Sirolimus Early Treatment in Vascular Anomalies Leads to a Better Response

Abstract

Objectives: 

Sirolimus has become a relevant drug in the treatment of vascular anomalies, initially relegated to rescue therapy for refractory lesions but more recently has become first line. Our objective is to determine when it is best to start sirolimus treatment.

Methods: 

A retrospective review of patients with vascular anomalies treated with sirolimus during a 9-year period was performed. The variables analyzed included subtype of vascular anomaly, age at the start of treatment, sirolimus dosage and levels, response, and duration of treatment among others.

Results: 

One hundred twenty-two patients were included and classified according to ISSVA in: tumor (7), lymphatic malformation (60), venous malformation (16), combined malformations (18), associated overgrowth (8), and others (15). Overall response was positive in 90.8%. Median duration of treatment was 32 months (0–116). Response was not influenced by dosage or levels, neither by the subtype of vascular anomaly, location, extension, symptoms, genetic testing nor previous treatments. However, the lower the age of starting sirolimus, the better the response, mainly under 5 years of age (P = 0.004). At 6 months, 67% of patients had responded while at 12 months >84% did. By age, patients <5 years showed positive response at a median time of 2 months, compared to >5 months from older patients. Time until a positive response was not influenced by subtype or severity.

Conclusion: 

Overall response to sirolimus was good and most patients responded irrespective of their severity in location, extension or symptoms. Patients <5 years old respond better and faster making our goal to attempt for an early treatment.

Introduction

Historically, vascular anomalies were only treated when symptomatic or complicated and treatment was mainly surgical/interventional because medical therapies were limited and ineffective. For the past decade, clinical and experimental research together with technological advances have enhanced our knowledge regarding the physiopathology, the histopathology, and the genetics of vascular anomalies, expanding at the same time our arsenal of therapeutic options.

From a predominantly surgical approach, we have shifted to a less aggressive focus in the management of these patients. Given the limitations of surgical/interventional approaches to achieve complete cure of vascular anomalies, physicians are choosing a more conservative path to reduce the symptoms and avoid progression, decreasing morbidity and/or mortality and long-term sequelae as well as improving the quality of life of these patients. Recent pharmacological discoveries and repurposing of drugs from cancer treatments have allowed the development of new targeted therapies in vascular anomalies.

Sirolimus, previously used as immunosuppressive and antiproliferative agent, was first reported in a patient with a kaposiform hemangioendothelioma in 2010.¹ Since the first publication of a series of patients with good response to sirolimus in 2011, it has been used “off-label” for numerous vascular anomalies.²

Over the last 15 years, sirolimus has become the third pillar in the treatment of vascular anomalies, along with surgical and interventional treatment. Initially sirolimus was relegated as rescue therapy for refractory lesions, but lately, in view of the promising response and the limited adverse effects in the short term, many authors recommend its use as first line before more aggressive treatments.

Despite growing experience in the field of vascular anomalies, many questions remain unanswered because most of the publications in vascular anomalies are case reports or case series with a paucity of more stringent evidence such as prospective studies or clinical trials due to the rarity and complexity of these diagnoses.³ Additionally, the increase in compassionate use of newer agents has also raised new interrogatives.⁴

When it is best to start sirolimus: at diagnosis to avoid future symptoms and complications, as first-line treatment in symptomatic patients, or as a rescue option in refractory patients?

How long should we maintain sirolimus: for at least 6 months to confirm response, for a year and then withdraw, for a year and then aim for the minimum effective dose, or indefinitely?

The main objective of this review is to respond to these rising questions.

Materials and methods

A retrospective review of patients with vascular anomalies treated with sirolimus from 2011 until 2020 was performed.

Inclusion criteria: all patients with vascular anomalies treated with sirolimus and followed in the vascular anomalies unit from Plastic Pediatric Surgery in our center. Sirolimus treatment was recommended according to our center’s protocol developed by our vascular anomalies unit. All patients, parents, and/or legal guardians accepted the treatment as compassionate use after being informed of the therapeutic options as well as the benefits and risks of sirolimus. Indications for treatment included refractoriness to standard therapies with persistent symptoms, risk of severe sequelae with surgical or interventional approach, bone involvement or visceral involvement. Initial dosage was 0.8 mg/m²/12 h, although lower doses were chosen for neonates or children younger than 1 year old and single daily dose was allowed for children with poor therapeutic compliance. Target levels between 5 and 10 ng/mL were selected for less symptomatic patients while levels between 10 and 15 ng/mL were the goal for larger and more complicated vascular anomalies. Treatment was maintained for at least 1 year, except for severe adverse effects or patients or family choice, and after that dose was progressively reduced until withdrawal or a minimum effective dose when symptoms recurred. Prophylaxis of Pneumocystis pneumonia (PCP) was not protocolized.

Exclusion criteria: patients excluded from the study included those that did not received >1 dose of sirolimus or did not follow the protocol described above. The variables analyzed for the review included the subtype of vascular anomaly, the age at the start of treatment, sirolimus dosage and levels, the response, the time until a positive response, the duration of treatment, and the side effects among others.

The subtype of vascular anomaly was assigned according to the ISSVA classification from 2018,⁵ dividing patients in tumors, malformations, and others.

•  Tumors: kaposiform hemangioendothelioma (KHE) or tufted angioma, composed hemangioendothelioma and congenital hemangioma.
•  Malformations: capillary malformation, common lymphatic malformation, complex lymphatic anomalies (including generalized lymphatic anomaly [GLA], Gorham-Stout disease [GSD], kaposiform lymphangiomatosis [KLA], and central conducting lymphatic anomaly), lymphedema, common venous malformation, diffuse or syndromic venous malformation, arteriovenous malformation, combined malformation, malformations with overgrowth (included here PROS).⁶
•  Others: PTEN hamartoma tumor syndrome (PHTS) and others provisionally unclassified vascular anomalies.

The variable age at the start of treatment was divided into 4 age groups: younger than 1 year of age, between 1 and 5 years, between 5 and 15 years, and older than 15 years of age. These 4 groups were comparable in sex, location, extension, and symptoms of the vascular anomalies.

Since the most appropriate measure of treatment response has not been established in a standardized way, 4 response categories were considered based on the classification determined by Adams et al⁷ in 2016: complete response, partial response, stability and progression. (Table 1) Performance score was assessed through Karnofsky (>16 years old) and Lansky (<16 years old) performance status scale.

Table 1. - Demographics and Type of Vascular Anomalies

Patients		N (%)
Sex	Male	52 (45%)
	Female	70 (55%)
Age	<1 y old	21 (17,1%)
	1–5 y old	22 (17,8%)
	5–15 y old	35 (28,7%)
	>15 y old	44 (36,4%)
Vascular tumors	Kaposiform hemangioendothelioma	5 (4.1%)
	Composite hemangioendothelioma	1 (0.8%)
	Congenital hemangioma	1 (0.8%)
Vascular malformations	Capillary malformations	3 (2.5%)
	Lymphatic malformations	60 (49.1%)
	Common	28 (23%)
	Complex	30 (24.5%)
	GLA	17 (14%)
	GSD	9 (7.4%)
	KLA	1 (0.8%)
	CCLA	1 (0.8%)
	Lymphedema	2 (1.6%)
	Venous malformations	16 (13.1%)
	Common	8 (6.5%)
	Syndromic	8 (6.5%)
	Arteriovenous malformations	1 (0.8%)
	Combined malformations	18 (14.7%)
	Associated overgrowth (PROS)	8 (6.5%)
Unclassified (PUVA)	PHTS	6 (4.9%)
	Unclassified	3 (2.5%)

Abbreviations: CCLA, central conducting lymphatic anomaly; GLA, generalized lymphatic anomaly; GSD, Gorham-Stout disease; KA, kaposiform lymphangiomatosis; PHTS, PTEN hamartoma tumor syndrome; PROS, PIK3CA-related overgrowth spectrum; PUVA, provisionally unclassified vascular anomalies.

-

Type of Response	Definition
Complete response	No evidence of the lesion in imaging tests, no evidence of organic dysfunction secondary to the lesion, no signs or symptoms associated to the lesion and normalization of the quality of life
Partial response	Decrease of at least 20% of the size of the lesion either in physical exam or imaging tests, improvement of the organic dysfunction secondary to the lesion at least 1 grade, improvement of signs and symptoms associated to the lesion allowing to carry out daily life activities and improvement of the quality of life at least 1 grade
Progression	Increase of at least 20% of the size of the lesion either in physical exam or imaging tests, worsening of the organic dysfunction secondary to the lesion at least 1 grade, worsening of signs and symptoms associated to the lesion and worsening of the quality of life at least 1 grade
Stability	None of the previous categories

For the statistical analysis, these categories were regrouped in 2:

•  Positive response: complete response, partial response, and stability in those vascular anomalies whose natural evolution is to progression (thus stability or no progression equals good response such as arteriovenous malformations).
•  Negative response: progression and stability in those vascular anomalies whose natural evolution is not to progression (thus stability or no progression does not equal good response, such as lymphatic malformations).

To facilitate the analysis, 4 categories of severity were created regarding location, extension, symptoms and adjuvant treatment.

•  Five categories for location were considered, a patient being able to belong to one, some or all of them: head and neck, thorax, abdomen and pelvis, upper limbs, and lower limbs.
•  Four categories for extension were considered, a patient being able to belong to one, some or all of them: skin and subcutaneous tissue, muscle, viscera, and bone.
•  Thirteen categories for symptoms or associated complications were created, a patient being able to belong to one, some or all of them: no symptoms, pain (evaluated by visual analogue scale), infection or inflammation, coagulopathy, bleeding, thrombosis (intralesional, deep vein thrombosis or thromboembolism), effusion (pleural, pericardial, or ascites), lymphorrhea, deformity, leg length discrepancy (LLD) >1 cm, respiratory impairment (need of accessory measures to breathe), digestive impairment (need of accessory measures to be fed), protein-losing enteropathy and others (such as pathological fractures, headaches, dysphonia, or hearing loss).
•  Previous and adjuvant treatments were collected and grouped in 3 categories: pharmacological, interventional, and surgical.

Genetic testing was performed in most patients at the Institute of Medical and Molecular Genetics (INGEMM) and the Vascular Anomalies Center at La Paz Hospital, Madrid (Spain). Deep high-throughput sequencing studies were performed with an average expected reading depth of >500X using a hybridization-based custom panel of 167 vascular-related genes. An in-house bioinformatic pipeline, designed to detect low mosaic variants, and subsequent validation of variants by ddPCR, was performed as described before by Rodriguez-Laguna et al.⁸

Data about sirolimus treatment were collected including blood levels (maximum and minimum for each patient), duration of treatment and side effects. Side effects considered for this study were grouped in three categories (mild, moderate, and severe) based on CTCAE (Common Terminology Criteria for Adverse Events)⁹ and depending on the need of reduce or suspend the treatment, the necessity of hospitalization or adjuvant treatment and the posterior sequelae.

Data were analyzed with SAS program version 9.4 (SAS Institute Inc. 2013. Base SAS 9.4 SAS/STAT; Statistical analysis. Cary, NC) and statistically significant differences were considered when presenting an error probability <5% (P < 0.05). For the description of continuous quantitative variables, the mean was used together with the standard deviation. The qualitative variables were described using absolute frequencies and relative frequencies expressed as a percentage.

Comparisons between continuous quantitative variables between independent groups were assessed mainly using non-parametric tests, specifically the Kruskal–Wallis test or the Mann–Whitney U test. In some cases, the post hoc test was used to correct the level of significance in multiple comparisons. The frequency analysis between qualitative variables was performed using the chi-square test or Fisher exact test when necessary. The time until a positive response was studied using the Kaplan and Meier technique with the Log Rank test for qualitative variables, and univariate Cox regression for quantitative variables.

Results

One hundred twenty-two patients were included in the study, being 52 male (45%) and 70 female (55%). Patients were divided in groups by the subtype of vascular anomaly according to the latest ISSVA classification in 2018.⁵ Seven patients presented with vascular tumors (5.7%), the most frequent being kaposiform hemangioendothelioma (n = 5, KHE). Sixty patients presented with lymphatic anomalies (49.1% LAs): 28 common lymphatic malformations (LMs), 30 complex LAs (17 GLA, 9, GSD, 1 KLA, 1 central conducting lymphatic anomaly), and 2 lymphedemas. All common LMs were either microcystic or mixed, there were no patient with pure macrocystic malformation. Sixteen patients presented with venous malformations (13.1%), 8 common and 8 diffuse or syndromic (including 2 patients with blue rubber bleb nevus [BRBN] syndrome and 1 diffuse glomovenous malformation). Eighteen patients had combined malformations (14.7%), all capillary-lymphatic, capillary-venous, and capillary-lymphatic-venous malformations. Eight patients presented with associated overgrowth (6.5%) and were considered part of the PIK3CA-related overgrowth spectrum (PROS), including 2 patients with CLOVES (Congenital Lipomatous Overgrowth, Vascular Malformations, Epidermal Nevi, Scoliosis/Skeletal and Spinal syndrome) and 4 patients with CLAPO (Capillary malformation of the lower lip, LM of the face and neck, Asymmetry and Partial or generalized Overgrowth). Nine patients were considered unclassified (7.3% provisionally unclassified vascular anomalies), including 6 patients with PHTS (4.9%) and 3 patients with unclassified vascular life-threatening lesions. Data on demographics and type of vascular anomalies are described on Table 1.

The area most frequently involved was head and neck (46.5%) followed by lower limbs (33.3%), abdomen and pelvis (31.8%), thorax (28.7%), and upper limbs (20.9%). While most patients presented with involvement of skin and subcutaneous tissue (86%) and muscle (57.4%), fewer had visceral (28.7%) and/or bone involvement (22.5%). Almost every patient presented with symptoms due to the vascular anomaly (99.2%), and only one patient with Sturge-Weber syndrome was asymptomatic, but presented high risk for seizures and was started on prophylactic sirolimus with good response.¹⁰ Pain, deformity and inflammation/infection were the most common symptoms (69%, 53.5%, and 51.2%, respectively). Some patients also presented with bleeding (14.7%), respiratory impairment (11.6%), digestive impairment (11.6%), lymphorrhea (10.1%), effusions (9.3%), coagulopathy (7.8%), LLD (7%), and thrombosis (6.2%).

Genetic testing was performed when there was available affected tissue (41.8%), with a gene mutation found in 84.3% of the patients tested. PIK3CA variant was identified in 72.1%, specifically in 44.4% of LMs and 48.1% in combined malformations and malformations with overgrowth. Two patients presented another variant besides PIK3CA: one TSC2 and one MAP2K2. Other variants found were PTEN (9.3%), GNAQ (7%), RASA1 (4.6%), and TEK (2.3%).

Most patients had received some types of previous treatment (82.2%), either pharmacological, interventional, or surgical. Pharmacological treatments were only antiplatelet and anticoagulation drugs to improve the symptoms of intralesional coagulopathy and prevent deep vein thrombosis. Only 17% of patients had sclerosis (mean number of sessions 2), but 57.3% of patients underwent one or more surgical resections (mean number of interventions 4). Adjuvant treatment at the same time of sirolimus was less frequent: 4.6% pharmacological, 6.2% interventional, and 10% surgical.

Data on location, extension, symptoms, genetics, and previous treatments are described in Table 2.

Table 2. - Location, Extension, Symptoms, Genetics, and Previous Treatments

Patients		N (%)
Location	Head and neck	58 (46.5%)
	Thorax	35 (28.7%)
	Abdomen and pelvis	39 (31.8%)
	Lower limbs	40 (33.3%)
	Upper limbs	26 (20.9%)
Extension	Skin and subcutaneous tissue	105 (86%)
	Muscle	70 (57.4%)
	Visceral	35 (28.7%)
	Bone	27 (22.5%)
Symptoms	Asymptomatic	1 (0.8%)
	Pain	84 (69%)
	Deformity	65 (53.5%)
	Inflammation/infection	62 (51.2%)
	Bleeding	18 (14.7%)
	Respiratory impairment	14 (11.6%)
	Digestive impairment	14 (11.6%)
	Lymphorrhea	12 (10.1%)
	Effusions	11 (9.3%)
	Coagulopathy	9 (7.8%)
	Leg length discrepancy	8 (7%)
	Thrombosis	7 (6.2%)
Genetics (performed in 41.8%)	No mutation found	19 (15.7%)
	PIK3CA	88 (72.1%)
	PTEN	11 (9.3%)
	GNAQ	8 (7%)
	RASA1	5 (4.6%)
	TEK	3 (2.3%)
Previous treatments	Interventional	21 (17%)
	Surgical	70 (57.3%)
Adjuvant treatments	Pharmacological	5 (4.6%)
	Interventional	7 (6.2%)
	Surgical	12 (10%)

Treatment with sirolimus was started at a median age of 9.7 ± 13.9 years old, with the youngest being a newborn on the first day of life. Patients were divided in four groups regarding the age at the start of treatment: younger than 1 year of age (17.1%), between 1 and 5 years (17.8%), between 5 and 15 years (28.7%), and older than 15 years of age (36.4%).

Response was categorized as complete in 3.1%, partial in 76.7%, stability in 11.6%, and progression in 1.5%. Complete response was achieved in one patient with KHE, one patient with composite hemangioendothelioma, one patient with multifocal lymphangioendotheliomatosis with thrombocytopenia and one patient with a cervical LM. Overall response was positive in 90.8% and negative in 9.2%. Median duration of the treatment was 32 months (0–116) and at the time of this study, 34.1% continue on treatment.

Withdrawal was decided in most cases due to partial response after a year of treatment (40%), in 10.6% due to side effects and only in 2.3% due to progression. From 85 patients that finished the treatment, only 2 (2.3%) required restarting sirolimus due to recurrence of symptoms after at least one month from withdrawal. The first patient was a 5-month-old girl with a cervicofacial LM that, after 4 months of sirolimus with a decrease in size of the malformation, presented with an episode of inflammation and after partial resection parents interrupted treatment. Sirolimus was restarted after 3 years due to progressive enlargement of sublingual malformation and recurrent lymphangitis with good response. The other patient was a 19-year-old woman with a GLA that, after 6 months of sirolimus with decrease of number of episodes of inflammation and pain, decided to abandon treatment due to side effects (vomiting). Sirolimus was reintroduced after 6 months due to enlargement of genitalia malformation and recurrent lymphangitis with good response.

Ten patients (8.1%) started afterward compassionate treatment with a specific PIK3CA inhibitor, alpelisib.

Half of the patients did not report any side effects. About 33.3% presented mild side effects, mainly mucositis 38.4% and dyslipidemia 21.5%; and 6.5% presented severe side effects, including 5 patients with pneumonia, 3 due to Pneumocystis jirovecii. Patients that suffered pneumonia due to Pneumocystis were a 1-year-old girl with a hepatic congenital hemangioma, a 1-year-old boy with BRBN, and a 14-year-old boy with KLA. All 3 did not have any other comorbidities, concomitant medication or known risk factors, presenting sirolimus levels between 5 and 8 ng/mL at the time of this episode. They received outpatient antibiotic treatment and continued on sirolimus without further complications. Two patients presented with pulmonary thromboembolism, considered a complication secondary to their underlying pathology due to persistence of embryonic veins and not related to sirolimus treatment.

Data on sirolimus treatment, response, and side effects are described in Table 3.

Table 3. - Sirolimus Treatment, Response, and Side Effects

Patients		N (%)
Response	Complete	4 (3.1%)
	Partial	94 (76.7%)
	Stability	14 (11.6%)
	Progression	2 (1.5%)
Overall response	Positive	111 (90.8%)
	Negative	11 (9.2%)
End of treatment	Continue	44 (34.1%)
	Withdraw	83 (64.4%)
	Recurrence	2 (1.5%)
Side effects	None	61 (50%)
	Mild	40 (33.3%)
	Moderate	2 (1.5%)
	Severe	8 (6.5%)

Initial analysis revealed no correlation between dosage and blood levels, neither globally nor studied in groups by age. Response was not significantly influenced by dosage or levels. Response was neither influenced by sex, type of vascular anomaly, location, extension, symptoms, genetic testing, or previous treatments. We only found a negative correlation between leg length difference discrepancy (LLD) and response (P = 0.036). There were not statistically significant differences between response and severity, neither in location, extension, symptoms nor adjuvant treatment.

The higher the dosage and blood levels, the more frequent the adverse effects were (P = 0.04, P = 0.2) although there was no correlation with the severity of side effects. There were no statistically significant differences between the presence of adverse effects and age, sex, type of vascular anomaly, location, extension, symptoms, genetic testing, or previous treatments.

Subsequent analysis showed the lower the age of starting sirolimus, the better the response, mainly in the group under 5 years of age (P = 0.014) (Figure 1).

Figure 1.: Response by age (boxplot).

Focusing on the time until a positive response, at 4 months of treatment 50% of patients had responded while at 6 months of treatment 67% had responded and at 12 months over 84% had responded. In addition, the lower the age of starting treatment, the faster the response: patients <5 years of age showed a positive response at a median time of 2 months of treatment, while patients >5 years of age showed a positive response at a median time of 6 months (P = 0.004) (Figure 2).

Figure 2.: Time until a positive response by age (boxplot).

Repeating this analysis in three groups by age, patients under 5 years of age showed positive response at a median time of 2 months, patients >15 years at 5 months and patients between 5 and 15 years at 7 months (Figure 3).

Figure 3.: Time until a positive response by age (Kaplan–Meier).

Time until a positive response was not influenced by severity neither by type of vascular anomaly, location, extension, symptoms, genetics or by the number of previous interventional (surgery + sclerotherapy) treatments. There were not statistically significant differences between time until a positive response and severity, neither in location, extension, symptoms, nor adjuvant treatment.

Discussion

Indications for sirolimus treatment and likelihood to respond is nowadays well-established, including LMs, venous malformations, and combined malformations. As we had a very high rate of positive response (90.8%), we could not find any statistically significant differences regarding type of vascular anomaly, location, extension, symptoms or severity. We only found LLD as a risk factor for no response. However, patients usually present with multiple associated symptoms so LLD should not be a reason to avoid giving sirolimus.

Currently, there is no consensus on when to start sirolimus treatment in patients with vascular anomalies and for how long to maintain it. In our series, the lower the age of starting sirolimus, the better and faster the response, mainly in patients <5 years of age. Furthermore, after 4 months of treatment, half the patients had responded and after 12 months >84% did; although we had to wait until >24 months to reach a positive response in the last patients.

The first case reports and case series published between 2011 and 2017 described sirolimus treatment mainly in refractory patients, with a variable time until response but usually a few months after achieving adequate blood levels. Time to start sirolimus and length of treatment were not defined and withdrawal was not standardized, they were determined individually depending on the symptoms and response.¹¹

An early treatment in these patients was not addressed until after 9 years from the first use of sirolimus in vascular anomalies. A multicenter study conducted in 2019 suggested that the early use of sirolimus should be considered in order to reduce morbidities, decrease the number of aggressive interventions and improve prognosis.¹² This had been previously proposed by Nadal et al in 2016,¹³ but there were no numbers to support this hypothesis and there has not been any other study looking further into it so far. Our data support an early treatment, best before 5 years of age, and opens the door to the use of sirolimus as a therapy even before symptoms, as it has been described in neonates with risk of upper airway obstruction due to LMs.¹⁴

On the other hand, the time until a clinical response has been extensively discussed in the literature. The initial papers on the topic described a wide range of time delay until a successful response with a median duration of 10 weeks (1–16).¹⁵ These numbers were lower in the first systematic review published in 2020 with 35.7% of patients showing a clinical response before 7 days of treatment and 75% before 3 weeks.¹⁶ Results were more similar to our own stated above in the series of kaposiform hemangioendothelioma by Ji et al where 75% reduction in tumor size was achieved in 46% of patients at 6 months of treatment and 71% of patients at 12 months, as well as in the prospective phase II study on slow-flow malformations where 78% of patients presented reduction in size at 3 months on top of functional improvement in 100%.^17,18 The most recent results on this issue are the preliminary data from the VASE trial, a prospective phase III study on refractory slow-flow vascular malformations, where 87% of patients showed clinical response at 6 months of follow-up.¹⁹

Despite the great advances achieved during these years, we are still lacking a protocol on how long to maintain treatment and what to do after obtaining a successful response. Our data suggest that sirolimus should be continued at least one year, maybe even 2, because some patients respond after the mark of 6 months (17%) and some others respond after 12 months (6%). If after 2 years of treatment, patient shows no response, patient can be declared non responder. If patient has responded, prevailing practice is to progressively reduce the dose until finding the minimum effective dose because many patients present recurrence or progression of the disease after complete sirolimus suspension, occurring in up to 49% of patients in a review by Geeurickx and Labarque, requiring resumption of treatment.²⁰ In our series, only 2 patients showed recurrence when there was worsening and need of restart sirolimus at least after 1 month of withdrawal. The reason for small number of recurrences is probably because most of our patients slowly decrease dosage and many of them continue on sirolimus with a minimum effective dose if there is immediate worsening after withdrawal.

The longest duration of treatment reported in the literature is from a patient with BRBN syndrome that was on sirolimus for 9 years,²¹ as well as one of our first patients with GSD (110 months).

There are several limitations in this study including the wide variety of phenotypes in our series of patients (which makes it difficult to extrapolate these results), the data gap especially in blood levels (which is probably why we did not find any correlation between dose and levels) and the generally good response to sirolimus (making it harder to find statistically significant differences in the analysis).

Conclusion

Overall response to sirolimus was good, but the low rate of progression made it difficult to find statistically significant differences. Risk factors for a negative response were not found, as most patients responded irrespective of their severity in location, extension or symptoms. Patients <5 years old showed a better and faster response than older patients, highlighting the advantages of early treatment. We recommend that patients continue sirolimus treatment for at least 1 year before declaring success or failure. Further research on the durability of the response, recurrence after withdrawal and long-term side effects is necessary to clarify the future of treatment for our patients, as well as standardizing practice for individualize dosage, levels, length of treatment, and slow taper for each patient.

Ethics declaration

This study was approved by the ethics committee. The data and samples used in this study were obtained with written informed consent.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.