Page 276 - Read Online
P. 276
Table 8: Complications and obesity status - matched analysis
Obese Non-obese P-value
n % n %
Overall 1,464 1,464
Any complication 91 6.2 72 4.9 0.1456
Surgical complication 22 1.5 24 1.6 0.8828
Wound complication 68 4.6 45 3.1 0.0334
Medical complication 9 0.6 5 0.3 0.4227
Return to operating room 22 1.5 24 1.6 0.8828
Superficial SSI 55 3.8 36 2.5 0.0536
Deep SSI 8 0.5 5 0.3 0.5465
Organ/space SSI 3 0.2 0 0.0 0.2482
Wound dehiscence 4 0.3 4 0.3 1.0000
Venous thromboembolism 2 0.1 3 0.2 1.0000
Unplanned reintubation 0 0.0 0 0.0 NR
Urinary tract infection 1 0.2 2 0.1 1.0000
Other bleeding 3 0.2 0 0.0 0.2482
Hospital length of stay, median and
range 0 0-32 0 0-234 < 0.0001
A total of 2,928 patients (1,464 per group) were matched using propensity scores. The unmatched patients were discarded from the analysis.
McNemar’s test and the Wilcoxon signed rank test test were used to compare the two matched groups. The rate of wound complication (P = 0.0334)
and the distribution of length of stay (P < 0.0001) was found to differ between the matched groups. SSI: surgical site infection; NR: not reported
the number of events are low and there are multiple Multivariate analysis among the non-obese, overweight,
confounders. [22] and three classes of obesity showed statistically significant
differences in demographics, comorbidities, and
Many authors have tried to definitively determine complication rates [Tables 4-6]. In our unmatched analysis
the correlation between obesity and adverse events [Table 6], overall complications, wound complications,
after surgery. Although many studies consistently superficial SSI, and wound dehiscence were significantly
demonstrate the deleterious effect of obesity, nearly increased in the obese population compared to the non-
all analyses are confounded by the effects of associated obese cohort after multivariable analysis controlling for
medical conditions on outcomes. One such study did not significantly different variables between obese and non-
find a statistical difference in obese versus non-obese obese cohorts. Comorbidities may confound the isolated
patients in relation to complication and hospital length risk of obesity on complication rates. The distinguishing
of stay. Another did not find significant differences feature of our study was matching obese patients
[23]
in complications attributable to age, BMI, size of to non-obese patients with similar preoperative and
resection, smoking status, comorbidities, or surgical operative variables, thus eliminating the confounding
technique, even in the morbidly obese. Other studies effect of associated comorbidities on outcomes.
[16]
similarly found no statistically significant difference in While multivariable analysis attempts to control for
comorbidities via advanced statistical techniques, 1:1
complication rates among the obese. [14,15,17]
matching is a dramatically more powerful technique that
matches each study patient with a near-identical “control”
However, contradictory findings exist in the literature as [24]
well, supporting obesity as a risk factor. [6,9-13] Chun et al. patient, in spite of detractors of this technique. After
[13]
analysis of matched cohorts, only wound complications
identified a threshold of BMI 35.6 at which postoperative were increased in the obese population [Table 8]. On
complications were increased two-fold, the most further analysis, the difference was mainly attributed to
common complication being infection. The pioneering a risk of increased surgical site infection in the obese
study using NSQIP data to analyze BMI and breast cohort. Of note, length of hospital stay was found to be
[6]
reduction complications by Nelson et al. included significantly increased in the normal-weight cohort. On
4,545 patients between 2005 and 2011. This study used close examination, this was due to a statistical aberrancy
logistic regression to account for demographics and (in that the range of values for length of stay for non-
comorbidities. They found an increased rate in overall obese patients was greater than for obese patients).
complications, wound complications in all obesity classes,
and major surgical complications in class III obesity. In previous studies, dissatisfied patients had frequently
266 Plast Aesthet Res || Volume 3 || July 28, 2016