• Childhood obesity;
  • fructose;
  • fructose malabsorption;
  • multicentre study


  1. Top of page
  2. Summary
  3. Author contributions
  4. Conflict of interests statement
  5. Funding
  6. References


Studies in animals and in man have demonstrated that excessive consumption of fructose can cause all components of the metabolic syndrome.


To investigate the impact of a condition resulting in decreased absorption of fructose, on obesity.


In a multicentre study, we analyzed a cohort of paediatric patients with suspected primary fructose malabsorption (FM). Patients with chronic intestinal diseases were excluded. The final cohort comprised 628 patients.


302 patients were diagnosed with primary FM (48.1%). The proportion of obese patients was lower among FM patients, compared to non-FM patients (2.3 vs. 6.1%, P = 0.029). Logistic regression analysis with inclusion of various covariates showed that FM was negatively associated with obesity (OR 0.35, 95% CI [0.13; 0.97]). We discuss several mechanisms involving the metabolic, endocrine and gastrointestinal system.


Our data indicate that primary FM is negatively associated with childhood obesity.


body mass index


fructose malabsorption


hydrogen breath test


high-fructose corn syrup


interquartile interval


lactose intolerance


short-chain fatty acids


standard deviation score

Fructose malabsorption (FM) is defined as a reduced capacity to transport fructose across the intestinal epithelium. The absorption capacity varies widely between healthy individuals [1] and is influenced by age [2] as well as numerous dietary and hormonal factors [3]. In clinical practice, the diagnosis is based on increase of breath hydrogen and induction of typical symptoms (abdominal pain, diarrhoea, flatulence), following ingestion of a fructose meal with 1 g fructose per kilogram body weight [1]. A fructose-reduced diet completely reverses these symptoms. Because the capacity of fructose absorption is limited in most individuals, FM has also been suggested to be a frequent physiological variant [3]. Numerous studies in animals and in man (adults and children) have investigated the impact of nutritional fructose intake on body weight/obesity [4], as well as on metabolic [5, 6] and cardiovascular [5] comorbidities. In summary, excessive fructose consumption, either as sucrose or as high-fructose corn syrup (HFCS) [5], is an established risk factor for all components of the metabolic syndrome. The relationship between a condition resulting in a reduced capacity to absorb fructose and obesity has never been investigated.

We used a cohort of paediatric patients to analyze the relationship between primary FM and obesity. All patients had presented to our outpatient clinics with a history of suspected food intolerance, diarrhoea or abdominal pain. Inclusion criteria: (i) documentation of a hydrogen breath test (HBT) following lactose (2 g kg−1 body weight, max. 50 g) and/or fructose provocation (1 g kg−1 body weight, max. 25 g); (ii) complete anthropometric data including body weight and height; (iii) age at performance of HBT between 0 and 18 years. The diagnosis of FM or LI (lactose intolerance) was confirmed if breath hydrogen rose ≥ 20 ppm.

17 patients of the initial sample (n = 645) had to be excluded, due to either confirmation of a gastrointestinal disease potentially leading to secondary FM by medical history, lab tests and endoscopic exam (n = 14) or due to incomplete diagnostic procedures (n = 3). For the definition of childhood obesity, we used current national guidelines (AGA, 2009), i.e. a body mass index (BMI) > 97th age- and sex-specific percentile [7]. The study protocol was reviewed and approved by the Ethics Committee of the Justus-Liebig University Giessen. Data analysis was exploratory. Fisher's exact test was used to test homogeneity of both groups (FM yes/no) with regard to categorical variables. As the assumption of normal distribution could not be held, the distribution of continuous variables was described as median and interquartile interval. For comparison between the groups with and without FM, respectively, Mann–Whitney U-test was used. The potential influence of FM on being obese was examined by means of logistic regression analysis with inclusion of various covariates (sex, age, LI and an interaction term between FM and LI). For age, we defined two age groups: <10 years and ≥10 years. In the group ≥10 years, the rate of positive HBTs is equivalent to adults [2].

In this study, 302 of 628 patients (48.1%) were diagnosed with primary FM. Mean age was 9.3 years (range 0.6–17.8 years). Comorbidities were equally distributed between both groups (Table 1). Only 7 FM patients (2.3%), but 20 non-FM patients (6.1%), fulfilled the criteria for obesity (P = 0.029). BMI was 0.6 kg m−2 lower in FM patients (Hodges–Lehmann estimator, 95% CI [0.16; 1.03]). Despite the fact that the majority of patients enrolled in our study were normal weighted, BMI-SDS was 0.13 points lower in FM patients. Multiple logistic regression analysis for the risk of obesity in FM patients, adjusted for sex, age (as a categorical variable), LI and combined occurrence of LI and FM yielded an odds ratio of 0.35 (95% CI:[0.13;0.97]; Table 2). The effect was neither modified by LI alone nor by a combined occurrence of LI and FM.

Table 1. Clinical characteristics of FM and non-FM subjects
ParameterFM (n = 302)no FM (n = 326)P-value
  1. For age and anthropometric variables, median and interquartile intervals (IQ) are indicated in the first line. Mean values and standard deviations are indicated in the second line.

  2. For BMI, standard deviation score is given in the third line.

  3. For sex and comorbidities, numbers and percentages are indicated.

  4. **P < 0.01.

Sex  0.42
Boys131 (46.3%)152 (53.7%) 
Girls171 (49.6%)174 (50.4%) 
Age (years)8.8,IQ (5.9;11.5)10.3,IQ (6.3;13.0)0.009**
8.9 ± 3.69.7 ± 4.3
Anthropometric variables   
Body weight (kg)28.4,IQ (21.1;39.4)34.4,IQ (21.2;49.2)0.006**
32.0 ± 14.937.0 ± 19.6
Body height (cm)134.0,IQ (118.2;150.0)143.2,IQ (119.3;158.0)0.005**
133.3 ± 21.1137.2 ± 26.6
Body mass index (kg m−2) (BMI-SDS)16.2,IQ (14.8;18.3)16.8,IQ (15.0;20.0)0.007**
17.0 ± 3.218.1 ± 4.2
−0.16 ± 1.11−0.027 ± 1.18
Lactose intolerance48 (15.9%)39 (12.0%)0.17
Underweight18 (6.0%)20 (6.1%)1.00
Asthma15 (5.0%)24 (7.4%)0.25
Gastritis6 (2.0%)6 (1.8%)1.00
Food allergy5 (1.7%)6 (1.8%)1.00
Table 2. Logistic regression analysis for the risk of obesity depending on FM
VariableRegression coefficientP (two-sided)Odds ratio95% CI
  1. FM × LI, interaction term for a combined occurrence of FM and LI.

  2. *P < 0.05; **P < 0.01.

FM−1.0440.044*0.352[0.128; 0.971]
Age (categorical)−1.3400.005**0.262[0.103; 0.662]
Sex+0.2850.4761.329[0.608; 2.908]
Lactose intolerance (LI)−0.3260.6720.722[0.159; 3.273]
FM × LI+0.9680.4022.632[0.274; 25.271]

The finding has not been described in literature yet. It may be explained as follows:

  1. Individuals with FM may be partially protected from the detrimental health effects of excessive fructose consumption, i.e. increased insulin resistance and plasma triglycerides [5, 8, 9]. These negative health effects are due to unique properties of fructose metabolism [6, 8, 10].
  2. Fructose can constitute up to 10% of a person's daily energy intake. As a consequence, malabsorption of these energy-rich carbohydrates may result in an energy deficit. The amount of lost calories in FM due to unabsorbed fructose, or potential compensatory mechanisms, has not been investigated.
  3. We hypothesize that in primary FM, larger amounts of fructose reach the lumen of the large intestine, compared to non-FM subjects, favouring growth of a gut flora more favourable for weight maintenance.

Supporting our hypothesis, studies in animals and in man have shown that, between normal-weighted and obese individuals, distinct differences exist regarding the composition of the gut microbiome [11, 12]. In germ-free mice, transmission of gut microbiota from the caecum of ob/ob mice-fed Western diets resulted in a significantly greater increase of total body fat, compared to mice who had been transmitted bacteria from the guts of lean mice [12]. A recently published review by Payne et al. [13] shows a potential mechanism for fructose-induced interactions between host and gut microbiota: malabsorbed fructose is fermented to short-chain fatty acids (SCFA), which, in turn, alter colon transit time, lipid metabolism and glucose homeostasis. The metabolic consequences of increased SCFA production remain controversial; nevertheless, in mice, numerous beneficial effects have been documented, such as protection against diet-induced obesity [14] and insulin resistance [14, 15]. There is a limitation to our study: our sample is not representative of the general paediatric population because all subjects presented to our outpatient clinics with a history of suspected food intolerance or abdominal pain. Nevertheless, the prevalence of obese patients in the non-FM group is in line with current estimates for the whole German paediatric population [16] (6.3%). Further research is needed to show whether our results are applicable to adults. Also, the impact of FM on other obesity-related variables remains to be investigated. In conclusion, our exploratory study shows that, in a cohort of paediatric patients, primary FM is associated with a decreased prevalence of obesity.

Author contributions

  1. Top of page
  2. Summary
  3. Author contributions
  4. Conflict of interests statement
  5. Funding
  6. References

KPZ designed the study. AB collected the data. SD planned evaluation procedures and wrote the manuscript. SD and RHB did the statistical analysis. All authors critically reviewed the manuscript and approved the final version for submission.

Conflict of interests statement

  1. Top of page
  2. Summary
  3. Author contributions
  4. Conflict of interests statement
  5. Funding
  6. References

All authors declare that they have no conflicts of interest.


  1. Top of page
  2. Summary
  3. Author contributions
  4. Conflict of interests statement
  5. Funding
  6. References