1Division of Plastic Surgery, Department of Surgery, Baylor Scott & White Health, Temple, TX 76513, USA.
2Department of Biostatistics, Baylor Scott & White Health, Temple, TX 76513, USA.
Dr. Jon P. Ver Halen is currently an Associate Professor with the Texas A&M School of Medicine, Department of Surgery. He is also Associate Program Director of the Plastic Surgery Residency, and Program Director of the Microvascular Surgery Fellowship.
This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License (http://creativecommons.org/licenses/by-nc-sa/3.0/), which allows others to remix, tweak and build upon the work non-commercially, as long as the author is credited and the new creations are licensed under the identical terms.
Aim: Reduction mammaplasty is a commonly performed procedure for the treatment of symptomatic macromastia and is increasingly desired by the obese population. With the increasing prevalence obesity in the population, it is imperative to understand its effect on postoperative outcomes. The purpose of this study is to evaluate obesity as an independent risk factor for postoperative complications in breast reduction surgery using 1:1 patient matching through propensity scores between obese patients and non-obese controls.
Methods: Between 2005 and 2013, the National Surgical Quality Improvement Program dataset identified a total of 6,016 patients as having undergone primary reduction mammaplasty with 30-day postoperative follow-up. Patients were divided into obese [body mass index (BMI) of 30 or more] vs. not obese (BMI below 30). Patients were initially analyzed using standard multivariable analysis. Using propensity scores obtained from a logistic regression model, patients were subsequently matched 1:1 according to preoperative and operative variables to truly isolate the effect of obesity on surgical outcomes. Outcomes were compared between the matched cohorts using McNemar’s test and the Wilcoxon signed rank test.
Results: In unmatched multivariable analysis, rates of overall complications (7.2% vs. 5.3%, P = 0.0024), wound complications (5.5% vs. 3.6%, P = 0.0004), superficial surgical site infection (4.1% vs. 2.8%, P = 0.0050), and wound dehiscence (0.3% vs. 1.1%, P = 0.0005) were found to be statistically different between obese vs. non-obese, respectively. However, when comparing 1:1 matched obese and non-obese patients, only wound complications (4.6% vs. 3.1%, P = 0.0334) were significantly increased in the obese cohort.
Conclusion: Using the most robust statistical tools available, obesity was determined to affect wound complications after breast reduction without increased detriment on other major complications when compared to the non-obese. Obesity should be a considered with other preoperative comorbidities, rather than an independent contraindication to surgery. Breast reduction appears to be safe in the obese patient who is otherwise healthy.
Obesity, breast reduction, reduction mammoplasty, National Surgical Quality Improvement Program, propensity score
Breast reduction surgery, or reduction mammaplasty, is a commonly performed procedure for the treatment of symptomatic macromastia. Over 101,000 breast reductions were performed in 2014. Patients seek relief from back and neck pain, intertriginous rashes, shoulder grooving, ill-fitting clothing, and dissatisfaction with breast appearance. Breast reduction has been shown to improve physical, psychosocial, and sexual well-being. Patients experience enhanced quality of life and are highly satisfied with the procedure.[4,5]
The incidence of postoperative complications in reduction mammaplasty is relatively low, approximately 6%. Problems range from minor wound complications and infections to significant bleeding and thromboembolic events. Thorough preoperative assessment is imperative to patient safety and avoiding poor surgical outcomes.
Many women suffering from symptomatic macromastia are obese. Given the increasing number of obese patients in the general population, the role of body mass index (BMI) as a preoperative assessment factor remains of great interest to the surgical community. Obese patients are more likely to have medical comorbidities, including hypertension, diabetes, chronic respiratory disease and obstructive sleep apnea. They are 35% more likely to have an emergency department visit or hospital admission 30 days after outpatient surgery. Many surgeons require obese patients to lose weight prior to undergoing surgery, and certain insurance carriers use higher weights as refusal criteria for coverage. The role of obesity in postoperative complications following reduction mammaplasty is inconsistently defined in the literature. Some studies associate obesity with increased postoperative complications,[9-13] whereas others find no statistically significant correlation.[14-17]
In 2014, Nelson et al. studied obesity and reduction mammaplasty using the 2005-2011 American College of Surgeons - National Surgical Quality Improvement Program (ACS-NSQIP) datasets. NSQIP is a nationally-validated, risk-adjusted surgical outcomes database to measure and improve the quality of surgical care. The authors reported an increased rate of overall complications in the early 30-day postoperative period among obese patients on multivariable analysis. However, no study to date has harvested the statistical power of the NSQIP dataset with the use of propensity score matching to evaluate the effect of obesity as an independent risk factor on breast reduction outcomes. Multivariable analysis attempts to control for heterogeneity between patient cohorts via advanced statistical techniques. Patient matching, however, eliminates heterogeneity between patient cohorts by 1:1 matching each experimental group patient with a control group patient with similar characteristics. The goal of this study is to isolate the effect of obesity on breast reduction outcomes using 1:1 patient matching.
Patients undergoing primary reduction mammaplasty were identified from the 2005-2013 ACS-NSQIP registry. Methods for data acquisition involved trained research nurses from participating institutions in the United States who collected data through systemic sampling of surgical procedures, as previously described. A total of 240 variables were collected for each case. Further information can be accessed via the ACS-NSQIP website at http://www.acsnsqip.org/. Data are depersonalized and Health Insurance Portability and Accountability Act compliant.
The NSQIP registry was queried using Current Procedural Terminology code 19318 to identify patients who had undergone reduction mammaplasty. Patients were then characterized according to the World Health Organization (WHO) classification of obesity: non-obese (BMI < 30 kg/m2), class I obesity (BMI 30-34.9 kg/m2), class II obesity (BMI 35-39.9 kg/m2), or class III (BMI ≥ 40 kg/m2). Inclusion criteria included primary bilateral breast reductions.
Primary outcomes of interest were analyzed through several pre-defined NSQIP variables, including patient demographics and comorbidities, as well as early surgical complications, defined as adverse events occurring within 30 days after surgery. Demographics included race and age. Comorbidities included diabetes (further classified into insulin dependent and non-insulin dependent), active smoking, alcohol use, dependent functional status, respiratory disease (chronic obstructive pulmonary disease and dyspnea), hypertension, wound infection with in the prior 30 days, heart disease (previous cardiac surgery and history of angina), recent weight loss, bleeding disorder, preoperative sepsis, and prior operation within 30 days.
Surgical complications included wound complications, unplanned return to the operating room and graft/flap failure. Wound complications encompassed superficial surgical site infection (SSI), deep SSI, organ/deep space SSI, and wound dehiscence. Medical complications were defined as renal (renal failure and renal insufficiency), neurologic (stroke, coma, peripheral nerve injury), cardiac (myocardial infarction and cardiac arrest), sepsis, death, venous thromboembolism, failure to wean from ventilator, unplanned reintubation, pneumonia, bleeding, and urinary tract infection. Multivariable analysis of postoperative outcomes was performed to control for those preoperative and intraoperative variables with n > 10 events, and P < 0.05 on bivariate screen.
Obese and non-obese patients were then 1:1 propensity score matched to control for preoperative and intraoperative variables, in order to isolate the effect of obesity on postoperative outcomes. Patient characteristics were matched if n > 10 (i.e., greater than 10 events) and P < 0.05 on bivariate screen. Based on these criteria, matched characteristics included the following: age; diabetes mellitus; active smoking; alcohol use; hypertension; respiratory disease; heart disease; history of transient ischemic attack or stroke; bleeding comorbidity; preoperative wound infection; steroid or immunosuppressant use; recent weight loss > 10% of total body weight in 6 months prior to surgery; total number of comorbidities (none, one, or two or more); American Society of Anesthesiologists (ASA) class; inpatient versus outpatient status; operative time; and total work relative value units.
Characteristics of the sample were summarized using descriptive statistics. Medians and ranges were reported for continuous variables; frequencies and percentages are reported for categorical variables. The chi square test, Fisher’s exact test and the Kruskal-Wallis test were used to determine association between BMI groups and various demographic, comorbidity and outcome variables. If a statistically significant association was detected between a BMI group and a variable, a subgroup analysis was performed using the same tests to determine which of the groups were significantly different from each other. Multivariable analysis of postoperative outcomes was performed for those preoperative and intraoperative variables with n > 10 events, and P < 0.05 on bivariate screen.
Data were then separated into two groups: patients who were obese (BMI of 30 or more) and patients who were not obese (BMI below 30). Patients were matched on a 1:1 basis using propensity scores from a logistic regression model (as described above). Outcomes were then compared between the matched cohorts using McNemar’s test and the Wilcoxon signed rank test. Statistical significance is indicated by P < 0.05.
Between 2005 and 2013, the NSQIP datasets identified a total of 6,016 patients who underwent primary reduction mammaplasty with 30-day postoperative follow-up. The patients were predominantly white, comprising 46.1% of the cohort, and 15.3% were black. Fifty percent were younger than 45 years of age, 44.3% were between 45 and 65 years, and only 5.7% were older than 65 years.
From the total group of patients, 28.7% had at least one preoperative comorbidity, and 9.5% had two or more. Common comorbidities included hypertension (21.7%), active smoking (11.2%), and diabetes (4.7%) [Table 1]. Other factors to assess preoperative risk included ASA classification, with 28.0% in class 1, 60.4% in class 2, 11.4% in class 3, and 0.2% in class 4. A majority of cases (85.4%) were outpatient, and median operative time was 148 min, with a range of 13 to 739 min [Table 2].
|Dependent functional status||16||0.3||1||0.3|
|Chronic obstructive pulmonary disease||40||0.7||4||1.1|
|30-day prior wound infection||23||0.4||4||1.1|
|Previous cardiac surgery||31||0.5||2||0.5|
|History of angina||4||0.1||1||0.3|
|Recent weight loss||12||0.2||1||0.3|
|Prior operation within 30 days||8||0.1||0||0.0|
|2 or more||569||9.5||58||15.3|
|Year of procedure|
|Operative time (min), median and range||148||13-739||148||30-484|
Overall complications within the early postoperative period were rare, at a rate of 6.3%. These were comprised mostly of wound complications (4.6% of total, 72.8% of all complications). The most common wound complication was superficial SSI, occurring in 3.5%. Surgical complications occurred in 1.7%, and medical complications occurred in only 0.6% [Table 3].
|Return to operating room||99||1.6||99||26.2|
|Deep venous thrombosis||3||0.1||3||0.8|
|Urinary tract infection||8||0.1||8||2.1|
BMI data were then assessed according to WHO obesity classification. Overall, 3,054 of the patients (50.8%) were obese, with 1,708 (28.4%) classified as class I, 830 (13.8%) as class II, and 516 (8.6%) as class III. Analysis among the non-obese, overweight, and three classes of obesity showed statistically significant differences in demographic values and several comorbidities. Black patients comprised an increasingly large proportion with each class of obesity (5.8% underweight/normal, 9.3% overweight, 16.1% class I, 25.2% class II, and 37.2% class III) [Table 4].
Comorbidities by body mass index group
|Underweight and normal||Overweight||Class I||Class II||Class III||P-value||Sub-analysis|
|Any diabetes||10||1.1||49||2.4||90||5.3||55||6.6||76||14.7||< 0.0001||abcdefgij|
|Respiratory disease||5||0.5||26||1.3||45||2.6||42||5.1||51||9.9||< 0.0001||bcdefghij|
|30-day prior wound infection||3||0.3||5||0.2||12||0.7||1||0.1||2||0.4||0.1641|
|Prior operation within 30 days||0||0.0||6||0.3||1||0.1||1||0.1||0||0.0||NR|
|Total comorbidities||< 0.0001||abcdefghij|
|2 or more||31||3.3||109||5.4||183||10.7||116||14.0||130||25.2|
Regarding comorbidities, there was a significant increase in the rate of diabetes with increased obesity class: 1.1% in the underweight/normal, 2.4% in the overweight, 5.3% in class I, 6.6% in class II, and 14.7% in class III (P < 0.0001). Hypertension (8.5% underweight/normal, 15.9% overweight, 25.4% class I, 30.6% class II, and 42.6% class III) and respiratory disease (0.5% underweight/normal, 1.3% overweight, 2.6% class I, 5.1% class II, 9.9% class III) increased as well (P < 0.0001). As the class of obesity increased, there were greater total comorbidities (3.3% of underweight/normal patients had at least two comorbidities, compared to 25.2% of class III obese patients) (P < 0.0001). Smoking and alcohol use rates did not increase proportionally with increasing obesity class [Table 4].
Multivariable analysis of postoperative outcomes was performed for those preoperative and intraoperative variables with n > 10 events, and P < 0.05 on bivariate screen [Table 5 and 6]. After controlling for preoperative and interoperative differences by multivariable analysis, a significant increase was noted in any complication in class III obese patients (12.2%), when compared to underweight/normal (4.4%), overweight (5.7%), class I (6.1%) and class II (6.4%) patients (P < 0.0001). Surgical complications were significantly greater when comparing class III (3.3%) with underweight/normal (1.5%), overweight (1.6%) and class I patients (1.2%) (P < 0.0214). Regarding wound complications, class III patients had significantly increased rates (9.3%) compared to all other categories. However, they were also found to be greater in class I (4.7%) and class II patients (4.8%) when compared to underweight and normal weight patients (2.8%) (P < 0.0001). An unexpected return to the operating room occurred more frequently in class III patients (1.6%) relative to underweight/normal, overweight and class I patients (P < 0.0156). Superficial SSI and wound dehiscence also occurred significantly more in class III patients (7.2% and 2.7%, respectively) compared to all other categories; wound dehiscence occurred more in class I obese patients compared to the underweight and normal (P < 0.0001) [Table 5].
Complications and body mass index group
|Underweight and normal||Overweight||Class I||Class II||Class III||P-value||Sub-analysis|
|Any complication||42||4.4||115||5.7||105||6.1||53||6.4||63||12.2||< 0.0001||dgij|
|Wound complication||27||2.8||79||3.9||81||4.7||40||4.8||48||9.3||< 0.0001||bcdgij|
|Return to operating room||14||1.5||33||1.6||19||1.1||16||1.9||17||3.3||0.0156||dgi|
|Superficial SSI||26||2.7||56||2.8||62||3.6||27||3.3||37||7.2||< 0.0001||dgij|
|Wound dehiscence||1||0.1||8||0.4||13||0.8||6||0.7||14||2.7||< 0.0001||bdgij|
|Urinary tract infection||0||0.0||3||0.1||1||0.1||3||0.4||1||0.2||NR|
|Hospital length of stay, median and range||0||0-234||0||0-31||0||0-32||1||0-6||1||0-15||< 0.0001||bcdefghij|
Complications and obesity status - unmatched analysis
|Return to operating room||52||1.7||47||1.6||0.8011|
|Urinary tract infection||5||0.2||3||0.1||0.7266|
|Hospital length of stay, median and range||1||0-32||0||0-234||< 0.0001|
Again on multivariable analysis, obese patients (BMI 30 or more) were compared to the non-obese (BMI < 30) in an unmatched analysis. Rates of overall complications (7.2% vs. 5.3%, P = 0.0024), wound complications (5.5% vs. 3.6%, P = 0.0004), superficial SSI (4.1% vs. 2.8%, P = 0.0050), and wound dehiscence (0.3% vs. 1.1%, P = 0.0005) were found to be statistically different. Total hospital length of stay was found to change with obesity status (P < 0.0001) [Table 6].
Using propensity scores, obese patients were then matched to non-obese patients according to preoperative and operative variables, totaling 1,464 patients in each group. After matching, none of these variables were found to differ between the two groups. When comparing the matched obese vs. non-obese patients, only wound complications (4.6% vs. 3.1%, P = 0.0334) and hospital length of stay (P < 0.0001) were significantly increased in the obese cohort.
Obesity continues to be an epidemic not only in North America, but globally as well. Thirty-six percent of the population is considered obese, with a greater proportion of women than men.[19,20] Symptomatic macromastia is a common condition which afflicts many women, particularly the obese population. Although obesity has been correlated with increased complication rates,[9-13] this population also has a propensity towards having greater medical comorbidities. With literature demonstrating improved longevity in overweight patients compared to normal weight patients, BMI and obesity must therefore be assessed independent of these confounding comorbidities.
Obesity is an often assumed risk factor for postoperative complications following breast reduction surgery. However, its effect on risk outcomes remains incompletely understood. Our study hopes to better define obesity as a preoperative risk factor for breast reduction. Multivariate analysis both before propensity score matching [Tables 5 and 6] and after matching [Tables 7 and 8] was utilized to isolate the effects of obesity alone on postoperative outcomes. Propensity score matching produces estimates that are less biased, more robust, more precise, and with greater empirical power than logistic regression when the number of events are low and there are multiple confounders.
Using propensity scores, obese patients were matched to non-obese patients on the variables listed
|Full cohort||Matched cohort|
|% of patients||P-value||% of patients||P-value|
|n = 2,962||n = 3,054||n = 1,464||n = 1,464|
|Respiratory disease||1.0||4.5||< 0.0001||1.5||1.2||0.5708|
|ASA class||< 0.0001||0.3593|
|1 or 2||94.2||82.7||93.0||93.8|
|3 or 4||5.7||17.3||7.0||6.2|
|Total comorbidities||< 0.0001||0.4571|
|2 or more||4.7||14.0||5.7||50.0|
|Inpatient status||11.6||17.5||< 0.0001||11.8||13.9||0.0831|
|Total RVU, median (range)||16.0 (16.6-54.7)||16.0 (15.6-52.0)||< 0.0001||16.0 (15.6- 49.2)||16.0 (15.6-51.9)||0.7769|
|Operating time, min, median (range)||133 (13-739)||163 (14-636)||< 0.0001||146 (14-543)||146 (14-488)||0.3134|
Complications and obesity status - matched analysis
|Return to operating room||22||1.5||24||1.6||0.8828|
|Urinary tract infection||1||0.2||2||0.1||1.0000|
|Hospital length of stay, median and range||0||0-32||0||0-234||< 0.0001|
Many authors have tried to definitively determine the correlation between obesity and adverse events after surgery. Although many studies consistently demonstrate the deleterious effect of obesity, nearly all analyses are confounded by the effects of associated medical conditions on outcomes. One such study did not find a statistical difference in obese versus non-obese patients in relation to complication and hospital length of stay. Another did not find significant differences in complications attributable to age, BMI, size of resection, smoking status, comorbidities, or surgical technique, even in the morbidly obese. Other studies similarly found no statistically significant difference in complication rates among the obese.[14,15,17]
However, contradictory findings exist in the literature as well, supporting obesity as a risk factor.[6,9-13] Chun et al. identified a threshold of BMI 35.6 at which postoperative complications were increased two-fold, the most common complication being infection. The pioneering study using NSQIP data to analyze BMI and breast reduction complications by Nelson et al. included 4,545 patients between 2005 and 2011. This study used logistic regression to account for demographics and comorbidities. They found an increased rate in overall complications, wound complications in all obesity classes, and major surgical complications in class III obesity.
Multivariate analysis among the non-obese, overweight, and three classes of obesity showed statistically significant differences in demographics, comorbidities, and complication rates [Tables 4-6]. In our unmatched analysis [Table 6], overall complications, wound complications, superficial SSI, and wound dehiscence were significantly increased in the obese population compared to the non-obese cohort after multivariable analysis controlling for significantly different variables between obese and non-obese cohorts. Comorbidities may confound the isolated risk of obesity on complication rates. The distinguishing feature of our study was matching obese patients to non-obese patients with similar preoperative and operative variables, thus eliminating the confounding effect of associated comorbidities on outcomes. While multivariable analysis attempts to control for comorbidities via advanced statistical techniques, 1:1 matching is a dramatically more powerful technique that matches each study patient with a near-identical “control” patient, in spite of detractors of this technique. After analysis of matched cohorts, only wound complications were increased in the obese population [Table 8]. On further analysis, the difference was mainly attributed to a risk of increased surgical site infection in the obese cohort. Of note, length of hospital stay was found to be significantly increased in the normal-weight cohort. On close examination, this was due to a statistical aberrancy (in that the range of values for length of stay for non-obese patients was greater than for obese patients).
In previous studies, dissatisfied patients had frequently experienced postoperative soft tissue necrosis. The pathophysiology of wound healing in obese patients is currently being studied. Obesity has been shown to inhibit bone marrow-derived vasculogenic progenitor cell mobilization, trafficking and function. This in turn impairs the normal response to tissue injury and the proliferation of blood vessels. Adipocytes in fat also produce macrophage migration inhibitory factor, a factor which decreases wound healing through impairment of macrophage polarization/activation and inhibition of adipocyte progenitor cells.
Like other NSQIP-based analyses of reduction mammaplasty, there are limitations to this study.[6,27] Follow-up was only 30 days, a relatively short period of time. NSQIP does not include complications such as seroma, hematoma, fat necrosis, altered nipple sensation, aesthetic outcomes, or hypertrophic scarring. Setala et al. report complication rates amongst normal BMI, overweight, and obese, respectively, as follows: seroma, 8.6% vs. 10.0% vs. 3.0%; hematoma, 8.6% vs. 5.4% vs. 3.0%; and fat necrosis, 1.7% vs. 2.0% vs. 6.1%. These are significant complications for this operation, which vary amongst different BMI classes and may explain a lower overall complication rate in our analysis. These datasets also do not report pedicle design/skin incision or resection weights, which may also affect complication rates. Although NSQIP provides a powerful dataset, further investigation is warranted through prospective analysis, longer follow-up, and more comprehensive collection of operative and complication data.
In conclusion, the increasing number of obese patients accompanied by their desire for breast reduction surgery poses a significant challenge to surgeons. To provide optimal care and minimize surgical risk, understanding the role of obesity in postoperative outcomes is essential. This study was able to independently assess obesity as a surgical risk factor for postoperative wound complications following reduction mammaplasty using multivariate analysis and propensity score matching. Obesity alone should not be the sole determining factor of a patient’s surgical candidacy, but rather as a component of a complete preoperative evaluation. We recommend thorough risk stratification and patient counseling prior to surgical intervention.
There are no conflicts of interest.
1. American Society of Plastic Surgeons. 2014 Plastic Surgery Statistics Report. Available from: http://www.plasticsurgery.org/Documents/news-resources/statistics/2014-statistics/plastic-surgery-statsitics-full-report.pdf [Last Accessed on January 6, 2016].
2. Coriddi M, Nadeau M, Taghizadeh M, Taylor A. Analysis of satisfaction and well-being following breast reduction using a validated survey instrument: the BREAST-Q. Plast Reconstr Surg 2013;132:285-90.DOIPubMed
3. Blomqvist L, Brandberg Y. Three-year follow-up on clinical symptoms and health-related quality of life after reduction mammaplasty. Plast Reconstr Surg 2004;114:49-54.DOIPubMed
4. Carty MJ, Duclos A, Gu X, Elele N, Orgill D. Patient satisfaction and surgeon experience: A follow-up to the reduction mammaplasty learning curve study. Eplasty 2012;12:e22.PubMedPMC
5. Davis GM, Ringler SL, Short K, Sherrick D, Bengtson BP. Reduction mammaplasty: long-term efficacy, morbidity, and patient satisfaction. Plast Reconstr Surg 1995;96:1106-10.DOIPubMed
6. Nelson JA, Fischer JP, Chung CU, West A, Tuggle CT, Serletti JM, Kovach SJ. Obesity and early complications following reduction mammaplasty: an analysis of 4545 patients from the 2005-2011 NSQIP datasets. J Plast Surg Hand Surg 2014;48:334-9.DOIPubMed
7. Sieffert M, Fox JP, Abbott LE, Johnson RM. Obesity is associated with increased health care charges in patients undergoing outpatient plastic surgery. Plast Reconstr Surg 2015;135:1396-404.DOIPubMed
8. Nguyen JT, Wheatley MJ, Schnur PL, Nguyen TA, Winn SR. Reduction mammaplasty: a review of managed care medical policy coverage criteria. Plast Reconstr Surg 2008;121:1092-100.DOIPubMed
9. Zubowski R, Zins JE, Foray-Kaplon A, Yetman RJ, Lucas AR, Papay FA, Heil D, Hutton D. Relationship of obesity and specimen weight to complications in reduction mammaplasty. Plast Reconstr Surg 2000;106:998-1003.DOIPubMed
10. Blomqvist L. Reduction mammaplasty: analysis of patients' weight, resection weights, and late complications. Scand J Plast Reconstr Surg Hand Surg 2006;30:207-10.DOI
11. Gamboa-Bobadilla GM, Killingsworth C. Large-volume reduction mammaplasty: the effect of body mass index on postoperative complications. Ann Plast Surg 2007;58:246-9.DOIPubMed
12. Baldwin CJ, Kelly EJ, Batchelor AG. The variation in breast density and its relationship to delayed wound healing: a prospective study of 40 reduction mammoplasties. J Plast Reconstr Aesthet Surg 2010;63:663-5.DOIPubMed
13. Chun YS, Schwartz MA, Gu X, Lipsitz SR, Carty MJ. Body mass index as a predictor of postoperative complications in reduction mammaplasty. Plast Reconstr Surg 2012;129:e228-33.DOIPubMed
14. Wagner DS, Alfonso DR. The influence of obesity and volume of resection on success in reduction mammaplasty: an outcomes study. Plast Reconstr Surg 2005;115:1034-8.DOI
15. Setala L, Papp A, Joukainen S, Martikainen R, Berg L, Mustonen P, Härmä M. Obesity and complications in breast reduction surgery: are restrictions justified? J Plast Reconstr Aesthet Surg 2009;62:195-9.DOIPubMed
16. Roehl K, Craig ES, Gomez V, Phillips LG. Breast reduction: safe in the morbidly obese? Plast Reconstr Surg 2008;122:370-8.DOIPubMed
17. Cunningham BL, Gear AJ, Kerrigan CL, Collins ED. Analysis of breast reduction complications derived from the BRAVO study. Plast Reconstr Surg 2005;115:1597-604.DOIPubMed
18. Shiloach M, Frencher SK Jr, Steeger JE, Rowell KS, Bartzokis K, Tomeh MG, Richards KE, Ko CY, Hall BL. Toward robust information: data quality and inter-rater reliability in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg 2010;210:6-16.DOIPubMed
19. Ogden CL, Carroll MD, Kit BK, Flegal KM. Prevalence of childhood and adult obesity in the United States, 2011-2012. JAMA 2014;311:806-14.DOIPubMedPMC
20. Wang Y, Beydoun MA. The obesity epidemic in the United States - gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev 2007;29:6-28.DOIPubMed
21. Flegal KM, Kit BK, Orpana H, Graubard BI. Association of all-cause mortality with overweight and obesity using standard body mass index categories: a systematic review and meta-analysis. JAMA 2013;309:71-82.DOIPubMedPMC
22. Cepeda MS, Boston R, Farrar JT, Strom BL. Comparison of logistic regression versus propensity score when the number of events is low and there are multiple confounders. Am J Epidemiol 2003;158:280-7.DOIPubMed
23. Tadiparthi S, Liew SH. Use of patient body mass index as a rationing tool in breast reduction surgery. Plast Reconstr Surg 2008;122:35-6.DOIPubMed
24. Kitsios GD, Dahabreh IJ, Callahan S, Paulus JK, Campagna AC, Dargin JM. Can we trust observational studies using propensity scores in the critical care literature? A systematic comparison with randomized clinical trials. Crit Care Med 2015;43:1870-9.DOIPubMed
25. Wagner IJ, Szpalski C, Allen RJ Jr, Davidson EH, Canizares O, Saadeh PB, Warren SM. Obesity impairs wound closure through a vasculogenic mechanism. Wound Repair Regen 2012;20:512-22.DOI
26. Kim BS, Pallua N, Bernhagen J, Bucala R. The macrophage migration inhibitory factor protein superfamily in obesity and wound repair. Exp Mol Med 2015;47:e161.DOIPubMedPMC
27. Fischer JP, Cleveland EC, Shang EK, Nelson JA, Serletti JM. Complications following reduction mammaplasty: a review of 3538 Cases from the 2005-2010 NSQIP data sets. Aesthet Surg J 2014;34:66-73.DOIPubMed
Bouraoui Kotti et al., Plastic and Aesthetic Research, 2018
Victor Z. Zhu et al., Plastic and Aesthetic Research, 2016