded at the time of procedure; two patients had BMIs that varied between < 25 to > 25 among multiple procedures, and four patients had multiple procedures that varied between < 5 or ≥ 5 previous LPs.

^bEpidural hematoma (n = 1), reversal of lumbar lordosis (n = 1), and subdural fluid collection/CSF leak (n = 2).

^cBloody CSF was defined as > 10 red blood cells per mm³ of CSF.

3.3. Elevated body mass index

BMIs recorded at the time of the 64 procedures ranged from 14.07 to 49.53 (median: 30.12). Forty-five (70.3%) of the image-guided LPs were performed for patients with BMIs ≥ 25, of which 12 (18.8%) and 33 (51.5%) procedures were performed for patients with BMIs 25 to 30 or BMIs ≥ 30, respectively.

3.4. Anatomic abnormalities of the lumbar spine

Of the 33 patients in our study, 17 (51.5%) had at least one spinal abnormality (Table 5), and 7 (21.1%) had two spinal abnormalities, which was the highest frequency for any patient. Four (12.1%) patients definitively had no spinal abnormalities, and information pertaining to spinal abnormalities was not available for 12 (36.4%) patients. Three patients had a history of back surgery. One patient had a history of multilevel placement of graft material for posterior fusion, spinal hardware placement, and laminectomy. One patient had a history of lumbar spinal hardware placement, and one patient presented with a history of a thoracic laminectomy (partial resection of a mass).

3.5. History of previous lumbar punctures

The number of previous LPs was recorded for 57 LPs and ranged from 0 to 38 (median: 4). The number of prior LPs was ≥ 5 for 25 (39.1%) procedures, < 5 for 32 (50%) procedures, and unknown for 7 (10.9%) procedures (Table 5).

3.6. Infant age group

Three infants (age < 12 months) were included in the study: a 10-month-old female, a 10-month-old male, and an 11-month-old male. Of these infant patients, two presented with additional compounding risk factors. One patient had spinal abnormalities, including an epidural hematoma and stenosis of the central spinal canal. Another infant had a BMI of 27.6 and arachnoiditis.

3.7. Patients presenting with multiple factors associated with difficult lumbar punctures

Twenty-two patients (66.7%) presented with multiple factors associated with difficult LPs (Table 6). The highest number of factors in an individual patient was 4 (median: 2). Of the nine patients who presented with only one factor, the sole factor associated with difficult LPs in these patients was BMI > 25.

Table 6.

Summary of patient-related predisposing factors for difficult lumbar puncture.

Patient No.	Patient characteristics and diagnosis (number of LPs with imaging guidance)	No. factorsb	Elevated BMIc	Spinal abnormalityd	>5 LPs	Age <12 months	Back Surgery
1a	10 mo W M, ALL (1)	4	XA	X2	X	X
2	15 yo W M, Burkitt lymphoma (4)	4	XA	X1,2	X
3	22 yo AA F, Burkitt lymphoma (3)	4	XB	X1,2	X
4a	5 yo AA M, ALL (6)	3	XA	X1	X
5	16 yo W M, T-cell lymphoma (1)	3	XB	X1	X
6	19 yo W M, ALL (1)	3	XB	X1	X
7	21 yo W F, ALL-relapsed (6)	3	XB	X1	X
8a	17 yo AA M, ALL (1)	3	XB	X1,2
9	9 yo AA F, craniopharyngioma (1)	3	XB	X1,5
10	13 yo W F, ALL-relapsed (5)	3		X1,5	X
11	10 mo AA F, ALL (2)	3		X1,5		X
12	19 yo W F, diffuse large B-cell lymphoma (1)	2	XA	X3			X
13	15 yo AA M, ALL (2)	2	XA	X1
14	20 yo W F, medulloblastoma (3)	2	XB	X1
15a	20 yo AA F, non-Hodgkin lymphoma (1)	2	XB	X4
16a	13 yo W F, ALL (1)	2	XB		X
17	15 yo W F, AML (1)	2	XB		X
18	18 yo AA F, ALL (1)	2	XB		X
19	20 yo W F, medulloblastoma (6)	2	XB		X
20	20 yo AA M, ALL-relapsed (1)	2	XB				X
21	11 mo W M, ALL-refractory (1)	2		X5		X
22	16 yo W M, rhabdomyosarcoma (1)	2		X3,4			X
23	17 yo AA F, ALL (1)	1	XA
24	18 yo W M, ALL (1)	1	XA
25	14 yo W M, ALL-relapsed (1)	1	XB
26	16 yo W M , ALL-refractory (3)	1	XB
27	17 yo W F, Evans syndrome (1)	1	XB
28	17 yo AA M, sickle cell disease (2)	1	XB
29	20 yo W M, ALL-relapsed (1)	1	XB
30	21 yo AA F, HIV+ (1)	1	XB
31	23 yo AA M, glioblastoma & anaplastic astrocytoma (1)	1	XB
32	1 yo W M, biphenotypic leukemia-relapsed (1)	0
33	6 yo W M, ALL (1)	0

Abbreviations: AA, African American; ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; BMI, body mass index; F, female; LP, lumbar puncture; M, male; mo, month old; yo, year old; W, white.

^aPatients with unsuccessful image-guided LP.

^bIf patients had a spinal abnormality of surgical hardware/post-surgical changes in addition to a history of back surgery, these were not considered to be separate factors.

^cA = overweight (BMI > 25); B = obese (BMI > 30).

^d1 = spinal stenosis; 2 = arachnoiditis; 3 = surgical hardware/postsurgical changes; 4 = scoliosis; 5 = other spinal abnormality.

3.8. Success rate and complications

Although image guidance for LPs increases success rates, we found that procedural failure and complications occurred even with the assistance of image guidance. Of the 64 procedures attempted with fluoroscopy or US, 5 (7.8%) were unsuccessful, and complications occurred in 37 (57.8%), with bloody CSF being most common (Table 5). Ten (27%) of 37 LPs had two or more reported complications. We further analyzed the number of associated factors for each patient who experienced the following reported complications: bloody CSF (median: 3, range: 0–4), multiple attempts (median: 3, range: 1–4), multiple levels (median: 4, range: 2–4), and postprocedure headache (median: 3.5, range: 3–4).

Five patients with unsuccessful image-guided LPs in our study presented with multiple predisposing factors (Table 5), further emphasizing that these factors contribute to difficult LPs. Of note, all 5 (100%) of these patients had BMIs ≥ 25, 4 (80%) had at least one spinal abnormality, and 3 (60%) had a history of ≥ 5 LPs.

3.9. Additional procedure information

Of 64 procedures, 60 (93.8%) were performed with fluoroscopy, and 4 (6.3%) were performed with US. The fluoroscopy time recorded for 60 LPs ranged from 7 to 498 seconds (median: 34 seconds). Of the patients who received fluoroscopy-guided LPs, radiation exposure was recorded in 35 cases and ranged from 60.7 to 72,301 mGy·cm² (median: 1455 mGy·cm²).

Eight (12.5%) LPs were performed with local anesthetic infiltration alone, and the remaining 56 (87.5%) were performed under general anesthesia. The duration of anesthesia exposure was recorded in 52 cases, ranging from 20 to 116 minutes (median: 54 minutes). No anesthesia-related complications were reported.

3.10. Coagulation status

Platelet count was not available for one LP. The platelet counts for the remaining 63 LPs ranged from 27 × 10⁹ to 632 × 10⁹, and the median and mean platelet counts were 180 × 10⁹ and 210 × 10⁹, respectively, both within normal physiologic limits. Platelet counts were below the recommended 50 × 10⁹ for 4 (6.25%) LPs. Of the 20 reported INR values, none exceeded the recommended 1.5 limit.

4. Discussion

We evaluated the intrinsic factors predisposing difficult LPs in pediatric oncology patients. In addition, we examined the coexistence of multiple intrinsic factors and their effects on LP difficulty, success, and complications. Because LP difficulty is a known risk factor for TLPs, defining the predisposing factors is of considerable importance, particularly in this population.

Our findings indicate that three intrinsic factors are substantially related to difficult LPs: elevated BMI, preexisting anatomic abnormalities of the lumbar spine, and a history of multiple LPs. At least one of these factors was present in 96.9% of the LPs examined in this study, emphasizing the importance of early recognition of these factors. The presence of one and especially more than one of these factors markedly elevates the level of LP difficulty, thus increasing the likelihood of TLP. For this reason, the presence of one or more of these factors should prompt clinicians to consider using image guidance before difficulties are encountered.

BMI ≥ 25 was the most common intrinsic factor in patients who experienced difficult LPs. This finding is consistent with prior studies assessing the effects of elevated BMI on pediatric LP success; however, these studies used BMIs of 30 as a cutoff for difficult LPs, which may not be sufficiently sensitive.^2,6 The effect of elevated BMI on LP difficulty is primarily related to obscured landmark visualization and palpation. However, additional factors, such as the requirement of longer needles, with less ability to redirect, and difficulties with patient positioning may also play a role.¹⁷ Image guidance can help overcome LP difficulty secondary to body habitus.¹⁸

The presence of spinal abnormalities, whether congenital or acquired secondary to tumors or surgeries, was also a primary intrinsic factor predictive of difficult LPs. Similar to obesity, LP difficulty caused by spinal abnormalities may be related to challenging landmark visualization and palpation. However, anatomic abnormalities necessitate atypical needle routes for successful LPs. Even if the spinous processes are palpable externally, the internal anatomy may be altered, increasing the likelihood of the needle contacting bone or venous plexi during passage. This contributes not only to the likelihood of difficult LPs but also TLPs. This is perhaps the factor most easily rectified by diagnostic imaging and can often be anticipated based on a thorough patient history.^10,12,19

The final intrinsic factor we identified that increased LP difficulty is a history of multiple prior LPs. This factor has not been previously reported to cause difficult LPs and is less likely to occur outside of the oncology population because multiple LPs are less frequent. A surprising number of patients in our study experienced LP difficulty after five previously successful LPs. Although this is not intuitive because a history of successful LPs would imply future LP success, multiple prior LPs may cause trauma and scar tissue formation, thereby obstructing the subarachnoid space.¹² Subsequent fibrosis may also be amplified if increased inflammation is caused by introducing chemotherapeutic agents into the intrathecal space, which routinely occurs in the oncology population. Image guidance may assist with needle location determination to avoid fibrotic ligaments and subcutaneous structures.^10,12

Although the benefits of using image guidance for LPs are becoming more evident, difficulty and trauma are still encountered. In pediatric oncology patients with a history of difficult LPs, a 7.8% failure rate and a 42.1% TLP rate occurred despite using image guidance. A 10.7% rate of TLPs in image-guided LPs was previously reported; however, these patients did not have a history of difficult LPs, as was the case with our population.¹⁹ The high rate of TLPs occurred although the vast majority (92.2%) of LPs performed in our study occurred in patients with platelet counts above the recommended guidelines for LP procedures.¹⁴ This suggests that the bleeding these patients experienced was more likely related to LP difficulty rather than coagulopathy. This is consistent with recent studies suggesting that the risk of hemorrhagic complications from LP procedures is not increased in patients with thrombocytopenia.^20,21

In the pediatric oncology population, we advocate for early implementation of image guidance for LPs to reduce the risk of TLPs, procedural complications including failed LPs, and exposure to general anesthesia. Fluoroscopy is the traditional imaging modality of choice for difficult LPs, yet recent studies suggest that US guidance is preferred to avoid exposure to ionizing radiation.¹⁰ Furthermore, US permits real-time visualization of needles in relation to the intrathecal space and adjacent structures, such as the vasculature, whereas soft tissue structures are not visible with fluoroscopy. US is beneficial to visualize areas of thecal sac narrowing, prompting practitioners to attempt accessing the space at an alternate vertebral level. It may also be particularly beneficial for identifying the level of the conus medullaris in infants, in which ossification of the posterior elements has not yet occurred.

Traditional fluoroscopic guidance may, nevertheless, be preferred in certain circumstances. US effectiveness can be limited once ossification of the lumbar region has occurred or by the presence of spinal hardware, causing acoustic shadowing and decreased penetration of the US beam. Adipose tissue in the lumbar region of patients with elevated BMIs also decreases penetration of the US beam. Therefore, fluoroscopy may still be a beneficial imaging modality in older patients, patients with spinal hardware, and/or patients with elevated BMIs.

Our study is unique in its exploration of the pediatric oncology population and risk factors associations with difficult LPs, and it is valuable in that it draws attention to the implications of difficult LPs in the context of cancer. The limitations of this study include its small sample size and retrospective design. The number of patients included in this study was limited because of the overall high success rate of LPs performed without image guidance at St. Jude. In a larger sample population, we hypothesize that a stronger association between the risk factors we identified and LP difficulty will be established. Furthermore, the retrospective study design may have led to incomplete data capture of procedural details. A study design including a control group for direct comparison would permit further statistical analysis. Additionally, examination of extrinsic factors, such as practitioner experience, in addition to the intrinsic factors identified in our study, may further elucidate the association of all risk factors with difficult LP procedures. The high frequency of patients with a history of ≥ 5 previous LPs in our study may be due to the need for multiple diagnostic and therapeutic LPs in pediatric oncology. Therefore, a history of previous LPs may not necessarily be a sole predisposing factor for LP difficulty. Finally, the low incidence of cancer in children aged < 12 months and of back surgery in the pediatric population led to few patients in our study having these factors. On the basis of previously reported findings,¹⁰ we hypothesize that these two factors have a stronger association with difficult LPs than our data demonstrated, and a larger sample size may better capture this association. Although our study did not clearly identify age < 12 months or history of back surgery as predisposing factors for difficult LPs requiring image guidance, we hypothesize that these predisposing factors will be identified in a larger study sample size.

Based on our findings, we propose the recommendations for a clinical algorithm:

1. In all patients with BMI ≥25, early image-guidance should be considered;

2. In all patients with BMI ≥25 AND at least one other predisposing factor (spinal abnormality, history of ≥5 LPs, age <12 months, previous back surgery), early image-guidance is strongly recommended.

5. Conclusion

Minimizing the risk of TLPs in pediatric oncology patients is of utmost importance to improve oncologic outcomes. Difficult LPs are a known risk factor of TLPs and may be related to previously poorly defined intrinsic, patient-related factors. The most common intrinsic factors associated with difficult LPs in the pediatric oncology population identified were BMI ≥ 25, anatomic abnormalities of the lumbar spine, and history of ≥ 5 previous LPs. We consider these factors to be predictive of LP difficulty, and their identification should prompt early implementation of image guidance. This strategy may yield higher LP success rates, lower complication rates, reduced exposure to general anesthesia, and decreased rates of TLPs. These effects have the potential to improve long-term outcomes in the pediatric oncology population.