Deepak Bhatnagar
Consultant in Metabolic Medicine & Clinical Biochemistry
Janet Morgan
Research Sister
Sohaib Siddiq
Medical Student
Michael I. Mackness
Lecturer in Biochemistry
J. Paul Miller
Consultant Gastroenterologist
Paul N. Durrington
Professor of Medicine

University of Manchester, Department of Medicine,
Manchester Royal Infirmary, Oxford Road,
Manchester M13 9WL

Department of Gastroenterology, University Hospital of South Manchester,
Nell Lane, Withington, Manchester M20 2LR

Correspondence to: Professor Durrington

Tel: 0161 276 4226
Fax: 0161 274 4833

pdurrington@hq.cmht.nwest.nhs.uk

ABSTRACT

Objectives

To assess the feasibility of detecting new cases of heterozygous familial hypercholesterolaemia (FH) using a nurse-led genetic register technique.

Setting

Case-finding amongst relatives of patients attending two Lipid Clinics in Central and South Manchester.

Design

Patients with a clinical diagnosis of definite FH attending the Lipid Clinics between 1987 and 1998 (probands) were interviewed by a trained nurse. Names and contact details of all first-degree relatives were sought. These relatives were asked to attend one centre (the Manchester Royal Infirmary) for the measurement of blood cholesterol or to arrange for their General Practitioner to obtain the blood and send it to the centre. Relatives and probands also completed a standard questionnaire and were examined by the nurse or general practitioner, who was provided with a pictorial guide to the recognition of tendon xanthomata.

Subjects

Of 259 (137 men and 122 women) probands 216 (83.4%) had at least one first-degree relative extant. There were 285 first-degree relatives of whom 200 (70.2%) were successfully tested.

Results

Of the 200 first-degree relatives tested, 121 (60%) had inherited FH. The newly diagnosed patients were younger than the probands and were generally detected before they had clinically overt atherosclerosis. Serum cholesterol was 8.4± 1.7mmol/l and 8.1± 1.9mmol/l in affected men and women respectively and in unaffected men and women 5.6± 1.0 and 5.6± 1.1mmol/l. Coronary heart disease (CHD) risk factor screening as recommended in recent guidelines for CHD prevention would have failed to identify the majority of affected relatives in whom hypertension, diabetes mellitus, cigarette-smoking and obesity were uncommon.

Conclusions

121 new cases of FH were discovered from 200 cholesterol tests. With a frequency of 1 in 500 in the UK population over 60,000 tests would be required to detect a similar number by population screening and only a few would have been detected had cholesterol testing been restricted to those with other CHD risk factors. There is a case for organising a similar genetic register approach linking Lipid Clinics nationally.

INTRODUCTION

Familial hypercholesterolaemia in its heterozygous form has a frequency in Europe and North America of around 1 in 500, making it the most common potentially lethal genetic disorder. The characteristic clinical syndrome in adulthood comprises a raised serum cholesterol, tendon xanthomata and premature coronary heart disease (CHD), the median age of onset for the latter being around 50 years in men [1] and some 9 years later in women [2]. Statin treatment and the opportunity for prompt access to cardiological services afforded to patients by the diagnosis of FH appear to have resulted in improved survival [3]. In coronary angiographic trials cholesterol-lowering treatment is at least as effective in FH as in other type of patients [4,5]. FH thus satisfies the main criteria for screening described by Wilson and Jungner [6]. Only a small proportion of the potential 100,000 patients in the UK with FH appear to attend Lipid Clinics [7] and the majority are probably undiagnosed. The same is also likely to be true in other countries [8]. Frequently the clinical syndrome of FH is due to a mutation of the low density lipoprotein (LDL) receptor, but because more than 200 different mutations have been described in association with the syndrome [9] genetic testing is not currently a feasible means of establishing the diagnosis, except perhaps in families with a known mutation or in societies with a more limited number of mutations due to a founder gene effects or where consanguinity is common [10-12]. A previous report based on our clinic population revealed the most common LDL receptor gene mutation to have a prevalence of only 3.9% [13].

It is generally agreed that screening the population for high cholesterol should only be undertaken as part of a multifactorial approach to the detection of high CHD risk so that the most cost-effective use of cholesterol-lowering and antihypertensive treatment can be made [14]. However, FH appears to be a condition in which a single risk factor (high cholesterol from the moment of birth) often leads well before middle-age to an absolute CHD risk in the range for statin treatment [1,3]. We have therefore conducted a study to assess the possibility of using a genetic register method to diagnose new cases of FH which has the potential to be adopted nationally.

METHOD

Probands aged 18 years or more attending two adjacent Lipid Clinics (at the Manchester Royal Infirmary and the University Hospital of South Manchester) for the first time between 1987 and 1998 were identified using the Simon Broome criteria for the diagnosis of definite FH [15]. For patients aged more than 16 years these are serum cholesterol exceeding 7.5mmol/l (or LDL cholesterol exceeding 4.9mmol/l) plus tendon xanthomata in the patient or in a first- or second-degree relative. In none of the probands enrolled did the diagnosis depend on tendon xanthomata in a second-degree relative.

Nurses were trained to identify corneal arcus, xanthelasmata and tendon xanthomata and to administer a questionnaire to probands and their first degree relatives which enquired about the presence of other risks factors for CHD and cardiovascular disease already evident. In the case of the probands a detailed family history was recorded including addresses and wherever possible telephone numbers of any first-degree relatives. In order to do this special time was set aside from routine clinic visits and the reason for this was explained in advance. The risk factors recorded were hypertension requiring drug therapy, cigarette-smoking and a physician diagnosis of diabetes mellitus. CHD was recorded as previous acute myocardial infarction diagnosed in hospital, a physician diagnosis of angina of effort or coronary artery bypass surgery. A previous diagnosis of stroke by a physician was also recorded and the possible presence of intermittent claudication was sought using the Rose questionnaire.

Each first-degree relative was then sent a personalised, standard letter explaining the reason for suspecting that he or she may have FH, its importance and the method of diagnosis. A daytime telephone number was provided for enquiries and to inform the nurse of whether the relatives preferred to visit her or to be seen at their general practice. If they chose to see the nurse they attended the Manchester Royal Infirmary where the questionnaire was completed and a fasting blood sample taken. Otherwise the questionnaire was sent by post and the blood sample was taken at their general practice and sent to the nurse in Manchester. General practitioners of these relatives received a letter explaining FH, its clinical features (including coloured photographs of corneal arcus, xanthelemata and tendon xanthomata in the Achilles tendons and in the dorsum of the hands) and details of the register together with a blood specimen container, venepuncture equipment and secure prepaid packaging to send the blood sample by first class post.

In every case the relatives’ serum cholesterol results were sent to general practitioners with a letter explaining why the test hand been done and the significance of the result. The location of the nearest Lipid Clinic was provided when the test proved positive. General practitioners also had the option of treating newly diagnosed patients themselves with advice, if requested. None, however, chose to do this, preferring to refer the patient to hospital. Relatives with newly diagnosed FH were sent a letter indicating that their cholesterol was raised and suggesting they make an appointment to see their general practitioner. Counselling was also available by telephone and the general practitioner was also given the same telephone number, should additional information be required. Relatives not inheriting FH were also informed and the significance of that explained: they too were invited to telephone should they wish for further explanation.

The research ethical committee at both hospitals considered that the register was an extension of usual clinical practice.

Cholesterol in serum and high density lipoprotein (HDL) and triglycerides in serum were measured enzymatically by the CHOD-PAP and GPO-PAP methods respectively (both from Roche Diagnostics, Lewes, UK). HDL was isolated from serum by heparin manganese precipitation of the other lipoproteins and low density lipoprotein (LDL) cholesterol was calculated using the Friedewald formula [16]. Serum apolipoprotein B (apo B) was determined using rate nephelometry using the Beckman Array and reagents (Beckman Instruments Ltd, Palo Alto, California, USA) and serum lipoprotein (a) (Lp(a)) by an IRMA method (Mercodia, Uppsala, Sweden). At the time of referral some probands were already receiving treatment with cholestyramine or statins. Despite this in every case the Simon Broome criteria for diagnosis were satisfied. The contemporary laboratory values are quoted.

STATISTICS

Variables with a gaussian distribution were compared using Student’s t test and those which were non-gaussian by the Mann Whitney U test. Frequency distributions were compared by the X² tests. Probabilities £ 0.05 were considered statistically significant.

RESULTS

Compliance of probands and availability of relatives
Of 262 FH probands identified all but 3 agreed to participate. Thus 259 (99%) (137 men and 122 women) provided details of their family tree. Of these 216 (83.4%) had at least one living first degree relative, the total number of whom was estimated to be 285. Of these 205 (71.9%) were tested. The reasons why some were not available for testing and why some probands could not provide relatives are shown in table 1. In 26% of cases more than one relative of a proband was tested. The majority of relatives (98%) preferred to visit the nurse often accompanied by the proband.

Detection of new cases
Of the 205 relatives tested cholesterol results were available in 200 of whom 121 (60%) (46 men and 75 women) proved positive (heterozygotes by definition) and 79 (40%) (37 men and 42 women) had serum cholesterol levels less than 7.5mmol/l (Table 2). Male probands were less likely to provide a co-operative relative than female probands. Thus 137 male probands yielded 46 new cases (ratio of probands to new cases 3.0:1) whereas 122 female probands produced 75 new cases (ratio 1.6:1). The difference in the number of new cases detected for male and female probands was statistically significant (P<0.0005).

Clinical Characteristics of Probands and Relatives
Tendon xanthamata were present in 91% of male and 87% of female probands whereas only 26% of the newly diagnosed men and 19% of the newly diagnosed women possessed them (P<0.0005) (Table 3). This was probably because of the younger age of the newly diagnosed relatives (P<0.0001) (Table 4). It is notable that many probands had still not developed either corneal arcus or xanthelasmata at a stage of their disease when tendon xanthomata were evident (Table 3). In neither the probands nor the affected and unaffected relatives was obesity prevalent (Table 4). Male and female probands and newly diagnosed relatives with FH had raised cholesterol, LDL cholesterol and apolipoprotein B and both affected groups of men also had lower serum HDL cholesterol levels than their unaffected relatives (P<0.05 and <0.01 respectively). Serum cholesterol and LDL cholesterol levels were similar in probands and affected relatives despite the fact that some probands were receiving cholesterol-lowering treatment albeit generally with low doses of statin or cholestyramine when they were first studied. It was considered unethical to discontinue these medications for the purpose of this investigation. It is possible that the cholesterol level off treatment in the proband would have been higher than in their younger affected relatives, because of the serum cholesterol rise which occurs in middle age. Lipoprotein (a) was also significantly lower in male and female affected and unaffected relatives than in probands (both P<0.01). The mean values for unaffected men and women was similar to those in a normal healthy populations previously studied by us [17]. The affected relatives were intermediate with respect to their serum lipoprotein (a) concentration. We have previously reported that when matched for age probands and affected relatives have similar levels [18]. Thus our present finding may lend some support to the view that Lp(a) rises with advancing arterial disease, which would explain its association with CHD in some case studies in FH [19].

Cardiovascular disease was significantly more common in probands than in newly diagnosed affected relatives and in unaffected relatives, particularly coronary heart disease (angina and/or myocardial infarction and/or CABG). Cardiovascular disease was significantly more common in probands than in newly diagnosed affected relatives and in unaffected relatives, particularly coronary heart disease (Table 5).

There was a low prevalence of cardiovascular risk factors apart from raised cholesterol in either the probands or their newly diagnosed relatives (Table 5). The serum cholesterol would thus be unlikely to be measured in the newly diagnosed relatives in the UK except in those with overt CHD. Furthermore the average CHD risk in the newly diagnosed relatives according to the risk equation on which the UK, US and European guidelines are based [20] was 6% over 10 years for men and 3% over 10 years for women.

DISCUSSION

The investigation indicated that a genetic register based on 262 FH probands attending a Lipid Clinic could identify 121 new cases. Heterozygous FH affects approximately 1 in 500 of the general population. Thus to attempt to identify 121 new cases by universal population screening for high serum cholesterol would require that more than 60,000 cholesterol tests were performed whereas only 200 tests were necessary in the present study. Selective screening for high cholesterol by confining cholesterol testing to patients in whom other cardiovascular risk factors or CHD are present would have missed all of the cases identified in our investigation with the exception of a small proportion with established CHD or hypertension. Furthermore, had a detailed family history been obtained from newly diagnosed relatives with FH discovered in this study, it is likely that additional first-degree relatives could have been discovered who could be tested and so on, amplifying the number of new cases detected. Applying a similar method in other Lipid Clinics nationally would give access to new probands and could lead to a very considerable increase in the number of known FH heterozygotes.

The typical male and female heterozygote discovered in the present investigation would appear to have a CHD risk of only 6% or 3% respectively over the next 10 years, if calculated according to the Framingham risk equation [20] upon which current European and New Zealand guidelines about CHD prevention are based [14,21,22], whereas it is known from other investigations that they were likely to develop clinical CHD at a similar age to their probands [2], which means their true risk was several times greater. The discovery of raised cholesterol in an heterozygote for FH is not therefore likely to lead to appropriate therapy unless the clinician assessing the significance of the finding is aware that the patient has FH rather than polygenic hypercholesterolaemia. Our strategy ensures that this is the case.

In the present investigation relatives were required to fast. This was in order to report their fasting serum triglycerides needed to calculate LDL cholesterol [23]. This would not be necessary simply to identify new cases in practice, because the total serum cholesterol which is unaffected by fasting provides enough biochemical information for the Simon Broome definition of FH [15]. However, it is important to realise that this definition does not rely simply on a cholesterol value exceeding 7.5mmol/l. This is relatively common in Britain and would only indicate the presence of FH, if it occurs in a patient related to a definite case of FH, established in our study by the presence of tendon xanthomata. Except in FH tendon xanthomata are exceedingly rare occurring, otherwise only in phytosterolaemia and cerebrotendinous xanthomatosis [24]. Patients with serum cholesterol levels of 7.5mmol/l or even higher who do not have FH and no other cardiovascular risk factors are generally at much lower risk than heterozygotes for FH particularly at the comparatively young age of the new cases of FH discovered in our study: they would thus seldom require treatment with statins. Again our strategy of screening only relatives of FH probands ensures that inappropriate treatment and advice is not offered to people with less severe syndromes associated with hypercholesterolaemia.

The finding that male probands were less likely to provide an affected relative than females was probably because men were less likely to provide sufficient details of a relative for them to be traced, perhaps because their wives write the Christmas cards. A possible improvement to the present strategy might therefore be to ensure that wives were, if possible, present when male probands are interviewed.

The high prevalence of cardiovascular disease in probands is likely to be the result of the older age of the probands compared to newly diagnosed relatives and because their hypercholesterolaemia was discovered as the consequence of presenting with vascular symptoms. It has previously been reported that the age of onset of symptomatic CHD is similar in affected first degree relatives within individual families [2]. The present findings thus suggest that this method of detecting new cases frequently identifies them before vascular disease is clinically overt which, given the mortality associated with a first myocardial infarction which is around 30% [25] and the subsequent morbidity is a potentially important advantage.

It has been calculated that the cost per life year gained from cholesterol reduction in FH is similar to that in patients after acute myocardial infarction [26,27], which is generally considered to be highly cost-effective: more so for example than the cost of a generic thiazide to treat hypertension [28]. There are potentially detrimental effects of screening [29,30]. Our approach avoids the adverse effects caused by general population screening leading to the discovery of huge numbers of asymptomatic people with more common less severe hypercholesterolaemia in which the health gain from such knowledge may be minimal. Furthermore although our decision not to employ DNA methods for the detection of FH was pragmatic, it meant that our approach also avoided the potential psychological harm caused by DNA testing [11,31]. Discovery of FH by case-detection probably has relatively brief adverse psychological effects [31,32], but such reassuring findings have generally been reported when counselling was available. It is likely that such counselling will be most effective when provided by healthcare workers who have frequent contact with patients with FH and this is another potential advantage of the detection of new case of FH through established Lipid Clinics using the genetic register approach reported here.

ACKNOWLEDGEMENTS

We are grateful for support from an NHS Research and Development Grant (PS004, North West Regional Health Authority) and from the NHS Research and Development Levy. We thank Ms C. Price for expertly preparing this manuscript and Sisters Mary Brady, Pat Lockely and Morag Ravenscroft who provided additional nursing support. Copies of the standard letters to relative and general practitioners are available from Professor Durrington on request.

REFERENCES

Table 2 Results of testing 200 first degree relatives of FH patients diagnosed on clinical criteria
 

Table 3 Physical signs of hypercholesterolaemia in FH probands and their affected and unaffected relatives
 

Table 4 Age, body mass index, lipids and lipoproteins in FH probands, affected and unaffected relatives. Mean ± SD or median (interquartile range)

Significantly different from probands *P<0.05 **P<0.01 ***P<0.005
 
 
 

Table 5 Prevalence of major cardiovascular risk factors and diseases amongst FH probands, affected and unaffected relatives
 

CABG = coronary artery bypass grafting CHD = coronary heart disease
* Angina and/or Myocardial infarction and/or CABG
** Rose questionnaire
*** Coronary heart disease and/or stroke and/or intermittent claudication
Significantly greater than affected or unaffected relatives (x² test) ⁺P<0.01 ⁺⁺P<0.001 ⁺⁺⁺P<0.0001