Common infectious diseases (CID) remain a considerable cause of morbidity and mortality worldwide(Reference Yoshikawa1) particularly in the ageing population. The elderly experience more frequent and severe community-acquired respiratory and gastrointestinal infections(Reference Gavazzi and Krause2). Influenza and pneumonia are the fourth most common cause of death among the elderly(Reference Yoshikawa1, Reference LaCroix, Lipson and Miles3). Moreover, 77 % of deaths due to gastrointestinal infections(Reference Djuretic, Ryan and Fleming4) and 90 % of deaths due to respiratory infections(Reference Mouton, Bazaldua and Pierce5) are reported to occur in patients over 65 years of age.
The increased susceptibility of the elderly has been ascribed to age-associated alterations of the immune system affecting both innate and adaptive immune responses. For example, the antibody and cell-mediated immune responses to influenza vaccine are reduced in the elderly(Reference Goodwin, Viboud and Simonsen6–Reference Targonski, Jacobson and Poland8), and the clinical effectiveness of annual influenza vaccination is 50–60 % in subjects over sixty-five as compared with 80–90 % in younger adults(Reference Nichol9). In addition, ageing-associated changes such as involution of the thymus, decline of naive T-cell numbers, reduction in T-cell repertoire diversity and accumulation of specific memory T cells result in disruption of T-cell population balance and decreases in protection against new pathogens(Reference Nikolich-Zugich10). Ageing is also associated with a reduction in the activity of different cell subsets from the innate immune system such as monocytes(Reference Delpedro, Barjavel and Mamdouh11, Reference McLachlan, Serkin and Morrey-Clark12), polymorphonuclear neutrophils(Reference Wenisch, Patruta and Daxbock13) and natural killer cells(Reference Mysliwska, Mysliwski and Romanowski14, Reference Solana and Mariani15).
Achieving optimal success in preventing and controlling CID among the elderly requires not only the development of more effective vaccines and antimicrobial drugs, but also new strategies to oppose to the age-associated alterations of the immune system. The latter includes improvement of the living conditions, promotion of physical activity and better nutrition including functional foods(Reference Riezzo, Chiloiro and Russo16, Reference Jones and Varady17). These foods include fermented dairy products such as yoghurt or fermented milk some of which contain probiotics that may improve the resistance of old people to infection(Reference Gill and Guarner18–Reference Meydani and Ha23). Products containing probiotics were also shown to have an immunomodulatory effect in different classes of age(Reference Fukushima, Miyaguchi and Yamano20, Reference Nova, Warnberg and Gomez-Martinez24–Reference de Vrese, Winkler and Rautenberg29), and in several studies, modifications of immune parameters have been correlated with a protective effect against CID of the gastrointestinal or respiratory tract(Reference Fukushima, Miyaguchi and Yamano20, Reference Kaila, Isolauri and Soppi27–Reference de Vrese, Winkler and Rautenberg29). The present study was prompted by a pilot observation that the consumption of a fermented dairy product containing the probiotic strain L. casei DN-114 001 (Actimel®) in the elderly population is associated with a reduction in the duration of winter infections (gastrointestinal and respiratory infections taken together)(Reference Turchet, Laurenzano and Auboiron30). In a higher sample size and a more extended observation period in the present study, we investigated the effect of daily consumption of the same product on the resistance to CID. To this end, 1072 elderly volunteers were included in a multicentre, randomised, double-blind, controlled study over a 3-month winter period.
Materials and methods
Design
A multicentre, randomised, double-blind, controlled parallel study was conducted to evaluate the effect of the consumption of a fermented dairy product containing a probiotic on the resistance to CID of the airways and gastrointestinal tract (Table 1) in free-living elderly volunteers. One group of volunteers was randomly allocated to the fermented product and the other to a control product. The study consisted of a 3-month (84 d) product consumption phase and a 1-month (28 d) follow-up phase without consumption of the study product. The product consumption phase was preceded by a 2-week period of dietary restriction that was maintained until the end of study. The trial was conducted between 2 November 2006 (first subject included) and 4 May 2007 (last subject completed the study) by general practitioners in 125 centres distributed in twenty-five departments in France.
Table 1 Common infectious diseases (CID) and associated symptoms
The study received approval from the independent ethics committee (Independent Ethics Committee/Institutional Review Board) Lyon A (Hôpital Hôtel Dieu, Lyon, France) on 23 August 2006 and was conducted in line with the principles of the Declaration of Helsinki, and of Good Clinical Practice, and European regulatory requirements.
Subjects
Inclusion criteria were: male and female individuals of at least 70 years of age who were free-living (not residing in an institution), with a Gerontological Autonomy Iso-resource Group (AGGIR) score between 5 and 6 (AGGIR score assesses the individuals' physical and psychological independence on a six-point scale, where a score of 6 represents total autonomy and a score of 1 represents total dependence); vaccination against the influenza virus at least 14 d before inclusion; a mini-mental state score of at least 24; a BMI between 17 and 25 kg/m2 (bounds included); compliance with a dietary restriction during the 2 weeks preceding the product consumption phase and throughout the study. The restriction implied the exclusion of fermented dairy products with probiotics other than those used in the study, yoghurts and over-the-counter medication containing probiotics, vitamins, minerals or other nutrients.
Written consent was obtained from all participants and all of them were registered with the national social system security or benefited from a similar programme.
Exclusion criteria included: residing in institutions; any current or past severe respiratory, gastrointestinal or metabolic pathology; chronic or iatrogenic immunodeficiency, any infection in the last 14 d, laxatives more than twice in the last week; food allergy or intolerance; any progressive or chronic disease such as unstable type 1 and 2 diabetes or cancer, cognitive, neurological, cardiac or renal diseases; major surgery with general anaesthesia during the last month or gastrointestinal surgery during the last 3 months; artificial nutrition within the last 2 months; special medicated diets; nutritional complements; eating or transit disorders; alcohol abuse; currently receiving or having received in the 4 last weeks; drugs likely to interfere with evaluation of the study parameters, including antibiotics, intestinal or respiratory antiseptics, antifungal (except topical), corticoids, vaccines (except influenza vaccine), anti-histaminic molecules, non-corticoid anti-inflammatory substances (except aspirin or equivalent at doses preventing from aggregation of platelets or blood clotting) and immunosuppressant treatment.
Intervention
Volunteers were randomised to receive either the fermented product or the control product. The fermented product was a sweetened, flavoured, fermented dairy drink (Actimel®) containing at least 1010 colony-forming units (CFU)/100 g of the probiotic strain L. casei DN-114 001 (international reference: CNCM I-1518, also named Lactobacillus paracasei subsp. paracasei following the current nomenclature), combined with a symbiosis of two cultures commonly used in yoghurt, Streptococcus thermophilus and Lactobacillus delbrueckii subsp. bulgaricus (at least 109 CFU/100 g for the whole symbiosis). The health effects are carried by the entire complex product containing not only the strains (L. casei DN-114 001 and yoghurt ferments), but also all of the metabolites resulting from the proprietary fermentation process. Therefore, a non-fermented, acidified, sweetened, flavoured dairy drink was chosen as an appropriate control to respect double blinding without the active components(Reference Aggett, Antoine and Asp31, Reference Farnworth32). The nutritional composition, appearance, taste and packaging of the fermented product and control were identical throughout the study in order to maintain blinding. The study products were provided and manufactured by Danone (Bierun, Poland) and distributed by the investigators every 14 d.
During the phase of consumption, volunteers had to ingest two bottles of 100 g/d of fermented or control product (one bottle at breakfast and one at dinner, preferably with an interval of at least 8 h).
Procedure
The total duration of the study was 4·5 months (112 d). At the selection visit, 2 weeks before initiation of the study, volunteers underwent a clinical examination and started the dietary restriction that was maintained during the entire study. Two weeks after the selection visit, a certificate of eligibility was issued by the investigator, and the volunteers were randomly allocated to fermented product or control product group on day 0 of the study. Symmetric randomisation was carried out in blocks on a 1:1 ratio. Volunteers were assigned to study groups using an individual randomisation number (study product allocation concealed) and were included sequentially in accordance with the randomisation list, which was stratified by centre. Participants consumed the product for 3 months and were observed for a further 1 month. During the consumption phase, planned evaluation visits were conducted at the end of months 1, 2 and 3. A final evaluation was conducted at the end of the 1-month follow-up phase. At all the planned visits, a clinical examination was undertaken, including measurements of weight, BMI, blood pressure and heart rate. Dietary restriction compliance was also assessed, adverse events and concomitant medication were recorded and the SF-36® quality of life questionnaire was completed. Compliance was assessed by asking volunteers to record their daily consumption in a personal diary and to return any unused bottles at each visit. For the analysis of biological parameters, blood samples were also taken for a subset of included volunteers at day 0 just before the first product consumption (for analysis at baseline), at month 2 (following 2 months of product consumption) and at the end of the 1-month follow-up visit. Stool samples were also taken at each planned visit for a subset of included volunteers.
In addition, volunteers attended an additional evaluation visit each time they presented clinical symptoms related to the defined CID classified by category: upper respiratory tract infections (URTI, defined as the following types: rhinopharyngitis; sore throat; sinusitis; otitis), lower respiratory tract infections (LRTI, defined as the following types: bronchitis; pneumopathy; flu and flu-like syndromes) and gastrointestinal tract infections (GITI, defined as gastro-enteritis) (Table 1). These evaluation visits were conducted 3 d ( ± 1 d) after the initiation of the symptoms. At each additional visit, a clinical examination was performed, the type of CID was diagnosed and the start date (first day of symptoms), end date (last day of symptoms) and global severity (defined as mild, moderate or severe) of the disease were reported. A CID for which data were captured at the additional visits was termed a ‘CID regular reported’ and was thus diagnosed by a doctor during the CID. Occasionally, data were captured at the planned visits and the infection defined as a ‘CID delayed reported’, which refers to a CID diagnosed by a doctor on the basis of symptoms reported by the volunteers after the end of CID event. During each CID, in addition to their symptoms, the volunteers also recorded their temperature and their medication taken. At each of the additional visit, biological samples were taken for the identification of specific pathogens according to the type of CID. In addition, blood samples were also collected in a subset of volunteers to assess the same biological parameters as for planned visits. All the analytical dosages in the study were performed in blind.
Data analysis
The primary outcome measure was a comparison between groups of the cumulative number of all CID (Table 1) reported during the 3 months of study product consumption. Additionally, the occurrence of CID (defined as the number of subjects experiencing at least one CID) was investigated as a secondary outcome. Only CID occurring after the start date of product consumption were considered. A new CID occurring after a previous CID of the same type was considered as a separate event only if there was at least 2 d between the two events. Evolution from one type of disease to another was recorded as two CID. Multiple events of different infectious diseases occurring jointly (e.g. an URTI and a gastro-enteritis) were counted separately and added.
Another secondary objective of the study was to determine whether fermented product consumption changes the duration of CID (both cumulative and average duration per episode) calculated from the first to the last day of symptoms as defined in Table 1 for each CID. Additional secondary objectives were to assess the impact of fermented product consumption on the time span to the first occurrence of CID (time of event), on the severity of CID assessed with a three-point scale (mild–moderate–severe, based on the degree of interference with the subject's daily activity), on fever associated with CID (occurrence, duration and maximum, with fever defined as temperature ≥ 38°C) and on the occurrence or duration of medications (prescribed and/or auto-administered). In addition, analyses of the micro-organisms responsible for CID were performed by measuring the occurrence and amount of specific pathogens in biological samples (laboratory: CDL Pharma, Marseille, France). For URTI viruses, identification and quantification in nasal fluid samples were performed by quantitative reverse transcription-PCR. For bacteria identification and quantification, in case of LRTI in expectorations or throat swabs (the latter in case of sore throat) or in case of GITI in stools samples, a microscopical examination and Gram staining were performed and samples were cultured on enriched and specific media (API® gallery method).
In a subset of the included population, secondary objectives were to assess changes in biological (haemogram, serum C-reactive protein) and immunological parameters and the presence of L. casei species in the stools (laboratory: CDL Pharma). The immune system was assessed by analysis of blood parameters defined as oxidative burst activity in monocytes, cytolytic activity and count of blood natural killer cells and cytokines production in serum (IL-1 and IL-6, interferon-α, interferon-β, interferon-γ, IL-12, IL-10, TNF-α and IL-8). All biological analyses were performed by using standard procedures. The quantification of L. casei species in stools has been assessed by measuring total L. paracasei (which include L. casei DN-114 001) by quantitative PCR following the procedure described by Haarman & Knol(Reference Haarman and Knol33) with some modifications. Briefly, total DNA was extracted from stools by using a commercial kit according to the manufacturer recommendations (QIAamp DNA Stool kit, Qiagen, Courtaboeuf, France). Purity and concentration of DNA were then evaluated by measuring absorbance at 260 and 280 nm in a spectrophotometer. L. paracasei group-specific real-time PCR (TaqMan universal master mix, Applied Biosystems, Nieuwerkerk a/d IJssel, The Netherlands) was then performed. For the selection of primer and probe sequences, the 16S–23S intergenic spacer regions of the different Lactobacillus species were retrieved from the GenBank, EMBL and DDBJ databases: L. paracasei AB035487, AF182724 and U32964. Sequences were aligned and the conserved regions were determined by using DNASIS for Windows version 2.5 (Hitachi Software Engineering Co., Ltd, Wembley, UK). Using Primer Express 1.5a (Applied Biosystems), specific sequences were identified to design primers and probes (primers: F_paca_IS 5′-ACA TCA GTG TAT TGC TTG TCA GTG AAT AC-3′ and R_paca_IS 5′-CCT GCG GGT ACT GAG ATG TTT C-3′; probe P: paca_IS 5′-TGC CGC CGG CCA G-3′). A standard curve has been included in the quantitative polymerase chain reaction by using a suspension of L. paracasei of known concentration allowing the quantification of L. paracasei in the samples in equivalent of CFU (eq. CFU)/gram of stools.
The impact of the fermented product on subjects' quality of life was also assessed using the SF-36® questionnaire(Reference Ware, Kosinski and Dewey34). The SF-36® questionnaire comprises thirty-six items used to construct scores across eight dimensions (physical functioning, role physical, bodily pain, social functioning, mental health, role emotional, vitality and general health) and two summary scores: the physical component score; the mental component score.
For all parameters, analyses were performed during the 3 months of study product consumption, during the 1-month follow-up and during the whole-study phases, for all CID (any type), or each category (URTI, LRTI, GITI) or each type of CID.
Throughout the study, tolerability and safety were assessed by recording spontaneously reported adverse events and by measuring evolution of vital signs, i.e. systolic and diastolic blood pressures, heart rate and weight during the 3-month product consumption phase, during the 1-month follow-up phase and during the whole-study overall.
The severity of adverse events was defined as mild (no interference with the subject's daily activity), moderate (moderate but acceptable interference with the subject's daily activity) or severe (marked and unacceptable interference with the subject's daily activity). A serious adverse event was defined as an adverse event that: resulted in death; was life threatening; was likely to result in disability or permanent invalidity; resulted in hospitalisation or prolonged hospitalisation; was medically significant. Any serious adverse events had to be reported up to a maximum of 24 h after the investigator became aware of it. CID were not reported as adverse events, except if considered serious.
Statistics
The study design planned to include 1000 volunteers. This sample size was determined based on the primary outcome of the study: the cumulative number of CID episodes occurring within the 3 months of product consumption. According to several publications(Reference Chavance, Herbeth and Lemoine35, Reference Graat, Schouten and Kok36), it was assumed that an average number of 1·5 events would be observed in the control group with an estimated 15 % relative decrease in the fermented product group. Also it was presumed that the winter period during which the study took place allowed the expectation of the maximum number of CID episodes. The distribution of the cumulative number of infectious events was assumed to be an overdispersed Poisson distribution, with overdispersion expected because the occurrences of CID events within a volunteer are not independent. With an expected rate of 1·5 events over 3 months in the control group and using a Poisson regression with a two-sided test at the 5 % α-level and assuming moderate overdispersion, about 450 evaluable volunteers in each arm were expected to be needed to detect a 15 % reduction rate with at least 80 % power. A 5 % dropout rate of volunteers was assumed and as such inclusion of approximately 500 volunteers in each arm was necessary.
For the assessment of biological parameters, blood was collected from only a certain number of subjects. According to available literature(Reference Parra, De Morentin and Cobo37–Reference Marcos, Warnberg and Nova39), a total of about 100 evaluable subjects were necessary to allow analysis of differences in immunomodulatory effects between groups. For the quantification of L. casei DN-114 001 in the stools, the number of subjects performing stool collection was fixed at forty-eight.
Data were treated to build up a cleaned and locked database, following data validation, including checking and plausibility assessments. Global database, including derivate parameters, has been locked and used for statistical analysis. Analyses were performed on the intention to treat (ITT) population, which comprised all volunteers who were included, randomised to the groups and having received the study product. Analyses were also performed on the per protocol population. The primary assessment was at 3 months, at the end of study product consumption, with additional assessments conducted for the follow-up phase and for the whole-study overall. Comparisons between the fermented product group and the control group were performed using two-sided statistical tests with a significance level of 5 % (P < 0·05). Appropriate parametric and non-parametric methods were employed according to the distribution of the data. Baseline characteristics were compared using a t test or a Mann–Whitney test. For the outcome parameters, comparisons of continuous data were made using a mixed ANOVA model (or appropriate non-parametric analysis). For comparison of qualitative data, a χ2/Fisher exact test, a logistic regression analysis with a binary response or a Cochran–Mantel–Haenzel test was used. Time to event analyses were conducted using a log-rank test and/or a Cox regression. For comparison between groups of count data, a Poisson regression model was used, taking into account the study duration.
Confounders, including centre, age, sex and AGGIR score, have been taken into account as appropriate. For the primary criteria, history of CID during the last month and study duration were also taken into account. Data were analysed using SAS® software (version 8.2; SAS Institute Inc., Cary, NC, USA).
Results
Volunteer demographics
The flow of the volunteers through the studies is displayed in Fig. 1. A total of 1089 individuals were selected for inclusion in the study, with 1072 volunteers randomised to the fermented product (537) or the control product (535). Out of these volunteers, 125 were selected for blood sample assessment (sixty-three randomised to fermented product and sixty-two randomised to control). Sixty-three volunteers were selected for assessment of L. casei in stools (thirty-two randomised to fermented product and thirty-one randomised to control).
Fig. 1 Volunteer flow in the study.
There were 208 (19·4 %) volunteers with a major protocol deviation in the study. This was well balanced between groups. Most frequently, the major deviations were relative to the time between planned visits (7·9 %), the concomitant treatments that could interfere with the study results (4·5 %) and the completeness of the protocol at the final visit (4·3 %).
The baseline characteristics for the ITT population were well balanced across the fermented product and control product groups (Table 2) regarding age (median age = 76·0 years, range 69·0–95·0), sex ratio, BMI, mini-mental state score, current diseases, concomitant treatments and history of CID. Current diseases were mainly vascular diseases, muscular–bone diseases and cardiovascular diseases. The results of the initial clinical examination were normal for all volunteers.
Table 2 Baseline characteristics and product compliance for volunteers in the intention to treat population
(Median and quartile values and number and percentage of subjects)
Q, quartile; MMS, mini-mental state.
* Mann–Whitney.
† χ2 test.
Compliance was 100 % in median for all volunteers (assessed by the number of consumed bottles of products), with no difference between the study groups (Table 2). The consumption duration was 84·0 (83; 85) (median (Q1; Q3)) days (expected value for the study: 84 ± 6 d), and the mean number of bottles consumed was 168·0 (164; 169) (median (Q1; Q3)) bottles per subject (expected value for the study: 168 ± 12 bottles per subject). Forty-six volunteers (4 %) withdrew from the study prematurely and each withdrawal was unrelated to the product consumption.
Primary outcome measure: cumulative number of episodes of common infectious diseases
The primary outcome measure was analysed in the ITT population using a generalised linear model assuming a Poisson distribution, with factors of age, sex, AGGIR score, history of CID in previous month, product group, centre (random effect) and taking into account the study duration. During the 3 months of product consumption, 217 volunteers (20·4 % of the whole population) experienced a total of 255 episodes of CID (Fig. 2).
Fig. 2 Percentage of volunteers by the number of common infectious diseases (CID) experienced (intention to treat population) for all CID (any type), upper respiratory tract infections (URTI), lower respiratory tract infections (LRTI) or gastrointestinal tract infections (GITI) during the product consumption phase. The y-axis represents the percentage of volunteers. The x-axis describes the specific infection and number of infection experienced. Volunteers receiving fermented product (n 537) are represented by ■; volunteers receiving the control product (n 535) are represented by 
. No volunteer experienced three URTI or LRTI whatever the group, 0·2 % of the volunteers in the control group and 0 % in the fermented product group experienced two or three GITI.
In the fermented product group, 105 volunteers reported 120 CID during the product consumption phase, compared with 112 volunteers who reported 135 CID in the control group. Considering the whole population, the mean rate for this period was 0·2 (sd 0·5) in the fermented product group and 0·3 (sd 0·5) in the control group (rate reduction = 10·6 %), with an adjusted relative risk of 0·89 (95 % CI 0·70, 1·14, P = 0·373). The difference between groups regarding the cumulative number of all CID was not statistically significant. In addition, there was no statistically significant difference between groups when the whole-study phase was considered (3-month consumption phase+1-month follow-up phase).
There were no statistically significant interactions between fermented product and the included factors, and therefore subgroup analyses were not undertaken (sex: P = 0·577, age: P = 0·555, AGGIR score: P = 0·974, CID history: P = 0·199).
Analyses were conducted based on when the CID were reported. Among the 217 volunteers who have displayed CID, 82·9 % immediately reported CID and 17·1 % reported with delay. Whatever the time of reporting, there was no significant difference between groups for the cumulative number of CID.
Secondary outcome measures
Episodes of common infectious diseases by category and type (upper respiratory tract infections, lower respiratory tract infections and gastrointestinal tract infections)
In ITT population, during the whole-study phase, URTI accounted for the majority of reported CID (54·5 %) followed by LRTI (29·1 %) and GITI (16·4 %), and this pattern was similar across groups. Rhinopharyngitis was the most frequently reported type of CID (50·3 %), followed by bronchitis (23·6 %), gastro-enteritis (16·4 %), flu and flu-like syndrome (5·5 %), sinusitis (3·1 %) and sore throat (1·0 %). Neither otitis nor pneumonia was reported during the study.
During the product consumption phase, the number of URTI, LRTI, GITI (Fig. 2) or of each type of CID (data not shown) was not statistically different between groups.
Common infectious diseases duration
The average duration of disease for each episode and the cumulative duration of CID were analysed in ITT population experiencing CID using a Mann–Whitney non-parametric test. When considering all the CID during the study product consumption phase, both the duration of CID episodes and the cumulative duration of CID were significantly lower in the fermented product group compared with the control group (P = 0·008 and 0·009, respectively; Table 3). For both episode and cumulative duration, significant differences between groups were also found when the whole-study phase (consumption phase plus follow-up phase) was considered (P = 0·019 and 0·018, respectively).
Table 3 Cumulative and episode duration of common infectious diseases (CID) during product consumption phase in intention to treat population with CID
(Mean values and standard deviations; median values and quartile ranges)
Q, quartile; URTI, upper respiratory tract infections.
* Mann–Whitney test, median and quartiles must be considered as a summary statistic.
These differences were also significant in favour of fermented product for all URTI and for rhinopharyngitis during the product consumption phase (Table 3) and during the whole-study phase (all URTI: episode duration and cumulative duration, P = 0·004 for both; rhinopharyngitis: episode duration, P = 0·01; cumulative duration, P = 0·009).
For all CID, all URTI and rhinopharyngitis, the median episode duration was 1–1·5 d shorter in the fermented product compared with the control group (Table 3).
Regarding the LRTI and GITI, there was no significant difference in the episode duration or the cumulative duration.
All significant results were confirmed in the per protocol population presenting CID.
Lactobacillus casei in stools
Analysis of the amount of L. casei species in stools was performed at each planned visit in the ITT population. Comparison between groups was analysed by considering both the actual values and the change from baseline for each visit in both groups by using a Mann–Whitney test. At baseline, the results show no statistical difference between groups in the amount of L. casei species in stools (P = 0·525; Fig. 3). After 1, 2 or 3 months of product consumption, a statistically significant difference in the amount of L. casei was observed between the two groups (according to evolution from baseline and actual values at each point – P < 0·001 for all comparison). The number of L. casei remained unchanged in control group throughout the whole study at about 4·9 × 105 eq. CFU/g of stools (baseline value; Fig. 3). In the fermented product group, the median value of L. casei amount increased from 2·9 × 105 (baseline) to 3·3 × 107 eq. CFU/g after 1-month consumption and was roughly maintained at the same level throughout the product consumption (2–3·8 × 107 eq. CFU/g of stools, P < 0·001). At the end of the follow-up phase, no statistical difference was found between groups and in each group compared with baseline value. This indicated a return to baseline level of L. casei amount in the fermented product group 1 month after the consumption has ceased. All results were confirmed in the per protocol population.
Fig. 3 Amount of Lactobacillus casei species in stools (in log10 equivalents of CFU/g of stools) during product consumption and after 1-month follow-up in intention to treat population (, fermented product (n 32); □, control product (n 31)). In each box, upper limit, central line and lower limit indicate the quartile 3 (Q3), median and Q1 values, respectively. The upper and lower limits of the dotted lines indicate the minimum and maximum values, respectively. P value at 0 month (for comparison between groups of the data at baseline) and P values at 1–4 months (for the comparison between groups of the change from baseline) were calculated using a Mann–Whitney test.
Other secondary outcome measures in intention to treat population
There was no difference between the two groups of participants regarding the occurrence of CID (defined as the number of subjects having at least one CID), the time to first event, the severity, the intensity or duration of fever or CID-associated medication (prescribed and/or auto-administered). Only 3·6 % of the participants (n 4 in each group) experienced severe CID during the product consumption phase. In the same population and period, the proportion of volunteers who had fever associated with CID was 28·2 v. 28·3 %, and the proportion of volunteers receiving at least one prescription of medication was 67·5 v. 64·2 % in the fermented and the control product group, respectively.
The analysis of pathogens in case of CID could not be compared between groups due to an insufficient number of positive samples for the presence of the specific micro-organisms studied. Biological and immunological parameters were comparable between the two groups. They could not be analysed in the case of CID due to the insufficient number of infections. Quality of life was also assessed by change from baseline in SF-36® scores. No significant difference was reported between the two groups regarding the eight dimensions and the physical and mental component scores of the questionnaire.
Safety
The safety profile of the fermented product was comparable with the control product. During the study, 276 volunteers reported 416 adverse events. A summary of reported adverse events is given in Table 4. The most frequent categories of reported adverse events were the muscular–bone system (19 %), the gastrointestinal system (18·5 %) and infections other than CID (16·3 %), and there was no significant difference between groups in number of adverse event for any of these categories. Although a total of twenty-five serious adverse events were reported, none were related to fermented product consumption.
Table 4 Summary table of adverse events (total number and percentage of volunteers with at least one event) during whole-study phase in intention to treat population
AE, adverse event; emergent AE, AE that began or worsened during the product consumption phase.
Regarding the vital signs during the study, there was no significant difference between the groups in the evolution from baseline of weight (P = 0·471), of systolic (P = 0·902) or diastolic blood pressure (P = 0·554) or of heart rate (P = 0·586), and no relevant clinical evolution of these parameters was found in each group.
Discussion
The aim of this double-blind, randomised, controlled study was to assess the effect of a fermented dairy product containing the probiotic strain L. casei DN-114 001 on the resistance to common respiratory and gastrointestinal infections in an elderly free-living population. The baseline and demographic characteristics of the volunteers, the duration of study and the compliance to the product consumption were similar for the fermented product and the control product groups. There was also no difference between groups and no relevant modification in each group regarding the safety parameters including adverse events and vital signs. Furthermore, withdrawals from the study were rare and unrelated to the product consumption.
The study was performed from October 2006 to May 2007 and thus included the winter season when the rate of CID was expected to be high. However, instead of the estimated 1·5 CID we observed only 0·3 (sd 0·5) CID per subject in the control group. This was likely due to a particularly low incidence of infections during the winter 2006/2007 (as documented by the Groupes Régionaux d'Observation de la Grippe(40)). We can thus speculate that the fact that no difference was found between groups for the primary outcome measure, the cumulative number of all CID during the product consumption phase, may be attributed to a lower than expected number of events. Additionally, the low incidence of severe CID (4·2 %) could have been associated with an underreporting of total infections, which is consistent with the 17 % of delayed reported CID observed.
It is particularly interesting that the fermented product was found to significantly reduce both the average duration of CID per episode and the cumulative duration of CID. These reductions were statistically significant in the product consumption phase and also for the duration of the study as a whole (product consumption phase plus follow-up phase) when all the CID were considered. Significant reductions were also reported for all URTI and specifically for rhinopharyngitis. Fermented product consumption was found to reduce by 1–1·5 d (median) the duration of CID observed in the control group, which was comprised between 8 and 8·5 d. The present results confirm the preliminary observations on a similar ageing population showing that administration of the same product reduced the duration of respiratory and gastrointestinal infections(Reference Turchet, Laurenzano and Auboiron30). Other studies also described the influence of probiotics on respiratory tract infections in children(Reference Hatakka, Savilahti and Ponka41) or adults(Reference de Vrese, Winkler and Rautenberg29, Reference Cox, Pyne and Saunders42). In an unblinded trial, elderly immunised by influenza and pneumococcal vaccination displayed a lower incidence of infections, especially respiratory, after consumption of a nutritional supplement containing L. paracasei and other nutrients such as vitamins(Reference Bunout, Barrera and Hirsch28). In hospitalised enterally fed elderly, consumption of a fermented milk containing a probiotic strain of Lactobacillus johnsonii was shown to reduce the duration of infections requiring antibiotics (respiratory, intestinal and urinary infections taken together) and to improve respiratory symptoms(Reference Fukushima, Miyaguchi and Yamano20). The present study is the first randomised, double-blind, controlled trial indicating that a dairy probiotic product alone can be active against URTI in free-living elderly. The reduction of the duration of CID episodes by 1·5 d is in the range of the efficacy of neuraminidase inhibitors (1·0 d reduction) in influenza treatment(Reference Cooper, Sutton and Abrams43). The observed reduction of duration of respiratory infections could be associated with another benefit that is a lower risk of medical complication leading to severe symptoms since human rhinoviruses, the most common etiologic agents of URTI, were found to be associated with high frequency of prolonged illness and lower respiratory tract disease in elderly(Reference Hayden44, Reference Wald, Shult and Krause45). Such an effect could be investigated in a larger cohort of subjects.
Duration and severity of CID could be related clinically. Here, we observed an effect of the fermented product on CID duration, but not on the severity of CID. This could be attributed to the very low number of severe episodes in the whole population, which might have prevented the detection of any significant difference between groups. The absence of a product effect on fever is also at variance with a previous result showing a significant decrease of the maximum temperature recorded during winter pathologies in the elderly consuming the same product(Reference Turchet, Laurenzano and Auboiron30). However, in the latter, the incidence of infectious disease was much higher (0·38 per subject for 3 weeks only) and thus, more severe pathologies may have been reported as exemplified by the occurrence of bacterial bronchopneumonia.
Increase of Lactobacilli level in faeces and survival of L. casei DN-114 001 in the gut were previously reported in young adults(Reference Oozeer, Leplingard and Mater46–Reference Rochet, Rigottier-Gois and Levenez48) and children(Reference Guerin-Danan, Chabanet and Pedone49) after consumption of a dairy product containing L. casei DN-114 001. In the present study, we provide the first evidence in elderly consuming the fermented product for an increase of L. casei species in the gut as quantified in the stool analysis. A 100-fold higher level was observed throughout the product consumption and decreased to baseline level when consumption was stopped, suggesting the possibility of a transient presence in the gut of the probiotic strain as previously described in adults using selective culture methods(Reference Oozeer, Leplingard and Mater46). In any case, this result indicates a good compliance of the subjects to the products consumption. Further investigations are needed to understand the origin of those Lactobacilli since they could be the probiotic strain L. casei DN-114 001 and/or commensal gut bacteria of the same species specifically favoured by the product.
The mechanism by which a fermented product taken orally can have an effect against infection of the airways is a matter of debate. A clinical study showed that consumption of a mix of probiotics decreases the occurrence in nasal cavity of potentially pathogenic bacteria known to be responsible for respiratory infectious disease(Reference Glück and Gebbers50). Another possible mechanism could be the effect of probiotics on the innate and adaptive immune system(Reference Corthesy, Gaskins and Mercenier51), as previously reported for dairy product containing L. casei DN-114 001(Reference Parra, De Morentin and Cobo37–Reference Marcos, Warnberg and Nova39, Reference Pujol, Huguet and Drobnic52, Reference Tiollier, Chennaoui and Gomez-Merino53). In the present study, none of the immune parameters tested was modulated during fermented product consumption. However, the same product can improve immune responses to influenza vaccination of the elderly(Reference Boge, Rémigy and Vaudaine54). Therefore, we cannot rule out the possibility of an action of the fermented product on other immune parameters than the one assessed. In addition, the analysis of the data on the blood immune parameters during infection could not be performed due to the too low number of subjects presenting a CID in the subset of population with blood samples collection. Several other mechanisms enhancing defence were described, which may have contributed to the clinical effect observed(Reference Sherman, Ossa and Johnson-Henry55, Reference Bourlioux, Koletzko and Guarner56). In vitro experiments on intestinal epithelium human cells showed that L. casei DN-114 001 can inhibit the cell adhesion and invasion by adherent-invasive Escherichia coli (Reference Ingrassia, Leplingard and Darfeuille-Michaud57) and the increase in paracellular permeability induced by enteropathogenic E. coli infection(Reference Parassol, Freitas and Thoreux58). Furthermore, innate defence may be increased by probiotics via enhancing defensin expression. Until now, however, this has been shown in vitro using E. coli Nissle only(Reference Wehkamp, Harder and Wehkamp59). To what extent an improvement of mineral absorption and vitamin production by lactic acid producing bacteria may play a role is uncertain. There is some evidence for an increase in absorption of minerals with potential impact on defence such as Zn and Se by prebiotics. For probiotics, however, the evidence is still scarce(Reference Scholz-Ahrens, Ade and Marten60).
All the active components of the product used in the present study have not been fully identified. The health effects are likely carried by the whole product containing not only the strains (L. casei DN-114 001 and yoghurt ferments), but also all of the metabolites resulting from the fermentation process. This explains the design of our control product as a non-fermented, acidified, sweetened, flavoured dairy drink with no bacteria. The contribution of the probiotic strain L. casei DN-114 001 is supported by two studies in children, which compared the effect of the same fermented product to the one of a standard yoghurt containing S. thermophilus and Lactobacillus bulgaricus (Reference Pedone, Bernabeu and Postaire61, Reference Pedone, Arnaud and Postaire62). The results showed that compared with jellied milk, the consumption of the probiotic fermented product significantly reduced the duration of diarrhoea, whereas standard yoghurt had no significant effect(Reference Pedone, Bernabeu and Postaire61). In addition, when directly compared with a standard yoghurt, the probiotic fermented milk also significantly decreased the incidence of diarrhoea(Reference Pedone, Arnaud and Postaire62). There is therefore evidence that L. casei DN-114 001 may play a role in the effect of the fermented product we used, providing superiority to the yoghurt symbiosis present in the product. However, this symbiosis could also provide beneficial effects as illustrated by several reports describing the health benefit of yoghurt consumption on immune system and gut, especially gastrointestinal infections(Reference Adolfsson, Meydani and Russell22, Reference Meydani and Ha23). A further clarification of the extent to which the components of the fermented product used in the present study contribute to its effect would require a much higher sample size and separate studies with adequate, prior statistical estimations.
In conclusion, consumption of a fermented dairy product containing L. casei DN-114 001 is associated with a significant decrease in the duration of CID, especially URTI and particularly rhinopharyngitis providing the first evidence that a dairy fermented product containing a probiotic may have a beneficial effect against respiratory infections in the free-living elderly. It is possible that the low number of CID observed in the study as a whole contributed to the lack of effect of the product on the primary and some secondary outcomes. Based on the present results and on previous observations(Reference Turchet, Laurenzano and Auboiron30) and given the high morbidity and mortality associated with infectious diseases especially of the airways in the elderly, dietary intervention using a dairy product containing the probiotic strain L. casei DN-114 001 could be considered as a mean to improve health status of this population.
Acknowledgements
We thank particularly Leila Mhamdi (Keyrus, Levallois-Perret, France) for statistical analyses, Bénédicte Borgies, Caroline Caty and Isabelle Seksek (Danone Research) for the study coordination, Stéphane Doat and Tanveer Hayat (Danone Research) for the logistic organisation of products supply, Thomasz Gawlas (Danone-Poland, Bierun, Poland) for management of products manufacture and Pierre Burguière and Jérôme Combrisson for their support on L. paracasei analysis. We also thank all the general practitioner members of MGRecherches who participated to the study as investigators and the members of the contract organisation research Advanced Drug Development Services (Boulogne-Billancourt-France) involved in the management of the general study organisation (Séverine Guigui), the data management (Viriginie Gomis) and the statistical calculations (Vincent Houé). We thank Sophie Costello and Ebony Samuels for their help in manuscript preparation (Heron Evidence Development Ltd, Letchworth Garden City, UK). E. G. (research scientist at Danone Research) participated in the development of study concepts and designs, the writing of the protocol, the analyses and interpretation of the data, the writing of the study report and of the publication. F. T. (biostatistician at Danone Research) participated in the study design, in the writing of study protocol, in statistical management, in the statistical analyses and interpretation of the data, in the writing of the study report and of the publication. F. L. received funding from Danone Research as main investigator of the study. J. S. contributed to the design of the study, to the evaluation and interpretation of all data and to writing of the manuscript. He received fees for contributing to the Advisory Board of Danone, Germany. The Department of Physiology and Biochemistry of Nutrition, MRI was funded by Danone Research for performing several studies. Its institute was also funded by Campina, Chr. Hansen, Danisco, Müller, Nestlé, Bauer and Merck for studies on probiotics, and he received fees as a speaker at a symposia sponsored by Yakult, Danone, Nestlé, Orthomol and Merck. The work was undertaken at General practitioner centres of Quintiles/MGRecherches, France. The study was sponsored and funded by Danone Research.
