JVIR twitter

Tuesday, July 17, 2018


Randomized Controlled Trial Comparing Radiologic Pigtail-Retained Gastrostomy and Radiologic Mushroom-Retained Gastrostomy” Kichang Han et al. Volume 28, Issue 12 (December 2017)

  • Prospective randomized controlled trial compared 14-F pigtail-retained gastrostomy (PG; 50 patients) and 20-F mushroom-retained gastrostomy (MG; 49 patients) (level of evidence 1b).
  • Technical success rate was 100% for both groups.
  • Minor complication rate was higher in the PG group (34%) than in the MG group (12.2%; P = .016) owing to higher rates of tube occlusion (16% vs 0%) or dislodgment (4% vs 0%) and peristomal leakage (6% vs 2%).
  • Major complication rate was 2% in the PG group and 0% in the MG group (P > .05).

Monday, July 16, 2018

A Comparison of Concomitant Tributary Laser Ablation and Foam Sclerotherapy in Patients Undergoing Truncal Endovenous Laser Ablation for Lower Lime Varicose Veins 


A prospective non randomized controlled study comparing the efficacy and safety of simultaneous tributary vein ablation with either Foam Sclerotherapy (FS) or Endovenous Laser ablation(EVLA) while performing EVLA of the greater saphenous vein was conducted. 418 patients were enrolled (CEAP score between 2 and 6), 163 underwent tributary vein EVLA at 8 Watts while 255 underwent FS with 3% Polidoconol using the Tessari technique. Outcomes were measured at 3 days, 4 weeks and 6 months using the Aberdeen Varicose Vein Questionnaire(AVVQ), EuroQol Group 5(EQ-5D) and Numerical rating scale for pain. At 4 weeks and 6 months respectively, the EVLA group exhibited poorer QoL than the FS group (P<0.001 for AVVQ ;P<0.01 for EQ-5D) and (P<0.5 for AVVQ; P<0.05 for EQ-5D). At 6 months the FS group had lower (4.9%) visible varicosities than the EVLA group (21.6%). FS group reported significantly higher post procedural pain only at 4 weeks (P<0.001. Persistent ecchymosis, pigmentation and local induration were significantly higher in the FS group at both 4 weeks and 6 months (9.3%vs 6.2%). The authors concluded that concomitant tributary ablation with FS during truncal laser ablation of the GSV is more efficacious and improves Qol compared to EVLA.


This paper evaluates the utility of concomitant treatment of tributary veins using 2 different techniques. Advantage of simultaneously treating tributary varicosities include lesser cost and reduced need for a second procedure. Simultaneous treatment is associate with longer procedure time and post-procedure discomfort. Supported by recent randomized studies favoring simultaneous treatment, the National Institute for Health and Care Excellence has set simultaneous tributary vein treatment as the standard of care. Phlebectomy is more invasive than EVLA and FS. Based on the paper, the treatment of choice for tributary veins appears to be FS.

Click here for abstract

Wang JC, Li Y, Li GY, et al. A Comparison of Concomitant Tributary Laser Ablation and Foam Sclerotherapy in Patients Undergoing Truncal Endovenous Laser Ablation for Lower Limb Varicose Veins. J Vasc Interv Radiol 2018; 29:781-9.

Post Author:
Anil K Pillai, MD
Associate Professor and Section Chief,
University of Texas Health Science Center, Houston, TX 77030
Anil Pillai@AnkupiMD

Thursday, July 12, 2018

Vaccum-Assisted Thrombectomy in AVF and AVG


This prospective study included 35 patients with acutely thrombosed AVF and AVG that underwent Thrombectomy using the Indigo device. The study sought out to evaluate if the Indigo system would be safe, effective, decrease procedure time and complications. No patient was excluded during the study period. The authors used systemic heparin (no TPA), obtained antegrade and retrograde access to treat the thrombus, with adjunctive use of angioplasty and or stents for stenosis, occlusion balloon for thrombus in the arterial anastomosis and Trerotola device for adherent thrombus in an aneurysmal segment. The authors evaluated technical and clinical success, primary patency, primary assisted patency, and secondary patency, procedure time, blood loss and complications. Technical success was 34 out of 35 patients, with clinical success in 32 out of 35 patients. Three complications were not attributed to the device. As expected 93% of patients required angioplasty, 2 patients stent grafts, 1 patient required Trerotola device and 2 patients occlusion balloon. Average procedure time was 38 minutes, average blood loss was 122 mL, 6 month primary patency, primary assisted patency and secondary patency was 71%, 80% and 88.5%. The results of this study demonstrate high technical and clinical success, comparable or better to other studies published. Average procedure time (38 minutes) was also improved when compared to Angiojet (78 minutes), mechanical thrombolysis (126 minutes). Similarly patency rates compare favorably compared to other systems such as Angiojet, Trerotola, etc. Changes in hemoglobin were not seen in 17 patients that had hemoglobin levels checked after the procedure. The biggest limitation is the small number of patients, lack of cost analysis and lack of control group for comparison. The authors conclude that the indigo system is safe and effective with promising patency rates and procedural times when compared to other systems or techniques.

Figure- Representative fistulogram demonstrates the initial injection with thrombus in the graft (A), fistulagram after antegrade thrombectomy but before angioplasty of the stenotic lesion (B), fistulagram after thrombectomy and angioplasty of the lesion (C), fistulagram from the brachial artery showing thrombus in the arterial end (D), and final fistulagram after retrograde thrombectomy.


This study provides promising data regarding the use of the Indigo system for the treatment of thrombosed dialysis access. With low procedural times, high patency rates at 6 months and low complications rates, this system has the potential to replace other systems. The authors required the use of adjunctive systems in 3 separate cases, and as expected angioplasty was required in the majority of the cases to threat the underlying lesions. The biggest limitation in this study was the lack of cost analysis and lack of control group. Future studies that compare the indigo system against other systems (mechanical thrombectomy, Angiojet, Trerotola, etc) and provides a cost analysis will elucidate if the Indigo system can become a first line system in the treatment of thrombosed dialysis access.

Click here for abstract

Marcelin C, D'Souza S, Le Bras Y, Petitpierre F, Grenier N, van den Berg JC, Huasen B. Mechanical Thrombectomy in Acute Thrombosis of Dialysis Arteriovenous Fistulae and Grafts Using a Vacuum-Assisted Thrombectomy Catheter: A Multicenter Study. J Vasc Interv Radiol. 2018 Jul;29(7):993-997.

Post Author: 
Carlos J. Guevara, MD
Assistant Professor
Department of Radiology, Interventional Radiology Division
University of Texas Health Sciences, Houston

Monday, July 9, 2018

Contralateral DVT after Iliac Vein Stent Placement in May-Thurner Syndrome: location, location, location... 


Investigators from Inha University Hospital in the Republic of Korea published their analysis on the incidence and potential causes of contralateral deep vein thrombosis (DVT) after common iliac vein (CIV) stenting in patient with May-Thurner syndrome (MTS). Retrospective analysis included 111 patients (women: 73%). Median follow-up was 36 months. Stent location was determined by venogram and classified as extended to the inferior vena cava (IVC), covered the confluence, or confined to the iliac vein. The relationship between stent location and contralateral DVT was analyzed. Potential causes of contralateral DVT were presumed based on venographic findings. Ten patients exhibited contralateral DVT at a median timing of 40 months (9%). Potential causes were venous intimal hyperplasia (VIH) in the distal IVC (n=7), “jailing” of the contralateral CIV (n=2), and indeterminate (n 1⁄4 1). All patients with VIH had previous CIV stents overextended to the IVC. Overextension of CIV stent was associated with contralateral DVT (P < .001). The authors concluded that contralateral DVT after CIV stenting has a relatively high incidence, and overextension of the stent to the IVC wall is associated with contralateral DVT, having VIH as a potential cause.

Fig.1: Categorization of left iliac vein stent position on venogram: (a) extended to the IVC, (b) covered the confluence, and (c) confined to the ipsilateral iliac vein. Arrows indicate tips of stents. Asterisks indicate the confluence. (d) A diagram illustrating the 3 categories of iliac vein stent: 1) overextension, 2) complete confluence coverage, and 3) without confluence involvement.


This study helps in understanding the incidence and potential associated factors for contralateral DVT after CIV stenting in patients with MTS. Endovascular stent insertion has been recognized as the treatment of choice for venous outflow obstruction given its safety and effectiveness. Appropriate stent placement is critical for satisfactory long term results, and this study suggests that overextension of the stent into IVC to ensure complete coverage of the lesion may not go unpunished. At the same time research provides plausible explanation for the increased incidence of contralateral DVT. First, persistent insult of the distal IVC wall by the stent struts, leading to VIH. Second, “jailing” of the contralateral CIV disrupting normal blood outflow. On the other hand, not extending the stent into the IVC may prevent complete coverage of the stenotic lesion, causing recurrent ipsilateral symptoms. It would be interesting to see a comparison of freedom from target lesion recanalization (FLR) between those 3 groups with different stent position. Nonetheless, the study highlights the importance of appropriate stent placement. Utilization of additional imaging guidance such as intravascular ultrasound should be common practice, and industry must continue to improve device design to ensure better biocompatibility and precise deployment.

Click here for abstract

Le TB, Lee TK, Park KM, Jeon YS, Hong KC, Cho SG. Contralateral Deep Vein Thrombosis after Iliac Vein Stent Placement in Patients with May-Thurner Syndrome. J Vasc Interv Radiol. 2018 Jun;29(6):774-780.

Post Author:
Ricardo Yamada, MD
Assistant Professor
Department of Radiology
Division of Vascular and Interventional Radiology
Medical University of South Carolina

Wednesday, April 18, 2018

From the SIR Residents and Fellows Section (SIRRFS)

Teaching Topic: Uterine Artery Embolization for Pedunculated Subserosal Leiomyomas: Evidence of Safety and Efficacy

Kim, YS, Han K, Kim M, Kim GM, Kwon JH, Lee J, Choi W, Won JY, Lee DY. Uterine Artery Embolization for Pedunculated Subserosal Leiomyomas: Evidence of Safety and Efficacy. 2018 Feb 22. doi: 10.1016/j.jvir.2017.11.022 [Epub ahead of print]

Click here for abstract

Uterine artery embolization (UAE), colloquially known as uterine fibroid embolization, has become a widely accepted and increasingly popular option for the treatment of symptomatic leiomyomas. Since its first report by Ravina et al. in 1995, the evidence supporting its safety and efficacy has become increasingly robust with large prospective, randomized controlled trials like the REST and EMMY trials. [1,2] Recently, emerging evidence is even broadening the conditions that UAE could be indicated for to include adenomyosis and premenstrual symptoms. [3] Despite these advances, the historical controversy surrounding UAE for pedunculated subserosal (PS) leiomyomas, particularly ones with narrow stalks, has persisted. Guidelines from the Cardiovascular and Interventional Radiological Society of Europe (CIRSE) list PS leiomyomas with stalks that are less than 50% the diameter of the leiomyoma as a relative contraindication because of the risk of torsion and ischemic necrosis of the stalk and resulting separation of the leiomyoma from the uterus. These concerns largely stem from two early cases where necrotic PS leiomyomas detached and became a source of sepsis which required hysterectomy and bowel resection. [4] Because of this risk, PS leiomyomas were generally excluded from the major UAE trials, including the REST trial. However, the guidelines from the SIR Standards of Practice Committee explicitly refutes this contraindication citing four retrospective series of 12, 16, 18, and 29 patients who underwent successful UAE for PS leiomyomas. [5] In this study, Kim et al. present the largest and most comprehensive series to date of UAE for PS leiomyomas in 55 patients.

In this retrospective review, the authors present 55 patients with a total of 66 PS leiomyomas treated with UAE between 2007 and 2016. They further categorized leiomyomas with stalk diameters 25% or less of the diameter of the leiomyoma as high-risk (n=11) and leiomyomas with stalk diameters 50% or less of the diameter of the leiomyoma low-risk (n=55). Magnetic resonance imaging was performed 3 months after UAE to compare infarction rate and volume reduction between high-risk and low-risk groups and between PS leiomyomas and leiomyomas in other locations. The authors successfully embolized the uterine arteries bilaterally in 54/55 (98.2%) of patients. There was no significant difference in the mean volume of reduction in PS leiomyomas compared to non-PS leiomyomas (38.2% vs 38.4%, p=.953), though the mean infarction rate in PS leiomyomas was significantly lower than the rate in non-PS leiomyomas (86.0% vs 99.9%, p=.0025). All 66 stalks continued to enhance after UAE. Clinically, 53/55 patients reported symptomatic improvement. One of the patients who did not was the aforementioned technical failure due to non-embolized inferior mesenteric artery collaterals. It was not known what prevented symptomatic improvement in the other patient as volume reduction rates were in line with the other patients.
MR images before and after UAE for a high-risk PS leiomyoma showing complete infarction and remaining stalk enhancement. 

Clinical Pearls

What are the different locations of uterine leiomyomas?

Leiomyomas are typically classified by their location. Intramural leiomyomas, the most common type, occur in the muscular wall of the uterus. Subserosal leiomyomas are the rarest form and occur under the uterine serosa on the surface of the uterus. These can be pedunculated or sessile. They typically do not cause infertility. Submucosal leiomyomas occur under the endometrium and also may be pedunculated or sessile. Submucosal leiomyomas have the largest effect on fertility.

What imaging is required before a UAE?

MR imaging is generally performed before and after the procedure to evaluate leiomyoma amount, size, location, presence of adenomyosis and response to treatment. 3D reconstructed MRA imaging can be used for preprocedural planning and mapping of the uterine and ovarian artery and checking for variant anatomy. The authors of this study did not report if they used MRA preprocedurally and had one technical and clinical failure due to a collateral inferior mesenteric artery that was not embolized.

Questions to Consider

What are the potential complications and adverse events of UAE for PS leiomyomas?

In this study, the authors reported no major adverse events and 3 minor adverse events. Two of these consisted of the expulsion of submucosal leiomyomas. The third adverse event was a presentation to the emergency room 2 days after UAE for pelvic pain which resolved after 2 hours. Stalk necrosis causing separation and the subsequent infection is the primary fear with PS leiomyomas. This study compared stalk enhancement to adjacent myometrium and found no difference in enhancement in any of the 66 cases.

What is the evidence level of the SIR recommendation?

According to the 1998, United States Presentative Services Task Force (USPSTF) guidelines on levels of evidence, the recommendations by the SIR Standards of Practice Committee would be considered Level III evidence as it is expert opinion based on descriptive studies. This study adds 55 more patients to the previously documented 75 cases from the four case series. Level II evidence describes recommendations based from cohort, case-control, or non-randomized controlled trials. Recommendations require at least one randomized controlled to be considered Level I.

Additional Citations

1. The Rest investigators (2007) Uterine-artery embolization versus surgery for symptomatic uterine fibroids. N Engl J Med 356:360–370

2. Hehenkamp WJK, Volkers NA, Birnie E, Reekers JA, Ankum WM (2008) Symptomatic uterine fibroids: treatment with uterine artery embolization or hysterectomy. Results from the randomized clinical embolization versus hysterectomy (EMMY) trial. Radiology 246:823–832

3. Jang D, Kim MD, Lee SJ, Kim IJ, Park SI, Won JY, Lee DY. The effect of uterine artery embolization on premenstrual symptoms in patients with symptomatic fibroids or adenomyosis. J Vasc Interv Radiol. 2014 Jun;25(6):833-838.e1. doi:10.1016/j.jvir.2014.01.036. Epub 2014 Mar 20. PubMed PMID: 24657088.

4. Braude, P., Reidy, J., Nott, V., Taylor, A., and Forman, R. Embolization of uterine leiomyomata: current concepts in management. Hum Reprod Update. 2000; 6: 603–608

5. Dariushnia, S.R., Nikolic, B., Stokes, L.S., and Spies, J.B. Quality improvement guidelines for uterine artery embolization for symptomatic leiomyomata. J Vasc Interv Radiol. 2014; 25: 1737–1747

Post Author:
Charles Hyman, MS4
Chair, Communications Committee, SIRRFS
Warren Alpert Medical School of Brown University

Friday, April 13, 2018

Embolization for RCC: shift in treatment paradigm?


Researchers from Karolinska University Hospital in Stockholm have recently published their results of a prospective controlled trial on renal cell carcinoma (RCC) embolization. RCC is typically managed through nephron-sparing surgery, radical nephrectomy, ablation, or active surveillance. In this manuscript, the authors compared doxorubicin eluting embolic (DEE) to bland TAE. A total of 12 patients with average tumor size of 3.2 cm +/- 0.62 were randomized to receive DEE or TAE before a planned surgery. Size of embolic was determined on basis of degree of vascularity and ranged from 75-150 micron, 100-300 micron, and 300-500 micron. A CT was performed prior to surgical removal to assess treatment response. DEE transarterial chemoembolization (n = 6) resulted in a significantly (P = .018) higher degree of necrosis with an average of 88.3% compared with TAE (n = 5), which resulted in an average of 29.4%, as evaluated by CT. Histopathologic evaluation showed similar results (P = .016) with an average necrosis of 87.5% for DEE transarterial chemoembolization (n = 4) versus 26% for TAE (n = 5). Percentage of necrosis seen on microscopy correlated significantly (P = .0005) with radiologic findings, as 4 tumors in each arm were evaluated by both CT and microscopy. No major complications were observed in either group. The authors concluded that TACE is safe to be performed for localized RCC and has a significantly superior effect when compared with TAE.
Images of RCC treated by transarterial chemoembolization in patient 12. (a) Subtraction angiography image obtained after selective catheterization of the left renal artery showed contrast-filled intratumoral vessels (arrowhead) before transarterial chemoembolization. (b)Subtraction angiography image obtained during transarterial chemoembolization showed no contrast-filled vessels in the center of the tumor owing to ongoing embolization. The area adjacent to RCC (arrowhead) shares its blood supply with the tumor. (c) CT imageobtained before transarterial chemoembolization showed contrast enhancement in RCC (arrowhead). (d) CT image obtained 4 weeks after transarterial chemoembolization showed no contrast enhancement in RCC (arrowhead). Infarcted renal tissue (arrow) is adjacent to the treated tumor.


This manuscript is noteworthy as it opens the avenue for more catheter-based research for RCC locoregional treatment. The authors were able to demonstrate a statistically significant increased response to doxorubicin DEB compared to bland embolic. While both embolics contained residual tumor on pathologic evaluation, DEB had significantly better tumor response. This is interesting as systemic doxorubicin is not known to have a significant impact on RCC. Further, many tumors can become resistant to the effects of doxorubicin when they are ischemic (as with embolization). Clearly, there is more happening at the tumor level than we currently understand. However, is this all a moot point given the current treatment paradigm in RCC? There is a relatively low likelihood that a patient would not be a candidate for partial nephrectomy or ablation and have a tumor undergoing rapid enough growth to require palliative treatment. As such, this manuscript may be more impactful in what it says about the local effects of embolization and image-guided drug delivery and is unlikely to change RCC treatment algorithms.

Click here for abstract

Karalli A, Ghaffarpour R, Axelsson R, Lundell L, Bozoki B, Brismar T, Gustafsson O. Transarterial Chemoembolization of Renal Cell Carcinoma: A Prospective Controlled Trial. J Vasc Interv Radiol. 2018; 12:1664-1672

Post Author:
Luke R. Wilkins, MD
Assistant Professor
Department of Radiology and Medical Imaging
Section of Vascular and Interventional Radiology
University of Virginia

Thursday, March 29, 2018

From the SIR Residents and Fellows Section (SIRRFS)

Teaching Topic: Value of Antibiotic Prophylaxis for Percutaneous Gastrostomy: A Double-Blind Randomized Trial

Ingraham CR, Johnson G, Albrecht EL, et al. Value of antibiotic prophylaxis for percutaneous gastrostomy: a double-blind randomized trial. J Vasc Interv Radiol. 2018; 29: 55-61.

Since its development in 1980, percutaneous radiologic gastrostomy (PRG), also known as radiologically inserted gastrostomy (RIG) or simply percutaneous gastrostomy (PG), has become a common, safe, and effective option for the delivery of nutritional support to treat or prevent malnutrition in patients where oral intake is contraindicated. While surgical gastrostomy, performed endoscopically, known as percutaneous endoscopic gastrostomy (PEG), is the traditional gold standard, there are many advantages for fluoroscopically placed gastrostomy. In a meta-analysis, radiologic gastrostomy had a higher rate of successful placement than surgical (99.2% vs. 95.7%, p < .001) with a reduced major complication rate (5.9% vs. 9.4%, p <.001). [1] Additionally, radiological placement does not require general sedation and can avoid passage through the oropharynx preventing tumor seeding in head and neck cancer cases. Likewise, the avoidance of the oral passage and the prevention of contamination with oral flora is frequently cited as a major benefit of radiological placement. [2,3] In this study, Dr. Ingraham et al. conducted a double-blinded, randomized controlled trial to compare peristomal infection rates after prophylactic antibiotics in percutaneous gastrostomy.

The guidelines from the SIR Standards and Practice Committee recommends routine administration of prophylactic antibiotics in “pull technique” gastrostomy tube placement (transoral access), but did not find a consensus for prophylaxis administration with the “push technique” (transabdominal access). This study measured the rates of peristomal infection in 122 patients who received push-type gastrostomy placement in three study arms- placebo (n=34), prophylactic antibiotic administration (n=34), or in an observational group if they were already receiving antibiotics for another indication (n=68). All patients received a 16-F Deutsch gastrostomy tube (Cook Medical, Bloomington, Indiana), a pigtail catheter, placed over a stiff wire through an 18-gauge needle centered within 3 absorbable gastropexy sutures. Patients in the treatment arm received 1 g of intravenous cefazolin or 600 mg of intravenous clindamycin in those with allergies. A blinded evaluator examined the stoma site for subjective signs of infection at 3-5 days, 7-10 days, 14-17 days, and 28-30 days.

Because patients were lost to follow up or started on antibiotics for unrelated infections, analysis was restricted to the early time period (<10 days). During this period, 4/34 (11.8%) patients in the placebo arm and 0/34 (0.0%) in the treatment arm were found to have stoma site infections. This trend was not found to be statistically significant under intention-to-treat analysis (p = .057). However, under per-protocol analysis the difference between infection rate of 4/30 (13.3%) patients in the placebo arm and the 0/32 (0.0%) in the treatment arm was statistically significant (p=.049). Only one patient was found to have an infection in the observation arm, but compared to the placebo this difference was not significant in intention to treat (p=.078) or per protocol (p=.063) analysis.

Clinical Pearls

How can you tell if a stoma site is infected? How common is stoma site infections?

The authors of this study used an 11-point scale based on erythema, induration, and exudate that was originally developed for studying antibiotic prophylaxis in PEG placement. The points were distributed based on the “presence of erythema (0, none; 1, ≤ 5 mm; 2, 6–10 mm; 3, 11–15 mm; 4, ≥ 15 mm), induration (0, none; 1, ≤ 10 mm; 2, 11–20 mm; 3, ≥ 20 mm), and exudate (0, none; 1, small serous; 2, moderate serous; 3, large serous ± sanguineous; 4, purulent).”

According to the review of the literature in the SIR Antibiotic Prophylaxis guidelines, infections in transoral access range from 4%-30%, but infections in retrospective series of transabdominal access without antibiotics ranged from 0-2%. [4,5]. However, a retrospective review of head and neck cancer patients who underwent transabdominal gastrostomy found 15% infection rate among the patients that were not treated with prophylaxis. [6] Patients with head and neck cancer were not excluded from the study discussed in this Teaching Topic- 9 patients received placebo, 8 received treatment, and 15 were in the observation arm. It is possible this played a part in the relatively high (11.8-13.3%) infection rate found in this study.

What is the difference between the methods of radiologic gastrostomy?

While the SIR antibiotic guidelines only differentiate between the “push” and “pull” methods, there are many different variations on fluoroscopically placed gastrostomy tubes. For antibiotic consideration, traversing the oral cavity is likely the biggest consideration.

The classical radiologically inserted gastrostomy is a Seldinger technique that consists of insufflating the stomach through a nasogastric tube then percutaneously passing a needle into the stomach. A wire is then passed and a gastrostomy tube is then passed over the wire. This technique was first described in 1981 by Preshaw and modified to include percutaneous gastropexy by Brown et al. in 1986. [7,8] This is called the “push” method as you push the gastrostomy tube percutaneously into the stomach.

Transoral access allows for larger tube sizes, the same as placed endoscopically. This method utilizes a similar technique to enter the stomach before passing a catheter retrograde through the esophagus and out the oral cavity where a gastrostomy tube is advanced over the wire. This is what is referred to as the “pull” method as you pull the gastrostomy tube out of the stomach into its final position. It is standard of care to give antibiotics with this method.

Questions to Consider:

What is the difference between intention-to-treat and per protocol analysis?

Intention to treat (ITT) and per protocol are different methods for analyzing results of randomized controlled trials. ITT includes all subjects as originally randomized, regardless of completion of treatment, to compare outcomes of the groups. Per protocol only includes subjects who completed treatment and outcome assessment. ITT reduces the possible bias if attrition, non-compliance, or protocol mistakes were non-random. Broadly speaking, ITT measures the efficacy of the treatment protocol, while per protocol measures the outcomes of the specific treatment. In this study, patients were given antibiotics for other infections, thus violating the protocol and introducing a possible bias. These patients were excluded from the per protocol analysis which was found to be statistically significant. It is important to consider both ITT and per protocol when making clinical decisions based on trial data.

Additional Sources:

1. Wollman B, D’Agostino HB, Walus-Wigle JR, Easter DW, Beale A. Radiologic, endoscopic, and surgical gastrostomy: an institutional evaluation and meta-analysis of the literature. Radiology. 1995;197(3):699–704.

2. Shin JH, Park A-W. Updates on percutaneous radiologic gastrostomy/gastrojejunostomy and jejunostomy. Gut Liver. 2010;4(Suppl 1):S25–31.

3. Sutcliffe, J, Wigham, A, Mceniff, N, Dvorak P, Uberoi R. CIRSE Standards of Practice Guidelines on Gastrostomy. Cardiovasc Intervent Radiol. 2016;39(7):973.

4. Venkatesan, A.M., Kundu, S., Sacks, D. et al. Practice guidelines for adult antibiotic prophylaxis during vascular and interventional radiology procedures. Written by the Standards of Practice Committee for the Society of Interventional Radiology and Endorsed by the Cardiovascular Interventional Radiological Society of Europe and Canadian Interventional Radiology Association [corrected]. J Vasc Interv Radiol. 2010; 21: 1611–1630

5. Wollman, B. and D’Agostino, H.B. Percutaneous radiologic and endoscopic gastrostomy: a 3-year institutional analysis of procedure performance. AJR Am J Roentgenol. 1997; 169: 1551–1553

6. Cantwell, C.P., Perumpillichira, J.J., Maher, M.M. et al. Antibiotic prophylaxis for percutaneous radiologic gastrostomy and gastrojejunostomy insertion in outpatients with head and neck cancer. J Vasc Interv Radiol. 2008; 19: 571–575

7. Preshaw RM. A percutaneous method for inserting a feeding gastrostomy tube. Surg Gynecol Obstet. 1981;152(5):658–60.

8. Brown AS, Mueller PR, Ferrucci JT. Controlled percutaneous gastrostomy: nylon T-fastener for fixation of the anterior gastric wall. Radiology. 1986;158(2):543–5.

Post Author:
Charles Hyman, MS4
Chair, Communications Committee, SIRRFS
Warren Alpert Medical School of Brown University