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
Long term use of analgesics and risk of osteoarthritis progressions and knee replacement: propensity score matched cohort analysis of data from the Osteoarthritis Initiative
Address correspondence and reprint requests to: N. Hafezi-Nejad, Russell H. Morgan Department of Radiology, Johns Hopkins University School of Medicine, 601 N. Caroline Street, JHOC 4240, Baltimore, MD 21287, USA. Tel: 1-443-600-3265.
Affiliations
Russell H. Morgan Department of Radiology, Johns Hopkins University, Baltimore, MD, USA
Quantitative Imaging Center, Department of Radiology, Boston University School of Medicine, Boston, MA, USADepartment of Radiology, University of Erlangen-Nuremberg, Erlangen, Germany
To determine the association between the long-term use of analgesics and progression of osteoarthritis (OA) as evidenced by up to 3-years follow-up worsening of radiographic Kellgren–Lawrence (KL) grade and incidence of knee replacement (KR).
Design
Using nearest neighbor matching of the propensity scores with caliper in the Osteoarthritis Initiative (OAI) cohort, 173 index (Analgesic +) and 173 referent (Analgesic −) subjects were included. Analgesic + and − subjects had analgesics in all and none of their visits, respectively. Analgesic + and − subjects were balanced in their demographics, baseline, first, second and third year body mass index (BMI), Western Ontario and McMaster (WOMAC) total score, Physical and Mental health summary scales (SF-12), Physical Activity Scale for the Elderly (PASE) and Charleston Comorbidity Scale. Analgesic + and − subjects were also matched for baseline radiographic KL grade. Interval increase in the KL grade and incidence of KR were defined as the outcome.
Results
Included subjects had average 6.5 years of follow-up. By the third year, 44 subjects had an interval increase in the KL grade; 29 in Analgesic + and 15 among Analgesic − subjects (P = 0.024). By the eighth-year, 41 subjects had their first KR; 29 in Analgesic + and 12 among Analgesic − subjects (P = 0.005). Hazard Ratio (HR) of OA progression and KR for Analgesic + subjects was 1.91 (1.02–3.57) and 2.57 (1.31–5.04), respectively.
Conclusions
Long-term use of analgesics may be associated with radiographic progression of knee OA and increased risk of future KR.
. Knee OA has a progressive course, which is associated with gradually worsening symptoms such as chronic pain, stiffness and restricted range of motion
Risk factors predictive of joint replacement in a 2-year multicentre clinical trial in knee osteoarthritis using MRI: results from over 6 years of observation.
An algorithm recommendation for the management of knee osteoarthritis in Europe and internationally: a report from a task force of the European Society for Clinical and Economic Aspects of Osteoporosis and Osteoarthritis (ESCEO).
An algorithm recommendation for the management of knee osteoarthritis in Europe and internationally: a report from a task force of the European Society for Clinical and Economic Aspects of Osteoporosis and Osteoarthritis (ESCEO).
American College of Rheumatology 2012 recommendations for the use of nonpharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and knee.
. Analgesics are the largest group of medications that are prescribed for OA patients. A variety of analgesics including non-steroidal anti-inflammatory drugs (NSAIDS), Cyclooxygenase II (COX II) inhibitors, acetaminophen, salicylates and narcotic analgesics are used for symptom relief
Is there an association between the use of different types of nonsteroidal antiinflammatory drugs and radiologic progression of osteoarthritis? The Rotterdam Study.
First-line analysis of the effects of treatment on progression of structural changes in knee osteoarthritis over 24 months: data from the osteoarthritis initiative progression cohort.
Longer use of COX-2-specific inhibitors compared to nonspecific nonsteroidal antiinflammatory drugs: a longitudinal study of 3639 patients in community practice.
. Results of recent observational studies were more controversial. While some reports have confirmed animal studies on negative effects, others have claimed a protective role, especially for the selective COX-II inhibitors
Longer use of COX-2-specific inhibitors compared to nonspecific nonsteroidal antiinflammatory drugs: a longitudinal study of 3639 patients in community practice.
Efficacy, tolerability, and upper gastrointestinal safety of celecoxib for treatment of osteoarthritis and rheumatoid arthritis: systematic review of randomised controlled trials.
. Interestingly, long-term use of NSAIDs was associated with decreased joint space narrowing (JSN). In contrast, another first line analysis of the Osteoarthritis Initiative (OAI) cohort by Pelletier et al. reported a lower joint space width among analgesic + subjects
First-line analysis of the effects of treatment on progression of structural changes in knee osteoarthritis over 24 months: data from the osteoarthritis initiative progression cohort.
Inadequate pain relief and large functional loss among patients with knee osteoarthritis: evidence from a prospective multinational longitudinal study of osteoarthritis real-world therapies.
In this study, our objective was to investigate the long-term use of analgesics in association with radiographic progression of OA and subsequent KR as the ultimate structural outcome of knee OA.
Methods
Study design, setting and participants
For this analysis, we used the data from the OAI. The OAI is a cohort of 4796 subjects with or at risk of OA, which are regularly followed on an annual basis. Datasets are available for public access through the OAI website at the http://www.oai.ucsf.edu. Details of the study participants, enrollments, evaluations and follow-ups are also available for public access from the OAI website. The primary hypothesis of the study was to find the maximum number of the index (Analgesic +) and referent (Analgesics −) subjects which have the most similar follow-up assessments except their exposure to analgesics.
Exposure: long-term use of analgesics
Our definition of Analgesic + was documented use of analgesics in all available follow-ups. Physician-prescribed medications were adjudicated and were present in the OAI datasets. Following a recently published study, we used prescription analgesics as it's more reflective of its regular use especially in clinically relevant doses (compared to non-prescription analgesics)
. For comparison, we aimed at selecting our referent subjects from those with no report of prescription analgesics in neither of their follow-up visits. Thus, we checked the Medication Inventory Form (MIF) of each subject. Out of the total 4796 participants, 2654 subjects had analgesics in their MIF in “some” of the visits. The remaining 2142 subjects had analgesics in their MIF in “all”, or “none” of the follow-up visits.
Outcomes: radiographic Kellgren and Lawrence (KL) grade and KR
We had two distinct outcomes of interest. First, increase in the follow-up KL grade compared to the matched baseline KL grade (only between grade changes) as a measure of radiographic OA progression
Comparative evaluation of three semi-quantitative radiographic grading techniques for hip osteoarthritis in terms of validity and reproducibility in 1404 radiographs: report of the OARSI-OMERACT Task Force.
Risk factors predictive of joint replacement in a 2-year multicentre clinical trial in knee osteoarthritis using MRI: results from over 6 years of observation.
, the increase in KL grade was evaluated up to the third year of follow-up. The higher KL grade of the two knees was selected as the variable of interest for each participant
. KR incidence was assessed throughout the 8 years of follow-up using the latest updated datasets that were present online (Datasets were accessed at November 2014). For the purpose of consistency and to avoid issues that may arise from inconsistent or unstandardized readings, we used OAI's central readings of X-rays for KL grade.
Matching variables
In order to define matched referents, for every index, we found one matched referent using nearest neighbor matching of the propensity scores with calipers. For every index, we selected the one best referent who matched with the index in not only the baseline, but the follow-up variables as well (except radiographic KL grade, which was matched only for the initial baseline values). Therefore, we attempted to find subjects with a similar course of OA while having different exposures (long term use of analgesic vs no analgesics: Analgesic +/−). Variables that could possibly affect the probability of using analgesics were considered in finding the matching referents
Risk factors predictive of joint replacement in a 2-year multicentre clinical trial in knee osteoarthritis using MRI: results from over 6 years of observation.
. Demographic variables, body mass index (BMI), Western Ontario and McMaster Questionnaire (WOMAC) total score, Physical Summary Scale (PSS) for the Medical Outcome Study (MOS) 12 item short form (SF-12) health survey, mental summary scale (MSS) for the MOS 12 item short form (SF-12) health survey, Physical Activity Scale for the Elderly (PASE) score, and Charlson comorbidity score included the matching variables. Index and referent subjects (Analgesic + and Analgesic −) were selected to have initial matched KL grades, to compare follow-up changes in radiographic KL grades. The greater WOMAC score of the two knees was selected as the variable of interest for each participant
Association of magnetic resonance imaging-based knee cartilage T2 measurements and focal knee lesions with knee pain: data from the Osteoarthritis Initiative.
. After excluding 384 subjects with missing values in either the demographics, KL grades or the Charleston comorbidity scales, 1758 subjects remained. No participant had missing values in neither the exposure nor the outcome. Following exclusion of the above incomplete index cases, 125 of the remaining 61,405 variable-values, which were accounting for 0.002 of all values, were missing. These 125 missing values were imputed using expectation maximization (EM) technique, which provided us with one completely imputed dataset (no remaining missing values)
The imputed dataset was used for the purpose of matching. In order to determine referent subjects (Analgesic −) for each index (Analgesics+), we searched the dataset for the one nearest neighbor, non-replaced match
. The index (Analgesic +) and referent (Analgesic −) subjects had analgesics in all and none of their visits, respectively. To select the best available referents, propensity score matching was utilized. A 1:1 greedy matching algorithm was employed similar to the method described by Rosenbaum et al.
. For every participant in the index (Analgesic +), one best-matched referent (Analgesic −) participant was selected. Using greedy matching algorithm, we found the first best match pair and then moved to find the next best match in a hierarchical manner. Best match was defined as subjects with the highest digit match on their propensity scores, which were calculated using logistic regression analyzes to predict the probability of exposure (long-term use of analgesics). A caliper of 0.1 was employed in finding the highest digit match of the propensity scores. As mentioned earlier, out of the total 4796 participants, 2654 subjects had analgesics in their MIF in “some” of the visits. The remaining 2142 subjects had analgesics in their MIF in “all”, or “none” of the follow-up visits; 384 subjects were excluded due to having missing values in either the demographics, KL grades or the Charleston comorbidity scales, and 1758 subjects remained. By exclusion of subjects outside common support of both groups and by considering the caliper, of 187 index subjects (Analgesic +) 14 index cases were excluded and best matches were selected for 173 Analgesic + index subjects, resulting in 173 Analgesic − referents. This method of propensity matching and use of calipers has been extensively reviewed by Austin et al.
and is demonstrated to induce minimal, if any, bias for the comparisons. Fig. 1 shows the inclusion of subjects from the 4796 OAI participants. Table I demonstrates the characteristics of the index and referent subjects (Analgesic + and Analgesic −). There was no significant difference in any of the included variables in the matching process between the two groups neither in the average, standardized average or average ranks.
Fig. 1Flowchart outlining inclusion of the study subjects.
Comparison of index and referent subjects (Analgesic + and Analgesic −)
Cumulative numbers of subjects with an increase in KL grade and/or KRs were compared between the two groups using Chi-square (X2). Fisher's exact test was used whenever the expected number of events was less than five in a two by two table. For a clear description of the number of participants with OA progression throughout the follow-up, year by year analyzes were performed. Number of subjects with OA progression at each time point were calculated from the total of 173 baseline subjects. Cox regression analysis was employed to extract the Hazard Ratio (HR) of increase in KL grade (OA progression) and KR associated with long-term use of analgesics. Additional covariate adjustment was performed using the previously obtained propensity scores
. However, we took a conservative stance (avoiding type I error) and calculated the adjusted HRs as well. A two-tailed P value of less than 5% was considered significant. Analyzes were performed using SPSS (v.18, Chicago, IL) and the R platform (v. 2.8.1). We note that the aim of this investigation was to examine an association (rather than depicting a cause and effect graph) while dealing with the imbalance in a series of known determinants; using a relatively matched sample of subjects, in the OAI cohort.
Results
Participants and descriptive data
The mean age was 61.1 ± 8.74 and 61.0 ± 8.62 among the Analgesic + and Analgesic − subjects, respectively. Sixty-eight percent of Analgesic + and seventy-four percent of Analgesic − subjects were females. Table I demonstrates the standardized mean difference of the matching variables between Analgesic + and Analgesic − subjects. The standardized mean difference is used to represent the size of each variable's difference, relative to its variability, between Analgesic + and Analgesic − groups. The standardized mean difference is calculated by dividing the mean difference in each variable's value by the standard deviation of that variable in the studied group. There was no statistical difference in the demographics, BMI, WOMAC score, PSS, MSS and comorbidity scales of the two groups; neither at the baseline, first year, second year or third year follow-ups (Table I). The mean KL grades at the baseline radiographs were 1.82 ± 1.25 and 1.92 ± 1.21 for Analgesic + and Analgesic − subjects, respectively (P value = 0.43) (Table I). There was no significant difference in the distribution of the subjects among different KL grades (P values > 0.05). For both the Analgesic + and Analgesic − subjects, median and inter-quartile ranges of KL grades were 2 (1–3). The mean KL grade at the third year of follow-up was 1.94 ± 1.29 and 2.00 ± 1.24 for Analgesic + and Analgesic − subjects, respectively (P-value: 0.61). Table II demonstrates the details about the type of analgesics as reported in the follow-up reports. As it is evident in Table II, many Analgesic + subjects receive more than one type of analgesics. Moreover, patients frequently change the type of their analgesics in long follow-ups.
Table IIDetails of the prescription analgesics during the years of follow-up
At the end of third year, 16.76% (29/173) of Analgesic + subjects, and 8.67% (15/173) of the Analgesic − subjects had OA progression in terms of increased KL grades (P value = 0.024). By the eighth year of follow-up, 16.76% (29/173) and 6.94% (12/173) subjects had KR among the Analgesic + and Analgesic − subjects, respectively (P value = 0.005). Long-term use of analgesics was associated with 1.91 (Adjusted HR 95% CI: 1.02–3.57) and 2.57 (Adjusted HR 95% CI: 1.31–5.04) times greater risk of OA progression and KR (Table III, Table IV). Long-term use of analgesics was associated with significantly higher risk of KR even when controlled for the effect of increase in the KL grade in the first 3 years (HR: 2.30 (1.16–4.54); P-value: 0.017). In other words, use of analgesics was associated with KR among subjects with and without KL grade increase in the first 3 years.
Table IIICumulative number of subjects with increase in the radiographic KL scores and KR according to the study group (Analgesic + vs Analgesic −)
We performed an additional analysis of the recently released eighth year follow-up readings to further test our hypothesis. In the recently updated OAI project, reports of the KL grading on 264 (out of 346; 76.3% of included subjects) subjects were not available for the eighth year of follow-up. Thus, we continued our analysis knowing that our study power may be diminished due to small sample size. However, results of the available eighth year KL grades showed that 16 (out of 43) subjects in the Analgesics + and seven (out of 39) subjects in the Analgesic − group had OA progression (increase in KL grade compared to the baseline radiography). The comparison was significant with a P value (using Chi–square test) of 0.05. Other determinants had no significant difference among the Analgesic + and − group: Age (58.35 ± 8.44 vs 57.23 ± 7.63; P-value: 0.53), Gender (72% vs 85%; P value: 0.17), BMI (27.06 ± 4.44 vs 26.39 ± 3.86; P-value: 0.47), PASE (164.91 ± 73.74 vs 166.63 ± 77.91; P-value: 0.92), WOMAC (9.32 ± 9.71 vs 12.16 ± 15.10; P-value: 0.32) and baseline KL grade (0.26 ± 0.44 vs 0.33 ± 0.48; P-value: 0.45).
We also investigated the OAI cohort for the availability of data on the use of steroid injections in the symptomatic knees. Such information is collected on a yearly basis in the annual follow-ups (from baseline to the seventh year follow-up). Subjects are asked about steroid injections within the last 6 months. In the whole OAI cohort (n = 4796), only 1.94% (baseline) to 4.45% (seventh year – the last available data on steroid injections) of subjects had reports of using steroid injections. In this study (n = 346), number of subjects who had steroid injections was 3.2% (baseline) and 6.4% (seventh year – the last available data on steroid injection), respectively; which was not statistically significant among Analgesic + and Analgesic − subjects (P > 0.05). Moreover, WOMAC scores for pain, stiffness, and disability were analyzed separately to investigate whether Analgesic + and Analgesic − subjects have significantly different sub-scores, or not. Our analyzes revealed that none of the sub-scores had neither statistically (all P values > 0.10) nor clinically significant difference among Analgesic + and Analgesic − subjects (Supplementary Table 1).
Discussion
Key results
In this study, our findings suggest that the increased risk of KR in the index (Analgesic +) subjects may be at least partly explained by OA progression. Subjects with long-term use of analgesics were more likely to undergo KR in comparison to subjects with similar OA severity (OA severity was similar between the Analgesic + and Analgesic − subjects at the baseline). The most important discussion point is that the indication for KR is made clinically and is not only based on radiographs. If the patient has knee OA but is clinically asymptomatic, he/she does not receive joint replacement while a patient with persistent pain and analgesics consumption will receive joint replacement much earlier. Interestingly, other studies show that prescription analgesic consumption may even increase after the KR surgery
. These findings suggest new determinants of analgesic consumption, regardless of OA severity and progression. In addition to their analgesic effects, different analgesics might be associated with different pathophysiologic effects. Local anti-inflammatory effects (in case of ibuprofen)
, are of the most investigated mechanisms. Besides pathophysiologic explanations, exploring the differences in the level of satisfaction and expectations of patients in decision making for KR may potentially explain the differences between Analgesic + and Analgesic − subjects
. Furthermore, especially with regard to the incidence of KR, economic discrepancies and impact of healthcare access, information and education are of the deterministic factors in the management of OA; as expressed by the European Society for Clinical and Economic Aspects of Osteoporosis and Osteoarthritis (ESCEO)
An algorithm recommendation for the management of knee osteoarthritis in Europe and internationally: a report from a task force of the European Society for Clinical and Economic Aspects of Osteoporosis and Osteoarthritis (ESCEO).
Our results are similar to previous reports that highlighted the deleterious effects of analgesics, particularly non-selective NSAIDs, in the course of OA
Is there an association between the use of different types of nonsteroidal antiinflammatory drugs and radiologic progression of osteoarthritis? The Rotterdam Study.
. In this study, Analgesic + subjects with similar WOMAC scores (including pain, stiffness and disability) and initial radiographic KL grade had a higher risk of OA progression and KR. Recently, Lapane et al. evaluated the use of NSAIDs on symptoms and OA progression among subjects with knee OA
. They analyzed the OAI progression cohort using marginal structural modeling to control for the time variant variables including pain, body weight, physical and mental health status. Interestingly, long-term use of NSAIDs was associated with decreased JSN, albeit not statistically significant. In contrast, another first line analysis of the OAI cohort by Pelletier et al. reported a lower joint space width among analgesic + subjects
First-line analysis of the effects of treatment on progression of structural changes in knee osteoarthritis over 24 months: data from the osteoarthritis initiative progression cohort.
Inadequate pain relief and large functional loss among patients with knee osteoarthritis: evidence from a prospective multinational longitudinal study of osteoarthritis real-world therapies.
. Compared to previous studies, we investigated the effect of the long-term analgesic use on changes in follow-up radiographic KL grade measurements as an objective metric for OA progression. Subjects included in our study had NSAIDs among their prescriptions for the majority of the follow-up period as well. Certain differences in the choice of exposure (Analgesics in general vs class specific drug effects; long-term exposure vs short term exposure) study design (only including subjects with balanced symptoms vs including subjects with improved symptoms) and outcome (KR as a clinical decision based on patients' preference; radiographic KL grade that includes joint space, sclerosis, osteophytes and joint deformity vs JSN) contribute to the different findings among different studies.
We observed a meaningful difference, between Analgesic + and − subjects, for OA progression and KR. Likewise, a previous report of the Rotterdam Study indicated an increased risk of OA progression in subjects who were using diclofenac for more than 180 days
Is there an association between the use of different types of nonsteroidal antiinflammatory drugs and radiologic progression of osteoarthritis? The Rotterdam Study.
. Our results may further expand this observation. Even with similar pain, stiffness and activity profiles, those who achieve these profiles with the aid of analgesics, accompanied an increased risk of OA progression in the OAI cohort; whereas, structural joint damages were also shown to predict the risk of future KR
. In other words, having similar WOMAC score while using analgesics may further contribute to the difference between the Analgesic + and Analgesic − subjects. Aside from the theoretical independent contribution of the analgesic drugs to progression of OA, our findings may also speculate the potential role of pain perception in protection from hazardous positions, activities and weight bearings
Subjects with higher physical activity levels have more severe focal knee lesions diagnosed with 3T MRI: analysis of a non-symptomatic cohort of the osteoarthritis initiative.
that can accelerate OA progression. Despite having balanced Analgesic + and − populations, this study does not explain why some of the subjects were prescribed NSAIDs over the others. We could show that the Analgesic + and − groups were balanced in their symptoms (pain, stiffness and disability). Whether index subjects were less tolerant to pain than referents and whether that was the reason why they chose KR more often, is not elucidated in this study. Even beyond OA progression, Analgesic + subjects were more frequent candidates for KR; further suggesting their analgesic seeking behavior and their need to accept new alternatives interventions in addition to analgesics for the treatment of painful OA. In this regard, Collins et al. argued that subjects with symptomatic knee OA (KL grade ≥ 2) had little changes in their pain; during their 6-year follow-up study of the OAI cohort. Depression, medical comorbidities, and lower education (among others) were suggested to be associated with pain trajectories
Inadequate pain relief and large functional loss among patients with knee osteoarthritis: evidence from a prospective multinational longitudinal study of osteoarthritis real-world therapies.
Inadequate pain relief and large functional loss among patients with knee osteoarthritis: evidence from a prospective multinational longitudinal study of osteoarthritis real-world therapies.
This study had several limitations. Because of the number of available cases we could not draw meaningful comparisons for different types of analgesics. Table II demonstrates the details of analgesics used in each follow-up. The same is true for the number of the analgesics that were used by each participant at each time point (each follow-up) and during the entire follow-up period. Our study design made it impossible to investigate the number of analgesics in relation to OA progression and KR. Future investigations, using a larger number of Analgesic + subjects, should be designed to address this issue. Considering the WHO three ladder categorization of analgesics, we used all three categories in our definition of analgesic use
An algorithm recommendation for the management of knee osteoarthritis in Europe and internationally: a report from a task force of the European Society for Clinical and Economic Aspects of Osteoporosis and Osteoarthritis (ESCEO).
American College of Rheumatology 2012 recommendations for the use of nonpharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and knee.
. While different analgesics may have heterogeneous anti-inflammatory and disease-modifying effects, they all share the common property of reducing pain perception
An algorithm recommendation for the management of knee osteoarthritis in Europe and internationally: a report from a task force of the European Society for Clinical and Economic Aspects of Osteoporosis and Osteoarthritis (ESCEO).
American College of Rheumatology 2012 recommendations for the use of nonpharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and knee.
First-line analysis of the effects of treatment on progression of structural changes in knee osteoarthritis over 24 months: data from the osteoarthritis initiative progression cohort.
Influence of flare design on symptomatic efficacy of non-steroidal anti-inflammatory drugs in osteoarthritis: a meta-analysis of randomized placebo-controlled trials.
. In addition to the common use of analgesic medications in OA patients, there is a large variation in physician's recommendations and patients' consistency in their choice of medications
. Moreover, many subjects use multiple analgesics simultaneously and commonly switch between the types of analgesics during the course of symptomatic OA
. For all the aforementioned reasons, it was not practical to find adequate samples of subjects with the long-term use of any specific analgesic who also share a similar course of OA. Moreover, this study compares the extremes ends (no use vs persistent and long term use) of a spectrum of subjects using analgesic medications. These findings may not be achieved in subjects with minimal or short-term use of analgesics. We believe that future studies should reveal whether there is a significant difference among different analgesics for OA progression owing to their differential actions. Another limitation is the lack of analgesic seeking behavioral metrics in patients who are on analgesics. Finally, we note that especial caution should be made when applying our conclusions to other, possibly community dwelling populations. The analgesic + population in our study had a history of long-term analgesic use. Meanwhile, many other subjects in the OAI cohort may have used analgesics only in some of the follow-up years [Fig. 1], and thus were excluded from this analysis. In fact, in a trade off, we achieved finely tuned comparisons by assuming a condition that subjects had used analgesics in all of their follow-up visits.
Considering recent studies elucidating the effect of different analgesics on different aspects of OA progression warrants further investigations using advanced imaging modalities (including Magnetic Resonance Imaging – MRI) for more sensitive and specific detection of the early signs of knee OA. Larger studies may unveil the specific KL grading component that may be affected in Analgesic + subjects (sclerosis vs JSN vs osteophytes). Our next planned study is to investigate the MRI findings of Analgesic + and Analgesic − subjects, looking for signs of subtle cartilage damage, bone marrow lesions, synovitis and other structural knee damages that may not be apparent in knee radiographs. The ultimate question is whether individuals with more severe degenerative changes on MRI take more analgesics or (vice versa) that analgesics lead to progression of OA (the direction we proposed in this investigation).
Conclusions
In conclusion, our results indicate that, after controlling the effect of potential time-varying confounders, long-term use of analgesics can be associated with radiographic progression of knee OA and increased risk of future KR.
Authors' contributions
All authors have made a substantial contribution in the conception, design or acquisition of data, interpretation of data, and critically reviewing the draft of the article and approving the final version. In addition, NHN had a role in the design, analysis and drafting the manuscript. AG and FWR had an additional role in the interpretation of data, critical review of the manuscript and provision of study materials. JE had an additional role for statistical revision, intellectual revision of the manuscript and interpretation of the results. BZ and SD had an additional role in the conception of the study, provision of the analyzes and final preparation of the manuscript.
Conflicts of interest
Nima Hafezi-Nejad has no conflicts of interest. Ali Guermazi is president and shareholder of Boston Imaging Core Lab, LLC. He is a consultant to Genzyme, MerckSerono, TissueGene and OrthoTrophix. Frank W Roemer is CMO and shareholder of Boston Imaging Core Lab, LLC. John Eng and Bashir Zikria declare no conflicts of interest. Shadpour Demehri has grants from GERRAF 2014–2016; Carestream Health Inc. 2013–2015 for Cone – Beam CT clinical trial. He is a consultant to Toshiba Medical Systems.
Acknowledgments
The OAI is a public-private partnership comprised of five contracts (N01-AR-2-2258; N01-AR-2-2259; N01-AR-2-2260; N01-AR-2-2261; N01-AR-2-2262) funded by the National Institutes of Health, a branch of the Department of Health and Human Services, and conducted by the OAI Study Investigators. Private funding partners include Merck Research Laboratories; Novartis Pharmaceuticals Corporation, GlaxoSmithKline; and Pfizer, Inc. Private sector funding for the OAI is managed by the Foundation for the National Institutes of Health. This manuscript was prepared using an OAI public use data set and does not necessarily reflect the opinions or views of the OAI investigators, the NIH, or the private funding partners.
Appendix A. Supplementary data
The following is the supplementary data related to this article:
Risk factors predictive of joint replacement in a 2-year multicentre clinical trial in knee osteoarthritis using MRI: results from over 6 years of observation.
An algorithm recommendation for the management of knee osteoarthritis in Europe and internationally: a report from a task force of the European Society for Clinical and Economic Aspects of Osteoporosis and Osteoarthritis (ESCEO).
American College of Rheumatology 2012 recommendations for the use of nonpharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and knee.
Is there an association between the use of different types of nonsteroidal antiinflammatory drugs and radiologic progression of osteoarthritis? The Rotterdam Study.
First-line analysis of the effects of treatment on progression of structural changes in knee osteoarthritis over 24 months: data from the osteoarthritis initiative progression cohort.
Longer use of COX-2-specific inhibitors compared to nonspecific nonsteroidal antiinflammatory drugs: a longitudinal study of 3639 patients in community practice.
Efficacy, tolerability, and upper gastrointestinal safety of celecoxib for treatment of osteoarthritis and rheumatoid arthritis: systematic review of randomised controlled trials.
First-line analysis of the effects of treatment on progression of structural changes in knee osteoarthritis over 24 months: data from the osteoarthritis initiative progression cohort.
Inadequate pain relief and large functional loss among patients with knee osteoarthritis: evidence from a prospective multinational longitudinal study of osteoarthritis real-world therapies.
Comparative evaluation of three semi-quantitative radiographic grading techniques for hip osteoarthritis in terms of validity and reproducibility in 1404 radiographs: report of the OARSI-OMERACT Task Force.
Association of magnetic resonance imaging-based knee cartilage T2 measurements and focal knee lesions with knee pain: data from the Osteoarthritis Initiative.
Subjects with higher physical activity levels have more severe focal knee lesions diagnosed with 3T MRI: analysis of a non-symptomatic cohort of the osteoarthritis initiative.
Influence of flare design on symptomatic efficacy of non-steroidal anti-inflammatory drugs in osteoarthritis: a meta-analysis of randomized placebo-controlled trials.
01374-6&id=gr1.jpg){kind=link}