Health effects of probiotics and prebiotics A literature review on human studies

Human studies on health effects of probiotics and prebiotics were reviewed and evaluated. The main results can be summaries as follows: Certain probiotic lactobacilli may improve lactose digestion and reduce symptoms of lactose intolerance. The effect of probiotics on serum cholesterol is still inconclusive. Animal studies showing triacylglycerol-lowering effects of prebiotics need confirmation in humans. Data on effects of probiotics on constipation are not convincing, whereas inulin has dose-related laxating effect. Effects of a probiotic drink have been reported on symptoms in irritable bowel syndrome, but more studies are needed for firm conclusions. A significant shortening of acute watery rotavirus-included diarrhoea has been demonstrated for two lactobacilli, whereas possible effects on the risk of getting traveller's diarrhoea need further studies. There are promising indications that probiotics could be useful against antibiotic-associated diarrhoea, and a yeast preparation has been shown to reduce the risk of relapsing Clostridium dificile diarrhoea. Promising results from studies on the effect of probiotic products in the treatment of gastritis and inflammatory bowel disease should encourage further studies with pro-, preand synbiotic foods. Certain prebiotic oligosaccharides may increase calcium absorption. Probiotics can be regarded as safe although occasional infections have been reported in immunosuppressed patients. Prebiotics such as fructans may cause dose dependent gastrointestinal side-effects. The documentation of health-promoting effects of probiotic and prebiotic products is rapidly increasing. The food industry that develops proand prebiotic products should increase their efforts to develop high quality research and well-designed clinical trials on ordinary food products. This area is of great importance for improving human health.


Introduction
There is increasing evidence that the composition and metabolic effects of the gastrointestinal microflora are of key importance for human health. In addition to promoting normal gastrointestinal functions and protecting from infections, the microflora also seems to exert important effects on systemic metabolism and immune functions.
The definitions of probiotics as "live microbial food ingredients that are beneficial to health",prebiotics as "non-digestible food components that beneficially affect the host by selectively stimulating the growth and/or activity of one or a limited number of bacteria in the colon, that have the potential to improve host health", and synbiotics as "mixtures of probiotics and prebiotics that beneficially affect the host by improving the survival and implantation of live microbial dietary supplements in the gastrointestinal tract" are generally accepted (1).
Microorganisms representing many different genera have been used as probiotics. Most of the efforts have been focused on lactobacilli and bifidobacteria, but also enterococci and the yeast Saccharomyces boulardii have received attention. The term "pre-biotics" was coined at a time when the first studies appeared showing that fructo-oligosaccharidesundigestible in the small intestinecould be utilised only by a few bacterial species, notably bifidobacteria, and that feeding such oligosaccharides to experimental animals and man increased the count of bifidobacteria in the intestinal content (2). Much research on oligosaccharides, notably fructo-oligosaccharides and inulin, has been performed recently, including European Commission sponsored projects (3). It should be noted, however, that the stimulation of certain microorganisms by different carbohydrates in vivo does not appear as selective as indicated by in vitro studies Thus, resistant starch and non-starch polysaccharides have also shown a capacity to stimulate bifidobacteria in vivo (4). A question which must be addressed, however, is whether bifidobacteria in thousands really improve host health. nosus), prebiotics, inulin, oligofructose, diarrhoea, lactose, blood lipids, cholesterol, irritable bowel syndrome (IBS), cancer, mucositis, inflammatory bowel disease (IBD), Chrohn's disease, inflammation, allergy, safety.
To find additional articles, and also as a control of our search, reference lists of several recent review articles about pro-and prebiotics were used. Furthermore, companies with pro-or prebiotic foods on the Swedish market were asked to send documentation already published or in press. The articles were sorted after the studied effect into the following groups: Lactose intolerance Blood lipids Hypertension Constipation Irritable Bowel Syndrome (IBS) Diarrhoea Gastritis and reflux disease Inflammatory Bowel Disease (IBD) Cancer prevention Prebiotics and mineral absorption Safety The papers were critically reviewed by the members of the group. The focus of the examination has been studies using probiotic or prebiotic food products, but also studies with lyophilised bacteria have been considered when relevant.
Some studies are listed in Tables 1-4 on pages 66-75. Those references are marked (*) in the following text and in the reference list. Articles are evaluated in the Tables (column 13) as + low quality; ++ medium quality; +++ high quality.
Due to the complexity of effects on immune functions, a more in-depth review on this topic is presented separately (5).

Lactose intolerance
Lactose is the predominant carbohydrate in milk. The presence of sufficient lactase activity in the small bowel mucosa is necessary for the newborn child to absorb lactose from breast milk. This enzyme is, however, dramatically reduced in adult life in the majority of people. The genes that enable the activity of lactase to remain high prevail mainly in the white Western populations. In Sweden, e.g., low lactase activity is relatively uncommon. It should be noted, however, that lactose malabsorption often occurs without symptoms of lactose intolerance. Table 1 (on pages 66,68,69) shows studies on lactose intolerance. Several studies have demonstrated that subjects with low intestinal lactase activity absorb lactose from yoghurt (6*,7*) or milk containing L. acidophilus (8*) better than from milk. In another study (9*), however, a four-fold difference in lactase activity between the products had no effect on the digestion and tolerance of lactose. Yoghurt is also better tolerated (gives fewer symptoms) than milk (7*,10*). In these studies breath hydrogen concentrations have been measured, as an indication of bacterial fermentation of undigested lactose. The method is only semiquantitative and there is considerable variation in how much hydrogen is absorbed and how much passes with the stools.
Mechanisms: Up to about half of the lactose content can be fermented after 11 days storage (1 1 *), and microbial lactase activity enters the small bowel with the fermented product. Subjects with lactose intolerance experience more symptoms also after ingesting fructo-oligosaccharides (12*), indicating increased sensivity to malabsorbed carbohydrates.
Conclusion: Despite the above-mentioned concerns, it seems logical to assume that fermented milk products with specific probiotic lactic acid bacteria improve lactose digestibility and absorption, and may reduce symptoms of lactose intolerance. This conclusion is in agreement with a recent review by Hove et al. (13).

Blood lipids Probiotics
A number of studies have examined the potential of probiotic products to reduce serum cholesterol levels, some are listed in Table 1 (on pages 66-70). Studies without suitable control or placebo groups, lack of run-in periods or administering very large doses of fermented dairy products (700-5000 mllday) were excluded. An important consideration in the evaluation of the studies is the fact that participation for a single week in a nutritional trial in itself may result in a reduction of serum cholesterol. Furthermore, the analytical precision of serum cholesterol determination has to be considered.
About half a litre of a yoghurt with L. acidophilus (14*,15*) or tablets with viable L. acidophilus and L. bulgaricus (16") did not reduce serum cholesterol according to one randomised controlled trial (RCT) (39 subjects) and two randomised crossover studies (219 subjects), but a small effect was seen in two other studies (17*,18*). In a further study, yoghurt did not reduce serum lipids (1 9).
Aproduct containing L. plantarum 299v (proviva@), tested in a parallel study with 30 subjects (20*) gave a significant reduction in serum cholesterol. However, the test group and the control group had similar cholesterol levels at the end of the study.
A fermented milk product containing Enterococcus faecium and Streptococcus termophilus (GAIOB) has been tested in three RCT studies with parallel designs (21 "-23") in altogether 214 subjects. The randomisation was not always optimal. A small but significant reduction was found at some time-periods, but no difference was found between the groups after six months in one of the studies (23%). GAIO@ has also been tested in a randomised crossover study (24"). A reduction in serum cholesterol was found after six weeks.

Mechanisms:
The mechanisms whereby probiotics may reduce serum cholesterol are largely unknown. Certain strains of bacteria, however, have the capacity to assimilate cholesterol in vitro. An ileostomy study has shown a reduced absorption of cholesterol from the small bowel after treatment with probiotics (2.5'").

Conclusion:
The effect of probiotics on serum cholesterol is still inconclusive. More long-term studies are required to document a sustained effect. These conclusions are supported by recent reviews (26)(27).

Prebiotics
Dietary carbohydrates represent a complex group of food components. Dietary oligosaccharides, strictly defined, are composed of two to nine monomers linked together, but inulin is often included in spite of a chain length larger than this. Nondigestible oligosaccharides may act as substrates for the colonic microflora.
Convincing serum lipid-lowering effects of inulin have been demonstrated in animals; attempts to reproduce similar effects in humans, however, have given conflicting results. One reason may be that animals were given much larger doses of inulin than tolerated in humans. Animal studies have identified inhibition of fatty acid synthesis as the major site of action for the triacylglycerid-lowering effect of inulin. This pathway is, however, relatively inactive in humans (28).
Two RCT studies (29*,30*), including altogether 30 subjects, have shown reductions in serum cholesterol and triglycerides with inulin. Moreover, in one small crossover RCT (3 1 *) on 12 subjects, serum cholesterol and triglycerides were reduced after inulin ingestion. In another crossover RCT in 21 subjects (32*), a reduction of serum cholesterol was found only relative to the control group. In three crossover RCTs (33*-35") including 96 subjects, no effect in serum cholesterol and triglycerides was found. In another study there was also no effect on plasma lipids, but a decreased basal hepatic glucose production was found with fructo-oligosaccharides (36"). In a recent double-blind RCT parallel design study and 54 subjects there was a trend for the Scand J NutrlN2ringsforskning 2/01 triacylglycerol level to be lower after 8 weeks on 10 g of inulin (37*). Ileostomy studies have not shown any increase in cholesterol or bile salt excretion from the small bowel by prebiotic treatment (38*), i.e. addition of 17 g of inulin or oligofructose to the diet.

Conclusion:
No convincing serum cholesterol-lowering effect can be ascribed to inulin or oligo-fructose from the present studies. The effect on the triacylglycerol level in man remains to be elucidated further.

Hypertension
Two Japanese studies have been published on the effect of fermented milk on blood pressure (Table 1, on pages 67,68,70). In the first study, Lactobacillus helveticus and Saccaharomyces cerevisiae were used (39*), in the second one L. casei TMC 0409 and Streptococcus thermophilus TMC 1543 were investigated (40*). In both studies significant reductions in systolic blood pressure were found, and in one study (39*), effects on diastolic blood pressure were also noted. It was suggested that the effect could be due to the formation of certain tripeptides that are inhibitors of ACE (Angio-tension-converting enzyme).
Further studies are needed to test the effect of these peptides and to establish the conditions for their formation.

Constipation Probiotics
Constipation means complaints with bowel evacuations, low bowel emptying frequency and a slow transit through the large bowel. Based on early reports, different fermented milk products have been claimed to alleviate constipation. Such self-reported information, however, is difficult to evaluate.
Two hundred ml of a L. rhamnosus GG fermented whey drink (41 *) did not change bowel movement frequency or hardness of stools in a small placebo-controlled study. Similar findings were made with a L. rhamnosus GG yoghurt (42). In another study the effect of milk fermented by L. acidophilus on constipation in elderly subjects was difficult to evaluate (43").

Conclusion:
The available data on effects on constipation of probiotics are not convincing. Further studies are needed to substantiate such an effect.

Prebiotics
Dietary fibre increases the faecal bulk by two mechanisms: incompletely fermented types of fibre bind water throughout the gastrointestinal tract, whereas readily fermented types of fibre contribute by increasing the microbial mass (44). Since oligosacharides are completely fermented, their bulking effect would be expected to occur through an increased microbial mass. Table 3, (page 74) lists three studies (45*-47*) in which the effect of fructooligosaccharides on faecal weight in humans has been measured. The quite limited data indicate an increase in faecal wet weight of around or somewhat more than 1 glg ingested oligofructose and 1.5-2 g/g ingested inulin. These figures are similar to those reported for pectin (1.2) but considerably lower than for wheat bran (around 5) (44). In a study on elderly constipated patients inulin reduced functional constipation and increased stool frequency (48).
Conclusion: Inulin and fructooligosaccharides seem to have dose-related laxating effects.

Irritable bo we1 syndrome Probiotics
Irritable bowel syndrome (IBS) is a very common gastrointestinal disorder, and is often the most frequent diagnosis at a gastroenterologist's clinic. Typical symptoms reported are flatulence, variations between diarrhoea (not in the night) and constipation, and abdominal pains.
Different factors are associated with the condition, e.g. food intake, malabsorption and psychosomatic influences. These factors can exert an effect on the motor function of the gastrointestinal tract. Generally, patients with IBS report pain with a lesser degree of abdominal distention than others. It is always difficult to evaluate self-reported symptoms particularly in a condition with psychosomatic influences like IBS.
In a very carefully performed double-blind RCT study on 61 subjects with IBS, no difference in tolerance was seen with unfermented milk containing L. acidophilus compared to ordinary milk (49*).
A rose-hip drink with L. plantarum 299v (400 mllday) was tested in two RCTs with parallel designs. In one of the studies (40 patients), improvement of symptoms was significantly greater in the study group than in the control group (50"). In the other study (52 patients), flatulence was reduced in the test group compared with the placebo group (5 1 *). Abdominal pain was reduced in both groups, even though the reduction was more rapid and pronounced in the test group. There was no major change in gas bloating.
Conclusion: An effect on some symptoms in IBS is reported with L. plantarum 299v. More controlled studies are needed for firm conclusions on the importance of probiotics in the treatment of IBS.

Diarrhoea
The effect of probiotics on diarrhoea1 disease of varying aetiology has been quite extensively studied and clinical trials have recently been reviewed (52,53). We will concentrate here on four conditions that have been subject to human studies using milkbased productsacute watery diarrhoea in children, travellers' diarrhoea, antibiotic-associated diarrhoea and relapsing diarrhoea due to Clostridium dificile infection.

Acute watery diarrhoea in children
Acute diarrhoea in children is mainly caused by rotavirus. L. rhamnosus strain GG is the probiotic strain which has been most extensively studied to treat this condition. Both milk products fermented with this bacterium and freeze-dried bacteria have been shown to shorten acute diarrhoea, especially when caused by rotavirus (54*,55*) ( Table 2, on pages 71-73). Since acute diarrhoea is self-limiting, seldom lasting for more than a week, the therapeutic effect is small. Thus, the duration of diarrhoea is usually shortened by approximately 1 day (Table 2, on pages 7 1 -73). However, the effect is reasonably well proven and has also been replicated in a number of other studies (for a review see 52,53).
Two studies have investigated the effect of Lactobacillus reuteri SD2112 on acute diarrhoea in childhood. Both demonstrate an effect in the same order of magitude as reported for L. rhamnosus GG (56*,57*), (Table 2, on pages 7 1-73).
It should be noted that only one of the above mentioned studies (54*) gives data on the extent of breast-feeding in the different study groups. Since breast-feeding effectively counteracts diarrhoea, it is important to control for this factor.
Conclusion: A significant shortening of acute watery rotavirusrelated diarrhoea in children has been demonstrated for both L. rhamnosus GG and L. reuteri SD2112.

Travellers' diarrhoea
A few placebo-controlled studies have all failed to show effective prevention of infectious diarrhoea in adults (53). These studies were performed with freeze-dried probiotic preparations, and no data have been published on the possible effects of probiotic food products in this respect.
Conclusion: More controlled trials with probiotic foods are needed and justified since some probiotics have shown an effect on diarrhoea in children.

Antibiotic-associated diarrhoea
Treatment with antibiotics results in diarrhoea and abdominal discomfort in a variable fraction of patients, depending on the age group and the antibiotic used. In most cases, the cause of the diarrhoea is unknown, but a varying proportion of the cases are caused by Clostridium dificile. This toxin-producing species is not uncommon in the normal intestinal microflora, but is usually present only in low numbers and without causing any harm. After treatment with certain antibiotics, the lack of competition from other microbes in the normal intestinal flora permits C. dificile to reach high numbers. The C. dificile toxins may cause anything from mild diarrhoea, which can be cured simply by terminating the antibiotic treatment, to the life-threatening disease pseudomembranous colitis.
Table 2 (on pages 7 1-74) includes four randomised placebocontrolled studies (58*-61") that investigate the effect of probiotic intake on gastrointestinal side effects of antibiotic treatment. A yoghurt containing bifidobacteria was shown to quite effectively reduce abdominal complaints in volunteers consuming erythromycin for 3 days (58"). In another study, using parallel groups, no clinical effects of yoghurt with bifidobacteria and Lactobacillus acidophilus were noted, but on the other hand, inulin was added to the yoghurt, which in itself may cause loose stools and abdominal discomfort (59*). An interesting observation was that C. diflicile was isolated from stool cultures of six of ten in the control group, but in only one of nine volunteers in the group given the active yoghurt preparation (p=0.08).
The effect of Lactobacillus rhamnosus GG was studied in volunteers taking erythromycin for 7 days (60"). The data are poorly described in the paper, but according to the authors the volunteers given placebo experienced diarrhoea for 8 days and the volunteers given lactobacilli had diarrhoea only for 2 days.
Arvola et al. studied the potential of L. rhamnosus GG to reduce the risk of antibiotic-associated diarrhoea in a clinical setting (6 1 *). Children receiving antibiotics against respiratory tract infection (amoxicillin being most frequent, followed by penicillin) were randomised to placebo or capsules with lactobacilli. The drop-out rate was quite high and the therapeutic effect was of borderline significance.
Conclusion: Despite the individual drawbacks of the studies cited above, they offer promising indications that probiotics could be useful against side effects of antibiotic treatment. Larger and better controlled studies with probiotic foods are urgently needed.

Relapsing Clostridium difficile infection
Severe C. dificile infection is treated with antibiotics active against anaerobic bacteria (vancomycin or metronidazol). The treatment is successful in most cases, but in some 20% of patients, C. dificile is not eradicated and the patient is plagued by recurrent episodes of diarrhoea. This condition, termed relapsing C. diflicile infection, is difficult to treat and new therapeutic alternatives are needed in this patient group.
Only a single placebo-controlled clinical study has been reported on the effects of probiotics on relapsing C. diflicileinduced disease. This study utilised the yeast Saccharomyces boulardii (62") ( Table 2, on pages 71,72,74). The study was designed, performed and evaluated in an excellent manner, which permits conclusions to be drawn with a high degree of certainty. By adding S. boulardii to the metronidazol or vancomycin treatment aiming to eradicate C. dificile, the risk of the patient relapsing was halved.
L. rhamnosus GG has only been evaluated in open trials against relapsing C. dificile disease. Five adult patients were treated with 10'' CFUIday, and four experienced no further relapses (63). Four children who were treated with L. rhamnosus GG became asymptomatic (64).
Bennet et al. (65%) studied a large series of adult patients, some of whom were referred to a specialist clinician because of relapsing C. diflicile infection. Some of the patients were residents in a nursing home consistently plagued with C. diflicile disease. The patients were given capsules containing L. rhamnosus GG without addition of antibiotics (65 *) ( Table 2, on pages 7 1,72,74). After a single treatment period which lasted for 10-2 1 days, 84% of the patients did not relapse within the follow-up period, which was 1 month for the ambulatory patients and 2 months for the nursing home patients. Since the trial was open, we do not know what the relapse rate would have been, had the patients not been given probiotics. However, it is reasonable to believe that half of the patients would have relapsed during that period without treatment, based on figures from the large multicentre study of McFarland (62*). The study thus indicates that L. rhamnosus could be a promising candidate for treating relapsing C. dificile infection.
Conclusion: One probiotic agent, S. boulardii, has been convincingly shown to reduce the risk of relapsing with C. dificile diarrhoea. For other microorganisms, we have only data from open trials. Since this is a very important potential application for probiotics, controlled studies with probiotic foods should be carried out.

Potential mechanisms involved in control of diarrhoea
It was originally assumed that the ability of probiotic bacteria to shorten diarrhoea was dependent on their ability to colonise the intestine and "alter the microbial balance" in such a way that the pathogen would be eliminated. Specific probiotic strains such as L. plantarum 299 and 299v (66), L. rhamnosus 271 (66) and L. rhamnosus GG (67) have been proven to colonise human volunteers. This might relate to the fact that these lactobacillus species are prevalent on the normal human intestinal mucosa (68). However, some probiotic bacteria do not colonise, but are eliminated at a rate similar to ingested inert particles (69,70). Nevertheless, probiotics that are not likely to colonise can still reduce diarrhoea. This is most strikingly demonstrated for the yeast S. boulardii that has been unequivocally proven effective against C. diflicile-induced symptoms without being able to colonise the intestine. But also bifidobacteria seem to reduce diarrhoea caused by antibiotic treatment (58*) without being able to colonise the individual (59*).
Another proposed effect has been that the probiotic induces an enhanced immune response against the microorganism causing the diarrhoea and that this leads to earlier resolution of the diarrhoea1 disease. This has been proposed to be the mechanism of the effect of L. rhamnosus GG against rotavirus diarrhoea (71). But differences in antibody titres between patients fed L. rhamnosus GG and controls appear only in the convalescence phase (7 I), when the virus has long since disappeared. Moreover, no differences in antibody titres were seen between patients fed L. reuteri and controls (56*), although this organism seemed equally as efficient as L. rhamnosus in controlling the diarrhoea (56*,57*).
The mechanism by which probiotic intake reduces diarrhoea must therefore be regarded as entirely unknown. Potential mechanisms include an influence on the enteric nervous system andlor immune system leading to the production of neuropeptides (72), cytokines (73) or hormones (74) that reduce the secretion of water and electrolytes across the intestinal epithelium.
In the case of C. dificile diarrhoea, the probiotic might act on the host, reducing the secretory response to the clostridial toxin, as outlined above. It is also possible that the probiotic changes the milieu in the intestine, leading to reduced toxin production by the clostridiae, since the toxin production of these organisms is strictly controlled by environmental conditions (75).

Gastritis and reflux disease
L. acidophilus strain LA1 was grown in milk and tested against Helicobacter pylori induced gastritis in a clinical study with promising effect. Suppression of the infection was determined by a standard breath test (76). However, the effect was reversible, as also seen in a Japanese study with a probiotic Clostridiurn preparation. More recently Canducci et al. (77) tested L. rhamnosus GG strain together with antibiotic therapy with fewer gastrointestinal side effects in the probiotic group, suggesting that probiotics could be designed to improve H. pylori treatment outcome and prevent side effects such as symptoms of increased acid reflux post-treatment.

Inflammatory bowel disease
Inflammatory bowel disease (IBD) may be caused or aggravated by alterations in the microbial flora. Thus the distal ileum and colon are most frequently affected by the inflammatory process in patients with IBD, sites which harbour the largest populations of intestinal bacteria. Early studies with probiotic lactic acid bacteria (LAB) (L. reuteri and L. plantarum 299v) strains showed protective effect in chemically induced colitis in rodents. More recently, studies in humans with ulcerative colitis given mesalazine or capsules, containing a well defined E. coli strain, showed no diffeence in relapse rates, but unfortunately no placebo group was included.
One placebo-controlled clinical study was performed recently on with pouchitis, a common long-term inflammation of the ileal reservoir created after surgical removal of the colon. Patients on this therapy for 9 months showed few relapses (15%) compared to 100% in the control group. This study further supports the potential role of probitics food products in IBD therapy and prophylaxis (78-80).
Recently, a complex probiotic preparation containing 200 billion per gram of viable freeze-dried bacteria of four LAB strains, three bifidobacteria and one strain of Streptococcus salivarim, subspecies thermophilus, was tested in a clinical trial in patients showing allergy or intolerance of other origin to classical therapy with mesalamine or sulphasalazine (78-80). The treated group showed reduced faecal pH and remained in remission. In another study, the same complex preparation (VSL x 3 from Yovis, Sigma Tau Pomezia, Italy) showed good effects on patients with chronic relapsing pouchitis combined with increased LAB and S. salivarius counts in pouch contents (80).
Mechanisms: Recent observations in murine knockout models for IBD suggest that an immunlogical up-regulated Thl cell response and breaking of the mucosal tolerance against the indigenous gut microflora are involved in various forms of IBD (79). Other recent observations indicate that patients with ulcerative colitis lack a normal indigenous LAB microflora in colonic biopsies, also supporting the hypothesis that IBD may be prevented by replacement by an appropriate pro-and prebiotic food based regime.
Conclusions: Well-designed large-scale randomised placebocontrolled clinical trials of different pro-and prebiotics preparations versus standard therapy in IBD and pouchitis should now be undertaken. Promising results of probiotic preparations have been reported and should encourage studies with pro-, pre-, and synbiotic foods.

Cancer prevention
Various enzymes in the gut microflora modify ingested foreign compounds such as nitro aromatics, azocompounds and nitrate, which can be metabolised to genotoxic and carcinogenic substances by enzymes of the anaerobic microflora of the colon (81). A number of studies have shown that diet and antibiotics can change the microflora-associated characteristics, and nondigestible oligosaccharides (NDOs) suppress carcinogen metabolising enzyme activities in rats (82). Furthermore, LAB and bifidobacteria have generally low activities of enzymes involved in carcinogen production. Supplementation of galactooligosaccharides (GOS) and the synthetic disaccharide lactulose have been shown to decrease faecal P-glucuronidases and increase lactobacill counts in rats (83,84).
However, despite these experimental facts, evidence is still missing from human studies that LAB and prebiotics such as GOS and FOS (galactose-and fructooligosaccharides) would decrease the risk of colon cancer development in humans. Two early studies in Japan on the treatment of human urinary bladder cancer by L. casei (Yakult) indicates that immunomodulatory effects of LAB and bifidobacteria cells may be used in the future to prevent and treat cancer of the human colon (84,85). However, further studies in this direction should be performed.
Two studies with probiotic milk products in patients undergoing radiotherapy for pelvic malignancies indicate that such products should be further tested to prevent therapy-related diarrhoea and clinical bowel discomfort symptoms (86,87).
Conclusion: Further identification and validation of biomarkers for risk of cancer is a prerequisit for further studies, to evaluate the potential of probiotics and prebiotics in man in relation to cancer.

Prebioticsmineral absorption
Experiments with rats have shown that non-digestible oligosaccharides like inulin and oligofructose can increase the absorption and retention of minerals such as calcium, magnesium, iron and zink (e.g. 88,89). Similar results have previously been obtained with pectin. From animal studies it is postulated that this enhanced absorption occurs in the colon, and that the mechanism is related to increased solubility of calcium due to lower pH of the colonic content induced by fermentation of oligosaccharides.
Three human studies (Table 4, page 75) testing this hypothesis have been published so far. In the first one Coudray et al. (90") used conventional balance technique during 28-day periods. Forty grams of inulin (successively introduced to obtain the maximum dose during the last 12 days of the period) increased the apparent absorption of calcium from 21 to 34% and the retention from -10 to +92 mglday, i.e. by about 100 mglday. A similar improvement in calcium balance was obtained with sugar beet fibre.
When using a double stable isotope technique, van den Heuvel et al. (9 1 *) did not find any effect on calcium or iron absorption after 9 days with 15 g oligofructose/day in young (20-26 y) male subjects. Absorption was measured during 24 h. In a subsequent study (92") the group used the same technique in boys (14-1 6 y), but extended the measurement period to 36 h. An increase in true fractional calcium absorption by 11 % was obtained. Table 1.Studies on lactose intolerance, blood lipids, hypertension and irritable bowel syndrome, columns 1-6 (cont).