Rebound Growth of Infantile Hemangiomas after Propranolol versus Atenolol Treatment: A Retrospective Study

Infantile hemangioma (IH) is the most common soft tissue tumor in children, occurring in 4–10% of infants. Its growth is classified to early and late proliferative phases and an involution phase [1‒4].

Approximately 10–15% of IHs require treatment due to risk for complications [5, 6]. Propranolol, a nonselective β-adrenergic blocker, is the first-line treatment. In recent years, atenolol, a selective β1-adrenergic blocker, was shown to be as effective as propranolol with lower rates of adverse effects [7‒12].

Rebound growth following discontinuation of propranolol is a known phenomenon, with reported incidence of 6–25% [7, 13‒17]. Several risk factors have been described, including involvement of deep dermal tissues, segmental pattern, and treatment discontinuation before the age of 9 months [5, 13‒15, 17‒19].

As atenolol is considered an effective alternative treatment, assessing the incidence of rebound growth is important. We aimed to compare the incidence of rebound growth following discontinuation of atenolol and propranolol and to identify risk factors.

We reviewed the records of all patients diagnosed with IHs and treated with oral propranolol or atenolol during 2009–2019 in a pediatric dermatology unit in a single tertiary medical center. Propranolol was the mainstay of treatment until 2015, at a dose of 2–3 mg/kg/day. Since April 2015, most patients were treated with atenolol at a dose of 1–2 mg/kg/day. This change was related to concerns regarding the safety profile of propranolol [7, 11, 20, 21]. Aside from the change in medication, there were no other changes in treatment protocols.

The study inclusion criteria were diagnosis of IH, completion of at least 3 months of oral β-blocker, and completion of at least 3 months of follow-up after treatment discontinuation. Exclusion criteria were prior corticosteroid treatment and a diagnosis of syndrome-related hemangioma. The study protocol was approved by the Institutional Review Board.

A Pearson’s chi-square test or Fisher’s exact test were used for analysis of differences in discrete variables, and analysis of variance was used for comparing continuous variables. A p value of ≤0.05 was considered significant.

A multivariate logistic regression model was used to identify factors that may be associated with rebound growth. Adjusted odds ratios (aORs), confidence intervals (CIs) (95%), and p values were calculated.

A flowchart of the study population is presented in Figure 1. The clinical and demographic characteristics are presented in Table 1. No significant difference was observed in rebound growth incidence between patients treated with propranolol and atenolol (59 [22.1%] and 40 [22.5%], OR: 1.02, 95% CI: 0.64–1.60). The demographic and clinical features were similar, except for IH type. Among patients treated with atenolol compared to propranolol, a higher proportion had superficial or deep IHs. No significant difference was recorded in rebound growth incidence between patients with superficial and deep IHs. This was observed in the general cohort and within the specific treatment groups.

Table 2 presents characteristics of the treatment among the 99 patients who experienced rebound growth. The duration of treatment until growth arrest was shorter for patients treated with atenolol compared to propranolol: 9.41 ± 5.61 versus 14.79 ± 10.02 months (95% CI, p < 0.001).

The results of the multivariate analysis are presented in Table 3. The medication choice (propranolol vs. atenolol) was not associated with the risk of rebound growth nor were any of the other clinical characteristics. However, the duration of initial treatment was found to be associated with the risk of rebound growth (aOR: 1.24, 95% CI: 1.10–1.38).

The main finding of our study is similar incidence of rebound growth following treatment with propranolol and atenolol. The overall incidence was 22.2%, in line with previous studies [13, 14] and higher than other reports [7, 16]. Different inclusion and exclusion criteria, such as age at enrollment and type of IH, may explain the disparate findings. Notably, in our cohort, the aOR for regrowth was lower among patients who initiated treatment with atenolol before the age of 5 months, compared to patients who initiated treatment with atenolol and propranolol after the age of 5 months. Although this difference was not statistically significant, it supports the notion that initiating treatment after a certain age may result in a longer proliferation phase. Propranolol could inhibit hemangioma-derived endothelial cells proliferation and induce apoptosis, possibly shortening the proliferation phase [22]. Propranolol might also alternate the natural history of the disease by interfering with the natural mechanisms of clearing the hemangioma stem cells [23]. This suggests that propranolol might have a different effect on hemangioma cells at different stages of its course. These findings require further investigation as they could have important implications for management.

The rebound growth incidence was similar after discontinuing atenolol and propranolol, as reported previously [7, 11]. The different proportions between the treatment groups of patients with deep dermal involvement could conceivably cause a bias. However, such bias seems unlikely as IH type was not found to be a risk factor in a multivariate analysis. Nevertheless, other investigators did identify deep dermal involvement as an independent risk factor for rebound growth [13, 14, 17]. Deep hemangiomas tend to grow later and longer than superficial hemangiomas [4]. Therefore, it is possible that rebound growth also appears later in deep hemangiomas, and patients in this cohort were missed due to insufficient follow-up. However, most patients were followed for years after treatment discontinuation to identify those who need further interventions to treat the sequelae. We cannot entirely exclude the possibility that even though 12% of our cohort had deep IHs, our sample size was not sufficient to identify this condition as an independent risk factor.

In a univariate analysis, the rebound growth period was shorter for patients treated with atenolol than for patients treated with propranolol. This raises the question as to whether atenolol, as a selective β-blocker, may have a more specific effect than propranolol in patients with rebound growth. Since this observation has not been previously reported and given the limitations of this study’s design, larger and prospective randomized studies are required to confirm this finding.

Surprisingly, we found that longer duration of initial treatment, with either atenolol or propranolol, was correlated with higher occurrence of rebound growth. Szychta et al. [24] recorded a similar result. As the mechanism of action of oral β-blockers in treatment for IH is not clear, over-treatment remains a possibility. However, it is more likely that this association is evident because treatments were sustained rather than indicating that sustained treatment causes more rebound. Further evidence from randomized controlled trials is necessary to validate this hypothesis.

The limitations of our study include its retrospective nature and selection bias due to the number of patients who were lost to follow-up. Notably, the loss of follow-up is likely due to the successful treatment outcome, in patients without rebound growth. Additionally, this study did not include randomized allocation, making it more susceptible to selection bias. However, all patients who met the inclusion criteria before and after 2015 were included and were allocated to the propranolol and atenolol groups accordingly. This likely mitigated potential allocation bias.

In conclusion, our study demonstrated similar IH rebound growth incidence rates following the discontinuation of two oral β-blockers, supporting the non-inferiority of atenolol. In addition, patients who were treated with atenolol required shorter treatment durations for rebound growth episodes, compared to patients treated with propranolol. These findings should be further investigated in randomized controlled trials as they may impact clinical decisions.

The incidence of IH regrowth is similar following atenolol and propranolol treatment.

This retrospective chart review study involving human participants was in accordance with the ethical standards of the institutional and national research committee and with the 1964 Helsinki Declaration and its later amendments and ethical standards. The Human Investigation Committee (IRB) of Schneider Children’s Medical Center of Israel approved this study and deemed it exempt from informed consent on September 10, 2019 (Approval No. 0502-19RMC).

The authors have no conflicts of interest to disclose.

No funding was received for this study.

Shoham Baruch: conceptualization, methodology, investigation, statistical analysis, data curation, and writing – original draft. Dan Ben-Amitai: conceptualization, methodology, and writing – original draft. Rivka Friedland: investigation, data curation, and writing – review and editing.

The data underlying this article cannot be shared publicly due to the privacy of individuals that participated in the study. Data will be provided upon reasonable request. Further inquiries can be directed to the corresponding author.