Utilization of medications for neuropathic pain in patients with and without knee and hip osteoarthritis in Sweden: A population-based cohort study.

The study was financed by Swedish governmental funding of clinical research (ALF), through which CH used the funding to offset salary costs for the time dedicated to working on the manuscript. Additional sources of funding were Region Skåne (SUS Donations), Sweden; The 10.13039/501100004359 Swedish Research Council, Sweden; Österlund Foundation, Sweden; Gustaf V 80-Year Birthday Foundation, Sweden; the 10.13039/501100007949 Swedish Rheumatism Association, Sweden; the Foundation for People with Movement Disability in Skåne, Sweden; the 10.13039/501100006075 Greta and Johan Kock Foundation, Sweden; and 10.13039/501100022244 NIHR Applied Research Collaboration West Midlands, United Kingdom . None of the latter have been used specifically for funding this article and no funder have had influence on study design, analysis, decision regarding publication or the like.

All authors were involved in the conceptualization and design of the study. The first author, CH, was responsible for writing the original draft. AT and CH conducted the statistical analyses. All authors contributed to the review and editing of the manuscript and have given final approval of the manuscript to be submitted. ME served a supervisory role in the project. CH ( [email protected] ) takes responsibility for integrity of the work.

Out of the original cohort of 769,223 individuals residing in Skåne as of December 31, 2018, a total of 645,991 individuals remained after the exclusion criteria were applied. Among these, 81,606 (12.7%) individuals had OA in the knee, hip, or both knee and hip, with the distribution further detailed in the flowchart of inclusion ( Fig. 1 ). Most OA patients were female (59.3%), compared to half of the individuals without OA (50.3%). Additionally, OA patients were older (mean age 69 years; standard deviation 12), compared to the group without OA (mean age 56 years; standard deviation 14) ( Table 1 ). The point prevalence of dispensed neuropathic pain medications, i.e. the use of these medications, was higher in patients with OA, 2.22% (95% CI 2.12, 2.33), compared to individuals without, 1.15% (95% CI 1.12, 1.18) and also higher in those with any comorbidity associated with the treatment ( Appendix 3 ). The use was slightly higher in individuals with both knee and hip OA compared to those with either knee or hip OA alone, with similar proportions observed between the latter two groups ( Table 2 ). Furthermore, the use of neuropathic pain medications varied considerably by sex, where the prevalence among male OA patients remained relatively stable across age groups, 1.50% (95% CI 1.37, 1.64). However, middle-aged females had a higher use compared to more elderly females; e.g., prevalence was 3.97% (95% CI 3.41, 4.60) for the 45-54-year-olds while it was 2.30% (95% CI 2.05, 2.58) for the 75-84-year-olds. In individuals without OA, the use was quite stable across age groups for both sexes, 0.80% (95% CI 0.77, 0.83) for males, and for the females the prevalence of use was 1.69% (95% CI 1.59, 1.78) and 1.39% (95% CI 1.25, 1.54) for the age groups 45–54 and 75–84 years, respectively ( Fig. 2 ). Furthermore, this pattern of use remained consistent when analyzing each OA group (knee, hip, or both knee and hip OA), separately ( Appendix 4 ). Table 2 Prevalence of any dispensed neuropathic pain medication across OA groups (knee, hip, both knee and hip, and knee and/or hip OA) and individuals without OA, in total and stratified by sex, with its 95% confidence intervals (CI). Table 2 Prevalence of any neuropathic pain medication Both sexes Males Females Total number Prevalence (%) 95% CI Total number Prevalence (%) 95% CI Total number Prevalence (%) 95% CI Knee and/or hip OA 81,606 1814 (2.22) 2.12–2.33 33,240 499 (1.50) 1.37–1.64 48,366 1315 (2.72) 2.58–2.87 Knee OA only 53,356 1144 (2.14) 2.02–2.27 22,043 326 (1.48) 1.32–1.65 31,313 818 (2.61) 2.44–2.80 Hip OA only 18,506 397 (2.15) 1.94–2.36 7812 103 (1.32) 1.08–1.60 10,694 294 (2.75) 2.45–3.08 Hip and knee OA 9744 273 (2.80) 2.48–3.15 3385 70 (2.07) 1.62–2.61 6359 203 (3.19) 2.77–3.65 No knee or hip OA 564,385 6489 (1.15) 1.12–1.18 280,377 2242 (0.80) 0.77–0.83 284,008 4247 (1.50) 1.45–1.54 Fig. 2 Prevalence of any dispensed neuropathic pain medication with 95% confidence interval in individuals with or without OA for each age group, stratified by sex. Prevalence of any dispensed neuropathic pain medication across OA groups (knee, hip, both knee and hip, and knee and/or hip OA) and individuals without OA, in total and stratified by sex, with its 95% confidence intervals (CI). Prevalence of any dispensed neuropathic pain medication with 95% confidence interval in individuals with or without OA for each age group, stratified by sex. The most frequently used neuropathic pain medication for OA patients was gabapentin, 0.76% (95% CI 0.70, 0.82), whilst duloxetine was the most common among patients without OA, 0.40% (95% CI 0.38, 0.42). Among patients with knee OA only, 18.9% (95% CI 18.6, 19.3) had undergone knee replacement, while 44.8% (95% CI 44.0, 45.5) of those with hip OA only had undergone hip replacement. In females, the difference in OA-related outcomes were influenced by age, necessitating the inclusion of age interactions in the analysis. Therefore, separate regression models were fitted for the sexes. In the logistic regression model adjusting for demographics (age, sex, educational status, and if born in Sweden) the RR of using a neuropathic pain medication was 1.7 (95% CI 1.5, 1.8) in males with OA compared to those without. However, after adjusting for comorbidities (list of comorbidities in Table 1 and Appendix 2 ), the RR was 1.0 (95% CI 0.9, 1.1). Similarly, the risk difference decreased from 0.53% (95% CI 0.40, 0.66) to −0.01% (95% CI -0.10, 0.08) between males with and without OA after adjusting for comorbidities. Furthermore, the pattern of use remained consistent regardless of OA site (knee, hip, both knee and hip OA, or no OA; Appendix 5 ) and when examining the use in OA males with knee or hip replacement ( Appendix 6 ). In females, interactions between OA and age were included in the analysis, with an interaction RR of 0.84 (95% CI 0.80, 0.88) when adjusting for demographics and of 0.88 (95% CI 0.84, 0.93) when also adjusting for comorbidities per 10 years of age difference. The risk of using a neuropathic pain medication was highest among middle-aged females with OA compared to females of the same age groups without OA, with gradually decreasing RR and risk difference in the older age groups. Notably for females, the higher risk of using a neuropathic pain medication persisted in the middle age groups, after adjusting for comorbidities. In contrast, for the oldest age groups, the RR approached 1 and the risk difference zero after adjusting for comorbidities ( Table 3 and Appendix 7 ). Table 3 Risk ratio and risk difference with its 95% confidence intervals (CI) of a dispensed neuropathic pain medication between females with and without OA adjusting for demographics and for demographics + comorbidities. Presented by age group (age in midpoints for the age group intervals in descriptive table). Table 3 Adjusted for demographics Adjusted for demographics and comorbidities Age (years) Risk ratio 95% CI Risk ratio 95% CI 40 3.0 2.6–3.5 1.6 1.4–1.9 50 2.5 2.3–2.8 1.4 1.3–1.6 60 2.1 2.0–2.3 1.3 1.2–1.4 70 1.8 1.7–1.9 1.1 1.0–1.2 80 1.5 1.4–1.7 1.0 0.9–1.1 90 1.3 1.1–1.5 0.9 0.8–1.0 Risk difference % 95% CI Risk difference % 95% CI 40 3.20 2.56–3.84 1.04 0.66–1.42 50 2.32 1.95–2.69 0.68 0.45–0.92 60 1.62 1.42–1.83 0.40 0.26–0.54 70 1.08 0.93–1.23 0.17 0.06–0.28 80 0.65 0.49–0.82 −0.01 −0.14–0.12 90 0.33 0.13–0.53 −0.15 −0.31–0.00 Demographics: age, sex, educational status, and if born in Sweden. Comorbidities: see Table 1 . Risk ratio and risk difference with its 95% confidence intervals (CI) of a dispensed neuropathic pain medication between females with and without OA adjusting for demographics and for demographics + comorbidities. Presented by age group (age in midpoints for the age group intervals in descriptive table). Demographics: age, sex, educational status, and if born in Sweden. Comorbidities: see Table 1 . This pattern of use remained consistent across all OA sites, with middle-aged females having a higher risk of using a neuropathic pain medication compared to the older ones, where the higher use persisted among middle-aged females after adjusting for comorbidities. Notably, the CI for use of neuropathic pain medications across the different OA subgroups largely overlapped ( Fig. 3 and Appendix 8 ). Fig. 3 Risk difference with 95% CI after adjusting for demographics and comorbidities, presented by age, for groups of OA (knee, hip and knee and hip OA), between females with and without OA. Risk difference with 95% CI after adjusting for demographics and comorbidities, presented by age, for groups of OA (knee, hip and knee and hip OA), between females with and without OA. In the analysis of use of neuropathic pain medications in female OA patients with or without joint replacement, an interaction effect was observed for females with OA but without knee and/or hip replacement compared to females without OA (interaction RR 0.89; 95% CI 0.84, 0.95), as well as for females with OA with knee and/or hip replacement compared to females without OA (interaction RR 0.80; 95% CI 0.72, 0.88). The pattern of neuropathic use remained consistent as for previous analyses in females, showing a persistently elevated risk of use among middle-aged females with or without joint replacement after adjusting for comorbidities ( Appendix 9 ).

We aimed to estimate the neuropathic pain medication use as of October 1, 2019. Therefore, we extracted data for all 769,223 individuals aged ≥35 years, residing in the Skåne region, Sweden, as of December 31, 2018, assuming no relocation from the region during 2019, as residential information is provided annually. Furthermore, we required being resident in the region during 2016–2019 and to have at least 4 more years of prior health data available, to ensure the capture of OA diagnoses and potential comorbidities. Furthermore, patients who died before October 1, 2019 (the date of prevalence estimation) and patients with a cancer diagnosis registered in the Skåne Healthcare Register [ 13 ] between September 30, 2014, and September 30, 2019, were excluded, as a cancer diagnosis could obscure or lead to underreporting of other comorbidities, particularly during its more acute phase (a list of cancer diagnoses was compiled based on previous work by Thorlund et al. [ 14 ]). Of the remaining 645,991 individuals, we identified all patients with at least one registered ICD-10 code for OA in the knee (M17) or hip (M16) from a registered physician's visit between January 1, 2004, and September 30, 2018, in the Skåne Healthcare Register. Patients with OA in the knee and/or hip, hereafter referred to only as ‘OA’ for simplicity, were classified as exposed, while those without OA (i.e. without knee and/or hip OA) were classified as unexposed. Information on population characteristics and socioeconomic variables was extracted from the registers held by Statistics Sweden. We did not exclude patients with total knee or hip replacement because we could not rule out that the patients were using neuropathic pain medications due to OA in the other hip or knee joint ( Fig. 1 ). Fig. 1 Flowchart of inclusion of study sample and distribution of individuals with or without OA. Flowchart of inclusion of study sample and distribution of individuals with or without OA. Prescribed and dispensed neuropathic pain medications of amitriptyline (ATC code N06AA09), duloxetine (ATC code N06AX21), gabapentin (ATC code N03AX12) or pregabalin (ATC code N03AX16) for all individuals with or without OA were extracted from the Swedish Prescribed Drug Register for the year 2019. In Sweden, treatment with neuropathic pain medications requires a physician's prescription. Consequently, the register captures dispensed prescriptions for these medications, including prescriptions that may be for on-label or off-label use. However, it does not record the clinical indication for each prescription and cannot confirm whether a dispensed medication was consumed. For the estimation of point prevalence of use of any of these neuropathic pain medications, an individual was categorized as a user on October 1, 2019, when this day was included in the interval between the dispensing date and the theoretical end date of the prescription. The theoretical end date of the prescription was calculated as the dispensing date plus the number of total daily doses dispensed (filled defined daily dose). If there were multiple active dispensations of the same medication on October 1, the one with the closest preceding dispensation date to the index date was chosen. In the case of multiple dispensations of the same medication on the same day, one was chosen randomly. ICD-10 codes for comorbidities associated with treatment of the included neuropathic pain medications were extracted from the Skåne Healthcare Register between September 30, 2016, to September 30, 2019 (i.e. 3 years prior to the index date). The Skåne Healthcare Register contains data on diagnoses, including both the primary reason for consultation and relevant comorbidities, assigned at physician visits within Region Skåne, in primary care as well as in both outpatient secondary care and inpatient hospital care. At each physician visit, the patient's personal identification number and the date of the visit are also recorded. The list of comorbidity diagnoses was compiled through a combination of FDA-approved indications, reporting of off-label use [ 12 , [15] , [16] , [17] ], and Swedish drug treatment guidelines [ 10 , 11 , 18 , 19 ] for these medications ( Appendix 1 , Appendix 2 ). The point prevalence of use of neuropathic pain medications was determined on October 1, 2019. This date was selected a priori as it fell in the middle of the fall semester, avoided school breaks, and occurred in the middle of the week (Tuesday), providing a representative time point for analysis. Prevalence proportions of neuropathic pain medication use with 95% confidence intervals (CI) were reported for persons with or without OA, by age group (in 10-years intervals), and sex (female/male was defined according to each person's Swedish personal identity number). Based on the results from descriptive analyses above, we used a logistic regression model separately for males and females to estimate the use of neuropathic pain medications in knee and hip OA patients compared to individuals without. The model was adjusted for demographics: age (continuous), highest educational level reached (categories), if born in Sweden (yes/no), and if living in an urban area (yes/no). Country of birth was missing for 0.02% and education for 1.14% of the cohort, and persons with missing data were excluded from the adjusted analyses. Additionally, in a later step, we adjusted for comorbidities registered within the 3-year period prior to the index date that could have a prescription of a neuropathic pain medication, e.g. painful diabetic neuropathy, depressive disorder and epilepsy ( Table 1 and Appendix 2 ). The analyses were adjusted for in two stages. The first regression model included adjustment for demographics and the second included adjustment for demographics and comorbidities associated with neuropathic pain medications. By adjusting for comorbidities that may represent indications for the use of neuropathic pain medications, we aim to better identify the proportion of patients with OA who may be using these medications specifically for OA-related pain. Given the low prevalence of the outcome, odds ratios from the logistic regression model can be interpreted as risk ratios (RR). Furthermore, we estimated differences in proportions (i.e. risk differences) with 95% CI from the fitted regression models using command margins . We also performed sensitivity analyses where patients were classified as having OA without joint replacement, or as having OA and joint replacement in at least one joint, to examine whether knee or hip replacement surgeries influenced the results. StataNow 18_5 was used for all analyses. Table 1 Population characteristics, socioeconomic variables, and prevalence of comorbidities for individuals with and without OA. Table 1 No OA Knee and/or hip OA Knee OA only Hip OA only Knee and hip OA Sample size 564,385 (87.4%) 81,606 (12.6%) 53,356 (8.3%) 18,506 (2.9%) 9744 (1.5%) Age (years), mean (SD) 56 (14) 69 (12) 68 (12) 71 (12) 75 (10) Male sex 280,377 (49.7%) 33,240 (40.7%) 22,043 (41.3%) 7812 (42.2%) 3385 (34.7%) Age groups (years)  35-44 144,401 (25.6%) 2019 (2.5%) 1548 (2.9%) 444 (2.4%) 27 (0.3%)  45-54 149,447 (26.5%) 8445 (10.3%) 6527 (12.2%) 1574 (8.5%) 344 (3.5%)  55-64 115,750 (20.5%) 17,071 (20.9%) 12,749 (23.9%) 3100 (16.8%) 1222 (12.5%)  65-74 92,619 (16.4%) 25,303 (31.0%) 16,457 (30.8%) 5814 (31.4%) 3032 (31.1%)  74-84 46,297 (8.2%) 20,110 (24.6%) 11,482 (21.5%) 5185 (28.0%) 3443 (35.3%)  85+ 15,871 (2.8%) 8658 (10.6%) 4593 (8.6%) 2389 (12.9%) 1676 (17.2%) Highest educational level reached, n (%)  Education up to 9 years 95,905 (17.2%) 23,089 (28.6%) 14,595 (27.7%) 5273 (28.7%) 3221 (33.3%)  Education 10–12 years 239,079 (42.9%) 35,879 (44.4%) 23,912 (45.3%) 7798 (42.5%) 4169 (43.1%)  Education 13–14 years 78,799 (14.1%) 9578 (11.9%) 6316 (12.0%) 2239 (12.2%) 1023 (10.6%)  Education 15+ years 144,111 (25.8%) 12,204 (15.1%) 7907 (15.0%) 3047 (16.6%) 1250 (12.9%) Born in Sweden 442,023 (78.3%) 68,822 (84.3%) 44,126 (82.7%) 16,230 (87.7%) 8466 (86.9%) Lives in urban area 271,560 (48.1%) 34,995 (42.9%) 23,365 (43.8%) 7692 (41.6%) 3938 (40.4%) Comorbidities, n (%) . (.) . (.) . (.) . (.) . (.) Arthritis, other 12,759 (2.3%) 4630 (5.7%) 2937 (5.5%) 985 (5.3%) 708 (7.3%) Back pain, most common; including disc hernia, sciatica and back pain unspecified 68,657 (12.2%) 19,036 (23.3%) 11,549 (21.6%) 4532 (24.5%) 2955 (30.3%) Back pain, other 12,365 (2.2%) 5799 (7.1%) 3185 (6.0%) 1480 (8.0%) 1134 (11.6%) Chronic pain 26,816 (4.8%) 33,653 (5.2%) 4490 (8.4%) 1399 (7.6%) 948 (9.7%) Diabetes mellitus with complications 19,848 (3.5%) 6494 (8.0%) 4112 (7.7%) 1441 (7.8%) 941 (9.7%) Headache 35,703 (6.3%) 6130 (7.5%) 4104 (7.7%) 1271 (6.9%) 755 (7.7%) Joint disease, other 34,263 (6.1%) 13,639 (16.7%) 8967 (16.8%) 2713 (14.7%) 1959 (20.1%) Neurological disease 22,266 (3.9%) 6929 (8.5%) 4087 (7.7%) 1677 (9.1%) 1165 (12.0%) Neuropathy 28,116 (5.0%) 7886 (9.7%) 4964 (9.3%) 1678 (9.1%) 1244 (12.8%) OA in other joints a 19,993 (3.5%) 14,906 (18.3%) 8979 (16.8%) 3059 (16.5%) 2868 (29.4%) Other diseases 26,067 (4.6%) 5569 (6.8%) 3682 (6.9%) 1140 (6.2%) 747 (7.7%) Pain or discomfort, unspecified 63,324 (11.2%) 19,133 (23.4%) 12,125 (22.7%) 4192 (22.7%) 2816 (28.9%) Psychiatric disease, most common; including depression, anxiety, stress 115,888 (20.5%) 23,017 (28.2%) 14,870 (27.9%) 5063 (27.4%) 3084 (31.7%) Psychiatric disease, other 35,403 (6.3%) 5555 (6.8%) 3623 (6.8%) 1265 (6.8%) 667 (6.8%) Soft tissue disease, other 56,254 (10.0%) 17,301 (21.2%) 11,156 (20.9%) 3569 (19.3%) 2576 (26.4%) Ulcers 8458 (1.5%) 3097 (3.8%) 1746 (3.3%) 798 (4.3%) 553 (5.7%) a Not included in the regression analyses as it was considered part of generalized OA, potentially also encompassing knee and hip OA. Population characteristics, socioeconomic variables, and prevalence of comorbidities for individuals with and without OA. Not included in the regression analyses as it was considered part of generalized OA, potentially also encompassing knee and hip OA. The use of Swedish register data was approved by the Swedish Ethical Review Authority, Lund, (Dnr 2011_432 with amendment Dnr 2014_276 and Dnr 2018_233).

Using a comprehensive population-based data set, we found that the prevalence of neuropathic pain medication use was higher among patients with OA (in the knee and/or hip) compared to a non-OA population. However, after adjusting for comorbidities associated with use of neuropathic pain medications, to assess the likely use for OA pain management specifically, the risk was no longer elevated among the males and older females with OA. Interestingly, though, a higher prevalence of dispensed neuropathic pain medications persisted among middle-aged females with OA, even after adjusting for comorbidities. Furthermore, we found similar levels of neuropathic pain medication use irrespective of having OA in the knee or hip for both males and females. Our findings suggest that male OA patients primarily used the medications for reasons other than their OA-related pain and similar results were observed for the older females with OA. However, a noteworthy finding was that middle-aged females with OA may often use these medications for OA-related pain. Avoiding potentially harmful treatments, such as opioids, is particularly important in a younger population. This, combined with the limited pharmacological alternatives for the treatment of OA, may potentially explain why physicians explore other, less well-established treatments in this patient category. However, the same pattern was not seen in middle-aged male OA patients. Future research investigating potential differences in treatment effects between males and females, or a potentially higher prevalence of neuropathic pain among females, may provide further insights. Although the prescribing of medications for neuropathic pain has been reported to increase in patients with OA, few studies have specifically examined their use for OA-related pain. However, one such study estimated that approximately 9% of OA patients were prescribed gabapentinoids for OA-related pain. Furthermore, the authors reported that younger females with OA were the most likely to receive gabapentinoid prescriptions, although the specific indication for these prescriptions remained unclear [ 20 ]. Another study reported a lower, yet still notable, use of neuropathic pain medications in a subgroup of OA patients without comorbidities typically associated with such medications, compared to the broader OA population. For the broader OA population, female sex was associated with a higher likelihood of being prescribed antidepressants or anticonvulsants. However, in contrast to our findings, older age was also identified as a risk factor for such prescribing [ 7 ]. In our study, we included a comprehensive list of comorbidities, which may have accounted for residual confounding for males and older females with OA, compared to other studies. Furthermore, our analysis focused specifically on patients with knee and hip OA, utilizing a comprehensive data set from a well-defined geographical area and register resources of both exposure and outcome with high validity [ 21 ]. We encourage future research to replicate similar analyses in other populations and settings. Such efforts would help validate our findings and further investigate the potential factors contributing to the higher use of neuropathic pain medications among middle-aged females with knee and hip OA. This study is subject to certain important limitations. First, since we aimed to estimate the use of neuropathic pain medications specifically for OA-related pain, we adopted a conservative approach by including a broad range of comorbid conditions, thereby accepting potential over-adjustment rather than the risk of omitting relevant confounding diagnoses. By applying such a comprehensive list of comorbidities, we hope to have minimized the likelihood that the higher prescribing observed among patients with OA can be explained by comorbidities. However, we cannot entirely exclude the presence of residual confounding or unidentified interaction effects, potentially particularly relevant among middle-aged females, which may contribute to their persistently elevated use after adjustment. Conversely, this approach may also have led to an underestimation in use of neuropathic pain medications specifically for OA pain, if the drug was prescribed both for a comorbidity and OA pain. Furthermore, we included only patients with a knee and/or hip OA diagnosis recorded during a physician's visit. As a result, patients who have not sought medical care or who only consulted a physiotherapist or other healthcare professionals are not represented in our sample, which accounts for approximately 33% of patients with knee OA [ 22 ]. As we have likely missed patients with milder forms of symptomatic OA, the proportion of patients using neuropathic pain medications in our study might be higher than would be seen in a cohort where all severity levels of OA are included. In Sweden, the diagnosis of knee and hip OA is recommended to be made clinically based on patient history, symptoms, and physical examination of the joint [ 23 ]. Hence, it can be expected that the majority of patients recorded with a diagnosis of knee or hip OA in Sweden have symptomatic OA. However, we do not know how many of these patients also exhibited radiographic changes typically consistent with OA. In conclusion, we found a higher use of neuropathic pain medications among patients with knee and/or hip OA compared to individuals without. Middle-aged females seem to be using neuropathic pain medications for OA-related pain more than others.

During the preparation of this work the author(s) used ChatGPT to assist with the translation of briefer sections from Swedish to English, as well as to improve English grammar in other sections originally written by us in English. After using this tool/service, the author(s) reviewed and edited the content as needed and take(s) full responsibility for the content of the publication.

Nociceptive pain, which results from actual or pending damage to non-neural tissue and mediated by activation of nociceptors, is the most common and well-studied type of pain in patients with osteoarthritis (OA). However, in recent years, the recognition of a multifactorial pain pathogenesis in OA has emerged, including nociplastic pain, arising from altered nociception, and neuropathic pain, associated with structural nerve damage, where 8.2% of knee OA patients have been suggested to have a potential neuropathic pain component [ 1 ]. Previous studies have provided moderate-quality evidence in support of the effect of duloxetine, one of the medications used for neuropathic pain in knee OA [ 2 ]. Several clinical practice guidelines including Osteoarthritis Research Society International (OARSI), American College of Rheumatology, and The Royal Australian College of General Practitioners, have consequently issued conditional recommendations for the use of duloxetine in knee OA, particularly in cases with widespread and persistent pain [ 3 , 4 ]. In contrast, the quality of evidence regarding the effect of duloxetine on hip OA pain is insufficient, and the OARSI conditionally recommends against its use in hip OA [ 5 ]. There are reports of a global increase in prescribed neuropathic pain medications in recent years, both in the general population [ 6 ] and in patients with OA [ 7 ]. However, some of these medications, such as gabapentinoids, have been associated with risk of addiction and serious adverse events, warranting caution in their prescription [ 6 , 7 ]. At the same time, treatment options for OA pain are limited, which may lead to prescribing medications not formally recommended for OA pain, e.g., in patients with contraindications for NSAID use [ 8 ] or when opioid use is to be avoided. The rising prescription rates of neuropathic pain medications, coupled with rising concerns regarding the need for caution in their use, highlight the need to assess and monitor the usage of these medications in OA patients. Furthermore, although recent studies suggest a potential therapeutic effect, they are still only conditionally, if at all, recommended in the major treatment guidelines [ 9 ]. Additionally, these medications have multiple indications beyond pain management, including the treatment of epilepsy and depression [ 10 , 11 ], as well as off-label applications including insomnia and pruritus [ 12 ], meaning that a prescription to a patient with OA may not necessarily be attributable to OA-related pain. Therefore, it is important to gain new knowledge of their use in patients with OA and specifically for OA pain management. Thus, our study aimed to: - estimate the overall prevalence of prescribed neuropathic pain medications among patients with knee and/or hip OA compared to those without, irrespective of comorbidities. - estimate the use of prescribed neuropathic pain medications specifically for OA-related pain in the knee and/or hip by accounting for comorbidities. - examine potential differences in prescribing between patients with knee versus hip OA, as well as variations across sex and age groups. estimate the overall prevalence of prescribed neuropathic pain medications among patients with knee and/or hip OA compared to those without, irrespective of comorbidities. estimate the use of prescribed neuropathic pain medications specifically for OA-related pain in the knee and/or hip by accounting for comorbidities. examine potential differences in prescribing between patients with knee versus hip OA, as well as variations across sex and age groups.

ME reports past consultancy for Novo Nordisk, Genascence, Grünenthal Sweden AB and Key2Compliance AB. AT reports receiving fees as Associate Editor for statistics for Osteoarthritis and Cartilage. AD reports involvement as Chair of OARSI multimorbidity discussion group.