The primary purpose of this pilot study was to understand if the clinical safety and efficacy of hBO previously demonstrated in veterinary species (Beths et al., 2020; Pearson et al., 2009; Seabaugh et al., 2020; Whittem et al., 2021) would translate into decreased knee pain in a human population with X-ray diagnosed knee OA. Because of uncertainties in translating doses from animals to humans, we explored the safety and efficacy of three different doses of hBO. At the primary endpoint, all three doses of hBO resulted in changes in pain VAS that were distinct (no overlap of 95% CI) and highly significant (all p < 0.0001) vs. the change in the placebo group. Since statistically significant changes in measured clinical endpoints do not necessarily equate to clinical relevance, thresholds for meeting minimal clinically important improvements (MCII) have been determined in the context of hip and knee OA for both pain VAS and WOMAC function (Tubach et al., 2005). In addition, since absolute changes in pain and function are influenced by initial disease severity, MCII thresholds have been further refined based on tertiles of baseline knee pain (Tubach et al., 2005). In the current trial of hBO, the mean baseline pain VAS of approximately 78 placed this population into the high pain tertile (baseline VAS > 66.2), for which the MCII has been defined as (Mean [95% C.I.] -36.6 (-38.3 to -34.7). The LS mean changes at the primary endpoint in the pain VAS scores for each hBO group, but not the placebo, all fall within the 95% C.I for meeting the MCII threshold.
The exploratory endpoints are reported with p-values uncorrected for multiple testing and are presented as nominal, requiring independent prospective replication to confirm their significance. At study end, the changes from baseline in each hBO dose for mWOMAC total and subscale scores (pain, function, stiffness), as well as the change from baseline in the SF-36 total score and each SF-36 domain score (physical function, physical health, body pain, social function, mental function, emotional health, vitality, general health) were distinct (no overlap of 95% CI) and significant (all p < 0.0001) vs. the change in the placebo group. Since regaining function is of high importance to OA patients, an MCII threshold has also been defined for WOMAC function (Tubach et al., 2005). For patients in the severe baseline knee pain tertile, the MCII threshold was determined to be (Mean [95% C.I.] -20.4 (-22.5 to -18.1) on a 100-point WOMAC function scale (Tubach et al., 2005). The changes in mWOMAC function scores in the current study, normalized to a 100-point scale (Mean [95% C.I.]), were − 38.2 (-44.3, -32.2), -40.9 (-46.9, -34.7) and − 36.3 (-42.1, -30.6) for the hBO-HD, hBO-MD, and hBO-LD groups respectively. These all clearly exceed the MCII threshold for improved WOMAC function. In summary, the mWOMAC function and SF-36 data provide support for hBO not only decreasing pain as per the primary pain VAS outcome, but also improving joint function and patient quality of life.
The OMERACT–OARSI responder index, developed to distinguish between an efficacious active treatment and a placebo, has been validated in 14 human clinical trials (Pham et al., 2004). In the current study the percent responders was similar in all groups on Day 14; however, by Day 28 the % responders in the hBO groups was beginning to differentiate from the placebo group. By Day 56, the percent responders was numerically much greater in each hBO group in comparison to the placebo group (approximately 80% vs. 10.9%) and the distribution of responders in each hBO group was significantly different (nominal p values all < 0.0001) from the placebo group.
For each endpoint assessed longitudinally (i.e., pain VAS, mWOMAC and OMERACT-OARSI responder index), the onset of efficacy was gradual; however, between Days 28 and 56, there was an accelerated accumulation of efficacy with both maximal efficacy and differentiation from placebo achieved at study end. This is illustrated in the pain VAS time course (Fig. 2) and is consistent with observations from two prior canine OA studies (Beths et al., 2020; Whittem et al., 2021) in which efficacy of hBO increased with duration of dosing.
In accordance with the observed safety profile of hBO in veterinary species and in preclinical safety studies, there were no observed safety concerns at any dose over the 56-day study. All clinical safety measures and biochemical safety measures fell within normal limits at all time points. There were no SAEs, plus the AEs reported were mild in nature and could not be ascribed to the intervention. Over the dose range explored and the 56-day study period, hBO exhibited a favorable safety profile.
Limitations to the trial include focus on a population with relatively severe knee pain at baseline, an age range limited to 40–65, and a relatively recent (< 3 years) clinical diagnosis of OA. Although the baseline pain VAS scores were relatively high (approximately 78) in the current study, they are consistent with knee OA trials conducted in India with similar patient populations (Shep et al., 2019; Srivastava et al., 2016); however, further studies would be required to understand the efficacy of hBO in populations with less severe knee pain, and in patients outside the studied age range, and who had been diagnosed with OA for periods of time > 3 year. In addition, we did not study the efficacy of hBO vs. other forms of OA (e.g., hip and hand), follow-up studies in these populations would be required. In the current trial, the placebo response (16.8% on Day 28, declining to 12.6% on Day 56) was about half the placebo response observed in a meta-analysis of COX-2 inhibitor trials in patients with knee pain in predominantly Western populations (Reiter-Niesert et al., 2016). Other knee pain OA trials conducted in India and based on pain VAS score outcomes, have demonstrated similar placebo responses to those in the current trial and it is possible that there are differences in pain placebo responses with respect to societal factors, and/or geography and/or ethnicity (Ramakanth et al., 2016; Sengupta et al., 2010). There were no obvious differences in the use of rescue medication in any of the hBO groups vs. the placebo group; however, there was also no clear pattern in the use of acetaminophen, and it is possible that greater reductions in use after hBO treatment might be observed in longer duration trials of hBO. Finally, although three doses of hBO were explored in the trial, no clear dose response was observed, and all three doses were similarly efficacious vs. the placebo. At the study end, there were small numerical advantages in pain VAS scores in the hBO-MD vs. the hBO-LD and, given the once-a-day regimen and safety profile, this mid-dose will likely be utilized in confirmatory human studies.
Although efficacy vs. joint function and/or pain has now been demonstrated in OA trials in both canines (Beths et al., 2020; Whittem et al., 2021) and humans, neither the identity of the active species in hBO, nor the mechanism of action of hBO are clear. In both in vitro (Pearson et al., 2008) and in vivo (Pearson et al., 2009; Seabaugh et al., 2020) studies, there is evidence of reductions in PGE2 after treatment with hBO. The relatively slow onset of clinical efficacy vs. joint pain is consistent with a mechanism in which membrane proinflammatory lipid precursors are cumulatively replaced with lipids that either replace proinflammatory lipid substrates, or which act via alternate mechanisms (e.g., inhibition of MAPKs) to reduce the inflammatory response (Chen et al., 2015, 2012). Interestingly it has also been demonstrated that the seed oil of another conifer species, Sciadopitys verticillate that contains high levels of sciadonic acid also contains a 2-monoacylglycerol, 2-sciadonylglycerol, that acts as a potent CB1 agonist (Nakane et al., 2000). To date, it is unknown whether hBO contains 2-sciadonylglycerol, or whether this molecule is generated via metabolism after oral administration of sciadonic acid containing seed oils. Finally, there is also evidence that other mechanisms may be playing a role in the pharmacology of hBO that is of relevance to OA. In in vitro cartilage explants studies, hBO increased chondrocyte number (Pearson et al., 2008) while in a pilot equine model of surgically induced joint degeneration, hBO almost completely inhibited X-ray measured bone changes, including osteophyte growth (Seabaugh et al., 2020). It is unclear whether these activities on chondrocytes and/or bone are contributing to the observed clinical efficacy of hBO vs. joint pain and function, and whether longer term the intervention might lead to disease modifying activity. Since bone remodeling is a feature of OA (Zhu et al., 2021), and osteophytosis has been demonstrated to correlate with joint pain and/or function (Muraki et al., 2015), it is possible that hBO is also playing a role in mitigating pathological bone remodeling in joint tissues.