Fluoroscopically guided transforaminal epidural steroid injections at a quaternary-care teaching institution: effect of trainee involvement and patient body mass index on fluoroscopy time and patient dose

Clinical Radiology 71 (2016) e16ee20

Contents lists available at ScienceDirect

Clinical Radiology journal homepage: www.clinicalradiologyonline.net

Fluoroscopically guided transforaminal epidural steroid injections at a quaternary-care teaching institution: effect of trainee involvement and patient body mass index on fluoroscopy time and patient dose C.A. Tiegs-Heiden a, *, N.S. Murthy a, J.R. Geske b, F.E. Diehn a, B.A. Schueler a, J.T. Wald a, T.J. Kaufmann a, V.T. Lehman a, C.M. Carr a, K.K. Amrami a, J.M. Morris a, K.R. Thielen a, T.P. Maus a a b

Department of Radiology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA Biomedical Statistics and Informatics, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA

article in formation Article history: Received 12 March 2015 Received in revised form 16 September 2015 Accepted 5 October 2015

AIM: To investigate whether there are differences in fluoroscopy time and patient dose for fluoroscopically guided lumbar transforaminal epidural steroid injections (TFESIs) performed by staff radiologists versus with trainees and to evaluate the effect of patient body mass index (BMI) on fluoroscopy time and patient dose, including their interactions with other variables. MATERIALS AND METHODS: Single-level lumbar TFESIs (n¼1844) between 1 January 2011 and 31 December 2013 were reviewed. Fluoroscopy time, reference point air kerma (Ka,r), and kerma area product (KAP) were recorded. BMI and trainee involvement were examined as predictors of fluoroscopy time, Ka,r, and KAP in models adjusted for age and gender in multivariable linear models. Stratified models of BMI groups by trainee presence were performed. RESULTS: Increased age was the only significant predictor of increased fluoroscopy time (p<0.0001). Ka,r and KAP were significantly higher in patients with a higher BMI (p<0.0001 and p¼0.0009). When stratified by BMI, longer fluoroscopy time predicted increased Ka,r and KAP in all groups (p<0.0001). Trainee involvement was not a statistically significant predictor of fluoroscopy time or Ka,r in any BMI category. KAP was lower with trainees in the overweight group (p¼0.0009) and higher in male patients for all BMI categories (p<0.02). CONCLUSION: Trainee involvement did not result in increased fluoroscopy time or patient dose. BMI did not affect fluoroscopy time; however, overweight and obese patients received significantly higher Ka,r and KAP. Male patients received a higher KAP in all BMI categories. Limiting fluoroscopy time and good collimation practices should be reinforced in these patients. Ó 2015 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

* Guarantor and correspondent: C.A. Tiegs-Heiden, Department of Radiology, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA. Tel.: þ1 507 284 0440; fax: þ1 507 293 3680. E-mail address: [email protected] (C.A. Tiegs-Heiden). http://dx.doi.org/10.1016/j.crad.2015.10.003 0009-9260/Ó 2015 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

C.A. Tiegs-Heiden et al. / Clinical Radiology 71 (2016) e16ee20

Introduction Transforaminal epidural steroid injections (TFESIs) are an effective therapy in the treatment of radicular pain.1,2 During these procedures, fluoroscopic guidance is commonly utilised (versus, less often, computed tomography [CT]) to ensure the delivery of corticosteroid directly to the spinal nerve and root within the epidural space of the selected vertebral foramen.2 As with many other radiology examinations and procedures, radiation dose is an important consideration. In keeping with the principles of ALARA (as low as reasonably achievable), it is important to understand the factors that may influence radiation dose during these procedures. At teaching institutions, fluoroscopically guided spine injections are sometimes performed by a trainee under the supervision of a staff radiologist. Based on previous studies, it is anticipated that the fluoroscopy time and patient radiation dose are higher when a trainee is involved in the procedure, compared to a staff radiologist working alone.3,4 Another study reported significantly higher radiation doses for spine intervention procedures performed at university teaching hospitals compared to private practice settings, likely due to the presence of trainees.5 It has also been shown that patients with a larger body mass index (BMI) are exposed to a greater radiation dose compared to patients with lower BMIs during fluoroscopically guided procedures.6e8 Specifically, one study demonstrated longer fluoroscopy time and higher fluoroscopy dose in overweight patients during fluoroscopic-guided spine interventions.8 The purpose of this study was to measure whether there were differences in fluoroscopy time and patient dose between fluoroscopically guided TFESIs performed by staff radiologists alone versus with trainees. Additionally, the effect of patient BMI on fluoroscopy time and patient dose was evaluated, and any interaction between these variables was assessed.

Materials and methods Institutional Review Board approval was obtained prior to the review and the study population was limited to those patients who had given research authorisation. The review was also compliant with the Health Insurance Portability and Accountability Act. Patients who underwent fluoroscopic-guided, single-level, unilateral TFESIs from L1 to S1 between 1 January 2011 and 31 December 2013 were included in the study (n¼2108). Only one procedure per patient was included and limited to those patients with available BMI data within 2 months of the procedure (n¼1853). Patients under the age of 18 years were excluded as well as clear errors in fluoroscopy time and radiation dose entries. After applying the inclusion and exclusion criteria, a total of 1844 patients were identified. The practice consisted of musculoskeletal and neuroradiology fellowship-trained academic radiologists with specific spine pain intervention training. Trainees consisted

e17

primarily of musculoskeletal and neuroradiology fellows, with rare involvement by senior radiology residents. Trainee involvement was manually entered by the performing staff radiologist immediately after each procedure in a pre-existing quality assurance database as a simple yes/ no response. When the trainees were involved in the case, they typically performed the entire procedure and operated the fluoroscopy unit; staff radiologists were instructed to enter “No” if a trainee was present but not significantly involved. For all cases, the staff radiologist was present in the room for the entire procedure providing direct supervision, as is the required standard of care in the authors’ group. All procedures were performed on one of two fluoroscopy systems (both Multi Diagnostic Eleva, Phillips Healthcare, Andover, MA, USA). The displayed dose values of both systems were validated against an external dosimeter and found to be accurate to within 5%. Fluoroscopic guidance was typically performed at 15e30 frames per second (fps) with the smallest collimation possible while still being able to identify anatomical landmarks for documentation. Documentation of a typical TFESI consisted of five digital spot images: pre- and post-contrast in the anterioreposterior (AP) and lateral planes with an additional washout image in the AP plane following the administration of a test dose of lidocaine. Digital subtraction imaging was limited to cases where live fluoroscopic visualisation was suboptimal, including patients with iodinated contrast medium allergies where a gadoliniumbased contrast agent was substituted. The International Spine Intervention Society guidelines for image acquisition and documentation were followed.9 Fluoroscopy time, reference point air kerma (Ka,r), and kerma area product (KAP2) for all procedures performed during this time period were manually recorded in the database. Ka,r, and KAP are dose metrics, which can be used to estimate peak skin dose.10

Data analysis Descriptive and inferential statistics were performed in this observational study. Mean fluoroscopy time, Ka,r, and KAP were compared by trainee involvement (yes/no) using t-tests. Pearson correlations were used to examine associations of BMI with fluoroscopy time, Ka,r, and KAP. BMI and trainee involvement were examined as predictors of fluoroscopy time in multivariable general linear models adjusted for age and gender. BMI was considered as a class variable for these analyses by defining World Health Organization (WHO) categories of underweight or normal (BMI <25), overweight (BMI 25e30), and obese (BMI >30).11 Due to the increased variance in the fluoroscopy measures at higher BMIs, as well as possible confounding related to the various staff radiologists performing the procedure, heteroskedasticity-consistent standard errors and significant tests were performed.12 Similar models were constructed for fluoroscopy time, with the addition of interactions between BMI and fluoroscopy time and between BMI and gender. Stratified models of the three BMI

e18

C.A. Tiegs-Heiden et al. / Clinical Radiology 71 (2016) e16ee20

groups by trainee status were also performed. Statistical significance was set at p<0.05.

Results A total of 1844 cases were used for the primary analysis. There were 913 (49.5%) male patients. Mean age at the time of the procedure was 61.6 years (range 18e97 years). Trainees were involved in 359 (19.5%) procedures. There were 6e8 neuroradiology and musculoskeletal fellows each year during the study period and a single senior resident each year. There were 10 staff radiologists who each performed at least 44 (2.4%) of the total procedures, therefore allowing for methods to control for confounding by operator. The range of post-fellowship experience of the staff radiologists during the study period was 2e14 years. No staff radiologist transitioned from a fellow to staff during the study period. The most experienced staff radiologist performed the largest percentage of the cases (26.9%). The relative proportion of cases in which a trainee was involved was similar among underweight/normal (19.6%), overweight (20.3%), and obese (18.7%) BMI categories (p¼0.741). Mean (standard deviation [SD]) BMI was 30 (5.3) for males and 29.9 (6.9) for female patients (p¼0.621). Demographics and clinical summary statistics are presented in Table 1. Univariate analysis using t-tests regarding trainee involvement revealed that mean fluoroscopy time was not significantly different when a trainee was involved (meanSD: 2.250.98 min) compared to procedures with only the staff radiologist (2.251.20 min; p¼0.897). Likewise, Ka,r was similar for both groups (with trainee: 98.371.5 mGy, with no trainee: 102.570.2 mGy; p¼0.322). Mean KAP was significantly lower with trainee involvement (1073868 versus 1223910 mGy m2, p¼0.005). Univariate analyses of Ka,r and KAP with BMI revealed that higher Ka,r and KAP were positively correlated with increased BMI (dose Pearson coefficient¼0.496, KAP Pearson coefficient¼0.496; p<0.0001 for both). Multivariable models examined associations with trainee involvement and BMI adjusted for age, gender, and fluoroscopy time, and level of injection (Table 2). In the entire cohort, increased age was significantly associated Table 1 Demographics and clinical characteristics. Characteristic

n (%) or MeanSD

No. of procedures No. of operators No. of procedures/operator Age Gender (male) BMI Underweight/normal Overweight Obese Trainee involved Fluoroscopy time (min) Ka,r (mGy) KAP (mGy$m2)

1844 10 Range 44e500 (2.4e27.0%) 61.815.5 913 (49.5) 29.96.2 392 (21.3) 650 (35.2) 802 (43.5) 359 (19.5) 2.31.2 101.770.5 1193.9904.4

BMI, body mass index; Ka,r, point air kerma; KAP, kerma area product.

with increased fluoroscopy time (p<0.0001). Gender, trainee presence, and BMI were not significant predictors. There was a statistically significant difference in fluoroscopy times by level of injection; however, the difference in mean fluoroscopy times was clinically small. The median fluoroscopy times for L1, L2, L3, L4, L5, and S1 were 2, 2.15, 2.14, 1.98, 20, and 1.82 minutes, respectively. Ka,r was significantly higher with longer fluoroscopy time and in patients with a higher BMI. KAP was significantly higher in male patients, with longer fluoroscopy time, and with higher than normal BMI. KAP was significantly lower when a trainee was present. There were significant interactions of BMI and fluoroscopy time in predicting higher Ka,r and KAP, meaning the effect of increased fluoroscopy time was more significant in patients with higher BMIs. (Fig 1 demonstrates this in a sample male patient.) This interaction effect warranted further analyses using BMI stratified models (Table 3). Longer fluoroscopy time predicted increased Ka,r and KAP in all BMI groups. Increased age was a significant predictor of increased fluoroscopy time in the normal/underweight and overweight groups. KAP was higher in male patients in all groups. KAP was lower with trainees present in the overweight and obese groups, but was not statistically different in the underweight/normal BMI group.

Discussion In the present study, trainee involvement was not associated with a significant increase in fluoroscopy time or radiation dose during lumbar TFESIs. In contrast to prior studies,3,4 the present study did not demonstrate an increased fluoroscopy time or fluoroscopy dose when a trainee was involved in a spine injection procedure. For example, a study of fluoroscopy times for voiding cystoTable 2 Models of fluoroscopy time, reference point air kerma (Ka,r), and kerma area product (KAP) using predictors of age, gender, trainee involvement, body mass index (BMI), and the interaction term of BMItime. Outcome

R2

Effect

Fluoroscopy time

0.019

Age

Ka,r

0.572

KAP

0.602

Gender (male) Trainee BMI Level (L1-S1: 5 df) Age Gender (male) Trainee Time BMI BMItime Age Gender (male) Trainee Time BMI BMItime

Parameter estimate

Standard error

p-Value

0.010

0.002

<0.0001

0.024 0.017 0.001 e

0.054 0.068 0.004 e

0.6516 0.7740 0.7656 0.0178

0.103 3.580 1.146 11.413 2.429 1.510 1.784 128.855 114.313 256.570 22.970 23.439

0.073 2.186 2.809 6.203 0.494 0.224 0.818 26.921 35.009 82.561 6.902 3.055

0.1582 0.1016 0.6832 0.0659 <0.0001 <0.0001 0.0292 <0.0001 0.0011 0.0019 0.0009 <0.0001

C.A. Tiegs-Heiden et al. / Clinical Radiology 71 (2016) e16ee20

e19

Figure 1 Model-based estimated means BMI by time interaction on Ka,r and KAP for a male patient of study mean age, 61.6 years.

urethrogram (VCUG) procedures demonstrated that staff radiologists had significantly lower fluoroscopy times than senior residents and fellows.3 Another study found that resident physicians used approximately twice the fluoroscopy time as staff radiologists during central venous catheter placement.4 One explanation for the differences is that in the authors’ practice, the trainee was directly supervised by a staff physician throughout the entire procedure. A

supervising radiologist was also present for each case in the prior studies, but the degree of supervision varied.3,4 This direct staff supervision may also explain why KAP was actually lower in the entire cohort when a trainee was present. The trainees were likely constantly reminded of the importance of the ALARA principle and the attention to collimation. Furthermore, in the authors’ experience, spinal pain patients cannot tolerate lengthy procedures. The

Table 3 Body mass index (BMI) stratified models. Outcome

BMI category

R2

Effect

Parameter estimate

Standard error

p-Value

Fluoroscopy time

Underweight & Normal

0.051

Overweight

0.028

Obese

0.120

Underweight & Normal

0.310

Overweight

0.498

Obese

0.360

Underweight & Normal

0.530

Overweight

0.553

Obese

0.351

Age Gender (male) Trainee Age Gender (male) Trainee Age Gender (male) Trainee Age Gender (male) Time Trainee Age Gender (male) Time Trainee Age Gender (male) Time Trainee Age Gender (male) Time Trainee Age Gender (male) Time Trainee Age Gender (male) Time Trainee

0.016 0.168 0.010 0.012 0.047 0.033 0.005 0.087 0.016 0.301 5.874 20.282 1.110 0.012 0.084 29.125 3.396 0.163 3.123 41.877 6.066 2.845 79.338 255.343 35.988 2.916 101.759 398.244 149.754 2.738 140.996 551.878 197.770

0.004 0.125 0.129 0.002 0.090 0.087 0.003 0.081 0.099 0.130 4.724 2.187 5.572 0.084 2.700 1.660 3.863 0.163 4.917 2.179 6.23 0.792 34.493 20.053 44.190 1.007 33.392 21.237 44.981 2.062 61.625 36.418 78.031

<0.0001 0.1783 0.9407 <0.0001 0.6006 0.7070 0.0518 0.2883 0.3818 0.0305 0.2145 <0.0001 0.8422 0.8831 0.9751 <0.0001 0.3797 0.3181 0.5250 <0.0001 0.3302 0.0004 0.0220 <0.0001 0.4159 0.0039 0.0024 <0.0001 0.0009 0.1845 0.0224 <0.0001 0.0115

Ka,r

KAP

Numbers of patients in each category were n¼392 underweight and normal; n¼650 overweight; n¼801 obese.

e20

C.A. Tiegs-Heiden et al. / Clinical Radiology 71 (2016) e16ee20

potential for neural irritation must also be avoided, which can dramatically increase the patient’s pain and ability to tolerate the procedure. For both of these reasons, the attending staff is potentially more vigilant of the trainees to ensure a safe and relatively quick procedure. Additionally, the degree of trainee experience and the teaching methods used may account for the present results. The majority of trainees who rotate through the spine injection practice are musculoskeletal or neuroradiology fellows. These individuals are more experienced in image-guided procedures than residents, which could result in lower fluoroscopy times, similar to staff radiologists. Furthermore, the trainees are slowly graduated with respect to procedure complexity. They begin with relatively simple procedures, such as sacroiliac and lumbar facet joint injections. Once they have mastered these injections and become familiar with the needle techniques and fluoroscopy equipment, they are invited to perform lumbar TFESIs. At this stage, their skill level and efficiency are likely greater than when they were performing the joint injections. The BMI did not affect fluoroscopy time in the present study; however, patients with a higher BMI did receive significantly higher Ka,r and KAP. This result is consistent with multiple prior studies, which demonstrated increased radiation doses to patients with higher BMIs during fluoroscopic procedures.6,7 Patients with larger BMIs attenuate X-rays to a greater degree, which results in the fluoroscopic automatic dose rate control increasing the kilovolt peak and milliamperes of the X-ray beam in order to maintain a constant exposure at the image receptor entrance. In addition, the increased path length through a larger BMI patient increases radiation scatter, lowers subject contrast, and reduces image quality. Other studies have found an increase in fluoroscopy time in overweight patients likely due to this reduced image quality.8 Increased age predicted increased fluoroscopy time in the normal/underweight and overweight groups in this study. These cases may be more complex secondary to agerelated degenerative changes, which makes the anatomy more difficult to visualise and access. Radiologists may also be less vigilant about applying ALARA in older patients. Increased KAP was seen in male patients in this study, despite relatively equal numbers of male and female patients in each BMI category. This is of uncertain cause: it may be because male patients often carry more weight in their abdomen and/or an increase in lean body mass compared to female patients, thus leading to increased radiation dose during lumbar spine TFESIs. One limitation of this study is its retrospective nature. An attempt was made to overcome potential confounding related to patient factors (age, sex, and patient size) through various regression and stratification methods. Residual confounding related to caseetrainee matching, however, cannot be ruled out. This limitation would require prospective experimentation in a way that would be challenging logistically and potentially ethically. The use of five standardised digital spot films could have contributed to uniformity in dose and reduced confounding. The degree of

trainee involvement was not completely characterised for each case; however, as described in the methods, at this stage of their graduated spine intervention instruction, the trainees typically performed the entire procedure themselves, and staff radiologists have been instructed to categorise such procedures as having trainee involvement. For the Mayo Clinic, the data indicate that trainee involvement does not contribute to increased fluoroscopy time or patient dose. In the overweight group, trainee involvement was associated with a decrease in KAP, likely related to greater attention to collimation and the ALARA principle. Overweight and obese patients, however, received a higher radiation dose than their underweight/ normal weight counterparts. Radiation dose was also greater in male patients for all BMI categories. Limiting fluoroscopic time and good collimation practices should be reinforced in all patients, especially for males and in the overweight and obese populations who are susceptible to receiving increased radiation doses.

Acknowledgements The authors thank Ms Jane Gagnon for assistance with manuscript preparation.

References 1. Kennedy DJ, Plastaras C, Casey E, et al. Comparative effectiveness of lumbar transforaminal epidural steroid injections with particulate versus nonparticulate corticosteroids for lumbar radicular pain due to intervertebral disc herniation: a prospective, randomized, double-blind trial. Pain Med 2014;15:548e55. 2. MacVicar J, King W, Landers MH, et al. The effectiveness of lumbar transforaminal injection of steroids: a comprehensive review with systematic analysis of the published data. Pain Med 2013;14:14e28. 3. Lim R, Khawaja RD, Nimkin K, et al. Relationship between radiologist training level and fluoroscopy time for voiding cystourethrography. AJR Am J Roentgenol 2013;200:645e51. 4. Xu BJ, Duszak Jr R, McGinnis RS, et al. Increased fluoroscopy time for central venous catheter placement by radiology residents versus staff radiologists. J Am Coll Radiol 2013;10:518e22. 5. Zhou Y, Singh N, Abdi S, et al. Fluoroscopy radiation safety for spine interventional pain procedures in university teaching hospitals. Pain Physician 2005;8(1):49e53. 6. Ector J, Dragusin O, Adriaenssens B, et al. Obesity is a major determinant of radiation dose in patients undergoing pulmonary vein isolation for atrial fibrillation. J Am Coll Cardiol 2007;50:234e42. 7. Hsi RS, Zamora DA, Kanal KM, et al. Severe obesity is associated with 3fold higher radiation dose rate during ureteroscopy. Urology 2013;82:780e5. 8. Smuck M, Zheng P, Chong T, et al. Duration of fluoroscopic-guided spine interventions and radiation exposure is increased in overweight patients. PM R 2013;5:291e6. quiz 296. 9. Bogduk N. Practice guidelines for spinal diagnostic and treatment procedures. 2nd ed. San Francisco: International Spine Intervention Society; 2013. 10. Kwon D, Little MP, Miller DL. Reference air kerma and kerma-area product as estimators of peak skin dose for fluoroscopically guided interventions. Med Phys 2011;38:4196e204. 11. Obesity: preventing and managing the global epidemic. Report of a WHO consultation. World Health Organ Tech Rep Ser 2000;894(ixii):1e253. 12. White HA. heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 1980;48:817e38.