Philip S. Lewis

Inequity in recording of risk in a local population-based screening programme for cardiovascular disease.

BACKGROUND Screening for cardiovascular disease is an important primary preventive measure, yet research has documented that not all population groups receive the same quality of preventive healthcare. DESIGN Longitudinal analysis of cardiovascular disease risk factor recording. METHODS Data were made available from a local population-based screening programme for cardiovascular disease (1989-1999), whereby residents aged 35-60 years were invited for screening every 5 years (n=84,646). Data were recorded for major risk factors including blood pressure, cholesterol, body mass index, smoking status, and alcohol consumption. Completeness of risk factor recording was compared between groups in the screened population defined by gender, ethnicity (Caucasian/South Asian) and employment status (employed/unemployed). RESULTS Recording of risk in the screened population was significantly less complete for women and South Asian participants over the duration of the screening programme, compared with men and Caucasian participants respectively. Conversely, recording of risk was significantly more complete for the unemployed compared with the employed participants. CONCLUSIONS These findings present evidence of a less systematic screening procedure for women and South Asians, whilst it seems that men, Caucasian participants and the unemployed were appropriately screened. Inequalities at the primary preventive level will likely influence outcome, because equitable identification of risk is important for the provision of successful treatment measures, and to reduce inequalities in morbidity and mortality due to cardiovascular disease.

Deprivation status and mid-term change in blood pressure, total cholesterol and smoking status in middle life: a cohort study.

Background Individuals of lower socioeconomic status have an adverse cardiovascular disease risk factor profile. We examined whether deprivation status influences within-individual change over time in blood pressure (BP), cholesterol and smoking status during middle life. Methods Records of participants of a primary care-based cardiovascular disease risk factor screening programme who were aged 35-55 years and had a first screening episode between 1989 and 1993 and a subsequent screening episode, were analysed. Deprivation status was defined using quintiles of the Townsend score. Using regression, mean annual change in BP, and total cholesterol was calculated for each deprivation group; and the effect of deprivation group status was examined. The probability of quitting smoking was also examined by deprivation group. Results Of all participants, 13 812 (72.1%) men and 16 932 (77.0%) women had complete follow-up (i.e. two screening episodes). Mean annual increase in systolic BP was significantly greater with increasing deprivation group [by +0.24 and + 0.28 mmHg/incremental deprivation group in men and women, respectively (95% confidence interval: +0.09 to +0.39 men, and +0.13 to +0.42 women)]. Deprivation status did not influence change in cholesterol (P = 0.620, men, P = 0.289, women). The probability of quitting smoking was significantly greater with increasing deprivation group in women [odds ratio 1.06 (95% confidence interval: 1.01-1.12)], but no effect was observed in men (P = 0.389). Discussion The results are suggestive of a ‘mixed’ picture of widening (e.g. systolic and diastolic BP) as well as narrowing (e.g. smoking in women) socioeconomic inequalities in cardiovascular risk factor inequalities. Eur J Cardiovasc Prev Rehabil 14: 844-850

Plethysmography without venous occlusion for measuring forearm blood flow: comparison with venous occlusive method

Limb blood flow is widely used as an indicator of the human vascular properties. There are only few non‐invasive methods for its measurement such as venous occlusion plethysmography. However, several authors have questioned its validity. The problems appear to be related to the process of venous occlusion. We developed two methods to measure forearm blood flow by plethysmography without venous occlusion in combination with Doppler velocimetry (without imaging). Method 1: the gradient of a tangent drawn on the latter part of the down stroke of the plethysmographic volume pulse is an approximation of venous blood flow in the absence of diastolic blood flow. At equilibrium, it equals the average arterial flow in a cardiac cycle. The Doppler velocity waveform recorded simultaneously allows improvement of this approximation when there is diastolic blood flow. Method 2: the volume pulse detected by a plethysmograph calibrated in absolute volume is used to calibrate the velocity waveform recorded simultaneously to produce an approximation of arterial volumetric flow waveform. Bland‐Altman analysis shows both methods have good correlation and agreement with venous occlusion plethysmography at rest. Method 1: mean difference (blood flow measured by venous occlusion minus calculated flow) = 0·10 ml/pulse (±0·18), limits of agreement = −0·41 and 0·61 ml/pulse. Method 2: mean difference = −0·041 ml/pulse (±0·15), limits of agreement = −0·45 and 0·37 ml/pulse. During hyperaemia, venous occlusion plethysmography grossly underestimated relative to the new methods. The new methods are not dependent on venous occlusion and produce consistent results with or without hyperaemia.

Blood flow measurement from plethysmographic pulse waves without venous occlusion.

An air plethysmograph with a sensitive phototransducer was constructed so that plethysmographic volume-change pulsations could be displayed in detail without using venous occlusion. Software was developed to allow analysis of the pulses using a modification of the backward extrapolation technique. This allowed calculation of the forward arterial blood flow and noninvasive derivation of the resting arterial flow waveform. There is good reproducibility of the technique, with 8% variability between pairs of measurements at rest and 4% variability after hand exercise. Direct comparison made with blood flows measured by venous occlusion plethysmography showed good average agreement. The mean blood flow for venous occlusion (rest and exercise) was 0.76 +/- 0.07 mL/beat (mean +/- SEM), and the mean blood flow for backward extrapolation (rest and exercise) was 0.74 +/- 0.09 mL/beat (mean +/- SEM). This corresponds to 3.86 +/- 0.36 mL/min/100 mL and 3.76 +/- 0.46 mL/min/100 mL, respectively. Important assumptions when using this method are that venous return is constant and that forward arterial flow is over before the end of the cardiac cycle.

Effect of different cuff types on blood pressure measurement: Variation in BP values for different cuff types:

Accurate measurement of blood pressure is very important as high blood pressure is a major cause of cardiovascular disease. Most non-invasive blood pressure measurement systems depend on a blood pressure cuff mostly made of fabrics but these differ widely in geometry and mechanical properties. To investigate the effect of differences in cuff construction on the arterial pressure, a numerical model using DICOM data obtained from magnetic resonance imaging scan is presented. The model arm was loaded with experimentally obtained interface pressures for eight different types of cuffs made of coated woven and non-woven fabrics. Variations of up to 27 mmHg in estimated systolic and up to 17 mmHg diastolic BP were found. Data were collected at inflation pressures between 80 and 130 mmHg to reflect the normal range of human systolic and diastolic blood pressure. Pressure transmission from the arm to the arterial surface was less than 90% under all the blood pressure cuffs except one with a rubber bladder which showed a 95% transmission ratio. The results from this study demonstrate wide variations in the arterial pressure values between cuffs of different types and cuffs supplied by a single manufacturer. It is crucial that a blood pressure cuff applies a known value of pressure around the arm. This study shows that blood pressure measurements may be critically affected by the fabric types used to construct blood pressure cuffs, leading to potentially serious errors in the assessment of cardiovascular risk.

146 Could Placement of the Blood Pressure Cuff Over the Elbow Rather than the Upper Arm Increase the Accuracy of Automatic Blood Pressure Machines in the Detection of Systolic Blood Pressure

Automatic blood pressure monitors (oscillometers) with the cuff placed over the upper arm give different systolic BP (SBP) readings to the Korotkoff/Riva-Rocci auscultatory method, influenced by quantity of soft tissue and arm circumference. Other studies have examined whether SBP measurement at the wrist or forearm is more accurate, with limited success. No literature was found on oscillometer use at the elbow. We therefore examined whether oscillometric SBP at the elbow was more accurate than the upper arm and the influence of demographic and anthropometric variables, using simultaneous Doppler return-to-flow at the wrist as the comparative method for SBP. Triplicate SBP readings at the upper arm and elbow were recorded in 62 participants. Bland-Altman plots showed similarly poor agreement between the oscillometers and Dopplers at each cuff position, despite upper arm oscillometry being a clinically validated method of blood pressure measurement. We found a significant difference (p < 0.01) between the mean oscillometer SBP readings at the upper arm and elbow, and between the mean Doppler SBP readings at the upper arm and elbow. On further analysis, we noticed that the difference between the oscillometric readings at the upper arm and elbow had an inverse relationship with both log transformation of triceps skinfold thickness (p = 0.013) and age (p = 0.014), although correlation between these two variables was not significant (p = 0.193), therefore avoiding collinearity. No relationship was found between the difference between Doppler readings and any anthropometric variables measured. Placing the cuff over the elbow may therefore be a feasible alternative to a cuff over the upper arm, especially in those with a large quantity of non-muscular soft tissue and older people. Additional development of this method and further comparison is needed as we must adapt to the increasingly obese and elderly population.

The impact of screening for cardiovascular disease risk factors on population health and inequality. Evidence from the Stockport Screening Programme, United Kingdom

Background : Population based screening for cardiovascular disease risk factors can potentially reduce coronary heart disease mortality and morbidity. There is little contemporary evidence that has examined the actual impact of such a programme on population health and on reducing inequality in health between affluent and deprived areas. Methods : 82,015 residents of Stockport Health Authority, UK age 35-60 took up an invitation to be screened for cardiovascular disease risk factors from 1989 -1999. We compared uptake of screening and coronary heart disease (CHD) mortality and hospital admissions (1997 – 2003) between screened and unscreened male and female populations from affluent and deprived areas. Results : Males and females in the unscreened population were more likely to die from CHD (IRR=3.60; p<0.001, IRR=4.64, p<0.001) and to have a hospital episode (IRR=1.75, p<0.001, IRR=1.94, p<0.001) than those in the screened population. This was independent of age and deprivation. The highest rates of CHD mortality and hospital admissions were found for unscreened deprived populations, the lowest for screened affluent populations. For both males and females mean rates of CHD mortality and hospital admissions were significantly lower for those who were screened and living in deprived areas compared to those who were unscreened and living in affluent areas. Conclusions : Screening for cardiovascular disease risk factors improved the cardiovascular health of the population by targeting and treating ‘high risk’ groups, including those living in deprived areas. The potential of screening to reduce health inequality by promoting faster and more substantial health improvement in deprived areas was not observed in this study.