Keywords
BMI cutoff - obesity - total knee arthroplasty - knee arthroplasty - weight loss
As a result of the rising prevalence of obesity in North America, an increasing number
of morbidly obese patients with primary knee osteoarthritis will be evaluated in need
of total knee arthroplasty (TKA).[1]
[2] However, many studies have established a link between obesity and complications
from TKA.[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10] In 2013, the American Association of Hip and Knee Surgeons (AAHKS) Evidence-Based
Committee published the following statement: “The morbidly obese (BMI > 40 kg/m2) and the super obese (BMI > 50 kg/m2) have complication profiles that may outweigh the functional benefits of total joint
arthroplasty (TJA). These patients should be counseled regarding these risks prior
to any surgical intervention. It is our consensus opinion that consideration should
be given to delaying TJA in a patient with a BMI > 40 kg/m2, especially when associated with other comorbid conditions, such as poorly controlled
diabetes or malnutrition.”[11] As a result, many institutions have established a BMI cutoff, over which patients
are not eligible for TKA.[12] However, the degree to which surgeons adhere to their institution's BMI cutoffs
has not been well-defined in the literature, and it is possible that surgeons take
significant leeway.
The ostensible utility of an institutional BMI cutoff is to mitigate the increased
risk of complications associated with performing a TKA in the morbidly obese. Caring
for morbidly obese patients is costly.[13]
[14]
[15] With increasing emphasis on value-based care in the United States, these considerations
are especially important.[16] It is unclear whether an investment in clinic time and resources for evaluation
of the morbidly obese that initially do not meet a BMI cutoff will lead to patients
achieving weight loss goals to become surgical candidates. Few studies have examined
the percentage of patients over a BMI cutoff that clinicians can expect to return
for surgery once they have established care.[17]
[18]
In the current study, we aimed to identify the percentage of morbidly obese patients
(in other words, those over the institutional BMI cutoff) who present with a diagnosis
of primary knee osteoarthritis and ultimately undergo TKA at a BMI less than 40 kg/m2 within 2 years of their initial visit. As secondary goals, we aimed to elucidate
what percentage of patients underwent TKA at a BMI greater than 40 kg/m2 or went to an outside hospital to have surgery, and in patients who did not undergo
TKA, what percentage of patients reached a BMI less than 40 kg/m2 within 2 years of their initial visit. This information could help surgeons make
informed decisions about the deployment of clinic resources, aid in patient counseling,
and establish a baseline of expected outcomes for future studies looking to examine
the utility of a BMI cutoff and the fate the morbidly obese who are not offered a
TKA.
Materials and Methods
We performed an observational study in which we queried the institutional electronic
medical record (EMR) (Epic Systems Corporation, Verona, WI) to identify morbidly obese
patients 40 years of age and older with a diagnosis of knee osteoarthritis who presented
to a high-volume arthroplasty clinic for surgical evaluation. Morbid obesity was defined
as BMI greater than 40 kg/m2 calculated using the height and weight data gathered by the medical assistant at
the patient's initial clinic visit. The query was performed for patient presenting
between February 2014 (when the EMR was implemented) to March 2018 (to allow for 2
years of follow-up). We initially identified and screened 775 patients. Patients were
excluded if they presented with a diagnosis other than primary knee osteoarthritis,
were less than 40 years old, had BMI less than 40 kg/m2 on initial presentation, or had been seen outside of established date range. Our
clinic is a multi-subspecialty practice, including sports medicine surgeons and nonsurgeon
providers. We included only those patients who established care with an arthroplasty
surgeon or physician assistant (5 total surgeons and 1 physician assistant). In an
attempt to remove patients who continuously return to arthroplasty providers for injections
or medications without a desire for surgery, we also excluded patients whose first
EMR-recorded visit was listed as a “return visit” rather than a new patient encounter.
Ultimately, 487 patients were excluded and 288 patients were included ([Fig. 1]).
Fig. 1 Flow diagram showing the breakdown of patients excluded from and included in the
study and the pathways of the included patients. BMI, body mass index; EMR, electronic
medical record; TKA, total knee arthroplasty.
For these 288 patients, the EMR was used to record multiple demographic variables
including diagnosis of diabetes, current smoking status, and prior history of bariatric
surgery. In an attempt to characterize the frequency with which particular weight
loss interventions are recommended on an initial clinic visit, we recorded whether
there was documentation of patient counseling (including patients simply being told
they were above the BMI cutoff), physical therapy referral, dietitian referral, or
bariatric surgery referral. Our institution does not have an established weight loss
protocol for morbidly obese patients; it is up to each provider to make weight loss
recommendations and order interventions.
Patients' records were then examined for whether the patient ultimately underwent
TKA at our institution within 2 years of their initial clinic visit. The patient's
BMI at the time of surgery was recorded. Patients that did not undergo surgery at
our institution were contacted by telephone or their medical record was reviewed to
ascertain whether they had undergone TKA at an outside hospital within 2 years of
their initial clinic visit. Thirty-two of 288 patients (11.1%) were unable to be contacted
and were considered lost to follow-up ([Fig. 1]).
At our institution, height and weight are recorded for each clinical encounter, regardless
of specialty. Thus, for each encounter in the EMR after the initial clinic visit,
BMI was assessed. If BMI was recorded as less than 40 kg/m2 at any time point, the patient was classified as “TKA-eligible,” even if the visit
of record was not with an orthopedist or the weight loss was not sustained.
Sample characteristics were described using descriptive statistics. Frequencies and
percentages were used to describe categorical variables. Means and standard deviations
or medians and ranges were used to describe continuous variables.
Results
Patient demographics are shown in [Table 1]. The frequency of documentation of specific interventions is shown in [Table 2]. Documentation of referrals to a physical therapy, dietitian, or bariatric surgeon
was found in less than 6% of patients. Providers documented counseling patients on
weight loss strategies or weight loss goals to qualify for TKA 47.6% of the time.
Table 1
Patient demographics (n = 288)
|
Average age in years (range)
|
59 (40–84)
|
|
Average BMI in kg/m2 (range)
|
46.2 (40–65.4)
|
|
Sex
|
Number (% of total)
|
|
Male
|
83 (28.8)
|
|
Female
|
205 (71.2)
|
|
Diagnosis of diabetes?
|
|
|
Yes
|
125 (43.4)
|
|
No
|
163 (56.6)
|
|
Current smoker?
|
|
|
Yes
|
28 (9.7)
|
|
No
|
260 (90.3)
|
|
Prior history of bariatric surgery?
|
|
|
Yes
|
21 (7.3)
|
|
No
|
267 (92.7)
|
Abbreviation: BMI, body mass index.
Table 2
Frequency of interventions utilized by clinicians (n = 288)
|
Referrals placed
|
Number (% of total)
|
|
Physical therapy
|
6 (2.1)
|
|
Dietitian
|
2 (0.7)
|
|
Bariatric surgery
|
7 (2.4)
|
|
Was there counseling/discussion of BMI cutoff?
|
|
|
Yes
|
137 (47.6)
|
|
No
|
151 (52.4)
|
Overall, 32.4% of patients underwent TKA within 2 years ([Table 3]). The average BMI at initial visit in surgical patients was 43.6 kg/m2 (40–58.33) versus 46.8 kg/m2 (40–64.6) in the nonsurgical patients. For the surgical group, the average BMI decreased
to 42.3 kg/m2 (38.3–52.7) on the day of surgery. [Table 4] demonstrates the breakdown of surgical and nonsurgical patients who were able to
achieve the weight loss to meet the institutional goal of BMI less than 40 kg/m2. Of the 256 patients included in the study, 49 of 256 patients (19.1%) managed to
dip under the BMI cutoff and 12 of 256 actually underwent TKA at a BMI less than 40 kg/m2 (4.7%). Sixty-four of 256 patients (25%) underwent TKA at BMI greater than 40 kg/m2. Thirty-seven of 256 patients (14.4%) became TKA-eligible but did not undergo surgery,
while 136 of 256 patients (53.1%) neither met the institutional cutoff of BMI 40 kg/m2 nor had surgery within 2 years ([Fig. 1]). Seven patients underwent TKA at outside institutions and BMI at the time of surgery
could not ascertained. [Table 5] depicts our primary findings stratified according to BMI.
Table 3
Surgical pathways of all patients (n = 256)
|
Underwent TKA
|
Number (% of total)
|
Average BMI in kg/m2 at initial visit (range)
|
Average BMI in kg/m2 on day of surgery (range)
|
|
No
|
173 (67.6)
|
46.8 (40–64.6)
|
–
|
|
Yes
|
83 (32.4)
|
43.6 (40–58.33)
|
42.3 (38.3–52.7)
|
Abbreviations: BMI, body mass index; TKA, total knee arthroplasty.
Table 4
Clinical pathways of all patients relative to BMI cutoff (n = 256)
|
TKA performed within 2 years
|
Number (% of total)
|
|
BMI less than 40 kg/m2
|
12 (4.7)
|
|
BMI greater than 40 kg/m2
|
64 (25)
|
|
Surgery done at an outside hospital
|
7 (2.7)
|
|
TKA not performed within 2 years
|
|
|
Became TKA-eligible (BMI less than 40 kg/m2 at any time)
|
37 (14.5)
|
|
Did not become TKA-eligible (BMI never less than 40 kg/m2)
|
136 (53.1)
|
Abbreviations: BMI, body mass index; TKA, total knee arthroplasty.
Table 5
Pathways of patients based on BMI strata
|
BMI on initial presentation (kg/m2)
|
|
|
40–45
|
45–50
|
50+
|
Total
|
|
Number of patients who did not undergo TKA
|
79
|
58
|
36
|
173
|
|
Number who became TKA-eligible
|
31
|
5
|
1
|
37
|
|
Number who did not become TKA-eligible
|
48
|
53
|
35
|
136
|
|
Number of patients who underwent TKA
|
61
|
19
|
3
|
83
|
|
Average BMI at surgery (kg/m2)
|
41.4
|
44.7
|
46.6
|
–
|
Abbreviations: BMI, body mass index; TKA, total knee arthroplasty.
Discussion
The current study assessed the clinical pathways of morbidly obese patients with knee
osteoarthritis who establish care with an arthroplasty practice with an institutional
BMI cutoff of 40 kg/m2, by determining the percentage of patients who undergo surgery and the percentage
of patients that lose enough weight to be considered “eligible” according to an institutional
BMI cutoff. We also attempted to define provider tendencies in the initial management
of morbidly obese patients based on documented recommendations and interventions.
A few previous studies have examined similar patient populations. Springer et al prospectively
enrolled 289 morbidly obese patients who presented to clinic with both hip and knee
osteoarthritis. Their study differed in that they provided an intervention for all
patients: a referral to a bariatric surgeon. Only 23.2% of patients saw the bariatric
surgeon, and of those, just 20.9% had bariatric surgery. Ultimately, 20.1% of their
original cohort underwent TJA at an average BMI of 42.3 kg/m2. Of the patients who did have a TJA, 39.7% had a BMI less than 40 at the time of
surgery. The authors did not attempt to ascertain whether patients may have had surgery
at an outside hospital.[18] Shapiro et al retrospectively examined 99 morbidly obese patients who were denied
TKA. Of those, 20.2% met the target weight and underwent TKA at their institution.
Eight of 99 patients (8.1%) underwent TKA at another institution.[17] The current study reports findings similar to these historical precedents. However,
we also report the weight loss success of all patients who were reviewed, not just
those who underwent TKA. Of the 173 patients who did not undergo TKA, 37 patients
(21.4%) successfully brought their BMI to less than 40 kg/m2.
The current study is limited by its unclear generalizability to other practices. Regional
demographics and barriers to access affect the patient population that any given provider
sees in clinic. Our institution is a large academic medical center in the midst of
a relatively poor, rural area[19] in a state with one of the highest prevalence of obese patients.[20] Multiple nonsurgeon providers often serve as an initial screen to offload the clinical
burden of our surgeons. As a result, the patients that surgeons see in clinic typically
have expressed at least some level of inclination toward surgical intervention. Compared
with a practice without such screening measures, our clinical setup may therefore
overestimate the number of morbidly obese patients an arthroplasty surgeon may expect
to undergo surgery within 2 years. After all, 487 patients were excluded from analysis,
the majority for not having seen an arthroplasty surgeon or physician assistant.
The current study is also limited by its retrospective nature. While scheduling an
appointment with an arthroplasty surgeon when multiple nonoperative provider options
exist demonstrates a certain level of initiative, we cannot ensure that all patients
would have met surgical criteria for arthroplasty if not for their BMI, or if each
patient would have elected to proceed with surgery had it been offered at the initial
visit. However, the surgical discussion is often nuanced, and even a prospective study
would be subject to bias as operative criteria would inevitably vary subtly from surgeon
to surgeon, and there would be no way to identify which patients truly desire surgery
if they were precluded by their BMI.
The retrospective nature of this study also makes it difficult to draw conclusions
based on the percentages of patients who were provided with counseling or referrals
because of the variability in documentation and the use of EMR templates among providers.
Documentation of interventions or counseling may not accurately reflect the discussions
held with the patient. Even so, discussion of BMI cutoffs and weight loss counseling
would ideally approach 100%. Institutional restrictions necessitate institutional
solutions. As recent AAHKS symposium stated, “collaborative care pathways need to
be developed to provide a comprehensive optimization program for patients who need
total joint replacement.”[21] Our institution would likely benefit from an established weight-loss protocol from
morbidly obese patients who desire TKA.
While there is an abundance of literature on the general effectiveness of different
weight loss strategies, there is a relative dearth of studies examining interventions
to prepare patients for TKA. Dieting protocols and dietitian supervision have shown
promise.[22]
[23] Aquatic exercise may be an effective form of physical activity.[24] In addition to oral pain medications or steroid injections, genicular nerve ablation
has been proposed as an alternative pain control approach,[25] which could theoretically allow patients to exercise more effectively. Many patients
with a BMI over 40 kg/m2 are referred to bariatric surgery,[18]
[26] which has been shown to be effective in helping patients achieve weight loss goals
and may even obviate the need for TKA in some patients.[27] Our institution has not developed any systematic weight loss protocols for patients,
and therefore it is up to individual providers to counsel patients on effective weight
loss strategies and provide referrals for specific weight loss interventions. Adherence
to robust, evidence-based protocols has been effective in addressing other perioperative
challenges,[28] so development and implementation of weight loss protocols prior to elective TKA
may benefit morbidly obese patients.
There is very little evidence in the literature about provider adherence to institutional
BMI cutoffs. Shapiro et al offered surgery to patients who make a “concerted attempt”
at weight loss (defined as at least reaching two-thirds of the weight loss goal).[17] Our institution varies by provider, but in general, most of our surgeons do make
concessions for those patients who demonstrate motivation to lose weight, such as
achieving a significant portion of a weight loss goal or losing 10% of total body
weight, even if those patients ultimately do not reach the BMI cutoff of 40kg/m2. These habits help explain why the average BMI of the surgical patients was 42.4 kg/m2 in this study. Future research to define the strictness with which surgeons adhere
to their institution's BMI cutoffs would likely be of significant interest to the
arthroplasty community.
Interestingly, fewer patients in the current study went “doctor-shopping” than the
authors anticipated. Doctor-shopping involves visiting multiple doctors for the treatment
of the same health condition.[29] Morbidly obese patients have been shown to be up to 52% more likely to seek multiple
primary care providers during a 2-year period.[30] One might think that by withholding TKA, patients will simply find another surgeon
nearby to do the surgery. Of the 180 patients who were not offered surgery at our
institution, only 7 (3.9%) patients underwent surgery elsewhere. Shapiro et al found
that similarly, 8.1% of patients underwent TKA at an outside institution. The prevalence
of “surgeon-shopping” will depend in part upon the location of a practice and patients'
ease of access to other surgeons. Even so, surgeon-shopping may not be as pervasive
as one would expect, and the phenomenon merits further study.
The number of patients in the current study who were able to successfully lose weight,
either to undergo TKA safely or to continue to live with their knee osteoarthritis
was 19.1%. As we understand the limited efficacy of self-directed weight loss attempts
in the TJA population, delineation of evidence-based pathways to provide safe and
effective care for these patients become critical. Myriad studies have shown an increase
in complications, particularly infection, related to increasing BMI and TJA.[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10] However, it is also clear that morbidly obese patients benefit from TKA.[31]
[32] Weight loss prior to TJA has been shown to decrease length of stay.[33] While limited evidence suggests that weight loss would lower the risk of obesity-related
complications,[34]
[35] nonsurgical weight loss or undergoing bariatric surgery prior to TKA may have no
effect on risks[36]
[37] or actually increase such risks.[38]
[39]
[40] Studies currently underway may shed some more light on this conundrum.[41] Future studies will address the costs and benefits of an institutional BMI cutoff,
the utility of particular weight loss interventions, and the impact of weight loss
prior to TKA.
Conclusions
With the increasing incidence of obesity in the United States, it is important to
consider the potential clinical courses of morbidly obese patients with knee osteoarthritis.
This descriptive study shows that without strict enforcement of a BMI cutoff and a
lack of specific weight loss strategies for patients, the number is patients who reach
the sub-40 kg/m2 BMI milestone is low. This information may be used to further evaluate the utility
of an institutional BMI cutoff and investigate strategies to more effectively usher
patients to a lower BMI prior to surgery.
