Efficacy and safety of flurbiprofen cataplasms versus loxoprofen sodium cataplasms in knee osteoarthritis: a randomized controlled trial

Abstract

Background:

Clinical trial evidence is limited to identify better topical non-steroidal anti-inflammatory drugs (NSAIDs) for treating knee osteoarthritis (OA). We aimed to compare the clinical efficacy and safety of flurbiprofen cataplasms (FPC) with loxoprofen sodium cataplasms (LSC) in treating patients with knee OA.

Methods:

This is an open-label, non-inferiority randomized controlled trial conducted at Peking University Shougang Hospital. Overall, 250 patients with knee OA admitted from October 2021 to April 2022 were randomly assigned to FPC and LSC treatment groups in a 1:1 ratio. Both medications were administered to patients for 28 days. The primary outcome was the change of pain measured by visual analog scale (VAS) score from baseline to day 28 (range, 0–10 points; higher score indicates worse pain; non-inferiority margin: 1 point; superiority margin: 0 point). There were four secondary outcomes, including the extent of pain relief, the change trends of VAS scores, joint function scores measured by the Western Ontario and McMaster University Osteoarthritis Index (WOMAC), and adverse events.

Results:

Among 250 randomized patients (One patient without complete baseline record in the flurbiprofen cataplasms was excluded; age, 62.8 ± 10.5 years; 61.4% [153/249] women), 234 (93.6%) finally completed the trial. In the intention-to-treat analysis, the decline of the VAS score for the 24-h most intense pain in the FPC group was non-inferior, and also superior to that in the LSC group (differences and 95% confidence interval, 0.414 (0.147–0.681); P <0.001 for non-inferiority; P = 0.001 for superiority). Similar results were observed of the VAS scores for the current pain and pain during exercise. WOMAC scores were also lower in the FPC group at week 4 (12.50 [8.00–22.50] vs. 16.00 [11.00–27.00], P = 0.010), mainly driven by the dimension of daily activity difficulty. In addition, the FPC group experienced a significantly lower incidence of adverse events (5.6% [7/124] vs. 33.6% [42/125], P <0.001), including irritation, rash and pain of the skin, and sticky hair uncovering pain.

Conclusions:

This study suggested that FPC is superior to LSC for treating patients with knee OA in pain relief, joint function improvement, and safety profile.

Introduction

Osteoarthritis (OA) is one of the leading causes of disability in the middle-aged and elderly population, with more than 300 million cases worldwide.^[1] For those over 60 years of age, 9.6% of the males and 18.0% of the females are diagnosed with OA.^[2] Knee OA, the most common type of OA, is mainly characterized by joint pain and limited mobility. It is becoming one of the major global public health challenges given its high prevalence and disability rate.^[3]

Non-steroidal anti-inflammatory drugs (NSAIDs) are recommended by national and international guidelines as an important strategy for the early treatment of knee OA on account of their combined benefits of pain relief and reduction of inflammation.^[4,5,6] As recommended by the 2020 UK National Institute for Health and Clinical Excellence (NICE), NSAIDs should be prescribed before further systemic analgesics in patients with knee OA.^[7,8] Nevertheless, the use of oral NSAIDs has been hampered by the risks of gastrointestinal and renal damage, as well as cardiovascular adverse events. Previous studies have confirmed that topical NSAIDs have a moderate analgesic efficacy similar to oral NSAIDs but with much better safety profile.^{[9,10,11,12,13]}

Flurbiprofen cataplasms (FPC) and loxoprofen sodium cataplasms (LSC) are two of the most frequently prescribed topical NSAIDs for the treatment of knee OA in China. FPC is made of a hydrophilic polymer material, which has strong transdermal properties, less skin irritation, and high absorption characteristics.^{[14,15,16,17]} LSC is an NSAID of propionic acid precursor without biological activity, which is absorbed through the skin and metabolized into trans-OH drugs in the body to inhibit prostaglandin biosynthesis.^[18,19,20] On the other hand, the bidding price of FPC is lower than LSC. This economic benefit offers a potential for FPC to be a preferred topical NSAID candidate over LSC. However, there are few studies on previous comparison of analgesic effect and safety between FPC and LSC. In this study, we aimed to investigate the differences in clinical efficacy and safety between FPC and LSC for treating patients with knee OA to serve as robust evidence for guiding the clinical practice.

Methods

Ethical approval

The trial was carried out according to the principles of Good Clinical Practice guidelines and relevant regulations. The study protocol was approved by the Ethics Committee of Peking University Shougang Hospital (No.IRBK-2021-031-01) in accordance with the Declaration of Helsinki. Informed consents were signed by all patients in this study. The trial was registered on the Chinese Clinical Trial Registry (ChiCTR2100054822).

Study design

This was an open-label, non-inferiority randomized controlled trial. From October 2021 to April 2022, a total of 250 knee OA patients with radiographic confirmation (Kellgren–Lawrence stage II–III) at the orthopedic outpatient clinic of Peking University Shougang Hospital were recruited. All enrollees were evaluated by attending physicians not meeting the indications of surgery including severe pain and/or dysfunction caused by bone ischemic necrosis, knee joint instability, severe stiffness or deformity, etc.^{[21,22,23,24]} The patients were recruited according to the OARSI's^[8] recommendations for clinical trials in patients with knee OA, including the techniques of radiographic imaging for OA of the knee, key analytic principle involved in the conduct of pharmacologic OA trials, etc. Subjects with unilateral or bilateral knee OA were both included. And for those patients with bilateral knee OA, side with most pain would receive the investigational NSAID interventions while being selected as the index joint.

After enrollment, patients were followed up at the 1-, 2-, and 4-week interval. We recorded the maximum and minimum daily pain levels by adopting the visual analog scale (VAS) score method; in addition, joint function assessed by Western Ontario and McMaster University Osteoarthritis Index (WOMAC) score, adverse events, the use of remedial drugs, and the reasons for patients to self-reduce or stop the medication were also recorded. Adverse drug reactions were reported within 1 month after the discontinuation of the drug administration.

Inclusion and exclusion criteria

Inclusion criteria: (1) Age >40 years; (2) Kellgren–Lawrence stage II–III: grade II is defined as a small osteophyte and possible joint space narrowing on a standing knee X radiograph; while grade III knee OA is characterized by a large number of moderate osteophytes, definite joint space narrowing, and possible knee deformities; (3) presence of joint pain symptoms; (4) not using oral or topical NSAIDs or opioid analgesics within 1 week before enrollment; (5) no previous history of bleeding, asthma, severe cardiac, hepatic, renal insufficiency, or NSAIDs allergy, etc.; and (6) agree to participate in the current study by signing informed consent.

Exclusion criteria: (1) pregnant and lactating women or women of childbearing age who are preparing for conception; (2) enrolled in other clinical trials within 1 month; (3) treated with intra-articular drug injections within 1 month; (4) had hormone therapy 1 month before entry; (5) previous knee or hip surgery within 3 months, causing pain or affecting joint function; (6) other concomitant diseases or complications that may interfere with the observation of efficacy, such as local skin damage, infection, and rash, etc.

Interventions

Patients were randomly assigned to two groups according to the random number table method. FPC topical patch (Beijing Tide Pharmaceutical Co., Ltd, Beijing, China) was administered to the patients in the FPC group with 1 patch/site/session, ingredient flurbiprofen 40 mg per patch, twice daily, for 4 weeks; and the patients in the LSC group were administered with LSC topical patch (Hunan Jiudian Pharmaceutical Co., Ltd, Hunan, China) with 1 patch/site/session, ingredient loxoprofen 100 mg per patch, once daily, for 4 weeks. Prescriptive medication instructions include: (1) avoid applying to surfaces of damaged or incomplete skin; (2) avoid applying to areas with eczema or rash; (3) keep the applied area dry and clean; and (4) two daily applications should be applied 12 h apart for FPC. Apply once per 24 h for LSC. FPC and LSC are both post-marketing drugs. Dosage and frequency in this study were determined according to their drug instructions, which has been proved by clinical use and evidence-based medicine to achieve standard treatment efficacy.

Outcomes of interest

The primary outcome was the differences of the changes in VAS scores (baseline's value–last follow-up's value) for the most intense, resting, exercise and current pain levels in the past 24 h for knee OA patients between FPC group and LSC group based on non-inferiority test. If non-inferiority is demonstrated, a superiority test will be further performed. Note that the VAS score ranges from 0 to 10 points and a higher score indicates worse pain.

There were four secondary outcomes, including (1) extent of effectiveness in pain relief: very effective (VAS score drops ≥3 points from baseline to week 4), effective (VAS score drops 1–3 points from baseline to week 4), and not effective (VAS score drops <1 point from baseline to week 4); (2) the changing trend of VAS scores during the whole follow-up period; (3) WOMAC scores after 4 weeks, and changes of WOMAC scores from baseline to week 4; and (4) adverse events during the drug administration and within 1 month after the discontinuation of the drugs, including irritation, rash and pain of the skin, and sticky hair uncovering pain. The severity of each adverse event was measured using a scale ranging from 0 to 4, with higher points indicating more severe conditions.

Sample size estimation

Based on a previous trial,^[9] FPC is expected to introduce a 3.5 ± 2.0 decline in VAS score. Assuming the actual difference of the decline in VAS score between FPC and LSC is 0, and the standard deviation (SD) of the VAS score decline in LSC group is equal to that in FPC group, sample size calculation (by PASS 15 software, NCSS, LLC, Kaysville, Utah, USA) yielded 108 patients in each group (216 in total) to obtain a power of 90% with a one-sided type I error rate of 2.5%, when the non-inferiority margin is set at 1 and the drop-out rate is considered 20%.

Statistical analysis

The primary effectiveness analysis was based on the modified intention to treat (mITT) population comprised of patients who underwent randomization and had at least one documented VAS score. The secondary effectiveness analysis was based on the per-protocol (PP) population who completed all follow-up visits. Safety analysis was based on patients who had received at least one study treatment. Missing values of VAS and WOMAC scores in the intention to treat (ITT) analysis were imputed using the last observation carried forward method.

Frequency and proportion for categorical variables and mean with SD or median (interquartile range) for continuous variables were used to describe the data. Differences between means were tested using the Student t-test or Wilcoxon rank-sum test. Differences between proportions were tested using Pearson's chi-square test and Fisher's exact test if necessary. Specifically for the primary outcome, differences of the changes in VAS scores between the two study groups and their two-sided 95% confidence interval (CI) were estimated using the analysis of covariance (ANCOVA), corrected for baseline VAS scores. One-sided P-values were calculated for the non-inferiority test (margin: 1 point) and the superiority test (only if non-inferiority is demonstrated; margin: 0 point), with a P-value of less than 0.025 indicating statistical significance. The effectiveness would be declared superior if the lower limit of confidence interval of the FPC is above -1 and 0. The change trends of VAS scores across different visits were assessed primarily using one-way repeated-measures analysis of variance. In sensitivity analyses, the change trends of the difference of VAS scores from baseline (value of each visit–baseline value) at each visit were analyzed using repeated measures mixed effects model, corrected for baseline VAS score.

All statistical analyses were performed using SAS (version 9.4, SAS Institute Inc., Cary, NC, USA). Except for the primary outcome, all statistical tests were two-sided and P <0.05 indicated a statistical significance.

Results

Included patients and characteristics

Between October 2021 and April 2022, a total of 320 patients were screened for eligibility. Eventually, 250 patients were included in this study. Of these, 234 patients (93.60%) completed all the follow-ups (10 patients in the FPC group and 6 in the LSC group withdrew consent). A total of 249, 234, and 250 patients were included in the mITT, PP, and safety analysis, respectively [Figure 1]. Reasons for withdrawal and early discontinuation of the study drugs are detailed in Supplementary Table 1, http://links.lww.com/CM9/B663.

Figure 1.

Screening, randomization, and patient flow in a trial of flurbiprofen cataplasms vs. loxoprofen sodium cataplasms in knee OA. *One patient without complete baseline record in the flurbiprofen cataplasms was excluded. FPC: Flurbiprofen cataplasms; LSC: Loxoprofen sodium cataplasms; mITT: Modified intention to treat; OA: Osteoarthritis; PP: Per-protocol.

The baseline characteristics of age, sex, types of comorbidities, comorbid medications taken, VAS scores of 24-h most intense pain, and the WOMAC scores were homogeneous between the two groups [Table 1]. While the VAS scores for 24-h resting, exercise, and current pain in the FPC group were slightly higher than in the LSC group (baseline characteristics in the PP analysis set were given in Supplementary Table 2, http://links.lww.com/CM9/B663).

Table 1.

Baseline characteristics of the patients with knee OA in mITT study group.

Features	Total (N = 249)	Flurbiprofen (N = 124*)	Loxoprofen (N = 125)	Statistics	P-value
Age (years)	62.8 ± 10.5	63.4 ± 10.2	62.3 ± 10.9	0.870	0.385
Gender	–	–	–	1.567	0.211
Male	96 (38.6)	43 (34.7)	53 (42.4)	–	–
Female	153 (61.4)	81 (65.3)	72 (57.6)	–	–
Presence of comorbidities	108 (43.4)	54 (43.5)	54 (43.2)	0.003	0.956
Cardiovascular disease	75 (30.1)	41 (33.1)	34 (27.2)	1.017	0.313
History of drug allergy	19 (7.6)	12 (9.7)	7 (5.6)	1.468	0.226
Asthma	1 (0.4)	0	1 (0.4)	-	1.000
Digestive tract diseases	23 (9.2)	13 (10.5)	10 (8.0)	0.458	0.499
Kidney disease	4 (1.6)	2 (1.6)	2 (1.6)	0.0001	0.994
Endocrine system diseases	25 (10.0)	11 (8.9)	14 (11.2)	0.374	0.541
Combined drug use	70 (28.1)	35 (28.2)	35 (28.0)	0.002	0.968
Baseline VAS score
Passed 24 h most severe pain level	5.0 (4.0–6.0)	5.0 (4.0–6.0)	4.0 (4.0–5.0)	1.917	0.055
Passed 24 h resting pain level	3.0 (1.0–3.0)	3.0 (2.0–4.0)	2.0 (0–3.0)	4.239	<0.001
Passed 24 h exercise pain level	4.0 (3.0–5.0)	4.0 (4.0–5.0)	4.0 (3.0–5.0)	2.549	0.011
Current pain level	4.0 (3.0–4.0)	4.0 (3.0–4.0)	3.0 (3.0–4.0)	2.098	0.036
Baseline WOMAC score
Total Score	43.0 (27.0–71.0)	41.3 (26.5–73.0)	45.0 (27.0–67.0)	0.502	0.616
Pain	10.0 (6.0–15.0)	9.0 (6.0–16.5)	11.0 (6.0–15.0)	-0.186	0.853
Stiffness	4.0 (1.0–6.0)	4.0 (1.0–6.0)	4.0 (2.0–6.0)	0.047	0.962
Activity	31.0 (18.0–50.0)	28.8 (19.0–53.0)	32.0 (17.0–48.0)	0.794	0.427

Data are presented as mean±standard deviation or median (interquartile range). ^*One patient without complete baseline record in the flurbiprofen cataplasms was excluded. –: Not applicable. mITT: Modified intention to treat; OA: Osteoarthritis; SD: Standard deviation; VAS: Visual analog scale; WOMAC: The Western Ontario and McMaster University Osteoarthritis Index.

Primary outcome assessment

In the mITT analysis, the differences and 95% CIs of the decline in VAS scores for the 24-h most intense, resting, exercise, and current pain between the treatment groups (FPC-LSC) were 0.414 (0.147–0.681), 0.085 (-0.207 to 0.377), 0.366 (0.094–0.640), and 0.455 (0.201–0.709), respectively. The decline of most intense, exercise, and current pain of the FPC group were both non-inferior and superior to the LSC group in that all the corresponding low bound of 95% CI of the decline in VAS scores were higher than -1 and 0, respectively. The resting pain of the FPC group was also non-inferior to, but not superior to that of the LSC group (low bound of 95% CI of decline in VAS score was -0.207). The results from the PP set were consistent with the mITT set [Table 2].

Table 2.

Between-group differences of the decline of VAS scores from baseline to week 4 of the patients with knee OA.

Decline of VAS score	Least squares mean difference (flurbiprofen group-loxoprofen group)	95% CI (two-sides)	Non-inferiority test P-value (one-sided) *	P-value of superiority test (one-sided) †
mITT analysis
24 h most severe pain level	0.414	0.147–0.681	<0.001	0.001
24 h resting pain level	0.085	-0.207 to 0.377	<0.001	0.284
24 h exercise pain level	0.366	0.094–0.640	<0.001	0.004
Current pain level	0.455	0.201–0.709	<0.001	<0.001
PP analysis
24 h most severe pain level	0.454	0.229–0.680	<0.001	<0.001
24 h resting pain level	0.068	-0.204 to 0.340	<0.001	0.311
24 h exercise pain level	0.422	0.189–0.655	<0.001	<0.001
Current pain level	0.493	0.271–0.716	<0.001	<0.001

^*Non-inferiority margin: 10. ^†Superiority margin: 0. ANCOVA: Analysis of covariance; CI: Confidence interval; mITT: Modified intention to treat; PP: Per-protocol; VAS: Visual analog scale. ANCOVA was used to estimate between-group differences in the decline in VAS scores, correcting for baseline VAS scores.

Secondary outcome assessment

FPC group had a significantly higher proportion of very effective outcomes in pain relief, including 24-h most intense, resting, exercise, and current pain, than the LSC group both in the mITT and PP analyses (P <0.05) [Table 3].

Table 3.

Cross-tabulation and chi-squared test results of treatment efficiency in pain relief among patients with knee osteoarthritis.

Treatment effect	Flurbiprofen	Loxoprofen	Chi-Squared test	P-value
mITT analysis*	N = 124	N = 125
24 h most severe pain level			14.888	<0.001
Very effective	44 (35.48)	27 (21.60)
Effective	70 (56.45)	67 (53.60)
Not effective	10 (8.06)	31 (24.80)
24 h resting pain level			11.691	0.003
Very effective	45 (36.29)	33 (26.40)
Effective	49 (39.52)	36 (28.80)
Not effective	30 (24.19)	56 (44.80)
24 h exercise pain level			7.733	0.021
Very effective	47 (37.90)	31 (24.80)
Effective	64 (51.61)	68 (54.40)
Not effective	13 (10.48)	26 (20.80)
Current pain level			14.877	<0.001
Very effective	43 (34.68)	19 (15.20)
Effective	71 (57.26)	84 (67.20)
Not effective	10 (8.06)	22 (17.60)
PP analysis†	N = 115	N = 119
Change in 24 h most severe pain level score			22.280	<0.001
Very effective	44 (38.26)	27 (22.69)
Effective	68 (59.13)	66 (55.46)
Not effective	3 (2.61)	26 (21.85)
Change in pain level score at 24 h resting time			14.384	<0.001
Very effective	44 (38.26)	33 (27.73)
Effective	49 (42.61)	36 (30.25)
Not effective	22 (19.13)	50 (42.02)
Change in pain level score during 24 h exercise			11.313	0.004
Very effective	46 (40.00)	31 (26.05)
Effective	63 (54.78)	67 (56.30)
Not effective	6 (5.22)	21 (17.65)
Change in current pain level score			22.976	<0.001
Very effective	43 (37.39)	19 (15.97)
Effective	70 (60.87)	82 (68.91)
Not effective	2 (1.74)	18 (15.13)

Data are presented as n (%). ^*One patient without complete baseline record in the flurbiprofen cataplasms was excluded. ^†10 and 6 patients were lost to follow-up in flurbiprofen and loxoprofen groups, respectively. mITT: Modified intention to treat; PP: Per-protocol.

In the mITT analysis, the change trends of VAS scores across the follow-up visits between groups were shown in Figure 2. Repeated-measures ANOVA showed that all four VAS scores in the FPC and LSC groups had a descending pattern over time (P <0.001) and the FPC group declined faster (P <0.001). Similar trends were observed in the PP analysis [Supplementary Table 3, http://links.lww.com/CM9/B663]. After adjustment of baseline VAS score, repeated measures mixed effects model resulted in a significantly faster decline in the FPC group of the 24-h most intense, during exercise and current pain level (P-values = 0.027, 0.020, 0.016, respectively), but not of 24-h resting pain level (P = 0.207). Results from the PP analysis were consistent [Supplementary Table 4 and Supplementary Figure 1, http://links.lww.com/CM9/B663].

Figure 2.

Trends of VAS scores of 24 hour catastrophic resting pain (A), 24 hour resting pain (B), 24 hour movement pain (C) and current pain (D) in the ITT analysis. Data points are mean VAS scores for the corresponding visit, and the upper and lower horizontal lines indicate 95% CIs for the means. BL: Baseline; CIs: Confidence intervals; FPC: Flurbiprofen cataplasms; LSC: Loxoprofensodium cataplasms; VAS: Visual analog scale.

In the mITT analysis, the total WOMAC score at week 4 and change of the total WOMAC score from baseline to week 4 were significantly lower in the FPC group compared with those in the LSC group (P = 0.010 and 0.038, respectively). The differences were mainly driven by the dimension of daily activity difficulty. PP analysis showed similar results [Table 4].

Table 4.

Comparison of WOMAC scores after 4 weeks and changes of WOMAC scores from baseline to week 4 among patients with knee arthritis after treatment.

WOMAC rating	Flurbiprofen	Loxoprofen	Z-value	P-value *
mITT analysis†	N = 124	N = 125
Total score
Post-intervention (V4 visit)	12.5 (8.0–22.5)	16.0 (11.0–27.0)	-2.568	0.010
Value decline (baseline–V4)	23.0 (11.0–51.5)	17.0 (4.0–39.0)	2.072	0.038
Pain
Post-intervention (V4 visit)	3.0 (2.0–6.0)	4.0 (3.0–7.0)	-2.098	0.036
Value decline (baseline–V4)	4.5 (2.0–9.5)	4.0 (1.0–10.0)	0.914	0.361
Stiffness
Post-intervention (V4 visit)	0 (0–2.0)	1.0 (0–2.0)	-1.295	0.195
Value decline (baseline–V4)	2.0 (0–4.0)	2.0 (0–4.0)	0.477	0.634
Daily activity difficulty
Post-intervention (V4 visit)	8.0 (5.0–17.0)	12.0 (8.0–18.0)	-2.873	0.004
Value decline (baseline–V4)	17.5 (7.0–38.5)	12.0 (3.0–28.0)	2.404	0.016
PP analysis‡	N = 115	N = 119
Total score
Post-intervention (V4 visit)	12.0 (7.0–19.0)	15.0 (11.0–25.0)	-3.140	0.002
Value decline (baseline–V4)	25.0 (13.0–53.0)	19.0 (6.0–41.0)	2.287	0.022
Pain
Post-intervention (V4 visit)	3.0 (2.0–5.0)	4.0 (2.0–6.0)	-2.566	0.010
Value decline (baseline–V4)	5.0 (3.0–10.0)	5.0 (1.0–10.0)	1.078	0.281
Stiffness
Post-intervention (V4 visit)	0 (0–2.0)	1.0 (0–2.0)	-1.641	0.101
Value decline (baseline–V4)	2.0 (0–4.0)	2.0 (0–4.0)	0.700	0.484
Daily activity difficulty
Post-intervention (V4 visit)	8.0 (5.0–14.0)	11.0 (7.0–17.0)	-3.474	<0.001
Value decline (baseline–V4)	19.0 (9.0–40.0)	13.0 (4.0–29.0)	2.619	0.009

Data are presented as median (interquartile range). WOMAC: The Western Ontario and McMaster University Osteoarthritis Index; mITT: Modified intention to treat; PP: Per-protocol. ^*Using the Wilcoxon rank sum test. ^†One patient without complete baseline record in the flurbiprofen cataplasms was excluded. ^‡There are 10 and 6 patients were lost to follow-up in flurbiprofen and loxoprofen groups, respectively.

In the safety analysis population, there was a significantly lower incidence of reported overall adverse events in the FPC group than in the LSC group (5.6% vs. 33.6%, P <0.001). The incidences of each specific adverse event (irritation, rash and pain of the skin, and sticky hair uncovering pain) and the overall adverse event severity score were also lower in the FPC group (P <0.05, Table 5).

Table 5.

Safety analysis results of two groups of patients with knee arthritis.

Adverse events	Total (N = 250)	Flurbiprofen (N = 125)	Loxoprofen (N = 125)	P-value
Overall adverse events	49 (19.6)	7 (5.6)	42 (33.6)	<0.001
Skin irritation	16 (6.4)	4 (3.2)	12 (9.6)	0.039
Skin rash	15 (6.0)	2 (1.6)	13 (10.4)	0.003
Skin pain	11 (4.4)	2 (1.6)	9 (7.2)	0.031
Sticky hair uncovering pain	36 (14.4)	1 (0.8)	35 (28.0)	<0.001
Adverse event severity score*	0 (0–1.0)	0 (0)	0 (0–1.0)	<0.001

Data are presented as n (%) and median (interquartile range). ^*The severity of each adverse event was measured using a scale ranging from 0 to 4, with higher points indicating more severe conditions.

Discussion

Knee OA is one of the most common causes of chronic pain among middle-aged and elderly patients.^[1,2,3] In patients with early and mid-stage knee OA, acesodyne is the key method for effective intervention^[10,11]. Effective pain relief increases patient compliance with conventional treatment schemes and patient motivation for adopting non-medical interventions such as weight loss and exercise.^{[12,13,14,15]} Among various types of acesodynes, previous studies have revealed that topical NSAIDs are much safer than oral NSAIDs for pain alleviation, especially during the early treatment.^[13] However, high-quality studies such as RCT are still rare. Thus, it is difficult to identify effective topical NSAIDs. Overall, it is urgent to confirm the efficacy and safety of different topical NSAIDs for the treatment of knee OA. FPC and LSC are two of the most commonly prescriptive topical NSAIDs for treating knee OA in China. The current study findings suggest that the use of FPC as a treatment option for knee OA pain may provide more benefit to patients in terms of pain relief effect, improvement in joint function, and safety compared with LSC.

Our principle finding indicated that FPC was more effective than LSC in reducing pain levels, which was in line with the previous report.^[20,23] In addition to overall efficacy, rapid relief of painful symptoms is significant for knee OA patients.^[17] The results of the decline in VAS scores analyses support that patients using FPC can alleviate pain in a shorter time, while the decrease in the indicators of most intense, resting, exercise, and current pain level in the past 24-h are superior to LSC throughout the treatment, indicating that knee OA pain symptoms can be rapidly improved by using FPC. The rapid improvement in painful symptoms increases patient motivation to participate in the study, and this suggests that treatment with FPC may help knee OA patients to improve treatment adherence. The efficiency of reducing VAS scores of pain levels in both groups in our study exceeds 55% for the most intense, during exercise, and current pain levels in the last 24-h, which is higher than those findings in previous studies.^{[20,21,22,23]} However, the efficacy was not significant for pain levels during resting in the last 24-h, which is attributed to the fact that the resting pain mechanism of knee OA may be related to central sensitization and could not be improved by NSAIDs.^[21] During the implementation of this study, only one patient in the LSC group withdrew from our study due to unimproved symptoms of pain while the withdrawal rate of patients due to poor efficacy was low.

Our secondary outcome results show a significantly larger decline in the WOMAC scores in patients treated with FPC compared with LSC. The significant difference is only validated in the difficult activities of daily living (ADL). For the WOMAC score of knee joint pain, the difference between groups is not statistically significant (P >0.05), probably due to the inclusion of pain assessment in more mobile scenarios. The use of FPC for knee OA may have a reduced risk of developing adverse events during treatment compared with LSC. A previous study investigating the safety of topical NSAIDs found a low incidence of gastrointestinal, nephrotoxic, and cardiovascular events. These are consistent with the results of the present study.^[20] The adverse events that occurred in our study were allergic skin reactions and discomfort of application. This may be explained by the fact that the gel preparation process of FPC is superior to that of LSC.

In this study, the efficacy of FPC is superior to LSC with a faster effect. This advantage may be attributable to the potent COX inhibitor nature of FPC, and the improvement of the knee OA symptoms was dose-dependent. Moreover, the composition of FPC has better penetration of synovial tissue.^[18,19] LSC acts as a precursor drug and is metabolized by carbonyl reductase in local tissues, such as skin and liver, to form the anti-OH structure of the active metabolite (SRS ligand) before exerting its pharmacological effects. Therefore, our results showed that LSC provided slower pain relief than FPC over a 4-week treatment period, which may be due to a different pharmacological mechanism of reaction. However, LSC has been reported to act continuously for 24 h at the localized site of involvement,^[23] and the use of a more multidimensional approach to efficacy assessment may show its superiority in terms of long-lasting effects. In this study, the risks of having adverse events such as skin reaction or discomfort of application were significantly higher in the LSC group. This may be due to that the gel of FPC uses hydroxypropyl methylcellulose as a carrier, which has better skin adaptability and drug permeability over LSC.

Serval limitations remain in the current study: (1) this study is a single-center study, which to some extent weakens the generalizability of the findings; (2) the short observation period of the study limits the assessment in terms of long-term efficacy and long-term cardiovascular events, gastrointestinal adverse effects, or renal risk. However, previous studies have shown no increased risk of cardiovascular events or gastrointestinal events with long-term topical NSAIDs^{[10,11,20,24]}; (3) blinded process was not adopted in the current study and the sample may be subjective to the diagnostic suspicion bias; and (4) placebo arm was not included and therefore it was not possible to avoid a placebo effect.

In conclusion, topical NSAIDs are effective in relieving pain symptoms in patients with knee OA. The use of FPC has a higher therapeutic benefit for patients compared with LSC. More clinical trials with large sample size are needed to further evaluate the effect of long-term use of topical NSAIDs.

Acknowledgments

The authors acknowledge the collaboration and commitment of all investigators and their staff. The authors also thank all the patients who participated in this trial, and all participating institutions, particularly the research nurses and assistants.

Conflicts of interest

None.

Supplementary Material

cm9-136-2187-s001.pdf ^{(309.5KB, pdf)}