Jarlsberg cheese consumption is associated with favourable changes in bone turnover markers and BMD in postmenopausal women and older men: a randomised clinical trial

Abstract

Introduction

Osteoporosis remains a significant global health issue, and prevention of age-related bone loss is of particular interest.

Aim

To assess whether daily Jarlsberg cheese fermented by Propionibacterium freudenreichii (PF) is associated with changes in bone turnover markers and bone mineral density (BMD) in postmenopausal women and men aged ≥55 years.

Methods

The study was performed as a randomised semi-cross-over designed multicentre trial using delayed start-up. 57 postmenopausal women and men ≥55 years, both with BMD T-score between −2.5 and 0, who did not use any osteoporosis medication were included. All participants received daily calcium and vitamin D. One group was randomly allocated to Jarlsberg from start till 32 weeks, whereas the second had a 16-week delay before Jarlsberg start-up for 32 weeks. Bone turnover markers and BMD were assessed every 16 weeks.

Results

In the Jarlsberg group, total osteocalcin (OC) increased from 18.8 to 23.5 ng/mL at week 32 (p<0.01), carboxylated OC from 11.8 to 17.3 ng/mL (p<0.001) and the ratio between procollagen type 1 N-terminal propeptide (P1NP) and crosslinked C-terminal telopeptide of type 1-collagen (CTX) (P1NP/CTX) from 168.8 to 203.9 (p<0.01). Lumbar spine BMD increased from 1.151 to 1.163 g/cm² (p=0.04) and total hip BMD from 0.929 to 0.933 g/cm² (p=0.02). The differences between groups in BTM were significant in favour of Jarlsberg cheese. After Jarlsberg was introduced, OC-level, biomarkers and BMD increased.

Conclusion

Jarlsberg intake alongside calcium and vitamin D was associated with favourable increases in OC-related markers, P1NP/CTX ratio and BMD in postmenopausal women and older men.

WHAT IS ALREADY KNOWN ON THIS TOPIC.

WHAT THIS STUDY ADDS

• Jarlsberg cheese, containing Propionibacterium freudenreichii, increases bone formation markers as osteocalcin and P1NP/CTX ratio and shows a positive effect on bone mineral density, suggesting antiresorptive effects.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Jarlsberg cheese may be considered as a dietary supplement for osteoporosis prevention, alongside calcium and vitamin D. The findings support further research into long-term effects.

Introduction

Osteoporosis is a major global health problem. Although hip fracture incidence has decreased in Norway and internationally over the past two decades, Norway still reports among the highest fracture rates.^1–4 Pharmacological treatments are effective but adverse effects may limit adherence and treatment length.^{5 6} Nutrition and lifestyle measures, including adequate calcium and vitamin D, sufficient protein and physical activity, remain important for prevention and management.⁷

Vitamin K₂ and 1,4-dihydroxy-2-naphthoic acid (DHNA) have been linked to beneficial effects on bone.^{8 9} Fermented dairy products contain vitamin K₂. Jarlsberg cheese is fermented with Propionibacterium freudenreichii (PF), producing relatively high amounts of vitamin K₂ and DHNA. Vitamin K₂ supports osteocalcin (OC) carboxylation, and both vitamin K₂ and DHNA have been proposed to influence bone formation and resorption.^9–12 In prior studies, it has been demonstrated that Jarlsberg intake increases total OC (tOC) and carboxylated OC (cOC) levels and improves the ratio between procollagen type 1 N-terminal propeptide (P1NP) and crosslinked C-terminal telopeptide of type 1-collagen (CTX) and P1NP/CTX ratio in premenopausal women and cross-country skiers.^{13 14}

A recent dose-response study in postmenopausal women and men above 55 years supported the feasibility of using OC-related endpoints to estimate an optimal effective dose of Jarlsberg and suggested a potential shift in bone turnover markers, primarily driven by reductions in CTX.¹⁵ Building on that work, the present trial evaluates Jarlsberg intake in a population with bone mineral density (BMD) T-scores between −2.5 and 0 and extends outcomes beyond biomarkers by including repeated BMD measurements.

The aim of this study in postmenopausal women and older men was to compare Jarlsberg cheese given in addition to calcium and vitamin D (J-group) with calcium and vitamin D alone (C-group), with respect to bone turnover markers (BTM) and BMD during the randomised 16-week comparative phase; further to explore any changes in these variables in the C-group after introducing Jarlsberg cheese; and finally, to explore the changes in these variables over 32 weeks of Jarlsberg exposure in the J- and C-groups after delayed start.

Material and methods

Study population

The study population consisted of postmenopausal women and men aged ≥55 years with BMD T-scores between −2.5 and 0. Postmenopausal status was defined clinically, and individuals with osteoporosis (T-score ≤−2.5) were not eligible. Use of medications with major effects on bone metabolism like antiresorptive, osteoanabolic therapy and systemic glucocorticoids was not permitted.

Exclusion criteria included eating disorders, symptomatic gastrointestinal disease, abnormal liver function, severe renal impairment, allergy to milk products or confirmed cancer. Type 1 diabetes mellitus and poorly controlled type 2 diabetes were excluded.

All eligible individuals provided informed consent prior to participation. Recruitment targeted individuals with risk factors for low bone mass, such as osteoarthritis and/or arthritis, history of bariatric surgery, early menopause, anti-oestrogen treatment or a family predisposition to osteoporosis, and was supported by advertisements at participating medical centres.

The study sample

Study approval date was 19 January 2023, the first participant was enrolled on 19 January 2023 and the last participant completed the study on 14 October 2024. After DXA screening, 57 participants were recruited by seven general practitioners and assigned to either the J-group or C-group using a pre-randomisation list.^{16 17} The J-group included 19 females and nine males, while the C-group included 24 females and five males. One participant from each group withdrew in the first week. The final sample comprised 42 women and 13 men, with a mean age of 64.9 years (range: 27.8–84.5) and BMI of 27.3 kg/m² (range: 19.7–44.5)^{16 17} (figure 1). The youngest was a 27.8-year-old woman but had premature menopause several years prior to inclusion due to premature ovarian insufficiency and therefore met the eligibility criteria. At baseline, the J-group reported 66 illnesses across 15 diagnoses and the C-group 75 illnesses across 16 diagnoses (table 1). Treatments of these conditions remained unchanged during the study period, and both groups were similar in demographics, vital signs and BMD.

Figure 1. Study design.

Table 1. Demographics, vital signs, baseline bone mineral density (BMD), BMD T-score and concomitant diseases or claims.

Variable categories and names		Estimators	Jarlsberg (n=27)	Vit D and Ca (n=28)
Demographics	Age (years)	Mean (SD)	64.8 (11.9)	65.4 (8.7)
		Min-max	27.8 to 84.5	51.6 to 81.1
	BMI=weight/height2 (kg/cm2)	Mean (SD)	26.4 (4.3)	28.2 (5.2)
		Min-max	19.7 to 35.8	21.4 to 44.5
Vital signs	Systolic blood pressure (mm Hg)	Mean (SD)	133 (9.7)	130 (15.6)
		Min-max	107 to 153	100 to 160
	Diastolic blood pressure (mm Hg)	Mean (SD)	77 (7.1)	76 (8.2)
		Min-max	65 to 94	60 to 98
	Pulse rate (beats/min)	Mean (SD)	69 (12)	69 (9)
		Min-max	44 to 96	48 to 88
BMD	L1–L4 (g/cm2)	Mean (SD)	1.15 (0.21)	1.15 (0.17)
		Min-max	0.88 to 1.65	0.91 to 1.54
	Total hip (g/cm2)	Mean (SD)	0.92 (0.12)	0.93 (0.11)
		Min-max	0.72 to 1.17	0.72 to 1.12
	Femoral neck (g/cm2)	Mean (SD)	0.86 (0.09)	0.86 (0.10)
		Min-max	0.70 to 1.02	0.71 to 1.04
BMD T-score	L1–L4	Mean (SD)	−0.4 (1.6)	−0.3 (1.4)
		Min-max	−2.3 to 3.9	−2.3 to 3.0
	Total hips	Mean (SD)	−0.9 (0.8)	−0.7 (0.9)
		Min-max	−2.3 to 0.3	−2.3 to 0.2
	Femoral neck	Mean (SD)	−1.2 (0.7)	−1.1 (0.8)
		Min-max	−2.3 to −0.1	−2.3 to 0.5
Disease status at baseline	Cardiac disorders	N	0	2
	Endocrine	N	2	10
	Eye	N	4	2
	Gastrointestinal	N	4	7
	Immune system	N	9	6
	Infection and infestations	N	2	1
	Metabolic and nutrition	N	5	7
	Musculoskeletal	N	4	3
	Nervous system	N	4	2
	Psychiatric	N	4	2
	Renal and urinary tract	N	1	2
	Reproduction system	N	1	5
	Respiratory	N	1	1
	Skin	N	1	1
	Surgical and medical proc	N	12	7
	Vascular	N	12	17

BMI, body mass index.

Participant flow and follow-up

In the J-group, one participant discontinued at week 24. In the C-group, one participant discontinued at week 12, during the supplement period, and two discontinued at week 24 after Jarlsberg had been introduced (figure 1). Both intention-to-treat (ITT) and per protocol (PP) analysis were performed. The two analyses gave similar results. The results from the PP-analysis are presented in this manuscript. For other outcomes, participants with unavailable follow-up measurements were excluded from the relevant analyses, resulting in outcome-specific sample sizes.

Study design

The study was performed as a randomised semi-cross-over designed multicentre trial using delayed start-up^{18 19} (figure 1). All participants took 40 µg vitamin D and 500 mg calcium supplementation daily throughout the study period. Half were assigned to the J-group and consumed Jarlsberg from baseline for 32 weeks. After 16 weeks, the C-group added Jarlsberg for the subsequent 32 weeks.

Randomisation

The participants were assigned 1:1 to the J- or C-groups using block randomisation with random block size between 2 and 6 (14). Sequential allocation and enrolment were performed by the responsible investigator using a pre-generated list prepared by the study data manager.

Due to the nature of the dietary intervention, participants and treating clinicians were not blinded. Outcome laboratory analyses were performed by external laboratories according to their standard procedures.

Clinical intervention

The participants received the optimal efficacy dose estimated in the dose-finding study.¹⁵ Females in the J-group received 47 g (three slices) and males 67 g (4.5 slices) of Jarlsberg daily. After 16 weeks, the C-group started the same Jarlsberg regimen for 32 weeks. Jarlsberg was supplied by TINE SA; per 100 g, it provides 351 kcal of which 27 g fat, 27 g protein and contains 80.4 µg/100 g vitamin K₂; mainly MK-4, MK-8, MK-9 and MK-9 (4H).

The Jarlsberg curd is fermented by Lactococcus bacteria producing MK-7/8/9 and PF converting lactic acid into propionic acid. DHNA and MK-9 (4H) are produced as byproducts. Mature Jarlsberg contains >10⁸ CFU/g even after storage.

Clinical procedure

Participants were instructed to maintain their usual diet while avoiding other cheeses after introduction of Jarlsberg cheese. Cheese intake was recorded. Compliance and resupply of vitamin D/calcium tablets and Jarlsberg were checked every 4 weeks. Dietary intake was registered using an app based on the Norwegian Food Composition Table and monitored by a nutrition physiologist.

Clinical status and blood sampling were performed initially and every 16 weeks. Blood samples were drawn in the morning out of consideration for CTX. The participants were instructed to attend fasting when feasible, and sampling time was kept consistent within participants across visits to reduce circadian variability. BTM included tOC, cOC, ucOC, P1NP and CTX, with additional measures including thyroid stimulating hormone (TSH) and parathyroid hormone (PTH). DXA for measurement of L1-L4 (LS-BMD), total hip (TH-BMD) and femoral neck (FN-BMD) was performed every 16 weeks.

Variables

tOC, P1NP/CTX ratio and the BMD-variables were specified as primary outcomes. Secondary outcomes were cOC, ucOC, the ratio R_O=cOC/ucOC, CTX, P1NP, PTH, TSH and biochemical variables.

Measurement procedures

The vitamin K content of Jarlsberg cheese was determined by TINE SA.¹⁵ The cOC and the ucOC were measured in plasma using immunoassay kits (Takara Bio Ōtsu, Japan) by Vitas AS laboratories in Oslo.

P1NP, CTX and PTH were analysed at the Hormone Laboratory at Oslo University Hospital, Aker, Norway. The DXA scans were performed by Professor Erik Fink Eriksen's staff at the Osteoporosis Clinic, Oslo, Norway. The common blood samples and TSH were analysed by Fürst Medical Laboratories, Oslo, Norway.

Statistical analysis

With a significance level of α=0.05, 90% power and clinically relevant difference between groups of 1×SD, at least ≥23 participants per group should be included. 30 per group were planned to allow for dropouts.

The analysis was divided into three parts (figure 1). First is comparing the J-group with the C-group for 16 weeks. Second is to analyse the changes in the C-group after introducing Jarlsberg cheese at 16 and 32 weeks of Jarlsberg intervention. In case the results from the second analysis part confirm the results obtained for the J-group in the first main analysis, the pooled groups could be used for estimation of the 32-week Jarlsberg effect on BTM and BMD.

Continuous variables are reported as mean (SD) and 95% CI.²⁰ Assumption of variable distributions was investigated by applying the Shapiro–Wilk test. Skewed variables were log-transformed for analysis and the results retransformed for presentation. Within-group changes were analysed using analysis of variance with repeated measurements, and changes between groups adjusted for sex and baseline values by using analysis of covariance.^{20 21} The analysis was performed both by PP and ITT. Both analyses gave the same results, and the PP analysis is presented in this paper.

Statistical analysis system (SAS V.9.4) was used in the data analysis.

Approvals

The study was approved by the Norwegian South-East Regional Ethical Committee (Reference no. REK-537058), the Norwegian Centre for Research Data (SIKT-710434) and registered with EudraCT number: 2022–0 03 252-13 and ClinicalTrials.gov number NCT 05787808.

Results

Comparative phase

OC

In the J-group, tOC rose from 20.8 ng/mL (95% CI 17.2 to 24.5) at baseline to 23.1 ng/mL (95% CI 19.7 to 26.5) and 23.5 ng/mL (95% CI 19.5 to 27.4) after 16 and 32 weeks, respectively (p<0.01) (figure 2). The increase was greater than in the C-group (p=0.04) where tOC remained unchanged during the first 16 weeks. Similarly, cOC and R_O increased slightly during the first 16 weeks and significantly after 32 weeks in the J-group, while both were slightly reduced in the C-group (table 2). ucOC was non-significantly reduced from baseline to week 16 and week 32 in the J-group. No change was found in the C-group during the first 16 weeks, and no significant difference was detected between the groups in this period.

Figure 2. Development in total osteocalcin and P1NP/CTX ratio.

Table 2. Comparison of vitamin D and calcium tablets ± Jarlsberg cheese in development of osteocalcin (OC), bone turnover markers (BTM), hormones, lipids, biochemical variables, energy intake and Jarlsberg compliance during 16 weeks of daily cheese intake. The last column shows the status after 32 weeks with Jarlsberg cheese and vitamin D and Ca. The results are expressed by mean values with SD in brackets and 95% CIs.

Variables	Baseline		Week 16		Week 16 vs baseline		P values between groups	Week 32
	Jarlsberg (n=27)	Vit D and Ca (n=28)	Jarlsberg (n=27)	Vit D and Ca (n=28)	Jarlsberg+Vit D and Ca (n=27)	Vit D and Ca (n=28)		Jarlsberg (n=26)*
Carboxylated OC (ng/mL)	12.7 (7.5)	11.0 (3.9)	13.8 (6.1)	10.8 (3.2)	1.1 (4.2)	−0.2 (3.9)	p=0.04	17.2 (7.3)
	9.7 to 15.7	9.5 to 12.5	11.4 to 16.2	9.6 to 12.1	−0.5 to 2.7	−1.7 to 1.3		14.3–20.2
Under carboxylated OC (ng/mL)	8.1 (5.1)	6.2 (3.5)	7.5 (5.0)	6.5 (3.7)	−0.6 (2.6)	0.2 (2.6)	p=0.60	6.2 (4.3)
	6.1 to 10.2	4.9 to 7.6	5.6 to 9.5	5.0 to 7.9	−1.6 to 0.4	−0.8 to 1.2		4.5–8.0
Ratio: carboxylated/under carboxylated	1.91 (1.45)	2.30 (1.58)	1.94 (1.00)	2.20 (1.32)	0.03 (0.89)	−0.10 (1.15)	p=0.17	3.36 (2.16)
	1.30 to 2.53	1.68 to 2.93	1.52 to 2.36	1.68 to 2.72	−0.35 to 0.40	−0.56 to 0.36		2.42–4.26
P1NP: procollagen 1 N-term, propeptide (ng/mL)	57.0 (22.5)	59.4 (21.8)	56.3 (22.3)	57.6 (27.4)	−0.7 (13.4)	−1.71 (19.2)	p=0.89	51.6 (20.8)
	48.1–65.9	50.9–67.8	47.5 to 65.1	47.0 to 68.3	−6.0 to 4.6	−9.2 to 5.7		43.2 to 60.0
CTX-1: C-telopeptide 1 collagen (ng/mL)	0.40 (0.20)	0.35 (0.16)	0.31 (0.15)	0.34 (0.15)	−0.09 (0.13)	−0.02 (0.10	p=0.06	0.35 (0.18)
	0.32 to 0.48	0.29 to 0.42	0.25 to 0.37	0.28 to 0.40	−0.14 to −0.04	−0.05 to 0.02		0.27 to 0.42
Parathyroid hormone (pmol/L)	6.0 (3.0)	6.0 (2.8)	6.6 (3.1)	4.9 (2.7)	0.6 (2.1)	−1.2 (1.9)	p<0.01	5.8 (2.8)
	4.8 to 7.2	5.0 to 7.1	5.4 to 7.8	3.8 to 5.9	−0.2 to 1.4	−1.9 to −0.4		4.6 to 6.9
Thyroid-stimulating hormone (mU/L)	2.15 (1.23)	1.87 (1.11)						1.99 (0.95)
	1.66 to 2.63	1.44 to 2.30						1.59 to 2.39
HbA1c (mmol/mol)†	40.5 (3.1)	42.6 (11.1)	38.1 (2.7)	41.2 (12.3)	−2.4 (1.9)	−1.4 (7.6)	p=0.41	38.5 (3.6)
	37.7 to 43.3	38.3 to 46.9	37.0 to 39.2	36.5 to 46.0	−3.2 to −1.6	−4.3 to 1.6		37.1 to 40.0
LDL/HDL cholesterol ratio	2.52 (1.19)	2.14 (0.66)	2.46 (1.01)	2.09 (0.67)	−0.06 (0.58)	−0.05 (0.32)	p=0.45	2.18 (0.95)
	2.04 to 2.99	1.88 to 2.39	2.01 to 2.85	1.83 to 2.35	−0.29 to 0.17	−0.17 to 0.08		1.80 to 2.57
Creatinine (μmol/L)	73.4 (11.9)	72.4 (17.2)	72.9 (11.8)	74.4 (17.3)	−0.5 (5.8)	1.9 (6.2)	p=0.21	72.1 (12.7)
	68.7 to 78.1	65.8 to 79.1	68.3 to 77.6	67.7 to 81.1	−2.8 to 1.8	−0.5 to 4.3		67.0 to 77.2
Calcium (Ca+) (mmol/L)	2.24 (0.08)	2.28 (0.09)	2.29 (0.08)	2.31 (0.10)	0.04 (0.09)	0.03 (0.10)	p=0.53	2.28 (0.11)
	2.21 to 2.28	2.25 to 2.31	2.26 to 2.32	2.28 to 2.35	0.01 to 0.08	0.00 to 0.07		2.24 to 2.33
Magnesium (Mg+) (mmol/L)	0.84 (0.07)	0.83 (0.08)	0.83 (0.06)	0.81 (0.06)	0.00 (0.05)	−0.01 (0.07)	p=0.27	0.82 (0.06)
	0.81 to 0.87	0.80 to 0.86	0.81 to 0.86	0.79 to 0.84	−0.02 to 0.02	−0.04 to 0.02		0.80 to 0.85
Phosphate (mmol/L)	1.12 (0.18)	1.17 (0.16)	1.13 (0.19)	1.27 (0.18)	0.01 (0.18)	0.10 (0.18)	p=0.02	1.13 (0.17)
	1.05 to 1.19	1.11 to 1.23	1.06 to 1.21	1.20 to 1.34	−0.06 to 0.08	0.03 to 0.17		1.06 to 1.20
Food intake/day total energy (kcal)‡	1571 (370)	1611 (453)	1615 (340)	1585 (507)	45 (248)	−25 (318)	p=0.73	1595 (327)
	1397 to 1744	1399 to 1822	1456 to 1774	1348 to 1823	−72 to 161	−174 to 123		1442 to 1748
Jarlsberg compliance (%)			99.0 (26.4)					99.1 (22.4)
			98.1 to 99.9					98.3 to 99.9

^*One patient in the Jarlsberg group dropped out after 24 weeks

^†Data missing for two patient’s weeks 16 in the J-group.

^‡Food intake is recorded by n=20 patients in each group.

HDL, high density lipoprotein; LDL, low density lipoprotein.

P1np/CTX

In the J-group, the ratio rose significantly from 161.5 (95% CI 135.1 to 187.9) at baseline to 197.3 (95% CI 166.5 to 228.1) at 16 weeks and 211.5 (95%CI: 177.2 to 245.9) at 32 weeks (figure 2) (p<0.01). The C-group exhibited no change in P1NP/CTX-ratio from 185.7 (95% CI 157.4 to 206.3) to 176.9 (95% CI 153.8 to 200.0) over 16 weeks. The J-group exhibited a significantly greater increase in P1NP/CTX-ratio than the C-group during the first 16 weeks (p=0.03). P1NP decreased slightly in the J-group and more notably in the C-group during the first 16 weeks with no group difference, whereas CTX dropped (p=0.02) in the J-group over 16 weeks but remained stable in the C-group (table 2).

BMD

LS-BMD remained stable in the J-group the first 16 weeks, with a mean value of 1.148 g/cm² (95% CI 1.069 to 1.227) but increased to 1.154 g/cm² (95% CI 1.076 to 1.320) (p=0.05) after 32 weeks (figure 3). In the C-group, LS-BMD decreased from 1.151 g/cm² (95% CI 1.084 to 1.218) at baseline to 1.145 g/cm² (95% CI 1.078 to 1.212) at week 16. The difference between the groups was not statistically significant (p=0.35). TH-BMD in the J-group increased slightly from 0.921 g/cm² (95% CI 0.874 to 0.968) at baseline to 0.924 g/cm² (0.876 to 0.972) at 16 weeks and remained unchanged at 32 weeks (figure 3). The C-group displayed similar during the first 16 weeks, rising from 0.925 g/cm² (0.881 to 0.969) to 0.927 g/cm² (0.883 to 0.971). FN-BMD in the J-group rose from 0.861 g/cm² (95% CI 0.822 to 0.898) at baseline to 0.865 g/cm² (95% CI 0.825 to 0.905) at week 16 and 0.866 g/cm² (95% CI 0.825 to 0.905) at week 32 (p=0.04) (figure 3). In contrast, FN-BMD in the C-group decreased from 0.856 g/cm² (95% CI 0.817 to 0.895) to 0.846 g/cm² (95% CI 0.809 to 0.884) at week 16, without significant group differences (p=0.26).

Figure 3. Progression of Bone mineral density.

Hormones

PTH stayed stable in the J-group but dropped significantly in the C-group over 16 weeks, with a significant difference between groups (table 2).

Biochemical variables

HbA1c and low density lipoprotein/ high density lipoprotein (LDL/HDL) ratio decreased in both groups during the first 16 weeks, but only the J-group showed a significant reduction (p<0.01). No significant group differences were detected (table 2). Phosphate increased in the C-group but not in the J-group (p=0.03). The other biochemical variables remained unchanged.

Food intake and Jarlsberg compliance

Jarlsberg compliance was consistently very high, and the food intake stayed stable in both groups during the first 16 weeks, with no significant differences (table 2).

Delayed jarlsberg startup phase

OC

After Jarlsberg was introduced to the C-group, tOC increased (p<0.01) from 17.3 ng/mg (95% CI 15.4 to 19.2) to 22.1 ng/mL (95% CI 19.3 to 24.9) and 23.5 ng/mL (95% CI 20.5 to 26.6) after 16 and 32 weeks, respectively (figure 2). Similarly, cOC and R_O rise occurred (p<0.01), while ucOC decreased slightly (table 3).

Table 3. Development in osteocalcin (OC), bone turnover markers (BTM), hormones, lipids, biochemical variables, energy intake and Jarlsberg compliance in the control group after introducing Jarlsberg® cheese. Timepoints represent 16 and 32 weeks of Jarlsberg exposure in the delayed-start C-group (corresponding to weeks 32 and 48 overall). The results are expressed by mean values with SD in brackets and 95% CIs.

Variables	Baseline BL (n=28)	16 weeks (n=27)*	32 weeks (n=25)†	Week 32: BL (n=25)	P values
Carboxylated OC (ng/mL)	10.8 (3.2)	16.2 (5.5)	17.4 (5.3)	6.8 (4.8)	p<0.01
	9.6 to 12.1	14.0 to 18.4	15.2 to 19.6	4.9 to 8.8
Under carboxylated OC (ng/mL)	6.5 (3.7)	5.9 (3.0)	6.2 (3.6)	−0.1 (2.2)	p=0.86
	5.0 to 7.9	4.7 to 7.1	4.7 to 7.6	−1.0 to 0.8
Ratio: carboxylated/under carboxylated	2.20 (1.32)	3.28 (1.74)	3.36 (2.26)	1.37 (1.59)	p<0.01
	1.68 to 2.72	2.57 to 3.98	2.67 to 4.59	0.70 to 2.04
P1NP: procollagen 1 N-term, propeptide (ng/mL)	57.6 (27.4)	50.9 (21.6)	49.0 (18.0)	−6.4 (17.2)	p=0.07
	47.0 to 68.3	42.4 to 59.5	41.6 to 56.4	−13.5 to 0.7
CTX-1: C-telopeptide 1 collagen (ng/ml)	0.34 (0.15)	0.30 (0.16)	0.30 (0.15)	−0.04 (0.11)	p=0.04
	0.28 to 0.40	0.24 to 0.37	0.24 to 0.36	−0.07 to −0.01
Parathyroid hormone (pmoL/L)	4.9 (2.7)	4.9 (2.2)	5.2 (3.0)	0.0 (2.1)	p=0.95
	3.8 to 5.9	4.1 to 5.8	4.0 to 6.4	−0.8 to 0.9
Thyroid-stimulating hormone (mμU/L)	1.87 (1.11)		1.69 (0.96)	−0.24 (0.88)	p=0.19
	1.44 to 2.30		1.29 to 2.09	−0.60 to 0.13
HbA1C (mmoL/moL)	41.2 (12.3)	41.7 (7.7)	41.7 (7.5)	0.3 (1.9)	p=0.47
	36.5 to 46.0	38.6 to 44.7	38.6 to 44.8	−0.5 to 1.1
LDL/HDL cholesterol ratio	2.09 (0.67)	2.04 (0.62)	2.11 (0.69)	−0.01 (0.49)	p=0.96
	1.83 to 2.35	1.79 to 2.28	1.83 to 2.40	−0.21 to 0.20
Creatinine (μmoL/L)	74.4 (17.3)	71.6 (16.2)	71.6 (16.5)	−0.1 (6.8)	p=0.93
	67.7 to 81.1	65.2 to 78.0	64.6 to 78.6	−3.0 to 2.7
Calcium (Ca+) (mmoL/L)	2.31 (0.10)	2.35 (0.10)	2.37 (0.11)	0.09 (0.10)	p<0.01
	2.28 to 2.35	2.31 to 2.39	2.33 to 2.42	0.05 to 0.13
Magnesium (Mg+) (mmoL/L)	0.81 (0.06)	0.83 (0.06)	0.83 (0.05)	0.01 (0.06)	p=0.38
	0.79 to 0.84	0.81 to 0.86	0.81 to 0.86	−0.01 to 0.03
Phosphate (mmoL/L)	1.27 (0.18)	1.19 (0.14)	1.12 (0.16)	−0.04 (0.16)	p=0.17
	1.20 to 1.34	1.13 to 1.24	1.06 to 1.19	0.11 to 0.02
Food intake/day total energy (kcal)‡	1585 (507)	1712 (591)	1866 (1192)	281 (999)	p=0.22
	1348 to 1823	1436 to 1989	1292 to 2441	−200 to 762
Jarlsberg compliance (%)		95.3 (21.3)	95.7 (20.9)

^*One patient from the control group dropped out of the study before 12 weeks after addition of Jarlsberg cheese.

^†Additionally, two patients from the control group dropped out of the study before 32 weeks after addition of Jarlsberg cheese.

^‡Food intake was recorded by n=20 patients

P1NP/CTX

The ratio in the C-group increased from 176.9 to 193.9 (95% CI 162.2 to 225.6) and 195.3 (95% CI 170.9 to 219.7) after 16 and 32 (figure 2). The reduction in P1NP persisted non-significantly after adding Jarlsberg and CTX decreased significantly (p=0.04) over the following 32 weeks (table 3).

BMD

LS-BMD increased (p=0.02) from 1.145 g/cm² (95% CI 1.078 to 1.212) to 1.169 g/cm² (95% CI 1.097 to 1.242) and 1.172 g/cm² (95% CI 1.101 to 1.242) after 16 and 32 weeks, respectively (p=0.02) (figure 3). TH-BMD increased significantly to 0.938 g/cm² (0.891 to 0.985) after 16 weeks and further to 0.942 g/cm² (0.898 to 0.986) at week 32 (p=0.03) (figure 3). FN-BMD increased from 0.846 to 0.856 g/cm² (95% CI 0.814 to 0.898) at 16 weeks (p=0.02), but it dropped to 0.851 g/cm² (95% CI 0.811 to 0.891) by week 32 (figure 3).

Pooled data

OC

tOC rose from 18.8 ng/mL (95% CI 16.7 to 21.0) at baseline to 21.4 ng/mL (95% CI 19.1 to 23.6) at 16 weeks and 23.5 ng/mL (95% CI 21.1 to 25.9) at 32 weeks (figure 4), with a significant increase of 4.7 ng/mL (95% CI 3.1 to 6.3) (p<0.01). cOC and Rₒ showed similar trends, while ucOC decreased (table 4).

Figure 4. Estimated effect of 32 weeks of daily Jarlsberg intake.

Table 4. Pooled total material of patients treated with Jarlsberg cheese and vitamin D and calcium tablets for 32 weeks. The development in osteocalcin (OC), bone turnover markers (BTM), bone mineral density (BMD), BMD T-score, hormones, lipids, biochemical variables, energy intake and Jarlsberg compliance are expressed by mean values with SD in brackets and 95% CIs.

Variables	Baseline (BL) (n=51)	16 weeks (n=51)	32 weeks (n=51)	Week 32: BL (n=51)	P values
Carboxylated OC (ng/mL)	11.8 (5.9)	14.9 (5.9)	17.3 (6.4)	5.5 (5.4)	p<0.001
	10.1 to 13.5	13.2 to 16.6	15.5 to 19.1	4.0 to 7.0
Under carboxylated osteocalcin (ng/mL)	7.0 (4.4)	6.5 (4.0)	6.2 (3.9)	- 0.8 (2.6)	p=0.02
	5.8 to 8.3	5.3 to 7.6	5.1 to 7.3	−1.6 to −0.1
Ratio: carboxylated/under carboxylated	2.12 (1.40)	2.68 (1.59)	3.50 (2.19)	1.37 (1.51)	p<0.001
	1.71 to 2.53	2.21 to 3.15	2.85 to 4.14	0.93 to 1.81
procollagen 1 N-term, propeptide (ng/mL)	56.2 (24.8)	53.4 (22.2)	50.3 (19.4)	−5.9 (16.6)	p=0.014
	49.2 to 63.2	47.2 to 59.6	44.9 to 55.8	−10.6 to −1.2
CTX-1: C-telopeptide 1 collagen (ng/mL)	0.36 (0.18)	0.30 (0.14)	0.32 (0.17)	−0.04 (0.14)	p=0.05
	0.32 to 0.41	0.26 to 0.34	0.28 to 0.37	−0.08 to 0.00
Bone mineral density T-score (L1–L4)*	−0.331 (1.52)	−0.288 (1.46)	−0.310 (1.46)	0.02 (0.30)	p=0.63
	−0.773 to 0.110	−0.712 to 0.137	−0.734 to 0.113	−0.066 to 0.107
BMD T-score (total hips)*	−0.783 (0.85)	−0.766 (0.84)	−0.736 (0.85)	0.047 (0.10)	p<0.01
	−1.03 to −0.535	−1.01 to −0.518	−0.985 to −0.488	0.018 to 0.076
BMD T-score (femur neck)*	−1.213 (0.71)	−1.166 (0.72)	−1.168 (0.70)	0.045 (0.16)	p=0.05
	−1.423 to −1.003	−1.377 to −0.955	−1.374 to −0.962	−0.000 to 0.096
Parathyroid hormone (pmoL/L)	5.5 (2.9)	5.9 (2.7)	5.5 (2.9)	0.0 (2.2)	p=0.89
	4.7 to 6.3	5.1 to 6.6	4.7 to 6.3	0.7 to 0.6
Thyroid-stimulating hormone (mμU/L)	2.05 (1.16)		1.84 (0.96)	−0.21 (0.88)	p=0.11
	1.72 to 2.38		1.56 to 2.11	−0.46 to 0.05
HbA1c (mmoL/moL)	40.8 (9.2)	39.9 (6.0)	40.1 (6.0)	−0.7 (5.3)	p=0.13
	38.3 to 43.3	38.3 to 41.6	38.4 to 41.8	−2.3 to 0.9
LDL/HDL ratio	2.30 (0.98)	2.25 (0.85)	2.15 (0.83)	−0.14 (0.71)	p=0.15
	2.03 to 2.56	2.01 to 2.48	1.92 to 2.38	−0.34 to 0.06
Creatinine (μmoL/L)	73.5 (14.8)	72.3 (14.0)	71.8 (14.5)	−1.7 (7.1)	p=0.09
	69.8 to 77.8	68.4 to 76.1	67.7 to 76.0	−3.8 to 0.3
S-calcium (mmoL/L)	2.28 (0.10)	2.32 (0.10)	2.33 (0.12)	0.05 (0.11)	p<0.01
	2.25 to 2.31	2.29 to 2.34	2.29 to2.36	0.02 to 0.08
S-magnesium (mmoL/L)	0.83 (0.07)	0.83 (0.06)	0.83 (0.06)	0.00 (0.05)	p=0.58
	0.81 to 0.84	0.82 to 0.85	0.81 to 0.84	−0.01 to 0.02
S-phosphate (mmoL/L)	1.20 (0.19)	1.16 (0.17)	1.13 (0.16)	−0.06 (0.20)	p=0.03
	1.14 to 1.25	1.12 to 1.21	1.08 to 1.17	−0.12 to −0.01
Energy intake/day (kcal)†	1532 (439)	1664 (479)	1727 (864)	195 (748)	p=0.11
	1379 to 1685	1511 to 1817	1447 to 2007	−48 to 437
Jarlsberg compliance (%)		97.1 (24.0)	97.5 (21.8)

^*Three patients missed DXA, n=48 patients included in the analysis

^†Food intake was recorded by n=40 patients

P1NP/CTX

The ratio rose from 168.8 (95% CI 150.8 to 186.7) at baseline to 197.8 (95% CI 175.9 to 219.6) at 16 weeks and 203.9 (95% CI 189.4 to 218.4) at 32 weeks (figure 4). A significant (p=0.02) increase of 35.1 (95% CI 17.9 to 52.3).

BMD

LS-BMD increased from 1.151 g/cm² (95% CI 1.096 to 1.205) at baseline to 1.155 g/cm² (95% CI 1.103 to 1.208) after 16 weeks and 1.163 g/cm² (95% CI 1.110 to 1.216) after 32 weeks (figure 4). This was an increase of 0.012 g/cm² (95% CI 0.002 to 0.022) (p=0.04), 1.0%. TH-BMD increased from 0.929 g/cm² (95% CI 0.896 to 0.962) at baseline to 0.930 g/cm² (95% CI 0.897 to 0.964) at 16 weeks and 0.933 g/cm² (95% CI 0.900 to 0.966) at 32 weeks (figure 4), with an estimated increase of 0.004 g/cm² (95% CI 0.001 to 0.007) (p=0.015), 0.4%.

FN-BMD was unchanged. Values increased from 0.855 g/cm² (95% CI 0.826 to 0.883) at baseline to 0.860 g/cm² (95% CI 0.831 to 0.889) at 16 weeks and remained at 0.860 g/cm² (95% CI 0.832 to 0.889) after 32 weeks (figure 4). The estimated increase was 0.005 g/cm² (95% CI −0.0005 to 0.011) (p=0.07).

Both TH- and FN-T-scores increased significantly after 32 weeks of Jarlsberg intake. LS-BMD T-score also increased but not significantly (table 4).

Hormones and biochemical variables

PTH was unchanged. TSH, HbA1c, LDL/HDL and serum phosphatase decreased slightly while S-calcium increased (table 4).

Food intake and Jarlsberg compliance

Jarlsberg compliance was consistently very good throughout the study, but the energy intake increased slightly but not significantly (table 4).

Discussion

The promising effects on BTM in premenopausal women and on BTM and BMD in Nordic skiers after Jarlsberg intake prompted us to investigate whether Jarlsberg could influence bone outcomes in a population at increased risk of osteoporotic fracture.^{13 14} The development in the two groups was compared during the first 16 weeks. The second analysis was performed as change within the C-group from inclusion of Jarlsberg cheese to weeks 16 and 32. The comparative part of the analysis shows a significant increase in BTM in favour of the J-group like shown in the dose-finding study.¹⁵

After Jarlsberg was introduced in the C-group, development in tOC, cOC and Rₒ verified the results obtained in the J-group.

In the comparative phase P1NP/CTX-ratio decreased in the C-group and increased in the J-group. When including Jarlsberg to the C-group, the ratio increased as in the J-group, verifying the results from the comparative part. LS-BMD was unchanged in the J-group and decreased in the C-group. After introducing Jarlsberg to the C-group, LS-BMD increased significantly. TH-BMD followed the same pattern as LS-BMD. In the comparative phase FN-BMD increased significantly in the J-group, while it decreased in the C-group. After introducing Jarlsberg, the increase was significant. These findings verify the results obtained in the comparative analysis.

In summary, the results obtained in the comparative part regarding the effect on the OC-variables, BTM and even BMD variables were verified in the within C-group analysis. Even if the materials in the comparative study part are statistically large enough, it is of importance to increase the material to better estimate the Jarlsberg-effect. Following the verification of the comparative effects, it was reliable to pool the two materials for estimation of the Jarlsberg effects.

In the pooled material, the tOC, cOC, R_o and P1NP/CTX-ratio increased significantly during the 32 weeks of Jarlsberg intervention. The observed BMD increase of approximately 1% at the lumbar spine and 0.4% at the total hip were significant and in accordance with the BTM results. The changes are positive, but small, and close to the precision error of DXA depending on site, device and local quality control. It is important to avoid overinterpreting short-term BMD changes and note that the study did not assess fractures or other clinical endpoints.

Carboxylated OC is a protein that assists the osteoblasts to deposit calcium into the bone matrix. P1NP and CTX are widely used BTMs,²² and the P1NP/CTX ratio reflects the balance between bone formation and resorption.^{23 24} In this study, Jarlsberg intake was associated with an increased P1NP/CTX ratio, mainly driven by reductions in CTX, which is compatible with reduced bone resorption. The P1NP/CTX ratio could increase if P1NP increases, combined with unchanged or increased CTX. In fertile women, an increase in P1NP and a decrease in CTX were observed. In the skiers and the elderly, P1NP was kept constant, while there was a significant decrease in CTX.^{13 15 24} These effects are associated with favourable effects on bone metabolism; even if the changes are limited, they are supported by positive BMD changes.

In the pooled material, the increases in the P1NP/CTX ratio and the modest BMD changes observed over 32 weeks are consistent with a potential shift in bone turnover and bone density during Jarlsberg intake in a population where age-related bone loss would be expected.²⁵ However, the study is not designed to detect causal reasons for observed changes attributed to the specific components in Jarlsberg cheese like PF, vitamin K₂ or DHNA and should be interpreted as exploratory.

Most studies on vitamin K₂, DHNA and PF have been performed in animals or in vitro.

Vitamin K₂ has been described to suppress osteoclast genesis in vitro,²⁶ DHNA promotes increased trabecular thickness and BMD in oestrogen-depleted mice.^{9 12} PF inhibits the nuclear factor κB ligand-induced osteoclast differentiation and exhibits an antiresorptive effect, in addition to enhancing osteoblast differentiation and mineralisation by increasing the osteoprotegerin/RANK ratio.^{27 28} The present investigational product contains the described elements as part of a complex food matrix, and the study is not designed to draw any mechanistic conclusions. Across our previous studies, the pattern of findings suggests that the observed effects of Jarlsberg cheese are unlikely to reflect a general ‘cheese effect’. Jarlsberg has been compared with (i) Camembert, which contains neither vitamin K₂ nor PF,²⁹ and (ii) Norvegia cheese, containing K₂ but not PF. Norvegia cheese, containing K₂ but not PF. Jarlsberg contains both vitamin K₂ and PF but the impact of PF-associated metabolites, as well as the potential effects related to the PF surface protein, is unknown³⁰ (in review). While these food-based comparisons do not permit firm mechanistic attribution to individual components, they are consistent with the hypothesis that PF directly and/or via PF-associated metabolites may contribute to the favourable effects observed.

Strengths and limitations

Strengths of this study include the inclusion of both women and men, a pragmatic intervention with high compliance and repeated assessments of BTM and DXA. Additional strengths are the randomised design and the use of both between-group comparisons and within-group comparisons. This approach corroborated the findings observed in the Jarlsberg group during the initial 16 weeks and allowed estimation of intervention effects in a pooled sample.

Key limitations were that blinding was not feasible and that background diet, including intake of other fermented foods, could not be fully controlled. Although randomisation and stable concomitant medication reduced confounding, population heterogeneity may have influenced the results; longer-term studies are needed to assess whether the observed changes in BTM and BMD are sustained.

Conclusion

Jarlsberg cheese intake alongside calcium and vitamin D was associated with favourable increases in OC-related markers, P1NP/CTX ratio and BMD in postmenopausal women and older men. Jarlsberg cheese might be a beneficial component of a targeted dietary approach to support bone health in this population.

Data availability statement

Data are available upon reasonable request.