To the Editor,
The study by George and colleagues  examines the risks associated with obesity among patients who underwent total joint arthroplasty (TJA). We were similarly skeptical of the proportion of patients with obesity as published by Odum and colleagues  among TKA patients being lower than the general population . Having worked with the National Inpatient Sample (NIS) database , we understand the database’s limitations, including the inability to access patient-specific data.
Despite this limitation, the NIS database is appropriate for assessing longitudinal procedural trends. However, it is not ideal for assessing specific patient-level anthropometric data, such as BMI. The inaccuracy in Odum and colleagues  is the result of the way obesity is tracked in the database, which relies on obesity as a diagnosis code upon discharge. As this code is not directly reimbursable, there is minimal incentive to code it accurately.
The complex method employed by George and colleagues, while interesting, should not be construed as an adequate means for evaluating trends in obesity among patients who underwent TJA. The relative risk of obesity among this patient population may change over time, and the cited relative risks may not apply to alternative populations, such as NIS database patients.
The ideal means for addressing obesity trends among this group would be through a database that tracks BMI, such as the American College of Surgeons National Surgical Quality Improvement Project (NSQIP) database. We evaluated the change in BMI and obesity proportion among surgical patients from 2008 to 2016 using NSQIP. Mean BMI was compared between TJA patients and the overall NSQIP population by two-sample t-test. Patients undergoing TJA procedure for treatment of fracture were excluded. Mean difference in BMI from 2008 to 2016 was calculated, as was the odds ratio (OR) for proportion of obesity (BMI ≥ 30) among each TJA population relative the overall NSQIP population. All statistics were performed using Stata IC 14.2 (StataCorp LLC, College Station, TX, USA).
In 2008, the mean BMI for TKA patients was significantly higher than the overall NSQIP population, but the mean BMI for total hip arthroplasty (THA) patients did not differ significantly. By 2016, the mean BMI was higher for both THA and TKA patients compared with the overall NSQIP population. Mean BMI increased for the overall NSQIP population from 2008 to 2016 but did not meaningfully differ among THA and TKA patients (Table 1).
It is evident that the rate of obesity among surgical patients has been increasing, however, we did not note a considerable change from 2008 to 2016 among TJA patients, as estimated by George and colleagues . We caution against applying static relative risks from one study population to another. We advocate for use of the best available database for large epidemiologic studies rather than extrapolation using complex statistical methods, especially when using the results to govern healthcare policy decisions.
We thank the Agency for Healthcare Research and Quality for providing access to the Nationwide and National Inpatient Sample databases from 2000-2014, supplied by the Healthcare Cost and Utilization Project . We also thank the American College of Surgeons for access to the National Surgical Quality Improvement Project databases from 2006-2016.
1. George J, Klika AK, Navale SM, Newman JM, Barsoum WK, Higuera CA. Obesity epidemic: Is its impact on total joint arthroplasty underestimated? An analysis of national trends. Clin Orthop Relat Res. 2017;7:1798–1806.
2. HCUP Nationwide and National Inpatient Sample (NIS). Healthcare Cost and Utilization Project (HCUP). 2000-2014.Agency for Healthcare Research and Quality, Rockville, MD. Available at: http://www.hcup-us.ahrq.gov/nisoverview.jsp
. Accessed January 19, 2018.
3. Odum SM, Springer BD, Dennos AC, Fehring TK. National obesity trends in total knee arthroplasty. J Arthroplasty. 2013;8 Suppl:148–151.