Aesth Plast Surg (2007) 31:532539 DOI 10.

1007/s00266-007-0162-8

ORIGINAL ARTICLE

A Retrospective Analysis of 3,000 Primary Aesthetic Breast Augmentations: Postoperative Complications and Associated Factors
A. Araco G. Gravante F. Araco D. Delogu V. Cervelli K. Walgenbach

Received: 29 July 2006 / Accepted: 25 August 2006 Springer Science+Business Media, LLC 2007

Abstract Background: A large retrospective analysis was performed on a homogeneous group of patients undergoing primary aesthetic breast augmentations to dene complication rates and nd associated factors. Methods: Data were collected from the personal databases of two different surgeons working at the Crown House Hospital, Oldbury, Birmingham, United Kingdom. The period considered was January 1996 to December 2001. All patients who received primary breast augmentation with or without associated mastopexy for cosmetic purposes were recorded. Results: A total of 3,002 women were included in the study. Hematomas were present in 46 patients (1.5%), infections in 33 patients (1.1%), breast asymmetries in 23 patients (0.8%), rippling in 21 patients (0.7%), and capsular contractures in 14 patients (0.5%). The multivariate analysis found that implant placement and the technique

used for pocket creation were variables associated with complications (p < 0.05). Capsular contractures carried a progressive cumulative risk and, in our series, appeared 5 years after surgery. No association was found between contractures and hematomas or infections. Conclusions: The overall incidence of complications in our series was relatively high (4.6%). Surgical placement of prostheses and the technique used for pocket creation were associated with complications. However, few patients required reoperation (1.6%), and the overall satisfaction rate was acceptable (visual analog score, 7). Keywords Breast asymmetries Capsular contractures Hematomas; Infections Mammaplasty Rippling Although breast mammaplasty is the most frequently performed aesthetic operation worldwide and recent studies have documented a signicant incidence of local complications, few large-scale studies (>1,000 patients) have focused on aesthetic breast implant surgery complications and their risk factors [2, 5, 7, 8, 10, 11, 13, 14]. As suggested, large studies are difcult to perform for many reasons. Patients who are doing well do not return for follow-up evaluation. Those with complications seek opinions from other surgeons. Many patients live far from the surgeons facility or move to other cities [7]. Furthermore, some studies have investigated heterogeneous patients. For example, the introduction of breast reconstructions into statistical analysis adds many risk factors specic for these patients (e.g., chemo/radiotherapy, nodal dissection) that affect the ability of tissues to repair and counteract bacterial contaminations [7, 8]. On the other hand, a large prospective trial on the use of corticosteroids in pockets showed a reduced incidence of

A. Araco F. Araco K. Walgenbach Crown House Hospital, Oldburi, Birmingham, United Kingdom G. Gravante Department of General Surgery, University of Tor Vergata in Rome, Rome, Italy D. Delogu University La Sapienza of Rome, Rome, Italy V. Cervelli Department of Plastic Surgery, University of Tor Vergata in Rome, Rome, Italy G. Gravante (&) via U. Maddalena 40/a, 00043 Roma, Ciampino, Italy e-mail: ggravante@hotmail.com

123

A Retrospective Analysis of 3,000 Primary Aesthetic Breast Augmentations

533

capsular contractions, but introduced an important difference in treatment and could not be used for the analysis [14]. For all these reasons, we recently conducted a large retrospective survey on the activity of two different surgeons at a single-center institution. The study focused on primary aesthetic breast augmentations to secure a group of homogeneous and healthy patients for the purpose of dening the incidence of complications and nding specic risk or protective factors associated with complications.

Surgery All patients were administered a general anesthesia. An inframammary incision was used with breast augmentations alone, and a periareolar incision was made for associated mastopexy. Surgeons created the pocket manually or with diathermy. Four different approaches were used to prepare the pockets: subglandular, subfascial, submuscular, and dual-plane positioning. Surgeons washed all implants with a povidoneiodine solution. In some cases, pockets also were irrigated with antibiotics (cefuroximegentamicine for patients with specic referred allergies) before prosthesis insertion. All patients received silicon gelcoated prostheses with a texturized surface. The shape was round or anatomic, according to the patients preference. Three different brands were used: Eurosilicone (Eurosilicone Laboratoires, ` re, France), Mentor (Mentor Corporation, Santa La Peyrolie Barbara, CA, USA), and PIP (PolyImplant Prostheses, La Seyne-sur-Mer, France). Drains to suction uid collections were used according to the surgeons preference. Postoperative Care No additional doses of antibiotics were administered. Tramadol usually was given as an analgesic at the patients request. Early mobilization was solicited 1 to 3 h after the operation, and no elastic bands were maintained as the patient began mobilization. Patients free of complications were discharged home 24 h after the operation. Outpatient follow-up visits were planned for postoperative days 7 and 30, after 6 months, then every year, with a physical examination. This study aimed to nd the incidence of complications with aesthetic breast augmentations and associated factors. We recorded from medical charts and databases the type of implant positioning (subglandular, submuscular, dual plane), prostheses (Eurosilicone, Mentor, PIP), technique of pocket creation (manually or with diathermy), use of drains, and use of antiseptics/antibiotics for pocket washing. Furthermore, we recorded eventual complications and the number of days, months, or years that had elapsed since the operation. All these data were inserted into an appropriate Excel database (Microsoft Corporation, Redmond, WA, USA). In most cases, diagnoses of complications were clinical. Two qualied plastic surgeons who had not operated on the patients (A.F. and C.V.), and 1 general surgeon (G.G.) expressed a judgment. If judgments were different, we favored the diagnosis or grading expressed by the majority (2 surgeons). Hematomas were diagnosed with bruising, swelling, and breast rmness greater than expected. Infections were considered with clinical signs (swelling,

Materials and Methods Data were collected from the personal databases of two different surgeons working at the Crown House Hospital, Oldbury, Birmingham, United Kingdom. The period considered was January 1996 to December 2001. We recorded patients who had received primary breast augmentation for cosmetic purposes, with or without associated mastopexy, and who had complete follow-up data for at least 5 years. We excluded those who had undergone reconstructions after mastectomy, with previous breast surgery, with a clinical infection or had received a complete antibiotic course in the 6 months before the operation, and those with a follow-up period shorter than 8 years or with incomplete data. All surgical procedures were performed by one of the two plastic surgeons (A.A. or W.K.) working at the Crown House Hospital. There was consistency in the care provided, both from surgeon to surgeon and over time (i.e. similar operative techniques, uniformly administered antibiotics, similarly applied dressings, same postoperative instructions, and same scheduled follow-up visits). None of the patients were ever told to return as needed. They were given scheduled appointments every year and not charged for follow-up visits. Preoperative Care Patients recovered in the hospital for 1 night. According to standard prophylaxis measures for deep venous thrombosis (DVT), if no history of previous DVT was referred, no heparin was administered. Only elastic stockings and mechanical calf compression were used until mobilization. Patients older than 60 years and those with a family or personal history of DVT received a single dose of lowmolecular-weight heparin 2,000 U/day before the operation. Infection prophylaxis was administered with one dose of intravenous (IV) cefuroxime 750 mg (1 g of erythromycin was given if specic allergies were referred) 10 to 30 min before the operation.

123

534 Table 1 Descriptive statistics Variable No. of subjects Age (years) Breast augmentation: n (%) Alone With mastopexy Incision: n (%) Inframammary Periareolar Positioning: n (%) Retroglandular Subfascial Retropectoral Dual plane Prosthesis type: n (%) Eurosilicone Mentor PIP Implant size (ml): mean (95% CI) Pocket creation: n (%) Manual Diathermy Use of drains: n (%) Use of antiseptics/antibiotics for pocket washing: n (%) Patient satisfaction (110): mean (95% CI) Reoperations: n (%) 1,750 (58.3) 1,252 (41.7) 2,000 (66.6) 1,902 (63.4) 6.9 1.6; 6.97) 47 (1.6) 1,793 (59.7) 477 (15.9) 732 (24.4) 296.8 46.6 (295.1298.5) 750 (25) 752 (25) 1,000 (33.3) 500 (16.7) 2,824 (94.1) 178 (5.9) 2,824 (94.1) 178 (5.9) Result 3,002 32 12

A. Araco et al.

rmation of normal distribution. Normality assumptions have been demonstrated with histograms, Q-Q plots, skewness and kurtosis, and Kolmogorov/Smirnov and Shapiro Wilk testings. Univariate analysis using the Pearson chi-square test of association compared the complication analyzed, a categorical variable, with other categorical variables such as implant positioning (retroglandular, subfascial, retropectoral, dual plane), types of prostheses used (Eurosilicone, Mentor, PIP), technique of pocket creation (manually or with diathermy), use of drains, and use of antiseptics/ antibiotics for pocket washing. If cells in the contingency table had fewer than ve expected counts, Fishers exact testing was used. The point biserial correlation coefcient was used to compare the complication with continuous variables (implant size). Multivariate analysis was performed with the multiple logistic regression analysis, with the model including only independent variables that proved to be signicantly associated with dependent variables at the univariate analysis. All p values less than 0.05 were considered signicant. Survival analysis curves generated using the KaplanMeier methodology depicted the likelihood of a patient remaining free of the complication over time.

Results We reviewed a total of 3,002 women who met the inclusion criteria. The follow-up period ranged from 5 to 7 years, and the mean follow-up period per patient was 73.2 months. Descriptive statistics and frequencies are summarized in Table 1. The overall complication rate was 4.6% (137 patients). Hematomas Hematomas were found in 46 patients (1.5%). Their occurrence showed a temporal left skew distribution, with a median postoperative day of 1, a minimum of 0 days, and a maximum of 37 days (Fig. 1). Most patients who experienced hematomas had received a submuscular placement, using both classic and dual-plane approaches (34 patients, 74%) and less the subglandular approach (12 patients, 26%). For patients with hematomas, prostheses used were Eurosilicone for 50% (23 cases), PIP for 45.6% (21 cases), and Mentor prosthesis for 4.4% (2 cases) of the patients. Pocket creation was performed using the manual technique for 26 patients (56.5%) and diathermy for 20 patients (43.5%). There was no particular distribution for implant size, drains, or antibiotics used to wash pockets. The univariate analysis showed a signicant association of hematomas with the submuscular/dual-plane

PIP, PolyImplant Prostheses; CI, condence interval

erythema, tenderness, fever, and leukocytosis). Rippling, breast asymmetries, or capsular contractions were selfdiagnosed by patients at home (and referred to surgeons) or during follow-up visits. We considered as signicant capsular contractures only those with Baker 3 or 4 grades. Additionally, all abnormalities that required surgical revisions were reported as complications. At the postoperative follow-up visit, the date of onset was dened as the day the surgeon diagnosed the abnormality or the patient rst noticed it. Statistical Analysis The database was constructed with Excel (Microsoft Corporation). Statistical analysis and calculation of sample size were performed using the Statistical Package for the Social Sciences Windows version 13.0 (SPSS, Chicago, IL, USA). The descriptive statistics used for quantitative continuous variables (implant size) and qualitative ordinal variables were the mean and standard deviation after con-

123

A Retrospective Analysis of 3,000 Primary Aesthetic Breast Augmentations

535

patients, 54.5%) over submuscular/dual-plane placement (45.4%) was observed. In patients with infections (n = 33), prostheses used were Eurosilicone in 60.6% (20 cases) and PIP in 39.3% (13 cases). Pocket creation was performed manually for 22 patients (66.6%) and using diathermy for 11 patients (33.3%). Drains were used in 26 patients (78.8%). No particular distribution was present for implant size or antibiotics used to wash pockets. The univariate analysis showed a signicant association of infections with the use of drains, the use of antibiotics to wash pockets, and the Mentor prosthesis. The Mentor prosthesis and antibiotic pocket washing had a protective effect, whereas the use of drains increased the occurrence of infections (p < 0.05). The multivariate logistic regression analysis did not conrm any of the three variables found to be signicant (Table 2). No association between infections and capsular contractures was found using either the univariate or the multivariate analysis. Breast Asymmetry Breast asymmetry was found in 23 patients (0.8%). Their occurrence showed no particular distribution, with a median occurrence of 202 days (a minimum of 71 and a maximum of 438 days) (Fig. 2). Most patients who experienced breast asymmetry had received a submuscular placement using both the classic and dual-plane approaches (17 patients, 73.9%) as well as the subglandular approach (6 patients, 18.2%). For patients with breast asymmetry, prostheses used were PIP in 73.9% (17 cases) and Eurosilicone in 18.2% (6 cases) of the patients. A prevalence was observed in these patients for high-volume implants (280 ml: 19 patients, 82.6%) and the use of drains (91.3%, 21 patients). No particular distribution was found for the technique of pocket creation or the antibiotics used for pocket washing. The univariate analysis showed a signicant association of breast asymmetry with the submuscular/dual-plane approach, and with the PIP and Mentor prostheses. The Mentor prosthesis and the submuscular/dual-plane approach had a protective effect, whereas the PIP increased the occurrence of breast asymmetry (p < 0.05). The multivariate logistic regression analysis conrmed only the submuscular/ dual-plane approach, and its presence decreased the occurrence of breast asymmetry (p < 0.05; Table 2). Rippling Infections Infections were found in 33 patients (1.1%). Their occurrence showed a normal curve distribution, with a mean postoperative day of 22 and a standard deviation of 15 (Fig. 1). A slight prevalence of subglandular placement (18 Rippling was found in 21 patients (0.7%). The occurrence showed a left skewed distribution, with a median occurrence of 479 days (a minimum of 17 and a maximum of 2,254 days) (Fig. 3). Most patients who experienced rippling had received a subglandular placement (17 patients,

Fig. 1 KaplanMeier curve showing the occurrence of acute complications. Upper panel: Hematomas. Lower panel: Infections. Survival analysis curves generated using the KaplanMeier methodology depict the likelihood of a patient remaining complication-free over time. The cumulative survival is the percentage of patients free of the complication at a given time in the follow-up period. Its value, expressed as a fraction, ranges from 0 to 1 (0%100%) and corresponds to the number of noncomplicated patients over the number of all patients (n = 3002)

approach, the manual type of pocket creation, and the PIP and Mentor prostheses. The Mentor prosthesis had a protective effect, whereas others increased the hematoma occurrence (p < 0.05). The multivariate logistic regression analysis conrmed only the submuscular/dual-plane and the manual type of pocket creation as factors that increased hematoma occurrence, whereas the Mentor prostheses decreased it twofold (p < 0.05; Table 2). No association between hematomas and capsular contractures was found at either the univariate or the multivariate analysis.

123

536

A. Araco et al.

Table 2 Results of univariate and multivariate analysis for each complication analyzed Complication Hematomas Univariate analysis Mentor prostheses (protect.) PIP Submuscular/dual-plane approach Manual pocket creation Infections Antibiotics for pocket washing (protect.) Mentor prostheses (protect.) Drains Breast asymmetry Submuscular/dual-plane approach (protect.) Mentor prostheses (protect.) PIP prostheses Rippling Capsular contractions Subglandular approach PIP prostheses Manual pocket creation Drains Protect., protective factor; PIP, polyimplant prostheses Multivariate analysis Submuscular/dual-plane approach Manual pocket creation Submuscular/dual-plane approach (protect.) Subglandular approach

81%). Most patients received Eurosilicone prostheses (12 patients, 57.1%), followed by PIP (7 patients, 33.3%) and the Mentor prosthesis (2 patients, 9.5%). A total of 14 patients (66.6%) had manual pocket creation and drain insertion. Pockets of 12 patients (57.1%) were washed with antibiotics. No particular distribution was observed for implant size. The univariate analysis showed a signicant association of rippling with the subglandular placement of prostheses and the PIP (p < 0.05). The multivariate logistic regression analysis conrmed only the subglandular approach, which increased the probability of rippling twofold (p < 0.05; Table 2).

Capsular Contractions Capsular contractions were found in 14 patients (0.5%). The occurrence showed a normal curve distribution, with a mean occurrence of 5.5 years and a standard deviation of 1 year (Fig. 3). Most patients who experienced capsular contractions had received a subglandular placement (10 patients, 71.4%). The majority of patients received Eurosilicone prostheses (8 patients, 57.1%), followed by PIP (5 patients, 35.7%) and the Mentor prosthesis (1 patient, 7.14%). Nine patients (64.3%) had high-volume implants (325350 ml). Manual pocket creation and drain insertion were performed for 13 patients (92.8%). The pockets of nine patients (64.2%) were not washed with antibiotics before prostheses insertion. The univariate analysis showed a signicant association of capsular contractions with the manual technique of pocket creation and the use of drains (p < 0.05). The multivariate logistic regression analysis did not conrm any of the variables examined (Table 2).

Discussion This large single-center-based study of patients found complication rates similar to those already published in the literature for aesthetic patients and essentially conrmed some widely held beliefs about postoperative complications [2, 5, 7, 8, 10, 11, 13, 14]. Although some recent case reports have described sporadic late occurrences [3, 15, 19], in our series, hematomas

Fig. 2 KaplanMeier curve showing the occurrence of subacute complications including breast asymmetries. Cumulative survival: percentage of patients free of complications

123

A Retrospective Analysis of 3,000 Primary Aesthetic Breast Augmentations

537

Fig. 3 KaplanMeier curve showing the occurrence of chronic complications. Upper panel: Rippling. Lower panel: Capsular contractions. Cum survival: percentage of patients free of complication

appeared mostly in the immediate postoperative period (02 days; Fig. 1). They were not difcult to manage and resolved spontaneously in all patients. To date, no specic studies of associated risk factors have been published, possibly because the condition resolved spontaneously in most cases. In our series, implant types, retropectoral placement, and manual pocket creation seemed to inuence complication occurrence at the univariate analysis, but the multivariate analysis conrmed only the surgical technique and the retropectoral placement as signicant variables that increased the incidence of complications. This could be attributable to the increased vascularization of muscles beneath the fascia and the cut, but not to the coagulate ability of the nger technique. Finally, it was believed that the postoperative presence of hematomas could favor the occurrence of capsular contractures [4]. In our series, no association was found with infections or capsular contractures. Breast implant infections overall involve 2% to 2.5% of patients and represent the leading cause of morbidity after

reconstructive surgery [17]. Oncologic patients have many specic risk factors such as radiotherapy, lymph node dissection, subcutaneous mastectomy, immediate implant placements, adjuvant chemotherapy, and periareolar incisions. For all these reasons, the incidence of infections in reconstructive surgery is higher than in primary aesthetic surgery, occurring in 1% to 1.5% of cases [17]. Few specic studies have investigated aesthetic patients [7]. Our descriptive statistical analysis conrmed the low incidence of infections (1.1%) and found an increasing cumulative risk as patients were followed up within the rst 2 postoperative months (Fig. 1). The univariate analysis indicated that three factors out of all those analyzed were signicantly associated with their occurrence: the use of Mentor prostheses (a protective factor: their use was associated with a decreased occurrence of infections), drain insertion, and pocket washings with antibiotics. It is a widely held belief that drains favor the occurrence of infections because they create a communication with the skin. More discussed is the use of antibiotics to wash pockets. Recently, a study by Adams et al. [1] demonstrated that the use of triple antibiotic breast irrigation decreased capsular contracture rates. Although their median follow-up period was 14 months (and capsular contractures usually occur later), their ndings could relate to the subclinical infection theory, by which a large bacterial load becomes clinically apparent and produces infections, whereas a subclinical load remains silent and produces capsular contractures [7]. In this context, we found antibiotic irrigation to be associated with a diminished incidence of clinical infections. This relates well with the ndings of Adams et al. [1]. However, none of the factors individuated in our series passed the multivariate analysis. This can be interpreted in one of two ways. Either additional factors not investigated in this study are inuencing the occurrence of infections and for this reason must be found and inserted into the multivariate analysis (e.g., same surgeon, associated diseases) or the occurrence of infections in this particular subgroup of patients is so rare as to make it difcult, if not impossible, to nd factors that could prove signicant. Further larger studies, possibly multicentric investigations, are necessary to clarify the correct hypothesis. Finally, infections were not associated with capsular contractures. Breast asymmetry, rippling, and capsular contracture had the lowest incidence. As predicted intuitively, rippling was signicantly associated with subglandular placement and breast asymmetry with the retropectoral placement. Both were conrmed by the multivariate analysis. Probably, rippling manifests itself more with subglandular placement because implants in these cases are more mobile and covered with soft tissue (rarely can rippling be seen or felt through the muscle).

123

538

A. Araco et al.

However, in our series, no association with large implants was found, disproving another widely held belief. Preoperative breast asymmetries are widely diffused (89% of patients who undergo breast augmentations) and demonstrated to inuence aesthetic outcomes [6, 18]. Unfortunately, no specic study analyzed our patients in terms of preexisting breast asymmetry. In our series, the multivariate analysis conrmed retropectoral placement as a signicant factor associated with an increased risk of breast asymmetry. We can speculate that the retropectoral placement exposes implants to asymmetries and dislocations more than subglandular placement due to the specic inuence of muscle contractions. Similar results were found by Hudson [12] with oncologic patients. Capsular contractures were found in only 0.5% of the patients. Curves from KaplanMeier survival analyses conrmed that these conditions are a progressive phenomenon that increase their cumulative risks as patients are followed. In our study, they began to appear after 5 years and progressively developed until the end of the follow-up period (7 years) (Fig. 2). These data discourage, as also discussed by Handel et al. [4], the theory that capsular contractures likely develop within the rst 2 postoperative years [7]. No signicant associations were found with the previously discussed risk factors such as hematomas [7] or infections [17] or with any other additional parameters included in our analysis. This could be related to a real absence of specic associated factors or to the low incidence of this complication, which prevents an exhaustive statistical analysis. The temporal analysis of complications yielded interesting results. Except for hematomas, which appeared mostly within the rst postoperative days, the rst complications were infections that occurred from postoperative day 4 until 2 months (Fig. 1). This agrees with the reported onset of infections, which occurs up to 26 weeks with silicone implants [17]. Breast asymmetries were spawned from the second postoperative month until more than 1 year (Fig. 2). Rippling and capsular contractures were the latest, manifesting, respectively, within 5 years and afterward (Fig. 3). Capsular contractures seemed to have a progressive cumulative risk over time, as suggested by Handel et al. [7] (Fig. 3). The reasons for reoperations of our patients (47 cases, 1.6%) were capsular contractures, followed by infections and breast asymmetries. Despite a relatively high rate of complications and the need for reoperation to manage some of them, patients were largely satised. On a 10-cm visual analog scale (VAS) ranging from 0 (no satisfaction at all) to 10 (maximum satisfaction), they assigned a mean satisfaction value of 7 with a standard deviation of 1.6. The high level of satisfaction expressed by patients undergoing implant surgery had already been expressed in previous

studies, although arbitrary values such as satised or very satised were used [9, 16]. The limitations of our study are different. First, as in all retrospective analyses, specic data were missing including accurate evaluations of preoperative breast asymmetries, as suggested by Rohrich et al. [18]; decisions made on a patient-to-patient basis regarding the preoperative approaches and the surgical technique, and other factors that could be investigated for eventual association with complications. Second, the diagnosis of complications was often subjective. We used the Baker grading system for scoring contractures. Two qualied plastic surgeons (A.F. and C.V, who did not directly operate on the patients.) and one general surgeon (G.G.) expressed a judgment on postoperative patients with clinical suspicion of complications and graded the contracture. If judgments were different, we favored the grade expressed by the majority (2 surgeons). The same was done for breast asymmetries and rippling, although no score system was used in these cases. Furthermore, we diagnosed hematomas and infections on the basis of clinical symptoms or signs. Finally, the long period analyzed (7 years) meant that prostheses used were different over time. We analyzed the three types commonly used and found that some of them (Mentor) were protective against the occurrence of some complications. However, these results were not conrmed by the multivariate analysis, in which it seemed that surgical techniques adopted were more inuent. Conclusions The overall incidence of complications in aesthetic breast augmentations was relatively high and involved 137 patients (4.6%). However only 47 of these patients (1.6%) required a reoperation. Among all complications, capsular contractures have a progressive cumulative risk, that in our series began to appear 5 years after surgery, conrming the hypothesis of a progressive and chronic phenomenon posited by Handel et al. [7]. The univariate analysis found many factors associated with complications, including different implant types, but the multivariate analysis conrmed, among them, only subglandular versus retropectoral placement and manual versus the diathermic pocket creation. The overall satisfaction rate of our patients was 7 on a VAS scale.

References
1. Adams WP Jr, Rios JL, Smith SJ (2006) Enhancing patient outcomes in aesthetic and reconstructive breast surgery using triple

123

A Retrospective Analysis of 3,000 Primary Aesthetic Breast Augmentations antibiotic breast irrigation: Six-year prospective clinical study. Plast Reconstr Surg 117:3036 Biggs TM, Cukier J, Worthing LF (1982) Augmentation mammaplasty: A review of 18 years. Plast Reconstr Surg 69:445452 Brickman M, Parsa NN, Parsa FD (2004) Late hematoma after breast implantation. Aesth Plast Surg 28:8082 Camirand A, Doucet J, Harris J (1999) Breast augmentation: Compressiona very important factor in preventing capsular contracture. Plast Reconstr Surg 104:529538 De Cholnoky T (1970) Augmentation mammaplasty: Survey of complications in 10,941 patients by 265 surgeons. Plast Reconstr Surg 45:573577 Dionyssopoulos A (2005) The nonperfect results of breast implants. Ann Chir Plast Esthet 50:534543 Handel N, Cordray T, Gutierrez J, et al. (2006) A long-term study of outcomes, complications, and patient satisfaction with breast implants. Plast Reconstr Surg 117:757767 Handel N, Jensen JA, Black Q, et al. (1995) The fate of breast implants: A critical analysis of complications and outcomes. Plast Reconstr Surg 96:15211533 Handel N, Wellisch D, Silverstein MJ, et al. (1993) Knowledge, concern, and satisfaction among augmentation mammaplasty patients. Ann Plast Surg 30:1320 Henriksen TF, Fryzek JP, Holmich LR, et al. (2005) Surgical intervention and capsular contracture after breast augmentation: A prospective study of risk factors. Ann Plast Surg 54:343351

539

2. 3. 4.

5.

6. 7.

8.

9.

10.

11. Henriksen TF, Holmich LR, Fryzek JP, et al. (2003) Incidence and severity of short-term complications after breast augmentation: Results from a nationwide breast implant registry. Ann Plast Surg 51:531539 12. Hudson DA (2004) Factors determining shape and symmetry in immediate breast reconstruction. Ann Plast Surg 52:1521 13. Lavine DM (1993) Saline inatable prostheses: 14 years experience. Aesth Plast Surg 17:325330 14. Lemperle G, Exner K (1993) Effect of cortisone on capsular contracture in double-lumen breast implants: Ten years experience. Aesth Plast Surg 17:317323 15. Mauro S, Eugenio F, Roberto B (2005) Late recurrent capsular hematoma after augmentation mammaplasty: Case report. Aesth Plast Surg 29:1012 16. Park AJ, Chetty U, Watson AC (1996) Patient satisfaction following insertion of silicone breast implants. Br J Plast Surg 49:515518 17. Pittet B, Montandon D, Pittet D (2005) Infection in breast implants. Lancet Infect Dis 5:94106 18. Rohrich RJ, Hartley W, Brown S (2003) Incidence of breast and chest wall asymmetry in breast augmentation: A retrospective analysis of 100 patients. Plast Reconstr Surg 111:15131519 19. Veiga DF, Filho JV, Schnaider CS, et al. (2006) Late hematoma after aesthetic breast augmentation with textured silicone prosthesis: A case report. Plast Reconstr Surg 117:757767

123