Ageing is associated with significant changes in gastrointestinal function. Due to changes in gut physiology, immune system reactivity and diet, elderly individuals are more susceptible to gastrointestinal problems and diseases than young adults. Diseases such as peptic ulcer and gastric cancer, irritable bowel syndrome, diverticulosis and colon cancer are relatively common problems of ageing. Constipation, another common concern in the elderly, often results in increased use of laxatives with ageing(Reference Harari, Gurwitz, Avorn, Bohn and Minaker1).
Furthermore, the common use of non-steroidal anti-inflammatory drugs (NSAID) among the elderly may affect their intestinal health. NSAID have analgesic, antipyretic and anti-inflammatory effects. The most common adverse effect caused by NSAID is damage to the mucosa in the gastrointestinal tract(Reference Bjarnason, Hayllar, MacPherson and Russel2, Reference Laine, Connors, Reicin, Hawkey, Burgos-Vargas, Schnitzer, Yu and Bombardier3). According to a recent study by Hartikainen et al. (Reference Hartikainen, Mäntyselkä, Lohivuori-Laako and Sulkava4), 70 % of the over-75-year-olds in the city of Kuopio (Finland) take at least one analgesic, of which NSAID are the most commonly used.
During infancy, the intestinal microbiota changes dramatically(Reference Favier, Vaughan, de Vos and Akkermans5) after which it remains relatively stable during childhood and throughout adult life(Reference Zoetendal, Akkermans and de Vos6). However, during old age, it has been observed that changes in the composition of the microbiota occur again. Using culture-based methods, it is generally shown that ageing correlates with a decrease in Bifidobacterium levels(Reference Mitsuoka, Hayakawa and Kimura7–Reference Hopkins, Sarp and Macfarlane9). Using molecular biological methods, such a decrease in total faecal Bifidobacterium counts in the elderly has not been observed. In contrast, levels similar to those normally observed in young adults have been detected(Reference Hopkins, Sarp and Macfarlane9, Reference Bartosch, Woodmansey, Paterson, McMurdo and Macfarlane10). Other reported changes in the microbiota of seniors are an increase in the numbers of lactobacilli, clostridia and facultative anaerobes(Reference Mitsuoka8, Reference Mitsuoka and Hayakawa11). These changes have, however, received less attention in recent molecular analyses of the intestinal microbiota of the elderly, and appear to be less pronounced than previously considered(Reference Hopkins, Sarp and Macfarlane9, Reference Bartosch, Woodmansey, Paterson, McMurdo and Macfarlane10).
Weakened immunity in the elderly has been observed by many studies addressing different arms of immune responses both in animal models and, to a lesser degree, in human studies(Reference Gill, Darragh and Cross12). In the present study, faecal IgA, TNF-α and PGE2 were measured. Intestinal IgA has an important role, not only in neutralising intestinal pathogens but also in the maintenance of gut homeostasis(Reference Fagarasan and Honjo13). TNF-α mediates the inflammatory responses necessary for successful defence against intracellular infections. Prostaglandins have an essential role in mucosal functions. They inhibit gastric acid secretion, stimulate bicarbonate and mucus secretion and are involved in the regulation of motility and epithelial barrier functions(Reference Mohajer and Ma14). PGE2 has been found to be able to induce proliferative responses in colonocytes as well as in colon cancer cells(Reference Dembinski and Konturek15, Reference Sheng, Shao, Washington and Dubois16). Prostaglandins also perform an important role in the regulation of inflammatory responses(Reference Martin and Wallace17).
The aim of the present study was to compare gut function, immunological status and the key faecal microbiota components between the elderly subjects, both users and non-users of NSAID, and young adults. This information can aid in the development of products suitable for the different elderly target groups that suffer from gastrointestinal complications.
Materials and methods
Study subjects
The study was carried out in two stages. The first part consisted of elderly subjects living in a nursing home and young adults living at a boarding school. The second stage consisted of elderly individuals not living in a nursing home. In total, twenty-six elderly users of NSAID, twenty-nine elderly non-users of NSAID and fourteen young adults took part in the study in the cities of Kuopio and Helsinki (both in Finland).
The elderly were defined as NSAID users when they used the medication three or more times per week. The NSAID used included indomethacin, ibuprofen, ketoprofen, diclofenac, piroxicam, celecoxib, nimesulid and aspirin. The exclusion criteria were critical illnesses, inflammatory bowel disease, coeliac diseases and major malignancies in the gastrointestinal tract. Subjects using antibiotics during the study or within 2 months before the screening were excluded. The use of probiotic and prebiotic products was not allowed during the study. The inclusion and exclusion criteria were monitored during the run-in period before randomisation. Throughout the sampling both the elderly and the young study subjects followed their habitual diet and most of them used the uniform meal service provided at least once per d.
The study protocols were approved by the Research Ethics Committees of the Hospital Districts of Northern Savo and Helsinki-Uusimaa. The purpose of the study was explained to the participants, and written consent to the study was given by all of the subjects. Information on the demographic data, diseases, current medication and use of probiotics and prebiotics were obtained at the beginning of the study.
Faecal samples
Spot samples of faeces were obtained three times from each subject over 3 consecutive weeks. Subjects were personally instructed to take the faecal samples according to the written protocol. All the equipment and documentation papers were provided to the subjects before sampling. All faecal samples were initially stored at − 20°C, transferred to the laboratory within 12 h of defecation and stored at − 70°C until analysed.
Physico-chemical analyses
The concentration of ammonia in the faecal samples was analysed with an enzymic test kit (Boehringer Mannheim/R-Biopharm catalogue no. 11 112 732 035; R-Biopharm AG, Darmstadt, Germany). The pH of the faecal samples was measured with a Mettler Toledo InLab 427 (Mettler Toledo Inc., Columbus, OH, USA) electrode according to protocol 104A:1984 of the International Dairy Federation. For DM determination, about 1 g of faecal sample was weighed and dried (105°C for 16 to 18 h), cooled down in a desiccator to room temperature and reweighed.
The concentrations of acetic, propionic, butyric, valeric, isovaleric, capronic and isocapronic acid were determined by capillary GC (HP-6890; Hewlett Packard, Palo Alto, CA, USA) by adapting the method described by Hoverstad et al. (Reference Hoverstad, Fausa, Bjorneklett and Bohmer18) whereby faecal samples were diluted 1:10 with 0·1 m-potassium phosphate buffer (pH 7·0), homogenised for 1 min with a Stomacher blender (Seward, Thetford, Norfolk, UK) and filtered. The filtrate was sonicated for 1 min at 4°C and centrifuged (450 g; 15 min; 4°C). The samples were stored at − 20°C. For analysis, a 4 ml sample of faecal suspension was mixed with an internal standard (2-ethyl butyric acid; 5 mg/100 g), 250 μl sulfuric acid pro analysi, 10 ml diethyl ether and approximately 3·5 g NaSO4 and sonicated for 5 min. The 4 ml sample was then transferred into a heptane-regenerated solid-phase extraction Varian NH2 cartridge (Varian Inc., Palo Alto, CA, USA). The cartridge was then washed with 5 ml isopropanol–heptane mixture (1:3, v/v). The sample was ready for GC after elution with 5 ml of 3 % HCOOH in diethyl ether.
To analyse the d- and l-lactic acid concentrations, 1 g of the sample was mixed with 6 ml water. The sample was shaken thoroughly and centrifuged at 7000 rpm for 10 min and 0·4 ml of the supernatant fraction was transferred into a 1·5 ml microfuge tube. To precipitate protein, 0·4 ml of 0·4 m-HClO4 was added. The sample was kept on ice for 5 min and centrifuged (16 000 g; 5 min). To neutralise the sample, 0·6 ml of the supernatant fraction was transferred into a new 1·5 ml microfuge tube and 70 μl 2 m-KOH was added. The sample was kept on ice for 5 min and centrifuged (16 000 g; 5 min). d- and l-Lactic acid were then analysed enzymically (R-Biopharm E1112821; Darmstadt, Germany) from the supernatant fraction.
Microbial analyses
The presence and quantity of yeast, anaerobic and aerobic bacteria were analysed by cultivation methods. Faecal samples were thawed in an anaerobic chamber and diluted to a ratio of 1:10 with Wilkins–Chalgren medium (Oxoid Ltd, Basingstoke, Hants, UK) in plastic bags. The mixture was homogenised, a tenfold dilution series was made and the appropriate dilutions were plated on agar. Aerobic bacteria were cultivated for 3 d and the anaerobic bacteria anaerobically for 5 d on brain heart infusion agar (Oxoid) at 37°C. The yeasts were grown on yeast extract–glucose–chloramphenicol agar (YGC agar; International Dairy Federation standard 94B:1990) at 25°C for 5 d.
The total bacteria cell counts in faecal samples were determined by flow cytometry (FACSCalibur; Becton Dickinson, Franklin Lakes, NJ, USA) as previously described(Reference Apajalahti, Kettunen, Kettunen, Holben, Nurminen, Rautonen and Mutanen19). The bacterial fractions were recovered by suspending faecal samples in a buffer, followed by centrifugation and washing. The cell samples were diluted, fixed and stained with a fluorescent nucleic acid-binding dye (SYTO 24; Molecular Probes, Leiden, The Netherlands). The results are expressed as the quantity of bacteria/g wet faeces.
DNA was extracted from the recovered bacteria using the method described by Apajalahti et al. (Reference Apajalahti, Särkilahti, Mäki, Heikkinen, Nurminen and Holben20) whereby bacteria were subjected to five freeze–thaw cycles and subsequently treated with lysozyme and proteinase K. The recovered bacterial DNA was used to quantify the total bifidobacteria, as described by Gueimonde et al. (Reference Gueimonde, Tolkko, Korpimäki and Salminen21). The primers and probe for the detection of sulfate-reducing bacteria were designed using the adenosine-5′-phosphosulfate reductase α gene of Desulfovibrio intestinalis, and for the detection of Clostridium perfringens by using the α-toxin gene with PrimerExpress software (Applied Biosystems, Foster City, CA, USA). Quantitative real-time PCR was performed using 1 μg of isolated bacterial DNA with primers and probes specific for C. perfringens and sulfate-reducing bacteria (Table 1). A 25 μl amplification reaction consisted of 1 × TaqMan universal master mix (Applied Biosystems) with 300 nm of both reverse and forward primers, and 200 nm TaqMan probe (Applied Biosystems). All assays were run on an ABI Prism 7000 Sequence Detection System (Applied Biosystems) using the instrument's default settings for thermal cycling and fluorescence measurements. For standard curves bacterial genomic DNA from C. perfringens (ATCC 13124) and in-house isolated genomic DNA from D. intestinalis (DSM 11275) were applied. Based on the genomic sizes, the weight of one copy of each the C. perfringens and D. intestinalis genome was calculated, and the amount of chromosomes in 1 pg was estimated. Standard amplification curves were constructed by using 1, 10, 100 and 1000 pg of bacterial genomic DNA as a template. The results are expressed as the quantity of bacteria/g faeces (fresh weight).
Primers and probes used for the quantitative determination of Clostridium perfringens and Desulfovibrio intestinalis
For fluorescence in situ hybridisation (FISH) analysis faecal samples were diluted to the ratio of 1:10 with PBS and homogenised in filter bags for 2 min with a Stomacher blender. A quantity of 5 ml of the suspension was transferred into a tube containing glass beads (diameter 2 mm). The sample was mixed for 30 s with a test-tube mixer and centrifuged for 2 min (250 g) to remove the coarse material. Of the supernatant fraction, 375 μl were mixed to a ratio of 1:4 with fresh 4 % paraformaldehyde (pH 7·2) and fixed overnight at 4°C. The fixed bacterial cells were collected (13 000 g; 5 min), washed twice with PBS and re-suspended, first in 150 μl PBS, after which an equal volume of 94 % ethanol was added. The suspension was then mixed and stored at − 20°C until used. The cell suspension was diluted to a ratio of 1:20 with 0·9 m-NaCl-20 mm-tri(hydroxymethyl)-aminomethane–0·1 % SDS (pH 7·2) hybridisation buffer at 50°C and a fluorescent probe was added to the final concentration of 5 ng/ml. Hybridisation was carried out overnight at 50°C, then the washing and filtering were performed according to Franks et al. (Reference Franks, Harmsen, Raangs, Jansen, Schut and Wellin22). The filters were mounted on a slide with AntiFade reagent (Molecular Probes) and covered with glass. The cells were counted visually with an epifluorescence microscope, examining at least fifteen fields in every sample. The oligonucleotide probes used in the FISH analysis were Cy3-labelled Bfra602 described for the Bacillus fragilis group, and Bdis656 for the B. distasonis group(Reference Franks, Harmsen, Raangs, Jansen, Schut and Wellin22), Bif164 for the bifidobacteria(Reference Langendijk, Schut, Jansen, Raangs, Kamphuis, Wilkinson and Welling23), Erec482 for the C. coccoides–Eubacterium rectale group, Chis150 for the C. histolyticum group(Reference Franks, Harmsen, Raangs, Jansen, Schut and Wellin22), Lab158 for the lactobacilli and enterococci(Reference Harmsen, Elfferich, Schut and Welling24) and Fprau0645 for the Faecalibacterium prausnitzii (Reference Suau, Rochet, Sghir, Gramet, Brewayes, Sutren, Rigottier-Gois and Dore25).
Immunological analyses
The concentrations of IgA, TNF-α and PGE2 were measured from the soluble fraction of the faeces. The frozen samples were thawed and extracted with bovine serum albumin as described previously(Reference Peuranen, Tiihonen, Apajalahti, Kettunen, Saarinen and Rautonen26) and stored at − 20°C before analysis. Concentrations of IgA, TNF-α and PGE2 were determined using ELISA according to the respective manufacturer's instructions (E80-102; Bethyl Laboratories, Inc., Montgomery, TX, USA; Biosource Europe S.A., Nivelles, Belgium; Cayman Chemical Company, Ltd, Ann Arbor, MI, USA) and the results were expressed as μg or pg/g faeces (fresh weight).
Statistical analysis
The summary statistics (mean, minimum, maximum) were separately calculated for each group. After performing the initial analysis, the effects of NSAID and age were studied separately by comparing the mean levels of parameters in the various groups using the t test.
The relationships of the correlated variables were studied using principal component analysis. The eigenvalues were calculated, and the percentages of total variation explained by them and the first ten principal components. The statistical analysis was performed using SAS statistical analysis software (version 9; SAS Institute, Cary, NC, USA).
Results
The baseline characteristics, defecation frequency, quality of faeces and use of laxatives of the participants are presented in Table 2. The mean age of the elderly subjects was 75·6 years (without NSAID) and 77·8 years (with NSAID), compared with the mean age of 28·1 years for the young adults. The self-reported defecation frequency was highest in young adults (1·6 times/d) decreasing in order in the elderly NSAID users (1·1) and non-NSAID users (0·6). Hard faeces were reported by 38 % and laxatives were used by 35 % of the elderly but by none of the young adults. The results of the physico-chemical, microbiological and immunological measurements are presented in Table 3. In general, the most important variables, explaining 77 % of the variation in the present study, were, in descending order, the use of NSAID, age, IgA, PGE2, TNF-α, DM, pH, ammonia and acetic acid (data not shown). Since NSAID use and age were the variables that explained most of the variation in the study, the effects of NSAID use and age are considered in more detail below.
Baseline characteristics of the sixty-nine participants of the three groups
NSAID, non-steroidal anti-inflammatory drugs.
(Mean, minimum (Min) and maximum (Max) values)
cfu, Colony-forming units.
*P < 0·05, **P < 0·01, ***P < 0·001, ****P < 0·0001 (pairwise t test).
† Detection limit for isovaleric acid, isobutyric acid, valeric acid, capronic acid and isocapronic acid was 1 mg/100 g fw.
‡ Detection limit for l- and d-lactic acid was 10 μmol/kg fw.
§ Detection limit for yeasts was 2 log10 cfu/g and for C. perfringens and sulfate reducers was 2 log10 cells/g.
Effect of use of non-steroidal anti-inflammatory drugs
The effect of NSAID use was first studied by comparing the elderly NSAID users with the elderly without NSAID, and second with the young subjects, none of whom used NSAID (Table 3; groups 1 v. 2, and groups 2 v. 3).
The mean proportion of faecal DM and the concentrations of isobutyrate, isovaleriate and l-lactate were significantly lower in the elderly with NSAID than in the elderly non-NSAID users (P < 0·001, P < 0·01, P < 0·01; P < 0·05, respectively). The use of NSAID did not have an effect on the measured faecal immune parameters or cause any significant changes in microbiota when compared with the elderly non-NSAID users (Table 3; groups 1 v. 2).
The elderly NSAID users had significantly lower levels of total SCFA (P < 0·05), butyric acid (P < 0·05), propionic acid (P < 0·01), isocapronic acid (P < 0·0001), C. coccoides–E. rectale bacteria (P < 0·05) and PGE2 (P < 0·01) when compared with the young adults (Table 3; groups 2 v. 3). In addition, a trend for lower acetic acid (P = 0·0542) and valeric acid (P = 0·0675) concentrations and higher numbers of bifidobacteria (P = 0·0657) were detected in the elderly NSAID users when compared with the young adults.
The effect of age
The effect of age was assessed by comparing the elderly non-NSAID users with the young adults (Table 3; groups 1 v. 3). The elderly non-NSAID users exhibited higher proportions of faecal DM (P < 0·01) and higher numbers of aerobic bacteria than the young adults (P < 0·05). Moreover, they had higher concentrations of isobutyric acid (P < 0·05), isovaleric acid (P < 0·05), isocapronic acid (P < 0·0001) and l-lactate (P < 0·05) than the young adults. Furthermore the elderly non-NSAID users had lower concentrations of propionic acid (P < 0·05) and PGE2 (P < 0·05) than the young adults.
The combined effects of non-steroidal anti-inflammatory drugs and age
A comparison of the elderly subjects, irrespective of the use of NSAID, with the young adults revealed significantly higher quantities of aerobic bacteria (P < 0·05) and higher concentrations of isocapronic acid (P < 0·0001), but lower concentrations of PGE2 (P < 0·01) and propionic acid (P < 0·001) in the elderly (Table 3; groups 1 and 2 v. 3). Moreover, there were trends for higher quantities of Lactobacillus/Enterococcus (P = 0·0585) and lower amounts of C. coccoides–E. rectale (P = 0·0676) and total bacteria (P = 0·0686), lower concentrations of acetic acid (P = 0·065), butyric acid (P = 0·075) and the total SCFA (P = 0·0599) in the elderly, when compared with the young adults.
Discussion
Constipation is known to be a common problem among the elderly, as evidenced by the high rate of laxative use(Reference Harari, Gurwitz, Avorn, Bohn and Minaker1). In this study population, 35 % of the elderly used laxatives but none of the young adults. Generally the most common defecation frequency is once daily(Reference Heaton, Radvan, Cripps, Mountford, Braddon and Hughes27). According to this survey the defecation frequency was highest in young adults then decreasing in order in the elderly with NSAID and then without NSAID use. The impaired bowel habit in non-NSAID users supports our finding that non-NSAID users had drier faeces compared with NSAID users. The use of NSAID is known to be associated with gastric damage but not with constipation, which is a typical problem with opioids(Reference Pappagallo28).
The self-reported stool type changed from normal to hard with advancing age: young adults reported a normal stool type but over one-third of elderly reported hard faeces. This is also consistent with the occurrence of increasing DM content of faeces associated with ageing. The faecal DM also correlated negatively to the main SCFA (acetic, propionic and butyric acid; R − 0·391, P = 0·001; R − 0·335, P = 0·006; R − 0·304, P = 0·013, respectively) and the total SCFA (R − 0·313, P = 0·01) concentrations, giving support to the hypothesis that accelerated intestinal transit is associated with increased colonic SCFA concentration, as indicated by Lewis & Heaton(Reference Lewis and Heaton29). The non-NSAID users in particular had altered physical bowel functioning, i.e. significantly higher faecal DM and decreased defecation frequency, which could lead some bacterial groups to decrease or become less active. Such consequences could be caused by a loss of contact by microbes with intestinal mucosa, a lack of energy, or by increased concentrations of toxic substances(Reference Tongeren, Slaets, Harmsen and Welling30). The NSAID users had more moist faeces than the non-NSAID users, but compared with the young subjects, still lower C. coccoides–E. rectale counts and reduced SCFA production.
Clear differences in colonic fermentation were demonstrated: NSAID use and/or age decreased concentrations of SCFA (acetate, propionate, butyrate) but increased concentrations of branched-chain fatty acids and l-lactic acid. Decreased concentrations of the major SCFA may indicate decreased availability of butyrate to the mucosa. Butyrate is the main source of energy in the colonic mucosa(Reference Roediger31) and thus any dietary change that can increase its availability in the elderly could have important health implications. Branched fatty acids are themselves not toxic, but are metabolites produced during protein fermentation (for a review, see Blachier et al. (Reference Blachier, Mariotti, Huneau and Tomé32)), which are indicating the production of other potentially toxic metabolites, such as indolic and phenolic compounds, which have been linked with long-term adverse health effects such as the development of colon cancer(Reference Bone, Tamm and Hill33).
In the present study, relatively few significant differences in the microbiota of elderly subjects were found when compared with the young. The levels of aerobic bacteria were higher in the elderly non-NSAID users, which is consistent with observations by Woodmansey et al. (Reference Woodmansey, McMurdo, Macfarlane and Macfarlane34). Although the levels of Lactobacillus/Enterococcus have previously been reported to have increased in the elderly(Reference Mitsuoka and Hayakawa11, Reference He, Harmsen, Raangs and Welling35, Reference Mueller, Saunier and Hanisch36), conflicting results have also been published(Reference Hopkins, Sarp and Macfarlane9, Reference Woodmansey, McMurdo, Macfarlane and Macfarlane34). Furthermore, a recent study indicated that age-related differences in Lactobacillus/Enterococcus levels could be country specific(Reference Mueller, Saunier and Hanisch36). There does not seem to be a consistent trend regarding the changes in the levels of these taxonomic groups in the elderly. The present study indicated a trend for increased Lactobacillus/Enterococcus levels in the elderly compared with the young. Although animal studies have indicated increased small-intestinal colonisation with enterococci when treated with the NSAID indomethacin(Reference Dalby, Frank, St Amand, Bendele and Pace37), we did not observe changes in the levels of Lactobacillus/Enterococcus in the NSAID users group. This may relate to methodology which did not distinguish between enterococci and lactobacilli, or the fact that in the present study we analysed faecal rather then small-intestinal microbiota. Consistent with the findings of others(Reference Tongeren, Slaets, Harmsen and Welling30, Reference He, Harmsen, Raangs and Welling35, Reference Mueller, Saunier and Hanisch36) in the combined group of elderly NSAID users and non-users the present study found that the C. coccoides–E. rectale group, which is among the most numerous groups in adults, had decreased significantly. The commonly reported often-mentioned decrease in faecal bifidobacteria in the elderly(Reference Mitsuoka and Hayakawa11, Reference Saunier and Dore38) could not be demonstrated in the present study. A similar difference in results had also been reported earlier by Bartosch et al. (Reference Bartosch, Woodmansey, Paterson, McMurdo and Macfarlane10) and may partly be explained by the differing use of culture-based techniques(Reference Mitsuoka and Hayakawa11) or molecular-based techniques by Bartosch et al. (Reference Bartosch, Woodmansey, Paterson, McMurdo and Macfarlane10) and the present study.
In the elderly faecal PGE2 levels were significantly decreased to approximately half the levels of those in the young adults irrespective of the use of NSAID. The functions of PGE2 in the large intestine are not fully understood but it has been proposed that ageing in general is associated with the development of a cancer-promoting microenvironment characterised by chronically elevated levels of PGE2(Reference Schwartsburd39). Increased concentrations of PGE2 have been demonstrated previously in vitro in macrophages and in bronchoalveolar lavages obtained from elderly subjects(Reference Wu, Marko, Claycombe, Paulson and Meydani40, Reference Meyer41). In the present study, however, faecal concentrations of PGE2 were clearly decreased in the elderly. This supports previous findings suggesting that ageing decreases prostaglandin concentrations, and thus mucosal protection, in the stomach and duodenum(Reference Newton42).
The volunteers did not exhibit acute-type inflammation in the large intestine, as judged by the low faecal TNF-α concentrations in all test groups, and therefore no indication of increased background inflammation was found(Reference Lesourd and Meaume43).
Interestingly, the faecal concentrations of IgA were similar in all test groups. Although it has previously been suggested that cellular-type immune responses are especially compromised by advanced age, and furthermore that responses are deviated towards humoral immunity, characterised by an increase in serum, bronchoalveolar lavage, and saliva IgA and IgG concentrations(Reference Meyer41, Reference Arranz, O'Mahony, Barton and Ferguson44, Reference Lesourd, Mazari and Ferry45), ageing does not appear to influence the intestinal IgA concentrations(Reference Arranz, O'Mahony, Barton and Ferguson44). The difference between the intestine and other mucosal sites and serum may be explained by the differing types of regulation of immune responses at the different sites, or by the fact that different immunoglobulin subclasses may change differentially with age(Reference Lottenbach, Mink, Barenkamp, Anderson, Homan and Powers46, Reference Powers47). Furthermore, measurement of specific IgA concentrations after oral vaccination would give a more direct measure of the mucosal responsiveness(Reference Fujihashi, Koga and McGhee48).
In summary, the most significant change in gut content with ageing was the increase in faecal DM and, especially in the elderly NSAID users, a decrease in the activity of microbial metabolism, such as butyrate concentrations. In contrast, an increase in the concentrations of branched fatty acids was detected in both the elderly NSAID users and non-users. Evidently, the decreasing bowel function in the elderly contributes to the changes in colonic fermentation towards an unfavourable putrefactive microbial metabolism. Dietary products that can enhance bowel function and saccharolytic fermentation would be especially beneficial for the elderly. In contrast to systemic and respiratory immune responses, ageing results in a different type of immune profile in the intestine, which is characterised by no change in IgA concentrations and a clear decrease in PGE2. Therefore, dietary intervention that can normalise the faecal PGE2 concentrations in the elderly will probably reflect an improvement in the protection of the mucosa.
Acknowledgements
K. T. contributed to the planning of the study, data management, coordination of the manuscript preparation and data interpretation. S. T. contributed partially to the design of the study, microbiological analyses, and manuscript preparation and data interpretation. A. O. contributed to the manuscript preparation and data interpretation. T. A. contributed partially to the microbiological analyses. N. R. coordinated planning of the study, manuscript preparation and data interpretation. The authors thank Markku Saarinen, Brita Mäki, Kirsi Stenström, Jaana Oksanen, Päivi Nurminen, Pia Sigvart-Mattila, Anne-Maria Riihimäki, Tuula Vähäsöyrinki, Juha Laukonmaa, Pirkko Sirviö and Anu Surakka for the various laboratory analyses, Tarja Suomalainen, Essi Sarkkinen, Henna Karvonen and Kajsa Kajander for the practical arrangements, Akra-Numero Research and Consultancy Centre, India for the statistical analysis and John Cowasji for the linguistic form. We especially appreciate the help of the organisers of the sampling in the Pohjois-Savo Vocational Institute in Siilinjärvi and the Harjula Hospital in the District of Northern Savo. We are indebted to the participants of the study for their cooperation and assistance. The study was supported by the National Echnology Agency of Finland (TEKES). There are no conflicts of interest.
