Introduction

The high and increasing prevalence of obesity in the developed world represents a growing health care problem affecting the entire health care system. Among orthopaedic patients, it is the most important reason for premature osteoarthritis leading to the need for joint arthroplasty procedures [1,2,3].

Aside from the fact that a number of studies identified obesity to be an independent risk factor for various peri-operative complications and adverse outcomes, an increased BMI most certainly represents technical and logistic challenges to caretakers, including physicians, nurses, and other operating-room staff [4,5,6,7]. Reported difficulties range from challenges to secure vascular access, achieve adequate surgical conditions, and provide staff and equipment for transport and positioning, as well as account for difficult anaesthetic management, all of which require additional time and resources [3]. Obesity is associated with more implant failure by various studies [7,8,9,10]. However, optimization of obesity such as gastric bypass surgery has not been proven to lower surgical complications after arthroplasty [11].

To date, however, the differential impact of obesity on peri-operative resource utilization remains poorly defined. Information on this aspect of care is of importance, because of the large and increasing number of obese patients undergoing surgery and because hospitals of the need to account and quantify physical and time resources in order to prepare adequately for procedures involving high-BMI individuals.

Therefore, we undertook this study utilizing data collected for the National Surgical Quality Improvement Project to determine if patients with various BMI ranges utilized joint arthroplasty-related operating-room time and resources differentially. We hypothesized (1) that care of patients with increased BMI would be associated with prolonged surgical and anaesthetic times, and (2) that a positive correlation between increasing levels of BMI and an increase of resource, such as readmissions, would exist.

Materials and methods

This study was exempted by the institutional review board (Hospital for Special Surgery, no. 2017-0716, New York, NY 10021, USA).

Study sample

We acquired the data from the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) from 2006 to 2015 (http://site.acsnsqip.org). NSQIP prospectively collects data for over 140 variables under a standardized protocol. The data includes demographic information, comorbidities, pre-operative laboratory results, intra-operative variables, and 30-day post-operative complications. To define our study cohort, we only included patients with the principal Current Procedural Terminology (CPT) code for primary total knee arthroplasty (TKA) (CPT 27447) or primary total hip arthroplasty (THA) (CPT 27130). There were 167,201 and 104,732 subjects respectively. We first excluded patients categorized as “emergency” (255 and 791 subjects), or American Society of Anesthesiologist (ASA) Class 4 and Class 5 (2671 and 1992 subjects). We next eliminated patients with missing information on BMI or BMI < 18.5 kg/m2 (882 and 1672 subjects). Then we removed patients receiving bilateral arthroplasty procedures (4571 and 585 subjects). Last, we excluded patients with BMI > =30.0 kg/m2 and BMI < 40.0 kg/m2 to minimize cutoff bias in our study (75,377 and 37,906 subjects). The final study cohort included 83,445 and 61,786 subjects for TKA and THA, respectively.

Study variables

We focused on seven available variables in the database related to resource utilization during the hospitalization, including anaesthesia start to surgery start, end of surgery to end of anaesthesia, total time in the operating room, total length of hospital stays, percent of patients remaining in hospital after 30 days, and readmissions within 30 days.

Statistical analysis

Data analysis was executed using STATA 14.2 statistical software (StataCorp LP, College Station, TX). To examine the impact of BMI on resource utilization, patients were separated into three groups based on BMI (18.5–30 kg/m2, 40.0–44.9 kg/m2, and > =45.0 kg/m2). We intentionally excluded patients with BMI between 30.0 and 40.0 kg/m2 a priori in order to minimize bias introduced by the arbitrary BMI cutoff selection. We conducted multivariable logistic regression analysis while adjusting age, gender, race, and ASA classification. All analyses without specification treated the BMI 18.5–30.0 kg/m2 group as the control group. The adjusted odds ratio (OR) and 95% confidence intervals (CI) were reported. The statistical significance was adjusted using Bonferroni correction to two-sided p < 0.0056 (0.05/9 comparisons) to account for the multiple models examined in this study.

Results

For our samples, we identified N = patients that underwent primary TKA and THA. These samples were divided into three groups based on BMI (BMI ≥ 18.5 and < 30 kg/m2, BMI ≥ 40 and < 45 kg/m2, and BMI ≥ 45 kg/m2). Demographic information is shown in Table 1 (TKA) and Table 2 (THA).

Table 1 Demographic information and comorbidities of all TKA patients
Full size table
Table 2 Demographic Information and Comorbidities of All THA Patients
Full size table

The analyses showed that for both THA and TKA patients, the three BMI groups were associated with a statistically significant difference average age. Patients with higher BMI tended to be younger (p < 0.0001). Additionally, there was a significant difference in gender as more female subjects were represented in the high-BMI groups (p < 0.001). Differences in race distribution among the three groups with more African American in the high BMI groups were recorded. There was a higher prevalence of diabetes and chronic obstructive pulmonary disease as well as higher rates of ASA II status in the higher BMI categories (p < 0.001) (Tables 1 and 2).

Table 3 (TKA) and Table 4 (THA) summarized resource utilization and readmission rates across the three BMI groups. TKA and THA patients with higher BMI required significantly longer total duration of anaesthesia (p < 0.0001), total operation time (p < 0.0001), total time in the operating room (p < 0.0001), total length of hospital stay (p < 0.0001), and increased time from day of surgery to discharge (p < 0.0001). However, incidences of remaining in hospital after 30 days of surgery did not show statistical differences in neither TKA nor THA patients. Although a trend towards longer was noticed for anaesthesia induction and anaesthesia emergence time among TKA patients, this finding reached statistical difference only among individuals undergoing THA.

Table 3 Summary statistics of time utilization and incidence of events among TKA patients
Full size table
Table 4 Summary statistics of time utilization and incidence of events among THA patients
Full size table

Logistic regression analysis was applied to further test the independent influence of various BMI levels on these seven outcome variables. The most significant finding was the consistently higher odds of readmissions within 30 days with both the three-group regression analysis and pairwise regression analysis in both TKA and THA patients (Tables 5 and 6). Higher BMI patients were associated with higher odds of total length of hospital stay and time from end of surgery to discharge. However, the analysis could not show consistent statistical significance.

Table 5 Multivariable regression analysis of BMI on time utilization and incidence of events among TKA patients
Full size table
Table 6 Multivariable regression analysis of BMI on time utilization and incidence of events among THA patients
Full size table

Discussion

Our results suggest that obese patients require more peri-operative resources and carry higher odds of readmissions within 30 days in both TKA and THA patients. Recognition of increased resource requirements should be accompanied by management strategy adjustments to optimize resource allocation.

Peri-operative risk factor identification, risk stratification, and peri-operative complication optimization have drawn broad attention in recent years. Increased BMI alone has been a major risk factor of post-operative complications in TKA and THA, including mortality [4, 7, 9, 12,13,14]. Despite the increased risks for adverse events and failure rate, obese patients might still benefit from joint arthroplasty. One potential hurdle for obese patients to receive quality treatment is the punitive merit-based reimbursement system which disincentives operations on higher risk individuals, including the obese [15]. In this context, resource allocation and cost management are thus becoming more important in making management decisions especially for elective surgeries. Good patient selection, optimization, and appropriate resource allocation might be helpful in achieving better outcomes while minimizing potential risks among such higher risk patient population.

Several studies have compared length of stay (LOS) and hospital cost among obese and non-obese patients in THA and TKA surgery. D’Apuzzo et al. reported 3.1% higher total hospital cost and a small increase in LOS (3.6 days vs 3.5 days) when comparing morbidly obese patients and non-obese TKA patients [16]. Similarly, Kim reported 7–9% higher costs among morbidly obese receiving THA and TKA surgery [17]. However, Batsis et al. did not observe a correlation between BMI and hospital resource use among 5521 TKA patients [18] and 5642 THA patients [19]. Detailed analysis indicated potential correlation between increasing BMI and operating room cost in addition to anaesthesia cost [19]. However, no further details were provided.

Our study adds to the available body of information on peri-operative resource utilization of TKA and THA among the obese population. The results suggest the increased need for time to perform intra-operative tasks in obese patients. The only exception found was for induction time and emerge time among THA patients, which were only minimally influenced by BMI. Our study also indicated longer length of stay with increase of BMI. The incidences of readmission within 30 days increased with the increase of BMI levels in both TKA and THA patients. However, our analysis did not indicate that increased BMI was associated with incidence of excessive longer duration of hospital stay beyond 30 days.

Our study has limitations. First, this was a retrospective observational study with administrative data. Second, this study is limited by the number of variables listed in the NSQIP database. Third, NSQIP prohibits identifying hospital and physician, while studying the surgical volume and physician practice pattern might be informative.

In conclusion, our study indicated that the care of TKA and THA patients with higher BMI is associated with significantly more intra-operative time, longer length of hospital stays, and higher incidence of readmissions within 30 days. Special considerations need to be given and necessary resources allocated, when making the decision to operate on obese patients.