Otolaryngology–Head and Neck Surgery (2008) 138, 418-424

LITERATURE REVIEW

Nonsurgical therapies for lymphangiomas:

A systematic review

Jason L. Acevedo, MD, Rahul K. Shah, MD, FAAP,

and Scott E. Brietzke, MD, MPH, Washington, DC

OBJECTIVE: Systematically review the published literature re-garding the efficacy of nonsurgical therapies in the treatment ofhead and neck (H&N) lymphatic malformations (LM) in children.DATA SOURCE: MEDLINE.REVIEW METHODS: MEDLINE was searched for literaturerelating to nonsurgical treatments for H&N LM.RESULTS: The initial search returned 1876 articles, with 22meeting criteria. The majority (20) were case series. All therapieswere percutaneous, with OK-432 or bleomycin sclerotherapy be-ing most common. Random-effects modeling revealed 43% (CI �28.9%-57%) of patients undergoing OK-432 for LM achieved acomplete/excellent response, 23.5% (CI � 5.8%-41.3%) achieveda good response, 16.9% (CI � 10.3%-23.4%) achieved a fair/poorresponse, and 15.4% (CI � 8.6%-22.2%) observed no response. Inthe bleomycin group, the results were: 35.2% (CI � 15.7%-54.6%)excellent, 37.1% (CI � 22%-52.3%) good, 18.4% (CI � 2.7%-34.2%) fair/poor, and 11.6% (CI � 3.5%-19.6%) no response.Seven major complications were noted out of the 289 patients inthe series, including two mortalities.CONCLUSIONS: The literature indicates that sclerotherapy forH&N LM achieves excellent/good clinical response in a majorityof patients, with few complications, and anecdotally does notcomplicate future surgery.© 2008 American Academy of Otolaryngology–Head and NeckSurgery Foundation. All rights reserved.

Otolaryngologists frequently manage patients with headand neck lymphatic malformations (LM), also com-

monly referred to as lymphangioma and cystic hygroma.Approximately 75% of these malformations occur in thehead and neck. Roughly half are detected at birth, with 80%to 90% detected by the age of 2 years. The internationalincidence of LMs has been approximated to be anywherefrom 1 in 6000 to 1 in 16,000 live births. No racial or sexualpredilections have been demonstrated.1

From mild cosmetic deformity to life-threatening airwayobstruction, these malformations have myriad clinical pre-sentations and multiple classification schemes for LM exist.The most commonly used classification systems are those ofMulliken and Glowacki,2 deSerres et al,3 and McGill andMulliken.1 The Mulliken and Glowacki system is a general

Received August 28, 2007; revised October 17, 2007; accepted No-

vember 15, 2007.

0194-5998/$34.00 © 2008 American Academy of Otolaryngology–Head and Necdoi:10.1016/j.otohns.2007.11.018

system for categorizing vascular malformations, while de-Serres and McGill and Mulliken specifically delineate lym-phatic malformations.

Traditionally management involves surgical therapy,with the goal being removal of involved tissue withoutsacrificing vital structures. Due to the anatomical relation-ships in the head and neck, extirpation can be cosmeticallyand/or functionally morbid. Further, these lesions tend torecur, partially due to the fact that complete excision isfrequently impossible. These factors have been the drivingforce in the search for nonsurgical therapies for the man-agement of LM. Multiple nonsurgical therapies have beenproposed and include diathermy, cryotherapy, radiationtherapy, fibrin glue, and percutaneous sclerosants. Sclero-sants that have been used include sodium morrhuate, dex-trose, hypertonic saline, tetracycline, doxycyline, aceticacid, ethanol, boiling water, and OK-432.4 The purpose ofthis study is to systematically review the published literature

Figure 1 Literature search flowchart.

k Surgery Foundation. All rights reserved.

419Acevedo et al Nonsurgical therapies for lymphangiomas . . .

Table 1

Evidence table

Author Therapy Study type LOEMean

age (y) Complications Author’s conclusions

Peters et al5 OK-432 Case series 4 3.78 None “We have found OK-432 to bea useful primary treatmentfor children withmacrocystic lymphaticmalformations.”

Luzzatto et al12 OK-432 Case series 4 2.92 None “. . . OK-432 remains our firstline treatment forlymphangioma . . .”

Baskin et al13 Bleomycin Case series 4 0.96 None “Bleomycin injection is asimple and effective method. . .”

Sichel et al14 OK-432 Case series 4 12.9 None “. . . OK-432 is an effectivetreatment modality formacrocystic-typelymphangiomas in the headand neck . . .”

Mathur et al15 Bleomycin Case series 4 3.78 None “. . . we now considerbleomycin sclerotherapybefore considering surgicalexcision . . .”

Won et al4 Acetic acid Case series 4 12.5 None “. . . sclerotherapy oflymphangiomas withdiluted acetic acid iseffective without seriouscomplications.”

Kim et al (a)16 OK-432 Cohort study 2b 7.53 None “We recommend thatsclerotherapy be viewed asa first-line treatmentmodality for these lesions.”

Kim et al (b)16 Bleomycin Cohort study 2b 1.44 One death frompulmonarycomplications, 3months aftertreatment

“We recommend thatsclerotherapy be viewed asa first-line treatmentmodality for these lesions.”

Rautio et al18 OK-432 Case series 4 11.39 Posttreatmentswelling in onepatient

“These facts encourage us tocontinue with OK-432treatment trials.”

Claesson et al19 OK-432 Case series 4 10.4 None “. . . OK-432 has so far provedto be effective for LM . . .”

Giguere et al11 OK-432 RCT 1b 4.12 One patient requiredorbitaldecompressionafter injection of aperiorbital LM,another requiredan emergencytracheostomy afterinjection of a largecervical LM

“OK-432 should be efficaciousin the treatment oflymphangiomas.”

Sung et al20 OK-432 Retrospectivereview

4 6.9 Transient facialnerve paralysis inone patient withparotid LM

“. . . we suggest tryingPicinibil sclerotherapy as aprimary treatment forLMHN.”

420 Otolaryngology–Head and Neck Surgery, Vol 138, No 4, April 2008

in order to ascertain the safety and efficacy of nonsurgicaltherapies for LM of the head and neck.

METHODS

The literature was systematically searched in the followingmanner. MEDLINE was searched using the search terms“treatment AND lymphangioma,” “treatment AND lym-phatic AND malformation,” and “treatment AND cysticAND hygroma.” The literature was searched from 1966 topresent. Limits were set for English-language articles andhuman subjects. The following criteria were then applied:number of subjects greater than or equal to five, related to

Table 1

(continued)

Author Therapy Study type LOE

Larrane et al21 OK-432 Case series 4

Luzzatto et al22 OK-432 Case series 4

Brewis et al23 OK-432 Case series 4

Castenon et al24 Fibrin glue Case series 4

Zulfiqar et al21 Bleomycin Case series 4

Greinwald et al6 OK-432 Case series 4

Dubois et al9 Ethibloc Case series 4

Schmidt et al22 OK-432 Case series 4

Orford et al24 Bleomycin Case series 4

Sung et al23 Bleomycin Case series 4

LOE, level of evidence.

nonsurgical therapies for lymphatic malformations of the

head and neck, and ability to extract data specifically forhead and neck lesions from the study (Fig 1).

Evidence tables were then compiled using the retainedarticles (Tables 1 and 2).5-24 Annotated were the number ofsubjects, mean age and age range, level of evidence (www.cebm.net), treatment modality, and results of treatment.Based on its usage in many of the articles, the followingformat was used to stratify results: A complete response isdefined as the inability to clinically and/or radiologicallydetect the lymphatic malformation. A good response wasdefined as 50% or greater decrease in the size of the lesion.A fair/poor response was defined as a response with lessthan 50% decrease in the size of the lesion. No response wasdefined as a complete failure of the lesion to respond to the

n(y) Complications Author’s conclusions

1 None “OK-432 is safe andeffective.”

5 None “We consider OK-432 safeand easy . . .”

9 One child hadposttreatmentupper airwayobstructionrequiringintubationandaspiration

“OK-432 injection is effectiveand safe as a first or secondline treatment . . .”

8 None “Puncture, aspiration, andinjection of fibrin adhesivemay be highly useful . . .”

None “. . . bleomycin injection issafe and effective . . .”

7 None “Picinibil is a safe andeffective form of therapy. . .”

3 None “. . . we recommendsclerotherapy with ethiblocas the initial treatment inmacrocystic and mixedlymphangiomas.”

Temporaryintubationafter injectionof a largecervical LM

“. . . OK-432 represents analternative, safe andeffective treatment forlymphangiomas.”

None “. . . safe and effectivealternative to surgery.”

4 One patientdied frompneumonia

“. . . a useful medicalalternative, and we believethat it should be attemptedbefore excision of the lesion

Meaage

6.2

3.3

5.0

3.4

N/A

1.8

4.2

4.5

N/A

1.4

treatment. Also annotated were patients who failed therapy

421Acevedo et al Nonsurgical therapies for lymphangiomas . . .

and underwent surgical excision, as well as patients whosuffered complications of therapy (eg, infection, airwaycompromise, and death). Finally, the authors’ conclusionswere noted.

The data was then statistically analyzed using the fol-lowing techniques. Crude pooled analysis was first per-formed, followed by random-effects modeling to estimatesummary treatment success rates, taking into account inter-study and intrastudy variance. The standard error for eachstudy was estimated using the inverse of the sample size. Allstatistical analyses were performed with the assistance ofcomputer software (Intercooled STATA V8.2, College Sta-tion, TX).

Table 3

Random-effects modeling results

Sclerosant ResultResponserate (%)

OK-432 Excellent 43OK-432 Good/fair 23.50OK-432 Poor 16.90OK-432 None 15.40Bleomycin Excellent 35.20Bleomycin Good/fair 37.10Bleomycin Poor 18.40Bleomycin None 11.60

Table 2

Summary table of articles reporting on sclerosing ther

Author Year Therapy nCe

Peters et al5 2006 OK-432 10Luzzatto et al12 2005 OK-432 17Sichel et al14 2004 OK-432 11Kim et al16 2004 OK-432 25Ruatio et al18 2003 OK-432 10Claesson et al19 2002 OK-432 27Giguere et al11 2002 OK-432 29Sung et al20 2001 OK-432 20Larrane et al21 2002 OK-432 8Luzzatto et al22 2000 OK-432 13Brewis et al23 2000 OK-432 11Greinwald et al6 1999 OK-432 12Schmidt et al22 1996 OK-432 5Baskin et al13 2005 Bleomycin 7Mathur et al15 2005 Bleomycin 7Kim et al16 2004 Bleomycin 10Zulfiqar et al21 1999 Bleomycin 11Orford et al24 1995 Bleomycin 13Sung et al23 1995 Bleomycin 10Won et al4 2004 Acetic acid 6Castenon et al24 1999 Fibrin glue 18Dubois et al9 1997 Ethibloc 9Totals 289 1

RESULTS

The initial search returned 1876 articles (Fig 1). Thirty-threearticles remained after application of exclusion criteria. An-other 11 were excluded because head and neck–specific datacould not be extracted. After these exclusions, 22 articlesremained (Table 1). Looking at the quality of the evidencepresent in the existing literature, one study was level 1b(randomized controlled trial), one study was level 2b (co-hort study), and the majority (20) were level 4 (case series).Sample sizes ranged from 5 to 29, with an average samplesize of 13. Demographic data were first compiled from eachstudy, revealing that average age at treatment was 5.7 years

Confidenceinterval (%) P value

Q (test forheterogeneity)

28.9-57 0.001 210.495.8-41.3 0.009 339.556

10.3-23.4 �0.001 41.5978.6-22.2 �0.001 45.355

15.7-54.6 �0.001 27.32722-52.3 �0.001 16.4742.7-34.2 0.022 17.7133.5-19.6 0.005 2.136

te/nt Good Fair/poor

Noresponse

Furthersurgery

0 3 3 63 1 3 11 0 3 2

10 5 3 65 1 0 1

23 4 0 01 2 8 02 3 3 04 0 0 02 4 0 22 5 2 01 4 3 20 0 2 20 1 0 11 4 0 35 3 1 35 0 2 05 0 2 16 2 1 52 0 0 00 1 0 17 0 0 0

85 43 36 36% 29.41% 14.88% 12.46% 12.46%

apies

omplexcelle

4107740

1812472436214614

172

2543.25

422 Otolaryngology–Head and Neck Surgery, Vol 138, No 4, April 2008

(n � 263), 57% were male (n � 257), and 59.2% of LMwere macrocystic (n � 150) (note: n is the number ofsubjects in which this variable was annotated). Macrocysticversus microcystic LM status was specified in only 152/289patients. In studies where it was annotated, crude analysisdemonstrated that 92% of patients with macrocystic mal-formations had a good or excellent response and 50% ofpatients with microcystic or mixed malformations had asimilar good/excellent response. deSerres staging was an-notated in one study, with 6 patients being classified as typeI, 3 type II, 11 type III, 2 type IV, and 3 type V. McGill andMulliken staging was annotated in four studies, with 34being type I, 10 type II, and 21 type I and II. Crude pooledanalysis of sclerotherapy-treated subjects (n � 289) showedthat OK-432 had a complete/excellent response rate of41.4% and bleomycin a rate of 34.5%. This trend was notfound to be statistically significant (Fisher’s exact � 0.392).Of the pooled group of 289 children, only 36 (12.5%) werereported to require surgical salvage. In patients who under-went surgical salvage, use of a sclerosing agent was anec-dotally reported not to cause extensive scarring in any case,and did not make surgery more difficult. Several authors com-mented that patients with prior attempts at therapy (surgical,sclerotherapy, etc) tended to have a lesser chance at successfulsclerotherapy. It was also noted that infrahyoid macrocysticlesions had a higher cure rate that any other lesion.

Random-effects modeling of the overall success rates fortherapy was performed and results are summarized in Table 3.Analysis of the random-effects modeling results shows thatthe majority of patients were observed to have a good toexcellent response following sclerotherapy with either bleo-mycin or OK-432. Representative Forest plots for “excel-lent” clinical response are shown in Figure 2 (OK-432) andFigure 3 (bleomycin) (Forest plots for “good,” “fair,” and “noresponse” constructed but not shown). Fibrin glue, acetic acid,and ethibloc were not included in the statistical analysis due toa lack of data (single case series for each).

Seven major complications were noted. Two children whounderwent bleomycin therapy died posttreatment from pulmo-nary complications. It is unclear whether their deaths wererelated to pulmonary toxicity from bleomycin. In the OK-432

Figure 2 Forest plot of patients achieving an “excellent” clin-

ical result following sclerotherapy with OK-432.

group, two children required emergent intubation for airwaycompromise, one child required an emergent tracheostomy,one required orbital decompression after injection of a perior-bital LM, and one child had transient facial nerve paralysisafter injection of a parotid LM. Minor side effects were morecommon, including pyrexia (with OK-432), local pain, andlocal erythema being most frequent (specific numbers were notdenoted by authors for minor side effects).

DISCUSSION

The systematic review reveals that the majority of the cur-rent biomedical literature demonstrates a high level ofeffectiveness for the nonsurgical treatment of LM with theuse of percutaneous sclerotherapy. The most experience wasreported with OK-432 and bleomycin, with both producinggood to excellent responses in the majority of patients. As afirst-line therapy, complete regression was noted in 43.3%of cases. A good result was noted in another 29.4%. Overall87.5% of all patients in this review demonstrated a responseof some sort, and only 12.5% went on to require salvagesurgery. These facts, coupled with the fact that sclerother-apy was not associated with a significant increase in surgicaldifficulty or morbidity, suggests that sclerotherapy may bean appropriate first-line therapy for lymphatic malforma-tions, with surgery reserved for refractory cases.

The vast reported experience with percutaneous therapyfor LM was with OK-432 and bleomycin sclerotherapy.Fibrin glue, acetic acid, and ethibloc show promise, but theyhave only been studied in single case reports. OK-432showed a trend toward being the most efficacious sclero-sant. It had relatively few side effects, other than transientpyrexia, and demonstrated a higher excellent clinical re-sponse rate than bleomycin. However, in statistical analysis,a significant difference between excellent/complete re-sponses to OK-432 vs bleomycin was not demonstrated.OK-432 may be superior to bleomycin based on trends ofresults and less serious complications, but more study is

Figure 3 Forest plot of patients achieving an “excellent” clin-ical result following sclerotherapy with bleomycin.

needed.

423Acevedo et al Nonsurgical therapies for lymphangiomas . . .

Major complications reported from the use of sclerother-apy were infrequent. However, given their severity, carefulreview is warranted. Seven major complications were notedfrom seven studies, five of which were from OK-432 andtwo from bleomycin. Two children in the bleomycin groupdied from pulmonary complications, although a definitivelink to bleomycin therapy was not established.

There are several weaknesses that exist within this study.First, as is typical with systematic review and meta-analysis,there was significant heterogeneity even between differentstudies using the same sclerosant in terms of dosing, treat-ment procedures, and follow-up. Random-effects modelingwas selected to try to account for this heterogeneity incalculating summary effect estimates. Secondly, subjectiveendpoints of treatment effectiveness were utilized in themajority of studies (ie, “excellent response,” “good re-sponse,” “fair response,” etc) and hence were used withinthis systematic review to compile and compare resultsacross studies. There is a possibility for wide variability asto which category a patient would be assigned to aftertreatment, which would of course introduce variance anderror into the results. Third, the overall level of evidencewas fair to poor. Twenty of 22 studies were case series withonly 2 higher-level studies being available. Finally, a ques-tion could be raised as to whether aspiration of lesion is theinitial common pathway to regression, rather than the injec-tion of a sclerosing agent. None of the studies employed aplacebo control of this type.

Both OK-432 and bleomycin appeared to be efficaciousand safe in the majority of patients, but several airwayemergencies did occur, and there were two mortalities in thebleomycin group. Any discussion of the use of these agentsshould include these complications. Further, the two mor-talities in the bleomycin group may lead clinicians to favorOK-432.

Although sclerotherapy as the initial therapy for pediatricLM seems to hold promise, the current level of supportingevidence is only fair. Randomized, placebo-controlled trialscomparing surgery to sclerotherapy as well as betweendifferent sclerosants could further elucidate an evidence-based algorithm for the treatment of LM of the head andneck.

CONCLUSIONS

The published literature strongly suggests that the majorityof patients who undergo sclerotherapy with OK-432 orbleomycin as first-line therapy for head and neck LM willachieve a good to excellent clinical response. Serious com-plications and the need to progress to surgical salvage wereinfrequent. However, the preponderance of evidence is lim-ited to case series, with randomized studies being needed to

further elucidate optimal treatment algorithms.

ACKNOWLEDGEMENT

The opinions and assertions of the authors contained herein are the privateviews of the authors and are not to be construed as reflecting the views ofthe Department of Defense or the Department of the Army.

AUTHOR INFORMATION

From the Department of Otolaryngology–Head and Neck Surgery (DrsAcevedo and Brietzke), Walter Reed Army Medical Center; and the Di-vision of Otolaryngology (Dr Shah), Children’s National Medical Center.

Presented at the Annual Meeting of the American Academy of Otolaryn-gology–Head and Neck Surgery, Washington, DC, September 16-19, 2007.

Corresponding author: Rahul K. Shah, MD, FAAP, Division of Otolaryn-gology, Children’s National Medical Center, 111 Michigan Avenue, NW,Washington, DC 20010.

E-mail address: rshah@cnmc.org.

AUTHOR CONTRIBUTIONS

Rahul K. Shah, study design, data collection, writer, critical review; JasonL. Acevedo, study design, data collection, writer, critical review; Scott E.Brietzke, study design, data collection, writer, critical review.

FINANCIAL DISCLOSURE

Dr Shah: Daiichi-Sankyo, speaker’s bureau.

REFERENCES

1. McGill TJ, Mulliken JB. Vascular anomalies of the head and neck. In:Cummings CW, Friedrickson JM, Harker LA, et al, editors. Otolaryngology–head and neck surgery, vol 1. 2nd ed. St. Louis: Mosby; 1993. p. 333–46.

2. Mulliken JB, Glowacki J. Hemangiomas and vascular malformationsin infants and children: a classification based on endothelial charac-teristics. Plast Reconstr Surg 1982;69:412.

3. deSerres LM, Sie KCY, Richardson MA. Lymphatic malformations ofthe head and neck: a proposal for staging. Arch Otolaryngol HeadNeck Surg 1995;121:577–82.

4. Won JH, Kim BM, Kim CH, et al. Percutaneous sclerotherapy oflymphangiomas with acetic acid. J Vasc Interv Radiol 2004 Jun;15(6):595–600. [Erratum in: J Vasc Interv Radiol 2004 Aug;15(8):874.]

5. Peters DA, Courtemanche DJ, Heran MK, et al. Treatment of cysticlymphatic vascular malformations with OK-432 sclerotherapy. PlastReconstr Surg 2006;18(6):1441–6.

6. Greinwald J, Burke D, Sato Y, et al. Treatment of lymphangiomas inchildren: an update of Picibanil (OK-432) sclerotherapy. OtolaryngolHead Neck Surg 1999;121(4):381–7.

7. Dubois J, Garel L, Abela A, et al. Lymphangiomas in children: per-cutaneous sclerotherapy with an alcoholic solution of Zein. Radiology1997;204(3):651–4.

8. Giguere CM, Bauman NM, Smith RJ. New treatment options forlymphangioma in infants and children. Ann Otol Rhinol Laryngol2002;111(12 Pt 1):1066–75.

9. Luzzatto C, Lo Piccolo R, Fascetti Leon F, et al. Further experiencewith OK-432 for lymphangiomas. Pediatr Surg Int 2005;21(12):969–

72. Epub 2005 Oct 25.

424 Otolaryngology–Head and Neck Surgery, Vol 138, No 4, April 2008

10. Baskin D, Tander B, Bankaoglu M. Local bleomycin injection in thetreatment of lymphangioma. Eur J Pediatr Surg 2005;15(6):383–6.

11. Sichel JY, Udassin R, Gozal D, et al. OK-432 therapy for cervicallymphangioma. Laryngoscope 2004;114(10):1805–9.

12. Mathur NN, Rana I, Bothra R, et al. Bleomycin sclerotherapy incongenital lymphatic and vascular malformations of head and neck. IntJ Pediatr Otorhinolaryngol 2005;69(1):75–80.

13. Kim KH, Sung MW, Roh JL, et al. Sclerotherapy for congenitallesions in the head and neck. Otolaryngol Head Neck Surg 2004;131(3):307–16.

14. Rautio R, Saarinen J, Laranne J, et al. Treatment of lymphangiomaswith OK-432 (Picibanil). Cardiovasc Intervent Radiol 2003 Jan-Feb;26(1):31–6. Epub 2002 Dec 20.

15. Claesson G, Kuylenstierna R. OK-432 therapy for lymphatic malfor-mation in 32 patients (28 children). Int J Pediatr Otorhinolaryngol2002 Aug 1;65(1):1–6.

16. Sung MW, Lee DW, Kim DY, et al. Sclerotherapy with Picibanil(OK-432) for congenital lymphatic malformation in the head and neck.Laryngoscope 2001;111(8):1430–3.

17. Laranne J, Keski-Nisula L, Rautio R, et al. OK-432 (Picibanil) therapy

for lymphangiomas in children [review]. Eur Arch Otorhinolaryngol2002;259(5):274–8.

18. Luzzatto C, Midrio P, Tchaprassian Z, et al. Sclerosing treatment oflymphangiomas with OK-432. Arch Dis Child 2000;82(4):316 – 8.

19. Brewis C, Pracy JP, Albert DM. Treatment of lymphangiomas of thehead and neck in children by intralesional injection of OK-432 (Pici-banil). Clin Otolaryngol Allied Sci 2000;25(2):130–4.

20. Castanon M, Margarit J, Carrasco R, et al. Long-term follow-up ofnineteen cystic lymphangiomas treated with fibrin sealant. J PediatrSurg 1999;34(8):1276–9.

21. Zulfiqar MA, Zaleha AM, Zakaria Z, et al. Treatment of lymphangi-omas in children: an update of Picibanil (OK-432) sclerotherapy.Otolaryngol Head Neck Surg 1999;121(4):381–7.

22. Schmidt B, Schimpl G, Hollwarth ME. OK-432 therapy of lymphangi-omas in children. Eur J Pediatr 1996;155(8):649–52.

23. Sung MW, Chang SO, Choi JH, et al. Bleomycin sclerotherapy inpatients with congenital lymphatic malformation in the head and neck.Am J Otolaryngol 1995;16(4):236–41.

24. Orford J, Barker A, Thonell S, et al. Bleomycin therapy for cystichygroma. J Pediatr Surg 1995;30(9):1282–7.