Dear Editor

In their recent article,[1] Martin et al. describe an interesting modelling exercise to estimate the need for coronary revascularisation based on the incidence of indications. We have used this population perspective model to help inform our current review of angiography services in Northumberland by using local data and report here the experience of using this model as a planning tool for angiography commissioning.

Instead of comparing current practice to the model of need as in Martin’s paper, we compared maximum and minimum values of parameters available for all three indications up to Parameter 4 (number of those referred to secondary care undergoing angiography).

For Parameters 1 and 2, incidence and prevalence data was extracted from local Hospital Episode Statistics (HES) and the Northumberland MEDICS project[2] which is a primary care data collection system that regularly feeds back comparative reports of morbidity and clinical care indicators to practices. Where data was unavailable, four GP practices which record high quality ischaemic heart disease computerised data supplied rates which were extrapolated to the Northumberland population. Referral rates for chronic stable angina quoted in Martin’s paper were used as upper and lower limits. Parameter 1 was not computed for acute MI and unstable angina as the sources of data fed into Parameter 2 only. We assumed 100% survival for Parameter 3 in chronic stable angina. Local HES data survival rates in the other indications groups was compared to that supplied by Martin. Angiography referral rates in Parameter 4 for cardiac origin chest pain from rapid access chest pain clinic reports were utilised for chronic stable angina. The upper limit was kept at 51% following a previous review of angiography rates in Newcastle upon Tyne which showed a similar result. We had no local data for acute MI and unstable angina referral rates so Martin’s figures were used.

The range of angiography need was wide (269 to 1294). The highest figure would be the maximum number of angiographies needed if best practice is applied to all patients and is therefore aspirational. The key step where more detailed data would make the tool more precise for planning is the chronic stable angina referral rate to secondary care. If the lower referral limit of 33% (current practice based on the general Practice Research Database) and the maximum number of incident cases for chronic stable angina are used, the total number of angiographies required would be between 269 and 870. Our current projected angiography rate of 512 in the last financial year sits in the middle of this range. Better local recording of angiography rates for acute MI and unstable angina would also improve the model’s precision.

Using the model as a planning tool obviously has some limitations, which Martin acknowledges will either over or under estimate the need. In addition to these is the effect that the RITA 3[3] trial may have on increasing intervention during acute coronary syndromes. The changing diagnosis of Acute MI could also lead to the misclassification of diagnoses. The generalisability of extrapolating incident data from a few GP practices is questionable but the rates did agree with population data except for unstable angina where the GP rates were grossly less than the HES data and were therefore excluded. Because of the limitations of our local data, non-standardised rates were used throughout our model.

We felt overall the model was useful as a guide and helped focus our thoughts on angiography from a population perspective. It helped compare current with aspirational practice. However, from a commissioner’s perspective, the maximum rates are aspirational and major changes in the referral patterns from primary care would have to take place to meet them. We recommend further studies look at current and optimum referral rates for chronic stable angina based on need.

References

(1) Martin RM, Hemingway H, Gunnell D, Karsch KR, Baumbach A, Frankel S. Population need for coronary revascularisation: are national targets for England credible? Heart 2002;88:627-633.

(2) Edwards R, Murphy P, Allan K, Gordon S, Singleton S. Comparative morbidity data in primary care: the Northumberland MEDICS project. Primary Health Care Research and Development 2002;3:217-227.

(3) Fox KA, Poole-Wilson PA, Henderson RA, Clayton TC, Chamberlain DA, Shaw TR, Wheatley DJ, Pocock SJ, Randomized Intervention Trial of unstable Angina I. Interventional versus conservative treatment for patients with unstable angina or non-ST-elevation myocardial infarction: the British Heart Foundation RITA 3 randomised trial. Randomized Intervention Trial of unstable Angina. Lancet 2002;360:743-51.

Dear Editor

We are grateful to Dr Francis for his response to our analysis [1] of the real implications, both clinical and financial, of implementing a policy of implanting drug-eluting stents. He has stated the case rather more starkly than we did, but we would agree that the main problem is pricing. If these new stents were only slightly more expensive than the current (rather satisfactory) generation of conventional stents, they would be implanted universally. At the price at which they have been pitched, the manufacturers are forcing us to analyse which patients and lesions we should target with a policy of limited implantation. This requires a detailed knowledge of the risks and benefits inherent in tackling the problem of restenosis – a problem which we demonstrated, in our analysis, to be one which obeys a law of diminishing returns. For sure, limited extra expenditure will reduce the restenosis rate, but only marginally; stent expenditure would have to triple or quadruple to eliminate the last few patients with restenosis. Is this justified, when many more cases could be treated with conventional technology with the same budget?

We would not want to give the impression that current drug-elution costs are ‘never financially justifiable’; rather, they make us sit up and think about what we are actually achieving with this huge increase in expenditure. We are not sure NICE will help selectivity. NICE may well endorse the use of this product in the lesion types included in RAVEL [2] and SIRIUS [3] trials. We will then have a clinical mandate to use them in a large proportion of our patients, but this will not solve the financial dilemma. Competition with other devices is the most likely long-term solution to this conundrum.

References

(1) Gunn J, Morton A, Wales C, Newman C, Crossman D, Cumberland D. Drug-eluting stents: maximising benefit and minimising cost. Heart 2003;89:127-31.

(2) Morice MC, Serruys PW, Sousa JE,et al. A randomised comparison of a sirolimus-eluting stent with a standard stent for coronary revascularization. N Engl J Med 2002;346:1773-80.

(3) Moses J, Leon M, et al. The SIRIUS trial. Results presented at Transcatheter Cardiovascular Therapeutics meeting, Washington DC, September 24, 2002.

Dear Editor

In their interesting study, Illien et al.[1] recently reported that among 302 consecutive patients with nonrheumatic atrial fibrillation(AF), 8 were detected who had no clinical risk factors for stroke/embolism (age>65, previous stroke/embolism, arterial hypertension, diabetes mellitus, heart failure) but a transesophageal echocardiographic(TEE) finding of a “thrombogenic milieu”, either spontaneous echo contrast(SEC) or decreased left atrial appendage(LAA) flow velocities. From these findings, the authors conclude that AF patients without clinical risk factors and a thrombogenic milieu constitute a subgroup of patients with increased risk for stroke/embolism. Patients of this subgroup will be missed at clinical risk stratification. Thus, the practical consequences of these assumptions would be to perform TEE in all AF patients without clinical risk factors for stroke/embolism in order to look for a thrombogenic milieu. There are several objections that have to be raised against these hypotheses:
1. The authors did not demonstrate that in any of these 8 patients stroke/embolism had occurred, neither in their previous history nor at follow-up investigations, which were not carried out. Based on this lack of information, it is not justified to conclude that these patients have an increased thromboembolic risk.

2. We have previously shown in a prospective study including 409 AF patients that SEC and LAA thrombi were no independent risk factors for stroke/embolism during a follow-up period of 5[2] and 10[3] years.

3. According to the retrospective design of Illien’s study, thromboembolic risk was only assessed by presence or absence of clinical risk factors. It has been repeatedly confirmed that there is an association between the prevalence of spontaneous echo contrast, LAA thrombi, decreased LAA velocities and clinical risk factors for stroke/embolism. However, no prospective study demonstrated so far, that spontaneous echo contrast and decreased LAA flow velocities are risk factors for stroke/embolism independent from the clinical risk factors.

4. Additionally, no clinical information is provided about the 8 patients with thrombogenic milieu in the absence of clinical risk factors for stroke/embolism. Particularly, it is not mentioned, whether they were treated with betablockers, a condition reported to induce SEC, decrease LAA flow velocity and promote thrombus formation.[4]

5. Furthermore, measurements of LAA flow velocities in AF are flawed by influences like heart rate and especially the duration of diastole [5]. It is not mentioned, how many heart cycles were taken for calculation of the mean LAA flow velocity, which is of particular importance when investigating AF patients. Blood pressure values, which also influence LAA flow velocity, are not given [4,5].

From these objections we conclude that it is still reasonable to assess the risk for stroke/embolism of AF patients on well established clinical criteria and not on TEE findings. It would be useful, however, to assess the risk for stroke/embolism in patients without clinical risk factors for stroke/embolism, but with spontaneous echo contrast or decreased LAA flow, by means of a prospective long-term follow-up study.

References

(1) Illien S, Maroto-Järvinen S, von der Recke G, Hammerstingl C, Schmidt H, Kuntz-Hehner S, Lüderitz B, Omran H. Atrial fibrillation: relation between clinical risk factors and transesophageal echocardiographic risk factors for thromboembolism. Heart 2003;89:165-8.

(2) Stöllberger C, Chnupa P, Kronik G, Brainin M, Finsterer J, Schneider B, Slany J, for the ELAT Study Group. Transesophageal echocardiography to assess embolic risk in patients with atrial fibrillation. Ann Intern Med 1998:128;630-8.

(3) Stöllberger C, Chnupa P, Abzieher C, Länger T, Finsterer J, Klem I, Hartl E, Wehinger C, Schneider B. Mortality and rate of stroke or embolism in atrial fibrillation during long-term follow-up in the ELAT (Embolism in Left Atrial Thrombi) study. Clin Cardiol 2003 (in press).

(4) Bilge M, Güler N, Eryonucu B, Erkoc M. Does acute-phase beta blockade reduce left atrial appendage function in patients with chronic nonvalvular atrial fibrillation? J Am Soc Echocardiogr 2001;14:194-9.

(5) Noda T, Arakawa M, Miwa H, Ito Y, Kagawa K, Nishigaki K, Hirakawa S, Fujiwara H. Effects of heart rate on flow velocity of the left atrial appendage in patients with nonvalvular atrial fibrillation. Clin Cardiol 1996;19:295-300.

Dear Editor

Herlitz and colleagues [1] paint a rather gloomy picture of the impact of 20 years of effort in managing out of hospital cardiac arrest. Their large population observations are very valuable and will have a familiar feel to many practising cardiologists. The study showing survival rates over the last 20 years have not improved is comparable to previous studies.[2,3] If survival rates for victims of out- of-hospital cardiac arrest are to improve, then one possible avenue is to explore in greater detail the events leading up to the arrest. One aspect we feel should be highlighted more clearly is the critical impact of patient contact prior to any event.

Patients with known illness may contact support services early or may defer contact such that arrest occurs in an unmonitored setting clearly putting resuscitation attempts and survival at a disadvantage. This is poorly defined in what is otherwise an extensively researched area. Second, those patients with no prior medical contacts may react differently again either presenting with undiagnosed or ‘atypical’ symptoms to one or other health professional. They may be treated for a non-cardiac diagnosis that proves incorrect only in retrospect. Whilst data on previous medical history (acute myocardial infarction, angina pectoris, congestive heart failure and smoking status) were included in Herlitz database there was typically no data on victims with no documented previous medical history, and in particular whether these victims have any contact with medical services in the hours or days prior to the event. While information from relatives may be difficult to collect and perhaps painful for the individual families, they might provide useful details on circumstances prior to an individual arrest.

What we need to define by such work is whether there is a window of opportunity for early intervention in these events, which could be critical in defining individual outcome. It is not known what proportion of victims of out-of-hospital cardiac arrest have occult cardiac or non- cardiac disease whose first presentation is the cardiac arrest. This would be an important finding as it would mean that appropriate diagnostic effort could potentially abort these events. Unfortunately there was minimal data in the current report on a final pathological diagnosis on discharge from hospital.

The second useful point to emerge from Herlitz’s work involves the poorer survival associated with diuretic use. The authors understandably suggested the common association with cases of ‘congestive heart failure’. This is a common presumption but in many studies a significant proportion of people who are on diuretic therapy do not have any evidence of left ventricular dysfunction even with a presumed ‘heart failure’ diagnosis in the community.[4] That impaired systolic function is strongly associated with sudden cardiac death is not at issue. Thiazide diuretic treatment is far more often a marker of hypertension and this would be more likely to be associated with occlusive coronary events than primary or secondary arrhythmia in systolic dysfunction. Clearly this is important in terms of the best therapeutic strategies that for coronary occlusive disease would be accelerated revascularisation rather than for example device management of arrhythmia in an already impaired left ventricle. Unfortunately the population data presented by Herlitz or indeed from any comparable database do not allow this to be defined.

In summary Herlitz large epidemiological survey serves to underline the pressing need for new approaches to out of hospital arrest over and above the fine tuning of resuscitation algorithms which are basically showing little real progress despite the huge efforts of those concerned. The published data look bleak on first reading but the outcome may be improved by more detailed individual case recognition and should start with more research on the perception of illness by patients in these critical cases. Are all sudden arrests truly sudden/instantaneous or could there be an early pattern of atypical symptoms in some prior to the arrest? Surely Herlitz summary data shows well enough that continuing with the current patterns is just not good enough?

References (1) Herlitz J, Bång A, Gunnarsson J, Engdahl J, Karlson BW, Lindqvist J, Waagstein L. Factors associated with survival to hospital discharge among patients hospitalised alive after out of hospital cardiac arrest: change in outcome over 20 years in the community of Göteborg, Sweden. Heart 2003;89:25-30.

(2) Nichol G, Detsky AS, Steill IG, O’Rourke K, Wells G, Laupacis A. Effectiveness of emergency medical services for victims of out-of-hospital cardiac arrest: a metaanalysis. Ann Emerg Med 1996;27:700-710.

(3) Eisenberg MS, Horwood BT, Cummins RO, Reynolds-Haertle R, Haerne TR. Cardiac arrest and resuscitation: a tale of 29 cities. Ann Emerg Med 1990;19:179-180.

(4) Davies MK, Hobbs FDR, Davis RC, Kenkre JE, Raolfe AK, Hare R, Wosornu D, Lancashire RJ. Prevalence of left-ventricular systolic dysfunction and heart failure in the Echocardiographic Heart of England Screening Study: a population based study. Lancet 2001;358:439-44.