Article Text

Download PDFPDF

Dientamoeba fragilis and chronic abdominal pain in children: a case–control study
  1. Marin J de Jong1,
  2. Judith J Korterink1,
  3. Marc A Benninga2,
  4. Mirrian Hilbink3,
  5. J Widdershoven1,4,
  6. Judith M Deckers-Kocken1
  1. 1Department of Pediatrics, Jeroen Bosch Hospital, ‘s Hertogenbosch, Brabant, The Netherlands
  2. 2Department of Pediatric Gastroenterology & Nutrition, Emma Children's Hospital/Academic Medical Center, Amsterdam, The Netherlands
  3. 3Jeroen Bosch Academy, Jeroen Bosch Hospital, ‘s Hertogenbosch, The Netherlands
  4. 4Psychiatric hospital Herlaarhof, Vught, The Netherlands
  1. Correspondence to Judith Korterink, Department of Pediatrics, Jeroen Bosch Hospital, Henri Dunantstraat 1, 's Hertogenbosch, Brabant 5223 GZ, The Netherlands; judithkorterink{at}


Background The association between Dientamoeba (D.) fragilis and the aetiology of functional gastrointestinal disorders (FGID) in children is unclear.

Aim The aim of this retrospective case–control study is to clarify the clinical relevance of D. fragilis in children with chronic abdominal pain.

Methods From April 2011 until April 2013, a total of 132 patients with chronic abdominal pain (AP), aged 8–18 years, referred to a non-academic hospital, and 77 control patients, aged 8–18 years without gastrointestinal symptoms referred to a psychiatric hospital, were included in the study. D. fragilis was diagnosed by real-time PCR in faecal samples. Symptomatic children without a D. fragilis infection fulfilled the ROME III criteria for AP-related FGID (AP-FGID). Clinical data were retrospectively analysed by examining patients’ hospital records from the Jeroen Bosch Hospital and Herlaarhof in The Netherlands.

Results D. fragilis was detected in 57 patients with chronic AP (43.2%) and in 39 controls (50.6%) (p=0.255). No significant differences in symptomatology were found between D. fragilis-infected children and children fulfilling the criteria for AP-FGID. Parasitological eradication was achieved in 61.7% of patients after treatment with metronidazole or clioquinol, while clinical improvement occurred in only 40.4% of patients (p=0.435).

Conclusions There were no differences in symptoms comparing children with and without D fragilis infection. Furthermore, no relation was found between clinical and microbiological response after treatment for D. fragilis. This retrospective study suggests that there is no association between chronic AP and D. fragilis infection.

  • Epidemiology
  • Gastroenterology
  • Infectious Diseases
  • Microbiology
  • Parasitology

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

What is already known on this topic

  • Dientamoeba (D.) fragilis is a frequently encountered parasite in children suffering chronic abdominal pain (AP) as well as asymptomatic patients.

  • As D. fragilis infection and AP-related functional gastrointestinal disorders (AP-FGID) may present with similar symptoms, dientamoebiasis has been proposed as a pathophysiological mechanism underlying FGID.

  • The clinical relevance and, thereby, the need for treatment of D. fragilis infection in children suffering chronic AP remains unclear.

What this study adds

  • The prevalence of D. fragilis in children with chronic AP and asymptomatic controls was not significantly different.

  • No association was found between clinical and microbiological response after treatment.

  • This study suggests that D. fragilis infection is not associated with chronic abdominal pain.


With a prevalence of 15%, chronic or recurrent abdominal pain (AP) is a major problem in school-aged children in the USA and Europe resulting in school absence and frequent medical consultation.1–4 In the majority of children, no explanatory organic cause for their AP can be identified.5 Children suffer either from pain in the upper abdomen or from abdominal discomfort in the lower abdomen with or without an altered stool pattern. These different characteristics of AP are described in the Rome III AP-related functional gastrointestinal disorders (AP-FGID).6 The pathogenesis of AP-FGID remains unclear, although several mechanisms have been proposed, such as altered gut motility, visceral hypersensitivity, abnormal brain-gut interaction, psychosocial disturbance and immune activation.7 ,8 Another pathophysiological mechanism that has been proposed to underlie functional abdominal pain is an infection with Dientamoeba (D.) fragilis.9–13

D. fragilis is a flagellated protozoan found worldwide in the human gastrointestinal tract.11 ,14 ,15 Neither its epidemiology nor its transmission route is completely known.14 ,16 Prevalence rates vary widely from 0.4% to 52% depending on the population studied and diagnostic method used.11 ,14 ,15 D. fragilis infection is diagnosed most commonly at ages below 20 years, with a peak in children approximately 7 years of age.17–23 Probably due to adaptive immunity,23 children might be more susceptible to dientamoebiasis and, when infected, present with clinical symptoms, that is, AP and diarrhoea, more often than adults.17 ,19 ,24–26

Since the first description of this parasite in 1918 by Jepps and Dobell,27 its clinical relevance has been controversial. Most researchers assume D. fragilis as a pathogen9 ,11 ,16 ,17 ,19 ,21 ,22 ,25 ,28–30 because many studies show that in the setting of D. fragilis infection, gastrointestinal symptoms often subside after antimicrobial eradication.11 ,14 ,22 ,24 ,30 ,31 On the other hand, some consider it a commensal organism27 because D. fragilis is frequently found in asymptomatic individuals.26 ,31 Also, D. fragilis infection could have a self-limiting character because clinical improvement has been observed without treatment.26 ,32

Considering D. fragilis as a pathogen, by definition, the parasite should either be excluded or eradicated before the diagnosis of AP-FGID can be made. Therefore, elucidation of the clinical relevance of dientamoebiasis in children suffering AP-FGID is important.

For that reason, the first objective of the present study was to investigate the prevalence of D fragilis in a population referred for chronic AP and in a population without AP. The second aim of the study was to compare symptomatology between children with chronic AP and children fulfilling the AP-related Rome III criteria. Lastly, we investigated the association between chronic AP and the extent of D. fragilis infection, in terms of symptomatology, parasitical load and treatment.

Materials and methods

Patient population

This study is a retrospective case–control study performed at a non-academic centre, the Jeroen Bosch Hospital in The Netherlands. Since April 2011, the triple faeces test for detecting D. fragilis has been replaced in our hospital by the diagnostically superior PCR.33 ,34 All patients attending the outpatient clinic with chronic AP between April 2011 and April 2013 were eligible for inclusion when they were aged between 8 and 18 years, had symptoms of AP at least once per week for at least 2 months, and had no evidence of an inflammatory, anatomic, metabolic or neoplastic process that could explain the symptoms, that is, physical examination and laboratory tests were normal. Stool PCR on parasites was performed, and patients infected with Giardia lamblia, Entamoeba histolytica or Cryptosporidium species were excluded. Patients with a Blastocystis species were not excluded, because until now Blastocystis species has not been associated with clinical symptoms35 ,36 and, therefore, not considered as a pathogen. Patients referred to the psychiatric hospital Herlaarhof in Vught, The Netherlands, served as controls as they did not present with somatic symptoms. To reduce the risk of missing an underlying gastrointestinal complaint in all children suspected of a psychiatric problem, children were screened comprehensively by an experienced general paediatrician. A standardised history and a thorough physical examination were performed. Furthermore, blood and stools were routinely collected in all admitted patients, and were examined in the same laboratory as the children in the study group. Since no abnormalities were detected, these children were considered physically healthy and, therefore, suitable as a control group. Patients on antiparasitic treatment were excluded from the study.

Data collection

Clinical data were extracted from patients’ records. The following information was recorded: AP, loose stools, constipation, nausea, vomiting, ructus, flatulence, bloating abdomen, abdominal cramps, blood or mucus in stool, changed stool pattern, fatigue, weight loss, anorexia, fever, sleeplessness and other functional symptoms like headache, back pain or neck pain. Presence of eosinophilia, defined as peripheral blood eosinopilic leukocytes ≥0.4×109/L was recorded. Cycle threshold (Ct) values of parasitic DNA load before and after treatment were recorded to measure the quantity of D. fragilis in the colon. Antiparasitic treatment of choice, dosage and duration of treatment were noted. Clinical response was defined by patient or parental-reported improvement of AP and altered stool pattern. Microbiological response was defined by a negative stool PCR after treatment. Clinical and microbiological responses were evaluated approximately 2 months after treatment was finished.

Microbiological analysis

Laboratory detection of D. fragilis, Blastocystis species, Giardia lamblia, Entamoeba histolytica and Cryptosporidium species was performed by real-time PCR as previously described by Stark et al.34

Data analysis

The data were analysed using the Statistical Package for the Social Sciences (SPSS) V.19.0.0. For continuous variables, data were summarised as mean±SD. Categorical variables are expressed as percentages. With regard to the continuous variables, we first judged for fit to the normal distribution by using stem-and-leaf plots and quantile-quantile plots. Comparison of continuous variables was done by an independent samples t test for normally distributed data, or a Mann–Whitney U test for non-normally distributed data. As part of the t test analysis, the Levene's F test for equality of variances was used to test the assumption of equality of variances. For comparison of categorical variables, a logistic regression analysis, Pearson χ2 test or Fisher's exact test were used. All statistical tests were 2-tailed, and p<0.05 was considered statistically significant. For significant group differences at baseline, interaction was assessed. The interaction terms were formed by multiplication of the two parameters involved. When an interaction effect showed a p value <0.10, stratified analysis were performed.


Demographic characteristics

A total of 135 patients fulfilled the inclusion criteria of chronic AP. Three children were excluded because of infection with other parasites than D. fragilis and Blastocystis species.

Eighty patients without gastrointestinal symptoms, admitted to a psychiatric hospital, were eligible as controls, of which three were excluded because they were on antiparasitic treatment and, therefore, assumed to have gastrointestinal symptoms. As shown in table 1 significantly more males were included in the control group. Except for coinfection with Blastocystis species, no other pathogenous parasites were isolated from the stools.

Table 1

Patient characteristics of symptomatic and asymptomatic patients with and without Dientamoeba fragilis (DF) infection

Prevalence of D. fragilis infection

D. fragilis was demonstrated on stool PCR in 57 patients with chronic AP (43%), whereas, it was negative in 75 patients (57%). In the control group, D. fragilis was found in 51% of the cases. As shown in table 1, the prevalence of D. fragilis did not significantly differ between children with chronic AP and asymptomatic controls. As the groups showed a different gender distribution at baseline, a logistic regression analysis, including an interaction term (gender×setting) was performed. Regression analysis revealed a non-significant p value (p=0.465) for the interaction term.

Clinical characteristics of symptomatic patients

As summarised in table 2, presence of individual symptoms often related to AP-FGID such as bloating, nausea, altered defaecation frequency, constipation and flatulence were not significantly affected by presence of D. fragilis infection. Other functional complaints (ie, headache, back pain or neck pain) were significantly more found in children with AP-FGID and eosinophilia, and in children with chronic AP and a D fragilis infection. Fatigue (59.4%) was found the most prevalent symptom in the D fragilis group, followed by bloating (58.7%) and nausea (50.9%). Totally, 53.8% of the patients reported school absence more than once a month and 30.3% even once a week. No association was found in duration of symptoms between children with chronic AP and a D fragilis infection, and children suffering AP-FGID (p=0.406).

Table 2

Clinical features of symptomatic patients with and without Dientamoeba fragilis infection

Parasitic DNA load

There was no statistically significant difference regarding parasitic DNA load between symptomatic and asymptomatic infected children, Ct values were, respectively, 24.28 and 24.53 (p=0.830).

Microbiological and clinical response

A total of 52 out of 57 patients received antiparasitic treatment, of whom 39 patients were treated with metronidazol and 8 with clioquinol. For five patients, treatment data were not available. Of the patients treated with either metronidazole or clioquinol, 14 (35.9%) and 5 patients (62.5%) reported improvement of symptoms, respectively (p=0.163). In 23 patients treated with metronidazole (59.0%) and 6 patients treated with clioquinol (75%), there was a microbiological response, that is, PCR for D fragilis was negative after treatment (p=0.396). There was no association between clinical and microbiological response in patients with chronic AP treated for D fragilis regardless of which medication was used (p=0.435).


AP-related FGID (AP-FGID) can be considered a heterogeneous group of disorders, and D fragilis has been proposed as a possible mechanism underlying FGID. This retrospective case-control study, however, found no difference in the presence of gastrointestinal symptoms between D fragilis-infected children with chronic AP and children suffering from AP-FGID. Moreover, similar prevalence rates for D fragilis were found in children with chronic AP and healthy asymptomatic controls. Last, after treatment, only 40.4% of the D fragilis-infected children reached clinical improvement.

In this study, the prevalence of D. fragilis was 43.2% in chronic AP patients, and 50.6% in asymptomatic controls. Previous studies in symptomatic and asymptomatic children reported prevalence rates between 8% and 19.8%.18 ,20 ,22 ,37 The low sensitivity of light microscopy, as compared with PCR, might explain the high prevalence rates found in this study compared to earlier studies.34 Indeed, another study using the PCR technique as well, showed a comparable high prevalence among children and adults with gastrointestinal symptoms, with a peak among younger children;23 unfortunately, no prevalence rates among asymptomatic controls were investigated in this study.

By contrast with earlier studies, the presence of D. fragilis was not significantly associated with any gastrointestinal symptom.24–26 ,30 ,31 Moreover, this study demonstrated that based on symptomatology, no distinction could be made from children with AP-FGID. Not surprisingly, AP-FGID patients had significantly more other functional complaints, such as headache, back pain or neck pain. These latter findings are in line with the diagnostic criteria for childhood functional AP syndrome which further includes symptoms as limb pain, difficulty sleeping and some loss of daily functioning.6

This study showed that 50% of asymptomatic children were carrier of D. fragilis. Some researchers consider D. fragilis as a commensal since this parasite is indeed frequently found in asymptomatic individuals.26 ,27 The connection between presence or absence of disease could be due to different genotypes of D. fragilis, as has been described in other enteric protozoa, such as Giardia lamblia.11 Up until now, only one genotype for D. fragilis has been described in symptomatic and asymptomatic patients.14 ,32 ,33 Another explanation could be a difference in quantity of parasitic load in the colon. Hypothetically, a higher parasitic DNA load would be expected in patients with more clinical symptoms compared to those without symptoms. Our study, however, did not show a significant difference in the Ct values of parasitic DNA load between children with and without symptoms of AP.

Although a causal relation between D. fragilis and gastrointestinal symptoms is not clear, in daily clinical practice D. fragilis is frequently treated with antiparasitical medication. Numerous studies show clinical improvement following appropriate treatment.22 ,24 ,26 ,30 ,31 In our study, parasitological eradication was achieved in 61.7% after the first treatment with either metronidazole or clioquinol, but only 40.4% of patients reported clinical improvement. The eradication rate, as well as the observed clinical improvement percentage, are lower than reported in previous studies.22 ,30 ,31 This difference might be due to the different populations studied. In accordance with our study, Engsbro et al observed a clinical response after metronidazole, defined as adequate relief of symptoms, in 7 of 22 patients (32%), whereas, microbiological response was 68%. In a logistic regression analysis, however, the investigators were unable to show a significant association between clinical and microbiological response.12

Another retrospective study found an eradication rate of D. fragilis in 50% of affected children with AP. No significant difference in decrease in AP was, however, found between those children treated with metronidazole or tinidazole, and children receiving no treatment.26 Using a more heterogenous population (0–90 years), Vandenberg et al32 described a clinical success rate of 78.9% (12/19) in patients with and 71% (5/7) without treatment, suggesting a possible self-limiting character of the infection.

The strength of our study is the use of a well-defined AP -related FGID-population using the ROME III criteria and an asymptomatic control group. The limitations of this study are derived from its retrospective character. The description of the patients’ histories and clinical symptoms were sometimes subjective, which could have introduced some selection bias. On the other hand, such histories are quite representative for daily clinical practice. Furthermore, in some children with an underlying psychiatric disorder it might be difficult to obtain a reliable medical history. However, all patients and their parents were interviewed by an experienced general practitioner working in a general hospital as well as a psychiatric hospital, diminishing the risk of missing a gastrointestinal complaint.

In conclusion, our study showed that D. fragilis is a frequently encountered parasite both in children with chronic AP and in asymptomatic children. Additionally, presence of gastrointestinal symptoms did not differ between patients with chronic AP with D. fragilis infection, and AP-FGID patients without D. fragilis infection. Furthermore, no association was found between clinical and microbiological response after treatment. These findings suggest that the association between AP-related FGID and D. fragilis infection is doubtful. In keeping with this, well-designed large placebo-controlled studies in distinct subsets of AP-FGID-infected patients are warranted to establish clearly whether eradication of the D. fragilis infection improves specific AP-related symptoms.


The authors would like to thank T van Gool MD, PhD, from the Department of Microbiology, Parasitology Section, Academic Medical Centre, for his review of this manuscript.



  • MJdJ and JK contributed equally.

  • Contributors MJdJ and JJK wrote the proposal, conducted the research, wrote the manuscript and provided final approval. JW contributed to the design of the study. MH contributed by performing statistical analysis. JMD-K supervised the project, contributed to the study design, manuscript and provided final approval. MAB provided final approval.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.