Osteoarthritis (OA) is the most common arthritic condition world wide. To date, its treatment remains largely symptomatic.1 2 A number of studies have shown that drugs including diacerein, glucosamine and chondroitin sulphate may have disease-modifying effects in patients with knee or hip OA.3^–7 A recent study demonstrated that doxycycline, a drug that can inhibit matrix metalloproteinase synthesis and activity, could reduce the progression of knee OA as assessed by conventional x-ray examinations.8

Licofelone is a pyrrolizine derivative that acts as a competitive inhibitor of both 5-lipoxygenase and cyclo-oxygenases. It is a substrate analogue to arachidonic acid that competes at the active site of these enzymes and has been shown to be a very potent anti-inflammatory in preclinical and clinical studies in patients with OA.9^–11 In an experimental dog OA model,12 13 licofelone also demonstrated in vivo disease-modifying effects by reducing joint structural changes and the synthesis of several catabolic factors. These findings provide a strong rationale for the initiation of a clinical study investigating the disease-modifying effects of licofelone in patients with knee OA.

To assess disease progression over time, x-ray technology is still the most commonly used14 and recommended by regulatory agencies for phase II and III disease-modifying osteoarthritis drug (DMOAD) trials. However, in recent years, MRI technology has been extensively used to assess the structural changes of knee OA in a number of cross-sectional and longitudinal studies.14^–21 Quantitative MRI (qMRI) is a sensitive and reliable tool for the assessment of the structural changes that occur in knee OA over time and enables the identification of risk factors associated with cartilage volume loss, among which the most predominant are meniscal tear/extrusion and subchondral bone lesions.16 17 20 22^–24 Until now, there have been no double-blind, multicentre, clinical trials of drug treatment in which both x-ray examinations and qMRI were used to assess the progression of structural changes in patients with knee OA.

This study aimed to compare the potential DMOAD effects of licofelone and naproxen on the progression of knee OA cartilage loss over a 2-year period using both qMRI and x-ray examinations.

PATIENTS AND METHODS

This was a multicentre, randomised, double-blind and parallel group study comparing licofelone with naproxen in patients with knee OA. Subjects were treated for 24 months. Naproxen was chosen as a comparator treatment, as it is one of the most commonly prescribed non-steroidal anti-inflammatory drugs (NSAIDs) for knee OA.

Patient selection

Patients were recruited from outpatient rheumatology clinics throughout Canada. Inclusion criteria were age (40–80 years), primary knee OA of the medial femorotibial compartment diagnosed according to the clinical and radiological criteria of the American College of Rheumatology,25 intermittent or constant pain for at least 50% of the 2-month duration before baseline and treatment with NSAIDs indicated. A Western Ontario and McMaster Questionnaire (WOMAC) pain subscale index26 of at least 40 (scale: 0, least to 100, worst), after a 24-h washout of any analgesic drugs and a 7-day washout of any NSAIDs, was required, as well as a disease severity grade of 2 or 3 based on the Kellgren–Lawrence27 radiographic system and at least 2 mm of medial joint space width at the narrowest point in the medial compartment, as assessed locally by the investigator with a ruler. The patients were also required to have at least one of the following three risk factors for increased risk of radiographic progression: body mass index >30 kg/m², the presence of Heberden’s nodes or female gender.

Exclusion criteria were the presence of other rheumatic diseases that could be responsible for secondary OA, severe articular inflammation as confirmed by physical examination, traumatic knee lesions, metabolic diseases, intra-articular or systemic corticosteroids within the 3 months preceding enrolment, total knee replacement of the contralateral knee within 6 months before screening, isolated lateral compartment disease, class IV functional capacity based on American College of Rheumatology criteria,28 surgery in any lower limb joint within a year of the baseline visit, arthroscopy, aspiration or lavage in any lower limb joint within 6 months of the baseline visit and a history of a gastrointestinal (GI) disorder that could prevent the intake of NSAIDs for the duration of the study. The study was approved by the local ethics committees and all patients gave their oral and written informed consent to participate.

Treatments administered

Subjects were randomly assigned to receive either therapeutic dosages of licofelone (200 mg twice a day) or naproxen (500 mg twice a day) from baseline to month 24. The drug was taken orally with food at intervals of about 12 h. Randomised subjects started the study drug on the day of randomisation, beginning with the evening dose. The licofelone dose used in the study has been shown to be safe and effective in previous phase II and III knee OA clinical studies.10 11 The naproxen dose used corresponds to the labelling of this approved drug.

Blinding

The investigators, subjects and sponsors were not aware of the treatment assigned. Through central randomisation, sealed coded tamper-proof envelopes, specifying the treatment group for each study drug kit number, were provided to each centre. The randomisation code envelopes were to be opened only in an emergency.

Prior and concomitant treatment

Oral or parenteral anticoagulants (with the exception of aspirin (acetyl salicylic acid at a maximum daily dose of 325 mg), oral or topical NSAIDs (other than the study drug), immunosuppressive drugs, lithium carbonate, phenytoin, analgesic drugs including over-the-counter preparations (other than rescue drugs provided by the investigator), other antiarthritic drugs, including indometacin, or compounds containing non-approved agents for arthritis or agents claiming to possess disease/structure-modifying properties (eg, glucosamine or chondroitin sulphate, or both) were prohibited. Subjects prone to or experiencing GI problems during the course of the trial were allowed, at the investigator’s discretion, preventive treatment with a proton pump inhibitor.

Acetaminophen (Tylenol; McNeil Consumer Healthcare, Guelph, Ontario, Canada), up to 4 g/day, provided by the sponsor, was the sole analgesic drug permitted during the study and was stopped 24 h before each study visit. The consumption of acetaminophen, a secondary measure of efficacy, was documented. The drugs and other treatments in use for intercurrent illnesses at the time of the baseline visit were to remain constant for the duration of the study, as evaluated by the investigator.

Treatment compliance

The number of tablets of the study drug taken was calculated from the drug dispensation log. The compliance index was calculated as the percentage of the rated drug doses taken between the date of dispensation and the date of return of the study drugs.

Outcome measures

Knee MRI acquisition

High-resolution, three-dimensional (3D) MRI was obtained using 1.5 T with integrated knee coil. These examinations are optimised 3D FISP (fast imaging with steady-state precession) acquisitions with water excitation (Siemens, Erlangen, Germany) or 3D SPGR (spoiled gradient echo recalled) acquisitions with fat suppression (General Electric, Milwaukee, Wisconsin, USA), as previously described.18

Primary outcome

Cartilage volume changes over time

Cartilage knee joint volume was measured by two trained readers using a specially developed computer program (Cartiscope; ArthroVision, Montreal, Quebec, Canada).18 29 The readers were blinded to treatment and to MRI examination time point except for baseline. The change in knee cartilage volume was obtained by subtracting the follow-up volume from the initial (baseline) volume.18 The change in cartilage volume over time was calculated for the entire (global) and for each of the medial and lateral compartments of the knee. The reproducibility of the method has previously been demonstrated to be excellent.18

Associated predictors: meniscal lesions and subchondral bone hypersignal (oedema) at baseline

The meniscal lesion and subchondral bone hypersignal (oedema) evaluation was performed at baseline as previously described.22 In brief, the proportion of the menisci affected by the tear was scored using a semiquantitative scale: 0, no damage to three areas to 3, all three areas affected. The extent of meniscal extrusion on the medial or lateral edges of the femoral tibial joint space was evaluated for the anterior, middle and posterior horns of the menisci: 0, no extrusion to 2, complete extrusion (severe). For bone, the intensity and extent of the lesion were assessed in the medial and lateral tibiofemoral compartments using the following scale as previously described23: 0, absence of oedema to 2, severe, a large lesion. The results were summarised as the absence or presence of any oedema (grade 1 or 2) and absence or presence of one severe lesion (grade 2) regardless of the presence of additional smaller lesions. Reliability of the scoring systems for meniscal and bone changes was found to be excellent.23

Secondary outcomes

Knee x-ray changes over time (Lyon–Schuss method)

The minimum and mean joint space width (JSW) of the target knee was evaluated at baseline, 12 and 24 months in the medial tibiofemoral compartment using an automated computerised method.30 The mean joint space value is an arithmetic average of all the joint space measurements made throughout specific boundaries of the entire knee. This was calculated automatically with the software. The minimum joint space is represented by the smallest distance measured among all the aforementioned measurements.

Method of reading

All the films (radiographs and MRIs) were read within the same time frame upon completion of the study. For the radiographs, the readers were completely blinded to the study drug and the time sequence but all three films (baseline, 12 months and 24 months of follow-up) were evaluated as a triplet. For MRIs, the readers were also blinded to the study drug. All four MRIs (baseline, 6 months, 12 months and 24 months) were analysed simultaneously. The readers were aware of the baseline acquisition, as this was mandatory for the registration process.18 29 However, the readers had no knowledge of which MR acquisition was at 6, 12 or 24 months. A special double-coding process was put into place specifically to identify each film (standardised radiograph or MR acquisition) so that patient identification, treatment allocation and the film time sequence was only known by the CRO.

Clinical evaluation

Patients underwent clinical evaluation at baseline, 6, 12 and 24 months based on the WOMAC (pain, stiffness, function and total score),26 visual analogue scale for patient global assessment (0, very good; 100, very bad) and the pain they were having on the day of the visit (patient pain score: 0, no pain; 100, most severe pain). There was a 24-h washout of analgesic drugs before the clinical evaluation.

Secondary outcome measures were use of rescue drugs as recorded in a daily diary, withdrawal rates, occurrence of adverse events with special attention to GI and cardiovascular events and routine safety laboratory tests for all patients still receiving the study treatment.

Statistical analysis

Data were entered into a computerised database using a blinded double-entry procedure, after which descriptive statistics for patient characteristics were tabulated. The primary efficacy outcome measure for structure modification was the loss of knee cartilage volume in the medial compartment in the signal joint, by intention-to-treat (ITT) analysis for all randomised patients, after 24 months of enrolment.

The sample size estimation was based on the assumption that the naproxen group would experience a 7.6% reduction of cartilage volume loss of the medial compartment after 24 months with a standard deviation of 6.0. The licofelone-treated subjects were expected to have at least a 5.3% reduction of cartilage volume loss or less after 24 months with a standard deviation of 6.0, representing a 30% less cartilage volume loss in comparison with naproxen. The required sample size at the two-sided 0.05 significance level was calculated as 110 patients for each group for a power of 80%.

Results were reported as according to protocol (ATP) for those patients who had taken all the study medication provided and for whom all the outcome (clinical and structural) evaluations had been made. ITT analysis was carried out by imputing the missing data to the average change recorded (mean value imputed) among patients within their corresponding treatment group at a specific time point (6, 12 and 24 months), provided that the patients had had at least a baseline MRI. The knee cartilage volume changes were also assessed globally and for the lateral compartment. Values were as absolute changes in mm³ at all time points (6, 12 and 24 months) compared with baseline. The impact of licofelone in comparison with naproxen on the cartilage volume loss according to absence or presence of meniscal or bone lesion was further assessed by variable stratification.

As a secondary outcome measure for structure modification, joint space narrowing in the signal joint—that is, the change in the minimum and mean JSW (mm) after 12 and 24 months in the medial tibiofemoral joint compartment, was assessed by the Lyon–Schuss method.30

The 6-, 12- and 24-month changes in the WOMAC pain score were used for assessment of symptom modification, with the final changes in the physical function and stiffness subscales as well as the total WOMAC analysed as secondary end points.

All differences within a treatment group between specific time points in comparison with baseline were assessed using a paired-comparison Student t test. Between treatment groups variable changes at all time points were assessed using a two-sample Student t test. Adverse event and dropout rates were analysed by χ² or two-sided Fisher exact tests. All statistical analyses were done using SAS software, version 9.1. All tests were two-sided and a p value ⩽0.05 was considered significant. Analyses were not adjusted for multiple comparisons.

RESULTS

Patient disposition

Three hundred and fifty-five patients were enrolled in the study and randomly assigned to receive licofelone or naproxen (fig 1). Three hundred and one patients, 147 in the licofelone and 154 in the naproxen group, had baseline MRI. A similar number of patients in the two groups completed the two-year treatment course, without significant differences in reasons for withdrawal (eg, adverse event or lack of efficacy). The main reason for the absence of MR data for the remaining 54 patients was because of either a patient’s claustrophobia or inability to maintain a relatively still position in the MR magnet, despite several attempts, in many cases, to obtain interpretable images. We nonetheless retained these patients in the study as this trial also had radiological and symptom outcomes that provide extremely relevant data in this study. Moreover, the final number of 301 patients recruited who also had a baseline MRI was well above the number of patients needed to complete the study according to the power calculations. Of note, not all patients underwent all MRI examinations at all time points, resulting in a total of 70 completing all four examinations in the licofelone and 73 in the naproxen group.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Trial profile. ATP, according to protocol; ITT, intention to treat.

Baseline characteristics

Table 1 shows the baseline demographic and clinical characteristics. Patients at study entry had similar mean age, gender, mean body mass index and WOMAC pain score in both treatment groups. For knee structural assessment (table 2), the mean and minimum JSW as well as global, medial and lateral compartment cartilage volume were also similar in the licofelone and naproxen groups, indicating that the randomisation procedure was satisfactory. Moreover, similar baseline characteristics between the two treatments were seen despite a high dropout rate for the ATP cohorts (data not shown). Compliance with study treatment was good: the proportion of patients who reported over 75% drug intake was almost identical in the treatment groups; 78.5% for the licofelone and 77.5% for the naproxen group, respectively (p = NS, χ²).

Loss of cartilage volume between treatment groups over time

Table 3 shows that for the ITT group, starting at 12 months, there were significant differences in cartilage volume loss between the two treatments, favouring the licofelone over the naproxen group for the global knee, as well as the lateral and medial compartments. Significant differences were seen whether the results were expressed in absolute values or percentage.

For the ATP, as early as 6 months after the initiation of treatment, there were significant differences in the percentage of cartilage volume loss between the two treatments, favouring the licofelone over the naproxen group for the global knee and lateral compartment. When expressed as absolute values, although differences favouring the licofelone group were seen at all time points, statistical significance was reached for the global and lateral compartment only at 24 months. For the medial compartment, as for the global and lateral compartment, differences favouring the licofelone group were seen at all time points; however, none reached statistical significance.

Loss of cartilage volume according to the absence or presence of meniscal or bone lesions (ITT and ATP)

The patient cohorts were stratified according to the absence or presence of meniscal or bone lesions at baseline. The only positive finding was found when the patients were stratified according to absence or presence of severe medial meniscal extrusion. Data for the ITT group showed a statistically significant difference in cartilage volume loss in the medial compartment (table 4) between the two treatments favouring the licofelone group at 12 and 24 months, either in the absence or presence of meniscal extrusion. For the ATP, only patients with severe extrusion at 24 months and when expressed as percentage had a significant reduction in cartilage loss with licofelone treatment compared with naproxen, yet, a reduction of about 20% was recorded. However, in the lateral compartment, no such difference between the two treatment groups was seen using meniscal extrusion as a confounder. Other measures, including meniscal tear as well as bone lesions, were assessed as potential confounders but did not yield significant differences (data not shown).

Loss of JSW

Seventy-eight patients in the licofelone group and 77 in the naproxen group had a complete set of standardised radiographs at 24 months. Table 5 shows a smaller loss of mean or minimum JSW (mm) over 12 and 24 months for the licofelone group for both the ITT and ATP cohorts. None of these differences was, however, statistically significant.

Symptoms

There was an improvement for the ITT group at 24 months in the primary symptoms outcome represented by the WOMAC pain score compared with baseline for all patients receiving either licofelone or naproxen (both at p<0.001). Statistically significant initial improvement of the pain (ATP) and total WOMAC scores (ITT and ATP) favouring naproxen over licofelone was seen at 6 months (p⩽0.05) (table 6). However, the difference between the groups’ average changes was clinically similar and not statistical at all subsequent time points. There were similar and significant improvements in WOMAC total, pain, stiffness and physical function subscales at 6, 12 and 24 months for both treatment groups compared with baseline, but there was no significant difference between groups (table 6). Similar results were also found for the visual analogue scale patient global and pain assessment at all time points for both ITT and ATP (data not shown). Over this 2-year study, many patients took at least one dose of acetaminophen (a simple analgesic) in similar proportions for both treatment groups: 40% for the licofelone versus 42% for the naproxen group for the ITT cohorts and 48% versus 52% for the ATP cohorts, respectively.

Safety

Most patients reported at least one adverse event: 63.3% in association with licofelone and 75.2% with naproxen. Table 7 shows the incidence of the most common adverse events recorded, most of which were transient and of mild to moderate severity. There were no substantial differences between groups in the incidence or pattern of events, with the exception of the GI events where 115 were reported for naproxen and 95 for the licofelone group (p = 0.036). Higher incidences of dyspepsia and nausea were reported for the naproxen group and higher incidences of diarrhoea for the licofelone group. Renal events were also reported, mainly as new onset of hypertension or peripheral oedema and more frequently for the naproxen group (p = 0.044). No serious GI adverse events such as perforation, ulcer or bleeding were reported. One myocardial infarction was observed for the naproxen group. Routine laboratory tests did not show any significant abnormalities in system organs or metabolic functions in the two groups (data not shown).

DISCUSSION

This longitudinal study was designed primarily to examine the effect of licofelone compared with naproxen on cartilage volume loss and clinical symptoms in patients with moderate knee OA. The study provides new information on the protective effects of licofelone on the progression of knee OA structural changes, as well as its effectiveness in relieving the disease symptoms. Moreover, this study is the first to validate the use of qMRI in a multicentre study and to demonstrate the superiority of qMRI over x-ray examinations in assessing the effects of drug treatment on cartilage loss over time, particularly in the different anatomical regions of the knee. Data from qMRI have also broadened our understanding of the role of meniscal lesions, a risk factor for rapid disease progression, as an important confounding factor in response to DMOAD treatment.

The latter also raises the issue of the importance of considering patient inclusion based on risk factors for disease progression in the planning of a DMOAD trial, a perspective which has been greatly broadened with MRI technology. In addition, qMRI was shown to be sensitive enough to detect significant differences between treatment groups as early as 6 months after the initiation of the treatment. Altogether, these findings highlight the very significant advantages provided by this new technology in such DMOAD studies. Moreover, these results bring to light the importance of evaluating cartilage loss in the different regions of the knee, including the lateral compartment, in which significant cartilage loss and a positive effect of licofelone treatment were detected.

The baseline demographic and clinical characteristics of the patients in this study reflect those of patients with knee OA with moderate to moderately severe disease routinely seen in outpatient clinics. The two treatment groups at baseline were balanced in number of patients and other characteristics, including signs, symptoms and imaging, for both x-ray examinations and MRI. The patient dropout rate in this study was within the expected range (30–50%) based on previous similar studies.3^–5 8 A significant number of patients (∼15%) in both groups failed to complete the follow-up MRI examination, the main reason being patient movement during the examination. Moreover, the number of patients leaving the study for other reasons, such as adverse events or lack of efficacy, was similar in both treatment groups.

The cartilage volume loss for both the ITT and ATP groups was found to increase over time in the two treatment groups, with the greatest loss found in the medial compartment. These findings are in accordance with those from previous MRI studies in patients with knee OA and were expected, based on the patient group selected for this study.18 21 31 The loss of cartilage volume was significantly less in patients treated with licofelone compared with naproxen. Interestingly, in the ITT group, the sparing effect of licofelone on cartilage loss was found in both compartments, although was more pronounced in the lateral compartment. The predominant protective effect of licofelone treatment in the lateral compartment is most interesting; however, the exact reason for this is unknown. One hypothesis is that licofelone may be more effective in areas where the cartilage damage is generally less severe. The finding of a protective effect of the treatment detected by MRI as early as 6 months is most interesting and reflects the high sensitivity of this technology.

Data from the ITT group indicate that the extent of reduction in cartilage loss did not reach the primary end point set for the study (30%). The ATP group presented similar results to those obtained from the ITT group, with the exception of the loss of statistical significance in the medial compartment. This finding could be related, at least in part, to the smaller number of patients in the ATP group. Interestingly, the protective effect of licofelone was present in the global knee and lateral compartments as early as 6 months and persisted throughout the entire study period of 24 months. These data provide additional support for the likelihood that the differences seen between the groups are related to the effect of treatment, as the ATP population had continuous exposure to drug treatment for the entire study period. The difference found (about 20% less cartilage loss favouring licofelone) is likely to be clinically relevant as “any” prevention of cartilage loss is important, especially considering the relatively short time span of the study. Recent publications have shown that cartilage degradation in patients with knee OA is associated with worsening knee pain in these patients. Cartilage volume loss of 8–10% at 24 months was found to correlate with the worsening of the WOMAC pain variable.32 33 However, no data exist on what would be a clinically significant difference for cartilage loss prevention between the two interventions. Moreover, no data are available on the use of this qMRI surrogate for long-term outcomes such as the need for a total knee replacement. However, even if the long-term protection is not addressed here, one can easily suggest that the protection offered by licofelone, if sustained, might have a significant clinical impact.

The absence of significant difference between drug treatments in the medial compartment in the ATP group was at first quite intriguing. Post hoc analyses with confounding factors showed the role played by meniscal extrusion, a known risk factor for disease progression.22 Indeed, data showed that in this compartment, patients with severe meniscal extrusion clearly responded more favourably to licofelone treatment.

This latter finding raises a number of very important concerns about the DMOAD trial and the therapeutic effectiveness of licofelone. One of these concerns is the inclusion criteria used for the selection of patients in DMOAD trials. In such studies, particular effort is made to ensure that the treatment groups are well balanced in regard to important risk factors such as clinical information and x-ray findings. However, this clearly overlooks a number of other major risk factors which can only be detected by MR images. Inclusion of known risk factors for rapid disease progression could be very helpful in DMOAD trials as it would reduce the number of non-progressing or slowly progressing patients.

The findings of this study indicate that patients with severe meniscal extrusion, in both ITT and ATP analyses, presented a more severe loss of cartilage and were also more responsive to licofelone treatment. An explanation for the licofelone effect may be related to the fact that this drug has been shown in vivo to abrogate the synthesis of many OA cartilage catabolic factors and inflammatory cytokines.12 13 Indeed, the greater loss of cartilage found in patients with severe medial meniscal extrusion might be due, in addition to biomechanical factors, to the presence of a locally higher level of these catabolic factors. It is thus tempting to speculate that the differences in cartilage loss between the two treatment groups are related to a protective effect of licofelone. However, to confirm a unique DMOAD effect of licofelone, other drugs in the same family or other NSAIDs should be tested in a similar study design. Nevertheless, the findings of this clinical trial correspond extremely well with those of preclinical studies in experimental OA.12 13 Moreover, analysis of the ATP subgroup with severe meniscal extrusion indicates that the reduction in cartilage volume loss in the licofelone-treated patients came very close to reaching the main goal defined for the study with a reduction of about 20% at 24 months. One limitation of this study is the absence of a placebo arm. For such a lengthy study, it would have been unethical to treat patients with a placebo. Nonetheless, the mean cartilage loss found in the naproxen group was fairly similar to that seen in other longitudinal trials,18 31 suggesting the absence of a naproxen toxic effect toward cartilage loss.

The loss of JSW (mean, minimum) over 24 months correlates well with the findings of previous studies.4^–8 Although not significant, there was a trend for JSW reduction in the licofelone-treated patients. Moreover, no differences were found between the two treatment groups even when data were analysed based on the absence or presence of severe medial meniscal extrusion (data not shown). This again points to the lower sensitivity of x-ray examinations compared with MRI to detect cartilage loss over time and, more particularly, to discriminate the effect of drug treatment in the context of clinical trials.

The measurement of changes in JSW in the medial compartment as performed in this and other knee OA studies5^–8 24 has been shown to estimate changes in cartilage thickness in the central weightbearing areas of the condyles and plateaus.21 These JSW measurement changes are quite small with a variability that reflects the low sensitivity of x-ray technology for detecting differences in the loss of cartilage volume between two treatment groups over time. This would indicate that more patients would be needed in each group to obtain a significant difference. The absence of a correlation between qMRI cartilage volume change with change in JSW as seen in standardised radiographs has been previously described.18 23 33 The x-ray examinations actually address, in a two-dimensional manner, a small portion of the cartilage damage, mainly the weightbearing area.30 The greater sensitivity to change in cartilage volume assessed by qMRI in comparison with radiographs is once again demonstrated in this study. Moreover, this study was not powered to use radiographs as a knee structure outcome, which may explain the absence of treatment significance, even at 2 years of follow-up.

Treatment with licofelone was found to be as effective as naproxen for relieving the symptoms of knee OA. The extent of the effect was essentially equal for both treatments on all the major clinical evaluation scales used. The symptomatic effect was found to be persistent throughout the entire study period for both drugs. The only exception was the superiority of naproxen over licofelone at 6 months for WOMAC total and pain. This may indicate a slower mode of action of licofelone on disease symptoms, as often seen with DMOADs. Although this study had no true placebo control, the results are still informative, as not all drugs used in DMOAD studies have shown a consistent effect on disease symptoms.3 6 The similarity in mechanisms of action of licofelone and naproxen as inhibitors of the cyclo-oxygenase and prostaglandin synthesis is a possible explanation. Moreover, the ability of licofelone to reduce the severity of synovitis in vivo as shown in experimental dog OA12 may also have contributed to the symptomatic effect of the drug.

From a safety point of view, the absence of serious adverse events related to the drugs indicates a good level of safety for both treatments when administered chronically to patients with OA. Also of significance was the absence of major NSAID complications, such as perforation, ulcer or bleeding, during the study. This finding could be explained, at least in part, by the patients’ use of gastroprotective agents (proton pump inhibitors) when needed, based on the investigator’s judgment. Of particular interest was the significantly lower incidence of GI and renal side effects in patients treated with licofelone, in which the lower incidences of dyspepsia, nausea, hypertension and oedema appear the main reasons for these differences. The good GI safety profile of licofelone has been reported in previous clinical studies.10 11 It is hypothesised that this effect may be connected with the ability of licofelone to also inhibit the 5-lipoxygenase pathway.9 The lower incidence of hypertension in these patients is also interesting. However, the criteria used to establish the diagnostic of hypertension may not have been of today’s satisfactory standards. These preliminary findings need to be confirmed by better defined controlled studies.

In summary, this study provides new information, indicating that better reduction of the loss of cartilage in patients with knee OA is achieved with licofelone than with naproxen. The effect of licofelone is more pronounced in patients with rapidly progressive disease—that is, severe medial meniscal extrusion. Both drugs were clinically effective in reducing the disease symptoms and both presented good safety profiles when given for an extended period of time. The superiority of licofelone over naproxen in achieving a lower incidence of GI and renal side effects remains to be fully explored. To conclude, licofelone is a slow acting agent, which is a profile often seen in DMOAD drugs and which may explain why at 6 months the data on symptoms favour naproxen. Of note, however, is that the difference between the two treatment groups was barely significant. On the other hand, the unique 5-lipoxygenase and cyclo-oxygenase inhibition of this drug might also explain its better tolerability.9

In conclusion, qMRI in this DMOAD trial of knee OA proved to be instrumental in providing reliable and sensitive information about the effects of the drugs on cartilage loss, in addition to providing new insight into the roles played by risk factors such as meniscal lesions in the response of patients to treatment.