These recommendations are based on evidence on the most common causes of tiredness in primary care, on evidence on investigations for tiredness, and on expert opinion.

Offer investigations as indicated by history and physical examination

• This is a pragmatic recommendation, also based on expert opinion from international guidelines [Godwin et al, 1999; BPAC NZ, 2006] and narrative reviews [Ridsdale, 1989; Rodriguez, 2000].

Delaying investigations until tiredness has lasted for 1 month

• This recommendation is based on limited evidence from a randomized trial that postponing blood tests does not appear to miss serious diagnoses (provided there are no red flags or other symptoms and signs indicating the need for immediate investigation), with few people returning for postponed investigations [Koch et al, 2009]. It is also supported by a New Zealand guideline [Godwin et al, 1999].

Choice of investigations for unexplained tiredness

• In the absence of large, prospective, randomized trials using reference standards for diagnoses and which are based in primary care, there is insufficient evidence to recommend withholding any investigations when the cause of tiredness is not identified clinically.
• There is limited evidence, from one randomized trial [Koch et al, 2009] and three observational studies [Sugarman and Berg, 1984; Ridsdale et al, 1993; Gialamas et al, 2003], that blood tests have low diagnostic yield in people presenting to primary care with tiredness or fatigue, detecting a physical cause in just 8–11% of people.
• Harms have not been fully evaluated. Possible harms of under-investigation and false-negatives include missing serious or treatable causes, and damage to the therapeutic relationship. Possible harms of over-investigation and false-positives include unnecessary anxiety and cost [Koch et al, 2009], and damage to subsequent efforts to address psychological or psychosocial causes of tiredness.
• Recommendations on the choice of initial, routine investigations are based on:
• Evidence from a randomized trial that a limited set of blood tests (including haemoglobin, erythrocyte sedimentation rate [ESR], glucose, and thyroid stimulating hormone) is almost as useful diagnostically as a more extensive set of tests [Koch et al, 2009].
• National Institute for Health and Clinical Excellence guidelines, which recommend tests for anaemia, thyroid function, and coeliac disease in people with persistent fatigue, even when there are no other indications [National Collaborating Centre for Primary Care, 2007; NICE, 2009].
• Evidence that, although physical causes are less frequently identified compared with psychosocial causes, anaemia and diabetes mellitus are the more common physical causes of tiredness in people presenting to primary care with tiredness as a complaint. However, these illnesses individually account for only 0.6–6% of the total number of people presenting with tiredness.
• Studies have identified malignancy in less than 1% of all people presenting to primary care with tiredness or fatigue [Valdini, 1985], but the rate may be higher in older people, who are poorly represented in aetiological studies.
• Consensus of expert opinion in the published literature that a full blood count is an appropriate first-line investigation [Rodriguez, 2000; Ebell and Belden, 2001; Sharpe and Wilks, 2002; Murtagh, 2003; BPAC NZ, 2006; Moncrieff and Fletcher, 2007; Rosenthal et al, 2008; Simon, 2008], with most also recommending ESR. C-reactive protein is recommended as an alternative to ESR on the basis of feedback from CKS expert reviewers. There was no consensus with regard to other investigations.
• Recommendations on the choice of additional investigations to consider
• Creatinine and electrolytes, liver function tests, bone biochemistry
• These investigations are recommended for older people on the basis of New Zealand and Canadian guidelines on the investigation of people with fatigue [Godwin et al, 1999; BPAC NZ, 2006]. The New Zealand guideline recommends additional investigations in people older than 50 years of age, whereas the threshold in the Canadian guideline is 65 years of age; 60 years of age is a compromise between these two opinions.
• Creatinine and electrolytes, and liver function tests were recommended by several CKS external reviewers for all people with tiredness/fatigue.
• There is evidence from one cohort study that fatigue is associated with non-alcoholic fatty liver disease (NAFLD) [Newton et al, 2008]. Fatigue in people with NAFLD was unrelated to severity of liver disease or to insulin resistance. People who were cases and controls were matched for body mass index, to prevent obesity confounding the relationship.
• Ferritin in women of child-bearing age
• This recommendation is based on evidence from one randomized, placebo-controlled trial that 4 weeks of iron supplementation reduced fatigue in women 18–55 years of age without anaemia who presented to primary care with fatigue [Verdon et al, 2003]. Only women with ferritin concentrations of 50 micrograms/L or less improved with treatment.
• Testing for vitamin D deficiency
• Although CKS found no published evidence on testing for vitamin D deficiency in people presenting with tiredness, this was recommended by several CKS expert reviewers. Risk factors for vitamin D deficiency are derived from an evidence-based narrative review [DTB, 2006].
• Testing for glandular fever in people younger than 40 years of age
• This recommendation is based on evidence that glandular fever is a relatively common cause of tiredness, and is detected even in people in whom it is not suspected [Morrison, 1980; Koch et al, 2009]. However, it is the cause of tiredness in only 0.6–6% of people (of all ages) presenting to primary care with tiredness.
• Glandular fever is rare in people 40 years of age or older [Petersen et al, 2006], and one prospective study which detected glandular fever in 1.4% of people used 40 years of age as a cut-off [Ridsdale et al, 1993].
• HIV test, hepatitis serology, and tuberculosis testing
• This recommendation is based on expert opinion from a New Zealand guideline [BPAC NZ, 2006] and on the CKS topic on Tuberculosis.