ACPOC - The Association of Children's Prosthetic-Orthotic Clinics Founded in 1978

Member Locator

View Options - Click to expand
Print Options - Click to expand
E-Mail Options - Click to expand

Orthotic Treatment of Pectus Carinatum: A Secondary Retrospective Analysis


Determining the effectiveness of bracing for the treatment of pectus carinatum (PC) using quantitative methods has been a challenge in the field of orthotics. The purpose of this study is to draw intraand intersubject comparisons related to patient outcomes and factors that influence change through the use of quantitative data collected from the Insignia scanner and surveys. Seventy-six patients (males = 66; females = 10) ages 9-18 were prescribed dynamic chest compression (DCC) orthoses for the treatment of PC. Paired t-tests were used to determine that the most significant change occurred in the PC during the first three months of treatment. An Analysis of Variance (ANOVA) was used to determine that both age at initiation of bracing, and wear time significantly influenced the amount of correction achieved. Participant responses to a structured questionnaire indicated that they were generally satisfied and rated their PCs as less severe, both during and after treatment with the brace, and. Results of this study may be used to guide orthotic prescription and protocol for potential patients, and to serve as a basis for future studies on similar topics.


Pectus carinatum (PC), commonly known as "pigeon chest," is a chest wall anomaly that is characterized by an anterior protrusion from the chest wall caused by an overgrowth of the costal cartilages of the rib cage as a result of damaged growth plates in the costochondral junction (Mavanur & Hight, 1993; Meilke & Winter, 1993; Haje & Bowen, 1992).

Since the initiation of bracing pectus carinatum in Brazil by Haje and colleagues, in the 1990s, questions regarding the effectiveness of treatment and the factors that influence positive outcomes have been examined by a handful of researchers (Banever et al., 2006; Egan et al., 2000, Frey et al., 2006; Haje & Bowen, 1992; Mavanur & Hight, 2008; Meilke & Winter, 1993; Stephenson & Dubois, 2008). Across the available literature, the general consensus is that bracing is an effective and minimally invasive treatment for pectus carinatum (Banever et al., 2006; Egan et al., 2000, Frey et al., 2006; Haje & Bowen, 1992; Mavanur & Hight, 2008; Meilke & Winter, 1993; Stephenson & Dubois, 2008). However, the methodologies, outcome measures, and results are varied. As a result, it is difficult to compare results across studies.

Measuring objective outcomes with this population is a challenge because of patient growth and postural change throughout the course of treatment. While an anterior-posterior measurement of the pectus at its most prominent point to the middle of the back provides a quantitative measure for a snapshot in time, it is not effective for tracking change long term. In addition to brace effects, several factors may influence this measurement and must be considered when investigating the efficacy of this treatment approach (e.g. weight gain/loss, typical child growth and development, building muscle mass). The subjective accounts of change and qualitative data collected from patients in previous studies are helpful in justifying the effectiveness of bracing pectus carinatum. However, quantitative scientific and measurable outcomes combined with these methods will better serve to substantiate the use of this treatment for individuals with PC.

Background Information:

Clinical Presentation/Appearance
Pectus carinatum deformities are generally classified by the structures involved and the appearance of the deformity. Two common categories include chondrogladiolar (more common) and chondromanubrial. The chondrogladiolar type involves symmetric or asymmetric overgrowth of the inferior costal cartilages and gladiolus, while the other type affects the manubrium and superior costal cartilages (Frey, et al., 2006). Pectus carinatum may are also referred to by type: I, II, and III. Type 1 ("keel chest") has a symmetrical appearance, Type II ("arcuate") involves the manubrium and first two sternocostal cartilages, and Type III is generally an asymmetrical/lateral protrusion occasionally accompanied with a depression on the contralateral side (Meilke & Winter, 1993). Stephenson and Dubois (2008) have also identified a significant rotational component of the sternal body that may influence the overall presentation of the PC.
Associated Symptoms/Problems
Pectus carinatum is generally asymptomatic, but in some cases, individuals have reported shortness of breath, chest pain, limitations in tidal volume, and exercise intolerance (Mavanur & Hight, 2008; Banever, Konefal, Gettens, & Moriarty, 2006; Frey et al., 2006; Stephenson & DuBois, 2008)). During adolescence, there does not appear to be any significant compromise to cardiopulmonary structures; however, evidence suggests that decreased elasticity post maturation of skeletal structures may account for pulmonary changes in adulthood (Mavanur & Hight, 2008; Haje & Bowen, 1992). Although physical symptoms may be minimal, patients have reported that the psychological and social impacts associated with this condition are often significant. Issues related to cosmesis in this adolescent population may contribute to poor self-image and self-esteem, anxiety, diminished motivation, and a high degree of self-observation that may influence relationships with their peers and performance in other aspects of their lives may be affected (Mavanur & Hight, 2008; Banever et al. 2006; Frey et al., 2006; Haje & Bowen, 1992; Einsiedal & Clausner, 1999).


The cause of PC remains unknown. However, there are several factors that may influence the rate of growth of the costal cartilages. Some hypotheses include intrauterine pressures, shortening of the central tendon of the diaphragm, and disruption in Type 2 collagen development. PC is often co morbid with other connective tissue disorders, thoracic dystrophy, and congenital abnormalities related to development of the abdomen and chest cavity. A higher incidence is also noted in individuals with scoliosis (21%), Poland's Syndrome, and Marfan's Syndrome, (Banever et al., 2006) Meilke & Winter, 1993; Mavanur & Hight, 2008). Frey et al. (2006) also report that heredity may play a role (25%).


The incidence of PC is 1:300 births with an approximate prevalence of .06% of the population, and affects males more than females (4:1) (Haje & Bowen, 1992; Mavanur & Hight, 2008; Meilke & Winter, 1993). PC accounts for 5% of all chest wall deformities versus pectus excavatum (anterior depression in the chest wall) at 6%.


Typical treatment protocol for PC consisted of extensive reconstructive surgery of the chest wall. The first surgery, performed by Meyer in 1911, involved excision of the costal cartilage, internal fixation, and a posterior sternal osteotomy. (Mavanur & Hight, 2008; Banever et al., 2006). Several other techniques (e.g. Nuss, Ratvich) were developed years later, but no one approach has been deemed "ideal" (Haje & Bowen, 1992). Due to the invasive nature of these procedures, complications and residual effects have been noted to include lasting growth restriction of the chest wall, seroma, pneumothorax, significant bleeding, recurrence requiring revision, scarring, Jeune's syndrome, with generally poor long term outcomes (Banever et al., 2006; Haje & Bowen, 1992; Mavanur & Hight, 2008).

Bracing and casting techniques to treatment pectus carinatum have been developed and applied in recent years. Wolff's Law serves as the basis for these two approaches; it states that bone and related tissue growth and development are influenced by the external pressures placed on them (Haje & Bowen, 1992). In relation to PC, excessive pressure placed on the prominent area of the chest wall will theoretically remodel the growing tissues to prevent the protrusion from worsening and/or correct it. Overall this treatment approach was initially disregarded, bracing has gained increasing popularity as an attractive alternative to the surgical approach throughout the course of the past two decades (Frey et al., 2006; Mavanur & Hight, 2008; Meilke & Winter, 1993; Stephenson & DuBois, 2008).

Literature Review:


A total of eight studies were reviewed. Two studies described their inclusion of participants and their decision to begin bracing based on the flexibility of the chest wall (Frey et al., 2006; Haje & Bowen, 1992.) The types of braces used were either the bi-valve torso style brace with floating pads and leather straps or the West Coast style brace with metal uprights and opposing compressive pads (Banever et al., 2006; Frey et al., 2006; Haje & Bowen, 1992). In one case study, casting was also used for 6 weeks before the initiation of bracing (Meilke & Winter 1993). Prescribed wear schedules were generally full time (23 hours) for all of the studies except for the two by Mavanur & Hight (2008) and Frey et al. (2006) who suggested 12 hours and 14-16 hours, respectively. Frey et al. (2006) was also the only study to cite that most of their participants removed the brace to sleep at night. Duration of wear ranged from 7 months to 36 months across the studies or until the participants demonstrated completion of linear growth (with a minimum of two years) as in the study by Frey et al. (2006). Three of the studies also included breathing exercises to maximize the growth and expansion of the chest wall as a part of their treatment protocols (Banever et al., 2006; Egan et al., 2000; Stephenson & DuBois, 2008).

Outcome Measures

Some studies documented the use of photographs, questionnaires, and patient accounts as subjective forms of outcome measure (Frey et al. 2006; Haje & Bowen, 1992; Mavanur & Hight, 2008; Meilke & Winter, 1993). Others reported quantitative data and/or positive outcomes but did not document how these results were measured or compared pre and post treatment (Banever et al., 2006; Mavanur & Hight, 2008). Only one study clearly examined and documented i quantitative measures through the use of magnetic resonance imaging (MRI) (Stephenson and DuBois, 2008). They used the Haller index (relationship between the transverse and anterior-posterior inner diameters of the chest wall), angle of sternal rotation, and an asymmetry index (comparing differences between anterior-posterior diameter of the thorax for right/left sides) in addition to qualitative measures. They found the greatest indicator for successful bracing was a 50% reduction in sternal rotation angle from pre-treatment to post-treatment.


Because the outcome measures for some of these studies are vague, the results were not well-defined. For example, one study reported that 75% of the participants demonstrated "good" to "complete" correction (Banever et al., 2006). Some studies suggest that the amount, and type of correction may be related to a variety of factors including wear time, location of the PC, and the patient's age at the initiation of bracing. They reported a positive correlation between correction and amount of hours the brace was worn, and greater correction for PCs located inferolaterally in relation to the sternum (Egan, et al., 2000; Haje & Bowen, 1992; Frey et al., 2006; Stephenson & DuBois, 2008). While the study by Frey et al. (2006) suggest that there are more favorable outcomes for individuals who start bracing at a younger age (e.g. 12-14 vs. 14-16), Banever's study (2006) suggested that there was no significant correlation between age and outcome. With regard to duration of wear and visibility of results one study indicated that the greatest improvement occurred after 6 months of wear (Egan et al., 2000). Another reported that results can be seen after six weeks with the most change occurring within 12 months and then reaching a plateau after 18 months (Frey et al., 2006). Several studies cited issues related to bracing which included rashes, bulkiness of the brace, activity limitations, and non-compliance(Banever et al., 2006; Haje & Bowen, 1992; Stephenson & Dubois, 2008). Only one study reported that compliance was not an issue and documented a rate of 90% (Frey et al., 2006).


The purpose of this study is multifaceted: to examine retrospective data and identify trends/factors that influence change within our patient population in comparison with the available literature, and propose a simple and reliable protocol to accurately capture intra-subject quantitative change in the pectus carinatum over time. This data may in turn be used to draw meaningful comparisons between subjects with regard to their response to bracing. In addition to quantitative measures, subjective data will offer insight into the patient experience and indicate patients' satisfaction with the bracing process. The data derived from this study and related studies may be considered in revising the referral, data collection, and treatment process for individuals with this condition in the future. It will also be used to direct and design more rigorous controlled studies.


Based on the clinical experience with patients with PC examined in this study in combination with the available literature, several hypotheses were made including:

  1. Younger patients are likely to respond more favorably and quickly to bracing;
  2. There will be a positive correlation between the number of hours the brace is worn and the amount of correction achieved;
  3. Older patients are less likely to be compliant with bracing; and
  4. The Insignia scanner and technique used to measure change in these patients is an effective means of quantitatively collecting data to allow for comparison both intra- and intersubject.


The subjects in this study were 76 patients (males = 66; females = 10; ages 9-18) with pectus carinatum who were fit with a brace for treatment of their chest wall anomaly at an orthotic facility in Hartford, Connecticut, between the years of 2005 and 2009. This number does not reflect individuals who were referred for bracing, but opted not to begin treatment. Forty members of the group returned for follow-up after brace delivery yielding 56% compliance with an average age of 13.7 years. Thirty-two were lost to follow-up (average age 14.3), and four were considered new patients at the time of data analysis. The average age at brace delivery was 14.35 years. Two of the subjects had co morbid conditions: one with Poland's Syndrome and one with Scheuermann's kyphosis. Total duration of treatment for subjects examined ranged from three months to approximately 3 years.


The subjects in this study were referred for orthotic treatment by a physician and then evaluated, measured, and fit with a brace by the same certified orthotist. Subjects were considered eligible bracing candidates based on several factors including flexibility of the pectus deformity, age, and location of the pectus. Flexibility of the deformity was subjectively determined by the amount of pressure it required to "correct" the PC using one hand to apply force over the deformity and the other to stabilize at the back, similar to Frey et al. (2006) and Haje & Bowen (1992). Likelihood of patient compliance was also considered. However, the informed decision to initiate bracing was ultimately made by the subject and their parent/guardian.

All subjects were measured for the brace using a plaster mold. This mold was used to build a West Coast style low-profile dynamic compression orthosis (DCC) consisting of anterior and posterior sections constructed from two aluminum uprights contoured to the patient's body shape (see Appendix 1 ). Opposing silicone lined foam pads (one over the chest prominence and the other in the middle of the back) were attached to plastic mounts riveted to the uprights. The front and back sections were connected by plastic threads and buckles (similar to ski boot buckles) for easy donning/doffing and adjustment of pressure application. At the time of delivery, the brace was fit to the patient with adjustments made as needed. Subjects were instructed in donning/doffing, skin care, brace care, fitting issues, and wear schedule. Prescribed wearing time was a minimum of 12 hours/day, especially night time wear, and it was encouraged that the brace be worn as much as possible up to 23 hours/day. An initial scan was also obtained with the Insignia Scanner to serve as a baseline image for subjective comparison of shape and size of the PC. The scanner was calibrated and tested for accuracy (1mm at any circumference) and reliability. Patients were seen for follow-up three weeks after the brace delivery to make adjustments, answer questions, etc. The next follow-up was three months post brace delivery and then in three month increments there after. At each three-month follow-up, subjects were re-scanned using the same Insignia scanner to compare and document any changes. It is important to note that subjects were instructed to remove the brace 24 hours before each follow-up appointment so there would be no influence from the brace reflected in the scan. Follow-up appointments and scans were scheduled until the doctor discontinued bracing (e.g. skeletal maturity), the patient independently chose to stop wearing the brace, or the patient failed to return for follow-up appointments.

Quantitative Measurement

Data from the scans was used to determine the effectiveness of bracing throughout treatment via assessment of qualitative change and quantitative measures. Quantitative data collected from the scans included circumference at the most prominent point of the PC, maximum anterior-posterior (A/P) and medial-lateral (M/L) measurements at this circumference, as well as a point-to-point measurement from the most prominent point to the center of the back at the same level. Changes in the PC deformity were expressed as a ratio (and/or percentage) between the anterior-posterior point-to -point measurement in relation to the M/L. Due to patient growth throughout the treatment process, this method was used to capture changes related to overall growth, weight gain/loss, etc. as it would be expected that these dimensions would grow at similar rates throughout development. A decrease in the ratio between these two measures indicated a positive outcome. A ratio that stayed the same or increased in value indicated that the brace was preventing the PC from getting worse or ineffective, respectively. These percentages were also used to determine a relative change in millimeters between follow-ups and at the end of treatment based on the point-to-point measurement from the previous appointment. Manual measurements, patient wear-time, and patient satisfaction were also collected throughout treatment, but not consistently for individual subjects.

Qualitative Measures

Surveys seeking qualitative and additional quantitative information about the PC history, bracing experience, wear time, and satisfaction were sent to each subject and his/her parent(s) April 2009 (see Appendix 2 and Appendix 3 ). Self-addressed stamped envelopes were included in an attempt to increase the return rate. Return of the survey served as one means of consent to participate in the study. Of the sixty-nine surveys mailed (contact information was not accessible for some patients) and twenty were returned at a rate of approximately 28%. Included in the surveys were three continuous rating scales (based on a 10 cm line) related to the patient's perception of the severity of the pectus before and after/during treatment and overall satisfaction

Statistical analyses

The quantitative data collected was examined using various statistical analyses including paired t-tests at alpha level .05, Pearson correlation analysis, and analysis of variance (ANOVA).



Of the forty subjects who followed up, the data from seven was not used secondary to insufficient data, poor scans, confusing records, or conditions that may have affected the outcome or time spent in the brace. There are no known normal ratio (magnitude A/P: magnitude M/L at level of the PC) values published for the measurement technique utilized in this study. The ratios in this study are meant to serve as a basis for comparison of inter/intra-subject changes in the pectus deformity. The average ratio at the start of treatment for the entire group (n = 33) was 73%. Results for the entire group (calculated from the initial measurements to the most recent follow-up) indicate a 5.8% decrease or approximately 12.5 mm of change, suggestive of successful outcomes related to bracing. Between groups, subjects 14.5 (n=17) years and younger at the initiation of bracing started with an average ratio of 69% and saw a 5% decrease (approximately 9.9 mm) overall. Subjects over 14.5 years of age had an average ratio of 73% at initial scan and demonstrated a 5.2% decrease (approximately 11.89 mm) overall. Four subjects in the older group experienced less than 1 mm of change either positive or negative.

Statistical analyses were used to examine various trends in the data. Paired t-tests looking at the mean ratio of the entire group at initial scan (73%), 3 months (69%), and 6 months (67%) indicate that there is a statistically significant difference between the ratios at initial scan and three months ( p =.000 ) and mean ratios between initial scan and six months ( p=.000 ), but, there is no statistically significant difference between the ratios at three months and six months ( p = .157 ). This suggests that most of the change seems to occur during the first three months of treatment. This data was only analyzed for the three and six month intervals due to insufficient consistent data sets beyond this initial time period. Pearson correlation analysis indicated that there is a high correlation between these mean ratios 835, .871, and .837, for the tests described respectively at p = .000 ). These strong relationships support the use of this ratio method as it accounts for overall changes to trunk dimensions and shape during growth.

An analysis of variance (ANOVA) was also used to examine the relationship between age and number of hours worn and their effect on the mean ratios at initial scan and three month follow-up ( p = .017 ). The data indicates that the interaction of age and and number of hours worn do have a significant effect on the change in ratio between subjects with an effect size of .886 and .884, respectively. Despite the small sample size, the high effect size suggests that the sample size is adequate considering the significance of the results.


Results of the survey were generally informative. Of the twenty surveys returned, four were from subjects who were non-compliant and did not return for followup. Despite their limited satisfaction with bracing (2.73 out of 10), three of the four non-compliant respondents recommended bracing to others and suggested that bracing begin earlier, especially for girls. Those who were compliant, rated the severity of their PC before treatment as 7.2, and after treatment as 3.1 (out of 10). Overall satisfaction for this group was 7.9. Other information collected included average reported wear time (11.7 hours), average age the PC was noted (12.7 years), and age when bracing was initiated (13.25 years). The average reported duration of wear was 14 months. This figure may be skewed as one of the subjects who responded wore his brace for 36 months and is an outlier in this group. Five of the twenty reported a family history of PC and two reported associated symptoms (e.g. cracking/numbness, chest pains). In response to a question regarding reasons bracing was terminated, subjects provided several replies including maturity, puberty, image issues, and trouble breathing. Other factors that influenced wear time included problems with rashes, interference with other activities (e.g. dance), conflict with other brace wear (kyphosis brace), difficulty sleeping, and discomfort during warm weather.


Determining the effectiveness of pectus carinatum bracing and the factors that influence outcomes are integral to formulating appropriate prescription criteria and managing the course of treatment. Similar in nature to preceding studies, the results of this study indicate that bracing is a conservative and effective treatment for pectus carinatum. The difference between this study and the majority of the others is that it provides quantitative comparisons within each subject as well as between subjects through the use of the Insignia scanner and related software. Although there is no published verification substantiating the use of this measurement method (A/P:M/L) to document change, it was consistently used across subjects, and therefore was an effective means to draw comparisons. It also accounted for the challenges associated with capturing changes in the PC despite developmental changes in body configuration. Perhaps the most important result determined by this study is that for all subjects, the most rapid correction occurs during the first three months of wear time. It is important to stress this point with new brace patients to optimize outcomes from the outset. With regard to the hypotheses, it was surprising to see that the amount of correction was higher in the older subjects than in the younger subjects. One possible reason for this may be that the older subjects had a slightly higher ratio at the initiation of bracing, suggesting that their deformities were slightly more severe with greater potential for correctable change. Another influential factor to consider is that chronological age is not always consistent with skeletal age. A "younger" skeletal age may theoretically yield greater results. Data on skeletal maturity was not collected in this study.

As hypothesized, the demographic data suggests that the older subjects were less compliant with bracing; however, it is not clear if the difference in age is statistically significant. A few survey responses suggested that they would have found bracing more helpful had they started earlier. For reasons related to compliance, optimal outcomes, and patient experience, it is important that awareness of PC be raised so that children at risk can be referred to specialists and potentially commence treatment earlier.

In terms of the wear time and the amount of correction achieved, the results suggest that these variables are positively correlated. However, there is insufficient data available to conclude what the minimum requirements are to achieve maximum results. Lack of data leaves this area open for further exploration.

The qualitative data collected suggests that patient experience and satisfaction with bracing is positive. Sharing these insights with new bracing patients will allow them to make informed decisions about treatment options, and may increase compliance in this population. The results of this study may also influence treatment practices among orthotists and provide a consistent way to measure the effectiveness of bracing.


Unfortunately there were several limitations in this study with most significant being the loss of subjects to follow-up. As a result, a limited data set was available for analysis. In addition, due to significant gaps in data, analysis of results was limited to only three and six month time frames. There was also no clear protocol with regard to initiation and termination of bracing, nor was there information included regarding individuals who were referred, but opted not to brace, or those who went on to surgery. Information was skewed to those individuals who were compliant with bracing. All of this information would have been helpful to improve consistency and would have increased knowledge of all of the subjects being treated for PC within the practice of the prescribing physicians.

With regard to the method of quantitative measurement used in this study, although it was formulated based on a scientifically proven method of measurement (Haller Index) it has no previously accepted scientific merit on its own relative to PC. However, within the confines of this study, it served as a good method of quantitative comparison. One of the challenges of using this method is that there are no age-stated norms for what the relationship between the A/P and M/L measurements should be for the "normal" population, or for the amount of variability that exists within the population. This information would be helpful as a comparison between patients with and without PC.

This study also lacked the resources to conduct higher-level statistical analysis for further exploration of certain variables and the relationships between them. Consultation with more experienced researchers and data analysts may have proven helpful in identifying other statistically significant outcomes.

Future Considerations

With all of the research that has been conducted, questions remain about PC and bracing treatment and related methodologies. Overall, the management of this complex cartilaginous deformity deserves a standardized approach to treatment comparable to those established for other skeletal abnormalities (e.g. scoliosis). In order to accomplish this goal, several factors require further exploration. A future study may include a more stringent protocol, similar to scoliosis screening, involving x-rays to determine bone age as a guideline for brace prescription, treatment protocol, and termination of bracing. Use of compliance monitors to determine reported vs. actual wear time will assist in determining minimum threshold for wear while ensuring optimal results. The study may also include a group of "normal" patients who are measured at the same increments as the brace patients to determine how the PC group is different/similar. With regard to data, it should be collected more systematically and consistently for each subject at the time of follow-up so that the data is prospective instead of retrospective. Consistent manual measurements should also be taken with the scan to compare the accuracy between methods and avoid discrepancies. Comparisons should also be done among individuals who made other choices with regard to PC treatment (e.g. no treatment, surgery).

Appendix 1 | Appendix 2 | Appendix 3

Connecticut Children's Medical Center, Hartford, Connecticut