Document downloaded from http://www.elsevier.es, day 28/03/2024. This copy is for personal use.

Any transmission of this document by any media or format is strictly prohibited.

Exploration of the optimal time to

discontinue propranolol treatment in
infantile hemangiomas: A
prospective study
Luying Wang, MD, PhD,a,b Wenjing Wang, MM,b Zifu Zhou, MD, PhD,b Jing Li, MD, PhD,a Zhiyu Li, MD, PhD,b
Renrong Lv, MD,a Guangqi Xu, MD, PhD,a Jianhai Bi, MD, PhD,a and Ran Huo, MDa

Background: Relapse of infantile hemangiomas after withdrawal from propranolol treatment is common.
Early withdrawal is believed to increase the risk of relapse.

Objective: The objective of this study was to determine the optimal time to discontinue propranolol
treatment for infantile hemangiomas.

Methods: A prospective study conducted at a tertiary referral center.

Results: Compared to withdrawal after 1-month maintenance treatment, withdrawal after 3-month
maintenance, corresponding achieving maximum regression of infantile hemangiomas, was associated
with a lower major relapse rate (P = .041). The relapse (P = .055) and adverse event rates (P = .154)
between the 2 withdrawal modes were not statistically significant. Compared with direct withdrawal, the
relapse (P = .396), major relapse (P = .963), and adverse event rates (P = .458) of gradual withdrawal were
not statistically different. Patients with/without relapse could be best distinguished according to whether
withdrawal followed a 3-month maintenance and age [13 months (area under the receiver operating
characteristic curve = 0.603). Patients with/without major relapse could be best distinguished according to
whether withdrawal was accompanied by 3-month maintenance (area under the receiver operating
characteristic curve = 0.610).

Limitations: The limitations of this study are nonrandomization and single-center design.

Conclusions: The optimal propranolol withdrawal time to avoid relapse is when the patient is aged
[13 months and the lesion has maintained for 3 months after reaching maximum regression, while the
optimal time to prevent major relapse is after 3 months of maintenance. ( J Am Acad Dermatol
2024;90:783-9.)

Key words: adverse event; clinical research; infantile hemangioma; propranolol; relapse; withdrawal.

INTRODUCTION 10%.1,2 Although there is a spontaneous trend for

Infantile hemangiomas (IHs) are the most common regression in IHs, complications, such as ulceration,
benign tumors in infants, with an incidence of 4% to bleeding, and compression of adjacent tissues, may

From the Department of Burn and Plastic Surgery, Shandong Accepted for publication December 19, 2023.
Provincial Hospital Affiliated to Shandong First Medical Univer- Correspondence to: Ran Huo, MD, Department of Burn and Plastic
sity, Jinan, Chinaa; and Department of Burn and Plastic Surgery, Surgery, Shandong Provincial Hospital Affiliated to Shandong
Shandong Provincial Hospital, Cheeloo College of Medicine, First Medical University, No 324, Jing Wu Rd, Jinan, 250021,
Shandong University, Jinan, China.b China. E-mail: huoran@email.sdu.edu.cn.
Drs Wang and Wang contributed equally to this work. Published online December 29, 2023.
Funding sources: This work was supported by the National Natural 0190-9622/$36.00
Science Foundation of China [grant number 82172227]. Ó 2023 by the American Academy of Dermatology, Inc.
Patient consent: Not applicable. https://doi.org/10.1016/j.jaad.2023.12.034
IRB approval status: Reviewed and approved by the Ethics
Committee of Shandong Provincial Hospital; approval
#2020-020.

783
Document downloaded from http://www.elsevier.es, day 28/03/2024. This copy is for personal use. Any transmission of this document by any media or format is strictly prohibited.

784 Wang et al J AM ACAD DERMATOL

APRIL 2024

occur during the rapid proliferative phase. Treatment and follow-up

Disfigurement may occur after IHs regression.3 All the enrolled patients began routine propran-
Therefore, the development of effective treatments olol treatment at an initial dose of 1 mg/kg/d,
for problematic IHs is vital. As a first-line treatment for which was increased to 2 mg/kg/d after 1 week if
IHs, propranolol has an efficacy rate of more than no obvious adverse event (AE) occurred. The dose
90%.4,5 However, relapse after treatment termination of 2 mg/kg/d was maintained throughout treat-
is also notable, and has been reported to range 10% to ment, including the 1-3 months maintenance
30%.6-8 In general, it is period, until the medication
believed that earlier with- was stopped. Drugs were
drawal of propranolol is CAPSULE SUMMARY administered twice a day,
related to relapse of IHs,1 during or within half an
Prolonged propranolol treatment may
d

but the specific time for with- hour after meals. Color
reduce the risk of infantile hemangiomas
drawal has not been estab- Doppler ultrasound and
relapse, but the optimal time to
lished. This prospective study clinical examinations were
discontinue propranolol is unclear.
was conducted to compare conducted during outpa-
the effects of different with- The optimal time for withdrawing
d
tient reviews. IHs were eval-
drawal modes of propranolol propranolol was when the patients were uated by ultrasound with
on relapse and to explore the older than 13 months and the lesions regard to volume and blood
optimal time for withdrawal maintained maximum regression for flow signal, which referred
from propranolol. 3 months. to the proportion of blood
flow area out of the total
METHODS area and was divided into
Study design and patients grade 0 (0%), 1 (0% to 20%), 2 (20% to 50%), and 3
This nonrandomized concurrent controlled trail ($50%). The degree of IHs regression was evalu-
was conducted at Shandong Provincial Hospital from ated by 2 physicians independently based on color
June 2020 to May 2023. This study was approved by Doppler ultrasound and clinical examinations and
the hospital’s ethics committee (#2020-020) and was a consensus was reached through discussion with
registered in the Chinese Clinical Trial Registry the third investigator in case of disagreement. The
(ChiCTR2000033415). Patients with IHs who maximum regression was considered when the
required propranolol treatment were included in volume change in the ultrasound was less than
the study after informed consent was obtained from 10%, the blood flow signal grade remained un-
their parents. Exclusion criteria were as follows: (1) changed, and no significant change was found in
patients older than 6 months, (2) patients who had the clinical examination. Propranolol withdrawal
previously received other treatments, and (3) pa- mode was determined according to the degree of
tients with propranolol contraindications, such as regression of IHs and the patient group. Efficacy in
cardiac disease, pulmonary disease, and allergy to medication termination was evaluated using the
propranolol. The enrolled patients were assigned to Achauer system (grade I, 75% to 100% regression;
groups A, B, C, or D referred to their parents’ grade II, 50% to 75% regression; grade III, 25% to
intentions. Patients in groups A and C continued 50% regression; and grade IV, 0% to 25% regres-
the medication for 1 month after reaching maximum sion).9 Follow-up was conducted via Internet and
regression and then discontinued the treatment outpatient review occurred for up to 6 months after
directly or gradually, respectively. The latter referred medication withdrawal. Throughout the study, the
to halving the dosage per week and discontinuing investigators in charge of follow-up after with-
medication within 4 weeks. Patients in groups B and drawal were blinded to withdrawal modes.
D continued the medication for 3 months and sub-
sequently stopped it either directly or gradually, Outcome measurements
respectively. Considering the lower enrollment effi- The primary outcome was relapse rate up to
ciency impacted by the COVID-19 pandemic, patient 6 months after withdrawal. Secondary outcomes
enrollment was stopped 18 months after the begin- included the major relapse rate up to 6 months after
ning of the study to ensure patient homogeneity and withdrawal and the AE rate throughout the entire
to reduce bias. The general characteristics of the study. Relapse referred to any degree of flushing,
enrolled patients were collected, including sex, age enlargement, or increase in blood flow as measured
at treatment start, number and location of lesions, using color Doppler ultrasound. Major relapse
morphologic pattern and type of IHs, ulceration, and referred to significant changes in color, volume,
area and thickness of lesions. and blood flow, requiring further treatment.
Document downloaded from http://www.elsevier.es, day 28/03/2024. This copy is for personal use. Any transmission of this document by any media or format is strictly prohibited.

J AM ACAD DERMATOL Wang et al 785

VOLUME 90, NUMBER 4

and their lesions are shown in Table I. Ultrasound

Abbreviations used:
images of typical lesions are shown in
AE: adverse event Supplementary Figs 1 to 3, available via Mendeley
AUC: the area under the receiver operating char-
acteristic curve at https://data.mendeley.com/datasets/gwnx93d3
IHs: infantile hemangiomas vy/1. No significant differences in general character-
istics were found among the groups (Table I).

Primary outcomes
Statistical analysis
Relapse rates in the 4 groups were 41.3%, 34.0%,
Differences in general characteristics between the
44.8%, and 20.6%, respectively. Differences in
groups were analyzed to ensure comparability.
relapse rates among the 4 groups were not significant
Groups A and C were integrated into 1-month
(P = .149, Table II). The relapse rate in the group with
maintenance group, whereas groups B and D were
a maintenance period of 1 month (groups A and C)
combined into 3-month maintenance group.
was 42.4% and that in the group with a maintenance
Similarly, groups A and B formed a direct withdrawal
period of 3 months (groups B and D) was 28.4%. The
group, while groups C and D were integrated as
difference in relapse rates between the two groups
gradual withdrawal group. Relapse, major relapse,
was not significant (P = .055, Table II). The relapse
and AE rates were compared between the groups.
rate was 38.2% in the direct withdrawal group
In addition, we explored the effects of the general
(groups A and B) and 31.7% in the gradual with-
characteristics of patients and IHs, efficacy of pro-
drawal group (groups C and D). No significant
pranolol, and withdrawal timing on relapse and
differences were found in the relapse rate between
major relapse. Specifically, x 2 test, Fisher’s exact
the 2 groups (P = .396, Table II).
test, or Mann-Whitney U test was used to analyze the
effects of general characteristics and treatment effi-
Secondary outcomes
cacy. Mann-Whitney U test was performed to
The major relapse rates in the 4 groups were
compare the treatment duration and age at treatment
17.5%, 10.6%, 24.1%, and 5.9%, respectively. There
termination in patients with or without relapse and in
were no significant differences in major relapse rates
patients with or without major relapse. Moreover, to
between the 4 groups (P = .156, Table II). The major
explore the effects of different withdrawal criteria on
relapse rate in 1-month maintenance group (groups
relapse and major relapse, we grouped the patients
A and C) was 19.6%, which was higher than in 3-
according to whether the treatment duration was
month maintenance group (groups B and D, 8.6%,
longer than 6/7/8/9/10/11/12/13 months; whether
P = .041, Table II). The major relapse rates were
the age at treatment termination was more than 9/10/
14.5% and 14.3% in the direct withdrawal group
11/12/13/14/15 months; whether withdrawal was
(groups A and B) and in the gradual withdrawal
with 3 months of maintenance and with a medication
group (groups C and D), respectively. No significant
duration longer than 6/7/8/9/10/11/12/13 months;
differences were found in the major relapse rate
and whether withdrawal was with 3 months of
between the 2 groups (P = .963, Table II).
maintenance and with an age older than 9/10/11/
AE rates in groups A to D were 22.6%, 17.9%,
12/13/14/15 months. Finally, we evaluated the area
35.7%, and 16.7%, respectively. The difference in AE
under the receiver operating characteristic curve
rates among the 4 groups was not significant
(AUC) for the different withdrawal criteria to deter-
(P = .280, Table II). Meanwhile, there were no
mine the optimal withdrawal mode. Statistical ana-
significant differences in AE rates between the group
lyses were performed using SPSS v25.0 and MedCalc
with 1 month of maintenance (groups A and C,
v20.1.4 by researchers blinded to the regimens.
27.2%) and the group with 3 months of maintenance
treatment (groups B and D, 17.4%, P = .154, Table II).
RESULTS In addition, no significant difference between the
Study population direct withdrawal group (groups A and B, 20.7%)
In total, 164 patients were enrolled between June and the gradual withdrawal group (groups C and D,
2020 and November 2021. Among them, 5 patients 25.9%) was observed (P = .458, Table II). The AEs
withdrew from the study and 9 patients were lost to observed in our study are listed in Supplementary
follow-up. The reasons for withdrawal are listed in Table II, available via Mendeley at https://data.
Supplementary Table I, available via Mendeley at mendeley.com/datasets/gwnx93d3vy/1. All AEs
https://data.mendeley.com/datasets/gwnx93d3vy/1. were grade 1 or 2 according to Common
The remaining 150 patients with 173 lesions were Terminology Criteria for Adverse Events v5.0 and
analyzed. The general characteristics of the patients resolved spontaneously without medication
Document downloaded from http://www.elsevier.es, day 28/03/2024. This copy is for personal use. Any transmission of this document by any media or format is strictly prohibited.

786 Wang et al
Table I. General characteristics of enrolled patients and lesions
With 1 mo of With 3 mo of Direct Gradual
Characteristics Group A Group B Group C Group D P-values maintenance maintenance P-values withdrawal withdrawal P-values
Patients N = 53 N = 39 N = 28 N = 30 N = 81 N = 69 N = 92 N = 58
Sex, no (%) .580* .935* .821*
Male 18 (34.0) 17 (43.6) 12 (42.9) 9 (30.0) 30 (37.0) 26 (37.7) 35 (38.0) 21 (36.2)
Female 35 (66.0) 22 (56.4) 16 (57.1) 21 (70.0) 51 (63.0) 43 (62.3) 57 (62.0) 37 (63.8)
Age at .780y .410z .667z
treatment
initiation
Median 2.80 (2.05-4.40) 3.50 (2.20-4.40) 3.85 (1.73-5.55) 2.70 (2.30-5.20) 2.9 (2.05-4.60) 3.00 (2.30-4.90) 3.00 (2.10-4.38) 2.95 (2.28-5.20)
(IQRs), mo.
Number of .483x .289* .178*
lesions (%)
1 43 (81.1) 34 (87.2) 25 (89.3) 28 (93.3) 68 (84.0) 62 (89.9) 77 (83.7) 53 (91.4)
＞1 10 (18.9) 5 (12.8) 3 (10.7) 2 (6.7) 13 (16.0) 7 (10.1) 15 (16.3) 5 (8.6)
Lesions N = 63 N = 47 N = 29 N = 34 N = 92 N = 81 N = 110 N = 63
Morphologic .711x .733* .441*
pattern,
no (%)
Localized 47 (74.6) 31 (66.0) 20 (69.0) 26 (76.5) 67 (72.8) 57 (58.1) 78 (70.9) 46 (73.0)
Segmental 11 (17.5) 10 (21.3) 5 (17.2) 3 (8.8) 16 (17.4) 13 (16.0) 21 (19.1) 8 (12.7)
Indeterminate 5 (7.9) 6 (12.8) 4 (13.8) 5 (14.7) 9 (9.8) 11 (13.6) 11 (10.0) 9 (14.3)
Type, no (%) .674x .239x .572x
Superficial 18 (28.6) 7 (14.9) 5 (17.2) 4 (11.8) 23 (25.0) 11 (13.6) 25 (22.7) 9 (14.3)
Deep 10 (15.9) 7 (14.9) 6 (20.7) 6 (17.6) 16 (17.4) 13 (16.0) 17 (15.5) 12 (19.0)
Mixed 34 (54.0) 31 (66.0) 17 (58.6) 23 (67.6) 51 (55.4) 54 (66.7) 65 (59.1) 40 (63.5)
IH-MAG 1 (1.6) 2 (4.3) 1 (3.4) 1 (2.9) 2 (2.2) 3 (3.7) 3 (2.7) 2 (3.2)
Location, .912* .840* .494*
no (%)
Head and 30 (47.6) 25 (53.2) 17 (58.6) 20 (58.8) 47 (51.1) 45 (55.6) 55 (50.0) 37 (58.7)
neck
Trunk 18 (28.6) 13 (27.7) 8 (27.6) 8 (23.5) 26 (28.3) 21 (25.9) 31 (28.2) 16 (25.4)
Limb 15 (23.8) 9 (19.1) 4 (13.8) 6 (17.6) 19 (20.7) 15 (18.5) 24 (21.8) 10 (15.9)
Efficacy at .479y .359k .238k

J AM ACAD DERMATOL
treatment
termination,
no (%)
Grade I 39 (61.9) 35 (74.5) 20 (69.0) 20 (58.8) 59 (64.1) 55 (67.9) 74 (67.3) 40 (63.5)

APRIL 2024
Grade II 13 (20.6) 6 (12.8) 8 (27.6) 13 (38.2) 21 (22.8) 19 (23.5) 19 (17.3) 21 (33.3)
Grade III 6 (9.5) 4 (8.5) 1 (3.4) 1 (2.9) 7 (7.6) 5 (6.2) 10 (9.1) 2 (3.2)
Document downloaded from http://www.elsevier.es, day 28/03/2024. This copy is for personal use. Any transmission of this document by any media or format is strictly prohibited.

J AM ACAD DERMATOL Wang et al 787

VOLUME 90, NUMBER 4

adjustment. No AEs were observed after medication

.120z

.679z
.525*
withdrawal.

3.50 (0.97-8.38) 4.36 (2.26-8.96)

0.71 (0.43-1.08) 0.75 (0.51-1.07)

Factors affecting relapse and the optimal
withdrawal mode for predicting relapse
59 (93.7)
0 (0.0)

4 (6.3)

None of the general characteristics of the patients

and their lesions influenced the relapse
(Supplementary Table III, available via Mendeley at
https://data.mendeley.com/datasets/gwnx93d3vy/
1). No significant differences in treatment duration
100 (90.9)
7 (6.4)

10 (9.1)

(P = .364, Supplementary Fig 4, A, available via

Mendeley at https://data.mendeley.com/datasets/
gwnx93d3vy/1) or age at treatment termination
(P = .494, Supplementary Fig 4, B, available via
.439z

.610z
.757*

Mendeley at https://data.mendeley.com/datasets/
gwnx93d3vy/1) were found between patients with
3.04 (1.27-9.05) 4.41 (1.69-7.94)

0.74 (0.42-1.08) 0.74 (0.51-1.10)

and without relapse. Eight withdrawal criteria

75 (92.6)

affected the relapse rate (Table III). Among them,

2 (2.5)

6 (7.4)

the AUC of the classification grouped by ‘‘whether

withdrawal was with 3 months of maintenance and
with age older than 13 months’’ was the largest
(AUC = 0.603, Table III).
84 (91.3)
5 (5.4)

8 (8.7)

Factors affecting major relapse and the

optimal withdrawal plan for predicting major
relapse
.142y

.604y
.979x

No general patient and lesion characteristics

influenced major relapses (Supplementary Table
IV, available via Mendeley at https://data.
2.46 (0.85-7.80) 4.79 (1.60-9.28) 5.32 (2.14-12.00) 4.28 (2.28-7.12)

0.68 (0.40-1.08) 0.79 (0.51-1.13) 0.80 (0.52-1.09) 0.72 (0.51-1.05)

mendeley.com/datasets/gwnx93d3vy/1). No signifi-
32 (94.1)
0 (0.0)

2 (5.9)

cant differences in treatment duration (P = .931,

Supplementary Fig 4, C, available via Mendeley at
https://data.mendeley.com/datasets/gwnx93d3vy/
1) or age at treatment termination (P = .866,
Supplementary Fig 4, D, available via Mendeley at
27 (93.1)

https://data.mendeley.com/datasets/gwnx93d3vy/
0 (0.0)

2 (6.9)

1) were found between patients with and without

major relapse. Only 1 withdrawal criterion, that is,
whether withdrawal was with 3 months of mainte-
nance, affected major relapse, and the AUC of this
classification was 0.610.
43 (91.5)
2 (4.3)

4 (8.5)

DISCUSSION
Relapse after withdrawal from propranolol is an
unfortunate complication in the treatment of IHs,
which often affects the parents’ confidence in the
57 (90.5)
5 (7.9)

6 (9.5)

cure and the compliance with subsequent treatment.

Most physicians believe that the longer the duration
Mantel-Haenszel x2 test.

of medication, the less likely patients are to

Mann-Whitney U test.
Kruskal Wallis H test.

relapse.1,10 Conversely, others believe that relapse

(IQRs), cm2

Fisher’s exact test.

(IQRs), cm

is independent of the treatment duration.11,12

*Pearson x 2 test.
Grade IV
Ulceration,

Median

Median
Thickness

Currently, studies with high levels of evidence

no (%)

regarding this controversy are lacking.

Yes
No
Area

Furthermore, although most researchers suggest

that later withdrawal can reduce the risk of relapse,
y
z

k
x
Document downloaded from http://www.elsevier.es, day 28/03/2024. This copy is for personal use. Any transmission of this document by any media or format is strictly prohibited.

788 Wang et al J AM ACAD DERMATOL

APRIL 2024

Table II. Outcomes of different groups

With 1 mo of With 3 mo of Direct Gradual
Characteristics Group A Group B Group C Group D P-values maintenance maintenance P-values withdrawal withdrawal P-values
Lesions N = 63 N = 47 N = 29 N = 34 N = 92 N = 81 N = 110 N = 63
Relapse, .149* .055* .396*
no (%)
Yes 26 (41.3) 16 (34.0) 13 (44.8) 7 (20.6) 39 (42.4) 23 (28.4) 42 (38.2) 20 (31.7)
No 37 (58.7) 31 (66.0) 16 (55.2) 27 (79.4) 53 (57.6) 58 (71.6) 68 (61.8) 43 (68.3)
Major .156y .041* .963*
relapse,
no (%)
Yes 11 (17.5) 5 (10.6) 7 (24.1) 2 (5.9) 18 (19.6) 7 (8.6) 16 (14.5) 9 (14.3)
No 52 (82.5) 42 (89.4) 22 (75.9) 32 (94.1) 74 (80.4) 74 (91.4) 94 (85.5) 54 (85.7)
Patients N = 53 N = 39 N = 28 N = 30 N = 81 N = 69 N = 92 N = 58
Adverse .280* .154* .458*
events,
no (%)
Yes 12 (22.6) 7 (17.9) 10 (35.7) 5 (16.7) 22 (27.2) 12 (17.4) 19 (20.7) 15 (25.9)
No 41 (77.4) 32 (82.1) 18 (64.3) 25 (83.3) 59 (72.8) 57 (82.6) 73 (79.3) 43 (74.1)

*Pearson x 2 test.
y
Fisher’s exact test.

Table III. The effects of different withdrawal criterions on relapse

Withdrawal regimens P-values* AUC
Withdrawal with 3 mo maintenance and with the medication duration longer than 7 mo 0.049 0.577
Withdrawal with 3 mo maintenance and with the medication duration longer than 9 mo 0.022 0.583
Withdrawal with 3 mo maintenance and with the medication duration longer than 10 mo 0.045 0.567
Withdrawal with 3 mo maintenance and with the medication duration longer than 11 mo 0.005 0.585
Withdrawal with 3 mo maintenance and with the medication duration longer than 12 mo 0.048 0.553
Withdrawal with 3 mo maintenance and with the age older than 13 mo 0.004 0.603
Withdrawal with 3 mo maintenance and with the age older than 14 mo 0.042 0.566
Withdrawal with 3 mo maintenance and with the age older than 15 mo 0.015 0.560

AUC, Areas under the receiver operating characteristic curve.

*Pearson x 2 test.

it is not possible to prolong the treatment indefinitely withdrawal with 3 months of maintenance was
in clinical treatment. Therefore, it is necessary to preferable.
choose an appropriate withdrawal time to minimize Some researchers have pointed to the benefits of
the occurrence of relapse and simultaneously avoid gradual withdrawal, suggesting that it may reduce
an ineffective extension of the medication course. the relapse rate and the risk of withdrawal syndrome,
Based on our clinical experience, we believed that such as heart rate or blood pressure rebound.13,14
continued medication for 3 months after the lesion Our study also compared the relapse rates, major
reaches its maximum degree of regression was relapse rates and AE rates between direct and
helpful in reducing the risk of relapse. Therefore, gradual withdrawal approaches. We determined
this study aimed to compare the relapse and major that direct withdrawal did not increase the risk of
relapse rates of propranolol discontinuation relapse or major relapse, nor did it increase the rate
following 1 and 3 months of maintenance treatment. of AEs, indicating that direct withdrawal of propran-
Simultaneously, the AE rates of these 2 modalities olol was not inferior.
were compared to clarify the safety of the additional By stratifying patients according to different with-
2 months of treatment. The results showed that drawal criteria, we found that patients with or without
withdrawal after 3 months of maintenance tended major relapse could be effectively distinguished by
to reduce the relapse rate (P = .055) and could whether or not they had terminated treatment ‘‘with
effectively reduce the major relapse rate (P = .041), 3 months of maintenance,’’ while those with or
but did not increase the rate of AEs, indicating that without relapse could be distinguished to the greatest
Document downloaded from http://www.elsevier.es, day 28/03/2024. This copy is for personal use. Any transmission of this document by any media or format is strictly prohibited.

J AM ACAD DERMATOL Wang et al 789

VOLUME 90, NUMBER 4

extent by whether propranolol withdrawal was ‘‘with CONCLUSIONS

3 months of maintenance and with an age at treatment We believe that propranolol withdrawal after
termination older than 13 months.’’ Overall, we 3 months of maintenance is most beneficial for
believe that the optimal time to discontinue propran- reducing the major relapse rate, while withdrawal
olol was when the patient was older than 13 months with age older than 13 months and 3 months of
and had maintained maximum IHs regression for maintenance is most helpful for reducing the relapse
3 months. Although prolonging the course of the rate.
propranolol treatment on this basis may further
reduce the risk of relapse, it would unnecessarily
Conflicts of interest
increase the duration of the treatment for patients
None disclosed.
who would otherwise not relapse.
The overall AE rate in this study was 22.67%; REFERENCES
however, only 1 patient withdrew from the study due 1. Leaute-Labreze C, Harper JI, Hoeger PH. Infantile haeman-
to AE. We believe this was due to the exhaustive gioma. Lancet. 2017;390(10089):85-94.
informed consent of parents about potential AEs and 2. Rodriguez Bandera AI, Sebaratnam DF, Wargon O, Wong LF.
Infantile hemangioma. Part 1: epidemiology, pathogenesis,
relatively mild severity of AEs. In addition, a pediatric
clinical presentation and assessment. J Am Acad Dermatol.
consultation was requested when AE occurred to 2021;85(6):1379-1392.
evaluate the suitability of continued medication. 3. Sebaratnam DF, Rodriguez Bandera AL, Wong LF, Wargon O.
Previous studies have shown that head and neck Infantile hemangioma. Part 2: management. J Am Acad
IHs, segmental IHs, or IHs with deep components Dermatol. 2021;85(6):1395-1404.
are more likely to relapse.15,16 Among the 20 patients 4. Solman L, Glover M, Beattie PE, et al. Oral propranolol in the
treatment of proliferating infantile haemangiomas: british
with multiple lesions in our cohort, 8 patients had Society for Paediatric Dermatology consensus guidelines. Br
different treatment efficacy and 6 patients experi- J Dermatol. 2018;179(3):582-589.
enced different relapse conditions across different 5. Xu W, Zhao H. Management of infantile hemangiomas: recent
lesions, suggesting that propranolol efficacy and advances. Front Oncol. 2022;12:121064048.
relapse may be related to the location, size, thick- 6. Frongia G, Byeon JO, Mehrabi A, Gunther P. Recurrence rate of
infantile hemangioma after oral propranolol therapy. Eur J
ness, morphologic pattern, and type of IHs, or to Pediatr. 2021;180(2):585-590.
other underlying pathological mechanisms. 7. Pam N, Kridin K, Khamaysi Z. Propranolol for infantile heman-
However, due to the limited sample size in this gioma: evaluating efficacy and predictors of response and
study, we could not conduct a stratified analysis of rebound growth. Dermatol Ther. 2021;34(3):e14936.
8. Holmes WJ, Mishra A, Gorst C, Liew SH. Propranolol as first-line
these multifocal lesions. Besides, our results did not
treatment for rapidly proliferating infantile haemangiomas. J
show a higher rate of relapse in head and neck IHs, Plast Reconstr Aesthet Surg. 2011;64(4):445-451.
segmental IHs, and non-superficial IHs. This may be 9. Achauer BM, Chang CJ, Vander Kam VM. Management of
largely because the sample size was not sufficiently hemangioma of infancy: review of 245 patients. Plast Reconstr
large to identify such differences. However, we Surg. 1997;99(5):1301-1308.
suspect that this might also be due to a selection 10. Ginguerra MA, Saito O, Fernandes J, Castro DS, Matayoshi S.
Clinical and radiological evaluation of periocular infantile
bias caused by previous retrospective studies. hemangioma treated with oral propranolol: a case series. Am
Parents tend to be more likely to pay closer attention J Ophthalmol. 2018;185:48-55.
to head and neck IHs and large segmental IHs; 11. Ahogo CK, Ezzedine K, Prey S, et al. Factors associated with the
therefore, they usually seek referral when IHs relapse of infantile haemangiomas in children treated with
oral propranolol. Br J Dermatol. 2013;169(6):1252-1256.
relapse. However, other IHs may not be taken as
12. Sagi L, Zvulunov A, Lapidoth M, Ben Amitai D. Efficacy and
seriously by parents even if there is a relapse. safety of propranolol for the treatment of infantile hemangi-
Study limitations include its single-center nature oma: a presentation of ninety-nine cases. Dermatology. 2014;
and nonrandomization of patient allocation. The 228(2):136-144.
single-center design in a tertiary reference center 13. Shah SD, Baselga E, McCuaig C, et al. Rebound growth of
may limit the generalization of the study findings infantile hemangiomas after propranolol therapy. Pediatrics.
2016;137(4):e20151754.
and may cause referral bias. Although a randomized 14. Szychta P, Stewart K, Anderson W. Treatment of infantile
controlled trial was originally designed, some patients hemangiomas with propranolol: clinical guidelines. Plast
were allocated to other groups according to their Reconstr Surg. 2014;133(4):852-862.
parents’ intentions due to the decreased compliance of 15. O’Brien KF, Shah SD, Pope E, et al. Late growth of infantile
parents affected by the COVID-19 pandemic; thus, the hemangiomas in children [3 years of age: a retrospective
study. J Am Acad Dermatol. 2019;80(2):493-499.
study was ultimately analyzed as a non-randomized 16. Schmid F, Hoeger PH. Propranolol response in patients with
controlled trail. The validity of our findings needs to be segmental versus focal facial hemangiomas: a retrospective
tested in a larger sample size study. case-control study. J Am Acad Dermatol. 2022;87(2):490-491.