Article Text


Original research
Sclerotherapy for lymphatic malformations of the head and neck in the pediatric population
  1. Joanna H Tu1,2,
  2. Huy M Do3,
  3. Viraat Patel4,
  4. Kristen W Yeom5,
  5. Joyce M C Teng6
  1. 1College of Physicians and Surgeons, Columbia University, New York, New York, USA
  2. 2Department of Dermatology, Stanford University, Palo Alto, California, USA
  3. 3Departments of Radiology and Neurosurgery, Stanford University, Palo Alto, California, USA
  4. 4Department of Dermatology, Baylor College of Medicine, Houston, Texas, USA
  5. 5Department of Radiology, Stanford University, Palo Alto, California, USA
  6. 6Department of Dermatology, Stanford University, Palo Alto, California, USA
  1. Correspondence to Dr Joyce M C Teng, 700 Welch Rd 301G, Palo Alto, CA 94304, USA; jteng3{at}


Background Sclerotherapy is one of the most commonly used minimally invasive interventions in the treatment of macrocystic lymphatic malformations (LMs). Several different sclerosing agents and injection protocols have been reported in the literature, each with varying degrees of success. The safety and efficacy of the treatments have not been evaluated comparatively in the pediatric population.

Methods Chart review of pediatric patients with macrocystic/mixed head and neck LMs who underwent sclerotherapy using OK-432, doxycycline, or ethanolamine oleate at Lucile Packard Children's Hospital at Stanford during 2000–2014. Clinical evaluation and radiographic imaging were reviewed to assess lesion characteristics and response to sclerotherapy following each treatment session. The post-intervention clinical response was categorized as excellent, good, fair, or poor.

Results Among the 41 pediatric cases reviewed, 10 patients were treated with OK-432, 19 patients received doxycycline, and 12 patients received ethanolamine. In univariate analysis, different sclerosants had similar effectiveness after the first injection and final clinical outcome (p=0.5317). In multivariate analysis controlling for disease severity stage as well as disease characteristics (macrocystic vs mixed subtypes), different sclerosants also had similar effectiveness after the first injection (p=0.1192). Radiologic analysis indicated an 84.5% average volume reduction, with similar effectiveness between the different sclerosants (p=0.9910).

Conclusions In this series of LM cases treated at Stanford, we found that doxycycline, OK-432, and ethanolamine oleate sclerotherapy appear to have a similar safety and efficacy profile in the treatment of macrocystic and mixed LMs of the head and neck in the pediatric population.

  • Intervention
  • Malformation
  • Neck
  • Pediatrics
  • Vascular Malformation

Statistics from


Lymphatic malformations (LMs) are localized slow-flow vascular malformations that occur most commonly in the head and neck region. Complications include recurrent infections, pain, sudden swelling, hemorrhage, or airway compromise.1–3 LMs of the head and neck often require treatment because of complications, esthetic, or functional concerns. The optimal treatment remains unknown and currently there are no universal guidelines. Management options include conservative observation, surgical excision, sclerotherapy, and potentially disease-modifying medications such as sildenafil and the mTOR inhibitor, sirolimus.4 ,5 Surgical excision or debulking has previously been the mainstay for management of LMs, but complications including scarring, skin ulceration, and potential nerve damage have been reported. Unless the malformation is completely removed, recurrence is common following surgery. Recently, sclerotherapy has shown promising results compared with conventional surgical excision in treating macrocystic LMs in children.6 ,7 Several different sclerosing agents and injection protocols have been reported in the literature, each with varying degrees of success.6–18 A recent review of sclerosing agents could not identify a significantly superior sclerosing agent in terms of effectiveness, and the evidence highlighted that differences in complication rates may potentially be the deciding factor in the choice of sclerosant.19 This study reports a single-center experience of percutaneous sclerotherapy of macrocystic and mixed LMs using three different sclerosants (doxycycline, OK-432, and ethanolamine oleate) in pediatric patients with similar disease severity and characteristics and compares the safety and efficacy of these sclerosants.


Institutional review board approval was obtained for the study protocol. A chart review of the Stanford Translational Research Integrated Database Environment (STRIDE) database was performed to identify pediatric patients (≤18 years) with head and neck LMs who underwent percutaneous sclerotherapy between 2000 and 2014.

Lesion characteristics, the number of treatment sessions, and clinical examination were obtained through review of clinical records and radiographic imaging. The classification of LMs was based on the International Society for the Study of Vascular Anomalies (ISSVA) classification scheme.20 In this series, macrocystic LM was defined as malformation with >50% of cysts larger than 2 cm in diameter and mixed LM was defined as malformation with <50% of cysts larger than 2 cm in diameter. LMs were staged using the de Serres staging system.21 Stage I represents unilateral infrahyoid disease, stage II represents unilateral suprahyoid disease, stage III represents unilateral infrahyoid and suprahyoid disease, stage IV represents bilateral infrahyoid disease, and stage V represents bilateral infrahyoid and suprahyoid disease.

Post-intervention clinical response was evaluated by physical examination using previously reported sclerotherapy response criteria. The response was classified as excellent, good, fair, or poor.10 ,14 ,22–24 An excellent outcome was defined as lesions with tumid deformity that were extinguished and became symmetrical in appearance. A good response was defined as size reduction of >50% on examination (<50% of original size) and a fair response was equivalent to a slight reduction (>50% of original size). If the lesions were unchanged or worse, it was considered a poor outcome. Clinical success means either excellent or good response.

LM volume was obtained from MRI pre- and post-intervention. All patients underwent brain MRI at either 1.5 Tesla (T) (Signa; GE Medical Systems, Milwaukee, Wisconsin, USA) or 3.0T (Discovery MR750; GE Medical Systems). T1-weighted spin echo (SE) sequence with fat suppression (FS) and T2-weighted short tau inversion recovery (STIR) sequence were acquired. The MRI parameters at 1.5 T were: T1 SE FS (600/10) at 4 mm slice thickness 1 mm skip; T2 STIR (3100/40) at 4 mm slice thickness, 1 mm skip. The MRI parameters at 3 T were: T1 SE FS (650/10) at 5 mm slice thickness 0.5 mm skip; T2 STIR (5900/35) at 5 mm slice thickness, 0.5 mm skip. Post-gadolinium contrast T1 sequences were used to determine if the malformation was lymphatic, venous, or mixed venolymphatic, and T2 STIR sequences were used to measure volume. The volume of the LM was calculated by measuring the greatest dimensions in the transverse (T), anterior-posterior (AP), and longitudinal (L) axes and using the dimensions in the formula for ellipsoid volume:25 V=(T×AP×L)/2. A similar formula has also been used in other studies for volume calculation of various tumors.26–29 These dimensions were measured by an experienced observer using GE Medical System's Centricity program, an in-house image display tool. The entire component of the malformation visible on T2 STIR was measured. Percentage change was calculated between the pre- and post-intervention volumes. Statistical assessment for both clinical and radiologic response was completed with SAS and RStudio data software.


The choice of sclerosing agents depended on the specialty service physicians at Stanford University performing sclerotherapy, which included neurointerventional radiology, pediatric surgery, plastic surgery, and interventional radiology and the training that these subspecialists received, their personal experience, preference, and philosophy. The intervention protocols used were similar for all sclerotherapy procedures.

Treatments were given under general anesthesia for most cases with only a few cases done under moderate sedation. Sonographic guidance was used for procedures done in the operating room and sonographic and fluoroscopic guidance were used for procedures done in the neurointerventional radiology suite. During treatment the lesions were palpated, punctured under imaging guidance, and aspirated with a 16–25 gauge needle. An approximate equal volume of doxycycline (10 mg/mL of saline, maximum 20 mL), ethanolamine oleate 5% (5 mL per 1 mL of contrast, maximum 10 mL), or OK-432 (10 µg/mL of saline, maximum 20 mL) was then injected.

Sclerotherapy was generally an outpatient procedure unless airway compromise or pain control was a concern, in which case patients were admitted overnight for observation. Follow-up examinations were performed 6–8 weeks after the procedure for all patients.

Patients were evaluated with contrast-enhanced MRI, CT, or ultrasound pre- and post-treatment sessions. Post-treatment pain, erythema, swelling, and low-grade fever were not considered adverse events, but rather expected side effects of sclerotherapy. We defined major adverse events following sclerotherapy as wound infections requiring hospital admission and IV antibiotics, permanent nerve injury, hematomas requiring operative drainage, skin scarring, profound localized edema requiring hospital admission, airway obstruction requiring tracheotomy or intubation, or death directly related to treatment.


Forty-one pediatric patients (14 female, 27 male) underwent sclerotherapy during the study period because of intermittent swelling (n=19), anatomical asymmetry (n=15), pain (n=7), and functional impairment (n=5). The most common location was the neck (70.7%), elsewhere in the face (22.0%), oral (4.9%), and periorbital (2.4%). Ten patients had injections with OK-432, 19 had injections with doxycycline, and 12 had injections with ethanolamine (table 1). The median time to last follow-up was 11.1 months, with a range of 1–118 months. The median age of patients who had their first injection was 16 months for OK-432 (2.7 months–9.5 years), 12.8 months for doxycycline (0.07 months–15.2 years), and 29.5 months for ethanolamine (0.47 months–17.5 years).

Table 1

Sclerotherapy for head and neck lymphatic malformations (LMs) in children

The distribution of different LM stages (I–V) across the sclerosant groups were similar (p=0.0511). On pre-procedure imaging, 29 children (70.7%) were diagnosed with macrocystic disease and 11 children (26.8%) were diagnosed with mixed disease. There was a similar percentage of each subtype treated with each of the three sclerosants (p=0.3366).

The clinical success rate was similar between the three sclerosants. In univariate analysis, all three sclerosants had similar efficacy after the first injection, and the final clinical outcomes were comparable as well (p=0.5317). In multivariate analysis controlling for lesion stage and characters (macrocystic vs mixed), the three sclerosants were also shown to have similar efficacy after the first injection (p=0.1192), and no significant differences were seen in treatment responses between the macrocystic and mixed lesions (p=0.5116). There was no difference between LM stages (I–V) among those who achieved ‘excellent’ outcomes (p=0.4650) (table 1).

Volumetric analysis was performed on 27 of the 41 patients (65.8%) owing to lack of migration of imaging results to the electronic medical system for patients seen before 2004 or lack of standardized imaging protocols. The median time to post-treatment imaging was 8 months, with a range of 1.5–72 months. Of the 27 pre- and post-treatment images analyzed, 14 received doxycycline, 8 received ethanolamine oleate, and 5 received OK-432. Pre-intervention treatment volumes ranged from 5.75 mL to 966.8 mL, with an average volume of 132.5 mL. The average pre-treatment volume varied between the different sclerosants but was not statistically significant (p=0.6618). Volumetric improvement was similar between the three sclerosing agents, with an average reduction of 84.5%, consistent with published literature.8 ,9 Ethanolamine had an average reduction in volume of 85.7% and OK-432 and doxycycline had an average reduction of 84.6% and 83.7%, respectively. One-way analysis of variance of the volume reduction in each treatment group indicated that the final outcomes were comparable (p=0.9910). A categorical analysis was also performed, with a significant response defined as a volume reduction of ≥90%. Comparison of the sclerosing agents with significant versus insignificant responses also yielded similar results, with no statistically significant difference between the three sclerosing agents (p=0.0665).

Recurrence of LMs occurred in 3/10 patients (30%) in the OK-432 group, 2/19 patients (10.5%) in the doxycycline group, and 1/12 patients (8.3%) in the ethanolamine group (p=0.1367). All recurrences were partial recurrences, defined as an increase in size after an observed significant reduction in size with treatment, with the increase in size still less than the original LM (a complete recurrence). The time to recurrence ranged from 4 to 101 months. There were three peri-procedural complications: airway compromise necessitating intubation (n=1), infection (n=1), and hemorrhage (n=1) (table 1).

One patient did not receive an adequate amount of doxycycline due to a procedural complication resulting from needle dislodgement during injection. This patient experienced increased respiratory distress with stridor requiring intubation 9 days after the procedure but recovered fully without further complications.


Our retrospective analysis found no statistical difference in the effectiveness of the three sclerosants. OK-432 was found to have a good to excellent response rate in 80% of patients, followed by 78.9% with doxycycline and 67.0% with ethanolamine. In terms of volumetric improvement, ethanolamine had an average volume reduction of 85.7%, with OK-432 and doxycycline closely following at 84.6% and 83.7%, respectively. The outcome is consistent with that in the literature from many different sclerosants including OK-432, doxycycline, ethanolamine oleate, bleomycin, dextrose, ethanol, sodium tetradecyl sulfate, and acetic acid. The efficacy rates with these agents range from 62% to 100%.8 ,9 Published reports have shown that sclerotherapy is nearly four times more successful than surgical excision for treating LMs, at the same time carrying lower associated morbidity.10 ,17 ,30 The LM stage does not affect the clinical response (p=0.4650) to different sclerosants, and the overall complication rate was equally low in all groups. Previously published literature found advanced stages generally to be associated with an increased risk of postoperative complications when surgical excisions were performed;21 however, this was not the case in our current study of sclerosing surgery. The increased risks of complication due to conventional surgery may be explained by the extent and invasive nature of LMs in the head and neck region among patients with advanced disease. These lesions usually have poorly defined borders and are intertwined with nearby important anatomical structures, especially in the case of bilateral LMs (stage IV and V), therefore increasing the risk from surgical excisions. In situations where there is uncertainty or concern about possible high-flow components (ie, arteriovenous malformation or atypical hemangioma), head and neck angiography is performed for confirmation.

Although efficacious, the current use of sclerosants in the treatment of LMs is off-label with no Food and Drug Administration (FDA) approved specific indication for treatment of slow-flow vascular malformations.31 Ethanolamine, for example, is specifically indicated by the FDA for treatment of bleeding esophageal varices. Most sclerosants cause an acute endothelial inflammatory response that leads to transient and self-limiting symptoms such as swelling, mild-to-moderate discomfort, erythema, and low-grade fevers. The mechanism of sclerosing action for each sclerosant used in this study is discussed in more detail below.


Doxycycline is a tetracycline antibiotic that is relatively inexpensive and readily available as a sclerosant. The exact mechanism of action is not yet elucidated, but its sclerosing effects have been attributed to the inhibition of cell proliferation and matrix metalloproteinases, as well as the suppression of vascular endothelial growth factor induced during lymphangiogenesis and angiogenesis.32 It also leads to formation of dense adhesions and fibrosis as a result of deposition of collagen and fibrin.19 Sclerotherapy with doxycycline for LMs has produced encouraging results with efficacy ranging between 64% and 100%. There are more than 50% of patients in our series with excellent responses. It also has a favorable safety profile with complication rates below 10% of patients treated.7 ,10–12 The current series, including 11 patients previously reported by Nehra et al, showed sclerotherapy with doxycycline resulted in excellent or good results in 78.9% of patients with an average volume reduction of 83.7%. Generally, macrocystic LMs respond well to sclerotherapy whereas microcystic LMs are refractory to the treatment. However, some reports have shown that sclerotherapy with doxycycline can be effective in treating microcystic LMs as well.10


OK-432 (Picibanil) from an inactive strain of group A Streptococcus pyogenes is widely used outside the USA as a sclerosant for macrocystic LMs, and response rates have ranged from 58% to 91%.13–18 ,30 The proposed mechanism of sclerosing action with OK-432 is thought to be the induction of various cytokines which induce a local but sufficient inflammatory response to damage the endothelium.33 The results of the current study are consistent with previous reports; sclerotherapy with OK-432 in this series resulted in excellent or good responses in 80.0% of patients treated and an average volume reduction of 84.6%. Although this study reports on the results of patients treated with OK-432, it is not currently approved by the Federal Drug Administration and is only available in a trial setting.

Ethanolamine oleate

Ethanolamine oleate is a salt of an unsaturated fatty acid that has been approved by the FDA for sclerotherapy of esophageal varices. The proposed mechanism of sclerosing action is damage to the endothelium by inducing thrombosis.6 In the current series, sclerotherapy with ethanolamine oleate resulted in excellent or good results in 67.0% of patients but also had the greatest average reduction in volume at 85.7% (figures 1 and 2). Additionally, ethanolamine had fewer recurrences and complications in this series of patients. One of two children with a poor treatment response had multiple congenital anomalies. The patient's family decided to discontinue treatment after the initial session. The patient's airway was subsequently compromised as a result of disease progression. The results of some patients in this group have previously been reported by Alexander et al,6 which showed excellent or good results in about 80% patients. An average lesion reduction of 28% for all LMs was previous reported, which is significantly less than that observed in this series.

Figure 1

Image studies of a young child with painful swelling of the right cheek and face. Contrast-enhanced CT (A), axial short tau inversion recovery (STIR) MR (B), and coronal T1-weighted post-contrast MR (C) images depict a 6.9×4.7×8.5 cm cystic mass in the right parotid gland with enhancing septations and fluid levels consistent with a predominantly macrocystic lymphatic malformation (LM). Note the deep component of this LM approximating the oropharynx.

Figure 2

Follow-up MRI performed 3 months after percutaneous sclerotherapy with ethanolamine. Axial T2 (A) and coronal T1-weighted post-contrast (B) images show very minimal residual lymphatic malformation (>98% volume reduction). Mass effect and facial deformation have disappeared.

The limitations of this study include the small number of patients since head and neck LMs are rare, its retrospective design, and lack of standardized clinical data collection during follow-up, making comparisons between agents difficult. Furthermore, there are many agents currently in use, and the choice of sclerosing agent often varies based on institution and clinician training. For example, bleomycin has been reported as an effective sclerotherapy agent for LMs,34 ,35 but this institution has limited experience with bleomycin due to satisfactory results with the agents discussed in this paper. As data with the various sclerosing agents increase, more in-depth comparisons in our patient population will be possible.

Another limitation was the subjective grading system based on physical examination. With the more objective system of assessing volumetric differences using imaging pre- and post-treatment, a major limitation was the lack of standardization in post-treatment imaging time. However, both short and extended time to imaging follow-up could underestimate the volume reduction with sclerotherapy: a short follow-up time of 1 month may not give the treatment enough time to have effect while an extended follow-up time could demonstrate increased volume as the malformation grows with the patient, even though the relative size may be reduced. An additional key limitation in this study is the small number of patients with a standardized imaging protocol for comparison. The lack of clear imaging for 14 of the patients in this study highlights a unique challenge in LM research in the pediatric population; in some cases, parents were reluctant to have their child undergo additional general anesthesia for a post-treatment MRI when they could perceive an excellent response. This reluctance is understandable, and underscores the importance of creating a validated assessment of treatment response that may be less intrusive, such as ultrasound combined with clinical examination. Efforts have been made at this institution to incorporate bedside ultrasounds at the multidisciplinary vascular anomalies clinic to better track progress over time.


Doxycycline, OK-432, and ethanolamine oleate sclerotherapy appear to be safe and effective as a primary treatment modality for macrocystic and mixed LMs of the head and neck in the pediatric population, although the recurrence rate is slightly higher with OK-432. Doxycycline remains the agent of choice for microcystic disease when sclerosing surgery is considered, although non-invasive medical treatments have been shown to be promising. A larger scale clinical trial is needed to confirm the safety of doxycycline, OK-432, and ethanolamine oleate as well as their efficacy in comparison with other sclerosing agents.


Shufeng Li and Jack Cackler performed data analysis.


View Abstract


  • Contributors Each author made a material contribution to the article, its revision and the final approval of the article for submission.

  • Competing interests None declared.

  • Ethics approval Ethics approval was obtained from Stanford Institutional Review Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.