We reviewed a total of 947 patient encounters of which 304 were excluded: 78 were not admitted for an orthopedic indication, 72 were for multiple visits by the same patient, 68 were cancer-related, 32 were for fractures in more than two body areas, 23 were not admitted directly from acute care, 11 were discharged to the community in less than 7 days, 10 were for individuals who previous used high-potency opioids, and 9 where the patient returned to acute care before assessments could be completed.
Of 643 included patient encounters, 125 (19.4%) involved chronic opioid users, 416 (64.7%) were opioid-naïve patients who received opioids during orthopedic rehabilitation, and 102 (15.9%) were opioid-naïve patients who did not receive opioids during orthopedic rehabilitation. Table 1 presents the patient characteristics for these three groupings. Chronic users were younger, had a higher BMI, and had a higher rate of psychiatric comorbidity. They were typically admitted for surgery and had a high proportion of elective joint replacements. Opioid-naïve users were older, had a lower BMI, and had a lower rate of psychiatric comorbidity. They were slightly less likely to have been operated on, with substantially less elective surgery. Non-users of opioids were older, had lower BMIs, and had a higher rate of dementia and baseline functional impairment. They were less likely to have undergone surgery, with only a minority of these surgeries being elective in nature.
Table 1
|
Chronic user
|
Opioid naive
|
Non-opioid user
|
p
|
n
|
125
|
416
|
102
|
|
Age (median (IQR))
|
72 (65, 76)
|
79 (71, 87)
|
83 (73, 90)
|
<0.001
|
Female Sex (%)
|
78 (62.4)
|
293 (70.4)
|
64 (62.7)
|
0.124
|
BMI (median (IQR))
|
27.6 (23.4, 33.7)
|
25.3 (22.1, 29.7)
|
22.5 (19.8, 26.5)
|
<0.001
|
Dementia (%)
|
1 (0.8)
|
22 (5.3)
|
15 (14.7)
|
<0.001
|
Anxiety or Depression (%)
|
49 (39.2)
|
108 (26.0)
|
22 (21.6)
|
0.005
|
History of Substance Use (%)
|
14 (11.2)
|
28 (6.7)
|
14 (13.7)
|
0.044
|
Charlson Comorbidity Index (median (IQR))
|
3 (3, 4)
|
4 (3, 5)
|
4 (3, 5)
|
0.003
|
Independent in All Basic Activities of Daily Living (%)
|
101 (80.8)
|
341 (82.6)
|
73 (71.6)
|
0.043
|
Admitted Following Surgery (%)
|
113 (90.4)
|
337 (81.0)
|
64 (62.7)
|
<0.001
|
Elective Surgery (%)
|
78 (62.4)
|
162 (38.9)
|
16 (15.7)
|
<0.001
|
Admitted Following Non-Surgical Treatment
|
14 (11.2)
|
79 (19.0)
|
38 (37.3)
|
<0.001
|
Surgery Type (%)
|
|
|
|
<0.001
|
Hip Fracture
|
21 (16.8)
|
147 (35.3)
|
38 (37.3)
|
|
Hip Replacement
|
44 (35.2)
|
76 (18.3)
|
10 ( 9.8)
|
|
Knee Replacement
|
30 (24.0)
|
75 (18.0)
|
5 ( 4.9)
|
|
Other Surgery
|
16 (12.8)
|
39 (9.4)
|
11 (10.8)
|
|
Length of Stay in Acute Care Prior to Rehabilitation in Days (median (IQR))
|
4 (2, 7)
|
4 (3, 7)
|
5 (3, 8)
|
0.286
|
EGFR (median (IQR))
|
87 (63, 100)
|
83 (61, 96)
|
80 (64, 90)
|
0.029
|
FIM on Admission (median (IQR))
|
80 (75, 86)
|
80 (74, 85)
|
80 (75, 89)
|
0.583
|
Table 2summarizes patient outcomes. Chronic users had a higher median opioid dose than naïve-users (30.3 mg vs 6.9 mg OME daily) and used long-acting formulations more frequently. The distribution of opioid medications and formulations used are detailed in Supplement 2. Median lengths of stay were slightly less in chronic users, with similar functional and discharge outcomes between groups. Maximal pain on day 3 was highest in chronic users and lowest in non-users. Pain did not materially change from day 3 to 7 in any group. Opioid-related side-effects were distributed similarly among the groups, with constipation being the most common. Nausea was more common in chronic users versus naïve users, with non-users reporting the least nausea. Severe adverse events including acute kidney injury, injurious fall, and respiratory complications were rare and without significant intergroup differences (data not shown). Only 11 (1.7%) patients experienced a program interruption and 42 (6.5%) patients returned to acute care, with only 2 of these complications deemed potentially related to opioid use on retrospective chart review.
Table 2
|
Chronic user
|
Opioid naive
|
Non-opioid user
|
p
|
n
|
125
|
416
|
102
|
|
Median Opioid Use Daily in mg OME (median (IQR))
|
30.3 (7.4, 72.0)
|
6.9 (2.4, 16.1)
|
n/a
|
<0.001
|
Proportion Receiving Long Acting Opioid Formulations (%)
|
45 (36.0)
|
42 (10.1)
|
n/a
|
<0.001
|
Maximum Pain Reported on Day 3 (median (IQR))
|
5 (3, 7)
|
4 (2, 6)
|
3 (2, 4)
|
<0.001
|
Maximum Pain Reported on Day 7 (median (IQR))
|
5 (3, 7)
|
4 (2, 6)
|
2 (0, 4)
|
<0.001
|
Opioid-Related Complications
|
|
|
|
|
Urinary Retention (%)
|
8 (6.4)
|
37 (8.9)
|
5 (4.9)
|
0.328
|
Constipation (%)
|
71 (56.8)
|
238 (57.2)
|
53 (52.0)
|
0.627
|
Confusion or Delirium (%)
|
9 (7.2)
|
20 (4.8)
|
6 (5.9)
|
0.573
|
Nausea (%)
|
40 (32.0)
|
94 (22.6)
|
19 (18.6)
|
0.039
|
Rehabilitation Length of Stay in Days (median (IQR))
|
20 (12, 28)
|
22 (14, 35)
|
23 (14, 35)
|
0.109
|
FIM on Discharge (median (IQR))
|
111 (107, 115)
|
110 (103, 114)
|
111 (106, 115)
|
0.031
|
FIM Change (mean (SD))
|
29 (9)
|
28 (9)
|
27 (8)
|
0.371
|
FIM Efficiency (median (IQR))
|
1.75 (1.27, 2.20)
|
1.46 (1.08, 2.07)
|
1.43 (0.95, 2.00)
|
0.044
|
Discharged to a Community Setting (%)
|
111 (88.8)
|
388 (93.3)
|
89 (87.3)
|
0.075
|
Univariable regression results for factors correlating with opioid dose are shown in Supplement 3. Results of multivariable regression modeling for opioid dose can be seen in Table 3. In the final multivariable model, age was inversely correlated with amount of opioid administered. Reported pain, as defined by day 3 pain score, was associated with increased opioid use. Knee replacement was associated with increased opioid use while admission for a non-surgical indication was associated with lower opioid requirements.
Univariable and multivariable regression modeling results examining the factors contributing to day 3 pain scores can be seen in Supplement 4. The multivariable model identified underlying dementia as being negatively associated with reported pain and a history of psychiatric illness as being positively associated with reported pain.
A separate tabulation examining the amount of daily opioid administered over the first week, specifically to opioid naïve users (i.e. excluding chronic users and non-opioid users) by age showed that those aged 50-64 received a median of 19.4 mg daily, 65-74 received 12.6 mg, 75-84 received 6.0 mg and 85+ received 3.4 mg.
Table 3
Multivariable Regression Associations with Opioid Dose in Oral Morphine Equivalents (OME) (Final Model)
|
Coefficients (CI)
|
p
|
Age
|
-2.348 (-3.251, -1.437)
|
<0.0001
|
Female Sex
|
0.832 (-18.1, 24.2) 0.938
|
0.938
|
BMI
|
1.121 (-0.425, 2.691)
|
0.157
|
Dementia
|
-17.9 (-44.3, 21)
|
0.32
|
Psychiatric Disorder
|
8.884 (-13.7, 37.3)
|
0.473
|
Elective Surgery
|
40.7 (-7.94, 115)
|
0.115
|
Admitted Following Non-Surgical Treatment
|
-36.3 (-50.9, -17.4)
|
<0.001
|
Hip Replacement
|
8.141 (-33.8, 76.6)
|
0.755
|
Knee Replacement
|
71.2 (4.039, 182)
|
0.035
|
Other Surgery
|
12.8 (-22.4, 64.1)
|
0.527
|
Maximum Pain Reported on Day 3
|
19.2 (14.6, 24)
|
<0.0001
|