- Email: [email protected]
The Influence of Modifiable, Postoperative Patient Variables on the Length of Stay After Total Hip Arthroplasty
Accepted Manuscript The Influence of Modifiable, Post-operative Patient variables on Length of Stay after Total Hip Arthroplasty Kevin X. Farley, BS, Albert T. Anastasio, BA, Ajay Premkumar, MD, Scott D. Boden, MD, Michael B. Gottschalk, MD, Thomas L. Bradbury, MD PII:
S0883-5403(19)30005-1
DOI:
https://doi.org/10.1016/j.arth.2018.12.041
Reference:
YARTH 56978
To appear in:
The Journal of Arthroplasty
Received Date: 7 December 2018 Revised Date:
25 December 2018
Accepted Date: 31 December 2018
Please cite this article as: Farley KX, Anastasio AT, Premkumar A, Boden SD, Gottschalk MB, Bradbury TL, The Influence of Modifiable, Post-operative Patient variables on Length of Stay after Total Hip Arthroplasty, The Journal of Arthroplasty (2019), doi: https://doi.org/10.1016/j.arth.2018.12.041. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Title Page
1
Title: The Influence of Modifiable, Post-operative Patient variables on Length of Stay after
3
Total Hip Arthroplasty
4
Authors:
5
Kevin X. Farley, BS1 ([email protected])
6
Albert T. Anastasio, BA1 ([email protected])
7
Ajay Premkumar, MD2 ([email protected])
8
Scott D. Boden, MD1 ([email protected])
9
Michael B. Gottschalk, MD1 ([email protected])
M AN U
SC
RI PT
2
Thomas L. Bradbury, MD1 ([email protected])
11
Affiliation:
12
1
Department of Orthopaedic Surgery, Emory University School of Medicine, Atlanta, GA, USA
13
2
Department of Orthopaedic Surgery, Hospital for Special Surgery, New York, NY, USA
14
Address:
15
1
1648 Pierce Drive NE, Atlanta, GA 30307
16
2
535 E 70th St, New York, NY 10021
17
Corresponding author:
18
Michael B. Gottschalk, MD
19
1648 Pierce Dr. NE
20
Atlanta, GA 30307
21
Phone: (404) 251-1566
22
Email: [email protected]
AC C
EP
TE D
10
ACCEPTED MANUSCRIPT
Disclosures: Each author certifies that he or she has no commercial associations (eg,
24
consultancies, stock ownership, equity interest, patent/licensing arrangements, etc) that might
25
pose a conflict of interest in connection with the submitted article.
26
Word Count: 3,139 (including in-text references to tables and figures)
AC C
EP
TE D
M AN U
SC
RI PT
23
ACCEPTED MANUSCRIPT
1
Title: The Influence of Modifiable, Post-operative Patient variables on Length of Stay after
2
Total Hip Arthroplasty
3 ABSTRACT
5
Background
6
Many studies have examined strategies to reduce length of stay (LOS) after total hip arthroplasty
7
(THA), but few have focused on modifiable patient-specific information in the acute post-
8
operative period. This study investigates the determinants of LOS after THA, with a focus on
9
potentially modifiable factors.
M AN U
SC
RI PT
4
Methods
11
1,278 patients undergoing elective THA from 2012 to 2014 were extracted from our institutional
12
data warehouse at our academic orthopaedic specialty hospital. Data were collected on patient
13
demographics, comorbidities, inpatient opioid use, hypotensive events, and abnormalities in
14
laboratory values, all occurring on post-operative day (POD) 0 or 1. The main outcome was
15
hospital LOS. Multivariate regression analysis was performed to identify independent risk
16
factors for LOS over 3 days.
17
Results
18
The average age of patients undergoing primary total hip arthroplasty in our cohort was 62.3 (SD
19
10.7) years, and 52.7% were women. 6.3% (81/1278) of patients had a LOS more than 3 days.
20
Multivariate regression analysis demonstrated several statistically significant non-modifiable and
21
modifiable risk factors that influence LOS after THA. Non-modifiable risk factors included non-
22
white race (Odds Ratio [OR], 1.497), single marital status (OR, 1.724), increasing age (OR,
23
1.330), and increasing Charlson Comorbidity Index (OR, 1.411). Potentially modifiable risk
AC C
EP
TE D
10
1
ACCEPTED MANUSCRIPT
factors included every 10mg oral morphine equivalent consumption (1.069), every 5 post-
25
operative hypotensive events (OR, 1.232), low hemoglobin (OR, 3.265), high glucose levels
26
(OR, 1.887), and a high creatinine (OR, 2.874).
27
Conclusion
28
This study identifies potentially modifiable factors that are associated with increased LOS after
29
THA, including post-operative opioid use and hypotensive events. Efforts to control narcotic use
30
and initiatives aimed to reduce early postoperative hypotension could aid in reducing LOS.
31
Furthermore, attempts should be made to correct post-operative anemia, high glucose levels, and
32
a high creatinine when possible.
33
Key Words: Total Hip Arthroplasty; Opioid; Hypotension; Modifiable Risk Factors; Length of
34
Stay; Anemia
AC C
EP
TE D
M AN U
SC
RI PT
24
2
ACCEPTED MANUSCRIPT
35 36
INTRODUCTION Total hip arthroplasty (THA) is an effective and commonly performed procedure. Its relevance continues to increase in light of the aging American population. A recent study cited
38
the cost of a single primary THA at $22,076 [1]. Given high volume of these procedures, THA
39
represents a substantial cost to the healthcare system. One of the major drivers of the cost
40
associated with THA is the hospital length of stay (LOS) after the procedure [2].
Studies examining large cohorts of patients have estimated the LOS after total hip
SC
41
RI PT
37
arthroplasty to be roughly 2.97 days, down from 4.06 days in 2002 [2]. As the financial
43
landscape of healthcare is set to undergo significant changes, efforts aimed at decreasing LOS
44
after THA is one potential target to help contain and potentially decrease healthcare costs
45
associated with these procedures [3-6]. At the same time, efforts to reduce LOS that increase the
46
risk of readmission or post-operative complication could be counterproductive. Multiple authors have analyzed factors associated with increased length of stay in the
TE D
47
M AN U
42
hospital after TJA [6-11]. Additionally, several studies have been published looking at predictors
49
of LOS specifically for THA [12-23]. One such study found that illness classification indices like
50
the American Society for Anesthesiologists (ASA) physical classification system and the
51
Severity of Illness scoring system were somewhat predictive of LOS after THA, but that other
52
factors such as insurance status and race remained stronger predictors [20]. Another study found
53
that lower ASA, age, and male gender were correlated with decreased LOS [12]. While certain
54
patient characteristics, such as age, body mass index (BMI), various comorbidities, race, income,
55
and insurance status have been correlated with an increased LOS, less attention has been focused
56
on how potentially modifiable risk factors, such as post-operative opioid use, aberrant vital signs,
57
or abnormal post-operative laboratory values may affect LOS.
AC C
EP
48
3
ACCEPTED MANUSCRIPT
In particular, postoperative hypotension has been linked to increased LOS in other
59
surgical fields [24, 25]. While there are published clinical protocols that stress the importance of
60
fluid management to minimize risk of postoperative hypotension after THA, postoperative
61
hypotension and symptomatic post-operative orthostasis remains underreported in the
62
orthopaedic literature [26, 27].
RI PT
58
In this study, we analyzed inpatient hospital records for individuals undergoing primary
64
total hip arthroplasty at a single institution. We sought to identify independent modifiable risk
65
factors for delayed discharge that have been previously underrepresented in the literature. Our
66
primary outcome measure was postoperative hypotensive events. We hypothesized that, similar
67
to other surgical specialties, the number of hypotensive events on post-operative day 0 and 1
68
after THA is correlated to patient LOS over 3 days. Secondary outcomes measures included
69
post-operative opioid use and abnormal post-operative laboratory values. We also examined
70
patient characteristics such as demographic information, comorbidities, and insurance status, to
71
confirm if an independent link between these variables and LOS exists.
AC C
EP
TE D
M AN U
SC
63
4
ACCEPTED MANUSCRIPT
72 73
PATIENTS AND METHODS After Institutional Review Board approval (Approval Date 8/28/2015, IRB 76391), we identified all patients undergoing primary THA for degenerative osteoarthritis of the hip between
75
June 2012 and August 2014, by one of four hip arthroplasty specialists at our institution. Each of
76
the four surgeons is fellowship-trained and works in the same orthopaedic specialty hospital.
77
None of the 1,278 patients identified by the above criteria died during their post-operative
78
hospital stay, and none were subsequently excluded from our analysis.
SC
79
RI PT
74
The primary outcome measure of interest was hospital LOS, defined categorically as days. Each night spent in the hospital after surgery was considered an increase in LOS of one
81
day. For example, a patient discharged on the day following surgery spent one night in the
82
hospital and thus had a LOS of 1. Due to lack of available, reliably accurate data on the exact
83
time of discharge during each day, efforts were not made to distinguish if a patient was
84
discharged in the morning or the evening on a particular postoperative day. All patients were
85
admitted to the hospital the day of surgery.
TE D
M AN U
80
Our model examined predictors of LOS including both non-modifiable and modifiable
87
patient variables. Non-modifiable variables included patient demographics and comorbidities.
88
Modifiable variables included postoperative hypotension, inpatient post-operative opioid
89
medication use, and abnormal post-operative laboratory values. To assess comorbidities, the
90
Charlson Comorbidity Index (CCI) was computed for each patient using International
91
Classification of Diseases (ICD) categories [28]. Opioids were converted to oral morphine
92
equivalents (OME) for comparison and were recorded as the number of OMEs consumed on the
93
day of surgery (POD 0) and post-operative day 1 (POD 1). The number of postoperative
94
hypotensive events, defined as a systolic blood pressure less than 90 mmHg or a diastolic blood
AC C
EP
86
5
ACCEPTED MANUSCRIPT
95
pressure less than 60 mmHg for any single reading, was recorded as a continuous variable for
96
each postoperative day [29, 30].
97
Given the lack of universal consensus on laboratory threshold values, we determined abnormal values for laboratory results as the default threshold values embedded as clinical
99
support tools in our institution’s electronic medical record system (Cerner Systems). The specific
RI PT
98
cutoffs used for each laboratory value can be seen in Table 1. The presence of either an
101
abnormal high or abnormal low value was determined for each post-operative day by
102
retrospective review of all lab results for each patient during their hospitalization. Abnormal
103
values for specific lab results were thus coded as categorical variables, either ‘abnormal high’ or
104
‘abnormal low’, for each post-operative day.
M AN U
105
SC
100
In an effort to preserve the fidelity of our model, we attempted to remove very rare abnormal lab values from our analysis. To accomplish this, if an abnormal value appeared in at
107
least 5% of our sample, which we deemed a conservative threshold, that specific laboratory
108
result was included in our analysis. The presence of low calcium, high creatinine, high glucose,
109
low Hemoglobin, and low Sodium values were the only laboratory results to occur at a frequency
110
above this threshold, and thus were included in the model. A baseline characterization of
111
laboratory abnormalities and demographic characteristics for all patients in our sample can be
112
seen in Table 2.
EP
AC C
113
TE D
106
Given that 93.7% (1197/1278) of our cohort had a LOS ≤3 days and that the difference
114
between a LOS of 2 or 3 days may be dependent on what time of day surgery was performed, we
115
chose to evaluate risk factors for prolonged LOS past 3 days. This determination was made to
116
isolate a group of patients clearly above the average who would definitively benefit from more
117
thorough post-operative care pathways. See histogram (Figure 1) for LOS distribution.
6
ACCEPTED MANUSCRIPT
118
Statistical analysis was performed using the R:A language and environment for statistical computing (R Foundation for Statistical Computing, http://www.R-project.org). Student t tests
120
were performed for continuous data and χ2 or Fisher exact tests were performed for categorical
121
data, as appropriate. Variables with a significance of P < 0.05 were then entered into a
122
multivariate model. A binary logistic regression analysis was used to control for confounding
123
variables and to identify independent risk factors for LOS >3 days. Odds ratios (ORs) and 95%
124
confidence intervals (CIs) were calculated for associations between each risk factor and
125
outcomes of interest. All values are presented as a mean with standard deviation (SD) or 95%
126
confidence intervals (CI). Two-tailed p values <0.05 were considered statistically significant.
127
ORs for OME’s and age are presented per 10-unit increase (mg or years as applicable). ORs for
128
hypotensive events on POD 0 or 1 are reported per 5 events. ORs for CCI is presented per 1-unit
129
increase. Patients were considered to have a laboratory abnormality as a risk factor if they
130
experienced an abnormality on POD 0 or 1.
SC
M AN U
TE D EP
132
AC C
131
RI PT
119
7
ACCEPTED MANUSCRIPT
133
RESULTS The average age of patients undergoing primary total hip arthroplasty in our cohort was
135
62.3 (SD 10.7) years, and 52.7% were women. Both the median and mode for LOS was 2 days.
136
42.9% (548/1278) of patients were discharged on POD1, 35.3% (451/1278) on POD2, 15.5%
137
(198/1278) on POD3, and 3.6% (46/1278) on POD4, and 2.7% (35/1278) were discharged on
138
POD5 to POD7 (Figure 1). The average LOS was 1.89 (SD: 1.05, Range: 1,7).
Univariate analysis comparing patients with a prolonged LOS to those with a LOS of 3
SC
139
RI PT
134
days or less isolated numerous significant predictors for length of stay. These included both non-
141
modifiable and modifiable preoperative and postoperative variables included in Table 3. Not
142
statistically significant on univariate analysis were hypocalcemia, hyponatremia, and elevated
143
serum creatinine on POD 0/1.
144
M AN U
140
Multivariate logistic regression comparing patients with a prolonged LOS to those with a LOS of 3 days or less identified several significant non-modifiable independent predictors of
146
prolonged LOS (Table 4). Non-modifiable risk factors included an increased CCI score (41.1%
147
increase in LOS past 3 days with an increasing CCI by 1 point), non-white race (49.7% increased
148
odds with an increasing age by 10 years), and single marital status (72.4% increased odds).
149
Insurance payer type (Medicare, Medicaid, or private insurance) and sex were not independent
150
risk factors for a prolonged LOS.
EP
AC C
151
TE D
145
More importantly, our model also demonstrated several modifiable risk factors for a
152
prolonged LOS (Table 4). With all other variables constant, post-operative anemia on POD 0 or
153
POD 1 proved to be the greatest risk factor in our model, with a 3.27 increased odds for
154
prolonged LOS greater than 3 days compared to those without post-operative anemia.
155
Additionally, there was a 1.232 increased odds of LOS greater than 3 days for every 5 post-
8
ACCEPTED MANUSCRIPT
operative hypotensive events. Likewise, for every 10 mg increase in oral morphine equivalents
157
received on POD0 or POD1, there was a 1.069 increased odds of a LOS greater than 3 days,
158
equating to a 2.38 increased odds of LOS greater than 3 days for a patient receiving 200mg
159
OMEs on POD0 or POD1. Laboratory abnormalities associated with an extended LOS included
160
a high glucose, a high creatinine, and a low hemoglobin, with a low hemoglobin being the most
161
significant contributor to an extended LOS (Odds Ratio: 3.265). When controlling for the effect
162
of other variables in our model, several modifiable factors were not significant, including
163
hyponatremia and hypocalcemia.
M AN U
164
SC
RI PT
156
AC C
EP
TE D
165
9
ACCEPTED MANUSCRIPT
166
DISCUSSION Non-modifiable risk factors influencing LOS after THA have been investigated by a
168
number of studies and are comprised of a number of patient psychosocial and demographic
169
factors, including insurance payer type, non-Caucasian race, age, BMI, and patient comorbidity
170
and frailty measures, including the CCI and ASA [31-34]. While identifying patients with non-
171
modifiable risk factors can aid in screening patients for stratification to specific care pathways,
172
identifying potentially modifiable risk factors presents the opportunity for risk factor adjustment
173
and the development of additional care pathways to curtail the burden of identified factors.
174
Furthermore, focusing on risk factors that can be adjusted based on careful clinical monitoring
175
rather than fixed variables for which alteration may not be possible will allow for greater
176
modification of LOS. Nevertheless, modifiable factors influencing LOS after total hip
177
arthroplasty have been evaluated less aggressively, especially those that occur in the immediate
178
peri-operative period. Those that have been studied usually pertain to pre-operative factors,
179
including pre-operative anemia and patient expectations regarding length of stay [35-37]. Our
180
results, however, confirm additional and more modifiable risk factors associated with increased
181
hospital LOS after THA, and suggests risk factors that can be adjusted in the immediate peri-
182
operative period.
184 185
SC
M AN U
TE D
EP
AC C
183
RI PT
167
Modifiable Risk Factors
Our results confirm several modifiable risk factors for increased hospital LOS after THA
186
(Table 4). Increased postoperative opioid use during the day of surgery (POD 0) and POD 1 was
187
an independent risk factor for increased hospital LOS. While opioid use may be a surrogate for
188
pain, potentially explaining this association between increased LOS and opioid use, there is
10
ACCEPTED MANUSCRIPT
existing literature that links opioid use to deleterious events in the post-operative period. Nausea,
190
vomiting, pruritis, hypotension and lethargy are common acute side effects of opioid use. Such
191
side effects and their management can have a direct effect on discharge readiness after total hip
192
arthroplasty. In primary total hip and knee arthroplasty, patients who use opioids pre-operatively
193
not only have a longer length of stay, but also increased revision rates and 90-day readmissions
194
[38]. Our data add to the body of evidence indicating the need to focus on opioid reduction
195
strategies in an effort to reduce length of stay, curtail rising hospital costs and improve patient
196
safety. Management pathways utilizing pre-emptive, multimodal, non-narcotic medications have
197
proven effective means of reducing pain and post-operative narcotic requirements. Effective
198
options include the use of perioperative intravenous steroids, non-steroidal anti-inflammatories,
199
and regional anesthetic techniques that spare motor function [39, 40]. In addition, behavioral
200
therapy focusing on pain coping techniques and the avoidance maladaptive pain responses have
201
shown early promise, but have been extremely underutilized in the immediate post-operative
202
period after hip arthroplasty [41-43].
SC
M AN U
TE D
203
RI PT
189
In addition, our data suggests an association between increased post-operative hypotensive events and increased length of stay. Hypotensive events, including orthostatic
205
intolerance, has previously been linked to failed same-day discharge following primary THA
206
[44]. Strategies to minimize post-operative hypotension and symptomatic orthostasis include
207
aggressive intravenous volume repletion and the avoidance of medications which may blunt the
208
sympathetic pathways involved in the normal hemodynamic response to post-operative
209
mobilization. More work is needed to evaluate the efficacy of protocols designed to minimize
210
post-operative hypotension.
AC C
EP
204
11
ACCEPTED MANUSCRIPT
211
Our results also confirm that several laboratory values are independently associated with increased LOS, and protocols to correct these aberrant laboratory values could also prove
213
valuable to reduce LOS after THA. Among the modifiable variables we evaluated, anemia on
214
POD 0 or POD 1 had the highest correlation with increased LOS. Although intraoperative blood
215
loss during total hip arthroplasty is not completely modifiable, there are numerous techniques to
216
minimize early post-operative anemia. Protocols to identify and treat preoperative anemia,
217
hypotension anesthetic techniques, surgical techniques emphasizing efficiency and proper vessel
218
management, and the use of tranexamic acid have all been shown to reduce post-operative
219
anemia and transfusion rates [45, 46]. Without a defined transfusion trigger point, provider
220
tolerance of asymptomatic post-operative anemia can have a major influence transfusion rates
221
and length of stay. Furthermore, proper post-operative management strategies to minimize
222
hyperglycemia could further decrease length of stay. Other studies have linked peri-operative
223
glucose variability to increased length of stay, surgical site infection, prosthetic joint infection,
224
and 90-day mortality [47]. Furthermore, with an increasing utilization of peri-operative
225
corticosteroids, which have been linked to hyperglycemia, peri-operative glucose care pathways
226
should be implemented to better regulate glucose abnormalities [48, 49].
229
SC
M AN U
TE D
EP
228
Non-modifiable Risk Factors
AC C
227
RI PT
212
In this setting, our results confirm several previously identified risk factors for increased
230
hospital LOS after THA. Specifically, increased CCI score, non-Caucasian race, single marital
231
status, and increased age were all independently associated with increased hospital LOS in our
232
model (Table 4). Living alone has been correlated with increased LOS for THA in prior studies,
233
and this factor may account for our finding that married status decreased LOS [13]. Of note,
12
ACCEPTED MANUSCRIPT
when studied in isolation, both Medicaid and Medicare insurance holders had an increased LOS
235
compared to private insurance holders. However, when included in our multivariate model, these
236
associations regarding insurance status and an increased LOS past 3 days were no longer
237
statistically significant (Table 3 and 4). The specificity of our model, including accounting for a
238
number of patient-specific variables not included in large database studies, and the single
239
institution nature of our study, may account for this difference.
242
Strengths and Limitations
While more work needs to be done to determine appropriate protocols for reducing the
M AN U
241
SC
240
RI PT
234
impact of modifiable risk factors and potentially decreasing postoperative LOS, properly
244
identifying modifiable targets as a first step is imperative. In attempting to do so, this study has
245
several strengths. This study has a large sample size of over 1,278 patients giving it considerable
246
power to detect the relative impact of various risk factors. Additionally, unlike national health
247
databases with data aggregated from numerous facilities with varied protocols and surgeon’s
248
with heterogeneous caseloads, our patients were all treated by surgeons performing greater than
249
400 TJA procedures each year, at the same institution with the same protocols. Thus, more
250
standardized data such as vital signs, laboratory values, and medication use can be gleaned and
251
examined across our sample. Furthermore, our patients had a mean LOS of 1.89 days—lower
252
than published estimates, which have reported a mean LOS of over 3 days for TJA [3, 4]. The
253
shorter LOS observed in this patient group is multifactorial and likely due partly to each
254
surgeon’s subspecialty focus on hip and knee arthroplasty, uniform and accelerated care
255
pathways focusing on early mobilization developed within a single specialty orthopaedic
AC C
EP
TE D
243
13
ACCEPTED MANUSCRIPT
256
hospital, multimodal pain management protocols focusing on non-narcotic modalities, and
257
preoperative patient education to set appropriate discharge expectations.
258
Our study also has several limitations including its retrospective nature and that all patient information was derived from medical records. As such, the influence of some variables
260
influencing length of stay in this patient cohort is unknown. Among the most significant of such
261
variables may be the patient’s expected length of stay. However, our pre-operative patient
262
education protocol provides consistent messaging to all patients emphasizing readiness for early
263
discharge after hip arthroplasty. In addition, the time from surgery to the first postoperative
264
physical therapy session was not included in our model. Delays in patient mobilization could
265
have a significant influence on length of stay. However, our nursing protocols specifically
266
address the importance of early mobilization upon recovery from anesthesia. Our study also
267
employs a conservative threshold to define hypotension as any reading with a systolic pressure of
268
90 mmHg or less or a diastolic pressure of 60 mmHg or less, as defined previously in the
269
literature [29, 30]. It is possible that values above this threshold may also be abnormal on the
270
continuum of postoperative blood pressure, especially in the setting of a patient with known
271
hypertension, a common comorbidity in elderly patients undergoing THA. Future studies should
272
examine individual postoperative hypotension readings as a continuous variable and adjust for
273
preoperative values and operative variables to determine the most appropriate threshold to be
274
classified as abnormal in this clinical setting. Lastly, laboratory values were coded as abnormal
275
low, normal, or abnormal high as described above. The categorical presence of abnormal low or
276
high values was thus ascertained and included in our model; however, our model does not
277
consider the magnitude of laboratory derangements. As this study was intended to be an initial
278
examination to identify potentially modifiable risk factors for prolonged hospital LOS, we
AC C
EP
TE D
M AN U
SC
RI PT
259
14
ACCEPTED MANUSCRIPT
279
deemed our handling of laboratory values as appropriate; however, future work should explore
280
each abnormal value and its magnitude in addition to its direction, to further guide initiatives
281
aimed at reducing aberrant values and potentially reducing hospital LOS.
RI PT
282 Conclusion
284
As the financial landscape of US health care is rapidly evolving, identifying means of decreasing
285
hospital length of stay without compromising care after total joint arthroplasty could have a
286
significant financial impact. This study demonstrates that increased opioid use, hypotensive
287
events, acute blood loss anemia, a high glucose, and a high creatinine level in the acute
288
postoperative period are all independently associated with a LOS over 3 days after total hip
289
arthroplasty. As these variables are modifiable, it is important to identify and emphasize the most
290
effective means of their control.
AC C
EP
TE D
291
M AN U
SC
283
15
ACCEPTED MANUSCRIPT
292
References
293
[1.]
Palsis, J.A., et al., The Cost of Joint Replacement: Comparing Two Approaches to Evaluating Costs of Total Hip and Knee Arthroplasty. J Bone Joint Surg Am, 2018.
295
100(4): p. 326-333. [2.]
Arthroplasty from 2002 to 2013. J Bone Joint Surg Am, 2017. 99(5): p. 402-407. [3.]
[4.]
[5.]
[6.]
Hart, A., et al., Comparison of US and Canadian Perioperative Outcomes and Hospital Efficiency After Total Hip and Knee Arthroplasty. JAMA Surg, 2015. 150(10): p. 990-8.
305 306
Iorio, R., et al., Early Results of Medicare's Bundled Payment Initiative for a 90-Day Total Joint Arthroplasty Episode of Care. J Arthroplasty, 2015.
303 304
Cram, P., et al., Total knee arthroplasty volume, utilization, and outcomes among Medicare beneficiaries, 1991-2010. JAMA, 2012. 308(12): p. 1227-36.
301 302
M AN U
total hip arthroplasty, 1991-2008. JAMA, 2011. 305(15): p. 1560-7.
299 300
Cram, P., et al., Clinical characteristics and outcomes of Medicare patients undergoing
SC
297 298
Molloy, I.B., et al., Effects of the Length of Stay on the Cost of Total Knee and Total Hip
TE D
296
RI PT
294
[7.]
Dall, G.F., et al., The influence of pre-operative factors on the length of in-patient stay following primary total hip replacement for osteoarthritis: a multivariate analysis of
308
2302 patients. J Bone Joint Surg Br, 2009. 91(4): p. 434-40. [8.]
Prospective Cohort Analysis. J Arthroplasty, 2015.
311
313
Rissman, C.M., et al., Predictors of Facility Discharge, Range of Motion, and PatientReported Physical Function Improvement After Primary Total Knee Arthroplasty: A
310
312
AC C
309
EP
307
[9.]
Styron, J.F., et al., Patient vs provider characteristics impacting hospital lengths of stay after total knee or hip arthroplasty. J Arthroplasty, 2011. 26(8): p. 1418-26 e1-2.
16
ACCEPTED MANUSCRIPT
315 316 317 318
[10.] Winemaker, M., et al., Not all total joint replacement patients are created equal: preoperative factors and length of stay in hospital. Can J Surg, 2015. 58(3): p. 160-6. [11.] Halawi, M.J., et al., Preoperative predictors of extended hospital length of stay following total knee arthroplasty. J Arthroplasty, 2015. 30(3): p. 361-4.
RI PT
314
[12.] Byrne, P.A., et al., Patient Factors Associated with Shorter Length of Stay Following Total Hip Arthroplasty-A Retrospective Cohort Study. Surg Technol Int, 2017. 31: p.
320
197-200.
321
SC
319
[13.] den Hartog, Y.M., et al., Which patient characteristics influence length of hospital stay after primary total hip arthroplasty in a 'fast-track' setting? Bone Joint J, 2015. 97-
323
B(1): p. 19-23.
M AN U
322
[14.] Elings, J., et al., What preoperative patient-related factors predict inpatient recovery of
325
physical functioning and length of stay after total hip arthroplasty? A systematic
326
review. Clin Rehabil, 2015. 29(5): p. 477-92.
327
TE D
324
[15.] Hayes, J.H., et al., Are clinical and patient assessed outcomes affected by reducing length of hospital stay for total hip arthroplasty? J Arthroplasty, 2000. 15(4): p. 448-
329
52.
331 332 333 334
[16.] Inneh, I.A., et al., Role of Sociodemographic, Co-morbid and Intraoperative Factors in
AC C
330
EP
328
Length of Stay Following Primary Total Hip Arthroplasty. J Arthroplasty, 2015.
30(12): p. 2092-7.
[17.] Makela, K.T., et al., The effect of hospital volume on length of stay, re-admissions, and complications of total hip arthroplasty. Acta Orthop, 2011. 82(1): p. 20-6.
17
ACCEPTED MANUSCRIPT
335
[18.] Olthof, M., et al., Medication Use is a Better Predictor of Length of Hospital Stay in
336
Total Hip Arthroplasty Than the American Society of Anesthetists (ASA) Score. J
337
Arthroplasty, 2017. 32(1): p. 24-27.
339 340
[19.] Petis, S.M., et al., Perioperative Predictors of Length of Stay After Total Hip Arthroplasty. J Arthroplasty, 2016. 31(7): p. 1427-30.
RI PT
338
[20.] Ramkumar, P.N., et al., Evidence-Based Thresholds for the Volume and Length of Stay Relationship in Total Hip Arthroplasty: Outcomes and Economies of Scale. J
342
Arthroplasty, 2018. 33(7): p. 2031-2037.
344 345
[21.] Rudasill, S.E., et al., Do illness rating systems predict discharge location, length of stay,
M AN U
343
SC
341
and cost after total hip arthroplasty? Arthroplast Today, 2018. 4(2): p. 210-215. [22.] Sibia, U.S., J.H. MacDonald, and P.J. King, Predictors of Hospital Length of Stay in an Enhanced Recovery After Surgery Program for Primary Total Hip Arthroplasty. J
347
Arthroplasty, 2016. 31(10): p. 2119-23.
348
TE D
346
[23.] van Aalst, M.J., et al., Can the length of hospital stay after total hip arthroplasty be predicted by preoperative physical function characteristics? Am J Phys Med Rehabil,
350
2014. 93(6): p. 486-92.
352 353 354 355 356
[24.] Tassoudis, V., et al., Impact of intraoperative hypotension on hospital stay in major
AC C
351
EP
349
abdominal surgery. J Anesth, 2011. 25(4): p. 492-9.
[25.] Bundgaard-Nielsen, M., et al., Orthostatic intolerance and the cardiovascular response to early postoperative mobilization. Br J Anaesth, 2009. 102(6): p. 756-62.
[26.] Husted, H., et al., Why still in hospital after fast-track hip and knee arthroplasty? Acta Orthop, 2011. 82(6): p. 679-84.
18
ACCEPTED MANUSCRIPT
357 358 359
[27.] Berger, R.A., et al., Newer anesthesia and rehabilitation protocols enable outpatient hip replacement in selected patients. Clin Orthop Relat Res, 2009. 467(6): p. 1424-30. [28.] Austin, S.R., et al., Why Summary Comorbidity Measures Such As the Charlson Comorbidity Index and Elixhauser Score Work. Med Care, 2015. 53(9): p. e65-72.
361
[29.] Ben-David, B., et al., Minidose bupivacaine-fentanyl spinal anesthesia for surgical
364 365
[30.] Guichard, J.L., et al., Isolated diastolic hypotension and incident heart failure in older adults. Hypertension, 2011. 58(5): p. 895-901.
SC
363
repair of hip fracture in the aged. Anesthesiology, 2000. 92(1): p. 6-10.
[31.] Singh, J.A. and J.D. Cleveland, Medicaid or Medicare insurance payer status and
M AN U
362
RI PT
360
366
household income are associated with outcomes after primary total hip arthroplasty.
367
Clin Rheumatol, 2018. 37(9): p. 2489-2496.
369 370
[32.] Husted, H., et al., Traditions and myths in hip and knee arthroplasty. Acta Orthop, 2014. 85(6): p. 548-55.
TE D
368
[33.] Maradit Kremers, H., et al., Obesity increases length of stay and direct medical costs in total hip arthroplasty. Clin Orthop Relat Res, 2014. 472(4): p. 1232-9.
372
[34.] Lakomkin, N., et al., Higher Modified Charlson Index Scores Are Associated With
374 375 376 377 378 379
Increased Incidence of Complications, Transfusion Events, and Length of Stay
AC C
373
EP
371
Following Revision Hip Arthroplasty. J Arthroplasty, 2017. 32(4): p. 1121-1124.
[35.] Tanzer, D., K. Smith, and M. Tanzer, Changing Patient Expectations Decreases Length of Stay in an Enhanced Recovery Program for THA. Clin Orthop Relat Res, 2018.
476(2): p. 372-378. [36.] Rogers, B.A., et al., Identification and treatment of anaemia in patients awaiting hip replacement. Ann R Coll Surg Engl, 2008. 90(6): p. 504-7.
19
ACCEPTED MANUSCRIPT
380
[37.] Dunne, J.R., et al., Perioperative anemia: an independent risk factor for infection,
381
mortality, and resource utilization in surgery. J Surg Res, 2002. 102(2): p. 237-44.
382
[38.] Weick, J., et al., Preoperative Opioid Use Is Associated with Higher Readmission and Revision Rates in Total Knee and Total Hip Arthroplasty. J Bone Joint Surg Am, 2018.
384
100(14): p. 1171-1176.
385
RI PT
383
[39.] Lunn, T.H., et al., Effect of high-dose preoperative methylprednisolone on recovery after total hip arthroplasty: a randomized, double-blind, placebo-controlled trial. Br J
387
Anaesth, 2013. 110(1): p. 66-73.
389 390
[40.] Kardash, K.J., et al., Single-dose dexamethasone reduces dynamic pain after total hip
M AN U
388
SC
386
arthroplasty. Anesth Analg, 2008. 106(4): p. 1253-7, table of contents. [41.] Dissanayake, R., et al., Does Dexamethasone Reduce Hospital Readiness for Discharge, Pain, Nausea, and Early Patient Satisfaction in Hip and Knee Arthroplasty? A
392
Randomized, Controlled Trial. J Arthroplasty, 2018.
393
TE D
391
[42.] Riddle, D.L., et al., Do Pain Coping and Pain Beliefs Associate With Outcome Measures Before Knee Arthroplasty in Patients Who Catastrophize About Pain? A Cross-sectional
395
Analysis From a Randomized Clinical Trial. Clin Orthop Relat Res, 2018. 476(4): p.
396
778-786.
398 399 400 401
AC C
397
EP
394
[43.] Riddle, D.L., et al., Pain coping skills training for patients with elevated pain catastrophizing who are scheduled for knee arthroplasty: a quasi-experimental study.
Arch Phys Med Rehabil, 2011. 92(6): p. 859-65.
[44.] Fraser, J.F., et al., Identifying Reasons for Failed Same-Day Discharge Following Primary Total Hip Arthroplasty. J Arthroplasty, 2018.
20
ACCEPTED MANUSCRIPT
402
[45.] Wang, F., et al., The efficacy of oral versus intravenous tranexamic acid in reducing
403
blood loss after primary total knee and hip arthroplasty: A meta-analysis. Medicine
404
(Baltimore), 2018. 97(36): p. e12270. [46.] Holt, J.B., et al., Minimizing Blood Transfusion in Total Hip and Knee Arthroplasty
406
Through a Multimodal Approach. J Arthroplasty, 2016. 31(2): p. 378-82.
RI PT
405
[47.] Shohat, N., et al., Increased postoperative glucose variability is associated with adverse
408
outcomes following orthopaedic surgery. Bone Joint J, 2018. 100-B(8): p. 1125-1132.
409
SC
407
[48.] Lindberg-Larsen, V., et al., Preoperative High-Dose Methylprednisolone and Glycemic Control Early After Total Hip and Knee Arthroplasty: A Randomized, Double-Blind,
411
Placebo-Controlled Trial. Anesth Analg, 2018.
412 413
M AN U
410
[49.] Polderman, J.A., et al., Adverse side effects of dexamethasone in surgical patients. Cochrane Database Syst Rev, 2018. 8: p. CD011940.
AC C
EP
TE D
414
21
ACCEPTED MANUSCRIPT
Tables:
3.5 38 2 261 0.2 8.4 98 22 0.66 74 3.5 6.3 15 21 137 9 11.4 33.3
High* 5 126 12 280 1.3 10.5 107 30 1.25 106 5.1 8.2 46 72 145 20 16.1 46.5
AC C
EP
TE D
* to be classified as low or high, value must be below or above cut-off
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
RI PT
Albumin Alkaline Phosphatase Anion Gap Osmolality, Calculated Bilirubin, Total Calcium, Total Chloride Carbon Dioxide Creatinine Glucose Potassium Protein, Total Aspartate Aminotransferase Alanine Aminotransferase Sodium Blood Urea Nitrogen Hemoglobin Hematocrit
3
SC
Table 1. Laboratory Reference Ranges Low*
M AN U
1 2
ACCEPTED MANUSCRIPT
Table 2. Patient Demographics and Risk Factors
OME (mg)1; first 2 days
163.8 ± 122.1 [0, 2623]
Race Marital status CCI2
M AN U
HMO / Managed Care Medicaid Medicare Other Low Calcium High Creatinine High Glucose Low Hemoglobin Low Sodium
Insurance Type
24 25 26 27
Oral Morphine Equivalents, number of OMEs (in mg) consumed on POD 0 and POD 1 Charlson Comorbidity Index
EP
2
AC C
1
TE D
Laboratory Abnormality
Days of IV Narcotic Use
52.7% (673) 47.3% (605) 77.6% (992) 22.4% (286) 67.7% (865) 32.3% (413) 0.38 ± 0.73 (0, 3) 37.1% (472) 2.2% (28) 40.7% (518) 20.1% (256) 48.2% (616) 72.2% (923) 55.0% (703) 71.7% (916) 47.4% (606) 1.07 ± 0.54 [0, 4]
SC
Female Male White Non-White Married Not Married
Sex
RI PT
Age (years)
mean ± sd [min, max] % (n) 62.3 ± 10.7 [16, 94]
Characteristic
ACCEPTED MANUSCRIPT
Characteristic
LOS 0-1 days (n=548)
LOS 2-3 days (n=649)
LOS ≥3 days (n=81)
Hypotensive events 1 (POD 0/1)*
3.40 (3.60)
5.07 (5.07)
CCI 2
0.27 (0.62)
0.42 (0.73)
Female
219/548 (40.0%)
397/649 (61.1%)
Male
329/548 (60.0%)
252/649 (38.8%)
RI PT
Table 3. Univariable Logistic Regression for Predictors of Delayed Discharge
White
458/548 (83.6%)
484/649 (74.6%)
50/81 (61.7%)
0.518 (0.398, 0.673)
Non-White
90/548 (16.4%)
165/649 (25.4%)
31/81 (38.3%)
--
Married
426/548 (77.7%)
411/649 (63.3%)
28/81 (34.6%)
0.386 (0.304, 0.490)
<0.0001
238/649 (36.7%)
53/81 (65.4%)
63.6 (11.1)
67.1 (13.1)
1.451 (1.306, 1.613)
<0.0001
189.5 (140.9)
208.2 (157.7)
1.058 (1.045, 1.070)
<0.0001
198/646 (30.7%)
15/81 (18.5%)
0.924 (0.684, 1.248)
<0.0001
19/646 (2.9%)
4/81 (2.9%)
4.802 (2.147, 10.739)
316/646 (48.9%)
54/81 (66.7%)
2.878 (2.129, 3.890) --
Marital status
60.1 (9.2) 126.7 (71.9) HMO/Managed Care
Insurance
Laboratory Abnormalities
Medicaid Medicare
259/547 (47.4%) 5/547 (0.9%)
TE D
34
148/547 (27.1%)
5.74 (5.61)
1.526 (1.344, 1.733)
<0.0001
0.89 (1.07)
1.662 (1.425, 1.938)
<0.0001
57/81 (70.4%)
--
<0.0001
24/81 (29.6%)
0.409 (0.328, 0.509) <0.0001
Other
135/547 (24.7%)
113/646 (17.5%)
8/81 (9.9%)
Low Calcium
262/548 (52.2%)
312/646 (48.0%)
42/81 (51.8%)
1.044 (0.842, 1.293)
0.6959
High Creatinine
415/548 (75.7%)
442/649 (68.1%)
66/81 (81.5%)
0.817 (0.643, 1.038)
0.0981
High Glucose
278/548 (50.7%)
368/649 (56.7%)
75/81 (70.4%)
1.415 (1.140, 1.756)
0.0016
Low Hemoglobin
288/548 (52.6%)
554/649 (85.4%)
74/81 (91.4%)
5.522 (4.229, 7.211)
<0.0001
248/548 (45.3%)
311/649 (47.9%)
47/81 (58.0%)
1.207 (0.974, 1.496)
0.0854
Low Sodium
EP
OME (mg)*
122/548 (22.3%)
AC C
Age*
Not Married
3
P-value
SC
Race
M AN U
Sex
OR (95% CI)
* variables presented as mean ± sd 1 OR reported per 5 events. 2 Charlson Comorbidity Index 3 OR reported per 10 units (years or mg as applicable). 4 Oral Morphine Equivalents, number of OMEs (in mg) consumed on POD 0 and POD 1
ACCEPTED MANUSCRIPT
Table 4. Multivariate Analysis of Risk Factors Significantly Associated with Length of Stay Greater than 3 days OR (95% CI)
P-value
Hypotensive events on POD 0/1 * 1
1.232 (1.070, 1.419)
0.0037
RI PT
Characteristic
Race
Non-White Vs. White
1.497 (1.105, 2.028)
Marital status
Not Married Vs. Married
1.724 (1.307, 2.273)
<0.0001
1.330 (1.136, 1.556)
0.0004
Age* 2
Lab Abnormalities
<0.0001
Low Hemoglobin
3.265 (2.355, 4.526)
<0.0001
High Glucose
1.887 (1.325, 2.687)
High Creatinine
2.874 (1.821, 4.525)
3
1.411 (1.190, 1.419)
0.0004
<0.0001 <.0.001
M AN U
CCI
1.069 (1.054, 1.084)
SC
OME (mg)*
2
0.0093
AC C
EP
TE D
* variables presented as mean ± sd 1 OR reported per 5 additional events. 2 OR reported per 10 units (years or mg as applicable), Oral Morphine Equivalents. 3 OR reported per 1-unit increase. Charlson Comorbidity Index.
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
Figures
AC C
EP
TE D
Figure 1. Graph shows a histogram of LOS after THA. The dashed line is at 3 days and represents the cutoff used for statistical analysis.
S0883-5403(19)30005-1
DOI:
https://doi.org/10.1016/j.arth.2018.12.041
Reference:
YARTH 56978
To appear in:
The Journal of Arthroplasty
Received Date: 7 December 2018 Revised Date:
25 December 2018
Accepted Date: 31 December 2018
Please cite this article as: Farley KX, Anastasio AT, Premkumar A, Boden SD, Gottschalk MB, Bradbury TL, The Influence of Modifiable, Post-operative Patient variables on Length of Stay after Total Hip Arthroplasty, The Journal of Arthroplasty (2019), doi: https://doi.org/10.1016/j.arth.2018.12.041. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Title Page
1
Title: The Influence of Modifiable, Post-operative Patient variables on Length of Stay after
3
Total Hip Arthroplasty
4
Authors:
5
Kevin X. Farley, BS1 ([email protected])
6
Albert T. Anastasio, BA1 ([email protected])
7
Ajay Premkumar, MD2 ([email protected])
8
Scott D. Boden, MD1 ([email protected])
9
Michael B. Gottschalk, MD1 ([email protected])
M AN U
SC
RI PT
2
Thomas L. Bradbury, MD1 ([email protected])
11
Affiliation:
12
1
Department of Orthopaedic Surgery, Emory University School of Medicine, Atlanta, GA, USA
13
2
Department of Orthopaedic Surgery, Hospital for Special Surgery, New York, NY, USA
14
Address:
15
1
1648 Pierce Drive NE, Atlanta, GA 30307
16
2
535 E 70th St, New York, NY 10021
17
Corresponding author:
18
Michael B. Gottschalk, MD
19
1648 Pierce Dr. NE
20
Atlanta, GA 30307
21
Phone: (404) 251-1566
22
Email: [email protected]
AC C
EP
TE D
10
ACCEPTED MANUSCRIPT
Disclosures: Each author certifies that he or she has no commercial associations (eg,
24
consultancies, stock ownership, equity interest, patent/licensing arrangements, etc) that might
25
pose a conflict of interest in connection with the submitted article.
26
Word Count: 3,139 (including in-text references to tables and figures)
AC C
EP
TE D
M AN U
SC
RI PT
23
ACCEPTED MANUSCRIPT
1
Title: The Influence of Modifiable, Post-operative Patient variables on Length of Stay after
2
Total Hip Arthroplasty
3 ABSTRACT
5
Background
6
Many studies have examined strategies to reduce length of stay (LOS) after total hip arthroplasty
7
(THA), but few have focused on modifiable patient-specific information in the acute post-
8
operative period. This study investigates the determinants of LOS after THA, with a focus on
9
potentially modifiable factors.
M AN U
SC
RI PT
4
Methods
11
1,278 patients undergoing elective THA from 2012 to 2014 were extracted from our institutional
12
data warehouse at our academic orthopaedic specialty hospital. Data were collected on patient
13
demographics, comorbidities, inpatient opioid use, hypotensive events, and abnormalities in
14
laboratory values, all occurring on post-operative day (POD) 0 or 1. The main outcome was
15
hospital LOS. Multivariate regression analysis was performed to identify independent risk
16
factors for LOS over 3 days.
17
Results
18
The average age of patients undergoing primary total hip arthroplasty in our cohort was 62.3 (SD
19
10.7) years, and 52.7% were women. 6.3% (81/1278) of patients had a LOS more than 3 days.
20
Multivariate regression analysis demonstrated several statistically significant non-modifiable and
21
modifiable risk factors that influence LOS after THA. Non-modifiable risk factors included non-
22
white race (Odds Ratio [OR], 1.497), single marital status (OR, 1.724), increasing age (OR,
23
1.330), and increasing Charlson Comorbidity Index (OR, 1.411). Potentially modifiable risk
AC C
EP
TE D
10
1
ACCEPTED MANUSCRIPT
factors included every 10mg oral morphine equivalent consumption (1.069), every 5 post-
25
operative hypotensive events (OR, 1.232), low hemoglobin (OR, 3.265), high glucose levels
26
(OR, 1.887), and a high creatinine (OR, 2.874).
27
Conclusion
28
This study identifies potentially modifiable factors that are associated with increased LOS after
29
THA, including post-operative opioid use and hypotensive events. Efforts to control narcotic use
30
and initiatives aimed to reduce early postoperative hypotension could aid in reducing LOS.
31
Furthermore, attempts should be made to correct post-operative anemia, high glucose levels, and
32
a high creatinine when possible.
33
Key Words: Total Hip Arthroplasty; Opioid; Hypotension; Modifiable Risk Factors; Length of
34
Stay; Anemia
AC C
EP
TE D
M AN U
SC
RI PT
24
2
ACCEPTED MANUSCRIPT
35 36
INTRODUCTION Total hip arthroplasty (THA) is an effective and commonly performed procedure. Its relevance continues to increase in light of the aging American population. A recent study cited
38
the cost of a single primary THA at $22,076 [1]. Given high volume of these procedures, THA
39
represents a substantial cost to the healthcare system. One of the major drivers of the cost
40
associated with THA is the hospital length of stay (LOS) after the procedure [2].
Studies examining large cohorts of patients have estimated the LOS after total hip
SC
41
RI PT
37
arthroplasty to be roughly 2.97 days, down from 4.06 days in 2002 [2]. As the financial
43
landscape of healthcare is set to undergo significant changes, efforts aimed at decreasing LOS
44
after THA is one potential target to help contain and potentially decrease healthcare costs
45
associated with these procedures [3-6]. At the same time, efforts to reduce LOS that increase the
46
risk of readmission or post-operative complication could be counterproductive. Multiple authors have analyzed factors associated with increased length of stay in the
TE D
47
M AN U
42
hospital after TJA [6-11]. Additionally, several studies have been published looking at predictors
49
of LOS specifically for THA [12-23]. One such study found that illness classification indices like
50
the American Society for Anesthesiologists (ASA) physical classification system and the
51
Severity of Illness scoring system were somewhat predictive of LOS after THA, but that other
52
factors such as insurance status and race remained stronger predictors [20]. Another study found
53
that lower ASA, age, and male gender were correlated with decreased LOS [12]. While certain
54
patient characteristics, such as age, body mass index (BMI), various comorbidities, race, income,
55
and insurance status have been correlated with an increased LOS, less attention has been focused
56
on how potentially modifiable risk factors, such as post-operative opioid use, aberrant vital signs,
57
or abnormal post-operative laboratory values may affect LOS.
AC C
EP
48
3
ACCEPTED MANUSCRIPT
In particular, postoperative hypotension has been linked to increased LOS in other
59
surgical fields [24, 25]. While there are published clinical protocols that stress the importance of
60
fluid management to minimize risk of postoperative hypotension after THA, postoperative
61
hypotension and symptomatic post-operative orthostasis remains underreported in the
62
orthopaedic literature [26, 27].
RI PT
58
In this study, we analyzed inpatient hospital records for individuals undergoing primary
64
total hip arthroplasty at a single institution. We sought to identify independent modifiable risk
65
factors for delayed discharge that have been previously underrepresented in the literature. Our
66
primary outcome measure was postoperative hypotensive events. We hypothesized that, similar
67
to other surgical specialties, the number of hypotensive events on post-operative day 0 and 1
68
after THA is correlated to patient LOS over 3 days. Secondary outcomes measures included
69
post-operative opioid use and abnormal post-operative laboratory values. We also examined
70
patient characteristics such as demographic information, comorbidities, and insurance status, to
71
confirm if an independent link between these variables and LOS exists.
AC C
EP
TE D
M AN U
SC
63
4
ACCEPTED MANUSCRIPT
72 73
PATIENTS AND METHODS After Institutional Review Board approval (Approval Date 8/28/2015, IRB 76391), we identified all patients undergoing primary THA for degenerative osteoarthritis of the hip between
75
June 2012 and August 2014, by one of four hip arthroplasty specialists at our institution. Each of
76
the four surgeons is fellowship-trained and works in the same orthopaedic specialty hospital.
77
None of the 1,278 patients identified by the above criteria died during their post-operative
78
hospital stay, and none were subsequently excluded from our analysis.
SC
79
RI PT
74
The primary outcome measure of interest was hospital LOS, defined categorically as days. Each night spent in the hospital after surgery was considered an increase in LOS of one
81
day. For example, a patient discharged on the day following surgery spent one night in the
82
hospital and thus had a LOS of 1. Due to lack of available, reliably accurate data on the exact
83
time of discharge during each day, efforts were not made to distinguish if a patient was
84
discharged in the morning or the evening on a particular postoperative day. All patients were
85
admitted to the hospital the day of surgery.
TE D
M AN U
80
Our model examined predictors of LOS including both non-modifiable and modifiable
87
patient variables. Non-modifiable variables included patient demographics and comorbidities.
88
Modifiable variables included postoperative hypotension, inpatient post-operative opioid
89
medication use, and abnormal post-operative laboratory values. To assess comorbidities, the
90
Charlson Comorbidity Index (CCI) was computed for each patient using International
91
Classification of Diseases (ICD) categories [28]. Opioids were converted to oral morphine
92
equivalents (OME) for comparison and were recorded as the number of OMEs consumed on the
93
day of surgery (POD 0) and post-operative day 1 (POD 1). The number of postoperative
94
hypotensive events, defined as a systolic blood pressure less than 90 mmHg or a diastolic blood
AC C
EP
86
5
ACCEPTED MANUSCRIPT
95
pressure less than 60 mmHg for any single reading, was recorded as a continuous variable for
96
each postoperative day [29, 30].
97
Given the lack of universal consensus on laboratory threshold values, we determined abnormal values for laboratory results as the default threshold values embedded as clinical
99
support tools in our institution’s electronic medical record system (Cerner Systems). The specific
RI PT
98
cutoffs used for each laboratory value can be seen in Table 1. The presence of either an
101
abnormal high or abnormal low value was determined for each post-operative day by
102
retrospective review of all lab results for each patient during their hospitalization. Abnormal
103
values for specific lab results were thus coded as categorical variables, either ‘abnormal high’ or
104
‘abnormal low’, for each post-operative day.
M AN U
105
SC
100
In an effort to preserve the fidelity of our model, we attempted to remove very rare abnormal lab values from our analysis. To accomplish this, if an abnormal value appeared in at
107
least 5% of our sample, which we deemed a conservative threshold, that specific laboratory
108
result was included in our analysis. The presence of low calcium, high creatinine, high glucose,
109
low Hemoglobin, and low Sodium values were the only laboratory results to occur at a frequency
110
above this threshold, and thus were included in the model. A baseline characterization of
111
laboratory abnormalities and demographic characteristics for all patients in our sample can be
112
seen in Table 2.
EP
AC C
113
TE D
106
Given that 93.7% (1197/1278) of our cohort had a LOS ≤3 days and that the difference
114
between a LOS of 2 or 3 days may be dependent on what time of day surgery was performed, we
115
chose to evaluate risk factors for prolonged LOS past 3 days. This determination was made to
116
isolate a group of patients clearly above the average who would definitively benefit from more
117
thorough post-operative care pathways. See histogram (Figure 1) for LOS distribution.
6
ACCEPTED MANUSCRIPT
118
Statistical analysis was performed using the R:A language and environment for statistical computing (R Foundation for Statistical Computing, http://www.R-project.org). Student t tests
120
were performed for continuous data and χ2 or Fisher exact tests were performed for categorical
121
data, as appropriate. Variables with a significance of P < 0.05 were then entered into a
122
multivariate model. A binary logistic regression analysis was used to control for confounding
123
variables and to identify independent risk factors for LOS >3 days. Odds ratios (ORs) and 95%
124
confidence intervals (CIs) were calculated for associations between each risk factor and
125
outcomes of interest. All values are presented as a mean with standard deviation (SD) or 95%
126
confidence intervals (CI). Two-tailed p values <0.05 were considered statistically significant.
127
ORs for OME’s and age are presented per 10-unit increase (mg or years as applicable). ORs for
128
hypotensive events on POD 0 or 1 are reported per 5 events. ORs for CCI is presented per 1-unit
129
increase. Patients were considered to have a laboratory abnormality as a risk factor if they
130
experienced an abnormality on POD 0 or 1.
SC
M AN U
TE D EP
132
AC C
131
RI PT
119
7
ACCEPTED MANUSCRIPT
133
RESULTS The average age of patients undergoing primary total hip arthroplasty in our cohort was
135
62.3 (SD 10.7) years, and 52.7% were women. Both the median and mode for LOS was 2 days.
136
42.9% (548/1278) of patients were discharged on POD1, 35.3% (451/1278) on POD2, 15.5%
137
(198/1278) on POD3, and 3.6% (46/1278) on POD4, and 2.7% (35/1278) were discharged on
138
POD5 to POD7 (Figure 1). The average LOS was 1.89 (SD: 1.05, Range: 1,7).
Univariate analysis comparing patients with a prolonged LOS to those with a LOS of 3
SC
139
RI PT
134
days or less isolated numerous significant predictors for length of stay. These included both non-
141
modifiable and modifiable preoperative and postoperative variables included in Table 3. Not
142
statistically significant on univariate analysis were hypocalcemia, hyponatremia, and elevated
143
serum creatinine on POD 0/1.
144
M AN U
140
Multivariate logistic regression comparing patients with a prolonged LOS to those with a LOS of 3 days or less identified several significant non-modifiable independent predictors of
146
prolonged LOS (Table 4). Non-modifiable risk factors included an increased CCI score (41.1%
147
increase in LOS past 3 days with an increasing CCI by 1 point), non-white race (49.7% increased
148
odds with an increasing age by 10 years), and single marital status (72.4% increased odds).
149
Insurance payer type (Medicare, Medicaid, or private insurance) and sex were not independent
150
risk factors for a prolonged LOS.
EP
AC C
151
TE D
145
More importantly, our model also demonstrated several modifiable risk factors for a
152
prolonged LOS (Table 4). With all other variables constant, post-operative anemia on POD 0 or
153
POD 1 proved to be the greatest risk factor in our model, with a 3.27 increased odds for
154
prolonged LOS greater than 3 days compared to those without post-operative anemia.
155
Additionally, there was a 1.232 increased odds of LOS greater than 3 days for every 5 post-
8
ACCEPTED MANUSCRIPT
operative hypotensive events. Likewise, for every 10 mg increase in oral morphine equivalents
157
received on POD0 or POD1, there was a 1.069 increased odds of a LOS greater than 3 days,
158
equating to a 2.38 increased odds of LOS greater than 3 days for a patient receiving 200mg
159
OMEs on POD0 or POD1. Laboratory abnormalities associated with an extended LOS included
160
a high glucose, a high creatinine, and a low hemoglobin, with a low hemoglobin being the most
161
significant contributor to an extended LOS (Odds Ratio: 3.265). When controlling for the effect
162
of other variables in our model, several modifiable factors were not significant, including
163
hyponatremia and hypocalcemia.
M AN U
164
SC
RI PT
156
AC C
EP
TE D
165
9
ACCEPTED MANUSCRIPT
166
DISCUSSION Non-modifiable risk factors influencing LOS after THA have been investigated by a
168
number of studies and are comprised of a number of patient psychosocial and demographic
169
factors, including insurance payer type, non-Caucasian race, age, BMI, and patient comorbidity
170
and frailty measures, including the CCI and ASA [31-34]. While identifying patients with non-
171
modifiable risk factors can aid in screening patients for stratification to specific care pathways,
172
identifying potentially modifiable risk factors presents the opportunity for risk factor adjustment
173
and the development of additional care pathways to curtail the burden of identified factors.
174
Furthermore, focusing on risk factors that can be adjusted based on careful clinical monitoring
175
rather than fixed variables for which alteration may not be possible will allow for greater
176
modification of LOS. Nevertheless, modifiable factors influencing LOS after total hip
177
arthroplasty have been evaluated less aggressively, especially those that occur in the immediate
178
peri-operative period. Those that have been studied usually pertain to pre-operative factors,
179
including pre-operative anemia and patient expectations regarding length of stay [35-37]. Our
180
results, however, confirm additional and more modifiable risk factors associated with increased
181
hospital LOS after THA, and suggests risk factors that can be adjusted in the immediate peri-
182
operative period.
184 185
SC
M AN U
TE D
EP
AC C
183
RI PT
167
Modifiable Risk Factors
Our results confirm several modifiable risk factors for increased hospital LOS after THA
186
(Table 4). Increased postoperative opioid use during the day of surgery (POD 0) and POD 1 was
187
an independent risk factor for increased hospital LOS. While opioid use may be a surrogate for
188
pain, potentially explaining this association between increased LOS and opioid use, there is
10
ACCEPTED MANUSCRIPT
existing literature that links opioid use to deleterious events in the post-operative period. Nausea,
190
vomiting, pruritis, hypotension and lethargy are common acute side effects of opioid use. Such
191
side effects and their management can have a direct effect on discharge readiness after total hip
192
arthroplasty. In primary total hip and knee arthroplasty, patients who use opioids pre-operatively
193
not only have a longer length of stay, but also increased revision rates and 90-day readmissions
194
[38]. Our data add to the body of evidence indicating the need to focus on opioid reduction
195
strategies in an effort to reduce length of stay, curtail rising hospital costs and improve patient
196
safety. Management pathways utilizing pre-emptive, multimodal, non-narcotic medications have
197
proven effective means of reducing pain and post-operative narcotic requirements. Effective
198
options include the use of perioperative intravenous steroids, non-steroidal anti-inflammatories,
199
and regional anesthetic techniques that spare motor function [39, 40]. In addition, behavioral
200
therapy focusing on pain coping techniques and the avoidance maladaptive pain responses have
201
shown early promise, but have been extremely underutilized in the immediate post-operative
202
period after hip arthroplasty [41-43].
SC
M AN U
TE D
203
RI PT
189
In addition, our data suggests an association between increased post-operative hypotensive events and increased length of stay. Hypotensive events, including orthostatic
205
intolerance, has previously been linked to failed same-day discharge following primary THA
206
[44]. Strategies to minimize post-operative hypotension and symptomatic orthostasis include
207
aggressive intravenous volume repletion and the avoidance of medications which may blunt the
208
sympathetic pathways involved in the normal hemodynamic response to post-operative
209
mobilization. More work is needed to evaluate the efficacy of protocols designed to minimize
210
post-operative hypotension.
AC C
EP
204
11
ACCEPTED MANUSCRIPT
211
Our results also confirm that several laboratory values are independently associated with increased LOS, and protocols to correct these aberrant laboratory values could also prove
213
valuable to reduce LOS after THA. Among the modifiable variables we evaluated, anemia on
214
POD 0 or POD 1 had the highest correlation with increased LOS. Although intraoperative blood
215
loss during total hip arthroplasty is not completely modifiable, there are numerous techniques to
216
minimize early post-operative anemia. Protocols to identify and treat preoperative anemia,
217
hypotension anesthetic techniques, surgical techniques emphasizing efficiency and proper vessel
218
management, and the use of tranexamic acid have all been shown to reduce post-operative
219
anemia and transfusion rates [45, 46]. Without a defined transfusion trigger point, provider
220
tolerance of asymptomatic post-operative anemia can have a major influence transfusion rates
221
and length of stay. Furthermore, proper post-operative management strategies to minimize
222
hyperglycemia could further decrease length of stay. Other studies have linked peri-operative
223
glucose variability to increased length of stay, surgical site infection, prosthetic joint infection,
224
and 90-day mortality [47]. Furthermore, with an increasing utilization of peri-operative
225
corticosteroids, which have been linked to hyperglycemia, peri-operative glucose care pathways
226
should be implemented to better regulate glucose abnormalities [48, 49].
229
SC
M AN U
TE D
EP
228
Non-modifiable Risk Factors
AC C
227
RI PT
212
In this setting, our results confirm several previously identified risk factors for increased
230
hospital LOS after THA. Specifically, increased CCI score, non-Caucasian race, single marital
231
status, and increased age were all independently associated with increased hospital LOS in our
232
model (Table 4). Living alone has been correlated with increased LOS for THA in prior studies,
233
and this factor may account for our finding that married status decreased LOS [13]. Of note,
12
ACCEPTED MANUSCRIPT
when studied in isolation, both Medicaid and Medicare insurance holders had an increased LOS
235
compared to private insurance holders. However, when included in our multivariate model, these
236
associations regarding insurance status and an increased LOS past 3 days were no longer
237
statistically significant (Table 3 and 4). The specificity of our model, including accounting for a
238
number of patient-specific variables not included in large database studies, and the single
239
institution nature of our study, may account for this difference.
242
Strengths and Limitations
While more work needs to be done to determine appropriate protocols for reducing the
M AN U
241
SC
240
RI PT
234
impact of modifiable risk factors and potentially decreasing postoperative LOS, properly
244
identifying modifiable targets as a first step is imperative. In attempting to do so, this study has
245
several strengths. This study has a large sample size of over 1,278 patients giving it considerable
246
power to detect the relative impact of various risk factors. Additionally, unlike national health
247
databases with data aggregated from numerous facilities with varied protocols and surgeon’s
248
with heterogeneous caseloads, our patients were all treated by surgeons performing greater than
249
400 TJA procedures each year, at the same institution with the same protocols. Thus, more
250
standardized data such as vital signs, laboratory values, and medication use can be gleaned and
251
examined across our sample. Furthermore, our patients had a mean LOS of 1.89 days—lower
252
than published estimates, which have reported a mean LOS of over 3 days for TJA [3, 4]. The
253
shorter LOS observed in this patient group is multifactorial and likely due partly to each
254
surgeon’s subspecialty focus on hip and knee arthroplasty, uniform and accelerated care
255
pathways focusing on early mobilization developed within a single specialty orthopaedic
AC C
EP
TE D
243
13
ACCEPTED MANUSCRIPT
256
hospital, multimodal pain management protocols focusing on non-narcotic modalities, and
257
preoperative patient education to set appropriate discharge expectations.
258
Our study also has several limitations including its retrospective nature and that all patient information was derived from medical records. As such, the influence of some variables
260
influencing length of stay in this patient cohort is unknown. Among the most significant of such
261
variables may be the patient’s expected length of stay. However, our pre-operative patient
262
education protocol provides consistent messaging to all patients emphasizing readiness for early
263
discharge after hip arthroplasty. In addition, the time from surgery to the first postoperative
264
physical therapy session was not included in our model. Delays in patient mobilization could
265
have a significant influence on length of stay. However, our nursing protocols specifically
266
address the importance of early mobilization upon recovery from anesthesia. Our study also
267
employs a conservative threshold to define hypotension as any reading with a systolic pressure of
268
90 mmHg or less or a diastolic pressure of 60 mmHg or less, as defined previously in the
269
literature [29, 30]. It is possible that values above this threshold may also be abnormal on the
270
continuum of postoperative blood pressure, especially in the setting of a patient with known
271
hypertension, a common comorbidity in elderly patients undergoing THA. Future studies should
272
examine individual postoperative hypotension readings as a continuous variable and adjust for
273
preoperative values and operative variables to determine the most appropriate threshold to be
274
classified as abnormal in this clinical setting. Lastly, laboratory values were coded as abnormal
275
low, normal, or abnormal high as described above. The categorical presence of abnormal low or
276
high values was thus ascertained and included in our model; however, our model does not
277
consider the magnitude of laboratory derangements. As this study was intended to be an initial
278
examination to identify potentially modifiable risk factors for prolonged hospital LOS, we
AC C
EP
TE D
M AN U
SC
RI PT
259
14
ACCEPTED MANUSCRIPT
279
deemed our handling of laboratory values as appropriate; however, future work should explore
280
each abnormal value and its magnitude in addition to its direction, to further guide initiatives
281
aimed at reducing aberrant values and potentially reducing hospital LOS.
RI PT
282 Conclusion
284
As the financial landscape of US health care is rapidly evolving, identifying means of decreasing
285
hospital length of stay without compromising care after total joint arthroplasty could have a
286
significant financial impact. This study demonstrates that increased opioid use, hypotensive
287
events, acute blood loss anemia, a high glucose, and a high creatinine level in the acute
288
postoperative period are all independently associated with a LOS over 3 days after total hip
289
arthroplasty. As these variables are modifiable, it is important to identify and emphasize the most
290
effective means of their control.
AC C
EP
TE D
291
M AN U
SC
283
15
ACCEPTED MANUSCRIPT
292
References
293
[1.]
Palsis, J.A., et al., The Cost of Joint Replacement: Comparing Two Approaches to Evaluating Costs of Total Hip and Knee Arthroplasty. J Bone Joint Surg Am, 2018.
295
100(4): p. 326-333. [2.]
Arthroplasty from 2002 to 2013. J Bone Joint Surg Am, 2017. 99(5): p. 402-407. [3.]
[4.]
[5.]
[6.]
Hart, A., et al., Comparison of US and Canadian Perioperative Outcomes and Hospital Efficiency After Total Hip and Knee Arthroplasty. JAMA Surg, 2015. 150(10): p. 990-8.
305 306
Iorio, R., et al., Early Results of Medicare's Bundled Payment Initiative for a 90-Day Total Joint Arthroplasty Episode of Care. J Arthroplasty, 2015.
303 304
Cram, P., et al., Total knee arthroplasty volume, utilization, and outcomes among Medicare beneficiaries, 1991-2010. JAMA, 2012. 308(12): p. 1227-36.
301 302
M AN U
total hip arthroplasty, 1991-2008. JAMA, 2011. 305(15): p. 1560-7.
299 300
Cram, P., et al., Clinical characteristics and outcomes of Medicare patients undergoing
SC
297 298
Molloy, I.B., et al., Effects of the Length of Stay on the Cost of Total Knee and Total Hip
TE D
296
RI PT
294
[7.]
Dall, G.F., et al., The influence of pre-operative factors on the length of in-patient stay following primary total hip replacement for osteoarthritis: a multivariate analysis of
308
2302 patients. J Bone Joint Surg Br, 2009. 91(4): p. 434-40. [8.]
Prospective Cohort Analysis. J Arthroplasty, 2015.
311
313
Rissman, C.M., et al., Predictors of Facility Discharge, Range of Motion, and PatientReported Physical Function Improvement After Primary Total Knee Arthroplasty: A
310
312
AC C
309
EP
307
[9.]
Styron, J.F., et al., Patient vs provider characteristics impacting hospital lengths of stay after total knee or hip arthroplasty. J Arthroplasty, 2011. 26(8): p. 1418-26 e1-2.
16
ACCEPTED MANUSCRIPT
315 316 317 318
[10.] Winemaker, M., et al., Not all total joint replacement patients are created equal: preoperative factors and length of stay in hospital. Can J Surg, 2015. 58(3): p. 160-6. [11.] Halawi, M.J., et al., Preoperative predictors of extended hospital length of stay following total knee arthroplasty. J Arthroplasty, 2015. 30(3): p. 361-4.
RI PT
314
[12.] Byrne, P.A., et al., Patient Factors Associated with Shorter Length of Stay Following Total Hip Arthroplasty-A Retrospective Cohort Study. Surg Technol Int, 2017. 31: p.
320
197-200.
321
SC
319
[13.] den Hartog, Y.M., et al., Which patient characteristics influence length of hospital stay after primary total hip arthroplasty in a 'fast-track' setting? Bone Joint J, 2015. 97-
323
B(1): p. 19-23.
M AN U
322
[14.] Elings, J., et al., What preoperative patient-related factors predict inpatient recovery of
325
physical functioning and length of stay after total hip arthroplasty? A systematic
326
review. Clin Rehabil, 2015. 29(5): p. 477-92.
327
TE D
324
[15.] Hayes, J.H., et al., Are clinical and patient assessed outcomes affected by reducing length of hospital stay for total hip arthroplasty? J Arthroplasty, 2000. 15(4): p. 448-
329
52.
331 332 333 334
[16.] Inneh, I.A., et al., Role of Sociodemographic, Co-morbid and Intraoperative Factors in
AC C
330
EP
328
Length of Stay Following Primary Total Hip Arthroplasty. J Arthroplasty, 2015.
30(12): p. 2092-7.
[17.] Makela, K.T., et al., The effect of hospital volume on length of stay, re-admissions, and complications of total hip arthroplasty. Acta Orthop, 2011. 82(1): p. 20-6.
17
ACCEPTED MANUSCRIPT
335
[18.] Olthof, M., et al., Medication Use is a Better Predictor of Length of Hospital Stay in
336
Total Hip Arthroplasty Than the American Society of Anesthetists (ASA) Score. J
337
Arthroplasty, 2017. 32(1): p. 24-27.
339 340
[19.] Petis, S.M., et al., Perioperative Predictors of Length of Stay After Total Hip Arthroplasty. J Arthroplasty, 2016. 31(7): p. 1427-30.
RI PT
338
[20.] Ramkumar, P.N., et al., Evidence-Based Thresholds for the Volume and Length of Stay Relationship in Total Hip Arthroplasty: Outcomes and Economies of Scale. J
342
Arthroplasty, 2018. 33(7): p. 2031-2037.
344 345
[21.] Rudasill, S.E., et al., Do illness rating systems predict discharge location, length of stay,
M AN U
343
SC
341
and cost after total hip arthroplasty? Arthroplast Today, 2018. 4(2): p. 210-215. [22.] Sibia, U.S., J.H. MacDonald, and P.J. King, Predictors of Hospital Length of Stay in an Enhanced Recovery After Surgery Program for Primary Total Hip Arthroplasty. J
347
Arthroplasty, 2016. 31(10): p. 2119-23.
348
TE D
346
[23.] van Aalst, M.J., et al., Can the length of hospital stay after total hip arthroplasty be predicted by preoperative physical function characteristics? Am J Phys Med Rehabil,
350
2014. 93(6): p. 486-92.
352 353 354 355 356
[24.] Tassoudis, V., et al., Impact of intraoperative hypotension on hospital stay in major
AC C
351
EP
349
abdominal surgery. J Anesth, 2011. 25(4): p. 492-9.
[25.] Bundgaard-Nielsen, M., et al., Orthostatic intolerance and the cardiovascular response to early postoperative mobilization. Br J Anaesth, 2009. 102(6): p. 756-62.
[26.] Husted, H., et al., Why still in hospital after fast-track hip and knee arthroplasty? Acta Orthop, 2011. 82(6): p. 679-84.
18
ACCEPTED MANUSCRIPT
357 358 359
[27.] Berger, R.A., et al., Newer anesthesia and rehabilitation protocols enable outpatient hip replacement in selected patients. Clin Orthop Relat Res, 2009. 467(6): p. 1424-30. [28.] Austin, S.R., et al., Why Summary Comorbidity Measures Such As the Charlson Comorbidity Index and Elixhauser Score Work. Med Care, 2015. 53(9): p. e65-72.
361
[29.] Ben-David, B., et al., Minidose bupivacaine-fentanyl spinal anesthesia for surgical
364 365
[30.] Guichard, J.L., et al., Isolated diastolic hypotension and incident heart failure in older adults. Hypertension, 2011. 58(5): p. 895-901.
SC
363
repair of hip fracture in the aged. Anesthesiology, 2000. 92(1): p. 6-10.
[31.] Singh, J.A. and J.D. Cleveland, Medicaid or Medicare insurance payer status and
M AN U
362
RI PT
360
366
household income are associated with outcomes after primary total hip arthroplasty.
367
Clin Rheumatol, 2018. 37(9): p. 2489-2496.
369 370
[32.] Husted, H., et al., Traditions and myths in hip and knee arthroplasty. Acta Orthop, 2014. 85(6): p. 548-55.
TE D
368
[33.] Maradit Kremers, H., et al., Obesity increases length of stay and direct medical costs in total hip arthroplasty. Clin Orthop Relat Res, 2014. 472(4): p. 1232-9.
372
[34.] Lakomkin, N., et al., Higher Modified Charlson Index Scores Are Associated With
374 375 376 377 378 379
Increased Incidence of Complications, Transfusion Events, and Length of Stay
AC C
373
EP
371
Following Revision Hip Arthroplasty. J Arthroplasty, 2017. 32(4): p. 1121-1124.
[35.] Tanzer, D., K. Smith, and M. Tanzer, Changing Patient Expectations Decreases Length of Stay in an Enhanced Recovery Program for THA. Clin Orthop Relat Res, 2018.
476(2): p. 372-378. [36.] Rogers, B.A., et al., Identification and treatment of anaemia in patients awaiting hip replacement. Ann R Coll Surg Engl, 2008. 90(6): p. 504-7.
19
ACCEPTED MANUSCRIPT
380
[37.] Dunne, J.R., et al., Perioperative anemia: an independent risk factor for infection,
381
mortality, and resource utilization in surgery. J Surg Res, 2002. 102(2): p. 237-44.
382
[38.] Weick, J., et al., Preoperative Opioid Use Is Associated with Higher Readmission and Revision Rates in Total Knee and Total Hip Arthroplasty. J Bone Joint Surg Am, 2018.
384
100(14): p. 1171-1176.
385
RI PT
383
[39.] Lunn, T.H., et al., Effect of high-dose preoperative methylprednisolone on recovery after total hip arthroplasty: a randomized, double-blind, placebo-controlled trial. Br J
387
Anaesth, 2013. 110(1): p. 66-73.
389 390
[40.] Kardash, K.J., et al., Single-dose dexamethasone reduces dynamic pain after total hip
M AN U
388
SC
386
arthroplasty. Anesth Analg, 2008. 106(4): p. 1253-7, table of contents. [41.] Dissanayake, R., et al., Does Dexamethasone Reduce Hospital Readiness for Discharge, Pain, Nausea, and Early Patient Satisfaction in Hip and Knee Arthroplasty? A
392
Randomized, Controlled Trial. J Arthroplasty, 2018.
393
TE D
391
[42.] Riddle, D.L., et al., Do Pain Coping and Pain Beliefs Associate With Outcome Measures Before Knee Arthroplasty in Patients Who Catastrophize About Pain? A Cross-sectional
395
Analysis From a Randomized Clinical Trial. Clin Orthop Relat Res, 2018. 476(4): p.
396
778-786.
398 399 400 401
AC C
397
EP
394
[43.] Riddle, D.L., et al., Pain coping skills training for patients with elevated pain catastrophizing who are scheduled for knee arthroplasty: a quasi-experimental study.
Arch Phys Med Rehabil, 2011. 92(6): p. 859-65.
[44.] Fraser, J.F., et al., Identifying Reasons for Failed Same-Day Discharge Following Primary Total Hip Arthroplasty. J Arthroplasty, 2018.
20
ACCEPTED MANUSCRIPT
402
[45.] Wang, F., et al., The efficacy of oral versus intravenous tranexamic acid in reducing
403
blood loss after primary total knee and hip arthroplasty: A meta-analysis. Medicine
404
(Baltimore), 2018. 97(36): p. e12270. [46.] Holt, J.B., et al., Minimizing Blood Transfusion in Total Hip and Knee Arthroplasty
406
Through a Multimodal Approach. J Arthroplasty, 2016. 31(2): p. 378-82.
RI PT
405
[47.] Shohat, N., et al., Increased postoperative glucose variability is associated with adverse
408
outcomes following orthopaedic surgery. Bone Joint J, 2018. 100-B(8): p. 1125-1132.
409
SC
407
[48.] Lindberg-Larsen, V., et al., Preoperative High-Dose Methylprednisolone and Glycemic Control Early After Total Hip and Knee Arthroplasty: A Randomized, Double-Blind,
411
Placebo-Controlled Trial. Anesth Analg, 2018.
412 413
M AN U
410
[49.] Polderman, J.A., et al., Adverse side effects of dexamethasone in surgical patients. Cochrane Database Syst Rev, 2018. 8: p. CD011940.
AC C
EP
TE D
414
21
ACCEPTED MANUSCRIPT
Tables:
3.5 38 2 261 0.2 8.4 98 22 0.66 74 3.5 6.3 15 21 137 9 11.4 33.3
High* 5 126 12 280 1.3 10.5 107 30 1.25 106 5.1 8.2 46 72 145 20 16.1 46.5
AC C
EP
TE D
* to be classified as low or high, value must be below or above cut-off
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
RI PT
Albumin Alkaline Phosphatase Anion Gap Osmolality, Calculated Bilirubin, Total Calcium, Total Chloride Carbon Dioxide Creatinine Glucose Potassium Protein, Total Aspartate Aminotransferase Alanine Aminotransferase Sodium Blood Urea Nitrogen Hemoglobin Hematocrit
3
SC
Table 1. Laboratory Reference Ranges Low*
M AN U
1 2
ACCEPTED MANUSCRIPT
Table 2. Patient Demographics and Risk Factors
OME (mg)1; first 2 days
163.8 ± 122.1 [0, 2623]
Race Marital status CCI2
M AN U
HMO / Managed Care Medicaid Medicare Other Low Calcium High Creatinine High Glucose Low Hemoglobin Low Sodium
Insurance Type
24 25 26 27
Oral Morphine Equivalents, number of OMEs (in mg) consumed on POD 0 and POD 1 Charlson Comorbidity Index
EP
2
AC C
1
TE D
Laboratory Abnormality
Days of IV Narcotic Use
52.7% (673) 47.3% (605) 77.6% (992) 22.4% (286) 67.7% (865) 32.3% (413) 0.38 ± 0.73 (0, 3) 37.1% (472) 2.2% (28) 40.7% (518) 20.1% (256) 48.2% (616) 72.2% (923) 55.0% (703) 71.7% (916) 47.4% (606) 1.07 ± 0.54 [0, 4]
SC
Female Male White Non-White Married Not Married
Sex
RI PT
Age (years)
mean ± sd [min, max] % (n) 62.3 ± 10.7 [16, 94]
Characteristic
ACCEPTED MANUSCRIPT
Characteristic
LOS 0-1 days (n=548)
LOS 2-3 days (n=649)
LOS ≥3 days (n=81)
Hypotensive events 1 (POD 0/1)*
3.40 (3.60)
5.07 (5.07)
CCI 2
0.27 (0.62)
0.42 (0.73)
Female
219/548 (40.0%)
397/649 (61.1%)
Male
329/548 (60.0%)
252/649 (38.8%)
RI PT
Table 3. Univariable Logistic Regression for Predictors of Delayed Discharge
White
458/548 (83.6%)
484/649 (74.6%)
50/81 (61.7%)
0.518 (0.398, 0.673)
Non-White
90/548 (16.4%)
165/649 (25.4%)
31/81 (38.3%)
--
Married
426/548 (77.7%)
411/649 (63.3%)
28/81 (34.6%)
0.386 (0.304, 0.490)
<0.0001
238/649 (36.7%)
53/81 (65.4%)
63.6 (11.1)
67.1 (13.1)
1.451 (1.306, 1.613)
<0.0001
189.5 (140.9)
208.2 (157.7)
1.058 (1.045, 1.070)
<0.0001
198/646 (30.7%)
15/81 (18.5%)
0.924 (0.684, 1.248)
<0.0001
19/646 (2.9%)
4/81 (2.9%)
4.802 (2.147, 10.739)
316/646 (48.9%)
54/81 (66.7%)
2.878 (2.129, 3.890) --
Marital status
60.1 (9.2) 126.7 (71.9) HMO/Managed Care
Insurance
Laboratory Abnormalities
Medicaid Medicare
259/547 (47.4%) 5/547 (0.9%)
TE D
34
148/547 (27.1%)
5.74 (5.61)
1.526 (1.344, 1.733)
<0.0001
0.89 (1.07)
1.662 (1.425, 1.938)
<0.0001
57/81 (70.4%)
--
<0.0001
24/81 (29.6%)
0.409 (0.328, 0.509) <0.0001
Other
135/547 (24.7%)
113/646 (17.5%)
8/81 (9.9%)
Low Calcium
262/548 (52.2%)
312/646 (48.0%)
42/81 (51.8%)
1.044 (0.842, 1.293)
0.6959
High Creatinine
415/548 (75.7%)
442/649 (68.1%)
66/81 (81.5%)
0.817 (0.643, 1.038)
0.0981
High Glucose
278/548 (50.7%)
368/649 (56.7%)
75/81 (70.4%)
1.415 (1.140, 1.756)
0.0016
Low Hemoglobin
288/548 (52.6%)
554/649 (85.4%)
74/81 (91.4%)
5.522 (4.229, 7.211)
<0.0001
248/548 (45.3%)
311/649 (47.9%)
47/81 (58.0%)
1.207 (0.974, 1.496)
0.0854
Low Sodium
EP
OME (mg)*
122/548 (22.3%)
AC C
Age*
Not Married
3
P-value
SC
Race
M AN U
Sex
OR (95% CI)
* variables presented as mean ± sd 1 OR reported per 5 events. 2 Charlson Comorbidity Index 3 OR reported per 10 units (years or mg as applicable). 4 Oral Morphine Equivalents, number of OMEs (in mg) consumed on POD 0 and POD 1
ACCEPTED MANUSCRIPT
Table 4. Multivariate Analysis of Risk Factors Significantly Associated with Length of Stay Greater than 3 days OR (95% CI)
P-value
Hypotensive events on POD 0/1 * 1
1.232 (1.070, 1.419)
0.0037
RI PT
Characteristic
Race
Non-White Vs. White
1.497 (1.105, 2.028)
Marital status
Not Married Vs. Married
1.724 (1.307, 2.273)
<0.0001
1.330 (1.136, 1.556)
0.0004
Age* 2
Lab Abnormalities
<0.0001
Low Hemoglobin
3.265 (2.355, 4.526)
<0.0001
High Glucose
1.887 (1.325, 2.687)
High Creatinine
2.874 (1.821, 4.525)
3
1.411 (1.190, 1.419)
0.0004
<0.0001 <.0.001
M AN U
CCI
1.069 (1.054, 1.084)
SC
OME (mg)*
2
0.0093
AC C
EP
TE D
* variables presented as mean ± sd 1 OR reported per 5 additional events. 2 OR reported per 10 units (years or mg as applicable), Oral Morphine Equivalents. 3 OR reported per 1-unit increase. Charlson Comorbidity Index.
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
Figures
AC C
EP
TE D
Figure 1. Graph shows a histogram of LOS after THA. The dashed line is at 3 days and represents the cutoff used for statistical analysis.