Dietary patterns reflecting healthy food choices are associated with lower serum LPS activity


Gram-negative bacteria-derived lipopolysaccharides (LPS) are associated with various negative health effects. Whether diet is associated with LPS, is an understudied phenomenon. We investigated the association between diet and serum LPS activity in 668 individuals with type 1 diabetes in the FinnDiane Study. Serum LPS activity was determined using the Limulus Amoebocyte Lysate assay. Diet was assessed with a food frequency questionnaire (FFQ) section of a diet questionnaire and a food record. The food record was used to calculate energy, macronutrient, and fibre intake. In a multivariable model, energy, macronutrient, or fibre intake was not associated with the LPS activity. Using factor analysis, we identified seven dietary patterns from the FFQ data (“Sweet”, “Cheese”, “Fish”, “Healthy snack”, “Vegetable”, “Traditional”, and “Modern”). In a multivariable model, higher factor scores of the Fish, Healthy snack, and Modern patterns predicted lower LPS activity. The validity of the diet questionnaire was also investigated. The questionnaire showed reasonable relative validity against a 6-day food record. The two methods classified participants into the dietary patterns better than expected by chance. In conclusion, healthy dietary choices, such as consumption of fish, fresh vegetables, and fruits and berries may be associated with positive health outcomes by reducing systemic endotoxaemia.


Lipopolysaccharides (LPS), also frequently called endotoxins, are lipid-soluble outer-membrane components of Gram-negative bacteria1. Among these bacteria are many pathogens, but also much of the commensal population of the human gut (i.e. Bacteroides). Indeed, the bacteria colonized in the intestinal tract are a major source of LPS in humans2. Importantly, LPS may be translocated from the intestine into the systemic circulation. During the prandial state, for example, the lipid regions of the LPS particle are incorporated into the forming chylomicrons, thereby enabling its absorption3. The paracellular translocation of LPS, on the other hand, may be enhanced upon impairment of the integrity of the mucosal epithelial barrier.

Binding of LPS-protein complexes to the toll-like receptor 4 activates cellular NF-κB signalling pathway which, in turn, leads to production of various proinflammatory cytokines and chemokines4. Elevated serum LPS activity levels have been linked with a number of unfavourable changes in human health. In diabetes and in obese subjects, high LPS levels have shown to be negatively correlated with insulin sensitivity5. Accordingly, LPS has been associated with obesity and type 2 diabetes6, 7, as well as progression of non-alcoholic fatty liver disease8. The LPS-induced secretion of proinflammatory cytokines may also lead to endothelial dysfunction, plaque formation and rupture, oxidation of LDL-cholesterol particles, and thrombogenesis, thus contributing to the development of cardiovascular disease6, 9. In type 1 diabetes, we have previously shown that serum LPS activity is associated with the metabolic syndrome10, visceral fat mass11, and the development of diabetic nephropathy12.

Despite the obvious importance of LPS in human health, only relatively few studies have been conducted to assess whether LPS levels could be affected by dietary means. In one of these studies, Erridge et al. reported increase in circulating endotoxin levels, in healthy subjects, following a high-fat meal13. Similarly, cream consumption induced a rise in the plasma LPS concentrations, in healthy normal-weight adults, while isoenergetic intake of glucose had no such effect14. In contrast to these observations, however, a high-fat low-carbohydrate ketogenic diet exhibited lower proinflammatory cytokine levels in an LPS-induced fever model15. In vitro, fruit and vegetable extracts have shown the capacity to inhibit LPS-stimulated inflammatory responses16. Finally, orange juice consumption with a high-fat, high-carbohydrate meal prevented meal-induced increase in endotoxin concentration17.

Amongst the available studies, we were able to identify only one focusing on the association between dietary patterns and LPS levels18. In that study, a 6-month intervention with either a standard Healthy Eating diet or a modified Mediterranean diet was, insufficient to show any changes in the LPS-binding protein concentrations18.

In this paper we report the results of an observational study investigating the association between diet and serum LPS activity in a group of well-characterized individuals with type 1 diabetes. An approach was taken to study the dietary intake both at the level of macronutrient intake (food record data) and at the level of dietary patterns (food frequency questionnaire section of a diet questionnaire, FFQ). We hypothesized that a healthy diet would be associated with lower endotoxin levels. Moreover, we also report the results of a validation study, where the relative validity of the diet questionnaire against a 6-day food record was examined in a large population of individuals with type 1 diabetes.


Food frequency questionnaire data

A total of 668 individuals with type 1 diabetes completed the FFQ and fulfilled the inclusion criteria (Table 1). Of these participants, 47% were men, and the mean ± SD age was 44 ± 13 years. The median (interquartile range) serum LPS activity of the population was 0.37 (0.30–0.49) EU/ml.

Table 1 Basic characteristics of the study population.

Seven dietary factors (or patterns) with high degree of inter-correlation were formed from the FFQ (Table 2). Based on the contents of these clusters, they were named “Sweet”, “Cheese”, “Fish”, “Healthy snack”, “Vegetable”, “Traditional”, and “Modern”. The factor score of the Sweet -diet pattern correlated negatively with the LPS activity (r = −0.080, P = 0.039). In the model only adjusted for all the other dietary patterns, higher factor scores of the Fish and the Healthy snack patterns were associated with lower serum LPS activity (Table 3, Model 1). After further adjusting for sample storage time, age, sex, smoking, physical activity and diabetic nephropathy status, higher factor scores of the Fish [B (95% Confidence Interval), P-value; −0.020 (−0.035 – −0.006), 0.007], the Healthy snack [−0.023 (−0.043 – −0.002), 0.029], and the Modern [−0.027 (−0.050 – −0.004), 0.020] diet patterns significantly predicted lower serum LPS activity (Table 3, Model 3).

Table 2 Factor analysis-derived dietary patterns (n = 668).
Table 3 Associations between factor analysis-derived dietary patterns and serum LPS activity in 668 individuals.

Food record data

A total of 542 individuals (44% men, age 45 ± 13 years) completed the food record (Table 1). The median (interquartile range) energy [7694 (6468–9194) kJ)], carbohydrate [44.0 (39.3–48.1) E%)], fat [35.2 (30.9–38.9) E%)], protein [16.6 (14.8–18.5) E%)], and alcohol [1.1 (0–3.6) E%)] intake, of the population, was calculated. Of these dietary variables, carbohydrate intake correlated negatively with the serum LPS activity (r = −0.108, P = 0.012). Two models were constructed to study the association between macronutrient intake and serum LPS activity (Table 4). In the first model, dietary variables were entered as percentages of total energy intake, and in the second model as grams. Both models were additionally adjusted for energy and fibre intake, sample storage time, age, sex, smoking, physical activity, and diabetic nephropathy status. In these models, none of the dietary variables predicted serum LPS activity levels.

Table 4 Associations between dietary intake (energy, macronutrient, and fibre intake) and serum LPS activity in 542 individuals.

We then studied the association between LPS activity and macronutrient intake using the multivariable nutrient density substitution model (Table 5). In these analyses, neither the isoenergetic substitutions between macronutrient intakes, nor the substitutions between fatty acid intakes were associated with the LPS activity.

Table 5 Association between LPS activity and dietary macronutrient intake (substitution model) in 542 individuals.

Validation study

The basic characteristics of the study participants included in the validation study are shown in the Supplementary Table 1. The mean amounts of coffee, tea, milk, sour milk, yoghurt or curd, and bread, reported in the diet questionnaire and in the 6-day food record, were significantly correlated (P < 0.001, all) (Supplementary Table 2). Also the proportions of individuals reporting the consumption of filtered coffee, fat free milk, fat free sour milk, low fat yoghurt or curd, rye bread, and vegetable based spread were similar between the two methods (P < 0.001, all).

The mean consumption frequencies of the food groups reported in the FFQ part of the diet questionnaire and the 6-day food record are presented in Supplementary Table 3. The consumption frequencies of all the food groups were significantly correlated between the two methods (P < 0.001, all).

A total of six reasonably well matching dietary patterns were produced in the factor analysis from the FFQ and the 6-day food record (Supplementary Table 4). As the food items in the Cheese pattern (high-fat cheese and low-fat cheese) were observed to load in the opposite directions, in the two analyses, we applied inverse scoring for the food record-derived factor scores when classifying participants into the quartiles. For all the dietary patterns, the cross-classification to the same quartile, as well as to the same or adjacent quartile, was higher than expected by chance (Supplementary Table 5). Also the proportion of individuals misclassified to the opposite quartile was lower than expected by chance in all dietary patterns. Moreover, all factor scores of the six dietary patterns were observed to correlate significantly (Supplementary Table 4) (P < 0.001, all).

Finally, correlations of energy and macronutrient intakes between the two separate 3-day food records were investigated, and all variables were observed to correlate significantly (P < 0.001). Highest correlation was observed for the reported energy intake (r = 0.73), and lowest for the polyunsaturated fat intake (r = 0.43) (data not shown).


The potential role of bacterial endotoxins as proinflammatory mediators in the development of cardiovascular disease is increasingly recognized19. Notably, gram-negative bacteria, the source of LPS, are known to colonize the gastrointestinal tract of even apparently healthy individuals20. Indeed, the human gut is a major reservoir of LPS and, for these endotoxins, an important site of entry to the systemic circulation. Given that diet has the potential to modulate, not only the composition of the gut microbiota and its metabolic activity21, but also the integrity of the intestinal epithelium22, we hypothesized that dietary intake is associated with the level of endotoxaemia. In a population of individuals with type 1 diabetes, we observed that dietary patterns, reflecting food choices that are generally considered healthy, were associated with lower serum LPS activity. Amongst some of the food items, included in these dietary patterns, were fish, fresh vegetables, fruits and berries, yoghurt, pasta and rice, and poultry. Instead, no association was observed between dietary intake, measured at the level of macronutrients or fatty acids, and LPS.

Of the macronutrients, the role of dietary fats has gained particular interest in many of the previous LPS studies. Amongst other mechanisms, fat intake may induce endotoxaemia via enhanced intestinal absorption related to chylomicron formation3. High fat intake has also been shown to favour the numbers of LPS-containing Gram-negative bacteria in the gut at the expense of Gram-positive bacteria23. In human studies, an increase in postprandial endotoxaemia has been observed after a single high-fat meal14, 24. It could be speculated, however, that in the short-term feeding studies the amounts or proportions of fat consumed may be significantly higher than what is typically consumed as part of a regular diet. Other macronutrients, fibre, antioxidants and phytochemicals, present in the diet as a whole, may contribute to the complex associations that take place between the diet, gut microbes, and endotoxaemia. Finally, it has been suggested that total energy intake, rather than fat intake per se, confounds the previous observations related to dietary fat and endotoxaemia25.

Beyond studies focusing on specific macronutrient intake, only relatively few studies have been conducted on overall diet and endotoxaemia. In one such study, over a 6-month intervention with a standard Healthy Eating diet or a modified Mediterranean diet, no change in the plasma LPS-binding protein concentrations was observed, despite substantial increase in the intakes of fruits and vegetables and simultaneous reduction in fat intake18. The authors suspected that such interventions may show improvements in the LPS-binding protein levels only when instituted in populations with initially poor quality diets. In another study, a two-week intervention with 12 servings of fruits and vegetables per day, compared to period of 2 daily servings, exhibited a significantly lower ex vivo production of interleukin-6 in LPS-activated mononuclear cells26. As the two intervention diets were matched with respect to energy, macronutrient and fibre intake, the authors concluded that the effect was due to the difference in the phytochemical composition of the diets. In support of this hypothesis, isoflavone27 and quercetin28 consumption have been shown to produce beneficial responses in LPS-evoked endotoxaemia.

The current observations suggest the importance of an overall healthy diet in tackling the endotoxaemia. Many constituents of a healthy diet may contribute to the observed beneficial effects. For example, abundant intake of fibre may reduce LPS exposure not only via its ability to modify the microbial composition, but also through the production of short-chain fatty acids. Indeed, these end-products of microbial fermentation have shown to reduce the gut permeability by facilitating tight junction assembly29. Also, microbially produced butyrate stimulates the expression of intestinal alkaline phosphatase, which is involved in the regulation of chylomicron transport and detoxification of bacterial endotoxins30. Notably, in patients with type 2 diabetes, fibre supplementation exhibited a significant reduction in the LPS concentrations compared to the control group31. Consumption of fish, as part of a healthy diet may attenuate the disruption of the intestinal barrier function, as seen in studies of LPS-induced inflammation32, 33. Finally, potentially through the reduction in luminal cytotoxicity, lactobacilli growth and reduction of intestinal mucosal damage, dietary calcium appears to positively influence the composition of the gut microbiota and the intestinal barrier integrity34.

While our observations provide important insights into the association between dietary intake and endotoxaemia, the current study has also limitations that should be acknowledged. Importantly, the cross-sectional study design cannot give a definite answer to the question whether the diet affects LPS levels. Instead, our observations offer hypotheses that should be tested in intervention trials. The methods to study dietary intake have their unique limitations, as previously discussed35. Shortly, for a number of reasons, individuals may over- or under-report their dietary intake. Generally, such misreporting is thought to occur for food items perceived healthy and unhealthy, respectively. Should this have taken place, in the current study, it has most likely diluted the observations. Although the number of included participants was relatively large and the patients were well-characterised, some selection-bias may have taken place, which could limit the generalizability of the results. Indeed, compared to individuals volunteering to participate in a study, those declining may be less healthy or less health-conscious. Such selection-bias may have, for example, excluded proportionally higher number of individuals with diets perceived less optimal for health which, again, has a potential to dilute our observations. Finally, we cannot exclude that some residual confounding could explain our observations. One important confounder, not taken into account in the current study, is the potential impairment of the integrity of the gut. Some dietary constituents, such as gliadin36, has the potential to increase gut permeability which, again, could be associated with increased LPS absorption. Due to a small number of individuals adhering to a gluten-free diet, we were not able to address this issue in the current analyses.

Amongst the strengths of this study is that we measured dietary intake using both a macronutrient and a dietary pattern approach. With this dual approach, we were able to identify that diet, at the level of food items, is associated with serum LPS activity. However, dietary intake measured at the macronutrient level had no such association. Moreover, we investigated the role of macronutrient intake using two separate methods. First, all macronutrients were entered in a multivariable model in order to study whether their intake, as reported in the food records, was associated with the LPS activity. Next, we used the multivariable nutrient density substitution model, with a different rationale compared to the previous analysis. The idea behind the substitution model is that, by making different food-level choices, we can decrease (or increase) the intake of a particular macronutrient. In order to satisfy the energy requirement, however, one has to simultaneously increase (or decrease) the intake of another macronutrient. Indeed, the multivariable nutrient density substitution model takes into account this shift in the relative distribution of dietary macronutrients. Thus, we were able to calculate whether a five per cent increase in the consumption of one of the macronutrients, accompanied by a five per cent decrease in another macronutrient, had any effect on the level of LPS activity. While this is an appealing approach, occasionally used in the nutrition research37, 38, it has to be acknowledged that it is a mathematical model, and does not represent a real-life condition.

Finally the relative validity of the diet questionnaire, and subsequently the FFQ -section of the questionnaire, was investigated. The obtained results revealed that all the dietary variables investigated with the diet questionnaire and the 6-day food record were significantly correlated, and the questionnaire was able to correctly classify individuals in the formed dietary patterns. Moreover, data collected by the two separate 3-day food records, with a 2–3 month interval, showed significant correlation, suggesting that the no major changes in energy and macronutrient intakes take place in this population over this relatively short period.

In conclusion, our observations support the view that a healthy diet, as a whole, is associated with lower levels of endotoxaemia in type 1 diabetes. Instead, the habitual intake of energy, fibre, individual macronutrients or fatty acids has no relation to the serum LPS activity. The role of specific dietary patterns, in modifying endotoxaemia, should be investigated in an interventional setting. Moreover, whether dietary intake may, via its association with serum LPS levels, reduce the risk of diabetic complications, should be prospectively assessed.


The study subjects were participants in the Finnish Diabetic Nephropathy (FinnDiane) Study ( Since 1997, data from more than 5000 individuals have been collected. However, the collection of dietary data did not start until 2007. Therefore dietary data are available only from a subset of the FinnDiane Study population. Included in the current LPS-diet study were all participants in the FinnDiane Study with serum LPS activity measurements who had completed the diet questionnaire within two years from the study visit. The study protocol was approved by the Ethics Committee of Helsinki University Central Hospital, and the study was performed in accordance with the relevant guidelines and regulations. All subjects provided written informed consent prior to study participation.

During the regular visits to the attending physician, participants’ weight and height were measured, and body mass index (BMI, kg/m2) was calculated. Following a 10-minute rest, blood pressure was measured while seated. The second measurement followed after a 2-minute interval, and a mean of these measurements was calculated. Fasting (or following a light breakfast) blood samples were drawn and HbA1c was determined locally using standardized assays. Serum lipid and lipoprotein concentrations were measured as previously described39. Serum LPS activity was measured with the Limulus Amebocyte Lysate assay from 1:5 diluted serum samples at 405 nm as an end-point assay (LAL, Hycult Biotechnology, the Netherlands). Data on smoking and insulin dose were self-reported. The reported insulin dose was divided by body weight to obtain daily insulin dose in IU/kg. The attending physician recorded data on patient’s medication and diabetes complications on a standardized form. Data on physical activity were collected as previously described40.

Urinary albumin excretion rate (AER) in at least two out of three timed 24-hour or overnight urine collections was used to assess the participants’ renal status. The following classifications were made: normal albumin excretion rate (AER < 20 µg/min or <30 mg/24 h), microalbuminuria (AER ≥ 20 and <200 µg/min or ≥30 and <300 mg/24 h), macroalbuminuria (AER ≥ 200 µg/min or ≥300 mg/24 h), and end-stage renal disease (ESRD) (in dialysis or with kidney transplant). Diabetic nephropathy was defined as macroalbuminuria or ESRD.

Diet questionnaire

Dietary intake was measured using two separate methods, as previously described41. First, at the study visit, participants were provided a self-reported diet questionnaire to be completed. The food items included in the diet questionnaire and the types of data collected with the form are presented in Supplementary Table 6. The questionnaire also included a FFQ, where the consumption frequencies of a number of basic food items were queried (fish dishes, meat dishes, poultry, sausages and cold cuts, eggs, legumes, fresh vegetables, cooked vegetables, potatoes, pasta and rice, fruits and berries, high-fat cheese, low-fat cheese, yoghurt, ice cream, soft drinks, sweet pastries, sweets and chocolate, and fried and grilled foods). In this section, seven frequency response options were provided: several times per day, once a day, 4–6 times per week, 2–3 times per week, once a week, 1–3 times per month, and less frequently or never. In the current LPS-diet study, only data from the FFQ-section of the diet questionnaire are used. Here, the FFQ-derived data were used to form dietary patterns, representing the dietary intake at the level of food items, as described in the Statistical analyses in more detail. The factor scores of these dietary patterns were subsequently entered to the multivariable models to study the association between dietary patterns and serum LPS activity.

Food record

Upon returning the diet questionnaire, participants were sent a 3-day exercise and food record to be completed. In this record, participants reported all foods and drinks consumed during the three allocated consecutive days (two weekdays and one weekend day). The 3-day record-keeping was repeated after 2–3 months in order to take some seasonal variation into account. AivoDiet software (version, AIVO, Turku, Finland), based on the Finnish National Food Composition Database, was used to calculate the mean daily energy, macronutrient (both as grams and as percentages of total energy intake, E%), and fibre intakes from the records. The intake of trans fatty acids was estimated by subtracting saturated, monounsaturated, and polyunsaturated fatty acid intakes from the total fat intake. Subsequently, trans fatty acid intake was only used to adjust the fatty acid models, and not as an independent variable. Data collected with the food records were used in the multivariable models to study the association between dietary intake, at the level of energy and macronutrient intake, and serum LPS activity. In the current LPS-diet analyses, data from individuals who completed the record for a minimum of three days were included.

Validation study

For the validation study, we included data from all participants in the FinnDiane Study who had completed both the diet questionnaire and the 6-day food record (n = 1171). From the food record entries, data on all food items for which an unambiguous match was found in the diet questionnaire were extracted for subsequent analyses. For example, the numbers of cups of coffee, reported in the food record, were calculated and divided by the number of days the record was kept (six) to obtain the average of daily cups of coffee consumed. These values were then compared with the entries made in the diet questionnaires. Similarly, for the food record entries of fish dishes, meat dishes, poultry, and so on (the items listed in the FFQ part of the diet questionnaire), a daily average consumption frequency was calculated. Based on these calculated consumption frequencies, each food group was recoded according to the seven-point scale of the FFQ part of the diet questionnaire, to allow the comparison between the two methods.

Statistical analyses

Descriptive statistics are reported as percentages for categorical data, median (interquartile range) for non-normally distributed continuous data, and mean ± standard deviation (SD) for normally distributed continuous data. The respective statistical comparisons were performed using Chi squared test, Mann-Whitney U-test, and t-test.

Exploratory factor analysis (maximal likelihood and varimax rotation) was conducted to reveal underlying constructs within the FFQ in both the LPS-diet study and the validation study. In these analyses, the number of factors identified was based on eigenvalues >1.0, and food items with factor loading |≥0.20| with a particular factor, were included. The factor score was the sum of the scores for all items associated with that particular factor multiplied by its corresponding factor loading. In the validation study, the factor analysis was additionally conducted for the 6-day food record data. Within each dietary pattern, for which we could identify a reasonable match between the FFQ and the food record derived patterns, quartiles were formed based on the factor scores. The proportion of individuals classified into the same quartile, the same or adjacent quartile, or opposite quartile by the FFQ and the food record, was calculated.

The Spearman correlation coefficient was used to calculate the correlation between the factor scores of the matched dietary patterns in the validation study. Moreover, it was used to explore the correlations between serum LPS activity and dietary variables (both the factor scores of the FFQ-derived dietary patterns, and the energy, macronutrient, and fibre intake calculated from the food records). The independent associations between these dietary variables, and serum LPS activity were investigated in a multivariable generalized linear model.

Multivariable nutrient density substitution model was used to examine the associations between macronutrient intakes and LPS activity, as previously described42. In these analyses, LPS activity was entered as a dependent variable in the generalized linear regression analysis and, together with energy intake and selected confounders, dietary macronutrients (per 5 E%) were entered as independent variables. However one of the macronutrients, in each analysis, was left out. The results can be interpreted as an increase or decrease in LPS activity related to the isoenergetic (5 E%) substitution of a given macronutrient in the model with the macronutrient that was excluded from the model. As an example, in an equation: LPS activity = β0 + β1 (5 E% from carbohydrates) + β2 (5 E% from proteins) + β3 (5 E% from alcohol) + β4 (kcal), β1 can be interpreted as the change in the LPS activity when dietary carbohydrate intake is increased by 5 E% at the expense of fats, which was excluded from the model. Full models were corrected for sample storage time (as the LPS activity may reduce with increasing sample storage time), age, sex, smoking, physical activity, and diabetic nephropathy status. The substitution models with fatty acid intake were additionally adjusted for trans fatty acid intake.

A two-tailed P value < 0.05 was considered statistically significant. All data were analysed using IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp, Armonk, NY, USA).


  1. 1.

    Rhee, S. H. Lipopolysaccharide: basic biochemistry, intracellular signaling, and physiological impacts in the gut. Intest Res 12, 90–95 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  2. 2.

    Brun, P. et al. Increased intestinal permeability in obese mice: new evidence in the pathogenesis of nonalcoholic steatohepatitis. Am J Physiol Gastrointest Liver Physiol 292, G518–525 (2007).

    CAS  Article  PubMed  Google Scholar 

  3. 3.

    Kelly, C. J., Colgan, S. P. & Frank, D. N. Of microbes and meals: the health consequences of dietary endotoxemia. Nutr Clin Pract 27, 215–225 (2012).

    Article  PubMed  PubMed Central  Google Scholar 

  4. 4.

    Hawkesworth, S. et al. Evidence for metabolic endotoxemia in obese and diabetic Gambian women. Nutr Diabetes 3, e83 (2013).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  5. 5.

    Liang, H., Hussey, S. E., Sanchez-Avila, A., Tantiwong, P. & Musi, N. Effect of lipopolysaccharide on inflammation and insulin action in human muscle. PLoS One 8, e63983 (2013).

    ADS  Article  PubMed  PubMed Central  Google Scholar 

  6. 6.

    Manco, M., Putignani, L. & Bottazzo, G. F. Gut microbiota, lipopolysaccharides, and innate immunity in the pathogenesis of obesity and cardiovascular risk. Endocr Rev 31, 817–844 (2010).

    CAS  Article  PubMed  Google Scholar 

  7. 7.

    Pussinen, P. J., Havulinna, A. S., Lehto, M., Sundvall, J. & Salomaa, V. Endotoxemia is associated with an increased risk of incident diabetes. Diabetes Care 34, 392–397 (2011).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  8. 8.

    Gangarapu, V., Yildiz, K., Ince, A. T. & Baysal, B. Role of gut microbiota: obesity and NAFLD. Turk J Gastroenterol 25, 133–140 (2014).

    Article  Google Scholar 

  9. 9.

    Kallio, K. A. et al. Endotoxemia, nutrition, and cardiometabolic disorders. Acta Diabetol 52, 395–404 (2015).

    CAS  Article  PubMed  Google Scholar 

  10. 10.

    Lassenius, M. I. et al. Bacterial endotoxin activity in human serum is associated with dyslipidemia, insulin resistance, obesity, and chronic inflammation. Diabetes Care 34, 1809–1815 (2011).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  11. 11.

    Lassenius, M. I. et al. Endotoxins are associated with visceral fat mass in type 1 diabetes. Sci Rep 6, 38887 (2016).

    ADS  CAS  Article  PubMed  PubMed Central  Google Scholar 

  12. 12.

    Nymark, M. et al. Serum lipopolysaccharide activity is associated with the progression of kidney disease in Finnish patients with type 1 diabetes. Diabetes Care 32, 1689–1693 (2009).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  13. 13.

    Erridge, C., Attina, T., Spickett, C. M. & Webb, D. J. A high-fat meal induces low-grade endotoxemia: evidence of a novel mechanism of postprandial inflammation. Am J Clin Nutr 86, 1286–1292 (2007).

    CAS  PubMed  Google Scholar 

  14. 14.

    Deopurkar, R. et al. Differential effects of cream, glucose, and orange juice on inflammation, endotoxin, and the expression of Toll-like receptor-4 and suppressor of cytokine signaling-3. Diabetes Care 33, 991–997 (2010).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  15. 15.

    Dupuis, N., Curatolo, N., Benoist, J. F. & Auvin, S. Ketogenic diet exhibits anti-inflammatory properties. Epilepsia 56, e95–98 (2015).

    CAS  Article  PubMed  Google Scholar 

  16. 16.

    Lee, S. Y. et al. Effects of plant-based Korean food extracts on lipopolysaccharide-stimulated production of inflammatory mediators in vitro. Nutr Res Pract 8, 249–256 (2014).

    Article  PubMed  PubMed Central  Google Scholar 

  17. 17.

    Ghanim, H. et al. Orange juice neutralizes the proinflammatory effect of a high-fat, high-carbohydrate meal and prevents endotoxin increase and Toll-like receptor expression. Am J Clin Nutr 91, 940–949 (2010).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  18. 18.

    Umoh, F. I. et al. Markers of systemic exposures to products of intestinal bacteria in a dietary intervention study. Eur J Nutr 55, 793–798 (2016).

    CAS  Article  PubMed  Google Scholar 

  19. 19.

    Stoll, L. L., Denning, G. M. & Weintraub, N. L. Potential role of endotoxin as a proinflammatory mediator of atherosclerosis. Arterioscler Thromb Vasc Biol 24, 2227–2236 (2004).

    CAS  Article  PubMed  Google Scholar 

  20. 20.

    Berg, R. D. The indigenous gastrointestinal microflora. Trends Microbiol 4, 430–435 (1996).

    CAS  Article  PubMed  Google Scholar 

  21. 21.

    Rios-Covian, D. et al. Intestinal Short Chain Fatty Acids and their Link with Diet and Human Health. Front Microbiol 7, 185 (2016).

    Article  PubMed  PubMed Central  Google Scholar 

  22. 22.

    Bischoff, S. C. et al. Intestinal permeability–a new target for disease prevention and therapy. BMC Gastroenterol 14, 189 (2014).

    ADS  Article  PubMed  PubMed Central  Google Scholar 

  23. 23.

    Cani, P. D. et al. Selective increases of bifidobacteria in gut microflora improve high-fat-diet-induced diabetes in mice through a mechanism associated with endotoxaemia. Diabetologia 50, 2374–2383 (2007).

    CAS  Article  PubMed  Google Scholar 

  24. 24.

    Laugerette, F. et al. Emulsified lipids increase endotoxemia: possible role in early postprandial low-grade inflammation. J Nutr Biochem 22, 53–59 (2011).

    CAS  Article  PubMed  Google Scholar 

  25. 25.

    Amar, J. et al. Energy intake is associated with endotoxemia in apparently healthy men. Am J Clin Nutr 87, 1219–1223 (2008).

    CAS  PubMed  Google Scholar 

  26. 26.

    Yeon, J. Y., Kim, H. S. & Sung, M. K. Diets rich in fruits and vegetables suppress blood biomarkers of metabolic stress in overweight women. Prev Med 54(Suppl), S109–S115 (2012).

    CAS  Article  PubMed  Google Scholar 

  27. 27.

    Ferguson, J. F. et al. Dietary isoflavone intake is associated with evoked responses to inflammatory cardiometabolic stimuli and improved glucose homeostasis in healthy volunteers. Nutr Metab Cardiovasc Dis 24, 996–1003 (2014).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  28. 28.

    Chang, Y. C., Tsai, M. H., Sheu, W. H., Hsieh, S. C. & Chiang, A. N. The therapeutic potential and mechanisms of action of quercetin in relation to lipopolysaccharide-induced sepsis in vitro and in vivo. PLoS One 8, e80744 (2013).

    ADS  Article  PubMed  PubMed Central  Google Scholar 

  29. 29.

    Peng, L., Li, Z. R., Green, R. S., Holzman, I. R. & Lin, J. Butyrate enhances the intestinal barrier by facilitating tight junction assembly via activation of AMP-activated protein kinase in Caco-2 cell monolayers. J Nutr 139, 1619–1625 (2009).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  30. 30.

    Lallés, J. P. Intestinal alkaline phosphatase: multiple biological roles in maintenance of intestinal homeostasis and modulation by diet. Nutr Rev 68, 323–332 (2010).

    Article  PubMed  Google Scholar 

  31. 31.

    Dehghan, P., Gargari, B. P., Jafar-Abadi, M. A. & Aliasgharzadeh, A. Inulin controls inflammation and metabolic endotoxemia in women with type 2 diabetes mellitus: a randomized-controlled clinical trial. Int J Food Sci Nutr 65, 117–123 (2014).

    CAS  Article  PubMed  Google Scholar 

  32. 32.

    Liu, Y. et al. Fish oil enhances intestinal integrity and inhibits TLR4 and NOD2 signaling pathways in weaned pigs after LPS challenge. J Nutr 142, 2017–2024 (2012).

    CAS  Article  PubMed  Google Scholar 

  33. 33.

    Zhu, H. et al. Fish oil enhances intestinal barrier function and inhibits corticotropin-releasing hormone/corticotropin-releasing hormone receptor 1 signalling pathway in weaned pigs after lipopolysaccharide challenge. Br J Nutr 115, 1947–1957 (2016).

    CAS  Article  PubMed  Google Scholar 

  34. 34.

    Gomes, J. M., Costa, J. A. & Alfenas, R. C. Could the beneficial effects of dietary calcium on obesity and diabetes control be mediated by changes in intestinal microbiota and integrity? Br J Nutr 114, 1756–1765 (2015).

    CAS  Article  PubMed  Google Scholar 

  35. 35.

    Ahola, A. J., Freese, R., Mäkimattila, S., Forsblom, C. & Groop, P. H. Dietary patterns are associated with various vascular health markers and complications in type 1 diabetes. J Diabetes Complications 30, 1144–1150 (2016).

    Article  PubMed  Google Scholar 

  36. 36.

    Hollon, J. et al. Effect of gliadin on permeability of intestinal biopsy explants from celiac disease patients and patients with non-celiac gluten sensitivity. Nutrients 7, 1565–1576 (2015).

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  37. 37.

    Skilton, M. R., Laville, M., Cust, A. E., Moulin, P. & Bonnet, F. The association between dietary macronutrient intake and the prevalence of the metabolic syndrome. Br J Nutr 100, 400–407 (2008).

    CAS  Article  PubMed  Google Scholar 

  38. 38.

    Harris, C. et al. Dietary Fatty Acids and Changes in Blood Lipids during Adolescence: The Role of Substituting Nutrient Intakes. Nutrients 9, doi:10.3390/nu9020127 (2017).

  39. 39.

    Thorn, L. M. et al. Metabolic syndrome in type 1 diabetes: association with diabetic nephropathy and glycemic control (the FinnDiane study). Diabetes Care 28, 2019–2024 (2005).

    Article  PubMed  Google Scholar 

  40. 40.

    Wadén, J. et al. Leisure time physical activity is associated with poor glycemic control in type 1 diabetic women: the FinnDiane study. Diabetes Care 28, 777–782 (2005).

    Article  PubMed  Google Scholar 

  41. 41.

    Ahola, A. J. et al. Energy and nutrient intakes and adherence to dietary guidelines among Finnish adults with type 1 diabetes. Ann Med 44, 73–81 (2012).

    CAS  Article  PubMed  Google Scholar 

  42. 42.

    Ahola, A. J. et al. The association between macronutrient intake and the metabolic syndrome and its components in type 1 diabetes. Br J Nutr 1, 7 (2017).

    Google Scholar 

Download references


This study was supported by grants from Academy of Finland, Novo Nordisk Foundation (#NNF14SA0003), Signe and Ane Gyllenberg Foundation, Folkhälsan Research Foundation, The Helsinki University Central Hospital Research Funds, Wilhelm and Else Stockmann Foundation, Päivikki and Sakari Sohlberg Foundation, and Diabetes Wellness Finland. The skilled technical assistance of Anna Sandelin, Jaana Tuomikangas, Mira Korolainen, and Satu Kinnunen is gratefully acknowledged. The authors also acknowledge all the physicians and nurses at each centre participating in the collection of patients (online appendix).

Author information




A.J.A., M.I.L., C.F., V.H., M.L., and P.H.G. contributed to the conception and design, acquisition of data and interpretation of data. A.J.A. conducted the analyses and drafted the initial manuscript. M.I.L., C.F., V.H., M.L., and P.H.G. critically revised the manuscript, and gave final approval of the version to be published.

Corresponding author

Correspondence to Per-Henrik Groop.

Ethics declarations

Competing Interests

P.H.G. has received research grants from Eli Lilly and Roche, is an advisory board member for AbbVie, Boehringer-Ingelheim, Eli Lilly, Janssen, Medscape, Novartis, Novo Nordisk, and Sanofi. He has received lecture fees from Astra Zeneca, Boehringer-Ingelheim, Eli Lilly, Genzyme, M.S.D., Novartis, Novo Nordisk, and Sanofi. A.J.A. has received a grant from Diabetes Wellness Finland. All other authors declare that there is no conflict of interest.

Additional information

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Ahola, A.J., Lassenius, M.I., Forsblom, C. et al. Dietary patterns reflecting healthy food choices are associated with lower serum LPS activity. Sci Rep 7, 6511 (2017).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.