A, Grade 1: superficial soft tissue triangle and columella infantile hemangioma (IH). B, Grade 2: superficial and subcutaneous nasal tip IH, involving 1 nasal subunit. C, Grade 3: superficial and subcutaneous nasal IH, involving 2 subunits (nasal dorsum and left nasal side wall). D, Grade 4: superficial and subcutaneous nasal IH, involving 3 or more subunits (nasal tip, dorsum, and bilateral nasal side walls) and nasal obstruction.
A, Grade 1 lesion (superficial right nasal side wall) treated with 2 mg/kg/d of propranolol for 2 months with tumor shrinkage and decreased erythema. B, The progress halted, so the parents opted for surgical excision. C, Four months after the operation, it is still a clinical grade 1 lesion.
A, Grade 4 lesion (mixed tip, superficial dorsum, side wall, and ala) treated with 2 mg/kg/d of propranolol for 4 months (B) and 10 months, with resolution of the deep components, but the superficial component remains, making it a clinical grade 1 lesion. Pulsed-dye laser may be warranted in the future to decrease the superficial erythema and thickness. C, After 8 months of propranolol treatment, at 14 months postpropranolol treatment. No laser treatment performed.
A, Grade 4 lesion (mixed tip, dorsum, and bilateral nasal side walls). B, Treatment with 2 mg/kg/d of propranolol for 10 months. There was residual fibrofatty tissue, so surgical excision was offered. C, Four months after the operation.
A, Grade 3 lesion (mixed tip and columella) treated with 2 mg/kg/d of propranolol for 24 months. B and C, Now a grade 2 lesion isolated to the nasal tip with minimal residual deep irregularities seen on anteroposterior and basal views.
A, Grade 3 lesion involving subcutaneous and superficial nasal dorsum and side wall treated with 2 laser treatments followed by 2 surgical excisions. B, Postoperatively, this is a grade 1 lesion (residual erythema). C, Grade 4 lesion with subcutaneous and superficial involvement of the nasal dorsum, both sidewalls and soft tissue triangles, tip, and columella. This was treated with 3 laser treatments initially, followed by 2 surgical excisions, resulting in a grade 1 lesion. D, Six months following the second nasal surgery. Note red mark is a small unrelated abrasion.
Perkins JA, Chen BS, Saltzman B, Manning SC, Parikh SR. Propranolol Therapy for Reducing the Number of Nasal Infantile Hemangioma Invasive Procedures. JAMA Otolaryngol Head Neck Surg. 2014;140(3):220-227. doi:10.1001/jamaoto.2013.6524
Copyright 2014 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
Propranolol therapy is changing the treatment paradigm for infantile hemangioma. This study addresses the effect of propranolol therapy on the treatment of nasal infantile hemangioma (NIH), an area that often does not respond to medical therapy.
To determine if propranolol treatment is associated with fewer invasive treatments for NIH.
Design, Setting, and Participants
Retrospective cohort study conducted within a single pediatric institution’s multidisciplinary vascular anomaly program for patients with NIH treated between January 1, 2003, and December 31, 2011. Three NIH cohorts were compared: prepropranolol (20 in group 1; 2003-2009), propranolol (25 in group 2; 2009-2011), and nonpropranolol (13 in group 3; 2009-2011) treatment.
Analysis of systemic medical, laser, or surgical therapies for NIH.
Main Outcomes and Measures
The study plan was created to detect a change in invasive therapy for NIH. Data collected included presenting age, sex, affected nasal subunits, infantile hemangioma morphologic characteristics, treatment type and number, and primary treating service. An NIH grading system, based on nasal subunit involvement, helped quantify treatment change. Descriptive statistics summarized data, and a Cox proportional hazards regression model evaluated propranolol use and the likelihood of invasive treatments (surgical excision or laser).
Of the 95 patients identified, 58 met inclusion criteria: 20 in group 1 (mean age, 4.8 months), 25 in group 2 (mean age, 4.9 months), and 13 in group 3 (mean age, 4.9 months). Nasal infantile hemangiomas involved the nasal tip subunit in 33 of 58 patients (56.9%). Eight of 13 patients (61.5%) in group 3 frequently had small NIH (grade 1). Patients in group 2 were less likely to undergo any invasive treatments (relative risk, 0.44; 95% CI, 0.27-0.73), have surgical excision only (0.45; 0.15-1.38), or undergo laser treatment only (0.44; 0.27-0.78) compared with those in group 1. Patients with higher-grade NIH had more medical or invasive therapy, but invasive procedures were carried out in each subgroup defined by grade.
Conclusions and Relevance
Patients with isolated propranolol-treated NIH were less likely to undergo invasive treatment, but despite its implementation, the need for invasive treatment was not totally supplanted by its use.
Infantile hemangioma (IH) is the most common benign tumor of childhood, affecting 10% of children younger than 1 year.1 Historically, IH was treated with corticosteroids in an attempt to slow aberrant angiogenesis presumed to be occurring in “proliferating” IH.2 Treatment outcomes with corticosteroids were inconsistent and associated with a high incidence of adverse effects.3 Since the serendipitous discovery of the pharmacologic nonselective β-adrenergic receptor blockade, propranolol consistently stops IH growth and induces almost complete regression in some.4 Propranolol has become the primary medical treatment modality for IH.5
Nasal IH (NIH), in particular, occupies the center of an infant’s face and is disfiguring because it distorts nasal shape and often causes skin discoloration.6- 9 Treatment with corticosteroids has not caused a significant reduction in NIH size or skin color. It is believed by some that an IH in the central face has different biologic development than IH in other areas of the face and body, making it less responsive to medical therapy and natural regression.10,11 Growth cessation of NIHs is typical following a period of growth, but rather than completely involute, larger “mixed” lesions leave behind subcutaneous fibrofatty tissue that causes persistent disfigurement. Surgical excision for NIH has been done to remove this subcutaneous portion, either during NIH growth or after NIH growth has ceased. Because of the lack of consistent NIH regression with medical therapy, previous reports of NIH treatment have focused primarily on surgical and/or laser techniques and the effect that timing of invasive treatment may have on success.7,9,12- 20
Propranolol is reported to induce regression of NIH. We undertook studies to describe our multidisciplinary multimodal (medical, laser, and surgical excision) experience with NIH treatment before and after the initiation of IH propranolol therapy.15,21 To determine whether the introduction of propranolol changed the clinical and treatment course of NIH, we selected a representative NIH cohort treated before our knowledge of propranolol IH therapy and compared it with patients with NIH after this time. Following 2008, rapid Internet information dissemination regarding propranolol and IH appeared to change our practice in several ways. We began seeing many patients with minor NIH at younger ages than were typical before 2008. The parents of many patients with NIH actively inquired about propranolol treatment, while some refused propranolol therapy for patients who were candidates. In addition, as we gained experience with IH propranolol therapy, we became less aggressive in our treatment approach to small NIHs and accepted that some NIHs did not need therapy. With this understanding, we conducted this study to examine our NIH treatment patterns.
Following approval by the Seattle Children’s Hospital Institutional Review Board, where consent was waived due to the retrospective nature of the study, we conducted a review of all patients evaluated for isolated NIH between January 1, 2003, and December 31, 2011, at Seattle Children’s Hospital. We hypothesized that (1) since the advent of IH propranolol treatment, the number of NIH operations has decreased, and (2) NIHs with subcutaneous extension persist, even following propranolol treatment.
We included patients with IH who were younger than 18 months at the initial evaluation, had isolated NIH, obtained complete photodocumentation, and received treatment between April 1, 2003, and March 31, 2012. Potential participants were excluded if they had another vascular anomaly, extensive IH with potential systemic effects, and less than 6 months of follow-up after the initial treatment. Extensive IHs include hemangiomatosis and PHACES syndrome (posterior fossa abnormalities, hemangiomas of cervical or facial regions, arterial cerebrovascular anomalies, cardiac defects, and eye anomalies); patients with significant nasofacial IH extension were excluded. Data collected on each patient included age at presentation, IH morphologic characteristic, number and dates of laser and/or surgical treatments, availability of propranolol at the time of evaluation, use of propranolol, and other medical treatments used (corticosteroids) and the primary managing service. We divided the patients into 3 groups: those who sought treatment in the prepropranolol era (group 1) and those did (group 2) and did not (group 3) receive propranolol in the postpropranolol era.
In addition, we developed a grading system for NIH to better characterize these lesions and help quantify therapeutic change. We created 4 grades based on the depth of lesion, nasal subunit involvement, and functional impairment (Figure 1). Grade 1 lesions were superficial IH only and involved any number of nasal subunits. Grade 2 lesions involved only 1 nasal subunit, having both superficial and subcutaneous components. Grade 3 lesions were both superficial and subcutaneous and involved 2 nasal subunits. Last, grade 4 lesions were both superficial and subcutaneous and involved 3 or more subunits or caused functional impairment, such as nasal stenosis. Photographs taken at initial presentation were graded separately from posttreatment photographs. Grade change was defined as the difference between the grades assigned before and after treatment, as judged through an independent review by 2 of us (B.S. and S.R.P.) not involved in treatment planning or medical or surgical therapy. When looking at posttreatment photographs, residual deep or subcutaneous components were defined as obvious subcutaneous irregularity with or without overlying skin changes.
To take into account the censoring that resulted from variable follow-up time, we used Cox proportional hazards regression models to evaluate the association between propranolol use and the need for subsequent treatment: surgical excision, laser treatment (Vbeam Pulsed Dye Laser; Candela Corporation), or both. In this type of analysis, a hazard ratio of less than 1 indicates a lower likelihood of a treatment in that population vs the comparison group. A hazard ratio of more than 1 means there was a greater likelihood of a subsequent treatment than in the comparison group; a 95% CI that does not include 1 indicates P < .05. We evaluated the potentially confounding effects of lesion grade and age at presentation and the application of other medical treatments on the association between propranolol uses (or treatment in the propranolol era) and need for subsequent treatments. Since we found no substantive difference between the adjusted and unadjusted risk estimates, we present the unadjusted risk estimates here. We separately evaluated the association within groups defined by age at presentation, initial grade of the lesion, and use of other medical treatments to account for potentially modifying effects of these factors on the association between the use of propranolol and the need for other treatments. We evaluated the risk of further treatment (surgical excision, laser therapy, or both) in patients treated with propranolol (group 2) vs patients in the prepropranolol era (group 1). In an effort to account for the fact that not all patients would have chosen propranolol treatment had it been available earlier, and since we cannot identify patients who would have taken propranolol had it been available, we also carried out an analysis comparing the need for subsequent treatment among patients seen in the era when propranolol was available (groups 2 and 3) with those who were seen before it was available (group 1). The rationale and benefits of this approach are similar to those of an intent-to-treat analysis in a randomized clinical trial.
Of the 95 patients with NIH who were identified, 58 met inclusion criteria. The median presenting age was 3.7 months (range, 1.3-14.3 months), and 45 (77.6%) were female. The morphologic characteristic of NIH was superficial only (grade 1) in 15 patients (25.9%), superficial and subcutaneous in 1 subunit (grade 2) in 23 (39.7%), superficial and subcutaneous in 2 subunits (grade 3) in 7 (12.1%), and evidence of functional compromise or involvement of 3 or more subunits (grade 4) in 13 (22.4%). In the study population, 66 laser procedures and 19 surgical excisions were performed. Twenty patients with NIH (34.5%) were treated before the advent of IH propranolol therapy (group 1). Of the 38 patients with NIH who sought treatment after propranolol therapy became available, 25 (65.8%) received propranolol (group 2). The mean length of propranolol treatment in group 2 was 7.6 months, all treated at a total dosage of 2 mg/kg/d. Patients with NIH were managed primarily by the following services: otolaryngology (38 [65.5%]), dermatology (13 [22.4%]), and a combination of dermatology and otolaryngology (7 [12.1%]) (Table 1 and Table 2).
The median age at presentation was similar between groups 1, 2, and 3 (Table 1). Nasal IH most commonly involved the nasal tip in all groups. Medical treatments with corticosteroids and/or surgical excision were less frequent in groups 2 and 3 compared with group 1. After the advent of IH propranolol therapy, no patient had more than 1 NIH surgical excision. The group of patients who used propranolol (group 2) had the lowest mean number of laser treatments per patient: 0.8 treatments per NIH (range, 0-3) vs 1.5 (range, 0-5) and 1.6 (range, 0-5) in groups 1 and 3, respectively.
Grade 3 and 4 NIHs were managed most frequently by the following services: otolaryngology or dermatology in combination with otolaryngology, as opposed to dermatology alone (Table 2). Higher grades of NIH were more likely to be excised surgically than were lower grades of NIH. Multiple laser treatments were used more frequently in NIHs managed primarily by a dermatologic or a combination of a dermatologic and otolaryngologic service than those handled by a otolaryngologic service alone.
Treatment patterns according to NIH grade are shown in Table 3. The median presenting NIH grade was higher in groups 1 and 2 than in group 3 (median, 2 [range, 1-4]; median, 2 [range 1-4]; and median, 1 [range 1-4], respectively). After treatment, NIH grade was reduced by at least 1 category in 8 of 15 (53.3%) group 1 NIHs, as well as 7 of 18 (38.9%) group 2 and 2 of 9 (22.2%) group 3 NIHs undergoing any therapy (medical, propranolol, or invasive procedures; Figures 2, 3, 4, and 5). In group 2, only 4 of 17 (23.5%) NIHs had complete regression of subcutaneous IH with propranolol treatment (Figure 3). Among patients who had an operation, complete subcutaneous excision of NIH was achieved in 8 of 9 (88.9%) (Figure 6). Patients with grade 1 NIH had invasive therapy less frequently than children with grade 2 or 3 NIH, usually based on parental choice. Multimodal therapy of medical, laser, and/or surgical therapy was used more frequently in grade 3 and 4 NIHs. In group 1, 9 of 20 (45.0%) patients with NIH had surgical excision; this changed to 5 of 25 (20.0%) in group 2 and 2 of 13 (15.4%) in group 3.
Patients treated after the availability of IH propranolol therapy (group 2 and combined groups 2 and 3) were less likely to receive any invasive treatment compared with those with NIH managed before IH propranolol therapy was introduced. When the risk of having a surgical excision was evaluated separately, children in group 2 had half the rate of subsequent surgical treatment compared with children in group 1 (hazard ratio, 0.45; 95% CI, 0.15-1.38), but these findings are statistically imprecise. This lower rate of surgical excision was not limited to group 2 but was seen among all children with IH who sought treatment after the availability of propranolol (hazard ratio, 0.39; 95% CI, 0.14-1.04). Similarly, children who required care after the introduction of propranolol were less likely to undergo laser treatment, and this was true among those treated with propranolol as well as groups 2 and 3 combined (hazard ratio, 0.44; 95% CI, 0.27-0.78 and hazard ratio, 0.75; 95% CI, 0.46-1.25, respectively; Table 4).
This study demonstrates that the advent of IH propranolol therapy had an effect on the treatment of NIH at our institution. While children with IH were less likely to require invasive treatments, propranolol therapy did not eliminate invasive therapy altogether. Higher grades of NIH treated with propranolol may not completely regress, and surgical excision is still done to remove persistent cutaneous and subcutaneous IH. Similarly, following propranolol therapy, there is still a role for lasers in the treatment of persistent cutaneous NIH.
The age at which our study population had their initial evaluation did not change with the advent of propranolol therapy at our center. Other studies of NIH treatment have shown similar presentation ages for NIH.6 Our impression that the shift from corticosteroids to propranolol for IH medical therapy had decreased the age of NIH presentation is not supported by these data. This is probably due to the frequent NIH involvement of the nasal tip subunit, since this is the most visible area of the face and is unlikely to regress spontaneously.22 For these reasons, historically and in our study, surgical excision for NIH often has been directed at removal of nasal tip IH7,12 (Figure 6).
Duration of IH propranolol therapy is currently determined by treatment response and patient family observation.5 This treatment principle and a standard propranolol dose were used in group 2 patients. Unpredictable IH regrowth can occur with propranolol cessation, so treatment is continued until IH proliferation ends. For these reasons, propranolol is used for more than 4 months in all cases. Once propranolol was introduced, corticosteroid use dramatically decreased in our and other vascular anomaly programs.5 Anecdotally, in our study population, most families opting for medical therapy wanted only propranolol. Further study is necessary to determine the optimal duration of propranolol therapy for NIH and IH.
The size and extent of NIH, as determined by NIH grade, were associated with invasive therapy when comparing propranolol-treated and untreated lesions. One important change was that propranolol-treated NIH (group 2) required a single surgical excision. Similar to other experiences, patients with NIH treated during the prepropranolol era (group 1) often had multiple surgical and laser procedures performed on the same lesion.6,16 Our NIH grading system has not been validated as a means of measuring treatment response, so results surrounding our data and changes in NIH grade need to be interpreted with caution. Following medical treatment with propranolol, the NIH grade was decreased in most patients with small superficial lesions (grades 1-2). When the presenting NIH grade was larger and involved subcutaneous tissues, posttreatment grade decreased more in group 2 than in group 3. It is our impression that propranolol stops NIH growth and usually causes partial NIH regression, particularly in larger mixed, cutaneous and subcutaneous tissues, which may be the reason posttreatment grade decreased more in group 2 NIH. It is likely that propranolol NIH treatment results in a detectable reduction of cutaneous NIH, but our grading system is not sensitive to detect change in the subcutaneous portion of the NIH, or this portion of the NIH does not regress as much as cutaneous IH. It seems that restricting NIH growth through propranolol treatment reduces the extent of NIH surgical excision or laser therapy, since lesion size, vascularity, and volume are reduced. These data support the idea that propranolol therapy is not always necessary for small (grade 1) NIH and that larger (grade 2 or 3) NIH partially regressing with primary propranolol therapy can be safely managed with adjunctive surgical excision, if desired.
When interpreting the results of this study, the reader should keep in mind several possible limitations: the number of patients in our study cohorts is small, resulting in statistical imprecision. We were unable to analyze secondary NIH treatment outcomes, such as treatment differences between specialties. The grading system used to group NIH is not validated; it was created to categorize the data and attempt to quantify NIH change with therapy. This is a retrospective study relying on medical record abstraction; this may have resulted in incomplete or imprecise documentation of the primary lesion, comorbidities, or previous treatment. We have no reason to suspect that this would result in systematic misclassification of either the exposure or the outcome. Therefore, any resulting misclassification would arise from a diluted treatment effect in which the true association is actually stronger than we report. We likely have some confounding by treatment indication since the severity and growth trajectory of a child’s lesion influence the parents’ decisions about what kind of treatment to pursue. However, results from our sensitivity analyses comparing all children treated in the era when propranolol was available (groups 2 and 3) were broadly similar to the findings of comparisons among children treated with propranolol (group 2) and those managed before the availability of propranolol (group 1).
The findings from this study suggest that the introduction of propranolol as a treatment for NIH corresponds to a reduction in the number of invasive treatments in this patient population. Future work to identify the optimal propranolol dose and treatment duration is warranted.
Submitted for Publication: June 26, 2013; final revision received November 5, 2013; accepted December 4, 2013.
Corresponding Author: Jonathan A. Perkins, DO, Division of Pediatric Otolaryngology–Head and Neck Surgery, Department of Surgery, Seattle Children’s Hospital, 4800 Sandpoint Way NE, Seattle, WA 98105 (firstname.lastname@example.org).
Published Online: February 20, 2014. doi:10.1001/jamaoto.2013.6524.
Author Contributions: Dr Chen and Mr Perkins had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Perkins, Chen, Manning, Parikh.
Acquisition of data: Perkins, Chen, Saltzman, Manning.
Analysis and interpretation of data: Perkins, Chen, Saltzman, Manning.
Drafting of the manuscript: Chen.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Saltzman, Manning.
Administrative, technical, or material support: Chen, Manning.
Study supervision: Perkins, Saltzman.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This study was presented at the meeting of the American Society of Pediatric Otolaryngology; May 26, 2013; Washington, DC.
Additional Contributions: Eden Palmer prepared the images for submission and Carrie Dimico assisted with manuscript preparation.