Table 1 Knee OA antioxidant supplements based on turmeric, avocado and Boswellia

Kuptniratsaikul et al. 2014 [63, 64]

RDB^a: compared curcumin (n = 171) with ibuprofen (n = 160)

Thai modified WOMAC, 6 min walk and patient satisfaction

All improved in both groups (p <.001) with no intergroup differences

Trial only 4 weeks

Safety profile better for curcumin

Belcaro et al. 2010 [71]

Open: best available treatment + Meriva versus best available treatment (n = 100)

Treadmill walking test, WOMAC and Karnofsky, and oxidative stress levels, inflammatory markers.

8 months - all measures improved with Meriva (p <0.05)

Not a blind study

Belcaro et al. in 2014 [66]

Open: Meriva and glucosamine (n = 63) vs chondroitin and glucosamine (n = 61)

Treadmill walking test, WOMAC and Karnofsky scales

4 month - similar improvements in both groups. Use of NSAIDS decreased in both groups

Not a blind study

Panahi et al. in 2014 [66, 68]

RDBP: curcumin + bioperine (n = 21) vs placebo (n =19)

WOMAC, VAS, Lequesne’s pain and stiffness score

6 weeks - improvement in WOMAC, VAS, Lequesne’s pain scores (p <.0010) but not in stiffness score

Mild gastrointestinal symptoms reported in both groups

Pinsornsak 2012 [65]

DB: (Curcumin 1000 mg + diclofenac 75 mg)/day (n = 44) vs (diclofenac 75 mg + placebo) (n = 44)

VAS AND Knee Injury and Osteoarthritis Outcome Score

No difference between groups for pain and function.

Small group size, submaximal doses The effects of two treatments is not additive

Henrotin, Y et al. 2014 [67]

Open: Flexofytol (curcumin with polysorbate) (n = 100) in real life situations

Serum Coll-2-1, Coll-2-INO₂, Fib3-1, Fib3-2, CRP, MPO, CTX-II. VAS pain

6 months - Coll-2-1 decreased. No change in pain

Not a blind study, no control

Appelboom et al. 2014 [72]

Open: Flexofytol- physicians in real life situation (n = 820)

Pain severity, flexibility and quality of life

6 months - improved in all (p <.0001), use of other treatments decreased (p <.0001)

Not a blind study

Significant improvements started in 6 weeks

Kertia 2012 [62]

RDB: C. domestica 3 × 30 mg (n = 34) vs diclofenac 3 × 25 mg (n = 39)

COX 2 levels in sinovial fluid at time = 0 and at 4 weeks

All improved no difference between groups

Short trial, no report of change in clinical symptoms

Madhu 2013 [70]

RSBP: NR-INF-02 1 g vs glucosamine 1.5 g vs NR-INF-02 + glucosamine vs placebo

VAS and WOMAC

6 weeks - all treatments showed significant improvement over baseline and placebo

Small study (<30/ group) and over short time period. NR-INF-02 is curcuminoid free extract of C. longa

Blotman 1997 [82]

RDBP: ASU (n = 83) vs placebo (n = 80)

Lequesne’s, Initially all groups received NSAID

6 months - Lequesne’s improved and NSAID use decreased

Over time pain similar in both groups

Maheu et al. 1998 [81]

RDBP: piascledine 300 mg (n = 85) vs placebo (n = 79) (knee and hip OA)

Lequesne’s, VAS for pain, and NSAID usage

6 months + 2 month follow up. Decrease in Lequesne’s (p <0.001), pain (p <0.003) and NSAID usage.

Improvement more marked in hip OA.

Lequesne et al. 2002 [95]

RBDP: piascledine 300 mg (n = 55) vs placebo (n = 53)

JSW, VAS pain, global assessment

2 years - no statistical difference in JSW or clinical parameters.

Posthoc analysis - when OA was severe, ASU slowed the disease progression

Pavelka et al. 2010 [96]

RDB: piascledine 300 mg vs chondroitin 1200 mg (n = 263)

WOMAC, Lequesne’s, VAS, global assessment, use of rescue medication

6 months + 2 month follow up. All parameters improved during treatment. Stabilized or improved in the follow up. No differences between groups.

Statistical significance not achieved due to large variability.

Maheu et al. 2014 [97]

RDBP: piascledine 300 mg (n = 345)

JSW, WOMAC, Lequesne’s

3 years - fewer progressors in the ASU group. No differences in clinical measurements.

Kimmatkar et al. 2003 [84]

RDB crossover: Boswellia serrata extract vs placebo (n = 30)

Knee pain, inflection, walking distance, frequency of swelling, radiology

8 weeks - 3 weeks washout 8 weeks crossover. Significant improvement in all parameters except radiology

Small group size

Sengupta 2008 [85]

RDBP: 5-Loxin (B. serrata extract enriched with 30 % AKBA) (n = 75)

Pain and function VAS, WOMAC and Lequesne’s

90 days - improvement in stiffness, function and pain scores, decreased MMP-3 (p <0.0001).

Serum biochemistry also improved, bioavailability of AKBA is low

Sengupta et al. 2010 [86]

RDBP: Aflapin 100 mg vs 5-Loxin 100 mg vs placebo (n = 20/group)

WOMAC, Lequesne’s, VAS, and serum biochemical, hematological and urine changes

90 days - improvement in pain and physical function in both treatment groups.

Small study over short time. Aflapin inhibits MMP-3 and ICAM-1

Gupta et al. 2011 [98]

Open: Shallaki tablet (6 g/d) or tablet and ointment together (n = 56 total)

Pain, stiffness and swelling, mental state (Jung scales). Radiology, hematology and biochemistry

2 months - both groups reported significant improvements in pain, stiffness and swelling and by radiology

Not a blind study. No control group.

Measurements were subjective and results not clear

Vishal et al. 2011 [88]

RDBP: Alfapin vs placebo (n = 30/group)

WOMAC, Lequesne’s and VAS for pain and serum biochemical, hematological and urine changes

30 days - significant improvement in pain and function but not in biochemistry

Small short trial to assess safety of new formulation

Kulkarni et al., 1991 [76]

RDBP crossover: Articulin-F capsule (W. somnifera, B. serrata C. longa Zinc complex) (n = 42)

Pain, stiffness, grip strength, Ritchie articular index, disability score, radiology.

3 months 2 weeks washout crossover 3 months - pain and disability were significantly improved over placebo.

Small prospective study.

Chopra et al. 2004 [99]

RDBP: RA-11 (W. somnifera, B. serrata, Z. officinale, and C. longa) vs placebo (n = 90)

WOMAC, VAS and hematological, urine and biochemical tests

32 week - WOMAC, VAS improved over placebo

Very high dropout rate

Chopra et al. 2013 [89, 100]

RDBP: Glucosamine vs celecoxib vs SGCG^c vs SGC (n = 440)

Weight bearing pain modified WOMAC

24 weeks - improvement in pain and function in all groups. For WOMAC pain, SGCG worked marginally better than SGC.

Some patients had adverse hepatic effects of SGCG and SGC.