The naturally occurring matrix of collagen peptides and CS derived from hydrolyzed porcine cartilage was provided by Rousselot BV and is commercially available as Colartix (further described as HCM). The dosage form was hydroxypropyl methylcellulose capsules containing either 0.5 g of HCM or 0.5 g of maltodextrin as placebo. Both supplements had a matched appearance.

The specifically developed mobile app Ingredients for Life, which was available free of charge in the app store for both Android and iOS operating systems, was promoted to people known to experience joint pain after exercise by selected ambassadors (physiotherapists, personal fitness trainers, sports coaches, and nutritionists). Potential study participants did not suffer from any diagnosed medical conditions but would seek nutritional supplementation to support joint health. Participant recruitment, allocation, data collection, and follow-up were performed through the mobile app, and participants were guided through the necessary study information (privacy policy, research protocol, criteria, supplement ingredients information, and supplementation instructions) and taken through the weekly reporting of joint pain via remote demonstrations. Upon consent, participants received the blinded study supplements with nutrition information and directions for use via post. At week 12, the app reminded the participants to continue reporting joint pain scores while no longer taking supplementation. Upon study completion, participants received an individual report via email. All communication between researchers and participants was done via the app to avoid direct contact during the COVID-19 pandemic.

Consumers seeking to reduce activity-related joint pain with a nutritional supplement were recruited through the networks of various sports and health clubs across the United Kingdom via the Ingredients for Life mobile app. Eligibility criteria were as follows: older than 18 years, no medical diagnosis, no medication, frequent physical activity, and not pregnant or breastfeeding. Physical activity included all types and levels from gardening, brisk walking to marathon training. Individuals already using any type of supplementation with vitamin D, curcumin or turmeric, other collagens, green mussel extract, Boswellia, glucosamine, CS, and hyaluronic or folic acid were excluded from the recruitment. The app was downloaded 277 times; 36 people did not meet the inclusion criteria and were notified through the app, 17 declined to participate, and 11 did not start the study for other (unknown) reasons. Therefore, 213 healthy individuals, aged 18-72 years, were randomized into the study, of whom 201 participants completed the study.

The study was conducted in accordance with the Declaration of Helsinki. All data are protected under GDPR. The study was a 2-arm parallel design with participants randomly allocated (5:1 randomization) to either the HCM group (n=177) or the placebo group (n=36). Participants took 2 capsules with a total of 1 g of supplement or placebo per day for a duration of 12 weeks followed by a washout period of 4 weeks without supplementation but continuous logging of joint pain scores, amounting to a total study duration of 16 weeks. Participants did not receive any dietary or lifestyle advice either via a web platform or during the video calls but were given guidance on how to take the supplement and log their data. An overview of the study design is shown in Figure 1.

At the study start, demographic data including gender, age (in ranges), height, weight, and waist measurement were collected. The type and duration of the weekly physical activities were captured, providing a preset of choices for type (eg, gardening, brisk walking, running, cycling, tennis, swimming, strengthening sports, weightlifting [average weight squat 70 kg], flexibility or balance, yoga, Pilates, option to fill in any unlisted) and duration (eg, 0-5 hours, 6-10 hours, 11-15 hours, 16-20 hours, or >20 hours) presenting the total weekly hours of all activities combined. Each participant indicated one area of major joint pain in a body silhouette (Multimedia Appendix 1) and reported their baseline joint pain score using a visual analog scale (VAS; Multimedia Appendix 2) from 0 (pain free) to 10 (unspeakable) [35]. Each week, participants reported the level of pain using the same tool, which has been shown to retrieve accurate results on subjective pain measures of participants [36-38]. Participants’ identities were anonymized, allowing researchers to retract data at any time throughout the study if necessary.

To assess the impact of physical activities of different intensity levels, a weighting matrix was developed (Multimedia Appendix 3) categorizing participants by performing low- (L), moderate- (M), and high-intensity (H) activities and by the number of performed activities (1×, 2×, 3×). Walking, swimming, flexibility or balance, gardening, cycling, dancing, and horse riding were considered activities of low joint impact. Weight lifting, gym cardio, and boxing were considered activities of moderate joint impact. Running, ball sports, skipping, and heavy weight lifting were considered activities of high joint impact. This classification was based on the UK National Health Services guidelines [39]. As an example, an LLL participant practiced 3 different types of low-intensity sports, an LLM practiced 2 different types of low- and 1 type of moderate-intensity sports, and an LLH practiced 2 different types of low- and 1 type of high-intensity sports. The other subgroups follow the same rationale. Given that only 1 participant fitted into the category MHH (1× moderate and 2× high), this participant was considered in category 5 (LHH or MMH or MHH; Multimedia Appendix 4).

The intention-to-treat population included all participants who had taken at least 1 supplement and reported at least 1 weekly VAS measure after exercise. Continuous variables were described using mean value, SD, and range, and categorical variables were described using frequencies and percentages. The effect of treatment (HCM vs placebo) on mean joint pain scores was compared with the baseline score (week 1) using a generalized linear mixed-effects (LME) model with random intercept (b_i) for each participant. Follow-up time was treated as a factor variable. The full LME model with interaction is described by the following equation:

The interaction term represents the difference between the changes in the expected joint pain, comparing a participant from the HCM group with a placebo group participant for a given follow-up time. For nonsignificant interaction terms, it can be concluded that there was not enough evidence for differences in the joint pain profile over time between the supplementation groups. In that case, a second analysis was performed including only the main effects. The effect of gender on mean joint pain score was investigated by adjusting the model for gender and looking at 2-way and 3-way interaction terms between treatment, time, and gender. First, the full model was evaluated, and nonsignificant interaction terms were stepwise dropped from the model. Two exploratory subgroup analyses were performed for the HCM group only. The first analysis focused on the effect of age, and the second on the effect of physical activity intensity on mean joint pain score over time. For both analyses, a full LME model was constructed. Of note, the placebo group was not considered because the number of participants was too low for some categories of the age groups and the physical activity intensity.

Systematic departures from the normality assumptions were verified using QQ plots. Scatterplots of the residuals versus the predictions were used to check for systematic departures from the mean model. The LME analyses were performed using the lmne package in the R environment (R Core Team, 2021) [37]. For the main analysis, a P value of <.05 was considered as statistically significant. For the subgroup analyses, a P value of <.01 was considered statistically significant. All analyses were performed using R software (version 4.1.2; R Core Team, 2021). No imputation of missing data was performed.

A total of 277 individuals downloaded the app and completed the questionnaire. Of these, 36 people did not meet the inclusion criteria and were notified through the app, 17 declined to participate, and 11 did not start the study for other (unknown) reasons; 213 healthy individuals, aged 18-72 years, were randomized into the study. The consort flowchart with the study details is shown in Figure 2.

Of the 213 enrolled participants, 177 were assigned to the HCM group and 36 to the placebo group. Both groups present a similar age range (HCM group: mean 40.2, SD 11.0 years vs placebo group: mean 42.1, SD 11.0 years) and gender distribution (Table 1).

Over 90% (201/213) of enrolled participants completed the study. Notwithstanding, the percentage of participants who completed the study in the HCM group (172/177, 97%) was higher than that in the placebo group (29/36, 81%; Table 2). Although no data were collected on the reasons that led to dropout, it is important to note that of the 5 participants in the HCM group who did drop out, 2 only quit the study after week 12 and thus after the supplementation had stopped.

To evaluate the effect of HCM supplementation on joint pain alleviation, joint pain scores were monitored over a period of 16 weeks (Figure 3A). Statistical analysis revealed that there is a significant interaction between supplementation and time (duration on supplementation; P<.001), suggesting significant differences between the effect of HCM and placebo on joint pain over time (Multimedia Appendix 5).

No differences in joint pain scores were detected at baseline between both groups (P=.12; Figure 3B). However, this pattern changed over time. Participants in the HCM group experienced a significant reduction in mean joint pain after week 3 (P=.009) compared with placebo. In addition, the estimated difference in joint pain scores between the HCM group and placebo group increased over time from 0.55 (95% CI 0.13-0.96, P=.009, week 3) to 2.74 (95% CI 2.30-3.17, P<.001, week 12). Although the mean joint pain score in the placebo group remained steady over time, it decreased from 5 (distracting to moderately strong pain) to 2 (minor) in 12 weeks in the HCM group (Figure 3A).

Supplement intake was stopped after 12 weeks. Subsequently, mean joint pain scores gradually increased in the HCM group (Figure 3A). Of note, the joint pain score in the HCM group was still significantly lower than that in the placebo group at week 16 (HCM group on average 0.8 lower vs placebo group, 95% CI 0.4-1.2, P<.001; Figure 3B). Although the joint pain score gradually increased after stopping supplementation, the participants’ joint pain levels remained significantly below the levels reported at baseline (week 16 on average 1.7 lower vs week 0, 95% CI 1.5-1.8, P<.001).

The joint pain score development for the HCM group and placebo group by gender is depicted in Figures 4A and 4B, and dropout rates by gender are reported in Multimedia Appendix 6. The profile plots for mean joint pain scores show that there was no evidence for a significant 3-way interaction between gender, supplementation, and time (P>.05). No interaction between gender and supplementation nor between gender and time (P>.05) was recorded. Furthermore, the LME model revealed that joint pain scores were statistically similar between genders (P=.47).

Several age categories in the placebo group presented only a small number of participants (Multimedia Appendix 7). Thus, the statistical analysis focused on the HCM group.

The LME model revealed a significant interaction between age and time (duration on supplementation; P=.005), highlighting that different age categories of the HCM group differed in their joint pain levels over time (Figure 5A). Therefore, the joint pain alleviation between age categories of the HCM group was compared using the participants younger than 30 years as reference (Figure 5B). After 10 weeks on supplementation, participants aged 40-49 years had on average higher joint pain scores than did participants younger than 30 years (mean difference 0.71, 95% CI 0.25-1.18, P<.01; Figure 5A). This difference, however, was only recorded late in the study. Study participants younger than 30 years reported the lowest joint pain scores among all age categories. Overall, all subgroups had joint pain levels significantly below the levels reported at the baseline of the study in all age categories (Multimedia Appendix 8).

Participants were divided into 5 subgroups based on the average weekly physical activity intensity (subgroups 5 and 6 were merged because group 6 only contained 1 participant). Because of a low number of participants (Multimedia Appendix 9) and some dropouts (Multimedia Appendix 10) in the placebo group, the statistical analysis focused on the HCM group. The demographic characteristics of HCM group participants were similar across physical activity intensity subgroups (Multimedia Appendix 11).

The mean joint pain score profiles for each physical activity intensity subgroup of the HCM group are shown in Figure 6A. There is not enough evidence to conclude on a significant interaction between the physical activity intensity and time (duration on supplementation; P=.04). The statistical analysis revealed that joint pain scores did not vary significantly among the different subgroups of physical activity intensity (P=.09), suggesting that the effect of HCM supplementation is consistent across different subgroups of physical activity intensity (Figure 6B; Multimedia Appendix 11).

This real-world study investigated the effect of an HCM-containing hydrolyzed collagen and CS on joint discomfort in a heterogeneous group of healthy and physically active consumers in their everyday life. The study was conducted during the COVID-19 pandemic, which represented an additional challenge. The specifically designed mobile app Ingredients for Life assured the successful recruitment and follow-up of the study participants and helped to minimize the impact of the challenges caused by the pandemic. Additionally, a fully digitally run study could positively impact the diversity and inclusivity of the research by reaching groups of participants that are typically underrepresented in clinical trials. People who have mobility or transportation issues, have busy work schedules, or are single parents (often woman) have a hard time participating in face-to-face studies but can easily enter a digitally run study. The use of a mobile app to monitor the candidates’ progress overcomes the known barriers of face-to-face studies including inability to access the research center and practical and logistical issues including hidden costs (transportation, childcare, etc) [40]. This contributes to the importance of including digital tools in research.

The digital approach of this study not only allowed researchers to gather first-hand feedback from consumers on the effectiveness of the supplement but also demonstrated that mobile apps can generate qualitative, quantifiable data. Participants weekly logged their joint pain scores in the app using a VAS enabling quantification of subjective feedback. Previous studies showed that VAS is a validated pain rating scale and no clinically relevant differences exist between digitally or traditionally obtained results [41], highlighting the value and feasibility of using digital tools in scientific research. The digital execution of the study and engaging character of the app might also explain the relatively low dropout rate with over 90% of participants completing the full study.

The fact that participants were not restricted to any diet or lifestyle highlights the relevance of HCM supplementation in a real-world setting where outside factors are not controlled (contrary to strictly controlled clinical trial designs). This design allowed us to investigate if the benefits of HCM observed in an OA mouse model extend to a human heterogeneous population [13].

The results demonstrate a significant improvement of joint pain in the HCM group compared with the placebo group as early as 3 weeks after starting supplementation. Joint pain reduction further continued over the full supplementation period of 12 weeks. With the start of the 4-week washout period, where HCM supplementation was stopped, joint pain scores gradually increased but remained significantly below those of the placebo group at week 16. These findings show that daily oral HCM supplementation has a clear potential to alleviate joint discomfort in consumers of different ages and genders and performing physical activities of different intensity levels. The perceived positive effects of HCM supplementation might also explain the fact that the dropout rate in the HCM group was lower than that in the placebo group. Although dropouts are a common issue in controlled trials [42-44], the dropout rates in this study were relatively low (3% for the HCM group, 19% for the placebo group), especially given the fact that this study was run at distance.

The first preclinical report on HCM supplementation in obese mice with posttraumatic OA demonstrated cartilage protection and a significant suppression of key proinflammatory microbial strains in the gut, suggesting that HCM might indirectly protect cartilage by modulating the gut microbiome [13]. Future randomized controlled trial (RCT) protocols are warranted to assess extended clinical outcomes, capturing biomechanical function, biomarkers of inflammation, or imaging techniques to get more insights into the mode of action of HCM.

Mohammed and He [45] reported a reduction of wrist joint discomfort by a daily supplementation of 2.5 g of hydrolyzed chicken collagen type II in healthy participants aged 40-65 years (65.5% women) after 4 weeks of supplementation. Despite the differences with this study, such as RCT versus real-world approach and duration, both studies showed the effectiveness of a low dose (2.5 g vs 1 g/day).

Different factors drive the risk for developing joint pain, namely gender, age, and sports intensity [5,7,46]. Compared with previous reports that commonly investigate rather homogeneous cohorts, this study recruited participants of different ages, genders, and performing sports activities of different intensity levels. Many trials focus on groups aged >40 years, as many parameters of aging contribute to joint pain [5]. Relevant factors involve a reduction in lubricating fluids, accelerated cartilage breakdown, stiffening of ligaments, and a decrease in strength of the surrounding muscle [47]. In line with previously mentioned trials, in this study, the subgroup of participants aged 40-49 years reported a higher average joint pain score than those aged <30 years. However, all age subgroups reported a clear improvement of joint pain in response to HCM supplementation. Interestingly, although women tend to have more risk factors that determine susceptibility for joint pain, the gender had no impact on the effectiveness of HCM supplementation. HCM further showed a consistent effect across subgroups of different physical activities.

A strength of this study is the joint pain score recording during the 4-week washout period. The effect of the washout was directly noticeable in the HCM group with a gradual increase of the joint pain score over the 4-week period, while remaining constant in the placebo group, suggesting a direct mechanistic correlation between supplementation and benefit, such as an anti-inflammatory effect. CS has been reported to have anti-inflammatory activity [24,25], and so has HCM in a preclinical study [13], which could explain the attenuation of the experienced benefit in the washout period. However, joint discomfort was still lower than in the placebo group at the end of the washout period, suggesting that other, more long-lasting mechanisms, might contribute to the pain-alleviating effect of HCM. Of note, the pain scores were still decreasing at 12 weeks of HCM supplementation when the washout phase had started. Future studies should investigate whether a longer supplementation time can stabilize the pain scores at lower levels. The supplementation of placebo during a washout period might further correct for the participant’s knowledge that the supplementation has stopped and its potential influence on the joint pain scores.

Real-world studies (RWSs) provide many advantages, but, like any study design, they can also offer limitations and have to be cautiously interpreted. Compared with RCTs, patient selection and follow-up in RWS are less strict, making it harder to check patient compliance (eg, correct use of the supplement) and adding the risk of selection bias [32,48]. With the use of the Ingredients for Life app for patient selection, randomization, and follow-up, we tried to overcome these limitations. Other biases inherent to RWS, like a lack of randomization or a control supplement, were avoided by designing the study plan as a randomized, single-blind, placebo-controlled approach [32,49]. Although one of the main strengths of RWS is the large size and heterogeneity of the study population, it can also form the base of a major limitation. Diversity in participants increases data variation and might dilute the treatment effect [50], making it harder to find differences between groups; this is also called the β-error problem [49]. Test population heterogeneity can also lead to missing data [48]. This could be observed in this study, especially when dividing the data set into subgroups; the placebo groups of several age categories or activity intensities were too small to draw a trustworthy conclusion based on statistical data.

It is clear that each study design has its own strengths, and RCT and RWS could be seen as complementary, allowing the measurement of both the efficacy and effectiveness, respectively. It is important to combine data from both study types to overcome the limitations of each of them.

In general, this study shows that the digital execution of a real-world study is a promising and relevant design to assess the effects of dietary supplementation on health.

To the best of our knowledge, this is the first digitally supported, real-world study showing the benefits of an HCM of collagen and CS in joint health in the general public across all age groups (18 years or older) and genders with a physically active lifestyle. This study provides evidence of the scientific value of user-friendly digital tools such as the Ingredients for Life mobile app that encourage consumer cooperation (as shown by the low dropout rate) and simplify participant tracking.

This real-world study also shows that adults of different backgrounds and lifestyles with joint discomfort, who were otherwise healthy, physically active, and not consuming other mobility-related supplementation, significantly benefited from HCM supplementation to reduce joint discomfort, regardless of age, gender, and physical activity intensity. This type of study design allows for a closer connection with the final consumer and assessing supplement effectiveness in circumstances of everyday life.