384

C OPYRIGHT Ó 2019 BY T HE J OURNAL OF B ONE AND J OINT S URGERY, I NCORPORATED

Prediction of Complications, Readmission, and

Revision Surgery Based on Duration of Preoperative
Opioid Use
Analysis of Major Joint Replacement and Lumbar Fusion
Nikhil Jain, MD, John L. Brock, BA, Azeem Tariq Malik, MBBS, Frank M. Phillips, MD, and Safdar N. Khan, MD

Investigation performed at The Ohio State University Wexner Medical Center, Columbus, Ohio

Background: Preoperative opioid use results in adverse outcomes and higher costs after elective surgery. However,
duration thresholds for higher risk are not entirely known. Therefore, the purpose of our study was to determine the number
and duration of preoperative opioid prescriptions in order to estimate the risk of postoperative adverse events after major
joint replacement and lumbar fusion.
Methods: National insurance claims data (2007 to September 30, 2015) were used to identify primary total knee
arthroplasties (TKAs), total hip arthroplasties (THAs), and 1 or 2-level posterior lumbar fusions (PLFs) performed for
degenerative disease. The effect of preoperative opioid burden (naive, £3 months, >3 to 6 months, >6 months but
stopped 3 months before surgery, and >6 months of continuous use) on the risks of various adverse outcomes was
studied using Cox proportional hazards analysis with adjustment for demographic and clinical covariates.
Results: A total of 58,082 patients stratified into 3 cohorts of 32,667 with TKA, 14,734 with THA, and 10,681 with 1 or
2-level PLF were included for this analysis. A duration of preoperative opioids of >3 months was associated with a higher
risk of 90-day emergency department (ED) visits for all causes and readmission after TKA. Preoperative opioid prescription
for >6 months was associated with a higher risk of all-cause and pain-related ED visits, wound dehiscence/infection, and
hospital readmission within 90 days as well as revision surgery within 1 year after TKA, THA, and PLF. Stopping the opioid
prescription 3 months preoperatively for chronic users resulted in a significant reduction in the risk of adverse outcomes,
with the greatest impact seen after THA and PLF.
Conclusions: Patients with a preoperative opioid prescription for up to 3 months before a major arthroplasty or a 1 or 2-
level lumbar fusion had a similar risk of adverse outcomes as opioid-naive patients. While >6 months of opioid use was
associated with a higher risk of adverse outcomes, a 3-month prescription-free period before the surgery appeared to
mitigate this risk for chronic users.
Level of Evidence: Therapeutic Level II. See Instructions for Authors for a complete description of levels of evidence.

T
he opioid epidemic in the United States has serious
consequences for both the health of Americans and the
nation’s economy. The financial burden of the opioid
epidemic was estimated to be $504 billion per year by the
White House Council of Economic Advisers, which far exceeds
previous estimates1-5. Prescription for orthopaedic conditions
constitutes one of the highest burdens of opioid consumption,
with a 660% rise in prescriptions for back pain from 1997 to
2006 and an estimated 40% of Medicare beneficiaries receiving
at least one opioid prescription for knee osteoarthritis in
20096-8.
The use of opioids is associated with cognitive dysfunction,
respiratory depression, impaired mobility, immunosuppression,
delayed wound-healing, dependence, and hyperalgesia9. These
negative effects directly and indirectly lead to adverse patient
health, especially in those undergoing major surgery. Prior lit-
erature has shown that preoperative opioid use results in poor
pain outcomes, complications, higher costs, delays in return to

Disclosure: The authors received no funding for this study. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online
version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena
outside the submitted work (http://links.lww.com/JBJS/F57).

J Bone Joint Surg Am. 2019;101:384-91 d http://dx.doi.org/10.2106/JBJS.18.00502

 385
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
V O L U M E 101-A N U M B E R 5 M A R C H 6, 2 019
d d
work, early revision, and a risk of chronic postoperative opioid
use after spinal fusion10-22 and major joint arthroplasty8,23-31.
While these adverse outcomes have been extensively studied,
the duration of preoperative opioid use that constitutes a
higher risk is not known. Additionally, varying definitions
and durations of opioid use in published studies preclude
meaningful correlation 8,10-18,21-31.
We used a single national database to identify duration
thresholds of preoperative opioid use for known adverse
outcomes after primary total knee arthroplasty (TKA), total
hip arthroplasty (THA), and 1 or 2-level posterior lumbar
fusion (PLF) performed to treat degenerative disease. While
some authors used total morphine equivalents to study the
association between opioid burden and risk14-16,24,28,30,31, this
calculation is not routinely practical. Therefore, the purpose
of our study was to estimate the number and duration of
preoperative opioid prescriptions that create a risk of adverse
events postoperatively. We also estimated the benefit of
stopping opioid prescription 3 months preoperatively for
chronic preoperative users. Such data will be beneficial for
providers in their preoperative decision-making and patient
counseling, especially as we work toward improving the quality
and value of care.
Materials and Methods
Data Source
T he Humana national claims database (2007 to September
30, 2016) covering 22 million commercial insurance and
Medicare Advantage beneficiaries was used for analysis. With
Fig. 1
Algorithm for study inclusion and exclusion of patients with codes for primary TKA, THA, and 1 or 2-level PLF in the Humana database (see Appendix
Table E-1 for code descriptions).
C O M P L I C AT I O N S , R E A D M I S S I O N , A N D R E V I S I O N S U R G E RY V E R S U S
D U R AT I O N O F P R E O P E R AT I V E O P I O I D U S E
this database, longitudinal research can be performed over the
full data set using 1 or a combination of identifiable fields on
claims records. These include, but are not limited to, Inter-
national Classification of Diseases, Ninth and Tenth Revisions
(ICD-9 and 10) diagnosis and procedural coding, Diagnosis-
Related Groups (DRGs), Current Procedural Terminology (CPT)
codes, generic drug codes, prescription National Drug Codes
(NDCs), discharge status, and physician specialty. Data are
de-identified and HIPAA (Health Insurance Portability and
Accountability Act, 1996)-compliant and were accessed
through the PearlDiver Technologies research platform32.
Data Extraction
Separate cohorts of primary TKAs, THAs, and 1 or 2-level PLFs
for degenerative disease were created using the first occurrence
of the respective ICD and CPT codes in the Humana national
claims database from 2007 to September 30, 2015. Procedures
were queried up to September 30, 2015 to allow analysis of 1-
year revision surgery until September 30, 2016. We included
only unilateral TKA and THA procedures, using CPT modi-
fiers, and instrumented PLFs with or without an interbody cage
for degenerative disease. Patients with a previous spinal fusion
or lower-extremity arthroplasty or a diagnosis of spine or
lower-extremity fracture, cervical spondylosis, rheumatoid
arthritis, scoliosis, or paraplegia or quadriplegia were excluded
from all 3 cohorts to reduce confounding on opioid use and
complications and readmissions. The algorithm for inclusion
and exclusion of patients is given in Figure 1 and Appendix
Table E-1.
 386
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
C O M P L I C AT I O N S , R E A D M I S S I O N , A N D R E V I S I O N S U R G E RY V E R S U S
V O L U M E 101-A N U M B E R 5 M A R C H 6, 2 019
d d
D U R AT I O N O F P R E O P E R AT I V E O P I O I D U S E

Opioid Use (CCI)34 were used for adjustment. In addition, a diagnosis of

Longitudinal tracking of generic formulations of Schedule-II and
III opioids taken by the oral route were used to record preoperative
opioid use. These opioids included hydrocodone/acetaminophen,
oxycodone/acetaminophen, oxycodone, hydrocodone, hydroco-
done/ibuprofen, fentanyl, hydromorphone, meperidine, mor-
phine sulfate, and acetaminophen/codeine 33. Patients were
categorized according to the duration of preoperative opioid
use by identifying prescriptions in each preceding 6-week interval
starting on the day before the surgery. The categories were defined
as £3 months, >3 to 6 months, and >6 months of opioid pre-
scription immediately before the surgery. Additionally, we iden-
tified an opioid-naive group consisting of patients who never
received an opioid prescription before the surgery. To study the
effect of stopping opioid prescription before the surgery, a sepa-
rate category of patients with >6 months of opioid use but no
prescription in the 3 months immediately before the surgery was
created (Table I). The assumption for this group was that the last
refill being ‡3 months before the surgery would result in either
some opioid-free period or less frequent consumption, mimick-
ing a weaning protocol before surgery.
Outcome Measures
We chose specific adverse outcomes that are related to the effect
of opioids on pain modulation, mobility, and immune func-
tion9. All-cause and pain-related emergency department (ED)
visits, falls, deep vein thrombosis and pulmonary embolism
(DVT/PE), wound dehiscence and infection, systemic infection
and sepsis, and hospital readmission, all within 90 days after the
index surgery, were studied using CPT or ICD codes. Revision
surgery within 1 year after the index procedure was also studied
using ICD procedure codes (see Appendix Table E-2).
Data Analysis
Descriptive analyses of quantitative variables are reported as
the frequencies and percentages, and quantitative variables are
described as means and standard deviations (SDs). The signifi-
cance of the duration of preoperative opioid prescription as a risk
factor was analyzed with a univariate log-rank test as well as with
a multivariate Cox proportional hazards analysis with adjustment
for various demographic and clinical variables. Demographics
such as age, sex, ethnicity, insurance type, and clinical comor-
bidity burden according to the Charlson Comorbidity Index
anxiety, depression, tobacco use, and obesity (as defined in the
Medicare Research Chronic Conditions Data Warehouse)35
within 1 year before the index surgery was included in the
multivariate model as these factors have been shown to be sig-
nificant confounders for opioid use and our outcome mea-
sures3,9,14,17,21,23,24,30,36. Adjustment for patients whose preoperative
opioid use was either intermittent or could not be classified ac-
cording to our definitions was also done. For all adverse events,
we report the adjusted hazard ratio (HR) with the 95% confi-
dence interval (CI) for each preoperative opioid-use group, with
the opioid-naive patients serving as the reference. The signifi-
cance of each Cox model was judged by significant p values on
likelihood ratio, Wald, and score log-rank tests. For all analyses, a
p value of <0.05 was considered significant. Statistical analysis
was carried out in R (The R Project for Statistical Computing)
through the PearlDiver interface.
Results
A total of 58,082 patients stratified into 3 cohorts of 32,667
with TKA, 14,734 with THA, and 10,681 with 1 or 2-level
PLF were included for this analysis. Demographic and clinical
profiles of the patients are given in Table II. The patterns of
opioid use varied according to the type of procedure, with a
higher proportion of patients undergoing PLF using opioids for
>6 months preoperatively (28.3%) compared with the THA
(15.0%) and TKA (14.7%) groups. Similarly, only 23.4% of the
patients undergoing PLF were opioid-naive as compared with
40.9% of the patients undergoing TKA and 40.3% of those
undergoing THA (Table III).
Univariate analysis was performed to assess 90-day com-
plications and readmissions and 1-year revision after TKA, THA,
and 1 or 2-level PLF (Table IV) as related to the duration of
preoperative opioid use. There was a general trend for increasing
incidence of adverse outcomes with increasing duration of
preoperative use.
ED Visits
After TKA, the adjusted risk of ED visits (for any cause) within
90 days was significantly increased in patients with >3 months
of preoperative opioid use; the threshold for a higher risk of a
pain-related ED visit was >6 months of opioid use. After THA
and PLF, the adjusted risk for all-cause and pain-related ED

TABLE I Methodology for Classifying Duration of Preoperative Opioid Use Before Primary TKA, THA, and 1 or 2-Level PLF in Humana
Database

Preoperative Opioid Duration Definition

Opioid-naive No opioid prescription at any time before surgery

£3 mo ‡2 opioid prescriptions within immediate 3 mo before surgery and no prior prescriptions
>3-6 mo ‡4 opioid prescriptions within immediate 6 mo before surgery and no prior prescriptions
>6 mo, stopped 3 mo before No opioid prescription within 3 mo before surgery and ‡6 opioid prescriptions in preceding >6-month period
surgery
>6 mo, continuous ‡4 opioid prescriptions within immediate 6 mo before surgery and ‡2 prescriptions immediately before 6 mo
 387
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
C O M P L I C AT I O N S , R E A D M I S S I O N , A N D R E V I S I O N S U R G E RY V E R S U S
V O L U M E 101-A N U M B E R 5 M A R C H 6, 2 019
d d
D U R AT I O N O F P R E O P E R AT I V E O P I O I D U S E

TABLE II Demographic and Clinical Profile of Patients in Humana Database Who Underwent Primary TKA, THA, or 1 or 2-Level PLF

No. (%)
TKA THA 1 or 2-Level PLF

Overall 32,667 (100) 14,734 (100) 10,681 (100)

Age
30-34 yr 1 (<0.1) 10 (0.1) 67 (0.6)
35-39 yr 12 (<0.1) 27 (0.2) 126 (1.2)
40-44 yr 88 (0.3) 78 (0.5) 221 (2.1)
45-49 yr 296 (0.9) 224 (1.5) 399 (3.7)
50-54 yr 814 (2.5) 454 (3.1) 660 (6.2)
55-59 yr 1,730 (5.3) 817 (5.5) 928 (8.7)
60-64 yr 2,616 (8.0) 1,181 (8.0) 1,133 (10.6)
65-69 yr 7,338 (22.5) 3,115 (21.1) 2,452 (23.0)
70-74 yr 9,270 (28.4) 3,869 (26.3) 2,490 (23.3)
75-79 yr 6,445 (19.7) 2,820 (19.1) 1,493 (14.0)
80-84 yr 3,323 (10.2) 1,715 (11.6) 628 (5.9)
85-89 yr 734 (2.2) 424 (2.9) 84 (0.8)
Sex
Female 19,814 (60.7) 7,992 (54.2) 5,942 (55.6)
Male 12,853 (39.3) 6,742 (45.8) 4,739 (44.4)
Ethnicity
White 26,673 (81.7) 11,862 (80.5) 8,311 (77.8)
African-American 2,532 (7.8) 951 (6.5) 766 (7.2)
Other (Hispanic, Asian, Native American) 3,462 (10.6) 1,921 (13.0) 1,604 (15.0)
Charlson Comorbidity Index
0 9,038 (27.7) 4,668 (31.7) 2,601 (24.4)
1 7,386 (22.6) 3,321 (22.5) 2,470 (23.1)
2 5,364 (16.4) 2,300 (15.6) 1,742 (16.3)
3 3,957 (12.1) 1,607 (10.9) 1,325 (12.4)
4 2,438 (7.5) 975 (6.6) 839 (7.9)
>4 4,484 (13.7) 1,863 (12.6) 1,704 (16.0)
Mean ± SD 2.14 ± 2.33 2.0 ± 2.35 2.34 ± 2.46
Anxiety disorder 3,984 (12.2) 1,568 (10.6) 2,041 (19.1)
Depression 5,104 (15.6) 1,974 (13.4) 2,442 (22.9)
Tobacco use disorder 1,645 (5.0) 1,075 (7.3) 1,344 (12.6)
Obesity 7,150 (21.9) 2,379 (16.1) 1,825 (17.1)

visits was significantly higher with >6 months of continuous DVT/PE

preoperative opioid use; the risk of all-cause and pain-related
ED visits by patients with >6 months of preoperative opioid use
but without prescription within 3 months before the surgery
was similar to that for opioid-naive patients (Table V).
Falls
The adjusted risk of a fall within 90 days after TKA was sig-
nificantly higher for the patients who had used opioids for >6
months preoperatively (with or without stopping 3 months
prior to surgery) than it was for the opioid-naive patients. The
association between opioid use and the risk of a fall was not
significant in the THA and PLF groups (Table V).
The association between preoperative opioid use and DVT/PE was
not significant for the patients undergoing TKA or THA. It was also
not significant on univariate analysis of those undergoing PLF, but
>6 months of continuous preoperative opioid use was a significant
risk factor in that group on multivariate Cox analysis (Table V).
Wound Dehiscence and Infection
Continuous preoperative opioid use for >6 months was sig-
nificantly associated with a higher risk of 90-day wound
dehiscence and infection after TKA, THA, or PLF on adjusted
Cox analysis. However, patients who used opioids for >6
months but had no opioid prescription within 3 months before
 388
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
C O M P L I C AT I O N S , R E A D M I S S I O N , A N D R E V I S I O N S U R G E RY V E R S U S
V O L U M E 101-A N U M B E R 5 M A R C H 6, 2 019
d d
D U R AT I O N O F P R E O P E R AT I V E O P I O I D U S E

TABLE III Duration of Preoperative Opioid Use Before Primary TKA, THA, and 1 or 2-Level PLF in Humana Database

No. (%)
Preoperative Opioid Use TKA THA 1 or 2-Level PLF P Value

Opioid-naive 13,359 (40.9) 5,942 (40.3) 2,495 (23.4) <0.001

£3 mo 405 (1.2) 440 (3.0) 429 (4.0) <0.001
>3-6 mo 213 (0.7) 247 (1.7) 313 (2.9) <0.001
>6 mo, stopped 3 mo before surgery 1,894 (5.8) 619 (4.2) 657 (6.2) <0.001
>6 mo, continuous 2,892 (8.9) 1,594 (10.8) 2,361 (22.1) <0.001

TKA, THA, or PLF did not show a higher risk of wound com-
plications (Table V).
Systemic Infection and Sepsis
The adjusted risk of 90-day systemic infection or sepsis was
significantly higher in the patients with >6 months of contin-
uous preoperative opioid use in the TKA and THA groups
(Table V).
Readmission
On adjusted Cox analysis, >3 months of opioid use before TKA
and >6 months of continuous opioid use before THA and PLF
were significantly associated with 90-day hospital readmission.
Patients who used opioids for 6 months preoperatively but did
not have a prescription within 3 months before THA or PLF
had a similar risk of readmission as opioid-naive patients in
those groups (Table V).
Revision Surgery
More than 6 months of preoperative opioid use (with or
without stopping for 3 months before the surgery) was a sig-
nificant risk factor for revision surgery within 1 year after TKA
on adjusted analysis. Whereas >6 months of continuous pre-
operative opioid use was a risk factor for revision surgery after

TABLE IV Cumulative Incidence and Log-Rank P Value for 90-Day Complications and Readmissions and 1-Year Revision Surgery According
to Duration of Preoperative Opioid Use in Humana Database

No. (%) [95% CI]

>6 Mo, Stopped 3 Mo

Category Naive £3 Mo >3-6 Mo Before Surgery >6 Mo, Continuous P Value

TKA 13,359 405 213 1,894 2,892

All-cause ED visits 1,668 (12.5) [11.9-13.0] 60 (14.8) [11.3-18.2] 40 (18.8) [13.4-23.9] 420 (22.2) [20.3-24.0] 595 (20.6) [19.1–22.0] <0.001
Pain-related ED visits 544 (4.1) [3.7-4.4] 16 (4.0) [2.0-5.8] 16 (7.5) [3.9-11.0] 171 (9.0) [7.7-10.3] 265 (9.2) [8.1-10.2] <0.001
Fall 376 (2.8) [2.5-3.1] 13 (3.2) [1.5-4.9] 6 (2.8) [0.6-5.0] 97 (5.1) [4.1-6.1] 136 (4.7) [3.9-5.5] <0.001
DVT/PE 629 (4.7) [4.3-5.1] 15 (3.7) [1.8-5.5] 13 (6.1) [2.8-9.3] 102 (5.4) [4.4-6.4] 158 (5.5) [4.6-6.3] 0.21
Wound dehiscence and infection 841 (6.3) [5.9-6.7] 27 (6.7) [4.2-9.1] 21 (9.9) [5.8-13.8] 163 (8.6) [7.3-9.9] 287 (9.9) [8.8-11.0] <0.001
Systemic infection and sepsis 589 (4.4) [4.1-4.8] 16 (4.0) [2.0-5.8] 12 (5.6) [2.5-8.7] 143 (7.6) [6.4-8.7] 197 (6.8) [5.9-7.7] <0.001
Readmission 926 (6.9) [6.5-7.4] 33 (8.1) [5.4-10.8] 24 (11.3) [6.9-15.4] 198 (10.5) [9.1-11.8] 276 (9.5) [8.5-10.6] <0.001
Revision surgery 147 (1.1) [0.9-1.3] 4 (1.0) [0.7-1.9] 4 (1.9) [1.0-3.7] 46 (2.4) [1.7-3.1] 68 (2.4) [1.8-2.9] <0.001

THA 5,942 440 247 619 1,594

All-cause ED visits 678 (11.4) [10.6-12.2] 46 (10.5) [7.6-13.3] 36 (14.6) [10.1-18.9] 89 (14.4) [11.6-17.1] 323 (20.3) [18.3-22.2] <0.001
Pain-related ED visits 210 (3.5) [3.1-4.0] 15 (3.4) [1.7-5.1] 9 (3.6) [1.3-6.0] 32 (5.2) [3.4-6.9] 138 (8.7) [7.3-10.0] <0.001
Fall 203 (3.4) [3.0-3.9] 16 (3.6) [1.9-5.4] 6 (2.4) [0.5-4.3] 27 (4.4) [2.7-6.0] 65 (4.1) [3.1-5.0] 0.46
DVT/PE 216 (3.6) [3.2-4.1] 19 (4.3) [2.4-6.2] 11 (4.5) [1.8-7.0] 27 (4.4) [2.7-6.0] 78 (4.9) [3.8-5.9] 0.22
Wound dehiscence and infection 233 (3.9) [3.4-4.4] 24 (5.5) [3.3-7.6] 16 (6.5) [3.4-9.5] 38 (6.1) [4.2-8.0] 125 (7.8) [6.5-9.2] <0.001
Systemic infection and sepsis 234 (3.9) [3.4-4.4] 25 (5.7) [3.5-7.8] 13 (5.3) [2.4-8.0] 39 (6.3) [4.4-8.2] 105 (6.6) [5.4-7.8] <0.001
Readmission 405 (6.8) [6.2-7.5] 32 (7.3) [4.8-9.7] 21 (8.5) [5.0-11.9] 57 (9.2) [6.9-11.5] 183 (11.5) [9.9-13.0] <0.001
Revision surgery 71 (1.2) [0.9-1.5] 6 (1.4) [0.3-2.4] 0 (0) 15 (2.4) [1.2-3.6] 41 (2.6) [1.8-3.3] <0.001

PLF 2,495 429 313 657 2,361

All-cause ED visits 395 (15.8) [14.4-17.3] 54 (12.6) [9.4-15.7] 52 (16.6) [12.4-20.6] 138 (21.0) [17.8-24.1] 551 (23.3) [21.6-25.0] <0.001
Pain-related ED visits 177 (7.1) [6.1-8.1] 22 (5.1) [3.0-7.2] 28 (8.9) [5.7-12.1] 69 (10.5) [8.1-12.8] 304 (12.9) [11.5-14.2] <0.001
Fall 73 (2.9) [2.3-3.6] 8 (1.9) [0.6-3.1] 7 (2.2) [0.6-3.9] 31 (4.7) [3.1-6.3] 83 (3.5) [2.8-4.3] 0.05
DVT/PE 56 (2.2) [1.7-2.8] 12 (2.8) [1.2-4.3] 7 (2.2) [0.6-3.9] 22 (3.3) [2.0-4.7] 74 (3.1) [2.4-3.8] 0.29
Wound dehiscence and infection 127 (5.1) [4.2-6.0] 24 (5.6) [3.4-7.7] 18 (5.8) [3.1-8.3] 50 (7.6) [5.6-9.6] 185 (7.8) [6.8-8.9] 0.001
Systemic infection and sepsis 160 (6.4) [5.4-7.4] 29 (6.8) [4.4-9.1] 26 (8.3) [5.2-11.3] 49 (7.5) [5.4-9.4] 175 (7.4) [6.4-8.5] 0.57
Readmission 183 (7.3) [6.3-8.4] 35 (8.2) [5.5-10.7] 33 (10.5) [7.1-13.9] 55 (8.4) [6.2-10.5] 270 (11.4) [10.2-12.7] <0.001
Revision surgery 59 (2.4) [1.8-3.0] 9 (2.1) [0.7-3.4] 5 (1.6) [0.2-3.0] 24 (3.7) [2.2-5.1] 86 (3.6) [2.9-4.4] 0.02
 389
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
C O M P L I C AT I O N S , R E A D M I S S I O N , A N D R E V I S I O N S U R G E RY V E R S U S
V O L U M E 101-A N U M B E R 5 M A R C H 6, 2 019
d d
D U R AT I O N O F P R E O P E R AT I V E O P I O I D U S E

TABLE V Cox Hazards for 90-Day Complications and Readmissions and 1-Year Revision Surgery According to Duration of Preoperative
Opioid Use in Humana Database*

Risk of Adverse Events: HR* (95% CI); P Value

>6 Mo, Stopped 3 Mo

Category £3 Mo >3-6 Mo Before Surgery >6 Mo, Continuous

TKA
All-cause ED visits 1.06 (0.81-1.37); 0.68 1.41 (1.03-1.94); 0.03 1.59 (1.43-1.77); <0.001 1.43 (1.30-1.58); <0.001
Pain-related ED visits 0.79 (0.48-1.31); 0.36 1.64 (0.99-2.72); 0.05 1.75 (1.47-2.08); <0.001 1.61 (1.38-1.88); <0.001
Fall 1.11 (0.63-1.94); 0.72 0.94 (0.41-2.12); 0.88 1.66 (1.33-2.08); <0.001 1.52 (1.24-1.87); <0.001
DVT/PE 0.73 (0.43-1.22); 0.23 1.37 (0.79-2.38); 0.27 1.05 (0.85-1.30); 0.64 1.12 (0.93-1.35); 0.22
Wound dehiscence and infection 0.95 (0.64-1.40); 0.80 1.52 (0.98-2.36); 0.06 1.18 (0.99-1.39); 0.06 1.35 (1.17-1.55); <0.001
Systemic infection and sepsis 0.81 (0.49-1.33); 0.40 1.28 (0.72-2.28); 0.40 1.50 (1.24-1.80); <0.001 1.41 (1.19-1.67); 0.01
Readmission 1.05 (0.74-1.50); 0.77 1.54 (1.02-2.33); 0.04 1.33 (1.14-1.55); 0.003 1.20 (1.04-1.38); 0.01
Revision surgery 0.73 (0.27-1.99); 0.54 1.44 (0.53-3.93); 0.48 1.69 (1.21-2.36); 0.002 1.40 (1.03-1.89); 0.03
THA
All-cause ED visits 0.81 (0.60-1.10); 0.18 1.29 (0.91-1.82); 0.14 1.03 (0.83-1.29); 0.77 1.49 (1.30-1.71); <0.001
Pain-related ED visits 0.88 (0.52-1.50); 0.64 0.94 (0.47-1.86); 0.86 1.13 (0.78-1.65); 0.51 1.85 (1.48-2.32); <0.001
Fall 1.08 (0.64-1.81); 0.77 0.64 (0.28-1.46); 0.28 1.05 (0.70-1.57); 0.83 0.97 (0.73-1.31); 0.87
DVT/PE 1.09 (0.67-1.76); 0.73 1.16 (0.62-2.16); 0.64 1.04 (0.69-1.56); 0.84 1.23 (0.94-1.61); 0.13
Wound dehiscence and infection 1.16 (0.75-1.80); 0.49 1.36 (0.80-2.29); 0.25 1.19 (0.84-1.68); 0.32 1.45 (1.16-1.82); 0.001
Systemic infection and sepsis 1.26 (0.83-1.93); 0.28 1.05 (0.59-1.87); 0.86 1.22 (0.87-1.72); 0.25 1.29 (1.02-1.64); 0.04
Readmission 0.97 (0.67-1.40); 0.85 1.16 (0.74-1.83); 0.51 1.10 (0.83-1.45); 0.51 1.38 (1.15-1.65); 0.001
Revision surgery 1.26 (0.55-2.89); 0.59 <0.001 (0-infinity); 0.99 1.67 (0.95-2.92); 0.07 1.58 (1.06-2.37); 0.03
PLF
All-cause ED visits 0.76 (0.57-1.01); 0.06 1.10 (0.82-1.46); 0.55 1.18 (0.97-1.43); 0.09 1.30 (1.13-1.47); <0.001
Pain-related ED visits 0.66 (0.42-1.03); 0.07 1.38 (0.92-2.06); 0.12 1.32 (1.00-1.74); 0.05 1.54 (1.28-1.86); <0.001
Fall 0.66 (0.31-1.37); 0.26 0.83 (0.38-1.81); 0.64 1.37 (0.91-2.08); 0.13 1.05 (0.77-1.44); 0.75
DVT/PE 1.31 (0.70-2.46); 0.40 0.90 (0.40-1.98); 0.79 1.34 (0.83-2.18); 0.23 1.43 (1.02-2.02); 0.04
Wound dehiscence and infection 1.04 (0.67-1.62); 0.84 1.03 (0.63-1.69); 0.91 1.32 (0.96-1.82); 0.09 1.30 (1.04-1.63); 0.02
Systemic infection and sepsis 1.00 (0.68-1.51); 0.96 1.34 (0.88-2.04); 0.18 1.07 (0.78-1.48); 0.66 1.13 (0.91-1.40); 0.28
Readmission 1.02 (0.71-1.47); 0.91 1.41 (0.97-2.05); 0.07 0.99 (0.74-1.34); 0.96 1.44 (1.19-1.74); <0.001
Revision surgery 1.02 (0.50-2.07); 0.96 0.68 (0.27-1.72); 0.42 1.50 (0.93-2.40); 0.10 1.42 (1.01-2.00); 0.04

*With the opioid-naive group as the reference.

THA and PLF, patients with no prescription within 3 months Definitions of opioid use before arthroplasty or spinal
before the surgery did not show a higher risk than opioid-naive fusion have varied in prior studies8,11-16,18,23-31,37-39. In a study of
patients (Table V). 17,695 THA patients in the Humana database, Bedard et al.23
defined preoperative opioid use as filling an opioid prescription
Discussion within 3 months before surgery; 36.7% of their patients met

W e sought to determine the risk of postoperative adverse

events according to the duration of opioid use before
commonly performed major elective orthopaedic surgery. We
found that considerable proportions of patients had used
opioids for >6 months before THA, TKA, and 1 or 2-level PLF
(15.0%, 14.7%, and 28.3%, respectively). Preoperative opioid
prescription for >6 months was associated with higher risks of
all-cause and pain-related ED visits, wound dehiscence/infec-
tion, and readmission within 90 days, as well as revision surgery
within 1 year, after TKA, THA, and PLF. There appears to be
significant potential benefit in stopping opioid prescription for
chronic users before surgery. For patients who used opioids for
>6 months before THA or PLF but did not have a prescription
for 3 months preoperatively, the risk of adverse events was
similar to that for opioid-naive patients.
those criteria and were more likely to require revision surgery
within 2 years. Cancienne et al.26 used the same database to study
113,337 patients who underwent TKA and defined preoperative
opioid use as filling an opioid prescription between 4 months and
1 month before surgery. They found that preoperative opioid
users had an increased risk of ED visits within 30 days postop-
eratively and readmissions, joint infection, stiffness, and revision
surgery within 1 year. Although these authors used a 3-month
period to identify opioid prescription, they did not quantify the
overall opioid burden or duration. While greater (unmeasured)
opioid burden was responsible for their findings, we filtered the
number and timing of prescriptions and defined various dura-
tions of preoperative opioid use to study the adjusted risk of
adverse outcomes. Authors of future studies should aim to dis-
tinguish between various durations of preoperative use as this will
 390
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
V O L U M E 101-A N U M B E R 5 M A R C H 6, 2 019
d d
be important for informed risk assessment and preoperative
counseling of patients.
Our results indicated a differential effect of preoperative
opioid use on adverse outcomes after TKA, THA, and PLF. While
>3 to 6 months of opioid use was the threshold for a higher risk
of ED visits and readmission after TKA, ‡6 prescriptions over >6
months immediately before the surgery increased this risk after
THA and PLF. More than 6 months of continuous use was
associated with a higher risk of wound complications and revi-
sion surgery after all 3 procedures. It is possible that prolonged
preoperative opioid use reflects a higher severity of disease and
therefore higher risk. Although the severity and duration of
disease are not discernable from claims data, our finding that
stopping opioid prescriptions for chronic users reduced the risk
of adverse outcomes implicates the role of opioids independent
from disease severity and other factors.
A detailed pathophysiologic description of the effect of
opioids is beyond the scope of this report; however, opioids are
known to cause hyperalgesia by nociceptive modulation, increase
the risk of falls and fractures due to cognitive impairment and
osteoporosis, increase the risk of wound complications by in-
hibiting angiogenesis and macrophage recruitment, and increase
the risk of infection/sepsis through immunosuppression9,40.
Although these pathophysiologic effects of opioids are not readily
apparent clinically, we found that ‡6 preoperative opioid pre-
scriptions over >6 months were associated with increases in the
risk of falls after TKA, DVT/PE after PLF, and systemic infection
after TKA and THA.
The effect of stopping opioid prescription within 3 months
before surgery for chronic users was not uniform. In the THA
and PLF groups, we found that the risk of all adverse outcomes
was no longer significant whereas all adverse outcomes except
wound complications and DVT/PE remained significantly higher
after TKA. We were unable to hypothesize a reason for such
conflicting findings from our analysis. However, a benefit of
preoperative opioid reduction before TKA and THA was shown
previously by Nguyen et al.41. In their study, patients who
decreased their opioid burden by at least 50% had outcomes
comparable with those for an opioid-naive group, with both
groups demonstrating significantly improved pain and functional
outcomes compared with patients who had not reduced their
opioid use preoperatively. Recently, the authors of another report
highlighted their institutional protocol of a 4 to 6-week opioid-
free period with routine drug testing before all elective joint
arthroplasty42. Because of the adverse consequences of abrupt
cessation by chronic opioid users, comprehensive opioid-
weaning programs have been suggested to minimize risk and
maximize outcomes following major elective surgery43,44.
Our analysis has some limitations. It is a retrospective
analysis of administrative claims data, which are dependent on
coding methodology and accuracy. This may lead to underes-
timation or overestimation of study measures. There may be
sources of opioids outside of active insurance that we were
unable to account for. Our definitions of duration of preop-
erative opioid use do not indicate actual days of opioid con-
sumption but rather the duration over which they were
C O M P L I C AT I O N S , R E A D M I S S I O N , A N D R E V I S I O N S U R G E RY V E R S U S
D U R AT I O N O F P R E O P E R AT I V E O P I O I D U S E
prescribed, although this may be more practically relevant.
Also, although we incorporated relevant demographic and
clinical variables in our Cox model, there may have been bias
due to unmeasured variables such as type and dose of opioids,
severity of disease, radiographic findings, operative time, and
blood loss, as these may impact the risk of adverse outcomes.
This study was a comparative risk analysis of 3 major
elective orthopaedic procedures in the same national population
with uniform inclusion criteria and definition of variables. The
ethical and implementation challenges in determining the risk of
opioid use/reduction in prospective clinical studies highlights the
importance of retrospective analysis such as ours43. With respect to
opioid use before major arthroplasties and 1 or 2-level lumbar
fusion, prescription for up to 3 months prior to the procedure
conferred a risk of adverse outcomes that was similar to that for
opioid-naive patients. While >6 months of opioid use was asso-
ciated with a higher risk of ED visits, wound complications, re-
admission, and revision surgery, a 3-month prescription-free
period before surgery appears to mitigate that risk for chronic
users. Surgeons should emphasize the downstream effects of
chronic opioid use at the outset of nonoperative management of
degenerative musculoskeletal disease. However, like efforts toward
smoking cessation and optimization of nutrition, diabetes man-
agement, and weight, efforts to wean established opioid users off
opioids before elective surgery have the potential to reduce the risk
of adverse outcomes and associated additional health-care costs.
Appendix
Tables showing codes and definitions used to include and
exclude patients in the primary TKA, THA, and PLF
cohorts and ICD-9/ICD-10 and CPT codes used to identify
primary outcome measures in the Humana database are
available with the online version of this article as a data sup-
plement at jbjs.org (http://links.lww.com/JBJS/F58). n
Nikhil Jain, MD1
John L. Brock, BA2
Azeem Tariq Malik, MBBS1
Frank M. Phillips, MD3
Safdar N. Khan, MD1
1The Ohio State University Wexner Medical Center, Columbus, Ohio
2Perelman School of Medicine, University of Pennsylvania, Philadelphia,
Pennsylvania
3Midwest Orthopaedics at Rush, Rush University Medical Center, Chicago,
Illinois
E-mail address for S.N. Khan: Safdar.Khan@osumc.edu
ORCID iD for N. Jain: 0000-0002-9091-6156
ORCID iD for J.L. Brock: 0000-0001-6295-6575
ORCID iD for A.T. Malik: 0000-0002-8776-9749
ORCID iD for F.M. Phillips: 0000-0003-3546-6073
ORCID iD for S.N. Khan: 0000-0002-7879-8177
 391
TH E JO U R NA L O F B O N E & JO I N T SU RG E RY J B J S . O RG
d
V O L U M E 101-A N U M B E R 5 M A R C H 6, 2 019
d d
References
1. Manchikanti L, Helm S 2nd, Fellows B, Janata JW, Pampati V, Grider JS, Boswell
MV. Opioid epidemic in the United States. Pain Physician. 2012 Jul;15(3)(Suppl):
ES9-38.
2. Council of Economic Advisors. The underestimated cost of the opioid crisis. 2017
Nov. https://www.whitehouse.gov/sites/whitehouse.gov/files/images/The%
20Underestimated%20Cost%20of%20the%20Opioid%20Crisis.pdf. Accessed
2018 Mar 20.
3. Oderda GM, Lake J, Rüdell K, Roland CL, Masters ET. Economic burden of pre-
scription opioid misuse and abuse: a systematic review. J Pain Palliat Care Phar-
macother. 2015;29(4):388-400.
4. Kirson NY, Scarpati LM, Enloe CJ, Dincer AP, Birnbaum HG, Mayne TJ. The eco-
nomic burden of opioid abuse: updated findings. J Manag Care Spec Pharm. 2017
Apr;23(4):427-45. Epub 2016 Dec 30.
5. Florence CS, Zhou C, Luo F, Xu L. The economic burden of prescription opioid
overdose, abuse, and dependence in the United States, 2013. Med Care. 2016 Oct;
54(10):901-6.
6. Levy B, Paulozzi L, Mack KA, Jones CM. Trends in opioid analgesic-prescribing
rates by specialty, U.S., 2007-2012. Am J Prev Med. 2015 Sep;49(3):409-13. Epub
2015 Apr 18.
7. Martin BI, Turner JA, Mirza SK, Lee MJ, Comstock BA, Deyo RA. Trends in health
care expenditures, utilization, and health status among US adults with spine prob-
lems, 1997-2006. Spine (Phila Pa 1976). 2009 Sep 1;34(19):2077-84.
8. Smith SR, Bido J, Collins JE, Yang H, Katz JN, Losina E. Impact of preoperative
opioid use on total knee arthroplasty outcomes. J Bone Joint Surg Am. 2017 May 17;
99(10):803-8.
9. Davis MP, Mehta Z. Opioids and chronic pain: where is the balance? Curr Oncol
Rep. 2016 Dec;18(12):71.
10. Faour M, Anderson JT, Haas AR, Percy R, Woods ST, Ahn UM, Ahn NU. Prolonged
preoperative opioid therapy associated with poor return to work rates after single-
level cervical fusion for radiculopathy for patients receiving Workers’ Compensation
benefits. Spine (Phila Pa 1976). 2017 Jan 15;42(2):E104-10.
11. Lawrence JT, London N, Bohlman HH, Chin KR. Preoperative narcotic use as a
predictor of clinical outcome: results following anterior cervical arthrodesis. Spine
(Phila Pa 1976). 2008 Sep 1;33(19):2074-8.
12. Walid MS, Hyer L, Ajjan M, Barth AC, Robinson JS Jr. Prevalence of opioid
dependence in spine surgery patients and correlation with length of stay. J Opioid
Manag. 2007 May-Jun;3(3):127-8: 130-2.
13. Anderson JT, Haas AR, Percy R, Woods ST, Ahn UM, Ahn NU. Chronic opioid
therapy after lumbar fusion surgery for degenerative disc disease in a Workers’
Compensation setting. Spine (Phila Pa 1976). 2015 Nov;40(22):1775-84.
14. Armaghani SJ, Lee DS, Bible JE, Archer KR, Shau DN, Kay H, Zhang C, McGirt
MJ, Devin CJ. Preoperative opioid use and its association with perioperative opioid
demand and postoperative opioid independence in patients undergoing spine sur-
gery. Spine (Phila Pa 1976). 2014 Dec 1;39(25):E1524-30.
15. Armaghani SJ, Lee DS, Bible JE, Shau DN, Kay H, Zhang C, McGirt MJ, Devin CJ.
Increased preoperative narcotic use and its association with postoperative compli-
cations and length of hospital stay in patients undergoing spine surgery. Clin Spine
Surg. 2016 Mar;29(2):E93-8.
16. Lee D, Armaghani S, Archer KR, Bible J, Shau D, Kay H, Zhang C, McGirt MJ,
Devin C. Preoperative opioid use as a predictor of adverse postoperative self-
reported outcomes in patients undergoing spine surgery. J Bone Joint Surg Am. 2014
Jun 4;96(11):e89. Epub 2014 Jun 4.
17. Jain N, Phillips FM, Weaver T, Khan SN. Preoperative chronic opioid therapy: a
risk factor for complications, readmission, continued opioid use and increased costs
after one- and two-level posterior lumbar fusion. Spine (Phila Pa 1976). 2018 Oct 1;
43(19):1331-8.
18. Villavicencio AT, Nelson EL, Kantha V, Burneikiene S. Prediction based on
preoperative opioid use of clinical outcomes after transforaminal lumbar interbody
fusions. J Neurosurg Spine. 2017 Feb;26(2):144-9. Epub 2016 Sep 30.
19. Webster BS, Verma SK, Gatchel RJ. Relationship between early opioid pre-
scribing for acute occupational low back pain and disability duration, medical costs,
subsequent surgery and late opioid use. Spine (Phila Pa 1976). 2007 Sep 1;32(19):
2127-32.
20. Mahmud MA, Webster BS, Courtney TK, Matz S, Tacci JA, Christiani DC. Clinical
management and the duration of disability for work-related low back pain. J Occup
Environ Med. 2000 Dec;42(12):1178-87.
21. Connolly J 3rd, Javed Z, Raji MA, Chan W, Kuo YF, Baillargeon J. Predictors of
long-term opioid use following lumbar fusion surgery. Spine (Phila Pa 1976). 2017
Sep 15;42(18):1405-11.
C O M P L I C AT I O N S , R E A D M I S S I O N , A N D R E V I S I O N S U R G E RY V E R S U S
D U R AT I O N O F P R E O P E R AT I V E O P I O I D U S E
22. Jain N, Brock JL, Phillips FM, Weaver T, Khan SN. Chronic preoperative opioid
use is a risk factor for increased complications, resource use, and costs after
cervical fusion. Spine J. 2018 Apr 27:S1529-9430(18)30126-8. Epub 2018 Apr
27.
23. Bedard NA, DeMik DE, Dowdle SB, Owens JM, Liu SS, Callaghan JJ. Does
preoperative opioid use increase the risk of early revision total hip arthroplasty? J
Arthroplasty. 2018 Jul;33(7S):S154-6. Epub 2018 Feb 13.
24. Goesling J, Moser SE, Zaidi B, Hassett AL, Hilliard P, Hallstrom B, Clauw DJ,
Brummett CM. Trends and predictors of opioid use after total knee and total hip
arthroplasty. Pain. 2016 Jun;157(6):1259-65.
25. Bedard NA, Pugely AJ, Dowdle SB, Duchman KR, Glass NA, Callaghan JJ. Opioid
use following total hip arthroplasty: trends and risk factors for prolonged use. J
Arthroplasty. 2017 Dec;32(12):3675-9. Epub 2017 Aug 31.
26. Cancienne JM, Patel KJ, Browne JA, Werner BC. Narcotic use and total knee
arthroplasty. J Arthroplasty. 2018 Jan;33(1):113-8. Epub 2017 Aug 17.
27. Rozell JC, Courtney PM, Dattilo JR, Wu CH, Lee GC. Preoperative opiate use
independently predicts narcotic consumption and complications after total joint
arthroplasty. J Arthroplasty. 2017 Sep;32(9):2658-62. Epub 2017 Apr 12.
28. Hernandez NM, Parry JA, Mabry TM, Taunton MJ. Patients at risk: preoperative
opioid use affects opioid prescribing, refills, and outcomes after total knee arthro-
plasty. J Arthroplasty. 2018 Jul;33(7S):S142-6. Epub 2018 Jan 16.
29. Hadlandsmyth K, Vander Weg MW, McCoy KD, Mosher HJ, Vaughan-Sarrazin
MS, Lund BC. Risk for prolonged opioid use following total knee arthroplasty in
veterans. J Arthroplasty. 2018 Jan;33(1):119-23. Epub 2017 Aug 24.
30. Kim KY, Anoushiravani AA, Chen KK, Roof M, Long WJ, Schwarzkopf R. Preop-
erative chronic opioid users in total knee arthroplasty-which patients persistently
abuse opiates following surgery? J Arthroplasty. 2018 Jan;33(1):107-12. Epub
2017 Aug 3.
31. Hansen CA, Inacio MCS, Pratt NL, Roughead EE, Graves SE. Chronic use of
opioids before and after total knee arthroplasty: a retrospective cohort study. J
Arthroplasty. 2017 Mar;32(3):811-817.e1. Epub 2016 Oct 4.
32. PearlDiver Technologies. Research capabilities. 2018. http://www.
pearldiverinc.com/researchinfo.html. Accessed 2018 Mar 21.
33. Schedules CS. Controlled substance schedules. 2018. https://www.
deadiversion.usdoj.gov/schedules/. Accessed 2018 Mar 21.
34. Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined
comorbidity index. J Clin Epidemiol. 1994 Nov;47(11):1245-51.
35. Chronic Conditions Data Warehouse. Condition categories. 2018. https://
www.ccwdata.org/web/guest/condition-categories. Accessed 2018 Mar 21.
36. Cicero TJ, Wong G, Tian Y, Lynskey M, Todorov A, Isenberg K. Co-morbidity and
utilization of medical services by pain patients receiving opioid medications: data
from an insurance claims database. Pain. 2009 Jul;144(1-2):20-7. Epub 2009 Apr
10.
37. Nguyen TH, Randolph DC, Talmage J, Succop P, Travis R. Long-term outcomes
of lumbar fusion among Workers’ Compensation subjects: a historical cohort study.
Spine (Phila Pa 1976). 2011 Feb 15;36(4):320-31.
38. Mayer TG, Gatchel RJ, Brede E, Theodore BR. Lumbar surgery in work-related
chronic low back pain: can a continuum of care enhance outcomes? Spine J. 2014
Feb 1;14(2):263-73. Epub 2013 Nov 12.
39. Ben-Ari A, Chansky H, Rozet I. Preoperative opioid use is associated with early
revision after total knee arthroplasty: a study of male patients treated in the Veterans
Affairs system. J Bone Joint Surg Am. 2017 Jan 4;99(1):1-9.
40. Wiese AD, Griffin MR, Schaffner W, Stein CM, Greevy RA, Mitchel EF Jr, Grijalva
CG. Opioid analgesic use and risk for invasive pneumococcal diseases: a nested
case-control study. Ann Intern Med. 2018 Mar 20;168(6):396-404. Epub 2018 Feb
13.
41. Nguyen LC, Sing DC, Bozic KJ. Preoperative reduction of opioid use before total
joint arthroplasty. J Arthroplasty. 2016 Sep;31(9)(Suppl):282-7. Epub 2016 Mar 17.
42. Edwards PK, Mears SC, Stambough JB, Foster SE, Barnes CL. Choices, com-
promises, and controversies in total knee and total hip arthroplasty modifiable risk
factors: what you need to know. J Arthroplasty. 2018 Oct;33(10):3101-6. Epub 2018
Feb 23.
43. McAnally H. Rationale for and approach to preoperative opioid weaning: a pre-
operative optimization protocol. Perioper Med (Lond). 2017 Nov 22;6:19.
44. Hassamal S, Haglund M, Wittnebel K, Danovitch I. A preoperative interdisci-
plinary biopsychosocial opioid reduction program in patients on chronic opioid
analgesia prior to spine surgery: a preliminary report and case series. Scand J Pain.
2016 Oct;13:27-31. Epub 2016 Jul 4.