If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Time trends in use of non-steroidal anti-inflammatory drugs and opioids one year after total hip arthroplasty due to osteoarthritis during 1996–2018: a population-based cohort study of 103,209 patients
Address correspondence and reprint requests to: A.N. Klenø Department of Clinical Epidemiology Aarhus University Olof Palmes Allé 43-45, 8200, Århus N, Denmark. Tel.: 45-20-99-28-93.
To examine time trends in the use of NSAIDs and opioids for patients with osteoarthritis undergoing total hip arthroplasty (THA) during 1996–2018.
Method
Using Danish population-based medical databases, we identified 103,209 THA patients. Prevalence rates of NSAID and opioid use among preoperative users and non-users were calculated in four quarters (Q1-Q4) after THA by calendar periods (1996–2000, 2001–2006, 2007–2012 and 2013–2018). Prevalence rate ratios (PRR) were adjusted for age and gender.
Results
Among preoperative NSAID users and non-users, NSAID use in Q1 increased from 32.6% in 1996–2000 to 48.0% in 2013–2018 (PRR = 1.49, 95% CI: 1.42–1.55) and from 12.9% to 32.0% (PRR = 2.49 (2.32–2.67)), respectively. Among preoperative opioid users and non-users, opioid use in Q1 increased from 42.7% in 1996–2000 to 76.9% in 2013–2018 (PRR = 1.81 (1.73–1.89)) and from 15.2% to 58.2% (PRR = 3.85 (3.65–4.05)), respectively. NSAID use in Q4 decreased from 24.5% in 1996–2000 to 21.4% in 2013–2018 (PRR = 0.88 (0.83–0.93)) and from 6.9% to 5.6% (PRR = 0.81 (0.73–0.91)) in preoperative NSAIDs users and non-users, respectively. Opioid use in Q4 increased from 26.6% in 1996–2000 to 28.6% (PRR = 1.08 (1.02–1.15)) in 2013–2018 and from 4.1% to 5.0% (PRR = 1.25 (1.11–1.40)) in preoperative opioid users and non-users, respectively.
Conclusion
We observed up to a 4-fold increase in NSAID and opioid use in Q1 during 1996–2018, while usage in Q4 did not change substantially. However, 5–6% of the preoperative non-users of NSAIDs and opioids were users in Q4, which might relate to inaccurate indication for or timing of THA and the post-surgical phasing out of analgesics use.
, and although the absolute number of prescriptions are lower in Europe compared to the USA, several European countries do have high defined daily doses of opioid consumption
, but the proportion of patients using opioids 1 year after THA is almost identical to the proportion of patients using opioids 1 year before THA, while a decrease of less than 50% is seen for NSAID use in Denmark, Finland and Norway
Socioeconomic status and use of analgesic drugs before and after primary hip arthroplasty: a population-based cohort study of 103,209 patients during 1996-2018.
Prescription trajectories and effect of total hip arthroplasty on the use of analgesics, hypnotics, antidepressants, and anxiolytics: results from a population of total hip arthroplasty patients.
. Furthermore, patients who use opioids before surgery are much more likely to use opioids 1 year after THA (47%), compared to opioid naïve patients (10%)
. This can potentially be a matter of great concern to patient safety due to the number of side effects associated with opioids, like physical dependence or respiratory depression, hyperalgesia, hormonal dysfunction, mortality, risk of falling and fractures
, while less is known about time trends of opioid use in relation to THA in Europe. Changes in prescription patters often happen as a consequence of renewed guidelines or media attention. For instance in Denmark, a recommendation for using the cheapest opioids was issued by the Danish Health Authority in 2011, while a reassessment of reimbursement for several opioids was introduced in 2013
The Danish Hip Arthroplasty Register (DHR) contains information on all THAs performed in Denmark since 1995 and by linking it with population-based community pharmacy prescription data at the individual level, it is possible to perform a detailed assessment of time trends in analgesics use overall and by individual analgesics
. We hypothesized that increased awareness of possible side effects and the development of clinical opioid guidelines have contributed significantly to a change in analgesic prescription patterns after THA towards a reduction in the use of strong analgesics over the last 22 years.
We therefore conducted a population-based cohort study to examine time trends in the prescription of analgesics for patients with osteoarthritis (OA) undergoing THA during 1996–2018, while focusing on differences between preoperative analgesic users and non-users, and individual analgesics.
Patients and methods
Study design and population
Denmark is a country of 5.8 million residents, who all have universal access to healthcare
We conducted a population-based cohort study using prospectively collected data from nationwide medical and administrative databases. All Danish patients undergoing primary THA between January 1st 1996 and June 1st 2018 with OA as main indication for THA, were identified through the DHR
. The registration completeness of primary procedures in the DHR is more than 95% and all public orthopedic departments and private hospitals report to the register
Registration in the Danish hip arthroplasty registry: completeness of total hip arthroplasties and positive predictive value of registered diagnosis and postoperative complications.
. Patients were excluded from the cohort if undergoing revision THA in the subsequent year after their primary THA (n = 2,309). This was applied to avoid misleading analgesics use related to revision rather than primary THA. The Danish Civil Registration System (CRS), established in 1968, assigns a unique civil registration (CPR) number to every Danish resident at birth or upon immigration
. The CPR number encodes age and sex and enables us to link individual-level data between multiple databases and to obtain information on vital status.
Analgesics use
We identified prescriptions for NSAIDs and opioids by using the Danish National Prescription Registry
and relevant Anatomical Classification System (ATC) codes and dispensation dates. The registry contains information on all prescriptions redeemed by Danish residents in community pharmacies since 1995 (excluding hospital dispensations). In Denmark, prescription reimbursement covers several types of analgesics, being assessed by the Reimbursement Committee. Reimbursement increases with the need for higher quantities of analgesics and automatically applies to all Danish residents
We defined use of NSAIDs and opioids separately, as well as a composite outcome (referred to as any analgesic use). At the time of THA, patients were considered preoperative users of NSAIDs or opioids if they had redeemed at least one prescription of the respective drug 0–6 months before THA
. Patients without such a redeemed prescription were considered preoperative non-users. Data on prescriptions within 1 year after THA were obtained and divided into quarters; Q1, Q2, Q3, and Q4. In each quarter, patients were defined as users if redeeming at least one prescription. ATC codes for NSAIDs and opioids are presented in the supplementary material (Table S1).
Patient characteristics
Information on sex and age at time of THA were available through all of the included databases, as we used this information to link data between databases.
To obtain a 10-year history of hospitalizations prior to THA, we used the Danish National Patient Registry (DNPR)
. The registry contains data on all non-psychiatric inpatient contacts since 1977 and all outpatient clinic and emergency room contacts since 1995. To measure comorbidity, we computed a Charlson Comorbidity Index (CCI) score for each patient, based on the 19 disease categories and corresponding ICD-10 codes (the International Classification of Diseases (eighth revision [ICD-8] until the end of 1993 and tenth revision [ICD-10] thereafter) adapted for administrative purposes (Appendix 1)
. We divided CCI scores into three categories: 0 (low), 1–2 (medium) and ≥3 (high).
Statistics
We described the study population according to the distribution of sex, age (in categories <46, 46–55, 56–65, 66–75, 76–85, and >85 years), and CCI score, tabulating the number and percentage of patients by calendar periods of THA (1996–2000, 2001–2006, 2007–2012, and 2013–2018). Further, selected results were additionally divided into 2 years periods.
First, we estimated prevalence rates for any analgesic use and NSAID and opioid use separately for each of six quarters (-Q1 and -Q2 before THA and Q1-Q4 after THA) by calendar periods. Second, we estimated prevalence rates of NSAID and opioid use stratifying on preoperative users and preoperative non-users to overcome limitations due to lack of indication for analgesics prescription. Prevalence rates were based on the number of patients redeeming at least one prescription for NSAIDs or opioids divided by the total number of patients in our study population, who were alive at the beginning of each quarter. Third, we used Poisson regression to estimate prevalence rate ratios (PRR), comparing the prevalence rate for calendar periods with 1996–2000 as a reference, adjusting for sex and age. For the PRR, we focused on analgesic use in Q1 and Q4 after THA. Assumptions for Poisson regression were assessed and considered fulfilled. Last, we calculated numbers and percentages of the four most commonly prescribed NSAIDs and opioids throughout every quarter for both preoperative users and preoperative non-users. For all results, we calculated the 95% confidence intervals (CI). The analyses were performed using Stata 16.0
All included patients were OA patients who underwent THA. Since patients with no redeemed prescriptions for NSAIDs or opioids were considered non-users, there were no missing data to report in this study. Data on patient characteristics sex, age and CCI were complete.
Ethics
According to Danish law, an ethics committee approval is not required for registry-based studies. The study was reported to the Danish Data Protection Agency through registration at Aarhus University (record number: AU-2016-051-000001, sequential number 880).
Results
We identified 105,518 patients with OA undergoing primary THA during 1996–2018. After excluding 2,309 patients due to revision THA within 1 year after a primary THA, the final study cohort consisted of 103,209 patients (Fig. 1). Overall, 59% of the patients were female, the median age was 70 years and four out of five had no comorbidities (79%). Patient characteristics are presented separately for preoperative NSAID and opioid users and non-users by four calendar periods (Table I).
Fig. 1Flowchart. THA: Total Hip Arthroplasty. Preoperative users: Patients who redeemed at least one NSAID or opioid prescription 0–6 months before THA. Preoperative non-users: Patients who did not redeem a NSAID or opioid prescription 0–6 months before THA.
Table IPatient characteristics for preoperative users and preoperative non-users of NSAIDs and opioids at the day of THA by calendar periods
Preoperative NSAID users (n, %)
Preoperative NSAID non-users (n, %)
1996–2000
2001–2006
2007–2012
2013–2018
1996–2000
2001–2006
2007–2012
2013–2018
n = 9,545
n = 16,172
n = 17,701
n = 13,771
n = 6,781
n = 8,767
n = 13,033
n = 17,439
Sex
Female
5,412 (57)
9,408 (58)
9,942 (56)
7,804 (57)
3,738 (55)
4,623 (53)
7,078 (54)
9,659 (55)
Male
4,133 (43)
6,764 (42)
7,759 (44)
5,967 (43)
3,043 (45)
4,144 (47)
5,955 (46)
7,780 (45)
Age
Median
70 (63–77)
70 (62–76)
69 (62–75)
69 (62–75)
71 (64–77)
71 (63–77)
70 (63–77)
71 (65–77)
<46
118 (1)
236 (1)
294 (2)
298 (2)
83 (1)
166 (2)
287 (2)
291 (2)
46–55
790 (8)
1,284 (8)
1,503 (8)
1,446 (10)
485 (7)
651 (7)
989 (7)
1,361 (8)
56–65
2,426 (26)
4,471 (28)
4,887 (28)
3,425 (25)
1,473 (22)
2,118 (24)
3,243 (25)
3,469 (20)
66–75
3,556 (37)
5,985 (37)
6,936 (39)
5,596 (41)
2,692 (40)
3,295 (38)
4,929 (38)
7,182 (41)
76–85
2,397 (25)
3,701 (23)
3,585 (20)
2,697 (20)
1,823 (27)
2,207 (25)
3,081 (24)
4,483 (25)
>85
258 (3)
495 (3)
496 (3)
309 (2)
224 (3)
330 (4)
504 (4)
653 (4)
Comorbidities
Low
8,409 (88)
13,029 (80)
14,108 (80)
11,206 (81)
5,727 (84)
6,713 (76)
9,771 (75)
12,916 (74)
Medium
1,043 (11)
2,722 (17)
3,105 (17)
2,190 (16)
948 (14)
1,719 (20)
2,694 (21)
3,684 (21)
High
93 (1)
421 (3)
488 (3)
375 (3)
106 (2)
335 (4)
568 (4)
839 (5)
Preoperative opioid users (n, %)
Preoperative opioid non-users (n, %)
1996–2000
2001–2006
2007–2012
2013–2018
1996–2000
2001–2006
2007–2012
2013–2018
n = 5,791
n = 8,813
n = 11,592
n = 11,204
n = 10,535
n = 16,126
n = 19,142
n = 20,006
Sex
Female
3,605 (62)
5,567 (63)
7,090 (61)
6,996 (62)
5,545 (53)
8,464 (53)
9,930 (52)
10,467 (52)
Male
2,186 (38)
3,246 (37)
4,502 (39)
4,208 (38)
4,990 (47)
7,662 (47)
9,212 (48)
9,539 (48)
Age
Median
71 (64–78)
71 (64–78)
71 (64–77)
71 (64–78)
70 (63–77)
69 (62–76)
69 (62–75)
70 (63–76)
<46
60 (1)
92 (1)
165 (1)
152 (1)
141 (1)
310 (2)
416 (2)
437 (2)
46–55
434 (7)
578 (7)
795 (7)
893 (8)
841 (8)
1,357 (8)
1,697 (9)
1,914 (10)
56–65
1,321 (23)
2,198 (25)
2,841 (25)
2,320 (21)
2,578 (25)
4,391 (27)
5,289 (27)
4,574 (23)
66–75
2,130 (37)
3,231 (37)
4,440 (38)
4,435 (40)
4,118 (39)
6,049 (38)
7,425 (39)
4,343 (42)
76–85
1,622 (28)
2,319 (26)
2,836 (25)
2,922 (26)
2,599 (25)
3,589 (22)
3,830 (20)
4,258 (21)
>85
224 (4)
395 (4)
515 (4)
482 (4)
258 (2)
430 (3)
485 (3)
480 (2)
Comorbidities
Low
4,421 (81)
6,373 (72)
8,242 (71)
7,940 (71)
9,415 (89)
13,369 (83)
15,637 (82)
16,182 (81)
Medium
959 (17)
2,023 (23)
2,736 (24)
2,600 (23)
1,032 (10)
2,418 (15)
3,063 (16)
3,274 (16)
High
111 (2)
417 (5)
614 (5)
664 (4)
88 (1)
339 (2)
442 (2)
550 (3)
Characteristics for preoperative users and non-users are calculated for NSAIDs and opioids separately.
Preoperative users: Patients who redeemed at least one prescription 0–6 months before THA. Preoperative non-users: Patients who did not redeem a prescription 0–6 months before THA. Median is presented with interquartile range.
THA: Total Hip Arthroplasty, Comorbidities: Low (CCI = 0); Medium (CCI = 1 or 2); High (CCI = 3+), CCI: Charlson Comorbidity Index
Patient characteristics were generally highly comparable across calendar periods. Approximately 56% of patients were females and median age was around 70. Fewer patients had comorbidities in 1996–2000 compared to 2001–2018 (Table S2).
Use of any analgesics in -Q2 before THA decreased about 30% from 61.4% in 1996–2000 to 43.1% in 2013–2018. The prevalence rate of any analgesics use in Q4 after THA was 25.7% in 1996–2000 and 31.4% in 2001–2006, decreasing to 22.4% in 2013–2018. In contrast, use of any analgesics increased continuously in the short-term postoperative management phase (Q1), with the lowest prevalence rate in 1996–2000 (42.1%) and the highest prevalence rate in 2013–2018 (77.9%) (Fig. 2, Table S3).
Fig. 2Time trends in prevalence rates of analgesics users 6 months before to 1 year after THA for any analgesics, NSAIDs and opioids. The black line in the graph marks time of THA. THA: Total Hip Arthroplasty, Q: Quarters with a duration of 3 months.
Time trends in NSAID use among preoperative users and non-users
For both preoperative NSAID users and non-users, sex distribution and median age remained highly similar throughout all four calendar periods. The proportion of patients without any comorbidities decreased from 88% in 1996–2000 to 81% in 2013–2018 among preoperative users and from 84% to 74% among non-users (Table I).
Among preoperative users, the prevalence rate of NSAID users in Q1 increased from 32.6% in 1996–2000 to 48.0% in 2013–2018 (PRR = 1.49, CI: 1.42–1.55). Similarly, among preoperative non-users, the prevalence rate of NSAID users in Q1 increased from 12.9% to 32.0% (PRR = 2.49, CI: 2.32–2.67). In Q4, NSAID use decreased from 24.5% to 21.4% (PRR = 0.88, CI: 0.83–0.93) and from 6.9% to 5.6% (PRR = 0.81 CI: 0.73–0.91) from 1996–2000 to 2013–2018 in preoperative NSAID users and non-users, respectively (Figs. 3 and 4, Table II, Table S4).
Fig. 3Time trends in prevalence rates of NSAID and opioid use 1 year after THA, among preoperative users and preoperative non-users. Preoperative users: Patients who redeemed at least one NSAID or opioid prescription 0–6 months before THA. Preoperative non-users: Patients who did not redeem a NSAID or opioid prescription 0–6 months before THA. The black line in the graph marks time of THA. THA: Total Hip Arthroplasty, Q: Quarters with a duration of 3 months.
Fig. 4Detailed description of changes in prevalence rates for every 2 years. Preoperative users: Patients who redeemed at least one NSAID or opioid prescription 0–6 months before THA. Preoperative non-users: Patients who did not redeem a NSAID or opioid prescription 0–6 months before THA. THA: Total Hip Arthroplasty, Q: Quarters with a duration of 3 months.
Time trends in opioid use among preoperative users and non-users
About 62% of the preoperative opioid users were female whereas 53% of the preoperative opioid non-users were female throughout all four calendar periods. The median age was similar for both preoperative users and non-users while the proportion of patients with no comorbidities was slightly higher among preoperative opioid users but decreased over time for both preoperative users and non-users (Table I).
Among preoperative opioid users, the prevalence rate of opioid users in Q1 increased from 42.7% in 1996–2000 to 76.9% in 2013–2018 (PRR = 1.81, CI: 1.73–1.89). Among preoperative non-users, the prevalence rate of opioid users increased from 15.2% in 1996–2000 to 58.2% in 2013–2018 (PRR = 3.85, CI: 3.65–4.05). In Q4, opioid use increased slightly from 26.6% to 28.6% (PRR = 1.08, CI: 1.02–1.15) and from 4.1% to 5.0% (PRR = 1.25, CI: 1.11–1.40) in preoperative opioid users and non-users, respectively (Figs. 3 and 4, Table II, Table S4).
Changes in specific medication
Ibuprofen was the most commonly used type of NSAIDs through all calendar periods for both preoperative users and non-users. In 1996–2000, ibuprofen accounted for 44% of all NSAIDs used in Q1 among preoperative users and for 27% among preoperative non-users increasing to 80% and 83%, respectively, in 2013–2018. In 1996–2000, ibuprofen accounted for 35% and 34% of all NSAIDs used in Q4 among preoperative users and non-users, which increased to 74% and 84%, respectively, in 2013–2018 (Fig. 5, Table S5). In 1996–2000, diclofenac accounted for 17% and 14% of all NSAIDs used in Q1 among preoperative users and non-users, which decreased to 4% and 1%, respectively, in 2013–2018. For Q4, these numbers were 18% and 19%, decreasing to 8% and 4%, respectively (Fig. 5, Table S5).
Fig. 5Time trends in use of specific types of NSAIDs and opioids from 0 to 3 and 9 to 12 months after THA among preoperative users and preoperative non-users. Preoperative users: Patients who redeemed at least one NSAID or opioid prescription 0–6 months before THA. Preoperative non-users: Patients who did not redeem a NSAID or opioid prescription 0–6 months before THA. THA: Total Hip Arthroplasty, Q: Quarters with a duration of 3 months.
Tramadol was the most prescribed opioid for preoperative users through all calendar periods and for preoperative non-users in 1996–2006. In 1996–2000, tramadol accounted for 45% of all opioids used in Q1 among preoperative users and for 58% among preoperative non-users decreasing to 33% and 23%, respectively, in 2013–2018. In 1996–2000, tramadol accounted for 41% and 59% of all opioids used in Q4 among preoperative users and non-users, which changed to 43% and 37%, respectively, in 2013–2018 (Fig. 5, Table S6). In 1996–2000, oxycodone accounted for 1% of the prescribed opioids in Q1 for both preoperative users and non-users, but increased markedly to 35% and 47%, respectively, in 2013–2018. In 1996–2000, these numbers were 2% in Q4, increasing to 29% and 23%, respectively, in 2013–2018 (Fig. 5, Table S6).
Discussion
Based on 103,209 primary THA patients, we found that the prevalence rate of patients using analgesics decreased in -Q2 (61–43%) before THA and in Q4 (26–22%) after THA during the 22-year period during 1996–2018.
Although we observed up to a 4-fold increase in NSAID and opioid use in Q1 after THA during 1996–2018, use of NSAIDs and opioids has not changed substantially since 1996 in Q4 in neither preoperative users nor non-users. Nevertheless, 5–6% of preoperative non-users of NSAIDs and opioids were users in Q4 after THA.
We observed a switch in the use of specific NSAIDs. Thus, the use of ibuprofen increased, whereas the use of diclofenac decreased substantially over time. For opioids, tramadol was the predominant opioid used in 1996–2006 and was mainly replaced by oxycodone and morphine in 2013–2018.
Comparison to other studies
The decrease in NSAID and opioid use among THA patients is most likely a reflection of the general decrease in the use of these drugs in Denmark
. However, an American study on THA patients by Raad et al. suggested that patients who are physically independent before THA experience a much better outcome post-surgery
. Our finding of a decrease in analgesic use over time might suggest that THA is offered to less severe and more mobile OA patients now compared to 1996–2000.
Stokes et al. investigated time trends in opioid and non-opioid medication use during 1999–2016
. The cross-sectional study included patients with functional limitations attributable to musculoskeletal conditions, which unquestionably differ from OA and THA, but still provide interesting comparisons. Stokes et al. found that use of non-opioid analgesics was decreasing, which is comparable to our findings regarding NSAID use, with the exception of an increase in Q1 in 2013–2018. However, this increase decreased to a notably lower prevalence rate in Q2 compared to Q2 use in 1996–2012, and might be a result of short-term prophylactic use of NSAIDs to prevent heterotopic ossification after THA, as suggested in more recent research
. Furthermore, Stokes et al. observed an increase in long-term opioid use in 2015–2016 compared to 1999–2000. Contrary to this, we found that the increase in the number of patients using opioids were mainly centered around Q1, suggesting more aggressive approach to treatment of acute post-operative pain now compared to previous years. A similar trend was seen in hip fracture patients in a Danish cohort study by Bonnesen et al. who found very little variation in the number of opioid users 3–12 months after surgery, despite a prevalence rate during 0–3 months after surgery of 39% in 2003 compared to 29% in 2016
. This is in line with our trend analysis for preoperative users, but not for preoperative non-users, for whom we observed a much more consistent use during the same period (Fig. 4).
In large parts of the world, chronic opioid use is considered a major problem, and surgery is considered a risk factor for long-term use
. Politzer et al. investigated time trends in opioid use among total knee arthroplasty (TKA) patients and found that 35% of preoperative users and 5% of preoperative non-users were using opioids 6 months after TKA
. This is highly comparable to the 33% and 6% we found for that specific time in our study. Moreover, Politzer et al. examined time trends in specific opioid medication and found that tramadol use increased significantly during 2007–2013. This might be problematic due to studies indicating that tramadol is a predictor of long-term opioid use
. Opposite to Politzer, we found that tramadol usage in Q1 decreased markedly from 1996–2000 to 2013–2018, among both preoperative users and non-users. However, when we compared Q1 to Q4 tramadol consumption in 2013–2018, the use was seemingly much more persistent compared to other opioid types. This is in line with the European Pharmacovigilance Risk Assessment Committee's decision to update the Summary of Product Characteristics for tramadol in January 2018. The Danish Medicines Agency contributed with proposals of bringing tramadol recommendations in line with other opioids, removing the low dependence statement, as well as clarifying that tramadol dependence not only can occur during long-term use
A cohort study from New Zealand by Wilson et al. found that patients undergoing joint surgery and using opioids in the immediate postoperative period were at a substantial risk of becoming long-term users
. We found that immediate opioid use (Q1) increased substantially during the study period from 25% in 1996–2000 to 65% in 2013–2018 but that it did not affect use in Q4 (Fig. 4, Table S2). This trend might be explained by the switch from tramadol to less addictive opioids like oxycodone or morphine. In addition, an increase in NSAID use in Q1 over time, and lower doses of opioids combined with NSAIDs in Q1 might have lowered the addictive effect of getting several opioid prescriptions, thereby reducing the risk of becoming a long-term user. In line with this, a study on hip fracture patients showed that the number of patients using >1 type of opioids was decreasing
This cohort study contributes with a large nationwide sample size of 103,209 patients. We were able to base our investigation on a 22 year time span with population-based community pharmacy prescription data at the individual person level in Denmark. This enabled us to perform an in-depth assessment of time trends in analgesics use overall and by NSAIDs and opioids separately.
A potential limitation to this study is the lack of indication for analgesics use, resulting in some medication being prescribed for other reasons than pain related to OA or poor THA outcome. We accounted for this by stratifying patients into preoperative users and non-users, assuming that most non-users will be using analgesics postoperatively related to their current THA. This is a highly clinically relevant aspect of our study. It is likely that clinicians have the highest impact on preoperative non-users because the post-surgical handling of the patients does not intervene with other ongoing analgesic treatments. Therefore, we believe that it is a strength of the study that we are able to distinguish between the impact of THA on NSAID and opioid use in preoperative users and non-users. Preoperative opioid users had more comorbidities compared to non-users, and we expect that these conditions require more analgesic treatment. Therefore we assume that the expectation to analgesic treatment as well as postoperative handling should be different compared to that of preoperative non-users. It is furthermore important to acknowledge that the number of patients without comorbidities during calendar period 1996–2000 might be lower than anticipated due to misclassification from incomplete data on in-patient visits before 1995. However, we believe that this misclassification is less severe since the implementation of ambulatory capacity first really accelerated in 1995.
We have excluded revision THA to avoid impact of new surgery on the analgesic use. However, we did not exclude potential other surgeries performed in the year after primary THA. Although this could also influence the use of analgesics after primary THA, it is also important to address that this would mainly be due to unforeseen surgeries. Since the patient group are mainly elderly, they have a longer period of rehabilitation, which most likely would limit the number of scheduled surgeries in the year after THA.
Lastly, it is possible to buy over-the-counter NSAIDs in Denmark, which could potentially have led to misclassification of NSAID users as NSAID non-users, since analgesic use in our study was based on prescription data only. However, in a study examining the rate of misclassification of NSAID non-users in Danish prescription registries
The validity of Danish prescription data to measure use of aspirin and other non-steroidal anti-inflammatory drugs and quantification of bias due to non-prescription drug use.
, the authors found that over the last 20 years this misclassification peaked at 5%. This suggests that Danish prescription registries should be considered a valid data source for assessing NSAID use.
Conclusion
The overall use of analgesics before and after THA have generally decreased. Although we observed up to a 4-fold increase in NSAID and opioid use in Q1 after THA during 1996–2018, use of NSAIDs and opioids in Q4 did not change substantially since 1996. However, 5–6% of preoperative non-users were using analgesics in Q4, which might relate to inaccurate indication for or timing of THA and the post-surgical phasing out of analgesics use. Further, use of ibuprofen, oxycodone, and morphine increased, while use of diclofenac and tramadol decreased.
Author contributions
Study conception and design: ANK, ABP, HTS; data analysis: ANK; drafting the article: ANK, ABP; Critical revision of the article: ANK, ABP, HTS; final approval of the article: ANK, ABP, HTS.
ANK ([email protected]) and ABP ([email protected]) takes responsibility for the integrity of the work as a whole, from inception of the study to the finished article.
Funding source
ANK, ABP, and HTS are supported by and employed at Department of Clinical Epidemiology at Aarhus University. The Department of Clinical Epidemiology at Aarhus University Hospital is involved in studies with funding from various pharmaceutical companies as institutional research grants. These companies had no role in the study design, analysis and interpretation of data, in the writing of the manuscript, or in the decision to submit the manuscript for publication.
Conflict of interest
The authors declare no conflict of interest.
Appendix A. Supplementary data
The following is the Supplementary data to this article:
Socioeconomic status and use of analgesic drugs before and after primary hip arthroplasty: a population-based cohort study of 103,209 patients during 1996-2018.
Prescription trajectories and effect of total hip arthroplasty on the use of analgesics, hypnotics, antidepressants, and anxiolytics: results from a population of total hip arthroplasty patients.
Registration in the Danish hip arthroplasty registry: completeness of total hip arthroplasties and positive predictive value of registered diagnosis and postoperative complications.
The validity of Danish prescription data to measure use of aspirin and other non-steroidal anti-inflammatory drugs and quantification of bias due to non-prescription drug use.
00798-1&id=gr1.jpg){kind=link}
00798-1&id=gr2.jpg){kind=link}
00798-1&id=gr3.jpg){kind=link}
00798-1&id=gr4.jpg){kind=link}
00798-1&id=gr5.jpg){kind=link}