Practice Diagnosis in General Practice

Acute diarrhoea in adults

BMJ 2009; 338 doi: (Published 15 June 2009) Cite this as: BMJ 2009;338:b1877
  1. Roger Jones, Wolfson professor of general practice1,
  2. Greg Rubin, professor of general practice and primary care2
  1. 1Department of General Practice and Primary Care, King’s College London School of Medicine, London SE11 6SP
  2. 2Wolfson Research Institute, School of Medicine and Health, University of Durham
  1. Correspondence to: R Jones roger.jones{at}

    A common condition provides an example of the use of the test of time in diagnosis, explained in the accompanying article by Susanna Almond and Nick Summerton (doi: 10.1136/bmj.b1878)

    Case scenario

    A 47 year old man consults because he has had diarrhoea for the past 10 days. Three weeks ago he travelled to Bangkok for an academic convention. He now has occasional cramping abdominal pain and feels generally off colour; he has not lost weight and has not noticed blood or mucus in his stools. He has taken loperamide for the past few days, without much benefit.

    The diagnostic dilemma

    The diagnostic dilemma is to distinguish between diarrhoea due to short term, probably infective, causes and diarrhoea that represents an early symptom of a more serious gastrointestinal disorder. Acute diarrhoea is somewhat arbitrarily defined as diarrhoea for less than four weeks.1 It is said to affect almost every adult in the United Kingdom every year,2 though most people do not consult a doctor about it.3 4 Viruses are the most common infectious cause in the community, and Campylobacter (12%) and rotavirus (8%) are the organisms most commonly isolated among people who consult a general practitioner,5 though isolation rates of norovirus have increased recently.6 Possible non-infective causes include drugs, alcohol, and anxiety as well as more serious problems such as inflammatory bowel disease and bowel cancer.

    In distinguishing between self limiting and more serious causes of acute diarrhoea, the doctor needs to know the patient’s background, including factors such as recent foreign travel; general health (comorbidities such as HIV infection, disorders such as hyperthyroidism and diabetes, and previous gastrointestinal surgery); “red flag” symptoms and signs, such as blood in the stool; recent hospital treatment or antibiotics; evidence of dehydration or intra-abdominal disease (a mass or marked abdominal tenderness, for example); and weight loss.

    The diagnostic approach: the test of time

    The investigation of acute diarrhoea in adults is a useful example of the test of time as a diagnostic approach in primary care.7 This approach is used in initial presentations of undifferentiated symptoms to discriminate the minority of patients with serious disease, who require urgent attention, from the majority with self limiting or less serious problems. In a careful initial assessment, red flag symptoms for serious disease and more subtle discriminators pointing to a range of other diagnoses are excluded. This assessment is followed by one or more structured, planned reassessments, whose nature and timing are informed by knowledge of the epidemiology and course of the presenting symptoms.

    Causes of acute diarrhoea

    Many cases of acute diarrhoea resolve within two to four days8; different infective agents cause slightly different durations of symptoms. For example, diarrhoea due to rotavirus lasts three to eight days, norovirus a median of two days, and Campylobacter and Salmonella diarrhoea two to seven days.9 Diarrhoea lasting over a week is often associated with Giardia.10

    Foreign travel and persistent symptoms raise the possibility of bacterial, viral, or protozoal infection, although in 25-50% of cases of “travellers’ diarrhoea” no specific causative agent is identified.11 Stool microscopy and culture for eggs and parasites are highly sensitive and specific,12 and will identify infection with amoeba and Giardia, as well as other gut pathogens, including Shigella, Salmonella, Yersinia, Campylobacter, and pathogenic Escherichia coli. If symptoms persist after infective agents have been identified and treated, chronic gastrointestinal disorders is more likely, including irritable bowel syndrome, lactose intolerance, inflammatory bowel disease, coeliac disease, and colorectal cancer.

    Red flags

    Red flags may have a number of meanings in patients with acute diarrhoea. They may represent the alarm symptoms of an early presentation of a serious gastrointestinal problem such as inflammatory bowel disease or colorectal cancer (change in bowel habit, rectal bleeding, weight loss, systemic illness, etc); they may represent symptoms indicating serious systemic complications (sepsis, dehydration, abdominal disease), or they may be regarded as “discriminators”—symptoms that discriminate between self limiting diarrhoea and infective diarrhoea, for which investigation and appropriate treatment are required.


    Duration of diarrhoea is important. Diarrhoea lasting for more than a week should lead to a decision to investigate, with the objective of identifying persistent infectious and non-infectious causes.10

    Travel history may raise suspicion of exposure to more exotic organisms. As at least 50% of travellers returning from non-European destinations, particularly central America and South America and the Indian subcontinent, are likely to have a bacterial or parasitic cause for their diarrhoea,13 this group also merit investigation.

    Blood in the stool may help in discriminating between non-self limiting infectious and non-infective colitis. Around 40% of patients with Campylobacter infection have blood in their stool (compared with 10% of those whose infections are due to other causes),8 and bloody stools are also common in infection with toxogenic E coli, Salmonella, Shigella, and Yersinia.10

    Fever is present in around half of patients with infective diarrhoea, particularly in Campylobacter and rotavirus infections, compared with around 10% in non-infective diarrhoea.10

    Headache seems to be most common in rotavirus infections,14 in which bloody stools are unusual, but it is a non-specific symptom and not a useful discriminator.

    C reactive protein is more likely to be raised in gastroenteritis with a bacterial, rather than a viral, cause. It is of less value in diagnosing inflammatory bowel disease because its concentration is commonly below 10 mg/l in patients with ulcerative colitis but is raised in 75% of patients with Crohn’s disease.15 16

    High risk groups include older people and those with comorbidities, in whom the test of time should be applied with caution, and those with other important health problems, such as HIV infection and immunosuppression, who should be investigated at an early stage.

    Problems with the test of time

    Potential pitfalls in the use of the test of time include inadequate initial assessment, the use of inapplicable discriminators (such as the presence or characteristics of abdominal pain), and failure to reassess patients adequately and objectively, according to Almond and Summerton.7 They point out the danger of inappropriate focus on a single organ system, which can lead to important physical or psychological disease being overlooked, and the danger of over-reliance on the accuracy of previous assessment made by others. In acute diarrhoea this translates into an emphasis on the importance of objective follow-up of progress or deterioration; being alert to the development of dehydration, sepsis, and systemic illness; and being aware of the possibility that the diarrhoea is related to a non-gastrointestinal disorder, such as hyperthyroidism or diabetes.

    Case reassessment: non-infective causes of diarrhoea

    In this patient, stool culture identified Giardia lamblia, which was treated with a course of metronidazole. Three weeks after treatment he returned as planned for reassessment, and still had diarrhoea, occasionally accompanied by cramping abdominal pain. Stool microscopy was now negative. With persistent symptoms, the likelihood of other gastrointestinal disorders is increased. The peak age for irritable bowel syndrome and inflammatory bowel disease is the third and fourth decades, and colorectal cancer is a possibility in a 47 year old man, so further investigations are needed to clarify the diagnosis.

    A tissue transglutaminase antibody (tTgA) test for coeliac disease is appropriate in primary care.17 A xylose tolerance test can identify lactose intolerance secondary to Giardia infection. If inflammatory bowel disease or colorectal cancer are suspected, a haemoglobin estimation, along with erythrocyte sedimentation rate or C reactive protein, may support the diagnosis, but lower bowel endoscopy (preferably colonoscopy rather than flexible sigmoidoscopy, to ensure a full bowel examination) will be needed if diagnostic suspicion is high, particularly if there is a family history of colorectal cancer.

    Case review

    Because of concerns about a serious underlying cause for the diarrhoea, and although the predictive value of acute diarrhoea for a diagnosis of inflammatory bowel disease or colorectal cancer is low,18 these investigations, including colonoscopy, were performed over the next three weeks. All had negative results, but the diarrhoea persisted. There was no suggestion of laxative misuse (a surprisingly common cause of persistent “factitious” diarrhoea19 20) or evidence of an underlying endocrine abnormality.

    This patient’s acute diarrhoea almost certainly came from an infection picked up while overseas; although Giardia was eradicated, symptoms persisted. Thorough investigation excluded non-malignant causes of persistent diarrhoea, such as coeliac disease, lactose intolerance, endocrine disorders, inflammatory bowel disease, and colorectal cancer. The final diagnosis was post-infective irritable bowel syndrome.

    Almost one in five patients with irritable bowel syndrome (particularly those with Campylobacter and Shigella infection) date the onset of symptoms to an episode of diarrhoea or acute gastroenteritis.21 This patient’s management proceeded on the basis of a clear explanation, an honest appraisal of prognosis, and the use of anti-diarrhoeal agents and occasional anti-spasmodics, with good effect.

    Learning points

    • Acute diarrhoea (lasting less than four weeks) has a range of causes, most commonly short term and usually self limiting viral, bacterial, and protozoal infections, but it can herald chronic gastrointestinal conditions, including colorectal cancer, or can be due to underlying systemic illness, including endocrine disorders

    • Use of the “test of time” in acute diarrhoea depends on knowing the epidemiology and course of the condition and provides a useful framework for management

    • Prompt investigation to identify or exclude an infective cause is important, and assessment based on the patient’s age, clinical history, and presenting symptoms and signs (checking for “red flags”) may suggest further investigation, particularly if symptoms persist

    • Acute diarrhoea has a low predictive value for inflammatory bowel disease and colorectal cancer, but persisting symptoms, particularly in older patients, merit prompt investigation

    • Post-infective irritable bowel syndrome is common: 20% of irritable bowel syndrome patients attribute the onset of their condition to an infective episode


    Cite this as: BMJ 2009;338:b1877


    • This series aims to set out a diagnostic strategy and illustrate its application with a case. The series advisers are Kevin Barraclough, general practitioner, Painswick, and research fellow in community based medicine, University of Bristol; Paul Glasziou, professor of evidence based medicine, Department of Primary Health Care, University of Oxford; and Peter Rose, university lecturer, Department of Primary Health Care, University of Oxford.

    • Contributors: RJ wrote the first draft; GR helped rewrite and finalise the paper.

    • Competing interests: None declared.

    • Provenance and peer review: Commissioned; externally peer reviewed.

    • Patient consent not required (patient anonymised, dead, or hypothetical).


    View Abstract

    Sign in

    Log in through your institution