961
THE.LANCET
Hyperviscosity
in Disease
Two years ago we reviewed the factors which influence blood viscosity in health and diseased Since then, several more reports have linked viscosity
changes to disease. But, intriguing as these are, we still await proof that changes in viscosity precede the clinical manifestations and are the prime factor. Usually a raised viscosity is secondary to some other lesion. In sickle-cell disease, for instance, though an increase of blood viscosity precedes peripheral occlusion,2the basis of the disease is an abnormal haemoglobin. Manipulation of blood viscosity, though it may relieve some of the complications, is unlikely to be a complete solution. Furthermore, we still know little about how to change blood viscosity for optimum blood-flow. Reducing the hxmatocrit or plasma-fibrinogen is not necessarily beneficial. This week a group from Queen Square present evidence that patients with a haematocrit over 46% tend to have a low cerebral blood-flow; and they attribute the depression of blood-flow to the increased peripheral resistance which results from raised viscosity. By reducing the haematocrit from 49.3% to 42-6% they raised the mean cerebral blood-flow by half-from 41.1 to 62.1ml/100 g/min. In the succeeding article, a St. Mary’s group likewise link hyperviscosity with haematocrit, in patients with angina. The relation between haematocrit and blood-flow is probably more complex than these workers suggest; viscosity may be only part of the mechanism. Cardiac output is determined approximately by arterial pressure divided by peripheral resistance. The main pressure drop occurs across the vessels of the microcirculation. Peripheral resistance is influenced by the geometry and size of these vessels and by blood viscosity. Both groups reporting this week assume that the change of peripheral blood-flow results directly from an effect of haematocrit on blood viscosity; but this may be wrong. Firstly, there is evidence that the haematocrit is lower in arterioles and capillaries than in large arteries and veins. There is no evidence that an increase in the arterial-venous haema1. Lancet, 1975, i, 113. Chien, S., Usami, S., Bertles, J. F. J. clin. Invest. 1970, 49, 623.
2
tocrit produces a corresponding increase in the haematocrit of the microcirculation. Even if it did, this would not greatly alter the blood viscosity in this region of the circulation. The influence of hxmatocrit on blood viscosity falls steadily with decreasing vessel bore,until in vessels of 4 to 15 (im bore it has no effect.3 Similarly the influence of shear-rate is unlikely to be important. Irrespective of conflicting views as to whether the shear-rate in capillaries is high or low, the effect of shear-rate decreases with hsematocrit, and hsematocrit, as we have said, is reduced. The physical nature of flow in small vessels is not the same as that in large arteries and veins. The relation of flow-rate to driving-pressure in small vessels is linear, irrespective of the haematocrit. 3,4 The suggestion that a reduction of haematocrit gives rise to an increase in cardiac output,S,7 owing to a reduction of blood viscosity, also requires more convincing evidence than we have yet seen. The increased blood-flow is more likely to be due to lowered oxygen transport, associated with the lowering of the haemoglobin content ; and to assess this accurately we need to know not only the haematocrit but also the shape of the ozygen-dissociation curve, the difference of arterial to venous Po2, and the pH changes involved. The clinical improvement reported by the Queen Square group may result, as they suggest, simply from increased blood-flow, but there are other possible explanations. A factor not so far considered is the improvement in temperature which accompanies increased flow. Perfusion in turn improves as vessels dilate and the blood viscosity falls (particularly at low shear-rates). The influence of temperature on blood viscosity has not been system-
atically investigated as a means of improving tissue perfusion and reducing venous thrombosis. There is good evidence that blood viscosity in the microcirculation is determined predominantly by the plasma viscosity and erythrocyte flexibility. In this respect the paper of BARNES and others8 suggesting an association of erythrocyte flexibility with the degree of diabetic microangiopathy is of particular interest. One difficulty in accepting these data is their earlier work9 with the same filtration technique, showing that erythrocyte flexibility does not contribute to whole-blood viscosity at any shear-rate. What factor, then, accounts for the variation of whole-blood viscosity observed in normal subjects and patients at high shear-rates? The other difficulty is that a similar range of variation of 3. Jay, A. W. L., Rowlands, S., Skibo, L. Can. J. Physiol. Pharmac. 1972, 50, 1007. 4. Gregersen, M. I., Bryant, C. A., Hammerle, W. E., Usami, S., Chien, S.
Science, 1967, 157, 825. 5. Dormandy, J. A. Br. med. J. 1971, iv, 716. 6. Prothero, J., Burton, A. C. Biophys. J. 1962, 2, 199. 7. Segal, N., Bishop, J. M. J. clin. Invest. 1966, 45, 1555. 8. Barnes, A. J., Locke, P., Scudder, P. R., Dormandy, T. L.,
Dormandy, J. A., Slack, J. Lancet, 1977, ii, 789. 9. Reid, H. L., Dormandy, J. A., Barnes, A. J., Lock, P. J., Dormandy, T. L. ibid. 1976, i, 666.
962
erythrocyte flexibility was observed in both patients and normal controls. If erythrocyte flexibility was the only factor involved, why were some of the consimilarly afflicted ? WARDLE, PIERCY, and earlier reported that plasmaabove 400 mg/dl in diabetic patients was fibrinogen associated with onset of major cardiovascular complications within three years. The involvement of fibrinogen again suggests that the blood viscosity trols
not
ANDERSONI0
in the microcirculation may contribute to the development of vascular lesions. A link between these observations has been reported by KNIGHT, RAMPLING, and SIRS,11 who found that the flexibility of erythrocytes in blood from diabetic patients was lower at high fibrinogen levels than in normal controls.12,13 Below a plasma-fibrinogen concentration of 350 mg/dl, however, the plasma viscosity and erythrocyte flexibility are normal. The red-cell flexibility of patients with diabetes is independent of fibrinogen levels. This implies that in patients with high fibrinogen levels, and corresponding increased plasma viscosity, the blood viscosity in the microcirculation is abnormally high. In whole blood, the viscosity in diabetics is also raised at low shear rates,probably owing to the increased level of plasma-globulins,14 and this would predispose to venous thrombosis. The change in erythrocyte behaviour in diabetes is unexplained; it does not seem to be influenced by treatment with drugs such as insulin. Perhaps manipulation of the viscosity by control of plasma-fibrinogen level will prove useful in diabetic complications. Attempts with clofibrate have not so far been successful, but clofibrate has other effects, and patients have not been selected for suitability by measurement of both fibrinogen and erythrocyte flexibility. No drug has been shown unequivocally to increase erythrocyte flexibility, and work along these lines may be fruitful.
Barefoot Yes, Hotfoot No ,
THE barefoot-doctor approach to medical care in the Third World has been widely accepted-in at least-by governments, health planand aid agencies. The notion is good: indeed, ners, how else could the rural poor get the benefits of preventive medicine and simple effective treatments ? Yet the eagerness with which such schemes are being promoted may actually put the barefootdoctor concept at risk. Nowhere is this better seen than in population control. After the failure of the crash programmes of the 1950s and 1960s, some agencies realised that the emerging primary-care schemes offered the best
principle
10. Wardle, E., Piercy, D., Anderson, J. Postgrad. med. J. 1973, 49, 1. 11. Knight, K., Rampling, M. W., Sirs, J. A. Biorheology (in the press). 12. Rampling, M. W., Sirs, J. A. J. Physiol. Lond. 1972, 223, 199. 13. Dupont, P., Sirs, J. A. Thrombosis and Hœmostasis (in the press). 14. McMillan, D. E. J. clin. Invest. 1974, 53, 1071.
hope of bridging the gap between those wanting to provide contraception and those who seemed awkward about receiving it. The International Planned Parenthood Federation and USAID vigorously supported programmes in which primary-health-care workers, distributed contraceptive pills, administered injectables, inserted intrauterine devices, and performed male and female sterilisations. Programmes which had formerly struggled along on meagre budgets found themselves plied with large funds to finance rapid expansion. But barefoot doctoring requires the acquisition of new attitudes as well as new skills, and this takes time. Some of those workers at the scene of the action are beginning to suspect that population statistics-numbers rather than human need--count too highly in the thinking of their sponsors. For example, how are they to react when they read (of a vasectomy camp in India): "Although gloves were used, they are not necessary. If instruments are carefully sterilised and a ’no touch’ technique is used, bacteria from the hands cannot enter a vasectomy incision"?1 This may well be true; but so long as, in the aid-giving country, every wound is stitched with gloved hands, such phrases will have a hollow ring. Some of those experts who press for an all-out drive to train primary health workers in sterilisation of women are similarly prepared to dispense with gloves; yet these paramedicals will be operating, not on well-npurished westerners, but on women
already
weakened
by childbirth,
tion, and anaemia; and such for
malnutri-
women are not
good
subjects experiment with gloveless surgerywhen the operator has only rudimentary especially knowledge of bacterial behaviour. We can safely assume that, for every woman whose wound becomes infected postoperatively, several potential clients will stay away. Primary health workers are an enormous asset to family-planning services, but they will remain so only if the organisers resist over-rapid expansion and reduction of standards. Population experts have latched on to just one aspect of the barefoot-doctor approach-that it can be established and expanded rapidly, owing to the short training of the workers. But much more important, in the long run, is that the service becomes a permanent part of the village sceneunlike the sporadic forays of city doctors which represent most of the villagers’ experience of modern medicine. Any type of contraception will be more acceptable when provided by health workers who are known in the village and who will be available for advice in case of side-effects or complications. The Lancet has already commented favourably on the project in Bangladesh known as Gonoshasthaya Kendra, where paramedicals pro1.
an
Population Reports, series D, no. 2. Department of Medical Affairs, George Washington University Medical Center, 1975.
and Public
THE.LANCET
Hyperviscosity
in Disease
Two years ago we reviewed the factors which influence blood viscosity in health and diseased Since then, several more reports have linked viscosity
changes to disease. But, intriguing as these are, we still await proof that changes in viscosity precede the clinical manifestations and are the prime factor. Usually a raised viscosity is secondary to some other lesion. In sickle-cell disease, for instance, though an increase of blood viscosity precedes peripheral occlusion,2the basis of the disease is an abnormal haemoglobin. Manipulation of blood viscosity, though it may relieve some of the complications, is unlikely to be a complete solution. Furthermore, we still know little about how to change blood viscosity for optimum blood-flow. Reducing the hxmatocrit or plasma-fibrinogen is not necessarily beneficial. This week a group from Queen Square present evidence that patients with a haematocrit over 46% tend to have a low cerebral blood-flow; and they attribute the depression of blood-flow to the increased peripheral resistance which results from raised viscosity. By reducing the haematocrit from 49.3% to 42-6% they raised the mean cerebral blood-flow by half-from 41.1 to 62.1ml/100 g/min. In the succeeding article, a St. Mary’s group likewise link hyperviscosity with haematocrit, in patients with angina. The relation between haematocrit and blood-flow is probably more complex than these workers suggest; viscosity may be only part of the mechanism. Cardiac output is determined approximately by arterial pressure divided by peripheral resistance. The main pressure drop occurs across the vessels of the microcirculation. Peripheral resistance is influenced by the geometry and size of these vessels and by blood viscosity. Both groups reporting this week assume that the change of peripheral blood-flow results directly from an effect of haematocrit on blood viscosity; but this may be wrong. Firstly, there is evidence that the haematocrit is lower in arterioles and capillaries than in large arteries and veins. There is no evidence that an increase in the arterial-venous haema1. Lancet, 1975, i, 113. Chien, S., Usami, S., Bertles, J. F. J. clin. Invest. 1970, 49, 623.
2
tocrit produces a corresponding increase in the haematocrit of the microcirculation. Even if it did, this would not greatly alter the blood viscosity in this region of the circulation. The influence of hxmatocrit on blood viscosity falls steadily with decreasing vessel bore,until in vessels of 4 to 15 (im bore it has no effect.3 Similarly the influence of shear-rate is unlikely to be important. Irrespective of conflicting views as to whether the shear-rate in capillaries is high or low, the effect of shear-rate decreases with hsematocrit, and hsematocrit, as we have said, is reduced. The physical nature of flow in small vessels is not the same as that in large arteries and veins. The relation of flow-rate to driving-pressure in small vessels is linear, irrespective of the haematocrit. 3,4 The suggestion that a reduction of haematocrit gives rise to an increase in cardiac output,S,7 owing to a reduction of blood viscosity, also requires more convincing evidence than we have yet seen. The increased blood-flow is more likely to be due to lowered oxygen transport, associated with the lowering of the haemoglobin content ; and to assess this accurately we need to know not only the haematocrit but also the shape of the ozygen-dissociation curve, the difference of arterial to venous Po2, and the pH changes involved. The clinical improvement reported by the Queen Square group may result, as they suggest, simply from increased blood-flow, but there are other possible explanations. A factor not so far considered is the improvement in temperature which accompanies increased flow. Perfusion in turn improves as vessels dilate and the blood viscosity falls (particularly at low shear-rates). The influence of temperature on blood viscosity has not been system-
atically investigated as a means of improving tissue perfusion and reducing venous thrombosis. There is good evidence that blood viscosity in the microcirculation is determined predominantly by the plasma viscosity and erythrocyte flexibility. In this respect the paper of BARNES and others8 suggesting an association of erythrocyte flexibility with the degree of diabetic microangiopathy is of particular interest. One difficulty in accepting these data is their earlier work9 with the same filtration technique, showing that erythrocyte flexibility does not contribute to whole-blood viscosity at any shear-rate. What factor, then, accounts for the variation of whole-blood viscosity observed in normal subjects and patients at high shear-rates? The other difficulty is that a similar range of variation of 3. Jay, A. W. L., Rowlands, S., Skibo, L. Can. J. Physiol. Pharmac. 1972, 50, 1007. 4. Gregersen, M. I., Bryant, C. A., Hammerle, W. E., Usami, S., Chien, S.
Science, 1967, 157, 825. 5. Dormandy, J. A. Br. med. J. 1971, iv, 716. 6. Prothero, J., Burton, A. C. Biophys. J. 1962, 2, 199. 7. Segal, N., Bishop, J. M. J. clin. Invest. 1966, 45, 1555. 8. Barnes, A. J., Locke, P., Scudder, P. R., Dormandy, T. L.,
Dormandy, J. A., Slack, J. Lancet, 1977, ii, 789. 9. Reid, H. L., Dormandy, J. A., Barnes, A. J., Lock, P. J., Dormandy, T. L. ibid. 1976, i, 666.
962
erythrocyte flexibility was observed in both patients and normal controls. If erythrocyte flexibility was the only factor involved, why were some of the consimilarly afflicted ? WARDLE, PIERCY, and earlier reported that plasmaabove 400 mg/dl in diabetic patients was fibrinogen associated with onset of major cardiovascular complications within three years. The involvement of fibrinogen again suggests that the blood viscosity trols
not
ANDERSONI0
in the microcirculation may contribute to the development of vascular lesions. A link between these observations has been reported by KNIGHT, RAMPLING, and SIRS,11 who found that the flexibility of erythrocytes in blood from diabetic patients was lower at high fibrinogen levels than in normal controls.12,13 Below a plasma-fibrinogen concentration of 350 mg/dl, however, the plasma viscosity and erythrocyte flexibility are normal. The red-cell flexibility of patients with diabetes is independent of fibrinogen levels. This implies that in patients with high fibrinogen levels, and corresponding increased plasma viscosity, the blood viscosity in the microcirculation is abnormally high. In whole blood, the viscosity in diabetics is also raised at low shear rates,probably owing to the increased level of plasma-globulins,14 and this would predispose to venous thrombosis. The change in erythrocyte behaviour in diabetes is unexplained; it does not seem to be influenced by treatment with drugs such as insulin. Perhaps manipulation of the viscosity by control of plasma-fibrinogen level will prove useful in diabetic complications. Attempts with clofibrate have not so far been successful, but clofibrate has other effects, and patients have not been selected for suitability by measurement of both fibrinogen and erythrocyte flexibility. No drug has been shown unequivocally to increase erythrocyte flexibility, and work along these lines may be fruitful.
Barefoot Yes, Hotfoot No ,
THE barefoot-doctor approach to medical care in the Third World has been widely accepted-in at least-by governments, health planand aid agencies. The notion is good: indeed, ners, how else could the rural poor get the benefits of preventive medicine and simple effective treatments ? Yet the eagerness with which such schemes are being promoted may actually put the barefootdoctor concept at risk. Nowhere is this better seen than in population control. After the failure of the crash programmes of the 1950s and 1960s, some agencies realised that the emerging primary-care schemes offered the best
principle
10. Wardle, E., Piercy, D., Anderson, J. Postgrad. med. J. 1973, 49, 1. 11. Knight, K., Rampling, M. W., Sirs, J. A. Biorheology (in the press). 12. Rampling, M. W., Sirs, J. A. J. Physiol. Lond. 1972, 223, 199. 13. Dupont, P., Sirs, J. A. Thrombosis and Hœmostasis (in the press). 14. McMillan, D. E. J. clin. Invest. 1974, 53, 1071.
hope of bridging the gap between those wanting to provide contraception and those who seemed awkward about receiving it. The International Planned Parenthood Federation and USAID vigorously supported programmes in which primary-health-care workers, distributed contraceptive pills, administered injectables, inserted intrauterine devices, and performed male and female sterilisations. Programmes which had formerly struggled along on meagre budgets found themselves plied with large funds to finance rapid expansion. But barefoot doctoring requires the acquisition of new attitudes as well as new skills, and this takes time. Some of those workers at the scene of the action are beginning to suspect that population statistics-numbers rather than human need--count too highly in the thinking of their sponsors. For example, how are they to react when they read (of a vasectomy camp in India): "Although gloves were used, they are not necessary. If instruments are carefully sterilised and a ’no touch’ technique is used, bacteria from the hands cannot enter a vasectomy incision"?1 This may well be true; but so long as, in the aid-giving country, every wound is stitched with gloved hands, such phrases will have a hollow ring. Some of those experts who press for an all-out drive to train primary health workers in sterilisation of women are similarly prepared to dispense with gloves; yet these paramedicals will be operating, not on well-npurished westerners, but on women
already
weakened
by childbirth,
tion, and anaemia; and such for
malnutri-
women are not
good
subjects experiment with gloveless surgerywhen the operator has only rudimentary especially knowledge of bacterial behaviour. We can safely assume that, for every woman whose wound becomes infected postoperatively, several potential clients will stay away. Primary health workers are an enormous asset to family-planning services, but they will remain so only if the organisers resist over-rapid expansion and reduction of standards. Population experts have latched on to just one aspect of the barefoot-doctor approach-that it can be established and expanded rapidly, owing to the short training of the workers. But much more important, in the long run, is that the service becomes a permanent part of the village sceneunlike the sporadic forays of city doctors which represent most of the villagers’ experience of modern medicine. Any type of contraception will be more acceptable when provided by health workers who are known in the village and who will be available for advice in case of side-effects or complications. The Lancet has already commented favourably on the project in Bangladesh known as Gonoshasthaya Kendra, where paramedicals pro1.
an
Population Reports, series D, no. 2. Department of Medical Affairs, George Washington University Medical Center, 1975.
and Public
