To study the effect of sea buckthorn berries on the number and duration of common cold (CC) infections. As secondary objectives the effects on digestive and urinary tract infections (DTI, UTI), and serum C-reactive protein (CRP) concentrations were also investigated.
A total of 254 healthy volunteers were randomly assigned to receive sea buckthorn or placebo product during the study, which 233 of them completed.
There were no significant differences in the number or duration of CC or DTI between groups (CC: relative risks (sea buckthorn vs placebo) for the number and duration were 1.15 (95% CI 0.90–1.48) and 1.05 (95% CI 0.87–1.27), respectively). In the sea buckthorn group, as compared to the placebo, the serum CRP concentrations decreased significantly (difference in median change −0.059 mg/l, P=0.039). The number of UTI was too small to draw solid conclusions, but the results indicate the subject merits further investigation.
Sea buckthorn berries did not prevent CC or DTI. However, a reductive effect on CRP, a marker of inflammation, and a risk factor for cardiovascular diseases, was detected.
The common cold (CC) is a mild upper respiratory illness caused by numerous viruses. Although the CC is usually a self-limited illness of short duration, it causes considerable costs in terms of absence from work, visits to health care providers and the amounts of drugs prescribed (Puhakka et al., 2000; Heikkinen and Järvinen, 2003). Sea buckthorn (Hippophaë rhamnoides L.) berries have traditionally been used in Eastern medicine and studies indicate that sea buckthorn has immunomodulatory properties and positive effects on health (Yang et al., 1999; Johansson et al., 2000; Eccleston et al., 2002; Geetha et al., 2002; Yang and Kallio, 2002; Dorhoi et al., 2006). To our knowledge no clinical trials have been conducted to study the efficacy of sea buckthorn in prevention and treatment of the CC. The primary objective of this trial was to determine whether regular consumption of sea buckthorn berries affects the number of CC infections or the duration of symptoms in healthy adults. The effects on digestive tract infections (DTI), urinary tract infections (UTI) and serum concentrations of C-reactive protein (CRP), an important inflammatory and infection marker, were also studied.
Participants and methods
Participants of the study were healthy 19- to 50-year-old men and women, recruited for the study by announcements on the Internet pages and campus of the University of Turku (exclusion criteria: see Supplementary Information at the European Journal of Clinical Nutrition website). Baseline information about the participants was collected using a questionnaire. Participants were informed about the study protocol. They gave written informed consent to the study procedures, which were approved by the ethics committee of the Hospital District of Southwest Finland. The study statistician carried out the randomization procedure in a double-blind manner (Figure 1). Stratified randomization was used, the stratification factors being the age of participant, body mass index, gender, smoking habits and use of vitamin or mineral supplements. The randomization code was developed using a computer random number generator to select random permuted blocks (block length 10).
Study design and outcome measures
The study was carried out in the Functional Foods Forum, an independent research unit of the University of Turku, Finland, as a randomized, double-blind, placebo-controlled trial. During the 90-day observation period, the participants were asked to assess daily their symptoms of (1) CC, (2) DTI, (3) UTI or (4) other infections in a logbook form. In the context of each infection type typical symptoms were listed. Participants were asked to report the severity of symptoms using a four-point scale with possible responses ranging from 0=‘none’ to 3=‘severe’. The upper respiratory infection symptoms listed were fever, runny nose, nasal congestion, sore throat, fatigue, headache, earache, nausea, muscular/joint pain or ‘other’ for which option the participant was asked to specify the symptom. They were asked to report a symptom only if, in their assessment, it was caused by an infection. The daily logbook assessment also included a daily estimation of overall health (‘do you feel healthy’: yes/no). To detect potential differences in the causative viruses between groups, and to confirm the CC diagnoses, the participants were asked to give a nasal swab and serum sample on the first to third day they had symptoms of CC (‘the CC samples’). A second serum sample was taken after 2 weeks. In addition, blood samples were obtained after a 12 h fast, at the beginning and end of the observation period for the analysis of serum CRP concentrations.
During the observation period, the participants daily took 28 g of frozen sea buckthorn puree or placebo product similar in appearance, taste and smell to the active product. The sea buckthorn berries used were of the H. rhamnoides spp. mongolica cv. Prozcharachnaya origin harvested in Ostrobothnia, Finland, in autumn 2004 and stored frozen. For the puree, berries were mashed in a process where also the seeds were ground fine. To facilitate the making of similar placebo product, additives were used (recipes of the sea buckthorn and placebo products, see Supplementary Information). During the manufacturing process, the temperature of the berries increased to about +5°C. The participants were advised to store the product in the freezer and take it frozen or thaw it in such a way it did not get warm. They were allowed to take the product as such or with other foodstuffs. Daily consumption was reported in the logbooks.
To assess the success of blinding, each participant was, at the end of the observation period, asked to guess which product he/she had been given. The participants were also asked if they thought the study product had been beneficial to them. Only after the preliminary statistical analysis was the treatment code opened for the researchers and the statistician.
The participants were allowed to use normal medication during the study. During the observation period and the 1 month wash-out period the participants were asked not to use products containing sea buckthorn. Vitamin and mineral supplements were allowed if no megadoses (defined as 10 times the daily intake recommended by the Finnish National Nutrition Council, 2005) were used. Fish oil supplements were also allowed but the use of all other nutrient supplements was restricted. Lactic acid bacteria-containing pills and certain probiotic product brands (Gefilus, Rela, Yosa), commonly used in Finland, were not allowed. The use of nutrient supplements and medication was also recorded in the logbooks.
Most CC cases are detected in winter. Typically, in Finland the CC season begins with a rhinovirus epidemic in September–October and ends with a second rhinovirus epidemic in April–May. Corona and influenza viruses cause infections typically during the winter (Puhakka et al., 2000). During a period from mid-January to mid-February 2005 about 10–15 participants per day began the 90-day observation period. The study for the participants ended during a period from mid-April to mid-May, respectively.
The main outcome measures of the logbook analyses were the number of CC infections and the duration of symptoms. Also the number and duration of DTI and UTI were analyzed. A participant was considered to have a case of CC if he/she reported at least one symptom at a severity of 1 on the severity scale for 1 day or longer. For the duration of the symptoms, every day the participant had recorded having one or more symptoms was calculated. The CC cases were treated as separate if there were at least 2 consecutive days between them where the participant had not recorded any symptoms. The other infection types were analyzed accordingly. In an effort to explore the sensitivity of the conclusions for the definitions of infections used, we made complementary analyses based on the logbook question ‘do you feel healthy’. The first 7 days of the observation period were excluded from the analyses to take account of the incubation period of the CC (Puhakka et al., 2000; Heikkinen and Järvinen, 2003). Complementary analyses concerning only the last month of the observation period were done in case the prerequisite for effects of sea buckthorn was reaching certain minimum levels in the active compounds of the body.
The statistical analyses of the logbook data were carried out using generalized linear models (SAS software, version 9.1.3 SP2, GENMOD procedure; SAS Institute Inc., NC, USA). The CC is a contagious disease, and the number and duration of it and DTI was assumed to follow negative binomial distribution. In both cases Poisson distribution was first fitted to the data, but the observed overdispersion suggested the use of more flexible negative binomial distribution. Due to the small number of UTI their number and duration was assumed to follow Poisson distribution. In all statistical models, a logarithmic link function was used to convert the estimates of the parameters into relative risks (RR; unadjusted, no covariates were used). The estimated RR for infections or symptoms were calculated always comparing the sea buckthorn group (SBG) to the placebo group (PG), that is RR significantly less than 1 would denote a desirable sea buckthorn effect. CRP concentrations were analyzed using rank analysis of covariance with the baseline measurement as a covariate (same software as above). Due to the abnormal distribution of values, data on the CRP measurements and duration of infections are expressed as median (range). Fisher's exact test was used to compare the number of CC samples taken in treatment groups (same software as above). Two-sided significance tests and significance levels of 0.05 were used throughout.
All randomized subjects who returned logbooks were included in the primary analysis (intention to treat (ITT) data). In addition, analyses including only participants who were compliant with the protocol (per protocol, PP, data) were conducted (criteria for excluding participants from the PP analysis, see Supplementary Information). CRP concentrations >10 mg/l are considered as a marker of acute infection or inflammation, which could obscure the use of CRP concentration as a risk marker of cardiovascular diseases (Pearson et al., 2003). Using both ITT and PP datasets complementary analyses excluding values >10 mg/l were also made.
Prestudy sample size estimation was based on simplifying assumptions made about the number of CC infections and duration of symptoms. It was assumed 50% of the participants in the PG would have one case of CC during the 3-month observation period (Heikkinen and Järvinen, 2003). We considered a reduction of CC incidence to 30% of the participants in the SBG having one infection clinically significant. With a sample size of 120 participants/group the trial would have a power of approximately 85% to detect this difference between treatment groups (two-sided tests, 0.05 significance level, 10% assumed dropout rate) if χ2-test was used. The number of symptom days was assumed to be 5–7 on average in the PG (Barrett et al., 2002). A 2 days reduction in the duration of CC in the SBG was considered clinically significant (Barrett et al., 2002). If the standard deviation was 5 days, 130 persons in both groups would be needed to detect a difference of 2 days in treatment groups at a power of approximately 85% if the Mann–Whitney U-test was used (two-sided tests, 0.05 significance level, 10% dropout).
Serum CRP concentrations were measured by TYKSLAB (Turku, Finland) with high-sensitivity particle-enhanced immunoturbidometric assay using Roche tina-quant reagents (Roche Diagnostics, GmbH, Mannheim, Germany) and a fully automated analyzer Roche Modular P800 (Roche Diagnostics). The flavonolglycosides and vitamin C content of the sea buckthorn puree were analyzed using high-performance liquid chromatograph (HPLC)-UV methods (Kallio et al., 2005; Tiitinen et al., 2005). For the analysis of vitamin E content an HPLC method with fluorescence detection was used (Kallio et al., 2002b). The oil content of the product was analyzed gravimetrically using a chloroform–methanol extraction method (Tiitinen et al., 2005). The fatty acids of the oil were esterified using a boron trifluoride method (Ågren et al., 1992) and the fatty acid methyl esters were analyzed by gas chromatography (Yang et al., 1999). The carotenoids were analyzed by the MTT Agrifood Research Finland (Jokioinen, Finland) chemistry laboratory using an in-house HPLC-UV/VIS method.
Sea buckthorn product
The daily dose of sea buckthorn product contained 16.7 mg of flavonol glycosides, the glycosides of isorhamnetin being the most abundant (detailed composition of the product, see Supplementary Information). Calculated as aglycones the total daily amount of flavonols was about 8.4 mg. The consumption of flavonols has been estimated to be about 20–25 mg/day in the United States, Denmark and Holland and somewhat higher in Italy, the mean value being 35 mg/day (Manach et al., 2004). The variability of intake even in the same country can be considerable, however (Manach et al., 2004). The vitamin C content of the product was 15.6 mg/day, which is about 21% of the recommended daily intake for adults in Finland (Finnish National Nutrition Council, 2005). The vitamin E and carotenoid content of the product was small compared with the recommended intakes (Finnish National Nutrition Council, 2005) and the daily dose contained oil and unsaturated fatty acids less than used in previous studies reporting positive effects of sea buckthorn oil (Yang et al., 1999; Johansson et al., 2000).
Participants randomized in the sea buckthorn and PGs had similar characteristics and there were no major differences in their eating habits and lifestyles (Table 1; more characteristics of the ITT population, see Supplementary Information). In both groups 12 participants reported having asthma or allergy. Based on the information gained from the logbooks the use of medication, vitamins, minerals and nutrient supplements during the observation period did not differ substantially between treatment groups. There were no major differences in the characteristics of the groups in the PP data either (data not shown).
Overall 9% of the randomized participants dropped out of the study and the number of withdrawals was identical between groups (Figure 1). Two participants in the SBG and one in the PG gave the reason for withdrawal as being symptoms caused by the study product (PG: rash; SBG: nausea, ulcers in the mouth). The number of noncompliant participants excluded from PP data was similar in both groups (logbook analyses: SBG 13, PG 14; CRP analyses: SBG 26, PG 27).
At the end of the observation period, 70% of the participants in the SBG guessed their treatment group correctly. In the SBG, 57% of the participants felt the product had positive effects during the observation period (for example, refreshing or energizing). In the PG, 50% of the participants guessed they had been receiving the sea buckthorn product and 47% felt the product had had positive effects.
Number and duration of infections
Sea buckthorn did not have an effect on CC or DTI risk or duration in the primary ITT data (Table 2). Analyses of the PP data yielded the same result (number of infections; CC: RR 1.12, 95% CI 0.87–1.45, DTI: RR 0.99, 95% CI 0.60–1.63; duration of infections; CC: RR 1.05, 95% CI 0.86–1.28, DTI: RR 1.12, 95% CI 0.85–1.46; RR values unadjusted, assumed distributions as described in ‘Statistical analysis’) as did the analyses restricted to the last month of the observation period (data not shown).
The participants had only a few UTI during the observation period (Table 2). Thus, the results can be considered as indicative only. The secondary PP analysis suggested a significant positive effect of sea buckthorn on the number, but not the duration, of infections (number of infections: RR 0.43, 95% CI 0.20–0.94; duration of infections: RR 0.60, 95% CI 0.28–1.29; RR values unadjusted, assumed distributions as described in ‘Statistical analysis’; in PP data 10 UTI cases in 7 participants). Analyses concerning the last month of the study yielded significant results for both the number and duration of UTI (data not shown).
Number of common cold samples
The number of CC samples confirms the conclusion of the groups not differing in both ITT and PP data (in ITT data the number of participants giving 1 sample: 37 in SBG, 40 in PG; 2 samples: 8 in SBG, 4 in PG; 3 samples: 2 in SBG, 0 in PG; P=0.37; PP data not shown). Thus the samples were not further analyzed.
Subjective sensation of health
Reports of participants’ subjective sensation of health (logbook question ‘do you feel healthy’) varied greatly. Some reported feeling healthy even if they had symptoms of an infection while others always reported feeling sick if they had any symptoms. The median number of sick days during the observation period in ITT data was 4 (range 0–45 days) and 6 days (range 0–37 days) in the sea buckthorn and PGs, respectively. The groups did not differ significantly (RR 1.06, 95% CI 0.76–1.49). PP analysis and analysis using only the last month of the observation period led to the same conclusion (PP: RR 1.02, 95% CI 0.72–1.46).
Compared to the placebo, there was a small but significant reduction in serum concentrations of CRP in the SBG during the observation period (Table 3). This reduction was evident in data containing all measurements and when highly elevated values (>10 mg/l) were excluded.
Due to its high acid content the sea buckthorn berry taste does not always attract consumers (Tiitinen et al., 2006). We chose a moderate daily amount of the product to have a realistic dose for the normal consumer's everyday life. As Finns consume on average only about 10 g fresh berries daily (Männistö et al., 2003) this was still a substantial addition to the berry intake.
It was expected that the compounds may have synergetic effects (Middleton et al., 2000; Puupponen-Pimiä et al., 2001) as the study product was the whole berry. The amounts of bioactive compounds in sea buckthorn berry vary depending on the subspecies, area and year of cultivation, and the maturity of the berries (Gao et al., 2000; Kallio et al., 2002a). It is possible that loss of bioactive compounds took place during the manufacture and storage of the product (de Ancos et al., 2000; Häkkinen et al., 2000). The amounts of vitamins C and E especially were far from the highest amounts detected in sea buckthorn berries (Yang and Kallio, 2002; Tiitinen et al., 2005).
When compared to common daily intakes and recommendations, the intake of flavonols was, by estimation, affected most by the sea buckthorn puree. Flavonols along with other flavonoids and polyphenolic compounds influence the immune system and inflammatory cells (Middleton et al., 2000) and have antimicrobial properties (Conti et al., 1998; Middleton et al., 2000; Puupponen-Pimiä et al., 2001). The most convincing evidence on the efficacy of vitamins in CC treatment or prevention concerns people whose immune system may not function optimally or who are under exceptional stress (Hemilä et al., 2002, 2004; Meydani et al., 2004). The participants in the current study were healthy adults who exercised moderately and very few smoked. On the grounds of the eating habit information, it is likely that they obtained reasonable intakes of flavonoids and vitamins in their normal diet.
Berry phenolics possess antimicrobial activity against bacteria causing DTI and UTI (Puupponen-Pimiä et al., 2001, 2005) and cranberry and other berry juices rich in phenolic compounds have been associated with a lowered risk of UTI (Kontiokari et al., 2001, 2003). In this study, the number of UTI and the median duration of infections were smaller in the sea buckthorn than in the PG. The complementary statistical analyses even suggested significant efficacy of sea buckthorn. However, the number of UTI was too small for reliable conclusions.
The definitions of infections were based on self-assessment of symptoms by the participants, a method frequently used in CC studies (Brinkworth and Buckley, 2003; Langkamp-Henken et al., 2004; Turner et al., 2004). The clinical diagnosis of CC is in most cases simple and can reliably be made by adult patients themselves (Heikkinen and Järvinen, 2003). Though allergic and vasomotor rhinitis can mimic the CC, these conditions usually can be easily differentiated (Heikkinen and Järvinen, 2003) and in this study, the number of participants having allergies was similar in both groups. Even with modern diagnostic methods, about 20–30% of cold cases remain without a proven viral cause (Heikkinen and Järvinen, 2003) which makes the CC syndrome in large part subjective (Gwaltney et al., 1996). However, relying only on self-assessments as the primary outcomes in our trial is a limitation. Randomization was used to distribute the bias and noise potentially introduced by self-reporting equally between groups, but they still interfere with the detection of true signal (Gwaltney et al., 1996).
The number of CC samples taken was smaller than the number of infections reported in logbooks. That is likely because samples could only be taken during office hours, which made it difficult for participants. Some participants also told they were too sick to leave home to get the samples taken. Another limitation of the study was the short duration. In the case of a natural product it is possible the effects need more time to develop.
On the assessment of blinding, more than 50% of the participants in the SBG guessed they had been receiving the active product during the study, raising the question of a possible placebo effect. However, the logbook analyses showed no significant beneficial effects of sea buckthorn. Despite the lack of positive results from the logbook analyses, more participants in the sea buckthorn than in the PG reported having benefited from the product when asked about it at the end of the study. Participants can only be expected to guess 50% right in both groups at the end of the study if the active product has no treatment effect and neither of the products have side effects in addition to the products being identical in physical characteristics (Desbiens, 2002). Although universally stated important, there is controversy about the ideal and sufficient way to find out if blinding has been achieved and maintained in a double-blind trial (Desbiens, 2002; Caruso and Gwaltney, 2005; Rees et al., 2005).
The detected serum CRP reductive effect of sea buckthorn berries is consistent with recent studies showing an inverse association between diets rich in fruit, and circulating concentrations of inflammation markers, including CRP (Nettleton et al., 2006; Wannamethee et al., 2006). Flavonoids and vitamin C are likely to be among the bioactive compounds responsible for the anti-inflammatory effect of fruit, possibly through synergetic mechanisms (Middleton et al., 2000; Nijveldt et al., 2001; Rein et al., 2006; Wannamethee et al., 2006). CRP even at serum concentrations <10 mg/l, not associated with an acute infection or inflammation, is considered not only a risk marker for cardiovascular diseases, but also a proatherogenic and proinflammatory agent contributing to the pathogenesis (Pearson et al., 2003; Pepys and Hirschfield, 2003; van Oostrom et al., 2004).
In conclusion, consumption of sea buckthorn berries reduced serum concentrations of CRP. Although the change was small the results encourage further investigation concerning the effects of sea buckthorn on inflammation and cardiovascular disease risk factors. The number of UTI was too small for solid conclusions, but indicative results suggest the subject merits further research. Future studies should also evaluate the effects of sea buckthorn berries on infections in trials with a longer study period and different dosage, and in populations other than healthy adults.
We thank the volunteers who participated in the study. We also thank Terhi Pohjanheimo and Katja Tiitinen for development of the treatment products; Nina Kainulainen, Raija Nurmi, and Anja Pirinen for excellent assistance; Jukka-Pekka Suomela for the flavonol glycoside analysis of the study product; Saija Hurme for the statistical analyses of CRP data; and Marjo Mäkinen-Aakula and Ilkka Liukas for their contribution. We acknowledge the Finnish Agency for Technology and Innovation, Pakkasmarja Inc., Riitan Herkku Inc., Valioravinto Inc., Vinkkilä Organic Product, Turku University Foundation and ABS Graduate School for financial support.
About this article
Effect of a low dose of sea buckthorn berries on circulating concentrations of cholesterol, triacylglycerols, and flavonols in healthy adults
European Journal of Nutrition (2009)