Introduction

Ageing is a degenerative process caused by a decline in physiological functions.1 Individuals of the same age experience ageing differently due to intrinsic factors (e.g., genetics) and extrinsic factors (e.g., pollution, UV radiation, chemicals, toxins, and lifestyle). As a result, body morphology and physiology vary.1-3 Skin ageing involves a decline in skin quality due to overlapping intrinsic and extrinsic factors.4,5 Visible signs of ageing often drive individuals, to invest in cosmetics and treatments aimed at preventing or reversing their effects.6

Skin ageing involves reduced levels of collagen, elastin fibres, and hyaluronic acid (HA).7,8 Collagen and HA form the extracellular matrix, providing skin structure, while HA maintains hydration, and elastin fibres preserve elasticity.7-9 Their decline leads to reduced skin integrity, moisture loss, wrinkles, and sagging.3,9 Sagging skin (94.3%) and wrinkles (100%) were observed in individuals aged 50-81.10 Kim et al. (2018) suggested that collagen supplements counteract this decline, promoting elastin fibre regeneration and increasing HA levels, thereby improving skin health.7

Hydrolysed collagen supplements act through two pathways. The first involves free amino acids supporting collagen, elastin, and HA synthesis while suppressing MMP-2, an enzyme that degrades type IV collagen, thereby increasing its levels.11 The second involves collagen oligopeptides (Gly-Pro-Hyp) binding to fibroblast receptors, stimulating collagen, elastin, and HA production.12 However, the link between collagen oligopeptides and fibroblast receptors remains unclear.

Research on collagen supplementation and its effects on skin is limited, with mixed findings from previous studies. This has led to interest in systematically evaluating its impact on skin hydration and elasticity. This study aims to assess collagen’s role in reducing age-related declines in skin hydration and elasticity and to determine the effective dosage for these benefits.

Methods

This systematic review and meta-analysis followed PRISMA 2020 guidelines. The study protocol was pre-registered per standard methods. Inclusion criteria: (1) Population: Males and females; studies in English or Indonesian (2010–2021); (2) Intervention: Oral collagen for skin hydration and elasticity; (3) Comparison: Placebo or no supplementation; (4) Outcomes: Skin hydration and elasticity before and after supplementation. Only randomised placebo-controlled trials were included. Studies with incomplete or missing data were excluded.

Major databases (PubMed, Cochrane Library, Clinical Key, and ProQuest) were searched for studies published from 2010 to April 2021. The search used MeSH and Freestyle terms, including (“collagen” AND “oral” AND “skin” AND “placebo” AND (elasticity OR wrinkle) AND (hydration OR moisture)). Minor adjustments were made to fit the syntax of each database. The search process and study selection are detailed in the PRISMA 2020 flow diagram [Figure 1]. Titles and abstracts were manually screened by two independent reviewers based on pre-defined criteria. Full-texts of eligible studies were retrieved and assessed. Disagreements were resolved through discussion or a third reviewer if needed. Study selection details have been presented in the PRISMA 2020 flow diagram [Figure 1].

Figure 1: PRISMA 2020 flow diagram search results.

Export to PPT

Two authors independently extracted data using a predefined form, including study (author, year, country, design, sample size), participant (age, sex, criteria), intervention (collagen type/dosage), outcomes (skin hydration/elasticity before and after), and statistical data (mean, standardised differences (SD), confidence interval (CI), effect sizes). Discrepancies were resolved by consensus. The Cochrane Collaboration’s Risk of Bias Tool was used to assess selection, performance, detection, attrition, and reporting bias. Each was rated as low, unclear, or high risk based on predefined criteria.

Statistical analysis

Meta-analysis was conducted using R software (4.1.1) with the meta and metafor packages. Skin hydration and elasticity were analysed using a random-effects model, with standardised mean differences (SMD) and 95% CIs. SMDs were calculated using the metafor package in R software, with the escalc function used to compute the change scores (post-minus-pre values).

Heterogeneity was assessed using I2 and Tau2 (τ2). I2 quantifies the proportion of variability due to heterogeneity across studies, with interpretation guided by Higgins’ thresholds: 25% (low), 50% (moderate), and 75% (high). Hydration (I2 = 55.5%) and elasticity (I2 = 56.5%) showed moderate heterogeneity.13 Tau2 values further supported moderate variability for hydration (0.1331, p = 0.0766) and slightly less for elasticity (0.096, p = 0.0729). The higher variability in hydration may reflect methodological differences, sample characteristics, or collagen formulations.

Prediction intervals were used to estimate effects in future studies. Forest plots visualised the results, and publication bias was assessed using funnel plots and Egger’s test. Sensitivity analysis with a leave-one-out approach was performed for hydration and elasticity to test robustness. The meta-analysis was re-run, excluding each study, and pooled SMDs were compared. All data and R code are available upon request.

Result

The results revealed 19 studies (PubMed), 23 studies (Cochrane Library), 33 studies (Clinical Key), and 263 studies (ProQuest), totalling 338 studies. Research screening was conducted, and 10 studies were included in this systematic review. The assessment of bias risk has been listed in Table 1 and the summary of the included studies is presented in Figure 1, Tables 2-5. The studies were conducted between 2014 and 2021 from Italy, Germany, Korea, Japan, China, the U.S., and Slovenia. Participants were mostly female, aged 31 to 62 years. Collagen sources varied, including chicken, bovine, fish, and porcine, with doses ranging from 246 mg/ day to 10,000 mg/day [Table 5].

Table 1: Risk of bias assessment

No. Author [year] Random sequence generation (selection bias) Allocation concealment (selection bias) Blinding of participants and personnel (performance bias) Blinding of outcome assessment (detection bias) Incomplete outcome data (attrition bias) Selective reporting (reporting bias) Overall risk of bias assessment (low, unclear, high) 1 Cerbo et al17 [2015] ? ? ? ? ? + unclear 2 Bolke et al9 [2019] ? ? ? ? + + unclear 3 Czajka et al21 [2018] ? + + + + + unclear 4 Tak et al15 [2021] + ? ? ? + + unclear 5 Zmitek et al19 [2020] + ? ? ? ? + unclear 6 Inoue et al20 [2016] + + + + + + low 7 Schwartz et al18 [2019] + + + + + + low 8 Yoon et al14 [2014] ? ? + ? + + unclear 9 Kim et al7 [2018] ? + ? ? + + unclear 10 Asserin et al16 [2015] ? ? ? ? + + unclear

Table 2: Characteristics of the studies (1)

No Author [year] Level of Evidence; Study Design; Risk of Bias Participant; Inclusion criteria; Intervention setting Intervention group (IG) Control Group (CG) 1. Cerbo et al.17 [2015] 1b; RCT; Unclear Female (n = 30); Moderate/ severe facial photoageing (measured by VAS/ visual analogue scale); Outpatient clinic VISCODERM® Pearls (IBSA FARMACEUTICI ITALIA Srl, Lodi, Italy) containing Pycnogenol 15 mg, collagen (124 mg), chondroitinsulfate 40 mg, glucosamine sulphate > 3%, low-molecular-weight HA 20 mg, coenzyme Q10 10 mg Placebo 2. Bolke et al.9 [2019] 1b; RCT; Unclear Female (n = 72); Age ≥ 35 years, any type healthy skin; Outpatient clinic ELASTEN® (QUIRIS Healthcare, Gütersloh, Germany) containing 2.5 g collagen peptide, 666 mg acerola fruit extract, 80 mg vitamin C, 3 mg zinc, 2.3 mg vitamin E, 50 μg biotin Placebo 3. Czajka et al.21 [2018] 1b; RCT; Unclear Female (n = 91), male (n = 29); Male/ female (volunteer), age between 21-70 years, any body mass index and ethnic, balanced diet, had the ability to understand the study related information and to give written informed consent; Outpatient clinic GOLD COLLAGEN® (MINERVA Research Labs, London, UK) containing hydrolised fish collagen type I 4,000 mg, HA, glucosamine hydrochloride, L-carnitine, maca, and black pepper extract Placebo 4. Tak et al.15 [2021] 1b; RCT; Unclear n = 84; Age between 40-60 years with transepidermal water loss (TEEL) score ≥4 (measured by Tewameter); Outpatient clinic CTP (collagen tripeptide) from Nile Tilapia (Oreochromis niloticus) containing hydrolised collagen 1000 mg Placebo 5. Zmitek et al.19 [2020] 1b; RCT; Unclear

Female (n = 31) age 55.0 ± 7.6 years (CG n = 15), age 53.7 ± 6.2 years (IG n = 16); Caucasian female age between 40-65 years, Fitzpatrick skin type II, III, with signs of skin ageing

(wrinkle/ poor skin tone, dry); Outpatient clinic

Hydrolised collagen syrup (Valens Int. d.o.o, Šenčur, Slovenia) containing hydrolised fish collagen 4000 mg, water-soluble CoQ10 50 mg, vitamin C 80 mg, vitamin A 920 μg, biotin 150 μg Placebo 6. Inoue et al.20 2016] 1b; RCT; Low n = 80 (IG1 n = 28, IG2 n = 26, CG n = 26); Age between 35-55 years, subject conscious of their dry and rough skin, body mass index <30kg/m2, not regularly using other supplements or health foods, no treatment with sex hormones over the prior 3 months, no pregnant; Outpatient clinic IG1: L-CP / low molecular collagen peptide (5000 mg) (Nita Gelatin Inc., Osaka, Japan); IG2: H-CP / high molecular collagen peptide (5000 mg) (Nita Gelatin Inc., Osaka, Japan) Placebo (maltodextrin TK-16, Matsutani Chemical Industry Co., Ltd., Itami, Japan) 7. Schwartz et al.18 [2019] 1b; RCT; Low n = 128 (IG n = 56, CG n = 56); Overall healthy, age between 36-59 years, willing and able to provide consent, any Fitzpatrick skin type, shows sign of skin ageing, doesn’t have skin’s disease, willing to prevent skin product that claim can increase skin hydration, brightening, anti-ageing, willing to arrive at the test center for the study’s scheduled visits, prior to which they would apply no topical treatments, including moisturizers and creams; Outpatient clinic BioCell Collagen (BioCell Technology, Anaheim, CA) containing hydrolised collagen type II 300 mg, glycosaminoglycan 100 mg, chondroitin sulphate, HA 50 mg Placebo 8. Yoon et al.14 2014] 1b; RCT; Unclear n = 44 (IG n = 22, CG n = 22); Healthy female volunteer, age ± 40 years with wrinkle skin level ± 2; Outpatient clinic Astaxanthin capsule (triglyceride moderate chain 480 mg, Haematococcus pluvialis microalgae extract 20 mg), fish collagen tablet (enzymatic hydrolised fish collagen 0.75 g) Placebo 9. Kim et al.7 [2018] 1b; RCT; Unclear Female (n = 53); Female healthy age between 40-60, diagnosed wrinkle in the eye area by dermatologist with global photodamage score between 2-6, willing to follow the research, cooperative during the study; Outpatient clinic LMWCP (low molecular weight collagen peptide) (LMWCP 1000 mg, vitamin C 100 mg, mixed fruit concentrate 3000 mg, mixed flavor 200 mg, expiens 1900 mg, sweetener 12,5 mg; water 43,7 ml) Placebo 10. Asserin et al.16 [2015] 1b; RCT; Unclear Female (n = 33 (IG1n = 11, IG2 n = 11, CG n = 11)); Japanese female age between 40 and 59 years, low skin water content, no pregnant, doesn’t have systemic disease, no intolerance of fish and gluten, doesn’t consume any drugs and food supplements; Outpatient clinic Peptan® fish (10 g) for IG1, porcine (10g) for IG2 (Rousselot BVBA, Ghent, Belgium) Placebo

Table 3: Characteristics of the studies (1)

No Author [year] Instrument Result (Hydration) Result (Elasticity) 1. Cerbo et al.17 [2015] Hydration & elasticity : Skin Tester (manufactured by Selenia Italia (Pisa, Italy), distributed by Dermal Medical Division (Bologna, Italy)) Significantly increase (IG), no significantly change (CG) Decrease (IG), no significantly change (CG) 2. Bolke et al.9 [2019] Corneometer CM 825 for hydration and Cutometer MPA 580 for elasticity (Courage - Khazaka, Cologne, Germany) Significantly increase (IG), significant change between IG & CG Significantly increase (IG & CG) 3. Czajka et al.21 [2018] Elasticity: SkinLab USB Elasticity Module (DermaLab® Series, Cortex Technology, Hadsund, Denmark). Doesn’t measure hydration. Doesn’t measure hydration. Significantly increase (IG), no change (CG) 4. Tak et al.15 [2021] Corneometer CM 825 for hydration and Cutometer MPA 580 for elasticity (Courage - Khazaka, Cologne, Germany) Increase for both groups (IG > CG) Decrease for both groups 5. Zmitek et al.19 [2020] Hydration & elasticity : DermaLab Series, SkinLab Combo Hydration & Elasticity probe (Cortex Technology ApS, Hadsund, Denmark) Decrease for both groups and no significant change between the groups Decrease for both groups and no significant change between the groups 6. Inoue et al.20 [2016] Elasticity: Cutomater SEM575 (Courage - Khazaka, Cologne, Germany). Doesn’t measure hydration Doesn’t measure hydration. Significantly increase (IG 1 & IG 2), no change (CG) 7. Schwartz et al.18 [2019] MoistureMeterSC (Delfin Technologies, Kuopio, Finland) for hydration and Cutometer MPA 580 (Courage - Khazaka, Cologne, Germany) for elasticity Significantly increase for both groups (IG and CG) Significantly increase for both groups (IG and CG) 8. Yoon et al.14 [2014] Corneometer CM 825 for hydration and Cutometer MPA 580 for elasticity (Courage - Khazaka, Cologne, Germany), Tewameter (Courage - Khazaka Electronic, Cologne, Germany) for transepidermal water loss (TEWL), Increase for both groups (IG and CG) Significantly increase (IG), no change (CG) 9. Kim et al.7 [2018] Corneometer CM 825 for hydration and Cutometer MPA 580 for elasticity (Courage - Khazaka, Cologne, Germany) Significantly increase for both groups (IG > CG) Significantly increase (IG), increase (CG) 10. Asserin et al.16 [2015] Hydration: Corneometer (Courage - Khazaka Electronic, Köln, Germany) for skin moisture & Tewameter (Courage - Khazaka Electronic) for TEWL. Doesn’t measure elasticity. Significantly increase (IG 1 & IG 2), no change (CG) Doesn’t measure elasticity

Table 4: Characteristics of the studies (2)

No. Study Country Study duration Total participant Subject sex Mean age (years) 1. Cerbo et al.17 [2015] Italy 1 month 30 F 47.5±1.6 2. Bolke et al.9 [2019] Germany 3 month 72 F 50.6±11 3. Czajka et al.21 [2018] Italy 3 month 91 F, M 43±13 4. Tak et al.15 [2021] Korea 3 month 84 F 49.9±6.5 5. Zmitek et al.19 [2020] Slovenia 3 month 31 F 54.4±6.8 6. Inoue et al.20 [2016] China 2 month 80 F 43±4 7. Schwartz et al.18 [2019] America 3 month 128 F 51±5 8. Yoon et al.14 [2014] Korea 3 month 44 F 51±5 9. Kim et al.7 [2018] Korea 3 month 53 F − 10. Asserin et al.16 [2015] Japan 2 month 33 F −

Table 5: Collagen supplement dosage

Study Dose/ day Total dose/ day Cerbo et al.17 [2015] 2x124 mg 246 mg Bolke et al.9 [2019]ab 2500 mg 2500 mg Czajka et al.21 l [2018] 4000 mg 4000 mg Tak et al.15 [2021] 4x1000 mg 4000 mg Zmitek et al.19 l [2020] 4000 mg 4000 mg Inoue et al.20 [2016]b 5000 mg 5000 mg Schwartz et al.18 [2019] 2x300 mg 600 mg Yoon et al.14 [2014]a 4x750 mg 3000 mg Kim et al.7 [2018]ab 1000 mg 1000 mg Asserin et al.16 [2015]a 10000 mg 10000 mg Discussion Systematic review

The geographic spread and steady publication of studies from 2014 to 2021 reflect an increasing international focus on oral collagen supplementation and its potential impact on skin aging.

Oral Collagen Supplements and Skin Hydration

Four studies (Yoon et al., Kim et al., Bolke et al., and Tak et al.) using the Corneometer CM825 (Courage-Khazaka, Cologne, Germany) consistently showed a significant increase in skin hydration [Table 6].7,9,14,15 Hydrolysed collagen enhances hydration due to its moisture-retaining properties.14 Gly-Pro-Hyp peptide levels peak in the bloodstream 1–2 hours post-consumption, stimulating dermal fibroblast migration and growth.7 While the exact mechanism remains unclear, Asserin et al. suggested fibroblasts recognise Gly-Pro-Hyp, triggering collagen, elastin, and HA synthesis in the epidermis and dermis, improving hydration and elasticity.7,14,16 This mechanism is supported by Tak et al., Cerbo et al., and Asserin et al.15–17 Additionally, a clinical study detected Pro-Hyp in urine after collagen ingestion, confirming its stability and ability to reach skin tissue.15

Table 6: Effects of oral collagen-based supplement on skin’s hydration

No Study Group Total subjects Hydration (mean±SD) Interpretation Before intervention After intervention 1. Cerbo et al.17 [2015] IG 15 20.3±1.6 33.2±3.2 Significantly increase* (p <0,0001) CG 15 20.07±1.5 19.9±1.6 No significant change 2. Bolke et al.9 [2019] IG 36 35.0±4.8 44.5±4.4

Significantly increase*

(p <0,0004)

CG 36 33.7±5.1 36.6±5.7 3. Tak et al.15 [2021] IG 36 192.5±21.4 209.0±18.7 Increase (>) CG 38 199.7±21.9 212.1±16.0 Increase 4. Zmitek et al.19 [2020] IG 16 255.5±65.6 200.0±52.3 Decrease CG 15 230.6±67.3 194±1±44.4 Decrease 5. Schwartz et al.18 [2019] IG 58 41.84±29.19 61.31±24.53 Significantly increase CG 55 39.59±23.95 62.08±28.41 Significantly increase 6. Yoon et al.14 [2014] IG 22 46.0±14.0 114.6±14.5 Significantly increase CG 22 44.5±13.2 98.2±10.4 Significantly increase 7. Kim et al.7 [2018] IG 26 47.79±12.48 61.14±11.31 Increase (>)*(p <0,003) CG 27 48.43±12.52 53.02±13.59 Increase 8. Asserin et al.16 [2015] IG1 11 51.37±2.79 57.49±3.01 Significantly increase IG2 11 50.68±1.75 64.83±2.44 Significantly increase* CG 11 50.55±2.61 52.51±2.6 No significant change

The increase in skin hydration may also be due to the structural similarity between hydrolysed bovine collagen and human collagen. Bovine collagen shares 31% similarity with type I collagen alpha chain I, 18% with type I collagen alpha chain II, and 13% with type III collagen alpha chain I.9 Among collagen sources, bovine collagen exhibits the highest similarity to human collagen.9 Schwartz et al. attributed increased hydration partly to higher water intake from supplements and placebos.18 Tak et al. suggested that seasonal changes from winter to summer (February–June) during Kim et al.’s study may have influenced hydration and elasticity levels.7,15

However, Zmitek et al. found no significant changes in skin hydration after 12 weeks, as measured by the DermaLab Series, SkinLab Combo hydration probe (Cortex Technology Aps, Hadsund, Denmark).19 This contrasts with other studies showing increased hydration. The decline observed in Zmitek et al.’s study is likely due to seasonal changes, transitioning from late autumn to winter.19 Low humidity, combined with colder outdoor temperatures and indoor heating, may have negatively impacted skin hydration.19

Oral Collagen Supplements and Skin Elasticity

Three studies (Yoon et al., Bolke et al., and Schwartz et al.) using the Cutometer MPA580 (Courage-Khazaka, Cologne, Germany) consistently reported a significant increase in skin elasticity [Table 7].9,14,18 This improvement is attributed to molecular changes in the dermis’s extracellular matrix.14 Yoon et al. and Kim et al. suggested that hydrolysed collagen enhances elasticity by promoting elastin fibre synthesis and inhibiting MMP-2 and MMP-9, enzymes that degrade elastin.7,14 MMP levels significantly impact elastin and, consequently, skin elasticity.7 Additionally, skin hydration plays a crucial role, as elastin conformation changes only occur in the presence of hydrated proteins.7 These findings align with Kim et al.’s conclusions, further supporting the increase in skin elasticity.7

Table 7: Effects of oral collagen-based supplement on skin’s elasticity

No Study Group Total subjects Elasticity (mean±SD) Interpretation Before intervention After intervention 1. Cerbo et al.17 [2015] IG 15 12.4±0.9 9.7±1.9 Decrease CG 15 12.3±0.98 12.5±0.9 No significant change 2. Bolke et al.9 [2019] IG 36 0.69±0.05 0.81±0.04

Significantly increase*

(p <0,0004)

CG 36 0.71±0.06 0.75±0.06 Significantly increase 3. Tak et al.15 [2021] IG 36 0.40±0.11 0.33±0.10 Decrease CG 38 0.42±0.10 0.36±0.09 Decrease 4. Zmitek et al.19 [2020] IG 16 1.49±0.38 1.38±0.33 Decrease CG 15 1.53±0.64 1.59±0.68 5. Schwartz et al.18 [2019] IG 58 0.20±0.07 0.40±0.08 Significantly increase*(p<0,05) CG 55 0.22±0.06 0.34±0.15 Significantly increase 6. Yoon et al.14 [2014] IG 22 0.61±0.06 0.63±0.02 Significantly increase* CG 22 0.64±0.07 0.62±0.02 No significant change 7. Kim et al.7 [2018] IG 26 0.64±0.14 0.71±0.16

Significantly increase*

(p <0,027)

CG 27 0.63±0.13 0.64±0.15 Increase 8. Czajka et al.21 [2018] IG 61 7.9±0.2 9.8±0.2

Significantly increase*

(p <0,001)

CG 59 6.9±0.3 7.0±0.3 No significant change 9. Inoue et al.20 [2016] IG1 28 0.739±0.058 0.749±0.039 Increase IG2 26 0.725±0.058 0.767±0.058