Version 2.4

Please contact us with your comments, so we can improve the website.

These reviews were published in the Journal of Clinical Pathology and a supporting series of cases in the British Medical Journal. See Authors and acknowledgments.

This website is intended as an information source for healthcare professionals, particularly those working in primary care, and may also be of use to individual patients who wish to find out more about the tests they undergo. Patients can also obtain further information on Lab Tests Online (UK).

The way in which the questions are presented are described in About this site. Starting from the top of the list the reader can select the relevant questions and access, firstly the guidance and secondly the rationale behind the guidance. The site does not attempt to cover every individual situation which may occur, but rather to provide core information on some of the most common topics.

Please choose an antenatal testing topic:

The recommendations for normal pregnancy given in this article are based largely on the guideline entitled “Antenatal care: routine care for the healthy pregnant woman” published in October 2003 [NICE, 2003a], commissioned by the National Institute for Health and Clinical Excellence (NICE) from the National Collaborating Centre for Women’s and Children’s Health.

The ethos of the current guideline is that pregnancy is a normal physiological process. Any interventions offered (including laboratory tests) should have known benefits and be acceptable to pregnant women.

The guideline also stresses the importance of communicating the purpose of tests and informing women of all results.

Some women will require additional care because of pre-existing medical conditions or risk factors for complicated pregnancy (see Table 1). This article does not address how to identify or manage these individuals. The merits of screening normal, healthy women for a number of conditions are not clearly established and this article highlights some areas where uncertainty remains.

We recommend the following:

All the following tests should be offered at the first antenatal appointment and if accepted, arranged before 16 weeks of pregnancy:

All the following tests should be offered at the first antenatal appointment and if accepted, arranged before 16 weeks of pregnancy:

This is offered and explained at the first antenatal visit. If accepted, ultrasound assessment is carried out at 11-14 weeks and serum testing 14-20 weeks. This will not be considered further as it is the subject of individual national screening policies and practices. New developments are being evaluated in the UK, in particular, serum testing with different marker combinations at an earlier stage in pregnancy (11-13 weeks). The current UK policy position from the National Screening Committee may be accessed at the NLH Screening Specialist Library website (www.library.nhs.uk/screening).

Pre-eclampsia is a multi-system disorder affecting between 2-10% of pregnancies and results in increased maternal and neonatal morbidity and mortality. It is defined as hypertension new to pregnancy manifesting after 20 weeks of gestation associated with a new onset of proteinuria, which resolves after delivery. Hypertension new to pregnancy without proteinuria and resolving after delivery is termed pregnancy-induced hypertension.

Reagent strip (dipstick) urinalysis is sufficient for initial testing but is prone to observer error. Strips must be used within the expiry date and according to manufacturers’ instructions. Automated reading devices significantly reduce both false positive and false negative rates but add to the cost [Shennan and Waugh, 2003]. An initial result of >= 1+ should be confirmed by laboratory measurement of protein/creatinine ratio on a random urine sample. A protein/creatinine ratio of 30 mg/mmol is regarded as significant [Ferrazzani et al, 1990; Waugh et al, 2007].

No other biochemical tests are necessary. In particular, universal screening for gestational diabetes mellitus (GDM), including dipstick testing for glycosuria (see also topic on Glucose measurements), is not currently recommended either by NICE [NICE, 2003a] or the UK National Screening Committee [UK National Screening Committee, 2006] because it meets only some of the well established criteria for such a programme [Scott et al, 2002]. This field is beset by uncertainty over definitions, diagnostic tests and optimal clinical management. The World Health Organization (WHO) defines GDM as any degree of carbohydrate intolerance with onset or first recognition in pregnancy, applying criteria from non-pregnant subjects to identify “abnormal” blood or plasma glucose levels. Unanswered questions remain:

The Australian Carbohydrate Intolerance Study (ACHOIS) [Crowther et al, 2005] has reopened the debate in the UK as it provides evidence that active intervention in women with relatively mild impairment of glucose tolerance results in improved perinatal outcomes. The Scottish SIGN guideline [SIGN, 2007] recommends screening in pregnancy in urine at every antenatal visit. It recommends a random venous plasma glucose if 2+ glycosuria is detected, and routinely at 28 weeks gestation. It adds that the WHO advise that a 75 g oral glucose tolerance test (OGTT) should be carried out if the blood glucose is > 5.5 mmol/L two hours or more after food, or > 7 mmol/L within two hours of food.

In addition the American Diabetes Association guidance [American Diabetes Association, 2005] recommends that testing may not be needed in women with no risk factors for developing GDM, but testing all others.

Recent guideline development suggest that an approach of screening routinely at 28 weeks with targeted screening of higher risk patients (Table 1) at booking or in the first trimester may emerge (Professor R Bilous, personal communication, Feb 2007).

Further valuable data is expected from the Hyperglycaemia and Adverse Pregnancy Outcome Study (HAPO) [HAPO Study Cooperative Research Group, 2002]. This is a large, epidemiological study of a heterogeneous, ethnically diverse cohort of 25,000 women designed to clarify the relationship between adverse pregnancy outcomes and various levels of glucose intolerance less severe than overt diabetes mellitus.

There is no consensus on which screening test would be most appropriate for a universal programme and a variety of policies currently prevail in UK obstetric practice [Scott et al, 2002]. It is to be hoped that a more coherent national approach can be recommended in the anticipated update to NICE guidance which is due to be published in 2007.

It should be noted that in practice urine protein dipstick tests frequently also measure glucose and that this level of screening facility is probably therefore in widespread use although is insensitive compared to fasting plasma glucose or OGTT.

This is to identify any possible transfusion problems which might arise, anticipate the need for anti-D prophylaxis in RhD negative women and to determine the risk of other types of haemolytic disease of the newborn.

FBC will detect anaemia (low haemoglobin), but also less commonly white cell and platelet abnormalities. The normal UK haemoglobin in pregnancy is accepted as >= 110 g/L at up to 12 weeks, and >= 105 g/L at 28-30 weeks. A significantly abnormal result will dictate further investigations, for example ferritin if haemoglobin and mean cell volume (MCV) are reduced. Serum ferritin is the most sensitive single screening test to detect adequate iron stores with a sensitivity of 90% at a cut-off of 30 µg/L.

There are also separate topics on Lymphocytes, neutrophils, eosinophils, Platelets, high haemoglobin and Iron deficiency anaemia.

An NHS Sickle Cell and Thalassaemia screening programme for pregnant women commenced in April 2005 [NHS Antenatal and Newborn Screening Programmes, 2004; UK National Screening Committee, 2006]. All pregnant women in Trusts defined as high prevalence (sickle cell disease expected to affect more than 1.5 per 10,000 pregnancies) should be offered antenatal screening. Women in low prevalence Trusts (sickle cell disease expected in less than 1.5 per 10,000 pregnancies) should, from April 2006, be offered screening based on family origin and formal inspection of blood count indices.

Testing for rubella susceptibility identifies women at risk of contracting infection and who need vaccination in the postnatal period to protect future pregnancies. Although syphilis is relatively rare in the UK, treatment as early as possible is beneficial to mother and fetus. Antenatal intervention in HIV positive women and postnatal intervention in hepatitis B positive women reduces mother to child transmission. Tests for all four of these infectious diseases can be carried out on a single blood sample, and there are well-established pathways for dealing with positive screening results of serological tests.

There are also separate topics on Mumps and chickenpox and EBV infection, mononucleosis, pregnancy exposure to rash.

All pregnant women should be offered routine screening for asymptomatic bacteriuria by midstream urine culture at their first antenatal visit. It may be helpful to repeat samples if contamination is suspected, If however there is a pure or predominant growth of 105 organisms per ml, the woman should be treated with an appropriate antimicrobial agent [NICE, 2003a]. A further urine culture should be performed as a test of cure and again at regular intervals (monthly) for the remainder of gestation.

The prevalence of asymptomatic bacteriuria in pregnancy is about 5%, the same as in non-pregnant women. During pregnancy there is a dilatation of the ureters and renal pelvices with decreased ureteric peristalsis, changes beginning as early as the seventh week of gestation [Sobel and Kaye, 2000]. This predisposes to infection and about a third of pregnant women with untreated bacteriuria develop acute pyelonephritis [Berrington and Bint, 1999]. Infection may also be complicated by low birth weight and pre-maturity, pre-eclampsia, maternal anaemia, amnionitis and intra-uterine death [NICE, 2003a]. More controversial is whether screening is cost effective. This depends partly on the nature of the population being screened (affluent populations have a lower prevalence of bacteriuria). The two most widely used strategies for diagnosing bacteriuria are the use of leucocyte esterase-nitrite dipsticks and quantitative urine culture. Urine culture is more expensive, but has a higher sensitivity and specificity and is therefore the recommended test [NICE, 2003a; Nicolle et al, 2005] (see also topic on Infection - urinary).

Based on the original work of Kass in 1957 [Kass, 1957,] bacteriuria has traditionally been diagnosed on the basis of a pure or predominant growth of 105 organisms (Enterobacteria, “coliforms”) per ml of a properly taken clean catch or mid stream specimen of urine [Berrington and Bint, 1999]. Kass recommended that positive cultures should be repeated for confirmation on the basis that two positive results increase the post-test probability of true bacteriuria from 80 to 95%. However, workers have relied on single samples [Berrington and Bint, 1999] and NICE currently recommend treatment based on the result of a single screening test [NICE, 2003a].

A midstream urine culture should be requested routinely at first antenatal visit. A urine culture should be performed 7 days after antibiotic treatment as a test of cure, and at regular intervals (monthly) for the remainder of gestation, as women whose bacteriuria fails to respond to treatment are at highest risk of developing symptomatic infection [Sobel and Kaye, 2000].

There is also a separate topic on Infection - urinary.

Antenatal screening for the Group B Streptococcus (GBS) is not currently recommended in the UK. Pregnant women with risk factors for early-onset neonatal GBS disease may be offered intrapartum chemoprophylaxis; therefore GBS cultured incidentally from antenatal samples, including urine samples, should be reported by the laboratory.

GBS is the most frequent cause of severe early-onset neonatal infection in the UK, causing septicaemia, pneumonia and meningitis. The incidence is estimated to be 0.5/1000 births and approximately 10% of infections are fatal [NICE, 2003a; RCOG, 2003]. GBS may also cause maternal infections such as amnionitis, endometritis or septicaemia. Risk factors for early-onset GBS disease are intrapartum fever, prolonged rupture of membranes, delivery at < 37 weeks gestation and a previous infant with GBS disease. GBS is carried asymptomatically by approximately 25% of pregnant women in the UK. In the USA, where the incidence of GBS disease was previously three times that in the UK, interventions have been associated with a reduction in GBS disease [CDC, 2002]. In the USA, a universal culture-based screening programme is used to detect GBS carriage, using both vaginal and rectal swabs at 35-37 weeks gestation. Intrapartum antimicrobial prophylaxis is given to GBS-colonised women and those with risk factors for GBS disease, supported by the use of algorithms [CDC, 2002]. The effectiveness of screening in preventing early-onset neonatal GBS disease has been estimated at 50-80% based on observational studies [NICE, 2003a; RCOG, 2003]. Since the introduction of universal screening in the USA in the 1990s, the incidence of early-onset GBS disease has fallen to a similar level to that seen in the UK.

In the UK, the NICE has recommended that women should not be offered routine antenatal screening for GBS because evidence for clinical and cost effectiveness remains uncertain [NICE, 2003a]. Screening and prophylaxis strategies have not demonstrated an overall effect on all-cause neonatal sepsis or neonatal mortality [RCOG, 2003]. However, since approximately 60% of early-onset GBS cases have risk factors for GBS, the Royal College of Obstetricians & Gynaecologists currently recommends that pregnant women should be considered for intrapartum prophylaxis on the basis of risk factors, rather than culture-based screening [RCOG, 2003].

Bacterial vaginosis (BV) is the most common cause of vaginal discharge and malodour. BV results from a change in the normal flora of the vagina with a relative overgrowth of anaerobic bacteria. The condition is not sexually transmitted, but is associated with sexual activity. The presence of BV during pregnancy is around 10-20% and 50% of these women are asymptomatic. BV is associated with pre-term birth: women with BV are twice likely to deliver pre-term than women without BV. There is no evidence that screening and treating asymptomatic women who are not at high risk improves outcomes such as pre-term labour or birth [US Preventive Services Task force, 2001; NICE, 2003a]. Further studies are required to define the role of screening and treating for BV in pregnant women who have experienced previous pre-term delivery, as treatment may reduce pre-term birth in these women.

There is also a separate topic on Vaginal discharge and Chlamydia testing.

In conclusion, a limited number of investigations are recommended for a newly pregnant woman who is healthy (Table 2). These should be offered and explained at the first antenatal visit.

Table 2 and Table 3 are adapted from the practice algorithm in [NICE, 2003a].

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

Please choose an allergy topic:

This short question series examines the use of total and allergen specific IgE, mostly in the context of investigation of allergy. The documents found provide fairly close consensus and several comment on the need to identify the appropriate situations for their use. There appears to be increasingly limited need for the measurement of total IgE, particularly in the investigation of allergy.

There is also a topic on Lymphocytes, neutrophils, eosinophils.

We recommend there is very limited need for this test in general practice, except as an adjunct to the diagnoses listed below.

IgE concentrations can be increased in non-allergic states such as parasite infections, Churg-Strauss vasculitis, certain immune deficiencies such as the hyper-IgE syndrome, IgE myeloma, Hodgkin lymphoma and atopic dermatitis [Sicherer, 1999]. These are all rare conditions, usually diagnosed in secondary care, and the measurement of IgE is not crucial to the diagnosis.

Total IgE alone can neither confirm nor exclude allergy: increased concentrations may suggest an atopic state, but normal or low values do not exclude allergy [Saarinen et al, 1982; AAAAI, 2000]. Normal or low IgE concentrations also cannot be used as a prescreening test for radioallergosorbent testing (RAST) [WHO, 1981; AAAAI, 2000]. Therefore, it is inappropriate to request total IgE in isolation from general practice in this context.

We recommend allergen specific IgE measurement in the presence of clinical suspicion of type 1 IgE mediated hypersensitivity/allergy, principally for inhaled antigens. There is no need to request total IgE when requesting RAST.

There is limited consensus guidance on use of RAST testing in particular, as distinct from allergy testing in general, and the guidance above is drawn principally from review articles and by extrapolation from clinical studies.

RAST refers to one of the first tests used to test allergen specific IgE, which is no longer in use; a more appropriate name is allergen specific IgE testing. It is used as a more accessible or more convenient alternative to skin prick testing [RCP and Royal College of Pathologists, 1994]. It is only useful for assessing type I IgE mediated reactions (immediate hypersensitivity); RAST tests are therefore not useful for assessing pseudo-allergic reactions that are not mediated by IgE (such as non-allergic food intolerance; reactions to radiocontrast media, morphine, aspirin; physical urticarias, etc.) [WHO, 1981; Kay and Lessof, 1992; RCP, 2003]. Angio-oedema without urticaria is usually not an IgE mediated allergic reaction [American College of Allergy and Immunology Board of Regents, 1991]. RAST testing and skin prick testing are of little value in chronic urticaria, which is usually not caused by IgE dependent mechanisms.

RAST tests must be requested for a specified antigen based on clinical history [WHO, 1981; American College of Allergy and Immunology Board of Regents, 1991; Kay and Lessof, 1992; RCP, 2003]. They are of no benefit as screening tests without specified antigens [American College of Allergy and Immunology Board of Regents, 1991; Kay and Lessof, 1992; RCP, 2003]. It follows from this that requests should not be for widespread antigen screening. Test results must be interpreted in conjunction with clinical findings [WHO, 1981; American College of Allergy and Immunology Board of Regents, 1991; Kay and Lessof, 1992; Sicherer, 1999; RCP, 2003].

The specificity and sensitivity of RAST results vary for different allergens tested (for example, poor for fruits and vegetables). Overall, RAST tests have relatively low sensitivity and can be negative in the presence of allergy [Saarinen et al, 1982; RCP and Royal College of Pathologists, 1994; National Asthma Education and Prevention Program, 1997; Li, 2002; Volcheck, 2001]. In addition, adverse reactions to foods are IgE mediated allergies in only about a third of patients [WHO, 1981; National Asthma Education and Prevention Program, 1997; Volcheck, 2001]. Therefore, RAST tests are of limited value in this situation. It follows from this that RAST testing in food intolerance is unlikely to be helpful, and we recommend that it used only in the initial investigation of severe acute food intolerance reactions where a specific food is suspected.

The efficacy of unconventional/alternative allergy testing has not been confirmed, and can neither substitute or complement RAST and other classic allergy tests [Kay and Lessof, 1992; RCP, 2003]. Therefore, these tests are not recommended.

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

Please choose an anaemia topic:

Tests used in the diagnosis of iron deficiency anaemia have changed in recent years, and there remains some variability between the tests that laboratories offer and recommend. Their interpretation can also be complicated in the presence of acute or chronic inflammation. The guideline base is relatively limited, although consensus exists and laboratories are recommended to review the tests they offer and recommend for the investigation of iron deficiency.

There is also a topic on Vitamin B12 and folate deficiency.

We propose that in patients with anaemia, as the mean cell volume (MCV) falls, the probability of iron deficiency increases, although no specific cut-off can be used. Even in patients with MCV below 75fl, only 38% will have iron deficiency [Guyatt et al, 1990]. In patients with anaemia and MCV above 95fl, there is a low probability of iron deficiency being present. Other causes should be investigated initially. Ferritin is superior to iron and iron-binding capacity or transferrin saturation [Guyatt et al, 1992].

There is also a topic on Laboratory investigation of chronic diarrhoea.

All patients with anaemia should be assessed to estimate the presence of iron deficiency. The history should include:

Similarly, other causes of anaemia such as chronic inflammation or underlying malignancy, can also be assessed prior to testing.

There is also a topic on Faecal occult blood testing in adults with bowel symptoms.

A serum ferritin concentration of < 15ng/ml confirms a diagnosis of iron deficiency and a ferritin of > 100 ng/ml rules out the diagnosis [Guyatt et al, 1992]. For patients with chronic iron deficiency, ferritin levels between 15 and 40 ng/ml provide sufficient evidence for the diagnosis to be made and treatment to be commenced. For patients with chronic inflammation, the same conclusion can be reached for ferritins up to the level of 70 ng/ml. For ferritins above this level, however, further assessment is required before a diagnosis can be reached, as ferritin levels rise with acute or chronic inflammation. A response to iron replacement treatment is definite confirmation of deficiency.

We recommend that remeasurement of ferritin is not necessary. Recovery of iron deficiency anaemia is assessed from haemoglobin levels, after 3 weeks, to confirm response, and 9 weeks to confirm recovery once the source of iron deficiency has been identified and corrected.

Once oral iron replacement treatment has been started, the haemoglobin level should rise by about 2 g/dl every 3 weeks. Therefore, it is suggested that a blood count is repeated after 2-4 weeks of treatment to make sure there is a response [Guidelines & Protocols Advisory Committee, 2004; CKS, 2005c]. A further full blood count should be performed at 2-4 months [Guidelines & Protocols Advisory Committee, 2004] to ensure that the haemoglobin level has returned to normal. Iron replacement should then be continued for 3-6 months once the haemoglobin level has normalised, to replenish iron stores (3 months, 4-6 months [Guidelines & Protocols Advisory Committee, 2004]). Once the diagnosis of iron deficiency has been confirmed, its cause of the iron deficiency will need to be established [Goddard et al, 2000].

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

There is little information from large scale consensus statements or large clinical trials. The guidance below is derived from a small number of reviews supplemented by extrapolations from knowledge of the physiology of vitamin B12 and folate. These answers do not address the further investigation of the cause of vitamin B12 or folate deficiency.

There is also a topic on Iron deficiency anaemia.

The following are generally agreed indications for the measurement of vitamin B12 and folate concentrations:

Note that coexisting conditions such as iron deficiency or thalassaemia trait may mask the development or presence of macrocytosis.

Macrocytosis in itself may or may not be preset with anaemia but those patients who have a macrocytosis with anaemia should have vitamin B12 and folate measurements performed [Lancet and Rapoport, 1999; CKS, 2005g].

The level of macrocytosis can predict the probability of vitamin B12 and folate deficiency being present. As the MCV increases above 100fl, so the probability of vitamin B12 and folate deficiency increases. This is particularly true of patients with an MCV above 130 fl, except for those receiving hydroxyurea. The probability of B12 deficiency is less with MCVs between 100 and 110 fl, which is more likely to be related to other causes of macrocytosis, such as alcohol abuse, liver disease, anti-neoplastic drugs, human immunodeficiency virus infection, and also haematological disorders, such as myelodysplastic syndromes or haemolytic conditions [Lancet and Rapoport, 1999; Schrier, 2004].

Blood film comments help to point to a diagnosis; hypersegmented neutrophils and macro-ovalocytes are associated with vitamin B12 and folate deficiency, a uniform macrocytosis with alcohol abuse, target cells with liver disease, polychromasia with haemolysis.

Several neuropsychiatric abnormalities have been described in association with vitamin B12 deficiency (including paraesthesia, ataxia, peripheral neuropathy and memory loss). These may occur in the absence of either anaemia or macrocytosis. Objective signs associated with vitamin B12 deficiency include impaired vibration, touch, pain and position sense, together with an abnormal gait [Stabler et al, 1990].

A number of cases of B12/folate/iron deficiency present with severe oropharyngeal ulceration without changes in the blood count [Porter et al, 1988; Field et al, 1995] and respond to vitamin B12 treatment [Wray et al, 1975].

Co-existing conditions such as iron deficiency (see topic on Iron deficiency anaemia) or thalassaemia trait may mask the development of macrocytosis and therefore, in patients who have anaemia without a raised MCV, dual conditions should be considered as a possible diagnosis, and vitamin B12 and folate should be measured if the origin of the anaemia cannot be established by initial tests [Lancet and Rapoport, 1999].

There is also a topic on Laboratory investigation of chronic diarrhoea.

We recommend:

Vitamin B12 and folate deficiency: full blood count measurement at 10-14 days to document a rise in the haemoglobin and a fall in MCV and a further check in 8 weeks to document that the blood count has returned to normal [Schrier, 2004]. No further monitoring is required once a full haematological response is achieved.

There is very little published guidance on the monitoring of vitamin B12 or folate deficiency.

Long term folate replacement is considered unnecessary in most cases [BNF 45, 2003] the deficiency responds to short term treatment unless the cause (e.g. malnutrition) persists.

Patients with pernicious anaemia receiving vitamin B12 replacement should by definition not become vitamin B12 deficient and further monitoring would seem unnecessary in most instances although practices vary and some haematologists would recommend annual full blood counts. There is no obvious merit in repeating vitamin B12 and folate levels during replacement unless lack of compliance (specifically folate) is suspected or anaemia recurs.

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

Please choose an anticoagulant monitoring topic:

The great variation in response to warfarin between and within patients dictates the need to monitor and maintain the desired therapeutic international normalized ratio (INR) in order to ensure the effectiveness and safety of warfarin. It is also important to realise that careful selection and continuing risk assessment of patients with regard to hazards and benefits of anticoagulation is as important as INR monitoring.

Patient self-monitoring of anticoagulant therapy may be restrictive because of (i) exclusion rates (> 60%, range 31-88%) based on trials evaluated [Heneghan et al, 2006] and (ii) the higher cost associated with self-monitoring, with one study [Fitzmaurice et al, 2002] quoting £90 as the cost for primary care monitoring compared to £425 per patient/year for the self-monitoring group. The pressure on primary care to continue providing anticoagulant monitoring [Fitzmaurice et al, 2002] is therefore likely to continue.

We recommend:

The various indications for oral anticoagulation and their respective target INR are summarised in the third edition (2005) update of the British Committee for Standards in Haematology Guidelines (BCSH) on oral anticoagulation (warfarin) [Baglin et al, 2006]. A recommendation for a target INR of 2.5 is made for most indications; the exceptions are a target INR of 3.0 for mechanical aortic valve prosthesis and 3.5 for mechanical mitral valve prosthesis and recurrence of venous thromboembolism while on warfarin therapy.

The efficacy and safety of warfarin is critically dependent on maintaining the INR in the therapeutic range. The aim of achieving target INR is to maximise reduction in thromboembolic risks while minimising the associated haemorrhagic risk. In patients who are at very high risk of bleeding, their optimal target INR may have to be lowered, sacrificing some efficacy for safety [Ansell et al, 2004].

Several regimens are available to guide clinicians in initiating and predicting maintenance dose of warfarin in individual patients [Baglin et al, 2006]. For patients who require rapid induction of oral anticoagulation (e.g in acute venous thromboembolism when patients will also be on heparin), the most commonly used regimen is the Fennerty regimen [Fennerty et al, 1984] (Table 1) which recommends a starting warfarin dose of 10 mg and daily INR monitoring during at least the first 4 days of treatment; after the desired therapeutic INR is achieved, weekly INR monitoring is recommended until the control is stable and the frequency of recall can then be extended [British Committee for Standards in Haematology, 1998]. Other regimens are available for situation when heparin may not be required and when less rapid or more cautious induction of warfarin anticoagulation is considered appropriate or desirable (e.g in chronic atrial fibrillation or elderly patients); such regimens are discussed in the BCSH guidelines on oral anticoagulation [Baglin et al, 2006]. One example is that of Tait & Sefcick regimen [Tait and Sefcick, 1998], (Table 2) which recommends an initial dose of 5 mg warfarin for the first 4 days and INR check on day 5 and day 8 to predict subsequent warfarin dose. Both Fennerty and Tait & Sefcick regimens rely on a normal baseline INR, defined as < 1.4, at induction. The Tait and Sefcick regimen and the other slow-loading regimens [Baglin et al, 2006] are suitable for initiating warfarin anticoagulation in patients with atrial fibrillation in the community. As patients with acute thromboembolism are increasingly being managed in the community, primary care physicians may also become involved in the use of the Fennerty regimen.

Apart from a demonstrable stable dose response to warfarin, the frequency of long-term INR monitoring depends on individual patient's characteristics, which include drug compliance, change in drug history and other co-morbid conditions. While the recommended frequency of INR monitoring for stably controlled patients in North America is no less than every 4 weeks [Ansell et al, 2004], intervals of up to 12 weeks is accepted or recommended practice [British Committee for Standards in Haematology, 1998; BNF 51, 2006] in the UK; longer intervals of at least 14 weeks have even been suggested [Lidstone et al, 2000].

While anticoagulant monitoring should be managed by trained personnel, computer-assisted dosage adjustment can also help to achieve better INR control [Poller et al, 1998; Ansell et al, 2004]. To help with dosing, Baglin [Baglin, 1998] refers to a simple algorithm for calculating a new dose of warfarin based on the degree of over-anticoagulation; he also provides a validated simple dose reduction scheme for management of over-anticoagulation (Table 3). The BCSH guidelines [British Committee for Standards in Haematology, 1998; Baglin et al, 2006] provide details on management of bleeding and excessive anticoagulation with warfarin; hospital admission will be required for some groups of patients.

We recommend:

Management of warfarin drug interactions is complicated by either lack of good quality information regarding drug interactions [Wittkowsky, 2001; Holbrook et al, 2005] or significant post-marketing drug interactions being reported in spite of negative pre-marketing drug interaction studies [Wittkowsky, 2001]. However, the number of reports of interactions between warfarin and drugs or foods is increasing, which reflects widespread use of anticoagulant therapy and its use with concomitant medications [Holbrook et al, 2005]. A similar challenge applies with herbal products [Wittkowsky, 2001]. Self medication with over-the-counter vitamin K1-containing multivitamins is also problematic [Kurnik et al, 2004]. A regular and detailed history of pharmacological and non-pharmacological medications from patients on warfarin is therefore always essential.

It is wise to assume that almost any drug can interact with oral anticoagulants [British Committee for Standards in Haematology, 1998]. When prescribing new medication for patients on warfarin, clinicians are strongly advised to refer to appropriate sources of information on drug interactions, such as the British National Formulary [BNF 51, 2006], available on-line (www.bnf.org/bnf), which list the drugs with known predictive pharmacological interactions with warfarin resulting in over or under-anticoagulation; if such drugs cannot be avoided, more frequent INR monitoring will be required and the dose of warfarin adjusted as necessary. On-line resources for herbal product drug interactions are also available from the National Institutes of Health (Office of Dietary Supplements http://dietary-supplements.info.nih.gov) and National Center for Complementary and Alternative Medicine (http://nccam.nih.gov). Moreover, while accepting some limitations, it would appear sensible to choose a drug that is least likely to interact with warfarin, based on knowledge of drug metabolism or negative result in drug interaction studies. For example, citalopram has the lowest general risk of interaction with warfarin amongst the selective serotonin reuptake inhibitors [Baglin, 1998] and rabeprazole would be preferred over omeprazole in choosing a proton pump inhibitor [Holbrook et al, 2005].

The decision as to when to check the INR when a new drug is prescribed depends on the duration of therapy and if the drug is known to interact with warfarin. For short courses of new drug therapy, dose adjustment of warfarin is not necessary, although a slight dose reduction or omission of one dose could be considered if a known potentiator is prescribed [British Committee for Standards in Haematology, 1998]. If the new drug is prescribed for more than 5 days, the INR should be checked 1 week after starting and the dose of warfarin adjusted accordingly [British Committee for Standards in Haematology, 1998]. Likewise, monitoring to a new maintenance dose of warfarin may be required when a drug is discontinued.

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

Please choose an arthritis, inflammatory topic:

These questions consider 2 common clinical scenarios in primary care: diagnosis and monitoring of rheumatoid disease in small joint pain (Dr PC Dore and Dr R Herriot), and the utility of the human leukocyte antigen (HLA) B27 in the investigation of low back pain (Dr WSA Smellie and Dr GA Spickett). They highlight the limitations of both of these tests as diagnostic markers, and the need for a clinical filter designed to distinguish inflammatory from other mechanical joint pain prior to further investigation. In particular, they also demonstrate that repeat measurement of RhF appears to have no role in patients with rheumatoid arthritis.

We recommend measurement of RhF:

Rheumatoid factor (RhF) can be measured by a number of differing assays with variation in quantitation, sensitivity and specificity.

The type of assay used will determine its usefulness. Agglutination assays are less sensitive and are mainly used for screening purposes. Nephelometric and ELISA assays are more sensitive and provide a quantitative measure of RhF. Your local laboratory will be able to provide further information on the assay used.

There is agreement that whilst RhF is present in 70-90% of patients with rheumatoid arthritis, its absence does not exclude rheumatoid arthritis although these patients have a milder disease [Shmerling and Delbanco, 1991; van Zeben et al, 1992; van Schaardenburg et al, 1993].

A positive RhF at low titre is found in a variety of diseases, particularly immunological disorders, infectious diseases and haematological malignancies [Shmerling and Delbanco, 1991; van Zeben et al, 1992; van Schaardenburg et al, 1993].

RhF is used as a criterion in the American Rheumatism Association guidelines for the diagnosis of rheumatoid arthritis, but the assay variability means that there is a requirement for any assay to be positive in < 5% of a control population [Arnett et al, 1988; Shmerling and Delbanco, 1991].

When the assay is performed at its optimum, it achieves a sensitivity of 80% and specificity of 95%. It is significantly affected by the pretest probability. The overall prevalence of rheumatoid arthritis in the population is 0.8%. Using an assay which complies with the American Rheumatism Association guidelines, this gives a positive predictive value of RhF of 10%; it is therefore unsuitable for population screening [Shmerling and Delbanco, 1991; van Zeben et al, 1992].

In instances where an inflammatory arthopathy is present (raised ESR (see also topic on Erythrocyte sedimentation rate) or CRP, joint swelling or damage) the positive predictive value rises to > 75% and the negative predictive value is around 90%, making the test a useful diagnostic aid, providing an appropriate assay method is used [Shmerling and Delbanco, 1991; van Zeben et al, 1992].

The prevalence of Rhf increases with age, and 10-20% of people > 65 years will be seropositive [Shmerling and Delbanco, 1991; van Zeben et al, 1992].

The titre of the RhF is, in addition, an important predictor of severity of rheumatoid arthritis, and high titres (> 100 IU) are associated with the development of radiographic erosions and extra-articular manifestations [van Schaardenburg et al, 1993; van der Heide et al, 1995; Mottonen et al, 1998; Bukhari et al, 2002].

Other antibody subtypes (IgG and IgA) have not found a place in the routine investigation of rheumatoid arthritis. There are reports of more severe disease associated with the presence of IgA RhF [van Zeben et al, 1992; Jonsson and Valdimarsson, 1993].

RhF is useful in the diagnosis of rheumatoid arthritis when inflammatory arthropathy is present but cannot be used as a screening test. It provides prognostic information on the severity of the disease and the likelihood of erosions and extra-articular manifestations. Other subtypes have not proved useful in managing the disease, but further studies are required.

We do not recommend use of RhF in monitoring rheumatoid disease

The measurement of RhF varies depending on the assay used. The only quantitative assays statistically consistent enough for monitoring purposes would be nephelometry and ELISA methods.

However, no studies are convincing enough to show that even using a precise assay the level of RhF has more than a weak association with the disease process [Wernick et al, 1983; Scott et al, 1987; Mikuls et al, 2004].

There is a slow change in RhF in association with disease-modifying drugs such as penicillamine. This drop, although important, does not correlate with changes in ESR or joint score [Wernick et al, 1983; Scott et al, 1987; Mikuls et al, 2004].

Possible reasons for this poor correlation are:

There has been a report of the loss of RhF before the development of lymphoma in patients with Sjogren's syndrome, but further studies are required before this use can be recommended [Anderson and Talal, 1972].

RhF therefore adds no information in the monitoring of rheumatoid arthritis, and alternative assays such CRP as a marker of inflammation should be used if a marker of disease activity is needed.

We recommend:

The human leukocyte transplantation antigen B27 is known to be associated with ankylosing spondylosis and is present in about 90% of white and 50% of black patients with ankylosing spondylosis and other spondyloarthritides (SpA), many of which may represent early forms of AS [Hawkins et al, 1981; Wener, 2001]. However, it is insensitive as a test alone (20% sensitivity), as many people with the B27 antigen do not develop ankylosing spondylosis and B27 has a high prevalence in many populations.

Whilst radiological evidence of sacroileitis is reported to be the most specific and consistent finding in ankylosing spondylosis, this may be absent in early disease and in SpA [van der Linden et al, 1984].

Several recommendations for B27 testing suggest restricting its use to atypical patients with symptoms of inflammatory back pain associated with radiological sacroileitis, or in patients without radiological sacroileitis otherwise likely to have ankylosing spondylosis (eg uveitis and back pain) [Baron and Zendel, 1989; Amor et al, 1990; Shojania, 2000; Khan, 2002].

The presence of inflammatory back pain used by the European Spondyloarthropathy group (ESSG) [Dougados et al, 1991] and based on Calin's questionnaire [Calin et al, 1977] can be defined by 3 or more of the following features:

A recent review of published findings for ankylosing spondylosis and SpA found that presence of inflammatory back pain appears to raise the average background probability of SpA from 5% to 14% in patients with chronic back pain [Rudwaleit et al, 2004]. When combined with positive B27 measurement, the positive predictive value for SpA was found in this review to be to 54%, rising to 90% and over if other clinical features (anterior uveitis, enthesitis) are present. This appears to have reopened the debate about the targeted use of B27 in clinically equivocal situations in suitable populations [Sheehan, 2004; Rudwaleit et al, 2005; Sieper and Rudwaleit, 2005].

However, the permutations of clinical scenarios and population differences render this test of limited use in a primary care context, particularly as the principal discriminator for SpA is inflammatory back pain.

Patients with SpA respond to the same appropriate anti-inflammatory and other newer specific treatments as those with ankylosing spondylosis. There is still, however, no clear evidence that early identification of patients with SpA, estimated to represent potentially 5% of patients with primary care chronic back pain [Underwood and Dawes, 1995], will lead to improved long term outcome, although accurate identification will help to focus treatment.

Use of the test is further limited by the fact that pretest and post-test probabilities also vary greatly between populations, however, because of large differences in both the prevalence of B27 in populations and the prevalence of B27 amongst patients with spondylarthropathy [Khan, 2002].

The test is comparatively expensive and is not suitable for use as a screening or diagnostic test.

Testing for HLA B27 in primary care is not currently appropriate and discussion with a specialist rheumatologist is advisable if a spondyloarthropathy is suspected. B27 testing if performed would normally form part of secondary care assessment [CKS, 2005e].

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

Please choose a blood count abnormalities topic:

This second series of blood count scenarios examines selected abnormalities of white cell counts, namely lymphocytosis, neutropenia and eosinophilia.

As a typical full blood count report may contain 10 or more results, values outside of quoted reference ranges occur frequently on a statistical basis. These questions and answers attempt to establish thresholds for clinical action or referral and identify situations that are likely to be of clinical importance. Some (e.g. lymphocytosis) are extensively reviewed in existing guidelines, cited as the primary reference sources, whereas other, such as eosinophilia, draw guidance mostly from extrapolation from observational studies.

We recommend the following criteria for referral:

The primary aim of this answer is to establish the need for referral of patients with possible leukaemia or lymphoma. Patients may be found to have a lymphocytosis in the course of routine investigations for unrelated symptoms or as part of health screening. The cause may be an increase in T lymphocytes that is often reactive to an acute illness and is rarely a reflection of malignancy [British Committee for Standards in Haematology et al, 2004]. Alternatively, a B cell lymphocytosis may be present, which may be polyclonal but more often is the result of a clonal lymphoproliferative disorder, usually CLL. Other conditions presenting with a lymphocytosis such as follicular lymphoma, marginal zone lymphoma, mantle-cell lymphoma or hairy-cell leukaemia, will often have clinical features such as anaemia, splenomegaly or lymphadenopathy, and the blood film appearances may not be compatible with CLL [British Committee for Standards in Haematology et al, 2004]. Although there is limited clear guidance, it would seem reasonable to wait for an acute viral illness to resolve and recheck the lymphocyte count when a lymphocytosis is associated with features of acute viral illness.

A definitive diagnosis of CLL is based on the combination of a lymphocytosis > 5 x 109cells/L and a characteristic lymphocyte morphology and immunophenotype. Immunophenotyping is required to accurately classify the nature of the lymphocyte proliferation, thus enabling an appropriate treatment plan to be made. Although there are some morphological features associated with the different types of lymphocyte proliferation, these are no longer acceptable as the only means of confirming the diagnosis that will influence the patient's management. Immunophenotyping is always indicated in patients requiring treatment, in patients with lymphocytosis, that on morphological review is not typical of CLL, and in patients in whom it is felt important to exclude a reactive lymphocytosis [British Committee for Standards in Haematology et al, 2004]. Blood films should be prepared for patients who have lymphocytosis > 5 x 109cells/L when reviewed for the first time [Barnes et al, 2005].

Patients with a lymphocytosis should be referred for a review by a haematologist if they have lymphadenopathy, splenomegaly, anaemia, thrombocytopenia, or where the blood film reports a lymphocytosis that is not consistent with CLL [British Committee for Standards in Haematology et al, 2004], or when the lymphocytosis is not explained clinically by an acute self-limiting viral illness (Table 1).

The management of patients with early (stage A) CLL requires a collaborative approach between primary and secondary care. Some patients in stage A are regarded as having 'smouldering CLL' characterised by a haemoglobin level > 13g/dl, lymphocyte count < 30 x 109/L, minimal or no lymphadenopathy and a lymphocyte doubling time > 12 months, and these patients have a low progression rate (15% at 5 years, 80% 10-year survival). By contrast, patients with stage B or C disease have a 40% 5-year survival and require early treatment [CLL Trialists' Collaborative Group, 1999].

By extrapolation from the guidelines of the British Committee for Standards in Haematology [British Committee for Standards in Haematology et al, 2004], patients with indolent stage A CLL could be monitored in the primary care setting with a FBC every 6 months for the first 2 years. If the lymphocyte count doubles during this time, then the patient should be referred to a haematologist for an assessment. For patients whose lymphocyte count remains stable for this time, there is a low likelihood of the disease progressing, and monitoring can be reduced to an annual FBC. The follow-up of the patients seen initially in hospital who do not require treatment may be organised in primary care, in hospital outpatients or through a home-care service, depending on local resources and patient wishes. Before patients are discharged from the hospital follow-up, a clear management plan should be documented, which should include criteria for re-referral to the haematology service.

Neutropenia is potentially associated with life-threatening infection and we recommend that the following situations are of significance and require referral to secondary care:

Neutropenia is classified as:

Patients with neutrophil count < 1 x 109cells/L and fever require urgent, parenteral, broad-spectrum antibiotics, as infection may rapidly progress to established septic shock.

We found no published guidance for referral of patients with mild neutropenia. This guidance is drawn from a consensus of author and reviewer opinion.

Benign ethnic neutropenia is relatively common in individuals of African descent (neutrophil counts down to 1.0 x 109cells/L) and is also seen in some of Middle Eastern extraction. Individuals are physically normal and lack a history of susceptibility to infection. Confirm neutropenia with repeat FBC and confirm normal morphology with blood film [Haddy et al, 1999].

Transient neutropenia not lasting > 2 weeks is usually related to viral infections and not associated with clinical problems [Lakshman and Finn, 2001]. Occasionally, these infections may contribute to mild neutropenia for several months after the illness.

Most severe neutropenias are associated with fever, oral ulceration and bacterial or fungal infections [Boxer and Dale, 2002; Bhatt and Saleem, 2004]. The following areas should be reviewed:

Recurrent fever and oral ulcerations can be due to cyclical neutropenia [Lange, 1983], a rare autosomal dominant disorder in which neutrophil counts oscillate between 0.1 x 109cells/L and 1.5 x 109cells/L every 21 days. Neutropenic periods last 3-6 days accompanied by malaise, anorexia, fever, lymphadenopathy and mucosal ulceration. The diagnosis is established by obtaining neutrophil counts twice weekly for a minimum of 6 weeks, but this should not hinder referral of severe neutropenia to secondary care [Boxer and Dale, 2002].

We recommend the following if a patient's eosinophil count is (persistently) > 0.35 x 109cells/L or > 1. 5 x 109cells/L.

Clinical history concentrating on allergy/atopy, gastrointestinal, skin, respiratory and joint symptoms, and any change in general health (malignancy) combined, depending on clinical context, with:

Non-specific allergy testing is not recommended (see also topic on IGE (Total and Allergen Specific) measurement).

The upper limit of the reference range for the eosinophil count is > 0.35 x 109cells/L [Spry, 1989]. Eosinophilia has been subdivided into mild (0.35-1.5) moderate (1.5-5.0) and severe (> 5.0 x 109cells/L) [Rothenberg, 1998]. In a review of 1862 cases studied in an Italian series [Lombardi and Passalacqua, 2003], the most common causes and frequencies (which include potential drug reactions in several categories) can be summarised as follows:

These patients were identified from detailed clinical history, and an escalating and extensive series of investigations were performed until a diagnosis was established. A further 2.7% of patients were defined as having eosinophilia of unknown significance.

An additional group of patients has been identified with idiopathic hypereosinophilic syndrome [Fauci et al, 1982], which by definition excludes all patients with eosinophilia in which a cause can be found [Brito-Babapulle, 2003]. A North American study of incidental eosinophilia [Brigden and Graydon, 1997] (defined as eosinophil count of > 5% or >0.7 x 109cells/L in this study) in 195300 automated haematology profiles found that of the 225 cases (0.1% of samples), almost all were attributable to either allergic processes or to known underlying diseases (advanced malignancy, connective tissues disease). Only 2 cases (patients receiving gold therapy) were both unanticipated and not associated with signs or symptoms of the cause of the eosinophilia. In 30% of cases, no cause was found by the patient's doctor, although these patients were not subject to the same escalating investigations as in Rothenberg's report [Rothenberg, 1998]. The authors concluded that repeat blood count and focused investigations (parasitology, skin prick testing depending on clinical context, foreign travel) are sufficient, and that extensive diagnostic testing is unnecessary. Although the great majority of cases will be attributable to common allergic or parasitic disease, a small proportion potentially reflect other serious lung, gastrointestinal, renal autoimmune or malignant disease, although these would be expected to exhibit other signs or symptoms of the disease. In addition, a small number of patients will present with either the hypereosinophilia syndrome or a haematological malignancy in which the only feature may be the eosinophilia. Therefore, it would be reasonable to recommend that patients with persistent (> 6 months) mild eosinophilia or a finding of moderate or increasing eosinophilia in which the above investigations do not reveal a cause be referred for immediate assessment. In patients with moderate eosinophilia, if any signs of organ damage are present, as indicated by cardiac or pulmonary symptoms then the referral should also not be delayed.

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

Clear diagnostic results (eg. acute leukaemia, agranulocytosis) arising from a full blood count rarely pose serious diagnostic consequences in primary care. The boundaries between the statistically abnormal and clinically relevant abnormalities and the further investigations recommended in these situations are, however, less clear. The following series of questions examines raised and low platelet counts and haemoglobin and attempts to identify situations in which further monitoring or investigation is required and those where immediate secondary care attention is appropriate.

There is also a topic on Lymphocytes, neutrophils, eosinophils.

The lower end of the UK population reference range for blood platelets is 140 x 109/L [Bain, 1995] although one review has recommended 150 x 109 /L (or 120 x 109/L during late pregnancy) as the threshold to trigger a repeat sample in a Caucasian population [EBM Guidelines, 2005a]. Platelet counts of 120 x 109/L may, however, be seen in healthy people of African origin and 110 x 109/L in those of Afro-Caribbean origin.

Symptoms of a low platelet count are bruising and ecchymoses, muco-cutaneous bleeding and prolonged bleeding after trauma or lacerations. Very low platelet counts cause petechiae.

See also topic on Methotrexate.

Platelet counts > 30 x 109/L rarely cause symptoms and receive no urgent treatment [George et al, 1996; Portielje et al, 2001; Friedmann et al, 2002; Neylon et al, 2003], Antiplatelet treatment should, however, be avoided or stopped, depending on the relative risks of haemorrhage or thrombosis [Friedmann et al, 2002]. Although no specific referral threshold was found, we recommend referral to a haematologist if the platelet count falls to < 70 x 109/L or is < 110 x 109/L and falling on 4-weekly checks.

It is now recommended that children with immune thrombocytopenia do not receive treatment even with severe thrombocytopenia (< 10 x 109/L), in the absence of clinically important bleeding [George et al, 1996; Bolton-Maggs and Moon, 1997; Vesely et al, 2000]. A further guideline recommends that hospital admission is not warranted in idiopathic thrombocytopenia if the platelet count is > 20 x 109/L and no treatment if > 50 x 109/L. For practical purposes, these patients would have been referred under the above recommendations and the decision on further action would be taken in secondary care.

Low platelet count in pregnancy exclude hypertensive disorders and gestational thrombocytopenia [Burrows and Kelton, 1993]. (Gestational thrombocytopenia requires no intervention and is defined by five criteria: mild and asymptomatic thrombocytopenia; no history of thrombocytopenia (except possibly during a previous pregnancy); occurrence during late gestation; no association with fetal thrombocytopenia; and spontaneous resolution after delivery.)

No excess bleeding in the mother has been detected at delivery with platelet counts > 50 x 109/L [British Committee for Standards in Haematology, 2003].

Excessive complications were not detected when epidural punctures were carried out with platelet counts > 70 x 109/L [Beilin et al, 1997]. Therefore, we adopted 70/l as a threshold for seeking haematological advice in asymptomatic people. Children have not suffered complications from lumbar punctures carried out with platelet counts > 10 x 109/L [Howard et al, 2000].

For a slightly reduced platelet count (between 110 and 150 x 109/L), the national guidelines clearing house recommend rechecking 'every few months' to ensure the count is not falling [Bain, 1995]. We recommend checking every 2 months. No guidance is given for counts < 100 x 109/L.

We recommend that if asymptomatic, in the case of a raised platelet count, the test should be repeated after an 8-week period so that any cause of reactive thrombocytosis settles or becomes overt.

The categories of a high platelet count (> 450 x 109/L) are primary, secondary and redistribution.

Patients can have thrombotic problems even when the platelet count is in, or just above, the normal range [Regev et al, 1997]. If the patient is below 60 years of age and has no other risk factors then primary thrombocythaemia should be considered (if in doubt, seek advice from haematologist).

Even platelet counts > 1,000 x 109/L are not associated with major thrombosis or haemorrhage [Buss et al, 1994].

Clinical evaluation - signs of inflammation (e.g. infection, rheumatoid arthritis) or bleeding

We recommend firstly repeating the measurement. Consider referral if a patient has a persistently raised venous haematocrit (> 0.52 for males, > 0.48 for females for > 2 months) [McMullin et al, 2005].

The major pathophysiology of a raised haemoglobin is increasing haematocrit causing an increase in whole blood viscosity. The incidence of large-vessel occlusion increases with haematocrits > 0.45 [Pearson, 1987].

A high haemoglobin level can occur with a normal whole body red cell mass - this is called apparent erythrocytosis [Brown et al, 1971]. True polycythaemia can be assumed to be present when the haematocrit is > 0.6 in males and 0.56 in females [Pearson, 2001].

A high haemoglobin can be caused by a secondary polycythaemia such as hypoxic lung disease or even chronic high levels of carboxyhaemoglobin owing to cigarette smoking. Other much less common causes include erythropoietin release from renal cysts and tumours, liver and parathyroid tumours, and uterine fibroids [McMullin et al, 2005].

A high haemoglobin level can be caused by a primary polycythaemia (incidence 2:100,000) in which the WBC and platelets may also be raised. These individuals are at increased risk of thrombosis even with haematocrits near, or occasionally in, the normal range, possibly caused by the pan-myeloproliferative element of the condition [Cortelazzo et al, 1995]. Aspirin reduces the incidence of thrombosis in primary polycythaemia [Landolfi et al, 2004], suggesting a key role for platelets in thromboembolic events.

Studies have shown that a haematocrit level in the upper normal range, or slightly raised, may be associated with an increase in thrombotic events and cardiovascular mortality compared to those with a haematocrit in the middle or lower part of the normal range [Lowe, 1999]. That the increase in mortality associated with apparent erythrocytosis is because of the high haematocrit level is not clear; nor are there randomised studies to show that reducing the haematocrit in apparent erythrocytosis reduces morbidity or mortality.

Serial measurements of the haematocrit in untreated patients with apparent erythrocytosis show that the haematocrit level returns to within the normal range in up to 30% of patients [Messinezy and Pearson, 1990].

Modifications in the factors which are associated with apparent or secondary erythrocytosis, such as obesity, smoking and hypertension, may lead to a reduction in haematocrit level [Pearson, 1991]. In those with severe hypoxic pulmonary disease, supplementary oxygenation may be required and a review by a specialist respiratory physician should be considered. In addition, a clinically relevant minority of patients with erythrocytosis has nocturnal oxygen desaturation due to obstructive sleep apnoea [Moore-Gillon et al, 1986], and such patients should be referred for appropriate investigation [Eisensehr and Noachtar, 2001].

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

Please choose a cancer testing topic:

The answers presented here do not deal with the current debate in laboratory medicine about the relative utility of different forms of measurement (free, total, calculated ratios) of prostate-specific antigen (PSA). The use of PSA in the general community has been widely debated on both sides of the Atlantic, although occasionally the implications and limitations of widespread use of PSA in unselected populations or in those at lower risk has been lost in the emotive and political debate. These answers attempt to put that debate in perspective.

In England and Wales, the number of cases of prostate cancer diagnosed has more than doubled in a decade, mainly because of the widespread use of prostate-specific antigen (PSA) testing [Office for National Statistics, 2003]. Despite being the most common cancer in men, with nearly 27000 new cases diagnosed in 2002, there is no evidence to support routine PSA screening of prostate cancer in asymptomatic men. Even though screening often detects early 'curable cancers' in most patients, there is inconclusive evidence from one randomised controlled trial that screening and early detection improves life expectancy or quality of life [Labrie et al, 2004], and one American study reported no increase in longevity, but increased morbidity in a two-centre comparison of PSA measurement with surgery versus PSA with 'watch and wait' approach [Lu-Yao et al, 2002].

A randomised study from Sweden has shown a modest survival benefit from radical surgery over watchful waiting [Bill-Axelson et al, 2005]. Longitudinal studies in UK and USA, however, indicate that most men with localized low-grade prostate cancer have only a small risk of dying from prostate cancer and hence are unlikely to benefit from an early diagnosis of their cancer [Albertsen et al, 2005; Kattan et al, 2005]. Only about a third of patients with prostate cancer will die from their disease [Kattan et al, 2005].

Even though, up to two-thirds of men over the age of 70 are found to have prostate cancer in autopsy series, prostate cancer is responsible for only 3.5% of all deaths in men [Sakr et al, 1994; Office for National Statistics, 2004]. PSA screening need not therefore be proactively discussed with asymptomatic men.

Two large randomized screening trials, the Prostate, Lung, Colorectal and Ovary (PLCO) cancer trial in the USA and the European Randomized Screening for Prostate Cancer (ERSPC) trial in Europe are currently assessing the benefits of screening for prostate cancer [Andriole et al, 2005; Schroder, 2005].

While awaiting the results of the PSA screening studies, the PSA test should not be denied to men who request it. This should be conducted after proper counselling about the drawbacks of PSA test (false positive and false negative results), the natural history of prostate cancer, the treatment options for cancer, and the physical, psychological and financial implications of detection of a cancer that might never have affected health had it been detected during the lifetime of the individual [Potosky et al, 2002; Steineck et al, 2002; Bradley et al, 2005].

Prostate cancer is rare in men under 50 years old [Office for National Statistics, 2004]. In asymptomatic men who elect to undergo screening, PSA tests therefore should be started at the age of 50 years (or 45 years in the higher risk population). Annual or biannual PSA alone, without a rectal examination, is sufficient to identify men who are at high risk of having prostate cancer. PSA testing is not usually recommended for an asymptomatic man with less than 10 years life expectancy [Schroder et al, 2001; Labrie et al, 2004].

PSA is both a diagnostic and a screening test. UK NHS guidance recommends that PSA should be measured, if clinically appropriate, in men with refractory lower urinary tract symptoms, erectile dysfunction, hard irregular prostate, haematuria or bone pain with weight loss or other clinical situations compatible with prostate cancer [NHS Cancer Screening Programmes, 2002].

Various procedures, activities and diseases can affect the PSA levels. Hence, the exclusion criteria of a PSA are:

If an asymptomatic man has a PSA level higher than the age specific range, he should be counselled about further options, which include a prostate biopsy.

Prostate-specific antigen (PSA) is a protein produced almost exclusively in the epithelial cells of the prostate gland. It is normal for men to have low levels of PSA in their blood, and there is no specific normal or abnormal PSA level. Both prostate cancer and benign conditions can increase levels of PSA.

An isolated increase in the PSA level should be confirmed several weeks later, before proceeding with further testing, including prostate biopsy.

If practical, the man should have the PSA test before the digital rectal examination (DRE). If not, it is recommended that the PSA test be delayed by 1 week. The blood sample should reach the laboratory (and be separated) < 16 hours after the sample is taken [NHS Cancer Screening Programmes, 2002; NICE, 2005].

Age is an important factor in increasing PSA levels. For this reason, the NHS Prostate Cancer Risk Management Programme recommends age-adjusted PSA levels to determine when diagnostic tests are needed. The programme recommends the following cut-off values for referral:

A very high PSA value is strongly suggestive of cancer.

A PSA level of more than 100 micrograms/L with an abnormal prostate on DRE is almost certainly caused by prostate cancer. In an elderly man with extensive comorbidities and grossly raised PSA, biopsy is not therefore necessary before embarking on hormone therapy [Gerstenbluth et al, 2002].

If an asymptomatic man has a PSA level higher than the age specific range, he should be counselled about further options, which include:

Hopefully, the ongoing NHS Health Technology Assessment funded ProtecT trial, evaluating the effectiveness of various treatment options, including active surveillance for clinically localised prostate cancer will benefit future decision making [ProtecT, 2002]

Changes in PSA levels over a 3-6 month period can be prognostic. A PSA which doubles in the 6 months of follow-up after surgery, indicates an aggressive cancer and points towards 'incurable' systemic relapse rather than a 'potentially curable' local relapse [Pound et al, 1999a]. We therefore recommend PSA checks every 3-6 months in men with prostate cancer.

In general, after radical treatment of prostate cancer, when there are no signs of recurrence, PSA checks should be measured every 3 months for 1-2 years, every 6 months until the fifth year, and after 5 years, annual checks should suffice [Oh et al, 1999; Niwakawa et al, 2002]. Unlike other cancers, late recurrence and progression can occur even after 10-15 years. Thereafter, it is recommended that periodic measurements of PSA be continued for life [Oh et al, 1999; Niwakawa et al, 2002].

The frequency of PSA checks may be altered in specific cases for various reasons. Elderly or frail men may require PSA levels measured less frequently and only if clinical intervention is practical and realistic.

After radical prostatectomy, the PSA should be undetectable and drop to less than 0.1. Failure to drop to this level is cause for concern and further evaluation or treatment may be necessary [Rogers et al, 2004; Shen et al, 2005].

After radiation with either seeds (brachytherapy) or external beam or cryotherapy, the PSA will drop slowly, often reaching its nadir after several years. After 18-24 months following radiation, PSA levels are usually below 1.0. A slow fall in levels of PSA is not a cause for concern; paradoxically, the slower the fall the better the prognosis [Zagars, 1994]. It is not uncommon after prostate radiation to see a temporary rise or "bounce" in the PSA level, which then falls spontaneously. The classic PSA bounce timing is about 1-2 years after treatment and may be seen in up to 30% of patients. PSA levels should later fall and remain below 1.0 [Hanlon et al, 2001].

After hormone therapy alone, nadir levels are quite variable and depend on the extent of the cancer and its aggressiveness. In general, nadir levels of PSA of less than 2.0 indicates good prognosis. After hormone therapy, stable levels are reassuring and any rise in PSA is a cause for concern [Kiriyama et al, 2005].

Digital rectal exams (DRE) are not necessary during follow-up of patients who have been radically treated. The usefulness of DRE is limited in the follow-up of patients who have undergone radiotherapy [Pound et al, 1999b; Doneux et al, 2005].

Patients on active surveillance ('wait and watch policy'), should have the PSA levels checked every 3-6months. If there is a major rise in PSA, the patient should be counselled about treatment options again. It is important to note that the rate of rise in PSA is more important than the absolute levels of PSA [D'Amico et al, 2004; Eggener et al, 2005].

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

In a laboratory setting most requests for urine cytology are received from hospital clinics, primary care specimens forming a minority. The guidance provided is taken mostly from studies of test utility rather than consensus guidance and relate specifically to the initial primary care investigation of the situation described.

We recommend urine cytology in patients in the following situations:

Gross or visible haematuria requires evaluation of the upper and lower urinary tract [Malmstrom, 2003; Shah, 2003; Yun et al, 2004]. Urine cytology has a supportive role in the evaluation of these patients [Chahal et al, 2001] in conjunction with upper tract imaging, cystoscopy and bladder biopsy. However, in one study [Chahal et al, 2001] urine cytology did not lead to the discovery of additional tumours that were not detected by any of the other investigations. Recent urinary tract instrumentation should be excluded as a cause of gross haematuria [Shah, 2003].

MH (by dipstick analysis) is more controversial. Although it is generally accepted that symptomatic MH requires microscopy, this has recently been questioned [Bove et al, 1999]. There are no reliable data on the incidence of underlying bladder neoplasms in patients with symptomatic MH. Nonetheless, one small study of women who had incontinence and/or irritative voiding as well as MH, showed cytology to be of value [Duldulao et al, 1997], and the American Urological Association (AUA) best practice guidance recommends urine cytology in patients with a history of irritative voiding symptoms [Grossfeld et al, 2001]. Urine cytology is not helpful in the evaluation of men with lower urinary symptoms [Malmstrom, 2003] because asymptomatic MH is a frequent finding in patients who have benign prostatic hyperplasia [Ezz el Din et al, 1996], and urine cytology is not an investigation of choice to detect prostatic malignancy. It has not been evaluated in this context.

Urinary tract infection (UTI) does not require investigation with urine cytology, but via microbiological testing [Lammers et al, 2001; Moore et al, 2001; Sultana et al, 2001; D'Souza and D'Souza, 2004]. Cytological samples in the setting of UTI may be obscured by polymorphs and it is sensible to exclude and treat infection prior to submitting samples for cytological evaluation if indicated in a patient presenting with a UTI.

Asymptomatic MH is common in adult primary care populations (2.5-4.3%) [Kryszczuk et al, 2004] and up to 11% of patients with asymptomatic MH have been reported to have underlying urothelial malignancy [Mansson et al, 1993; Fracchia et al, 1995; Froom et al, 1997]. However, in a review of 17 series, comprising a total of 5,000 patients, urological cancer was diagnosed in < 3% of cases. A Californian study found asymptomatic MH in 2.9% of 20,751 patients of whom 0.5% had urothelial cancer, yet cancer was also found in 0.5% without MH. Therefore, the authors concluded that the presence of asymptomatic MH was not significantly associated with urological cancer or other serious urological disease [Hiatt and Ordonez, 1994]. No studies have demonstrated improved outcomes from screening for asymptomatic MH [Messing et al, 1995].

Screening for asymptomatic MH cannot be recommended as a means to detect urological malignancy but in patients > 40 years with an incidental finding of MH, the American Urological Association (AUA) recommends complete urological evaluation (including cytology) [Grossfeld et al, 2001]. The significance of age as a risk factor is supported by a study in a subspecialised urological setting where 87% of patients in whose samples malignant cells were found were > 50 years of age. It should however, be noted that 72% had a history of gross haematuria. The AUA [Grossfeld et al, 2001] also recommends complete urological evaluation (including cytology) in younger patients with a history that is "suspicious of underlying urological disease". The relative merits of full urological investigation (including cytology) in younger patients with asymptomatic MH have not been evaluated and are therefore debatable [Sultana et al, 1996; Nabi et al, 2003].

A combination of microscopic haematuria and proteinuria in younger patients is a predictor of non-neoplastic primary renal disease [Topham et al, 2004]. In patients with this combination of findings on dipstick examination, the cytological identification of red cell casts and dysmorphic red blood cells may serve as a further indicator of renal parenchymal disease [Fracchia et al, 1995].

Cytology is advised in patients with risk factors for transitional cell carcinoma [Hofland and Mariani, 2004] although other screening methods [Hofland and Mariani, 2004; Topham et al, 2004] probably have higher combinations of sensitivity and specificity in patients with low-grade non-invasive tumours or patients with carcinoma in-situ [Ramakumar et al, 1999; Hofland and Mariani, 2004]. In patients with incidental MH, the risk factors listed by the AUA as indications for urine cytology include: smoking history, occupational exposure to carcinogenic chemicals or dyes (benzenes or aromatic amines), analgesic abuse (for example, phenacetin), cyclophosphamide and pelvic irradiation [Grossfeld et al, 2001]. Known schistomiasis bladder infection may reasonably be included in relevant populations as a predisposing factor for bladder cancer [WHO, 2003]. Cytology has been used in the follow up of patients who have been treated for bladder cancer in conjunction with testing for MH.

In the laboratory a single cytospin deposit [Burton et al, 2000] (as opposed to the preparation of duplicate slides) is adequate for cytological evaluation, saves resources and caused minimal loss of clinically relevant information.

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

Cancer referral guidelines now place the emphasis on rapid secondary care investigation of suspected cancer patients; the primary care use of investigations such as sputum cytology (in the context of suspected malignancy) is increasingly changing to higher performing imaging and endoscopic techniques. This question-answer set attempts to identify situations in which primary care requesting of sputum cytology may be appropriate.

We recommend:

The United States Preventive Services Task Force has found evidence that while lung cancer screening using a variety of techniques, including sputum cytology, can detect lung cancer at an earlier stage than without screening, it found poor evidence for a screening policy reducing mortality [U.S. Preventive Services Task Force, 2006].

The sensitivity of sputum alone has been reported to be between 10-20 % [Marfin and Schenker, 1991], and 42% in one series [National Cancer Institute, 1984], although many series have reported detection rates of 60% to up to 90-95% from sequential 5 day series of sputa with 5 day collections [Koss et al, 1964; Erozan and Frost, 1974], and one study cited poor specimen conditions as a reason to explain this. There appears good consensus that correctly obtained and processed samples are important to maintain sensitivity. When collected, a minimum standard of 3 [Bocking et al, 1992], or ideally 3-5 [Johnston and Bossen, 1981] early morning specimens are recommended [de May, 1996; American Society of Cytopathology, 2004], taken from symptomatic patients with productive cough, and submitted either immediately to the laboratory or preserved in accordance with local recommendations.

Mass sputum screening is also not recommended by the American Society of Cytopathology [American Society of Cytopathology, 2004], American Cancer Society and US National Cancer Institute, reviewed in [State of New Jersey Department of Health and Senior Services, 2007] and considered to be inefficient [Wright et al, 1986].

In patients suspected of suffering from lung cancer, the American College of Chest Physicians Guideline [American College of Chest Physicians et al, 2003], containing contributions from 13 American and International medical associations recommended that the diagnosis be obtained by "whichever method is the easiest...as dictated by the patient's presentation", from sputum cytology, fine needle aspiration and bronchoscopic methods of obtaining specimens. While it states that sputum cytology in a centre with a formal programme for using this in the diagnostic workup is a reasonable first step, it states that it is of limited sensitivity; negative results do not exclude the diagnosis and require further investigation.

CKS guidance [CKS, 2005f] based on the UK Department of Health Referral Guidelines [DH, 2000c] recommends that sputum cytology be rarely indicated prior to specialist referral, but that a chest X-ray is an appropriate primary care investigation. This guidance includes criteria for requesting a chest X-ray and for urgent specialist referral. A 2-week standard also exists in the UK for patients with suspected lung cancer to be seen by a specialist [DH, 1997]. In view of this and the above guidelines, it would appear appropriate in the specific UK context, to refer suspected cases immediately and either leave the decision on the means of obtaining a cytological diagnosis to the specialist centre or discuss with the centre before performing.

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

Faecal occult bloods have long been contentious because of false positive and, particularly false negative results and the implications of missing the potential diagnosis of malignancy. Cancer referral guidelines now place the emphasis on rapid secondary care investigation of suspected cancer patients, and the primary care use of investigations such as faecal occult blood testing is increasingly changing to endoscopic techniques. This question and answer set attempts to identify patients in whom this test may be appropriate in primary care and the means of obtaining the best results from the test.

CKS guidance [CKS, 2005a], which follows that of NICE [NICE, 2005] cites the following criteria for urgent referral of suspected colorectal cancer and recommends that apart from a full blood count, abdominal and rectal examination, no other tests be performed, in order not to delay referral.

The Scottish SIGN guideline differs slightly, using a threshold of 50 years old and specifically adds intestinal obstruction as an indication for referral [SIGN, 2003a]:

All patients with iron-deficiency anaemia (Hb < 11g/100ml in men or < 10g/100ml in postmenopausal women) without overt cause should be thoroughly investigated for colorectal cancer (see also topic on Iron deficiency anaemia).

It follows from this that only lower risk patients who would not require urgent referral should be considered for FOB testing, in order to expedite referral of positive cases, who are more likely to have bowel pathology.

We recommend:

Faecal occult blood testing (FOB) is a non-invasive, simple and rapid near patient test. Typical tests use a Guaiac impregnated paper, which produces a colour change in the presence of blood when a hydrogen peroxide developing solution is dropped onto the test.

A number of FOB kits are available, which vary in specificity and sensitivity. One study in 1990 compared three tests in symptomatic patients (Haemoccult, Fecatwin and E-Z Detect) [Tate et al, 1990]. These were used to test stool specimens from three sequential days. Using double contrast barium enema as a diagnostic test, the authors found Fecatwin to be the most sensitive. This test, however, gave three times as many false positive results as Haemoccult. The authors therefore concluded that a Haemoccult positive symptomatic patient had approximately a 50% probability of mucosal disease, and suggested this was the best of the three tests to use in the community, as long as the tester was aware that a negative result does not exclude serious pathology. E-Z detect is a patient interpreted test, where a sheet of benzidine impregnated paper is floated in the toilet in the presence of stool. This was less sensitive for blood that Haemoccult, and was not recommended in this study.

Other tests are available for patients to perform themselves, such as the Coloscreen Self-Test, a floating card the patient places in the toilet pan. However, when compared to the Haemoccult in symptomatic patients, whilst patient preference was greater for the self-test method, compliance was better with Haemoccult tests distributed for the patients to spot with faeces, again, over three consecutive days, and return to their practitioner. In this study, Coloscreen Self-Test was also less sensitive [Pye et al, 1990].

Immunological tests also exist. One, Hemeselect, was compared with Haemoccult, in a population with gastrointestinal symptoms on three consecutive daily bowel motions. The immunological test was more sensitive although produced a higher false positive rate, with poorer specificity. The authors, however, concluded that due to the increased sensitivity for carcinoma, trials in asymptomatic patients may be justified [Thomas et al, 1992a].

Faecal alpha1 - antitrypsin assay has also been described as a marker of gastrointestinal bleeding. This quantitative test is slightly more specific than Haemoccult, but neither test was considered sensitive enough to justify routine use in high risk patients [Moran et al, 1995]. Guaiac based testing appears to offer the best compromise between sensitivity and specificity, is most commonly used in the UK, and will be used in the planned asymptomatic screening programmes [UK Colorectal Cancer Screening Pilot Group, 2004].

We recommend:

The accuracy and value of Guaiac testing for symptomatic patients was investigated in 1983 using Haemoccult tests in a prospective study of 802 symptomatic patients referred from secondary care. The authors found a low false positive rate of 8.6%, although the false negative rate 45.4% when patients examined two samples from each of three consecutive stools. This could, they claim, be improved when combined with a 'proper digital anorectal and proctosigmoidoscopic examination' to identify rectal tumours [Leicester et al, 1983a]. There was, however, some debate about the interpretation of the data presented [Leicester et al, 1983b].

A study of symptomatic referrals to secondary care in 1993, where 3 consecutive days' stools were tested found 11% to be positive for occult blood using Haemoccult.

Of these, 63% were found to have colonic pathology. In comparison, fewer than 10% of those with a negative FOB test had significant findings when investigated. The authors concluded that although a positive FOB test was highly specific, a negative test did not "adequately exclude colonic pathology", and they suggest that the FOB test could be used as "guide to the urgency of investigation" [Falkson and Bates, 1993].

An older Australian study using Haemoccult tests on symptomatic patients also supported the belief that a positive FOB test can indicate higher likelihood, but not exclude, organic disease [Goulston and Davidson, 1980].

Equally, a study published in 1995 by an Israeli group studied symptomatic patients using Haemoccult FOB testing of 2 samples from each stool over 3 days and colonoscopy. For neoplastic lesions, sensitivity and specificity were both acceptable at 69.2% and 73.2%; however the positive predictive value of the test was only 27.6% The patients in the trial undertook an appropriate exclusion diet before testing, however, the authors did emphasise the differences between the typical diet of their patients and the classical 'Western' diet. The authors concluded that they could not recommend FOB testing in the investigation of the symptomatic patient [Niv and Sperber, 1995]. Overall, positive Haemoccult will detect around two-thirds of bowel cancers, and only approximately 25% of positive results will reflect malignancy.

Several factors, in addition to diet [Thomas et al, 1989], are known to interfere with FOB testing. The American Cancer Society [American Cancer Society, 2007] recommends avoiding:

However, it is emphasised that compliance is more important than strict dietary adherence.

Population screening is advocated by the American Cancer Society, National Cancer Institute and US Preventative Services Task Force in the US (summarised in [State of New Jersey Department of Health and Senior Services, 2004]) based on the large number (around 20%) of asymptomatic cancers; over the next three years national screening programs for colorectal cancer will be rolled out across the UK (www.cancerscreening.nhs.uk/bowel). These have been established following a successful pilot study in 2003 [UK Colorectal Cancer Screening Pilot Group, 2004]. This will involve screening programmes targeting specific patients who are not necessarily symptomatic. Screening of asymptomatic populations will not be considered further in this answer.

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

Please choose a cholesterol and lipids topic:

This short question set of partially interrelated questions examines some of the less common, but nevertheless not uncommon situations which may arise in lipid clinics and in primary care, and can cause diagnostic difficulty. They often require different forms of treatment. Good consensus guidance exists although some recommendations are drawn from primary research (observational studies).

We recommend screening for the causes of secondary hyperlipidaemia in all patients in whom lipid-lowering treatment is being considered.

The following investigations are recommended:

One study of 1190 people referred from primary care to a hospital lipid clinic reported a prevalence of secondary hyperlipidaemia of 1.8% when diagnosed by laboratory tests alone, despite 17% of the people having at least one abnormality of the four screening tests used (although the abnormality was not believed to be related to an underlying secondary hyperlipidaemia) [Evans and Gray, 1994]. Although the prevalence of secondary causes would therefore appear to be low overall, the prevalence of laboratory abnormalities that may be relevant to lipid management is far greater. However, the prevalence of secondary causes appears to be considerably greater as lipid parameters rise. This is considered below.

Alcohol overuse and liver disease are common conditions that present with dyslipidaemia, notably hypertriglyceridaemia [SIGN, 1999a; CKS, 2003; University of Michigan Health System, 2003; International Task Force for the Prevention of Coronary Heart Disease, 2004].

Raised triglyceride concentrations are typically seen when people are first diagnosed with type 2 diabetes [Nikkila, 1973] and may be in excess of 20 mmol/L. Hypertriglyceridaemia is also common in type 2 diabetes as part of the metabolic syndrome [Durrington, 1990], even when blood glucose levels are well controlled (see also topic on Glycated haemoglobin and microalbumin).

Average cholesterol concentrations rise as the thyroid-stimulating hormone increases outside of the reference range [Elder et al, 1990]. One study that screened 2,250 Scottish people found 90 people with a total cholesterol > 8 mmol/L, and 12% of these had biochemical evidence of hypothyroidism [Series et al, 1988]. Hypopituitarism should also be considered [Ishibashi et al, 1985].

Specific patterns of dyslipidaemia are associated with certain conditions (see Table 1) but these patterns are not found consistently and the associated conditions may not be clinically apparent.

Widespread consensus exists about which investigations to perform to screen for the causes of secondary hyperlipidaemia [Durrington, 1995b; SIGN, 1999a; DH, 2000a; CKS, 2003; University of Michigan Health System, 2003; International Task Force for the Prevention of Coronary Heart Disease, 2004], although the International Task Force also recommends specialised tests (e.g. to diagnose polycystic ovary syndrome) in specific situations. These investigations are recommended to exclude the most common conditions that cause hyperlipidaemia, irrespective of the pattern of the dyslipidaemia.

Hyperlipidaemia is rare as a presenting feature of nephrosis, but urine testing is a simple, non-invasive and inexpensive means of screening for this.

No evidence or clear expert consensus exists on when to screen for the causes of secondary hyperlipidaemia, and most guidelines do not provide absolute screening thresholds. CKS [CKS, 2003] recommends excluding secondary causes, in particular in people with cholesterol > 6.5 mmol/L or triglycerides > 8.0 mmol/L. The European Atherosclerosis Society [Second Joint Task Force of European and Other Societies, 1998] and SIGN [SIGN, 1999a] recommend exclusion of secondary causes before initiating lipid lowering treatment.

We recommend that triglycerides are measured:

Triglycerides should be measured at the same time as cholesterol in certain circumstances, because raised levels:

Triglyceride levels can be described as raised if > 2.3 mmol/L and greatly raised if > 5.6 mmol/L [NCEP, 2001], although thresholds for these descriptions vary.

Raised triglyceride levels are common in the general population. The Munster Heart study [Assmann et al, 1998] reported a prevalence of hypertriglyceridaemia (triglycerides > 2.3 mmol/L) of 5% in 20 year old men, rising to 20% at age 45, and 2% in 20 year old women, rising to 7% at age 60. Triglyceride concentrations > 4mmol/L are more unusual in the general population (the 95th percentile in adult British males being approximately 3.8 mmol/L) [Mann et al, 1988]. A raised value should be rechecked to take account of biological variability. Non-fasting values over 2.3 mmol/L would appear to justify rechecking on the basis of increased cardiovascular risk in this group regardless of fasting status [Eberly et al, 2003], although a local evidence-based UK guideline has recommended a threshold of 4.5 mmol/L [FATS Steering Group, 2002]. There would not appear to be a clear evidence-based answer to this question, as 4.5 approximates to the threshold for triglyceride treatment (see below), retesting on a fasting sample would appear logical.

Many laboratories offer measurement of cholesterol without simultaneous measurement of triglycerides. For reasons set out below, however, in certain circumstances we recommend measuring the triglyceride level simultaneously [Smellie et al, 2001].

Triglyceride levels have several implications for the measurement and treatment of cholesterol. Concentrations > 2.3 mmol/L reduce the accuracy of methods used by many laboratories to calculate LDL cholesterol. When triglycerides > 4.5 mmol/L, the method to calculate LDL becomes unreliable and target total cholesterol concentrations cannot be used as treatment goals [Austin et al, 1998]. The NCEP recommends the use of non-HDL cholesterol as a different treatment goal if triglyceride concentrations are > 2.3 mmol/L [NCEP, 2001].

Although it is difficult to separate triglycerides from other risk variables, raised triglycerides are associated with increased coronary risk [Austin et al, 1998], which is reduced by treatment increasing HDL (see also topic on High density lipoprotein (HDL-C) measurement and treatment) and lowering triglycerides [Rubins et al, 1999]. Therefore, we recommend measuring triglycerides in all people being assessed for CVD risk. Opinions are divided on the inclusion of triglycerides as formal targets in coronary prevention.

Hypertriglyceridaemia is also a frequent finding in undiagnosed secondary dyslipidaemia, notably diabetes [Durrington, 1990], and in association with high alcohol intake and certain drugs (see secondary hyperlipidaemia above). Values may rise exponentially in these situations, owing to saturation of triglyceride-removal system [Brunzell et al, 1973]. Therefore, we recommend measuring triglycerides during the assessment of all people being considered for lipid-lowering treatment.

Triglyceride concentrations can fluctuate considerably over time and with diet [Neil et al, 1990; Rifai et al, 1999]. The European Task force therefore recommends that raised fasting values (> 2 mmol/L) should prompt repeat measurements.

Serum triglycerides of:

We recommend:

Very high levels of triglycerides carry a risk of pancreatitis. No absolute risk threshold exists for this and thresholds for increased risk described in recent guidelines vary — 6.5-13 mmol/L [Cucuzzella et al, 2004], 10 mmol/L [McIntosh et al, 2002], or even 20 mmol/L [Durrington, 1995a].

One retrospective study in 56 women with hypertriglyceridaemia (triglycerides > 5.2 mmol/L, median 16.5 mmol/L) found that 30% had suffered an episode of pancreatitis during the previous 3-year period [Goldenberg et al, 2003]. An earlier study of severe hypertriglyceridaemia (> 9.8 mmol/L) associated with oestrogen replacement treatment reported that four of seven women with triglycerides > 19.5 mmol/L had suffered acute pancreatitis and a further two suffered acute abdominal pain thought to be pancreatitis over a period of 2.75 years [Glueck et al, 1994]. Triglycerides levels that were previously regarded as being associated with an increased risk of pancreatitis are probably too conservative (C. Glueck, personal communication relating to [Glueck et al, 1994; Goldenberg et al, 2003]). Additionally, on the basis of limited evidence available from these small studies, there would appear to be a very high risk of pancreatitis for people with triglyceride concentrations > 19.5 mmol/L; a smaller but still high risk for people with triglyceride concentrations > 9.8 mmol/L; and a probable increased risk for people with triglyceride concentrations > 5 mmol/L, as it is probable that risk increases incrementally.

Most guidance recommends treatment of raised triglycerides > 5mmol/L regardless of other factors [NCEP, 2001; Singapore Ministry of Health, 2001; McIntosh et al, 2002; CKS, 2003; De Backer et al, 2003; Cucuzzella et al, 2004] and consideration of secondary referral > 10 mmol/L [CKS, 2003], because of the increased likelihood of underlying medical causes and familial hyperlipidaemias.

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

Lipid guidelines are changing on a regular basis and the results of this search are current as of May 2005. However, the guidance may change. Most of the guidelines are derived from consensus documents. Although published separately, these documents are not fully independent of one another, and there is frequent cross reference between documents. Overall agreement was in general close and the minor differences, for example in testing intervals found, are unlikely to have significant relevance to clinical outcome.

However, following the Heart Protection Study results, the decision to make simvastatin available without prescription and the anticipated publication of new Joint British Societies guidance late in 2005, the summaries below should be seen as the guidance position as of spring 2005, and of all the sets of guidance to be examined is the one which will require updating probably within one year. Target or optimum cholesterol reduction levels are not addressed in these questions because of recent changes discussed below. An overview of the guidance is shown in Figure 1.

There are separate topics on Secondary hyperlipidaemia and triglycerides and High density lipoprotein (HDL-C) measurement and treatment.

We recommend that after starting lipid lowering drug treatment, lipids should be checked:

There is close consensus in national and international guidance that patients should have lipid parameters measured after starting statin treatment until a target is reached [SIGN, 1999a; DH, 2000a; Singapore Ministry of Health, 2001; McIntosh et al, 2002; NCEP, 2002; University of Michigan Health System, 2003]. The recommended measurement intervals all range between four and 12 weeks after starting drug treatment. The term treatment frequently refers specifically to statins in these documents, although we consider that it is reasonable to extrapolate this to other drugs used in lipid lowering.

The review interval after starting drug treatment will depend on practice resources and logistics, but it is unlikely that differences within this 12 week period will influence long term outcome.

The Heart Protection Study [Heart Protection Study Collaborative Group, 2002], published after the above guidelines, examined a large number of patients with minimal laboratory follow up. However, this was a selected population and it cannot be assumed that tolerability and compliance would be identical in an unselected population. We would expect future guideline revisions to consider this option, although this approach cannot be recommended at present on the basis of any consensus guideline or statement.

We recommend that once a patient has reached a defined target for lipids, that lipid tests be performed annually (unless specific reason for more frequent reviews).

There is close consensus in national and international guidelines [SIGN, 1999a; DH, 2000a; Singapore Ministry of Health, 2001; NCEP, 2002] that patients should continue to be monitored once they have reached a target lipid level on lipid lowering treatment. The recommended monitoring interval is most frequently one year [SIGN, 1999a; DH, 2000a; VHA/DoD, 2001; McIntosh et al, 2002; De Backer et al, 2003; University of Michigan Health System, 2003].

Some guidelines outside of the UK recommend more regular measurements, at intervals of 6 months [Singapore Ministry of Health, 2001] or 4-6 months [NCEP, 2002] once a patient has reached target.

The Heart Protection Study [Heart Protection Study Collaborative Group, 2002] published after the above guidelines, examined a large number of patients with limited laboratory follow up. Some have therefore questioned the need for laboratory follow up. However, this was a selected population and it cannot be assumed that tolerability and compliance would be identical in an unselected population. We would expect future guideline revisions to consider this option, although it cannot at present be recommended on the basis of any consensus guideline or statement.

We recommend that the decision to start lipid lowering drug treatment should be based on:

When monitoring risk in patients not receiving lipid lowering drug therapy we recommend:

There is close consensus in both national and international guidance that the decision to start lipid lowering therapy should not be based on more than one measurement, but two, [SIGN, 1999a; DH, 2000a; Singapore Ministry of Health, 2001] or three, [NCEP, 2002; De Backer et al, 2003].

A recent analysis of the imprecision of coronary risk assessment has suggested that in fact three measurements are needed to produce a precise estimate of risk [Reynolds et al, 2002]

Although methods used to assess coronary risk in the guidelines vary, all adopt a common approach based on testing at five yearly intervals in the lowest risk groups [American Heart Foundation, 1995; SIGN, 1999a; VHA/DoD, 2001; De Backer et al, 2003; University of Michigan Health System, 2003], although in Singapore it is every three to five years [Singapore Ministry of Health, 2001]. This rises to annually [SIGN, 1999a; DH, 2000a; De Backer et al, 2003] or one to two yearly [American Heart Foundation, 1995; Singapore Ministry of Health, 2001; VHA/DoD, 2001; University of Michigan Health System, 2003] in secondary prevention and medium risk primary prevention patients who have yet not reached the threshold for lipid lowering drug treatment.

To use current risk assessment tools, lipid assessments before treatment require total and high density lipoprotein cholesterol (and therefore also potentially at least a calculated LDL) and triglycerides. Regardless of risk assessment method used, all of the available guidance recommends these tests.

We recommend a baseline check before starting treatment with a statin, then eight weeks after starting a statin or after any dose increase, then if liver function tests (LFTs) are stable, annual checks thereafter.

•Lipid/cholesterol management in coronary prevention: Discussion

There is also a topic on 'Minor' liver function test abnormalities.

There is broad consensus that alanine aminotransferase and/or aspartate aminotransferase should be checked prior to starting a statin, and then rechecked one to three months after starting treatment [DH, 2000a; NCEP, 2002; Pasternak et al, 2002; CKS, 2003]. There is less agreement on how often LFTs should be measured subsequently. Some authorities recommend only annual checks [Pasternak et al, 2002], whereas others recommend checks every eight to 12 weeks during the first year of treatment, followed by annual checks thereafter [UKMI, 2002b]. SIGN recommends a single check after starting and none thereafter if LFTs are normal. Many of the guidelines we reviewed advised or based their recommendations on the monitoring requirements outlined by the manufacturers of the different statins.

Data from the largest statin study to date, the Heart Protection Study, are reassuring [Heart Protection Study Collaborative Group, 2002]. Over 20,000 people were allocated to treatment with a statin or placebo and monitored over five years. The incidence of raised liver enzymes was extremely low, with no excess in the statin treated group. In particular, there was no significant excess in the number of people who had treatment stopped in the statin compared to placebo group because of raised liver enzymes (0.5% versus 0.3%).

A systematic review of 48 statin trials is similarly reassuring [Law et al, 2003]. There were no cases of liver failure in the trials. Raised liver enzymes occurred in 1.3% of individuals receiving a statin compared with 1.1% of those receiving placebo.

On the basis of this, it has been argued that routine monitoring of LFTs is unnecessary in people without pre-existing liver disease. This is not in keeping with current monitoring requirements outlined in the product licences, and it is difficult to know how reliably safety data from carefully monitored clinical trials can be generalised to people commonly treated in primary care. However, postmarketing surveillance data are reassuring, suggesting one case of liver failure for each 10 million prescriptions or about one per million person years of use [Law et al, 2003].

The National Institute for Health and Clinical Excellence (NICE) has recently highlighted the requirement for further data on the need for biochemical monitoring of statins for adverse effects [NICE, 2002d].

We recommend that if transaminases become raised in a person taking a statin:

Raised hepatic transaminases occur in 0.5%-2.0% of people taking statins. It is uncertain whether increases in transaminases represent true hepatotoxicity. Transaminase levels often settle with dosage reduction, and often do not rise again with further increase in dosage or use of another statin [Pasternak et al, 2002].

Most of the guidelines we reviewed recommend stopping statins if transaminases rise to more than three times the ULN. Some suggest that dosage reduction with close monitoring may be an option, but do not provide specific instructions regarding frequency of monitoring. The drug manufacturers generally recommend that if transaminase levels become raised, these should be rechecked promptly, and if still raised, monitored closely. If concentrations continue to rise, particularly to above three times the ULN, the manufacturers advise reducing the dose or stopping the drug. Again, no specific advice is given on how frequently monitoring should take place. Our view is that a reasonable interval is four to six weeks.

A recent BMJ editorial discussing the criteria that should apply to monitoring drug treatments highlights the difficulty with making clear recommendations regarding monitoring of liver function tests in patients on statins [Pirmohamed and Ferner, 2003].

We recommend a baseline check before starting treatment with a statin.

If a person does not have any identifiable risk factors for myopathy:

If the person has identifiable risk factors for myopathy:

If a person develops muscle pain, weakness or cramps while taking a statin:

Myopathy is a rare but potentially serious adverse effect of statin therapy. The incidence of severe myopathy is reported to be 0.08%, and most of these individuals have recognisable risk factors for myopathy [Pasternak et al, 2002]. The CSM [CSM, 2002] highlights the following main risk factors:

Data from the largest statin study to date, the Heart Protection Study, are reassuring [Heart Protection Study Collaborative Group, 2002]. Over 20,000 people were allocated to treatment with a statin or placebo and monitored over five years. Reported muscular pain was no more common in the statin group than in the placebo group, and there was no difference in the number of treatment discontinuations due to muscle symptoms (0.5% in both groups). However, people were excluded from this trial if they were thought to be at risk of myopathy.

A systematic review of 48 statin trials is also reassuring [Law et al, 2003]. In over 35,000 people, with a total of 158,000 person years of treatment between them, rhabdomyolysis occurred in eight treated and five placebo patients (none with serious illness or death). A serum CK level greater than 10 times the ULN occurred in 55 treated (0.17%) and 43 placebo (0.13%) patients.

Postmarketing surveillance data suggests one death from rhabdomyolysis for each 10 million prescriptions, or about one for each million person years of use.

Most guidelines suggest that routine laboratory monitoring of CK is of little value in the absence of clinical symptoms or signs, although that a baseline measurement is helpful to identify asymptomatic CK increases and thereby avoid subsequent confusion [NCEP, 2002; Pasternak et al, 2002]. Therefore, all people receiving statins should be instructed to report muscle discomfort or weakness or brown urine immediately, which should prompt a CK measurement [CSM, 2002; NCEP, 2002; Pasternak et al, 2002] and reassessment of the benefit/risk ratio of treatment.

One subgroup which merits specific attention is people of Afro-Caribbean descent, in whom raised CK values of 1000 IU/L may be seen as a normal variant. No clear guidance exists for management of this group. The opinion of Heart UK (personal communication, A Wierzbicki, 2005) is that this should be treated using a modified threshold to stop treatment of five times the pretreatment value, assuming the patient has no muscle symptoms.

We recommend that if CK become elevated in a person taking a statin:

If more than five times the ULN:

If less than five times the ULN:

Apart from high risk patients (described above) the benefits of statin therapy where indicated are considered to far outweigh the risks [CSM, 2002].

The CSM [CSM, 2002] recommends not starting a statin if a baseline CK is greater than five times the ULN, and stopping the statin if the CK rises to greater than five times the ULN. If the CK returns to normal after stopping the statin, the CSM suggests that the statin may be introduced cautiously at the lowest dose with close monitoring. Management to follow in cases of smaller rises is less clear. Symptoms alone may require the statin to be stopped. If symptoms are not severe and the CK has risen, regular (for example, fortnightly) measurement of CK to assess the rate of rise would appear prudent. If a patient is symptom free or has mild symptoms and the CK stabilised at a level of less than 1000 IU/L it is frequently possible to continue the statin, although more regular checks of the CK would appear prudent, even in an asymptomatic patient. There is little guidance on the intervals for checking in such cases, and a period of six monthly is suggested pragmatically.

All references with links to [Free Full-text] are freely available online to users in England and Wales. This includes the full text of Department of Health papers and Cochrane Library reviews.

The high-density lipoprotein cholesterol (HDL-C) level is inversely related to coronary risk, and is a necessary measurement for use of coronary risk calculation tables and programmes. Until recently, the emphasis on treatment targets has been predicated on low density lipoprotein cholesterol (LDL-C) targets, although recent guidelines have highlighted the need to consider intervening when HDL-C is low. These question and answer sets examine indications for measuring HDL-C, and considers when intervention may be appropriate.

This guidance should be read in conjunction with reviews 1 and 3 of this series, which consider monitoring other aspects of cholesterol and triglyceride measurement [Smellie et al, 2005a; Smellie et al, 2006a], which are accompanied by a further discussion in a case study on lipids in a series of articles in the British Medical Journal [Smellie, 2006].

There are also separate topics on Lipid/cholesterol management in coronary prevention and Secondary hyperlipidaemia and triglycerides.

We recommend that when assessing cardiovascular risk:

In patients receiving lipid-lowering treatment, HDL-C should be measured:

Or as a minimum:

Large population studies have consistently demonstrated a strong inverse relationship between plasma HDL-C and the risk of CHD.

Low HDL-C levels may have a genetic cause and are also associated with increased triglycerides, obesity and very high carbohydrate intakes (> 60% of calories), physical inactivity, type 2 diabetes, cigarette smoking, and certain drugs (e.g. beta-blockers, anabolic steroids, and progestogens) [Thompson, 1989].

Studies and drugs available to examine the benefit of specifically treating low HDL-C levels in isolation are lacking. Most treatment recommendations are aimed at achieving target LDL-C levels for which there is convincing evidence of benefit.

Recent American [National Heart Lung and Blood Institute, 2006] and British guidelines [DH, 2000b; Joint British Societies, 2005] do not include specific recommendations on thresholds for initiating drug treatment, on treatment goals, and on monitoring intervals for people with low HDL-C levels and varying levels of cardiovascular risk, but make reference to the need for practitioners to be aware of the increased risk associated with low HDL in their treatment decisions.

Measurement of HDL is essential to accurately assess absolute cardiac risk and is necessary for the risk assessments recommended by the UK National Service Framework [DH, 2000b] and Joint British Societies [Joint British Societies, 2005].

HDL-C in the non-fasting state is lower by 5-10% than in the fasting state. Non-fasting measurements, therefore, slightly overestimate CHD risk but are regarded as sufficiently accurate to use in screening and are more convenient for patients [Pignone et al, 2001]. All people with an abnormal non-fasting screening lipids or at high risk should then have a fasting lipid profile [DH, 2000b; Joint British Societies, 2005].

The National Institute for Health and Clinical Excellence recommends that all people with diabetes should have an annual lipid profile (but make no recommendations on the management of a low HDL-C) [NICE, 2002c].

No clear guidance is available on this topic. As any intervention to raise HDL-C will normally run concurrently with other lipid management, we recommend that HDL-C levels for people undertaking interventions to raise HDL-C levels be maintained at the same intervals as for people who are receiving treatment to lower cholesterol and/or triglycerides - that is, 8 (+/- 4) weeks after starting treatment and then every 8 (+/- 4) weeks until on target.

The Joint British Societies guideline [Joint British Societies, 2005] recommends a full lipid profile (i.e. total cholesterol, HDL, triglyceride and measured or calculated LDL-C) be performed annually in patients on lipid-lowering treatment. It could be argued that repeated measurement of HDL-C is of limited benefit in those in whom specific intervention would be intended, if the total cholesterol adequately acts as a surrogate marker for LDL-C. As a minimum, we would therefore suggest measurement of HDL-C in patients in whom the total cholesterol figure may be an inaccurate surrogate for LDL-C (see [Smellie et al, 2006a] for brief review of calculated LDL-C or in whom HDL raising methods might be considered. In the absence of specific guidance, this is set arbitrarily at 1.2 and below (the potential intervention threshold in women) and 1.7 and above (when the LDL may be lower than expected as indicated by total cholesterol).

We recommend:

Modification: