eMethods. Statistical Analysis Plan
eTable. Classification System for Ulceration in Infantile Hemangiomas
eFigure. Distribution of Age at Ulceration
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Fernández Faith E, Shah S, Witman PM, et al. Clinical Features, Prognostic Factors, and Treatment Interventions for Ulceration in Patients With Infantile Hemangioma. JAMA Dermatol. 2021;157(5):566–572. doi:10.1001/jamadermatol.2021.0469
What are the prognostic factors and efficacy of treatment interventions in infantile hemangioma ulceration?
In this cohort study that included 436 patients, the median time for hemangioma ulceration to heal overall was 6.14 weeks; time to heal was shortest for topical treatment only, followed by systemic β-blocker, and by multimodal therapy and pulsed-dye laser, likely reflecting the severity of ulceration and treatments used for management. Among patients treated with propranolol, treatment at lower doses (≤1 mg/kg/d), but not higher doses was associated with a significantly faster healing time; the size of the infantile hemangioma was identified as a significant prognostic factor.
For ulcerated infantile hemangiomas that require systemic therapy, low-dose propranolol may be considered as an initial treatment approach.
Ulceration is a common complication of infantile hemangioma (IH), which leads to substantial morbidity. Ulceration in IH has not been systematically studied since the advent of β-blocker therapy for IH.
To examine treatment interventions used for ulceration in IH and identify clinical prognostic indicators of healing time.
Design, Setting, and Participants
A retrospective, multicenter cohort study was conducted on 436 consecutive patients with a clinical diagnosis of ulcerated IH and available clinical photographs. Patients receiving care at tertiary referral centers evaluated between 2012 and 2016 were included; statistical and data analysis were performed from February 7 to April 27, 2020.
Clinical characteristics, treatment interventions, course, complications, and resource use were analyzed. Treatment interventions for ulceration in IH included local (wound care, topical), systemic (β-blocker, corticosteroids), and procedural (pulsed-dye laser).
Main Outcomes and Measures
The primary end point was time to complete or nearly complete ulceration healing. Clinical characteristics were analyzed to determine the responses to most common interventions and prognostic factors for healing of ulceration.
Of the 436 patients included in the study, 327 were girls (75.0%); median age at ulceration was 13.7 weeks (interquartile range, 8.86-21.30 weeks). The median heal time was 4.79 weeks (95% CI, 3.71-5.86 weeks) with wound care alone, 5.14 weeks (95% CI, 4.57-6.00 weeks) with timolol, 6.36 weeks (95% CI, 5.57-8.00 weeks) with a systemic β-blocker, and 7.71 weeks (95% CI, 6.71-10.14 weeks) with multimodal therapy. After adjusting for IH size, a dose of propranolol less than or equal to 1 mg/kg/d was associated with shorter healing time compared with higher propranolol doses (hazard ratio, 2.04; 95% CI, 1.11 to 3.73; P = .02). Size of the IH was identified as a significant prognostic factor for healing time in multivariable analysis. Increasing size of IH portends a proportionately longer time to heal of the ulceration.
Conclusions and Relevance
Despite the use of β-blockers, this cohort study found that a subset of patients with IH ulceration continued to experience prolonged IH healing times. Larger IH size appears to be a poor prognostic factor for time to heal. For patients requiring systemic therapy, initiation of propranolol at lower doses (≤1 mg/kg/d) should be considered.
Ulceration is a common complication of infantile hemangiomas (IHs) and can result in significant morbidities including pain, bleeding, risk for infection, functional impairment, and scarring.1,2 The pathogenesis of IH ulceration is poorly understood; evidence regarding optimal management is lacking. To our knowledge, there has been no large study of IH ulceration since β-blocker therapy for IH has become widespread. Despite evidence that β-blocker therapy can decrease the frequency and severity of ulceration, IH ulceration continues to be a complication and management challenge in a substantial minority of patients.3-6 Our study aim was to characterize IH ulceration in a large cohort during the “β-blocker era,” including characteristics of IH ulceration, treatment interventions, and healing times.
This retrospective cohort study included patients with ulcerated IH seen between 2012 and 2016 at 8 collaborating centers of the Pediatric Dermatology Research Alliance. Institutional review board approval was obtained at each center. Informed consent was waived because the study involved no more than minimal risk to the subjects and the research could not be carried out practicably without the waiver due to the retrospective nature of the study. Inclusion criteria were clinical diagnosis of ulcerated IH and available clinical photographs to analyze the clinical features of the IH and ulceration. A study manual was created to ensure uniformity in the assessment of clinical features for data collection. Information regarding patient demographics, clinical characteristics, treatment of ulceration, course, and complications were collected in a standardized data collection form and transferred via REDCap to the data coordinating center (Nationwide Children’s Hospital, Columbus, Ohio). Patients with unreliable or incomplete data on ulceration healing were excluded in time-to-heal analysis but included in the analysis of demographics and clinical features. For patients with more than 1 ulcerated IH, only data for either the first or largest IH to ulcerate were analyzed. Data analysis was performed from February 7 to April 27. 2020. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
The primary outcome was time to heal in weeks, defined as complete or near-complete re-epithelialization. All photographs were reviewed by 2 of us (E.F.F., I.J.F.) to ensure reliable classification of the clinical characteristics. Healing time was measured from the first visit for ulceration to the follow-up visit or medical record documentation of healing.
A standardized classification of IH ulceration had not previously been established. To characterize ulceration type, we developed and used a classification for IH ulceration that was included in the study manual to support their use (eTable in the Supplement). Aggressive ulceration was defined as meeting 2 of the following 3 criteria: significant soft-tissue destruction, worsening despite multimodal therapy, and healing time longer than 12 weeks.
Five treatments groups were considered for analysis: wound care only (barrier ± dressing ± topical antibiotic), topical timolol (±wound care), systemic β-blocker (±wound care), multimodal (combination of at least 2 of the following: topical timolol, systemic β-blocker, and surgery), and pulsed-dye laser (PDL) alone or in combination with other therapies.
Data are summarized using frequencies with percentages and medians with interquartile range (IQR). Group comparisons were evaluated using χ2 or Fisher exact tests as appropriate for categorical variables and Kruskal-Wallis tests for continuous variables, with Benjamani-Hochberg corrections for multiple comparisons. Univariate and multivariable proportional hazards regression were used to assess factors associated with shorter heal times, and multivariable logistic regression was used to evaluate factors associated with the odds of complications, prior receipt of treatment, ulcer recurrence, and aggressive ulcers. Variables were considered for multivariable assessment if their univariate P value was <.20, and variables were retained in the model if their adjusted P value was <.10 or if their inclusion had a substantial impact on model goodness of fit (based on Akaike information criterion). Treatment group was retained in all models regardless of statistical significance because it was the primary independent variable of interest. For statistical analysis, treatment group was considered the primary risk factor of interest for healing of ulceration. All analyses were conducted using R, version 4.0.2 (R Foundation for Statistical Computing), with the aid of the following packages: compareGroups, MASS, Amelia, sjPlot, ggplot2, rms, survival, and survminer. The statistical analysis plan is available in the eMethods in the Supplement.
A total of 436 patients were enrolled including 327 girls (75%) and 109 boys (25%). The median age at ulceration was 13.7 weeks (IQR, 8.86-21.30 weeks). Ulceration developed within the first 4 months of life in 244 infants (55.9%) and 362 (83%) before age 6 months. Age at ulceration is detailed in the eFigure in the Supplement. Other demographics and clinical characteristics are detailed in Table 1.
Infantile hemangioma–specific treatment had been initiated before ulceration in 64 (17.6%) patients. Of these, topical timolol (42 [65.6%]) and systemic β-blockers (24 [37.5%]) were most common. The median lapse time between initiation of treatment and development of ulceration was 6 weeks (IQR, 3-15 weeks). In multivariable analysis, prematurity (adjusted odds ratio [OR]2.56; 95% CI, 1.35-4.76; P = .004) and older age at ulceration (adjusted OR, 1.05; 95% CI, 1.02-1.08; P = .001) were associated with greater odds of ulceration during active therapy.
After excluding 69 patients who were lost to follow-up or lacked information necessary to calculate time to healing of ulceration, 367 patients remained. Three additional patients were excluded: 1 who was missing treatment group information and 2 who received surgical intervention only for treatment of ulcerated IH. These exclusions left a total of 364 patients with appropriate data for the primary analysis. Demographic factors associated with longer healing time included preterm delivery, larger size of the IH, and size of the ulceration. However, after accounting for size of the IH, no other factors, including treatment group, remained significantly associated with healing time. Larger IH size was associated with longer time to heal in weeks. Compared with IH size less than or equal to 5 cm2, ulceration in IH size greater than 5 cm2 to less than or equal to 10 cm2 took 29% longer time to heal (hazard ratio [HR], 0.714; 95% CI, 0.54-0.94; P = .02), ulceration in IH size greater than 10 cm2 to less than or equal to 50 cm2 took 39% longer to heal (HR, 0.608; 95% CI, 0.47-0.79; P < .001), and ulceration in IH size greater than 50 cm2 to less than or equal to 100 cm2 took 60% longer to heal (HR, 0.402; 95% CI, 0.22-0.74; P = .003). Compared with IH size less than or equal to 5 cm2, those with very large (>100 cm2) IHs were not significantly different with respect to healing times, but this finding could be due to the relatively small sample size among those with very large IHs.
Treatment for IH ulceration varied widely; many patients had more than 1 treatment, either simultaneously or in succession. Timolol and wound care were more likely to be given for predominantly superficial IH. Multimodal and PDL treatment were less likely to be given for predominantly superficial IH. Wound care and PDL were more common for localized IH, timolol for indeterminate IH, and multimodal therapy for segmental IH. Systemic β-blockers were more commonly given for IH on the head and neck, while timolol was more commonly given for IH in the diaper area. There were no significant differences in size of IH or ulceration characteristics between the treatment groups.
Median time to heal overall was 6.14 weeks (IQR, 3.9-10.2 weeks). Unadjusted survival analysis showed significant differences between treatment groups (Table 2). The median heal time was 4.79 weeks (95% CI, 3.71-5.86 weeks) with wound care alone, 5.14 weeks (95% CI, 4.57-6.00 weeks) with timolol, 6.36 weeks (95% CI, 5.57-8.00 weeks) with a systemic β-blocker, and 7.71 weeks (95% CI, 6.71-10.14 weeks) with multimodal therapy. Patients in the timolol group had similar heal times as those in the wound care group, and shorter heal times compared with those in the multimodal and PDL groups. Differences in heal time by treatment group identified on univariate assessment remained after accounting for IH size. The median follow-up time for the care of IH ulceration was 6.7 weeks (IQR, 4-10.6 weeks).
Among the patients treated with propranolol plus or minus wound care, different dosing regimens were compared. There were 87 patients who received propranolol and had dosing information available. Unadjusted survival analysis did not demonstrate a significant difference between the dosing regimens. After adjusting for IH size, a dose of propranolol, less than or equal to 1 mg/kg/d, was significantly associated with shorter healing time; treatment with a propranolol dose greater than 1 mg/kg/d was associated with 2.04 times longer healing time compared with a propranolol dose of less than or equal to 1 mg/kg/d (HR, 2.04; 95% CI, 1.11-3.73; P = .02) (Table 3).
Data on ulceration recurrence were available for 361 patients. Recurrence was present in 41 patients (11.4%) and was more likely in patients presenting with ulceration at a younger age (OR, 0.96 per week increase in age; 95% CI, 0.92-1.00; P = .06) and located in the diaper area (OR, 7.08; 95% CI, 3.50-15.34; P < .001). Recurrence was less likely in IHs with localized vs indeterminate morphologic characteristics (OR, 0.27; 95% CI, 0.12-0.61; P = .002), localized vs segmental morphologic characteristics (OR, 0.28; 95% CI, 0.12-0.62; P = .001), head and neck location (OR, 0.23; 95% CI, 0.08-0.55; P = .003), and the presence of eschar in the initial ulceration (OR, 0.11; 95% CI, 0.02-0.38, P = .003). After adjusting for IH size, age at presentation, and presence of eschar, there were no differences in the odds of recurrence by treatment group.
Information on aggressive IH ulceration was available in 363 patients, with 35 (9.6%) meeting criteria for aggressive ulceration, as defined in the eTable in the Supplement. Aggressive ulceration was associated with multiple gestation (OR, 3.00; 95% CI, 1.29-6.61; P = .008), segmental vs localized morphologic characteristics (OR, 4.00; 95% CI, 1.81-9.08; P = .001), larger size of ulceration in diameter (>1 vs ≤1 cm2: OR, 3.54; 95% CI, 1.45-10.8; P = .010), and percentage of IH surface involved (>25% vs ≤25%: OR, 3.82; 95% CI, 1.80-8.89; P < .001). There was a greater likelihood of multimodal therapy in this group, reflecting the many treatments used to treat persistent severe ulceration.
Secondary complications, including pain, bleeding (usually mild), and interference with daily life were noted in most patients (Table 4). Higher rates of complications were noted in the diaper area (OR, 3.54; 95% CI, 1.71-8.39; P < .001), with ulceration size greater than 1 cm2 (OR, 4.22; 95% CI, 2.37-7.69; P < .001) and with percentage of ulcerated surface greater than 25% (OR, 3.12; 95% CI, 1.70-6.04; P < .001). Head and neck ulcerations were associated with a lower incidence of secondary complications (OR, 0.42; 95% CI, 0.24-0.75; P = .003).
This study provides insights into IH ulceration, with a study period after use of β-blocker therapy for IH that had become widespread. The results support previously known associations, such as age at ulceration, distribution, and higher risk locations, and add information regarding clinical features, complication rates, clinical prognostic indicators, and treatments.
Knowledge of time of onset and duration of healing can be helpful in anticipatory guidance. The median age of ulceration was 13.7 weeks, with most developing by age 4 months, highlighting that ulceration is a complication primarily of the proliferative phase of IH. However, some ulcerations developed later, implying that active proliferation is not the sole explanation for ulceration.
Treatment interventions for IH ulceration have evolved. Wound care is a cornerstone in the management of IH ulceration. Almost invariably, all patients with IH ulceration receive some form of wound care, which by itself may be sufficient to promote ulceration healing. Since the use of β-blockers for IH, some evidence has emerged that β-blockers are helpful in decreasing ulceration or improving healing time. This evidence includes one prospective study7 and a few small retrospective studies, case series, and case reports.8-10 The average time to heal in the retrospective studies ranged from 40 to 60 days, while the prospective study showed a shorter average time to heal of 17 days. Uncertainty about best treatment options is evident when other studies of non–β-blocker interventions are considered. The average time to ulceration healing without the use of β-blockers has varied from 40 to 86 days.11,12 More recently, a study comparing the efficacy of propranolol vs corticosteroids for IH ulceration did not show a clear advantage of either treatment intervention.5 A systematic review of medical, surgical, and wound care management for ulcerated IH failed to uncover a specific intervention associated with faster healing times.6
Ulceration in IH does not occur in isolation; it most often develops during the active proliferation phase of IH, a time in which active intervention to prevent other complications (eg, deformity, functional impairment) is often being initiated. Although a single intervention that provides the greatest efficacy in the healing of IH ulceration has not been clearly identified, the risk of other complications must be carefully considered. For patients with ulcerated IH who require systemic therapy, initiation of propranolol at a low dose (≤1 mg/kg/d) should be considered. At the same time, the risk of systemic absorption must be taken into account in ulcerated IH treated with topical β-blockers.
Nearly one-fifth of the patients in our study developed ulceration after active therapy was initiated. Associated risk factors for this incidence included prematurity and older age at presentation. Some authors have hypothesized that β-blocker therapy may induce or precipitate ulceration via vasoconstriction, leading to tissue ischemia with resultant skin breakdown.13 While our study does not establish a cause-effect relationship, it suggests that active treatment for IH proliferation does not always prevent ulceration from developing. In this context, the finding that doses of propranolol higher than 1 mg/kg/d resulted in longer healing times compared with lower doses is noteworthy. The concept that β-blockers can induce ulceration was suggested in a study of patients with PHACE (posterior fossa defects, hemangiomas, cerebrovascular arterial anomalies, cardiovascular anomalies including coarctation of the aorta, and eye anomalies) syndrome treated with propranolol in which some patients were noted to have induction or worsening of ulceration after initiating β-blocker therapy.13 Neither the exact mechanism of action of propranolol on IH nor why a lower dose might lead to improved outcomes for ulceration is well understood. Effects on vascular tone, inhibition of hemangioma stem cells, and modulation of angiogenesis, vasculogenesis, and apoptosis have been proposed to explain the benefits of propranolol in IH.14 Active therapy in these patients may have prevented more severe ulceration from developing; prospective studies would be required, ideally in patients at higher risk, to address these findings.
Pulsed-dye laser and other light devices are also used in the treatment of IH ulceration with variable efficacy.6 A recent randomized clinical trial did not find a significant difference in healing time between topical timolol and PDL for treatment of small ulcerations in IH.15 In our study, all patients treated with PDL also received wound care and β-blocker therapy in different combinations. There was also variability in the use of PDL with regard to timing of this intervention and number of treatments. Taking these factors into account, it is difficult to establish the benefits of PDL from this study.
The distribution of ulcerations, with two-thirds involving either the head and neck or diaper area, is relevant in understanding potential complications. Complications included pain, bleeding, interference with daily activities, and infections, and the diaper area had a higher risk of these complications. Although head and neck IH had a lower risk of these complications, ulceration almost invariably leads to scarring. This outcome was not captured in our study because we only looked at short-term complications. However, even small ulcerations in an anatomic location such as the central face can lead to disfigurement with potential negative effects on the psychosocial development.
There were several limitations to our study. The multicenter design and retrospective nature led to differences in clinical practices and timing of treatment interventions, thus affecting the measurement of healing times. Wound care and topical therapy are typically the first interventions in IH ulceration. Patients receiving other modalities were thus skewed toward more treatment-resistant cases. Differences in severity were addressed by controlling for IH size, although there may be other unrecognized factors associated with ulceration severity and time to heal. Timing of initiation of therapies, including systemic therapy, also likely varied based on local practices including the degree of monitoring used for initiation of systemic β-blocker treatment. The use of PDL to treat ulceration also varied. Although this study included a large number of infants, the number of patients in each treatment group was limited. Further prospective studies will be important to evaluate the efficacy of different treatment interventions in IH ulceration.
All study sites are tertiary referral centers, adding potential selection bias toward more severe ulcerations. The lack of a standardized classification schema of IH-induced ulceration severity was another potential limitation, one that we tried to overcome by creating one in which the ulceration was verified with photographs, allowing for more objective analysis of ulceration. We hope this classification may prove useful for future studies. Clinical photographs were taken as standard of care and were not standardized, which brings another limitation. In an effort to minimize variability in the clinical classification, all photographs were evaluated by 2 of us (E.F.F., I.J.F.). A category of uncertain was created to account for ulcerations that were difficult to classify based on the photographs.
This study assembled a large cohort to analyze specific clinical features of ulceration in IH, and in doing so to examine β-blocker therapy—the main form of IH treatment—in the setting of ulceration. We found that, although infants benefit from β-blocker therapy in managing IHs, many continue to experience prolonged healing times. For IHs requiring systemic therapy, low-dose propranolol (≤1 mg/kg/d) should be considered as an initial approach to IH ulceration, with slow upward titration of the dose to provide the known benefits for IH proliferation, ideally after healing is well under way. Controlled prospective trials are needed to examine dosing regimens and other aspects of ulcerated IH management to help improve outcomes for this significant complication of IH.
Accepted for Publication: February 9, 2021.
Published Online: March 31, 2021. doi:10.1001/jamadermatol.2021.0469
Corresponding Author: Esteban Fernández Faith, MD, Division of Dermatology, Department of Pediatrics, Nationwide Children’s Hospital, 700 Children’s Dr, Columbus, OH 43205 (email@example.com).
Author Contributions: Dr Fernández Faith had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Fernández Faith, Shah, Pope, Moore-Clingenpeel, Frieden.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Fernández Faith, Frieden.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Alsumait, Moore-Clingenpeel.
Obtained funding: Fernández Faith.
Administrative, technical, or material support: Shah, Bradley, Covelli, Nguyen, Hunt.
Supervision: Fernández Faith, Frieden.
Conflict of Interest Disclosures: Dr Frieden reported serving as a consultant for Venthera Biobridge, Noble Pharmaceutical, and Novartis, and as a member of the data safety monitoring board for Pfizer. No other disclosures were reported.
Funding/Support: This study was supported by Pediatric Dermatology Research Alliance grant 82233117.
Role of the Funder/Sponsor: The funding organization had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.