Thirteen studies fulfilled our inclusion criteria and were included in subsequent analysis. This produced a pooled dataset of 796 patients undergoing BCS following bariatric surgery. Characteristics of included studies are in shown in Table 1. Eight of these were prospective observational studies, four were cross-sectional studies and one was a retrospective cohort study. BCS procedures performed included dog-ear correction, abdominoplasty, panniculectomy, dermolipectomy, liposuction, brachioplasty, mammoplasty, breast reduction and thigh lift. The scores used to assess QOL in each study are outlined in Table 1. The length of follow-up ranged from 2 to 42 months.
BMI After Initial Bariatric Surgery
Five studies reported on the change in BMI following bariatric surgery only. Pooled analysis demonstrated a weighted mean decrease in BMI of 14 points (− 14.816, 95% CI [− 17.0, − 12.6]), with moderately high heterogeneity (I2 = 77.8%).
BMI After Body Contouring Surgery
Three studies reported on the change in BMI following BCS. Pooled analyses revealed a significant weighted mean decrease in BMI of 2 points (− 1.99, 95% CI [− 2.99, −0.98]), with no heterogeneity present (Fig. 2a).
Seven studies reported on changes in physical functioning following BCS. Pooled analysis demonstrated a significant improvement in physical functioning by 28.5% (p = 0.004, 95% CI [8.9, 48.1]), with a low degree of heterogeneity (I2 = 32%) (Fig. 2b).
Six studies reported on changes in psychological wellbeing following BCS. Pooled analysis demonstrated a significant improvement in psychological wellbeing by 45.7% (p = 0.029, 95% CI [4.7, 86.7]); however, a high degree of heterogeneity was present (I2 = 87%).
Eight studies reported on improvements in social functioning after BCS. Pooled analysis revealed a significant improvement in social functioning by 24% (p = 0.001, 95% CI [10.0, 38.0]), with a low degree of heterogeneity (I2 = 29%).
Eight studies assessed for improvements in perception of body image following BCS. Pooled analysis demonstrated an improvement in body image by 55%; however, this result was not significant (p = 0.12, 95% CI [− 14.3, 125.6]).
Five studies observed changes in sexual functioning after BCS. Pooled analysis demonstrated an improvement in sexual functioning by 49.7%; however, this result was not significant (p = 0.238, 95% CI [− 32.8, 132.1]), and heterogeneity was high (I2 = 97%).
Three studies assessed for improvements in pain following BCS. Pooled analysis demonstrated a non-significant improvement in pain by 18.5%, with high heterogeneity (p = 0.4, 95%CI [− 26.7, 63.6], I2 = 82.9%).
Three studies reported on changes in self-esteem following BCS. Analysis revealed a non-significant improvement in self-esteem by 17.6% (p = 0.4, 95% CI [− 27.8, 63]).
Global QOL Score
Five studies reported a change in the overall QOL using a global QOL score that included a range of physical and psychosocial outcomes. One of these studies used the SF-36 score, two studies used the Health-Related Quality of Life (HRQOL) score, one study used the Impact of Weight on Quality of Life Questionnaire (IWQOL) score and one study used the World Health Organization Quality of Life (WHOQOL) score. Pooled analysis demonstrated an improvement in overall QOL by 14.2% (p = 0.083, 95% CI [− 1.9, 30.2]).
All included studies were assessed for their methodological quality and risk of bias using a modified Newcastle-Ottawa Scale. Five studies were of moderate-high quality. Of these, only one study was scored as high quality (≥ 7) . Eight studies were scored as low quality. All studies met the criteria for ascertainment of treatment and clearly defining outcomes of interest. The majority of studies also met the criteria for adequately reporting follow-up procedures. However, many studies scored poorly on comparability. This may be attributable to the lack of randomisation, which increases vulnerability to selection bias. The methodological quality of included studies is shown in Table 2.
The Cochrane ROBINS-I tool for non-randomised studies was used to assess the risk of bias in all included articles (Table 3). With the exception of two studies, all included articles had a moderate overall risk of bias. Most studies scored poorly on measurement of outcomes. This was primarily due to a lack of patient and assessor blinding, a common caveat in surgical trials . However, measures to overcome this source of bias were unaddressed or unreported by most studies. A degree of performance bias was also present in most studies since peri-operative outcomes are particularly vulnerable to this . Another domain of concern was the risk of confounding. This may be attributable to QOL being our outcome of interest. QOL is multifactorial and subjective in nature; therefore, outcomes associated with QOL are influenced by a multitude of other factors, which were unable to be controlled for. Funnel plot assessment was used to assess the degree of publication bias in included studies. Statistical analysis using Egger’s test did not reveal any significant small-study effects.