Cost-Effectiveness of Chondroitin Sulphate (with or without Glucosamine) Prescription in the Treatment of Knee Osteoarthritis Compared to NSAIDs and COXIBs

Introduction

Knee osteoarthritis (KOA) is the most common type of osteoarthritis (OA) in adults, characterized by a high rate of disability and poor quality of life.^1–3 Chondroitin sulfate alone (CS) (Condrosan^®/Condrosulf^®/Condroitin sulfato Kern Pharma^®) or associated with glucosamine (CS+GLU) (Droglican^®) (CS/CS+GLU) is one of the main constituent elements of cartilage, being indicated (alone or associated with glucosamine) in the symptomatic treatment of osteoarthritis.^4,5 Treatment of KOA with CS/CS+GLU has been associated with improvement in pain^6,7 and joint function.^8,9 Its efficacy has been confirmed in systematic reviews and meta-analyses^6,7,9–12 as well as in the MOVES,¹³ MOSAIC¹⁴ and CONCEPT¹⁵ clinical trials, in which CS/CS+GLU proved to be as effective as celecoxib after 6 months, but with a better safety profile.

On the other hand, as demonstrated in a meta-analysis,¹⁶ CS/CS+GLU is well tolerated. In fact, CS/CS+GLU and glucosamine were not associated with an increased risk of adverse events (AE) compared to placebo.¹⁶ This good tolerability of CS/CS+GLU contrasts with the toxicity problems described for NSAIDs/COXIBs¹⁷ at gastrointestinal,¹⁸ cardiovascular,^19,20 cerebrovascular²¹ and renal²² levels. The detrimental effects of continued NSAID use in osteoarthritis patients are well described in scientific literature. A 2019 study²³ involving more than 7,000 patients with osteoarthritis showed they had a 23% increased risk of developing cardiovascular disease compared to controls; and that a 41% increase in this risk was directly linked to NSAID use.

CS/CS+GLU is currently recommended for the treatment of KOA by, among others, the guidelines of the Spanish Society of Rheumatology (SER),^24,25 the European League Against Rheumatism (EULAR)²⁶ and those published by the European Society for Clinical and Economic Aspects of Osteoporosis, Osteoarthritis and Musculoskeletal Diseases (ESCEO).²⁷ Moreover, cardioprotective and cerebroprotective effect, based on reduced risk of acute myocardial infarction and ischemic stroke, revealed by Mazzucchelli et al, adjusts to the safety profile of OA patients, which usually have comorbidities that imply cardiovascular risk, similar as metabolic syndrome patients.²⁸

In 2020, an economic analysis was published which concluded that, “due to its better tolerability profile, treatment with CS/CS+GLU is expected to prevent thousands of adverse events over the next three years in Spain, some of which may be life-threatening, while generating considerable savings for the Spanish National Health System. (NHS)”²⁹. This study, however, was limited to the analysis of the costs and consequences of the better tolerability of CS/CS+GLU, but did not assess its cost-effectiveness compared to non-steroidal anti-inflammatory drugs (NSAIDs) and cyclooxygenase 2 inhibitors (COXIBs), ie the cost of quality-adjusted life years gained (QALYs) (calculated from the so-called “utilities”, the quality of life perceived by the patient) and did not include the cardioprotective (reduced risk of acute myocardial infarction) and cerebroprotective (reduced risk of ischemic stroke) effects revealed in two nested case-control studies, published in 2021 and 2022, respectively.^20,21

The aim of the present study was to estimate the cost-effectiveness of CS/CS+GLU compared with NSAIDs/COXIBs in the treatment of KOA, because of avoiding mild-moderate or severe gastrointestinal adverse events (GIAE), ischemic heart disease (IHD), acute kidney injury (AKI), chronic kidney failure (CKF) and ischemic stroke (IS) in Spain.

Methods

Economic Model

Based on the current situation (part of KOA patients treated with CS/CS+GLU and part treated with NSAIDs/COXIB), two analyses were performed: (i) To determine the savings to the NHS from current CS/CS+GLU treatments, it was hypothesized that all patients treated with CS/CS+GLU would be switched to NSAIDs/COXIBs. The additional costs that could result from this change would be the savings currently being generated by CS/CS+GLU treatments, due to their improved safety profile. (ii) To determine the maximum savings that could be achieved by the NHS, it was hypothesized that all patients with KOA currently treated with NSAIDs/COXIB would be switched to CS/CS+GLU.

In both analyses the following calculations were performed: a) estimation of the number of AE avoided with CS/CS+GLU in the Spanish population with KOA; b) estimation of the economic impact derived from the AE avoided with CS/CS+GLU in the Spanish population with KOA and probability of savings for the NHS for this concept; c) savings per patient with KOA associated with the utilization of CS/CS+GLU; d) QALYs gained per patient with KOA associated with the use of CS/CS+GLU; e) cost of gaining 1 QALY with CS/CS+GLU versus the hypothetical non-utilization of the treatment; f) probability that treatment with CS/CS+GLU is cost-effective (cost of gaining 1 QALY below €25,000) and/or dominant (with lower costs and gain in QALYs than without CS/CS+GLU).

A probabilistic model was performed using second-order Monte Carlo simulations,^30–32 with a hypothetical cohort of 1,000 iterations (patients, lato sensu), for a 3-year time horizon. This methodology allows the analysis of the uncertainty of the model variables.³² The costs and utilities (continuous variables) were adjusted to gamma distributions; the probabilities of AE (dichotomous variables) were adjusted to beta distributions, in both cases according to the minimum and maximum values described in the literature.^30,31 The annual variables of the model are summarized in Table 1 (probabilities) and Table 2 (costs and utilities). The number of adverse events avoided was obtained from the population treated with CS/CS+GLU or NSAID/COXIB. The probability of each adverse event with each treatment was calculated, using mean values and their 95% confidence intervals, through a second-order Monte Carlo simulation with 1,000 analyses.

Table 1 Annual Probabilities of Adverse Events with the Treatments Considered (Beta Distributions)

Table 2 Costs and Utilities Considered in the Model (Gamma Distributions)

Population

The evolution of a hypothetical cohort of patients with KOA was modelled. The prevalence of KOA in Spain, in patients over 40 years of age, was obtained from the EPISER study of the Spanish Society of Rheumatology⁴⁶ (Table 3). The population over 40 years of age in Spain, projected for 2025–2027, was obtained from National Institute of Statistics database⁴⁷ (Table 3). The percentage of patients with KOA following pharmacological treatment was obtained from the EPISER study⁴⁶ (Table 3). The percentage of patients with symptomatic KOA treated regularly or occasionally with NSAIDs/COXIBs was obtained from two Spanish studies^48,49 (Table 3). “Regular” use was understood to mean a medication possession ratio (MPR) ≥ 50% and “occasional” use with an MPR ≥ 25% and <50%.⁴⁸

Table 3 Population Estimates for Spain Considered in the Study

Based on this data and the number of units sold in the year 2024 CS/CS+GLU (IQVIA market data, 2025), the number of patients with KOA treated with NSAIDs/COXIBs and CS/CS+GLU in Spain and in the Spanish regions was calculated (Table 3). Probabilistic analyses were performed based on the minimum and maximum values shown in Table 3. The details of the population calculations are presented in Table 3.

The two analyses were performed at the national level, with a sub-analysis also being conducted at the regional level.

Time Horizon

The simulation covered a period of 3 years (2025 to 2027). Probabilities, costs and utilities were calculated for annual cycles.

Perspective of the Analysis

The analysis perspective was that of the Spanish NHS, so only direct health costs were considered.

Costs Analyzed

The following costs were analyzed: (i) the cost of managing AE (mild-moderate or severe GIAE, IHD, AKI, CKF and IS); (ii) the drugs acquisition cost (CS/CS+GLU, NSAIDs/COXIBs). The costs are presented in euros (€) updated to 2024.

Probabilities and Costs of AE

The probabilities of the appearance of the different AEs are presented in Table 1. The unit costs of handling the AEs analyzed are shown in Table 2.

Gastrointestinal AE (GIAE)

The annual probabilities of suffering a mild-moderate GIAE (0.64% with CS/CS+GLU; 13.86% with NSAIDs/COXIBs) or severe (0% with CS/CS+GLU; 0.35% with NSAIDs/COXIBs) GIAE were obtained from the GI-REASONS study^18,33 (Table 1). These assumptions are more conservative than those adopted in the previously published VECTRA study.³⁴ The unit costs (updated to 2024) of a mild-moderate GIAE (€285) and a severe GIAE (€3,393) were obtained from the economic analysis published in 2020²⁹ (Table 2).

Ischemic Heart Disease (IHD)

The risk of suffering a myocardial infarction (MI) was obtained from the study by Mazzucchelli et al²⁰ In this nested case-control study, which included 23,585 incident MI cases and 117,405 controls (of which 89 cases and 757 controls received CS/CS+GLU), a cardioprotective effect was observed in patients treated with CS/CS+GLU, with an adjusted odds ratio (AOR) of MI of 0.57 (95% CI 0.46–0.72). The probabilities of IHD considered in the model are presented in Table 1.

The economic impact of IHD was calculated from the public prices of DRG 121, 122, 123 and 140 (coronary ischemia, 7,158 €), according to the frequency observed in the study by the Jordi Gol Primary Care Research Institute^19,50 and health Spanish public prices³⁹ (Table 2).

Acute Kidney Insufficiency (AKI) and Chronic Kidney Failure (CKF)

The frequency of AKI (0.31%) and CKF (0.21%) associated with NSAIDs/COXIBs was obtained from the study by Nelson et al,²² a retrospective study that included a large cohort of patients in the USA, treated with NSAIDs/COXIBs for at least 7 months of observation (Table 1). The economic impact of AKI and CKF was calculated from public health prices³⁹ (Table 2).

Ischemic Stroke (IS)

The risk of developing IS in patients treated with CS/CS+GLU or NSAIDs was obtained from two nested case-control studies.^21,51 The AOR of IS with CS/CS+GLU was 0.77 (95% CI 0.60–0.99)²¹ and with NSAIDs it was 1.20 (95% CI 0.90–1.60).⁵¹ Therefore, a possible protective effect of CS/CS+GLU on the risk of developing IS was observed (Table 1). The cost of IS (€9,285) was obtained from the public price of DRG 45³⁹ (Table 2).

The definitions of GIAE, IHD, AKI, CKF and IS were those described in the studies that provided the probabilities of such adverse events.^{18,20–22,33,51}

Drugs Costs

The average annual cost of treatment with CS/CS+GLU is indicated in Table 2.^4,5,35–37 The average cost of CS/CS+GLU and NSAIDs/COXIBs was calculated for different durations of treatment over the period of one year: 180 days (6 and 12 packs for CS and CS+GLU treatment, respectively) in the base case of the analysis and 90 or 240 days (3 or 8 packs and 6 or 16 packs for CS and CS+GLU treatment, respectively) for the sensitivity analysis. According to its summary of product characteristics, CS/CS+GLU treatment should be carried out for at least 3 months, although in patients with significant inflammatory symptoms, after a rest period of 2 months, treatment could be started again following the same cycle.⁴ Therefore, over the period of one year, CS/CS+GLU treatment could range from a minimum of 3 months to a maximum of 8 months. In the case of CS/CS+GLU with glucosamine, it is advisable to administer it for a period of at least 6 months.⁵ With respect to NSAIDs/COXIBs, the duration of the treatment is highly variable, depending on the different studies. In the GI-REASONS study,^18,33 a randomized clinical trial aimed at analyzing the GIAEs associated with celecoxib and the NSAIDs/COXIBs, treatment duration was 6 months. This study was used to obtain the average frequency of GIAEs with NSAIDs/COXIBs used in the economic model. Consequently, an average treatment duration of 180 days, between a minimum of 30 and a maximum of 240 treatment days per year, was considered in the base case.

The average cost per patient treated with NSAIDs/COXIBs in the base case (180 days of treatment) was estimated at €31.44.^35–37 It was calculated from the report on the use of NSAIDs in Spain during the period 2013–2016, published by the Spanish Agency of Medicines and Medical Devices³⁵ and from prices by homogeneous grouping published by the Ministry of Health.³⁶ The daily doses of the various NSAIDs/COXIBs were obtained from the VECTRA study.³⁴

Utilities

The utility value (the quality of life perceived by the patient, ranging from 1 -perfect health- to 0 -death-) of the patient with KOA (0.532) was obtained from the study by Martín et al⁴⁰ The utility loss associated with the different AEs was also obtained from the literature: −0.034 in mild GIAE and −0.449 in severe GIAE;⁴¹ −0.231 in IHD;⁴² −0.214 in AKI and −0.334 in CKF;⁴³ finally, −0.359 in IS^44,45 (Table 2).

Base Case and Sensitivity Analysis

In the two analyses carried out, a base case was considered for a treatment duration of 180 days. Sensitivity analyzes were performed for treatment durations of 90 and 240 days.

Results

National Results

First Analysis (Savings to the NHS from Current CS/CS+GLU Treatment)

In Spain, it is estimated that 538,342, 546,147 and 553,471 patients will be treated with NSAIDs/COXIBs with KOA and 112,426, 114,675 and 116,059 with CS/CS+GLU in 2025, 2026 and 2027, respectively (Table 3).

Due to better CS/CS+GLU tolerability, 45,087 mild-moderate GIAE, 3,217 severe GIAE, 211 IHD, 1,087 AKI, 746 CKF and 3,359 IS, would be avoided in 3 years (Table 4A). Therefore, it is estimated that for every 10,000 patients treated with CS/CS+GLU, 1,314 mild-moderate GIAE, 94 severe GIAE, 6 IHD, 32 AKI, 22 CKF and 98 IS, would be avoided. Discounting the cost of the drugs, the three-year net savings for the NHS would be 57.1 million euros, with a probability of savings with CS/CS+GLU of 80.7% (Table 4B). The savings per patient treated with CS/CS+GLU would amount to €38.02 (95% CI 14.06; €75.69), with a mean gain of 0.0023 (95% CI 0.0018–0.0027) QALY (Table 5). The probability that CS/CS+GLU treatment was dominant (lower costs and QALY gain) or cost-effective (cost per QALY gained less than €25,000) was 80.7% and 98.1%, respectively (Table 5). The result of the probabilistic analysis can also be seen in Figure 1.

Table 4 First Analysis (Population Level). Current Situation (Part of the Patients with KOA are Treated with CS/CS+GLU and Part are Treated with NSAIDs/COXIBs) Compared to a Hypothetical Situation in Which All Patients with KOA Currently Treated with CS/CS+GLU Would Be Treated with NSAIDs/COXIBS. A) Estimation of Adverse Effects Avoided and QALYs Gained with CS/CS+GLU (Treatment Duration: 180 Days). B) Economic Impact of KOA Treatment with CS/CS+GLU

Table 5 First Analysis (Patient Level). Cost-Effectiveness of Current Situation (Part of the Patients with KOA are Treated with CS/CS+GLU and Part are Treated with NSAIDs/COXIBs) Compared to a Hypothetical Situation in Which All Patients with KOA Currently Treated with CS/CS+GLU Would Be Treated with NSAIDs/COXIBS. Treatment Duration: 180 Days

Figure 1 Cost-effectiveness of CS Prescription in the Treatment of Knee Osteoarthritis Compared to NSAIDs and COXIBs. Cost-effectiveness plane. Treatment duration: 180 days. Abbreviations: COXIBs, cyclooxygenase 2 inhibitors; CS, chondroitin sulphate with or without glucosamine; NSAIDs, non-steroidal anti-inflammatory drugs; QALY, quality-adjusted life-years.

Considering a treatment duration of 90 days, the probability of cost-effectiveness and dominance with CS/CS+GLU was 96.8% and 78.0%, respectively. For a duration of 240 days, the values would be 98.0% and 83.0%, respectively. Considering the estimated minimum and maximum number of patients treated with CS/CS+GLU and NSAIDs/COXIBs, the result remained in the same line as the base case of the analysis.

Second Analysis (Maximum Savings That Could Be Achieved by the NHS if All Patients with KOA Currently Treated with NSAIDs/COXIB Were Switched to CS/CS+GLU)

In the purely hypothetical assumption that all KOA patients currently treated with NSAIDs/COXIBs were switched to CS/CS+GLU, a total of 306,460 mild-moderate GIAEs, 20,790 severe GIAEs, 1,392 IHD, 7,203 AKI, 4,967 CKF, and 22,838 IS episodes would be avoided. As a result of the avoided AEs, 30,366 QALYs would be gained in the KOA population, with a saving of more than 387 million euros (Table 6).

Table 6 Second Analysis (Population Level) (if All Patients with KOA Currently Treated with NSAIDS/COXIBS Were Treated with CS/CS+GLU Only). A) Estimation of Adverse Effects Avoided and QALYs Gained with CS/CS+GLU (Treatment Duration: 180 Days). B) Economic Impact of KOA Treatment with CS/CS+GLU

The savings per patient treated only with CS/CS+GLU would amount to €121.48 (95% CI 37.77; €280.69), with a mean gain of 0.0130 (95% CI 0.0140–0.0156) QALY (Table 7). The probability that CS/CS+GLU treatment was dominant (lower costs and QALY gain) or cost-effective (cost per QALY gained less than €25,000) was 99.9% and 100%, respectively (Table 7).

Table 7 Second Analysis (Patient Level) (if All Patients with KOA Currently Treated with NSAIDS/COXIBS Were Treated with CS/CS+GLU Only). Cost-Effectiveness of CS/CS+GLU in the KOA Treatment. Treatment Duration: 180 Days

Regional Results

The regional results, in line with the national results, are presented in Tables S1 and S2 for analyses 1 and 2, respectively.

Discussion

Drug-induced AEs represent a significant cause of morbidity in developed healthcare systems with a relevant economic impact.^52,53 Therefore, an important aspect for determining the efficiency of a new therapy in KOA is to evaluate the AEs cost related to the symptomatic pharmacological treatment. This aspect has been previously evaluated in other indications, such as oncology and migraine.^54,55

According to the present study, due to its improved tolerability profile, treatment with CS/CS+GLU would prevent thousands of AEs over the next three years. It is important to highlight the severity of some of these AEs, which can be life-threatening. This health impact would be accompanied by considerable savings for the NHS, making it cost-effective treatment.

To assess the reliability of these results, we must consider both the strengths and weaknesses of the study. The strength of the study is the consistency of the sources used to obtain the main variables for the analysis. Population data and AEs frequency data for the economic model were obtained, from epidemiological studies or databases^46,47,49 and from large-scale clinical trials^18,33 or observational studies.^20–22 The prevalence of KOA in Spain was obtained from the Spanish Society of Rheumatology, specifically from the EPISER study.⁴⁶ The population over 40 years of age in Spain was taken from the database of the National Institute of Statistics.⁴⁷ The percentage of patients with KOA who follow pharmacological treatment and the percentage of patients treated regularly or occasionally with NSAIDs/COXIBs were obtained from a study on the use of medicines in KOA in Spain,⁴⁸ which included 238,536 participants, followed between 2006 and 2010. Most of this data was used in a cost-consequences analysis published in 2020.²⁹

While it is important to remember that this is a theoretical model (a simplified simulation of reality), a probabilistic model was developed to explore the effect of uncertainty in the model variables and to calculate the 95% confidence interval for the average results of the analysis (prevented AEs, total savings from prevented AEs, cost savings per patient, QALYs gained per patient, cost per QALY gained, probability of cost-effectiveness, probability of CS/CS+GLU dominance). The design of this type of model allows for a better simulation of clinical reality.^30–32 The suitability of the economic model was evaluated using the CHEERS checklist (Table S3).

The results obtained in this study are consistent with those obtained in three economic analyses previously published with data from Spain.^29,34,38

Regarding weaknesses, the most notable is the calculation of the estimated number of patients treated with CS/CS+GLU or NSAID/COXIB, a calculation based on available epidemiological and population data and subjected to a sensitivity analysis. Furthermore, the study did not include patients receiving CS/CS+GLU and NSAIDs concomitantly, estimated at 12% of patients treated with CS/CS+GLU according to the VECTRA study³⁴ and 2.17% in the study by Wilson et al,⁴⁸ which should also be considered a limitation of the study. It should be noted that the average cost per patient for acquiring NSAID/COXIB was calculated based on data on the use of the different drugs in Spain.³⁵ The probability of adverse events was obtained from studies that analyzed NSAID/COXIB as a whole.^{18,19,21,22,29,33,34} Another aspect to note regarding the second analysis is that the “maximum savings” scenario assumes that all KOA patients taking NSAIDs/COXIBs will switch to CS/CS+GLU. Obviously, it is a purely theoretical scenario. This may not be possible because of contraindications, differences in how patients respond, and doctor and patient preferences.

In this economic analysis, prescribing CS/CS+GLU for the treatment of KOA would be more efficient than prescribing NSAIDs and COXIBs. CS/CS+GLU is the dominant treatment (since it is more effective at a lower cost) compared to not prescribing CS/CS+GLU. However, the gain of 0.0023 QALYs per patient may not be clinically relevant.^56,57 If this premise were acceptable, up to €41 per patient treated with CS/CS+GLU would be saved compared to not receiving CS/CS+GLU treatment. This demonstrates that, in addition to the clinical value derived from using CS/CS+GLU, this alternative will generate resource savings for the NHS of Spain.

A fundamental difference between the present study and the one published in 2020 is due to the cost-effectiveness analysis carried out. It is important to note that the probability that the current use of CS/CS+GLU in KOA is “dominant” over a hypothetical situation without CS/CS+GLU use, with a probability of 80.7%. This means that in the 1,000 Monte Carlo simulations carried out, in 807 cases the treatment with CS/CS+GLU obtained more QALYs than without the utilization of CS/CS+GLU, also generating savings for the Spanish NHS. Furthermore, treatment with CS/CS+GLU was “cost-effective” compared to the non-use of CS/CS+GLU, ie. the cost of gaining 1 QALY with CS/CS+GLU vs. without CS/CS+GLU was below 25,000 EUR (in Spain the cost-effectiveness threshold has been set between 25,000 and 60,000 EUR per QALY gained with the most effective drug).⁵⁸ The probability that CS/CS+GLU use is the cost-effective option would be 98.1% (it was cost-effective in 981 simulations out of 1,000 Monte Carlo iterations performed).

Conclusions

Due to its better tolerability profile, in the actual scenario CS/CS+GLU treatment is expected to prevent thousands of AEs over the next 3 years while generating considerable savings (57 million euros) for the NHS, being cost-effective treatment with a high probability. Considering the characteristics of osteoarthritis, other treatments should be considered (as the prescription of other pharmacological groups such as SYSADOAs to minimize the use of NSAIDs,⁵⁹ a hypothetical scenario where all the KOA patients actually treated with NSAIDs were treated with CS/CS+GLU is assessed: as a result of the avoided AEs, 30,366 QALYs would be gained in the Spanish population, with a potential saving for the NHS of more than 387 million euros.