Introduction

Pectus excavatum (PE) is one of the most frequent thoracic malformations that mainly affects males (with a ratio of 5:1 to females). It is estimated that it occurs in 1:400–1:1000 live births [1,2,3]. Upon inspection, it looks like a median or paramedian chest depression of variable shape and depth, the major axis of which is vertically oriented. It can also occur in a particularly lateralized way, involving one hemithorax to a greater extent than the other one, with very asymmetrical contours. The malformation primarily affects the 3rd to the 8th costal cartilages that cave in the body of the sternum in the chest, reducing the distance between the vertebrae and sternum. Haller’s thoracic index (transverse diameter/anteroposterior diameter of the chest) defines the severity of the malformation; value ≥3.2 is considered by some authors to be a pathology [1, 3]. Poston et al. [4] propose a correction of the Haller index, in fact, the “correction index” seems more accurate in reflecting the severity of the pectus excavatum even in patients with nonstandard chest morphologies. The Haller index correlates well with the correction index in pectus patients with standard chest wall dimensions, but is quite discrepant in the nonstandard chest.

According to the Chin classification system, there are three different levels of PE: [5]

  • Type I—deformity with severe depth localized at the sternum;

  • Type II—deformity with less severe depth but more extensive localization;

  • Type III—highly asymmetric or unilateral deformities.

Although controversial, the most reliable etiological theories suggest the presence of a retrosternal ligament (Brown’s theory) and a chondrocostal hyperplasia (Ombredanne theory) [6].

The surgical techniques described are divided into two broad categories:

  • Structural interventions aimed at modification of the osteo-chondral curvature and sternocostal complex with metal bars (like the Nuss technique) or synthetic mesh (such as Rathvich’s modified technique);

  • Filling of the volumetric deficit with autologous tissue (flaps or fat grafts) or heterologous tissue (custom-made silicone implants).

Generally, the malformation is not associated with functional disorders (neither ventilatory or cardiac) and merely constitutes an aesthetic alteration that may result in significant psychological distress, especially in adolescence, thus necessitating correction [7, 8]. To reduce the visibility of the residual scarring produced by corrective surgery and to improve the aesthetic outcome, the authors propose a new prosthetic implant technique through a sternal periareolar access under fiber optic guidance. In this paper, the authors make a retrospective examination of the results of the procedure performed in 11 consecutive cases and make a comparative analysis of the data with those described in the literature for other accesses: vertical and transverse sternal, inframammary and transumbilical.

Methods

From January 2005 to January 2015, 11 patients affected with PE underwent a procedure in which a sternal prosthesis was implanted through a periareolar access.

  • Inclusion criteria Adult, presence of PE, Haller index ≥3.2, malformation types I and II according to the Chin classification system.

  • Exclusion criteria Presence of co-pathologies, functional disorders correlated with PE, associated malformative syndromes (Marfan or Poland), type III PE according to Chin.

The pre- and postoperative photographic documentation was acquired in standard form, in digital format in 5 projections: front, profile and degrees. Information regarding personal data, medical history and operators was recorded on paper. Preoperative screening included blood and urine tests, ECG and chest X-rays in 2 projections: anterior–posterior and lateral side. Clinical examination and photographic documentation were routinely carried out for all the patients at 1, 3, 6 and 12 months. In addition at 1-year follow-up, a chest X-ray was performed. All operations were performed by the senior author. The preoperative planning was performed with patients standing up, using the plaster model to draw the perimeter of the pocket on the skin.

Ten procedures were performed under local anesthesia plus sedation in the presence of an anesthesiologist to monitor vital signs (SpO2, FC, heart activity, breathing, blood pressure) and to ensure adequate patient comfort. For the infiltration, a solution was used consisting of 200 cc physiological solution, 20 cc of 2% lidocaine, 10 cc of 2% carbocaine and 20 cc of 7.5% Naropin + 1 mg of adrenaline added both to reduce the intraoperative bleeding and to lengthen the washout time of the anesthetic drugs. The average amount of solution used was 120 cc, ranging from 80 to 160 cc. We used propofol for sedation during hydrodissection and intravenous midazolam thereafter. A single procedure under general anesthesia was carried out in combination with breast augmentation surgery with round silicone gel soft touch textured implants weighing 240 gr. In all cases, intravenous antibiotic prophylaxis was performed with 2.2 g of amoxicillin clavulanate 2 h before surgery; for pain relief, patients were given a 1-g paracetamol tablet every 8 h for 3 days.

All the sternal implants used were in soft touch silicone elastomer, manufactured using plaster models of the preoperative thorax. The prostheses were all placed in a subcutaneous presternal pocket created with the help of a fiberscope. In all cases, a localized access was made in the medial half of the areola on the right breast from the 12 o’clock position to 6 o’clock. To reduce the risk of infection, drains were not used in the presternal pockets (Figs. 1, 2, 3, 4).

Fig. 1
figure 1

A 21-year-old female affected by type II pectus excavatum deformity according to Chin classification system: a preoperative frontal view; b operative planning showing periareolar incision and plaster model; c postoperative frontal view without sternal scars

Full size image
Fig. 2
figure 2

Lateral view of the same 21-year-old patient

Full size image
Fig. 3
figure 3

A 21-year-old male affected by type I pectus excavatum deformity according to Chin classification system: a lateral preoperative view; b postoperative lateral view; c postoperative frontal view without sternal scars; d particular of periareolar scar

Full size image
Fig. 4
figure 4

Custom-made sternal implant, manufactured in soft touch silicone elastomer using a plaster cast of the thorax

Full size image

The preoperative evaluation of the perception of the malformation and post-operation results were made using different questionnaires:

  • The Pectus Excavatum Evaluation Questionnaire (PEEQ) [9];

  • A modification of the Breast Evaluation Questionnaire, designed by Anderson [10];

  • The Strasser System Evaluation Scale [11].

A systematic search of studies assessing pectus excavatum correction with silicone implants was conducted through Medline, PubMed, Cochrane Library, CINAHL and ProQuest databases.

The data collected in our series were compared with that reported in 4 different works in the literature where other forms of access were used: sternal (vertical and transverse), inframammary and transumbilical to implant similar prostheses in a sample population similar to ours (Table 3) [12,13,14,15].

Statistical Methods

After the descriptive analysis of the results obtained with the periareolar access of the implant and pre-/post-evaluation with statistical tests, the new access method was rated by comparing the results obtained with other results found in the literature. The articles taken into consideration were:

  • for horizontal and inframammary sternal access, papers by Snel et al. [13], Wechselberger et al. [15] and Margulis et al. [14];

  • for umbilical access, the work by Horch et al. [12].

The comparison was made through descriptions of reported cases and by means of a meta-analysis with fixed effects. In particular, the meta-analysis considers as an outcome the percentage of “Excellent” results and/or “Good” according to the Strasser index (considered successful) of total cases in the various studies. The proportions of success are compared and estimates of their variability in the articles are considered: Snel et al. [13], Wechselberger et al. [15] and Margulis et al. [14]. However, the latter is excluded from the meta-analysis as the estimated variability is zero. Therefore, the meta-analysis is limited to only two studies but makes one size (effect size) of the result in the case of presternal access.

Details of the Surgical Technique with Periareolar Access

The patients were all placed in a supine position on the operating table with their arms lying beside their chest. After percutaneous hydrodissection of the periprosthetic pocket, an incision was made in the skin with a number 15 scalpel along the medial border of the areola to the right from 12 o’ clock to 6 o’clock. The mammary glad was incised with an electrocautery to reach the fascial plane. The parenchyma was dissected to create a deep tunnel, following the fascial plane in the ipsilateral breast region up to the side edge of the preplanned location of the pocket. The pocket was then dissected out with scissors, meticulously respecting the perimeter of skin markers. Once the pocket was prepared, hemostasis was carried out with the aid of the fiberscope followed by thorough irrigation with saline and betadine solutions. The upper and lower corners on the left side of the prosthesis were anchored with transcutaneous silk 0 stitches. With the help of a retractor, the prosthesis was introduced into the pocket caudal edge first. The maneuver to correctly position the prosthesis is performed simultaneously by two operators: the first surgeon squeezing the implant into the parasternal tunnel, while the second operator guides the implant by pulling the anchoring sutures previously placed. The sternal pocket was separated from the tunnel by closing its access using 2/0 Vicryl stitches. Once the gland was sutured, the skin access was sutured in multiple layers using absorbable glyconate monofilament. No anchoring of the prosthesis was performed. A compressive dressing was made and kept on for 3 weeks.

Results

Out of 11 cases treated, 8 were male and 3 were female with a mean age of 20.27, ranging from 19 to 23 years old. No patient had functional, cardiac or respiratory disorders related to the malformation or had other clinically relevant associated comorbidities or malformations such as Marfan or Poland syndrome. The average weight of the implanted prostheses was 158 gr, ranging from 80 to 248 gr. The size of the prostheses used was from 10 cm × 7 cm to 22 cm × 15 cm. The average length of incision around the areola was 4.8 cm, ranging from 4 to 6 cm.

The information regarding patients’ personal data, preoperative characteristics, surgical data and number of days in hospital is listed in Table 1. The average time of follow-up was 25.9 months, ranging from 12 to 60 months.

Table 1 Patients’ data, type and classification of pectus excavatum malformation, surgical details including: postoperative complications, operative time and hospital staying
Full size table

The average duration of the procedure was 66.6 min, ranging from 58 to 78 min; the average length of stay in hospital was 1.46 days, ranging from 1 to 2 days.

There were no reported major complications such as tissue necrosis, implant extrusion or infection. No clinically detectable capsular contractures or cases of deformed or dislocated prosthesis occurred. There were no cases of rotation or total failure. No cases required surgical revision of the operative field, implant or surgical access scar. Among the complications, 6 postoperative seromas (54%) were reported.

The patients’ perception of improvement through the use of 2 questionnaires and an evaluation scale showed substantial improvement in all the aesthetic outcomes (Table 2).

Table 2 Preoperative evaluation of the perception of the malformation by the patients and post-operation results using three different questionnaires: Anderson, PEEQ and Strasser
Full size table

Patients classified as Chin 2 reported deeper malformations with a Haller index on average higher than 0.8 compared to Chin 1 (see Table 1). The average pre- and postoperative differences between the Anderson index and the PEEQ are 2.16 and 2.01, respectively (see Table 2). These differences are statistically significant with a p value practically zero both according to the t test for paired samples and for the nonparametric Wilcoxon test. The latter was performed given the non-normality of the data. The evaluation result of the implant through the periareolar using the Strasser index provides an excellent result in 90.91% of cases and is good in 100%.

The comparison between the different methods of access is made by comparison of the studies described in Table 3. The work of Margulis et al. [14], although interesting, shows a small case record (only 7 cases) with little information either regarding patients or the results obtained; the only reported results (the absence of complications and hospitalization of 1 day only) cannot be considered to be analogous to those reported by other two works, which are very similar to each other. In the descriptions reported in Table 4, the cases of the first two studies are considered. Periareolar rather than sternal access improves the aesthetic result (90.91% of “Excellent” results on the Strasser scale compared to 47.22% in the case of sternal incision); it decreases the hospital stay (up to 2 days with an average of 1.46), with a lower percentage of complications. Although the mean age of the cases of our study is lower, it is in any case comparable with the case histories of the other two studies. The comparison between the supra-umbilical incision and the areolar one can be made only on the aesthetic results and complications. The excellent results of the first access method should, however, be evaluated on a number of cases that is at least equal to those of the present study. In fact, this study is limited to only 3 cases though it reports the results of the evaluation with the Strasser method and the number of complications.

Table 3 Review of the literature
Full size table
Table 4 Descriptive statistics comparing different surgical accesses
Full size table

The result of the meta-analysis reports a proportion of success for the sternal access (the effect size, ES) estimated at 82% with a 95% confidence interval (95% CI) equal to 71–94%. The results obtained in this work, therefore, with a percentage of 100%, are considered excellent. The forest plot of the meta-analysis is shown in Fig. 5. In the meta-analysis study, Wechselberger et al. [15] has a weight of 77.37% while that of Snel et al. [13] is 22.63%. This is due to less variability of the estimate of the proportion of success in the study of Wechselberger et al. [15].

Fig. 5
figure 5

Forest plot from the meta-analysis: the red line represents the result in this study (100%); diamond represents the effect size and its variability resulting from meta-analysis of the studies

Full size image

Discussion

In the absence of obvious functional cardio and respiratory anomalies, PE remains substantially an aesthetic anomaly, often with serious psychological implications for the patients affected by it [8, 16,17,18]. They generally have very high expectations about their surgery, regarding both the final result and the size of the resulting residual scars. Many do not accept invasive thorax remodeling procedures that, in addition to not ensuring improvement in cardiac performance or breathing, are not free from major complications. Furthermore, due to strong tissue memory, the correction of cartilaginous structures is difficult to stabilize and the deformity tends to recur to some degree. Besides, in adults there may be technical limitations related to the lower osteo-structural malleability compared to younger people [19]. In addition to requiring a degree of technical complexity, trained surgical team, hospital stay and longer recovery period, the use of free flaps is complicated by a high morbidity of the donor area, both regarding function and scarring. The adipose tissue grafts require multiple surgical steps and do not always ensure stable results over time. For all these reasons, in selected patients, custom-made silicone implants are the most widespread method to date. The implants can be manufactured with a plaster cast of the thorax or by digital imaging. The latter technique identifies more accurately the limits of the structural defect, excluding integuments like the breasts, and designs more precise implants, reducing the use of intraoperative remodeling. Based on our knowledge, the surgical accesses described in the literature to introduce the prosthetic implants are the following: vertical or transverse sternal, inframammary and transumbilical. Direct sternal accesses are quick and easy to carry out, providing direct control of the surgical field but the resulting scars that remain are very visible and can adversely affect the perception of the final result, negatively affecting the patient on a psychological level. Moreover, the sternal region is statistically one of the most affected by pathological scars, and direct tension resulting from the introduction of the system increases the risk of dehiscence. Placing the scar away from the final point of implant placement minimizes visibility, removes the memory of the malformation, wards off surgical wound dehiscence and avoids future embarrassment for patients. For these reasons, the authors opt for a unilateral periareolar incision using fibroscopically controlled dissection. This approach enables use of the same access whether the introduction of a breast implant is required also. The curvilinear course of the access ensures a greater dimension compared to a linear path of the same extension; this, together with the consistency, high malleability and flexibility of the implants in the soft touch silicone elastomer, allows the introduction of the prosthesis of substantial size. Hydrodissection prepares the groundwork for surgical plans and allows the intervention to be carried out under local anesthesia; long lasting local anesthesia ensures good control of postoperative analgesia. The correct dissection of the pocket limits both dislocation and rotation of the implant. The fiberscope allows valid and reliable hemostasis which together with pressure dressings removes the need for use of drains. Submuscular positioning coverage of the prosthetic edges reduces both the visibility and the palpability of the implant. Probably the high incidence of postsurgical seromas (54% of patients) is due to the creation of a subcutaneous pocket, with dissection within the fat layer where the prosthesis, acting as a foreign body, was placed. The authors chose to avoid the use of drains to reduce the risk of infections in the presence of a subcutaneous implant. When postsurgical seromas occurred, they were all managed conservatively through percutaneous drainage with a 19 G needle in the office, since there was no risk of implant damage, the sternal implant being a solid elastomeric prosthesis.

From a comparative analysis of the data in the literature on the use of other cutaneous accesses for the introduction of the sternal prosthesis it emerged that:

  • the average duration of the procedure with periareolar access was 66.6 min, while the duration of operations described in the literature for sternal or inframammary access is 72 min [14];

  • the average hospital stay for the procedure with periareolar access was about 1.46 days, while procedures in other accesses required up to 5 days in the hospital;

  • according to the Strasser System Evaluation Scale, the periareolar and transumbilical accesses have reported the best results but the case study reported by Horch is based on a smaller sample of patients [12];

  • greater visibility and worse quality of sternal location scars negatively affect the final result [13];

  • in the periareolar access, the incidence of complications, which only comprised postoperative seromas, was comparable to other techniques taken into consideration.

Conclusions

Even if in cases of small areolas this method cannot be performed, it provides excellent cosmetic results compared to the sternal one. However, the meta-analysis conducted (limited to only two studies) shows an index of heterogeneity between studies (I-squares) of 82.5%. Therefore, much of the variability in the results is attributable to the variability between studies, at least as regards the considered outcomes (% of Strasser excellent and/or good). In addition, the periareolar access causes fewer complications and necessitates a shorter average hospital stay than the sternum access. To conclude, according to what is shown in the literature, periareolar access seems to be a quicker procedure, requiring a shorter hospital stay, and results in scars that can more easily be hidden and are more accepted by patients.