Nissen fundoplication in Cornelia de Lange syndrome spectrum: Who are the potential candidates?
Abstract
Cornelia de Lange spectrum (CdLSp) is a rare genetic condition characterized by intellectual disability, facial dysmorphisms, major malformations, growth impairment, and development delay. Approximately 80% of CdLSp patients have gastroesophageal reflux disease (GERD) with a varied clinical presentation. The aim of this study is to define potential clinical/genetic risk factors based on the clinical phenotype description of CdLSp patients with severe GERD who underwent surgical treatment.
We retrospectively collected data from 23 CdLSp patients, 13 females and 10 males. Mean age of the patients undergoing surgical treatment was of 4 years. 21/23 (91%) had a molecular characterization, of which 21/21 (100%) had a NIPBL gene mutation, while 2/23 (9%) did not have a genetical characterization, only a clinical diagnosis. Most of our patients presented a moderate–severe severity score (21/23, 91%) with limb malformations evidenced in 10/23 (44%) of our patients and a moderate–severe intellectual disability in 20/23 (87%). Therefore, CdLSp patients harboring NIPBL variants, upper limb malformations and severe psychomotor delay are more likely to suffer from severe GERD, not responsive to proton pump inhibitors treatment. These features should be considered as clinical markers for potentially severe GERD that might require surgical treatment.
1 - INTRODUCTION
Cornelia de Lange syndrome (CdLS) is a rare genetic condition characterized by intellectual disability, facial dysmorphisms, major malformations, growth impairment, and developmental delay. There is a wide clinical variability, with about 20–30%of the patients showing a mild phenotype (Liu & Krantz, 2009). In order to describe the phenotypic variability, an International Consensus Statement (ICS) has recently defined the whole CdLS phenotype as a spectrum (Kline et al., 2018).
The CdLS spectrum (CdLSp) involves classic and nonclassic CdLSphenotype. Most of CdLSp patients harbor a pathogenic variant in a cohesin function-relevant gene (NIPBL, SMC1A, SMC3, RAD21, BRD4, HDAC8, ANKRD11), defining CdLSp as a cohesinopathy. Nevertheless, for some patients the genetic characterization (causal variant) remains unknown (Deardorff et al., 2007; Deardorff et al., 2012; S. Huisman et al., 2017; S. A. Huisman, Redeker,Maas,Mannens, & Hennekam, 2013; Kline et al., 2018; Krantz et al., 2004; Musio et al., 2006; Olley et al., 2018; Parenti et al., 2016; Parenti et al., 2016).
CdLSp genotype–phenotype correlations suggested a great variability within and between the different gene mutations. The ICS proposed that individuals with NIPBL truncating mutations are more likely to present “classical” clinical features with severely impaired cognitive functions in comparison to those with other causal mutations or those for whom no mutations have been identified (Gillis et al., 2004; Mannini, Cucco, Quarantotti, Krantz, & Musio, 2013; Sarogni, Pallotta, & Musio, 2019). The scientific community has established a genetic severity scale for the other genes associated with CdLSp, with HDAC8 gene mutations presenting a more severe clinical presentation and SMC3 gene mutations with the least severe presentation. Patients with NIPBL missense mutations generally present a mild phenotype as well. However all these correlations are based on a small number of patients, further studies are needed for confirmation (Kaur et al., 2016).
Patients with CdLSp frequently present gastrointestinal malformations, such as duodenal atresia, annular pancreas (Wick, Simmons, Ludwig, & Kleinberg, 1982), imperforate anus (Mende, Drake, Olomi, & Hamel, 2012), Meckel diverticulum (Mariani et al., 2016), and congenital diaphragmatic hernia (Cunniff et al., 1993). Pyloric stenosis has been reported in up to 7% of the infants. Inguinal hernia is common in childhood (Kline et al., 2007).
About 80% of CdLSp patients are considered to have symptoms of gastroesophageal reflux disease (GERD). The prevalence of GERD is not correlated to the severity of the phenotype (Luzzani, Macchini, Valadè, Milani, & Selicorni, 2003) with a varied spectrum of clinical presentation, including feeding problems, failure to thrive, dysphagia, restlessness, or poor sleep. In CdLSp patients, GERD should also be suspected in case of atypical symptoms such as apnea, recurrent abingestis pneumonia, or when a persistent anemia is documented. Behavioral mood swings, usually described in disabled children, such as agitation, irritability, or self-harm injuries, may also correlate with GERD. These atypical symptoms could be a manifestation of pain in patients with poor verbal communication skills. This clinical equivalent has consistently been described in CdLSp patients (Kline et al., 2018; Luzzani et al., 2003).
Due to the difficulty encountered to conduct research in neurologically affected patients, the European Society for Paediatric Gastroenterology Hepatology and Nutrition (ESPGHAN) guidelines proposed a proton pump inhibitors (PPIs) trial therapy with close follow-up for these patients. Combined esophageal multichannel intraluminal impedance with pH monitoring (pH-MII) and/or upper esophagogastroduodenoscopy (EGD) should be performed in patients not responding to medical therapy (Romano et al., 2017).
Upper endoscopy is considered the gold standard approach to evaluate the esophageal involvement. It is recommended to perform multiple biopsies of the esophagus to rule out other causes of esophagitis and to diagnose and monitor a possible Barrett's esophagus frequently described in CdLSp patients with 10% incidence in adolescents and adults (Kline et al., 2007; Macchini et al., 2010). As a matter of fact, a step-up empiric trial with PPIs at proper dosage was suggested by the ICS for CdLSp patients with GERD (Kline et al., 2018). Only if patients are not-responsive to optimal nutritional and medical therapies an endoscopy should be considered because they may to qualify for surgical treatment (Kline et al., 2018; Luzzani et al., 2003; Rosen et al., 2018).
Few papers with small cohorts of CdLSp patients that underwent Nissen fundoplication (NF) are reported in scientific literature (Szyca & Leksowski, 2011). The aim of this study is to describe clinical CdLSp features that could be associated with a high risk of severe GERD which might need a NF surgical treatment.
2 - PATIENTS AND METHODS
We retrospectively collected data related from Italian CdLSp patients assessed in different Italian Clinical Genetic Centers who underwent surgery (NF) at different ages to treat GERD. All these patients were evaluated by a group of expert clinical geneticists (A.S., M.M., A.C., G.Z.) and diagnosed according to clinical criteria published in 2007 (Kline et al., 2007). The reliability of the clinical diagnosis was confirmed according to the new ICS criteria (Kline et al., 2018).
All patients were screened for the most common CdLSp genes: NIPBL, SMC1A, SMC3, HDAC8, RAD21, ANKRD11. The molecular study was performed with gene by gene Sanger approach or with a NGS panel depending on the year of diagnosis. In accordance with prior reports (S. A. Huisman et al., 2013) if the blood samples disclosed a normal genetic result, the tests were repeated with buccal swab samples. When the study was performed, BDR4 gene was still unknown as a CdLSp related gene, therefore, no patient was analyzed for mutations in this gene. Patients with a formal clinical diagnosis of CdLS whose specific CdLSp molecular analysis were normal, had array CGH analysis to rule out copy number variations CNVs as described by van Allen (van Allen et al., 1993). We classified NIPBL mutations into truncated ant not-truncated mutations (not-truncated mutations included splice site, missense, and in-frame deletions), as truncating NIBPL mutations were most frequently evidenced in patients with a severe clinical score (Selicorni et al., 2007).
A wide number of clinical variables were analyzed, such as auxological data, which were plotted on CdLS growth charts (Kline, Barr, & Jackson, 1993), major limb malformations, CdLS severity score, and the presence of global delay or intellectual disability diagnosed by a physician specialized in the field.
3 - RESULTS
3.1 - Patient's clinical features
A total of 23 patients with CdLSp clinical hallmarks with a mean age of 10 years old, of which 13 females and 10 males, were recruited. 21/23 (91%) of these patients had a molecular characterization, while 2/23 (9%) were only clinically diagnosed. All patients with a defined molecular defect showed a NIPBL gene mutation; 57% (12/21) carried 2 PARMA ET AL. a truncated NIPBL mutation, 43% (9/21) a not-truncated NIPBL mutation. Neither other mutation from the cohesin-pathway genes nor any CNVs were identified in the patients with only clinical diagnosis. The auxological data evidenced that the majority of our patients have growth parameters within normal ranges for CdLS population. Limb malformations were present in 10/23 (44%). Considering the severity score, 2/23 (9%) were mild (mean score: 11.5), 7/23 (30%) were moderate (mean score: 18), and 14/23 (61%) were severe (mean score: 30). Regarding psychomotor development/intellectual disability, the majority of our patients (13/23, 57%) evidenced a severe delay, 7/23 (30%) had a moderate delay and just 3/23 (13%) a borderline-mild retardation according to a specialist physician in the field evaluation (Table 1).
According to ICS criteria (Kline et al., 2018), all patients presented a score of ≥11, indicating a classic CdLSp phenotype (Table 2).
3.2 - Surgical treatment motivation
Mean age at the time patients underwent surgical treatment was of 4 years. The patient's ages at the time of NF surgery are shown in Table 3. Two patients (9%) underwent a NF surgery in neonatal period while correcting a diaphragmatic hernia at the same time. Three patients (13%) had NF surgery during the packaging of a percutaneous endoscopic gastrostomy (PEG). Surgery was indicated in other three patients (13%) due to episodes of “ab ingestis” pneumonia, related to persistent vomit/regurgitation. Indication to surgery came after a noneffective drug therapy (high PPIs dosage) for the other 15 patients (65%) who had a clinical diagnosis ofGERD, later confirmed by EGDin 13/15 (90%).
In these patients, the leading symptom was unvoiced pain (15/23), common in CdLSp patients, mostly evident by atypical behavior (agitation, irritability, or self-harm injuries) (Table 3).
TABLE 1 Phenotype and genotype of CdSp patients
Patient number | Gender | Age (aa) | Height | Weight | Gastrointestinal Malformations | Intellectual disability/psychomotor delay | Severity | score | Limb Malformations | NIPBL variation |
---|---|---|---|---|---|---|---|---|---|---|
1 | F | 4 | 50 | 50 | Congenital diaphragmatic hernia | BD | 15 | MD | N | T |
2 | F | 2 | 50–95 | 50 | – | M | 18 | MD | N | NT |
3 | M | 5 | 50 | 50 | Intestinal Malrotation | SD | 37 | SV | Y | T |
4 | F | 11 | 25 | 25–50 | – | SD | 28 | SV | N | T |
5 | F | 12 | 50 | 50 | – | SD | 34 | SV | Y | NT |
6 | F | 10 | 25–50 | 25–75 | – | SD | 32 | SV | N | T |
7 | M | 14 | <3 | 25 | – | SD | 28 | SV | Y | NM |
8 | M | 4 | 50 | 50 | – | M | 16 | MD | N | T |
9 | F | 20 | <25 | 25–75 | Intestinal Malrotation | SD | 34 | SV | Y | NM |
10 | F | 26 | 50 | 50 | – | SD | 24 | SV | N | NT |
11 | M | 9 | 5 | 10 | – | M | 22 | MD | N | T |
12 | F | 7 | 25 | 25 | Hypertrophic pylorics Stenosis | SD | 20 | MD | N | T |
13 | M | 5 | 50–95 | 75 | Congenital diaphragmatic hernia | M | 16 | MD | N | T |
14 | F | 17 | 50–95 | 95 | Hypertrophic pyloric stenosis | SD | 25 | SV | N | T |
15 | M | 27 | <5 | 5 | – | SD | 31 | SV | Y | T |
16 | M | 13 | 25–50 | 25–50 | – | M | 32 | SV | Y | T |
17 | M | 4 | 50 | 50 | – | BD | 11 | ML | N | NT |
18 | F | 3 | 25–50 | 25–50 | – | M | 12 | ML | N | NT |
19 | F | 3 | 3–10 | 3 | – | SD | 32 | SV | Y | NT |
20 | M | 22 | 5 | 10 | – | SD | 32 | SV | Y | T |
21 | F | 1 | 50 | 50 | – | M | 23 | SV | N | NT |
22 | M | 1 | 25–50 | 25–50 | Intestinal Malrotation | SD | 34 | SV | Y | NT |
23 | F | 12 | 50 | 50 | – | BD | 20 | MD | Y | NT |
Abbreviations: Gender: F, female; M, male. Intellectual disability: SD, severe-deep; M, moderate M, BD, borderline-mild. Severity score: SV, severe; MD, moderate; ML, mild (according to Selicorni et al., 2007). Limb malformations: Y, yes; N, N. NIPBL variation: T, truncating; NT, nontruncating, NM, no mutations.
TABLE 2
Specific phenotypical score of patients, according to International Consensus Statement criteria (Kline et al., 2018)
Patient number | Synophrys and/or thick eyebrows (2 pt) | Short nose, concave nasal ridge and/or upturned nasal tip (2 pt) | Long and/or smooth philtrum (2 pt) | Thin upper lip vermilion and/or downturned corners of mouth (2 pt) | Hand oligodactyly and/or adactyly (2 pt) | Congenital diaphragmatic hernia (2 pt) | Global developmental delay and/or intellectual disability (1 pt) | Prenatal growth retardation (<2 SD) (1 pt) | Postnatal growth retardation (<2 SD) (1 pt) | Microcephaly (prenatally and/or postnatally) (1 pt) | Small hands and/or feet (1 pt) | Short fifth finger (1 pt) | Hirsutism (1 pt) | Total |
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
1 | 2 | 2 | 2 | 2 | 0 | 2 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 17 |
2 | 2 | 2 | 2 | 2 | 0 | 0 | 1 | 1 | 0 | 0 | 1 | 1 | 1 | 13 |
3 | 2 | 2 | 2 | 2 | 2 | 0 | 1 | 1 | 0 | 0 | 0 | 0 | 1 | 13 |
4 | 2 | 2 | 2 | 2 | 0 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 15 |
5 | 2 | 2 | 2 | 2 | 2 | 0 | 1 | 1 | 2 | 1 | 0 | 0 | 1 | 16 |
6 | 2 | 2 | 2 | 2 | 0 | 0 | 1 | 1 | 1 | 1 | 1 | 0 | 1 | 14 |
7 | 2 | 2 | 2 | 2 | 2 | 0 | 1 | 1 | 1 | 1 | 1 | 0 | 1 | 16 |
8 | 2 | 2 | 2 | 2 | 0 | 0 | 1 | 0 | 1 | 0 | 1 | 0 | 1 | 12 |
9 | 2 | 2 | 2 | 2 | 2 | 0 | 1 | 1 | 0 | 1 | 0 | 0 | 1 | 14 |
10 | 2 | 2 | 2 | 2 | 0 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 15 |
11 | 2 | 2 | 2 | 2 | 0 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 15 |
12 | 2 | 2 | 2 | 2 | 0 | 0 | 1 | 1 | 0 | 0 | 1 | 0 | 1 | 12 |
13 | 2 | 2 | 2 | 2 | 0 | 2 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 17 |
14 | 2 | 0 | 2 | 2 | 0 | 0 | 1 | 0 | 1 | 1 | 1 | 1 | 1 | 12 |
15 | 2 | 2 | 2 | 2 | 2 | 0 | 1 | 1 | 1 | 1 | 0 | 0 | 1 | 15 |
16 | 2 | 2 | 2 | 2 | 2 | 0 | 1 | 1 | 1 | 1 | 1 | 0 | 1 | 16 |
17 | 2 | 2 | 2 | 2 | 0 | 0 | 1 | 1 | 1 | 1 | 1 | 1 | 1 | 15 |
18 | 2 | 2 | 2 | 2 | 0 | 0 | 1 | 1 | 1 | 0 | 1 | 1 | 1 | 14 |
19 | 2 | 2 | 2 | 2 | 2 | 0 | 1 | 1 | 0 | 0 | 1 | 1 | 1 | 15 |
20 | 2 | 2 | 2 | 2 | 2 | 0 | 1 | 1 | 0 | 1 | 1 | 1 | 1 | 16 |
21 | 2 | 2 | 2 | 2 | 0 | 0 | 1 | 1 | 0 | 1 | 1 | 0 | 1 | 13 |
22 | 2 | 2 | 2 | 2 | 2 | 0 | 1 | 1 | 0 | 1 | 1 | 1 | 1 | 16 |
23 | 2 | 2 | 2 | 2 | 2 | 0 | 1 | 0 | 0 | 0 | 1 | 0 | 1 | 13 |
TABLE 3
Clinical and instrumental findings, reasons for NF
Patient number | GERD symptoms | EGD | Esophagitis (macroscopic) | Hiatal hernia | Age of NF (months) | Reasons of NF |
---|---|---|---|---|---|---|
1 | NA | N | NA | N | 1 | Diaphragmatic hernia neonatal correction |
2 | Behavioral changes for unvoiced pain, vomiting/regurgitation | Y | N | Y | 12 | Clinical PPI nonresponder in hiatal hernia |
3 | Behavioral changes for unvoiced pain | Y | Y | N | 36 | Esophagitis PPI nonresponder |
4 | Behavioral changes for unvoiced pain | Y | Y | Y | 36 | Esophagitis PPI not responder in voluminous hiatal hernia |
5 | Vomiting, regurgitation | Y | N | N | 48 | Aspiration pneumonia |
6 | Behavioral changes for unvoiced pain | Y | N | Y | 48 | ph-MMI pathological in PPI treatment in hiatal hernia |
7 | Vomiting, regurgitation | Y | N | N | 12 | Aspiration pneumonia |
8 | NA | Y | N | N | 12 | NF+ PEG |
9 | Behavioral changes for unvoiced pain | Y | Y | Y | 96 | Esophagitis PPI nonresponder in voluminous hiatal hernia |
10 | Behavioral changes for unvoiced pain, vomiting/regurgitation | Y | Y | N | 36 | Esophagitis PPI nonresponder; Barrett esophagus |
11 | Behavioral changes for unvoiced pain | Y | Y | N | 108 | Esophagitis PPI nonresponder |
12 | Behavioral changes for unvoiced pain | Y | Y | Y | 36 | Esophagitis PPI nonresponder in hiatal hernia |
13 | NA | N | NA | N | 1 | Diaphragmatic hernia neonatal correction |
14 | Behavioral changes for unvoiced pain | Y | Y | N | 6 | Esophagitis PPI nonresponder |
15 | Behavioral changes for unvoiced pain | Y | Y | Y | 10 | Esophagitis PPI not responder in hiatal hernia |
16 | Behavioral changes for unvoiced pain | Y | Y | Y | 188 | Esophagitis PPI not responder in voluminous hiatal hernia |
17 | Behavioral changes for unvoiced pain, Vomiting | N | NA | N | 46 | Clinical PPI nonresponder |
18 | Behavioral changes for unvoiced pain, vomiting/regurgitation | N | NA | NA | 58 | Clinical PPI nonresponder |
19 | Behavioral changes for unvoiced pain | Y | Y | N | 40 | Esophagitis PPI nonresponder |
20 | Behavioral changes for unvoiced pain, vomiting/regurgitation | Y | Y | N | 84 | Esophagitis PPI nonresponder |
21 | Vomiting, regurgitation | N | NA | N | 18 | Aspiration pneumonia |
22 | NA | N | NA | NA | 8 | NF + PEG |
23 | NA | Y | Y | N | 144 | NF + PEG |
Abbreviations: EGD, esophagogastroduodenoscopy; N, no; NA, data not available; NF, Nissen fundoplication; PEG, percutaneous endoscopic gastrostomy; PPI, proton pump inhibitors; Y, yes.
4 - DISCUSSION
A deeper understanding of this rare disease allowed us to correlate different phenotypes to a particular natural history and prognosis. For this reason, the recently published ICS defined the disease as a spectrum. Multiple studies have suggested that loss-of-function variants in NIPBL gene cause a “classic” phenotype, which is defined by a combination of signs and symptoms, and more severe clinical manifestations.
Other cohesin genes variants or missense variants in NIPBL gene are generally associated with a less severe phenotype (Bhuiyan et al., 2006; Kline et al., 2018; Selicorni et al., 2007). According to these observations, patients with a classic phenotype and a NIPBL variant have a ICS clinical score between 12 and 16, when a score ≥11 indicates for classic CdLSp (Kline et al., 2018). The most frequent and severe digestive disorder evidenced in about 80% of CdLSp patients is GERD. A recent study noted that GERD is more common in individuals with NIPBL variants (71%) than in those with other genetic variants (S. Huisman et al., 2017). ESPGHAN guidelines suggest a trial with PPIs in neurologically affected patients with suspected GERD. If patients are nonresponsive to medical treatment, invasive studies with esophageal pH-MII and/or upper endoscopy are recommended (Romano et al., 2017). A pH-MII could be useful to correlate clinical signs to reflux index in order to distinguish PPIs-responders from PPIs-nonresponders, especially in patients with negative endoscopic findings. This approach may present some limitations due to the complexity of the study and poor compliance of these patients. The same strategy was confirmed by the ICS for CdLSp patients, suggesting a PPIs trial therapy and an upper endoscopy when poorly responsive in order to rule out any GERD complications. An anti-reflux (NF) surgery is limited to those patients who do not respond even when medical and nutritional therapies are optimized (Kline et al., 2018). To the best of our knowledge, this is the first study that thoroughly describes a CdLSp cohort that underwent surgical treatment for GERD. The clinical and genetic characteristics of our cohort suggest that all patients that underwent a NF surgery had a classic phenotype, with an ICS clinical score above 11 (Kline et al., 2018).
Moreover, all patients who were genetically characterized had a NIPBL gene mutation. 12/21 (57%) of these patients had truncated mutations that correlated with a more severe phenotype (Selicorni et al., 2007) also in accordance with the high percentage of patients with limb malformations (10/23, 44%) (Mehta et al., 2016) and the moderate/severe intellectual disability of our patients (20/23, 87%). As a matter of fact, 91% (21/23) of our patients had moderate/severe severity score. All these features are characteristic of CdLSp patients with a severe phenotype. This study is corroborative of these findings; in fact, patients underwent NF for severe GERD are primarily with NIPBL variants and classical phenotype.
From the patients who underwent surgery, a diaphragmatic hernia placed indication to NF in two patients (9%); three patients (13%) underwent NF during a PEG, case-dependent approach (Puntis, Thwaites, Abel, & Stringer, 2000) in other three patients (13%) surgery was indicated due to ab ingestis pneumonia (Szyca & Leksowski, 2011). In the majority of our patients, 15/23 (65%), indication to surgery came after a noneffective drug therapy (high PPIs dosage). In accordance with the ICS for CdLSp patients, an upper endoscopy was performed in 13 patients with persistent symptoms of GERDafter PPIs therapy. The findingswere esophagitis in 11/13 patients and hiatal hernia in six patients (five copresent with esophagitis and one with pathological ph-MMI), as described in Table 3. Regarding medical therapy, the three patients who underwent NF during the PEG packaging and the three patients with abingestis pneumonia history, were previously treated with PPIs, summarizing 21/23 patients (91%) received ineffectivemedical treatment.
In our cohort, the most frequent symptom of GERD was unvoiced pain (15/23; 65%). In fact, CdLSp patients, as subjects with poor communication skills, may express pain through agitation, irritability or self-harm injuries. Thus, a close follow-up of GERD typical and atypical symptoms is mandatory to carefully monitor the effectiveness of GERD medical treatment and to consider surgical treatment as a valuable option in specific cases.
5 - CONCLUSION
Our data show that CdLSp patients with such severe GERD requiring surgical therapy are mostly patients with a classic CdLSp phenotype, as recently defined by the ICS. These patients carry a NIPBL gene variant, mostly truncated, and a high percentage of them have other markers of a severity such as a severe psychomotor delay/intellectual disability and limb reduction defects. Nonspecific clinical features, as irritability and behavioral issues, despite optimal PPIs therapy, can be considered as indicators for formal GERD evaluation. Therefore, from a practical point of view, we can consider the above mentioned clinical and genetic features as risk factors for severe GERD with poor response to medical treatment that may benefit from surgical treatment.
ACKNOWLEDGMENTS
We would like to thank parents and children of the Italian “Associazione di Volontariato Sindrome di Cornelia de Lange” for the collaboration. We are grateful to Mariani Foundation, (Milan), Italy, and S.I.L.V.I.A. Association (Como), Italy, for their support for the clinical activity at U.O.C. Pediatria, ASST Lariana, related to patients affected by rare disease.
REFERENCES
A full list of references]] used in this study
CONFLICT OF INTEREST
The authors declare no potential conflict of interest.
DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.
AUTHOR CONTRIBUTIONS
Barbara Parma and Paola Cianci collected data and wrote the manuscript with support from Milena Mariani, Pietro Betalli, and Francesco Macchini. The authors Anna Cereda, Roberto Panceri, Chiara Fossati, Luciano Maestri, Roberta Onesimo, Giuseppe Zampino, and Maurizio Cheli helped supervise the project; Angelo Selicorni supervised the project. All authors discussed the results and contributed to the final manuscript.