Underweight Patients Are the Greatest Risk Body Mass Index Group for 30-Day Perioperative Adverse Events After Total Shoulder Arthroplasty

Taylor D. Ottesen, BS; Walter R. Hsiang, BS; Rohil Malpani, BS; Allen D. Nicholson, MD; Arya G. Varthi, MD; Lee E. Rubin, MD; Jonathan N. Grauer, MD


J Am Acad Orthop Surg. 2021;29(3):e132-e142. 

In This Article


With the increasing popularity of TSA because of its proven efficacy and an ever aging cohort,[1] it is important to identify the key risk factors to better conduct patient counseling and optimize quality improvement measures. BMI has previously been identified as a risk factor that is especially pertinent with the rising epidemic of obesity worldwide. Recent studies have challenged this assertion with mixed data, and no study to date has specifically investigated the underweight cohort as a at-risk group.[6–10,16–19] The current study thus investigated the effect of the full spectrum of BMI on 30-day complications after elective TSA using the large, national NSQIP database.

The presented analyses of 15,725 patients undergoing TSA found BMI <18.5 kg/m2 to be significantly associated with higher complications on both univariate and multivariate analyses. On multivariate analysis, underweight patients were more likely to experience AAE (OR = 2.22, P = 0.034), serious adverse event (OR = 3.18, P = 0.004), or have postoperative infections (OR = 2.77, P = 0.012) in comparison to normal-weight patients. By contrast, other BMI categories were not shown to have significant differences in 30-day adverse events relative to normal-weight patients (up to and including super morbidly obese patients).

As noted earlier, previous studies evaluating the association of elevated BMI with adverse perioperative complications after TSA have shown mixed results.[6,10,16,17,19] One retrospective chart review found BMI >35 kg/m2 and <25 kg/m2 were both predictive of higher medical complication rates after reverse TSA.[32] A national study of inpatient patients receiving TSA noted that obese patients had longer hospital stays, increased postoperative respiratory complications, increased costs, and a trend toward greater mortality.[10] Another single institutional study found super obese patients to have higher surgical site infection,[18] and a separate institutional joint registry of 4,567 patients found BMI to be correlated with increased risk of reoperation, superficial infection, and longer-term outcomes, such as revision, to be correlated with increased BMI.[8]

By contrast, a different shoulder registry study of 4,630 TSA patients also investigated the effect of BMI on long-term outcomes finding that arthroplasty in obese patients was not significantly associated with higher risk of aseptic revision, increased 1-year mortality, or 90-day readmission, although they did find a marginal association of increased BMI with increased deep surgical site infections in patients undergoing reverse TSA.[19] Another large national study found no difference in any 30-day or perioperative complication between normal and elevated BMI categories.[16] A single-site study looking at complications after TSA within 90 days found that BMI was not a predictor and that the Charlson comorbidity index was a more important factor.[6] In addition to studies showing no difference in BMI, one study even concluded that BMI >30 kg/m2 was associated with lower risk of 90-day postoperative mortality after shoulder arthroplasty.[33]

Similar to some of these ladder studies, we found that elevated BMI is not associated with increased 30-day complication rates. We attribute the variation in conclusions to inconsistent and nongranular definitions of BMI categories. Each of the abovementioned studies used different cutoff values for obesity, and many failed to stratify higher BMI categories, for example, opting to designate any patient with BMI >35 into an all-inclusive obese category.[16] In an effort to create consistency, we used the internationally recognized obesity categories as set forth by the World Health Organization.

In addition to these inconsistent BMI groups, none of the abovementioned studies specifically investigated the subcategory of patients with BMI <18.5 kg/m2, instead only focusing at the effects of elevated BMI on complications after shoulder arthroplasty. These studies examining the effect of BMI on shoulder arthroplasty events often categorize their lowest BMI division as patients <25 kg/m2 and have not looked specifically at very low BMI patients <18.5 kg/m2 or completely excluded them from analysis.[6,16,19] Although patients with a BMI <18.5 kg/m2 comprised a small cohort of our study, they had significantly more adverse events which suggests that these patients should not be lumped into the "normal" BMI category of <25 kg/m2 and should be separately risk stratified by physicians.

Possible explanations for the association of low BMI with elevated 30-day complications after shoulder arthroplasty are not well elucidated. Factors such as cancer, alcoholism, and nutritional deficiencies could potentially decrease BMI and negatively influence postsurgical recovery.[34,35] A study of cardiac surgery patients found that the thinnest patients had more risk factors for postoperative morbidity.[36] There is also a possibility that patients with low BMI may have less soft-tissue coverage of their implants and more complications with wound healing. Further understanding of the low-BMI patient cohort undergoing shoulder arthroplasty may help to illuminate the causes of their increased surgical complications and allow preoperative adjustments to decrease surgical risk profiles for this at-risk patient cohort.

Limitations of this study include its design as a large database retrospective cohort study. Although retrospective cohort studies can establish ORs and correlations, they do not show causation, and thus, low BMI itself may not be directly responsible for the elevated complications after shoulder arthroplasty in this patient subset. In addition, underweight patients made up a minority of the total cohort compared with overweight populations that are becoming increasingly more common, thus limiting statistical power. Furthermore, we looked at differences in 30-day complications and cannot comment on more long-term outcome differences. Finally, it is possible that our low-BMI subset of patients may have had a significant comorbidity responsible for their low BMI and the resulting complications. However, we controlled for these comorbidities through ASA scores and we specifically chose only elective cases, where major causes of morbidity would likely have been identified preoperatively and prevented the elective procedure from going forward.

Overall, the current study found that patients with BMI <18.5 kg/m2 have significantly higher risks of complications within 30 days of elective shoulder arthroplasty surgery. By contrast, obese patients, including the morbidly obese, were not found to have a higher incidence of complications. These findings provide surgeons with important preoperative planning information to help reduce risk of postoperative complications. Physicians should give underweight patients additional consideration because they already do for the other end of the BMI spectrum. Patients that have a very low BMI (<18.5 kg/m2) may benefit from additional counseling such as nutrition consultation or investigation into the reasons of their low BMI before surgical intervention.