- Email: [email protected]
Frequency of protein Z deficiency in patients with ischaemic stroke
CORRESPONDENCE
sometimes disagree. Therefore, human judgment is indispensable. Documents such as Helsinki 2000 are invaluable in advancing general principles, but they err when they are so specific as paragraph 29. Of the three major ethical considerations—beneficence, equity, and autonomy—medical culture prizes beneficence above all. Extreme beneficence, however, can become paternalism. The 2000 Helsinki Declaration seems to suffer from this extremism. Also troubling, in light of the complexity of these issues is the World Medical Association’s hubris in asserting primacy through the Declaration, over all national ethics systems and scientific journals. Fortunately, the Association seems to recognise some of the weaknesses and will revisit them.2 James D Shelton Center for Population, Health and Nutrition, Agency for International Development Washington, DC 20523, USA (e-mail: [email protected]) 1
2
Taskin DP, Kanner R, Bailey S, et al. Smoking cessation in patients with chronic obstructive pulmonary disease: a doubleblind, placebo-controlled, randomised trial. Lancet 2001; 357: 1571–75. Anon. Thinking again. Science 2001; 292: 1037.
Frequency of protein Z deficiency in patients with ischaemic stroke Sir—We are perplexed by the results of Marc Vasse and colleagues (March 24, p 933)1 since they are in sharp contrast to our own findings in a similar study on 157 patients with a history of ischaemic stroke and 192 healthy controls.2 Vasse and co-workers noted decreased protein Z antigen concentrations in 33 of 169 patients who had had ischaemic cerebrovascular stroke at least 3 months before blood sampling. Mean protein Z concentration in the 169 stroke patients was lower, although whether this was significant is unclear, than the mean value in 88 healthy controls. Vasse and colleagues suggest that protein Z deficiency might constitute a risk factor for ischaemic cerebrovascular disease. We showed a median protein Z concentration of 99% of normal plasma (range 17–254) in 157 stroke patients compared with 94% (19–189) in healthy controls, which was a nonsignificant difference. Rising protein Z concentrations increased the risk for stroke. Logistic regression showed that the association of high protein Z
840
concentration with stroke remained significant (p=0·04) after adjustment for established risk factors such as age, sex, arterial hypertension, diabetes mellitus, smoking habit, body mass index, hyperlipidaemia, and fibrinogen. The freqency of protein Z deficiency was the same in the stroke and control groups. We have no explanation for the contrast between our and Vasse and colleagues’ findings. The patients in the two studies were young, stroke had occurred at least 2 months before analysis, and we used the same ELISA kits to measure protein Z antigen, although we slightly modified the assay for proper use.3 As opposed to Vasse and colleagues, we did not exclude patients with arterial hypertension, diabetes mellitus, and hyperlipidaemia. To resolve this issue, we propose an exchange of plasma samples to reassay protein Z plasma concentrations and other vitamin-Kdependent plasma proteins (which we measured to formally rule out vitaminK deficiency). *Walter A Wuillemin, Franziska Demarmels Biasiutti, Heinrich P Mattle, Bernhard Lämmle Central Haematology Laboratory and Department of Neurology, Inselspital, University of Bern, Switzerland; and *Division of Haematology, Department of Internal Medicine, Kantonsspital, 6000 Lucerne 16, Switzerland 1
2
3
Vasse M, Guegan-Massardier E, Borg J-Y, Woimant F, Soria C. Frequency of protein Z deficiency in patients with ischaemic stroke. Lancet 2001; 357: 933–34. Kobelt K, Demarmels Biasiutti F, Mattle HP, Lämmle B, Wuillemin WA. Protein Z in ischaemic stroke. Br J Haematol 2001; 119: 169–73. Ravi S, Mauron T, Lämmle B, Wuillemin WA. Protein Z in healthy human and in patients with a bleeding tendency. Br J Haematol 1998; 192: 1219–23.
Authors’ reply Sir—Wuillemin and colleagues note no higher prevalence of protein Z deficiency in 157 patients with a history of ischaemic stroke than in healthy controls. By contrast, they conclude that high protein Z concentration is a risk factor for stroke. They also suggest that the high frequency of protein Z deficiency that we report could be due to vitamin K deficiency. For all the patients included in our study, prothrombin time was within the normal range (>75%), and factor VII and protein C, two vitamin K sensitive factors, were measured, respectively, in the Paris and Rouen groups, and both were higher than 70%. Wuillemin and colleagues slightly
modified the commercial assay to obtain a better reproducibility. We followed the manufacturer’s instructions and used an automated workstation. Under these conditions, the between-assay coefficient was 8·2% and similar to that obtained by Wuillemin and colleagues using their modified method. Furthemore, a technical error in our study seems unlikely, since a low concentration of protein Z was systematically controlled on another ELISA plate. In addition, protein Z deficiency was confirmed on new samples in 12 patients. Wuillemin and colleagues also modified the assay calibration by use of an normal plasma pool, calibrated with purified protein Z. Since distribution of protein Z concentration in healthy controls is very wide, a high number of donors is required to obtain a representative pool. This necessity could perhaps explain why the mean plasma concentration of protein Z in their control group was 1·5 mg/L, which is much lower than that obtained by Miletich and Broze1 in plasma in edetic acid from 455 blood donors (2·9 mg/L) or that of our series in citrated plasma (2·29 mg/L). The low concentration of protein Z in the Swiss control group could also represent a genetic selection of the control population studied. Consequently, Wuillemin and colleagues’ normal range being substantially lower than normal could explain why they did not detect protein Z deficiency. The stroke patients analysed by Wuillemin and colleagues differ from ours. They did not exclude patients with hyperlipidaemia and diabetes, common causes of ischaemic stroke. However, vitamin K dependent proteins are strikingly increased in patients with hyperor hypertriglycholesterolaemia2 ceridaemia.3 This effect could, perhaps, explain why Wuillemin and colleagues saw raised protein Z concentrations in patients. We suggest that analysis of the effects of hypertriglyceridaemia and hypercholesterolaemia on plasma protein Z concentration could be useful. *Marc Vasse, Evelyne Guegain Massardier, Jeanne-Yvonne Borg, France Woimont, Claudine Soria *Laboratoire d’Hématologie and Département de Neurologie, Hôpital Charles Nicolle, 76031 Rouen, France; and Service de Neurologie and Laboratoire d’Hématologie, Hôpital Lariboisière, Paris, France 1
Miletich JP, Broze GJ Jr. Human plasma protein Z antigen: range in normal subjects and effects of warfarin therapy. Blood 1987; 69: 1580–86.
THE LANCET • Vol 358 • September 8, 2001
For personal use. Only reproduce with permission from The Lancet Publishing Group.
sometimes disagree. Therefore, human judgment is indispensable. Documents such as Helsinki 2000 are invaluable in advancing general principles, but they err when they are so specific as paragraph 29. Of the three major ethical considerations—beneficence, equity, and autonomy—medical culture prizes beneficence above all. Extreme beneficence, however, can become paternalism. The 2000 Helsinki Declaration seems to suffer from this extremism. Also troubling, in light of the complexity of these issues is the World Medical Association’s hubris in asserting primacy through the Declaration, over all national ethics systems and scientific journals. Fortunately, the Association seems to recognise some of the weaknesses and will revisit them.2 James D Shelton Center for Population, Health and Nutrition, Agency for International Development Washington, DC 20523, USA (e-mail: [email protected]) 1
2
Taskin DP, Kanner R, Bailey S, et al. Smoking cessation in patients with chronic obstructive pulmonary disease: a doubleblind, placebo-controlled, randomised trial. Lancet 2001; 357: 1571–75. Anon. Thinking again. Science 2001; 292: 1037.
Frequency of protein Z deficiency in patients with ischaemic stroke Sir—We are perplexed by the results of Marc Vasse and colleagues (March 24, p 933)1 since they are in sharp contrast to our own findings in a similar study on 157 patients with a history of ischaemic stroke and 192 healthy controls.2 Vasse and co-workers noted decreased protein Z antigen concentrations in 33 of 169 patients who had had ischaemic cerebrovascular stroke at least 3 months before blood sampling. Mean protein Z concentration in the 169 stroke patients was lower, although whether this was significant is unclear, than the mean value in 88 healthy controls. Vasse and colleagues suggest that protein Z deficiency might constitute a risk factor for ischaemic cerebrovascular disease. We showed a median protein Z concentration of 99% of normal plasma (range 17–254) in 157 stroke patients compared with 94% (19–189) in healthy controls, which was a nonsignificant difference. Rising protein Z concentrations increased the risk for stroke. Logistic regression showed that the association of high protein Z
840
concentration with stroke remained significant (p=0·04) after adjustment for established risk factors such as age, sex, arterial hypertension, diabetes mellitus, smoking habit, body mass index, hyperlipidaemia, and fibrinogen. The freqency of protein Z deficiency was the same in the stroke and control groups. We have no explanation for the contrast between our and Vasse and colleagues’ findings. The patients in the two studies were young, stroke had occurred at least 2 months before analysis, and we used the same ELISA kits to measure protein Z antigen, although we slightly modified the assay for proper use.3 As opposed to Vasse and colleagues, we did not exclude patients with arterial hypertension, diabetes mellitus, and hyperlipidaemia. To resolve this issue, we propose an exchange of plasma samples to reassay protein Z plasma concentrations and other vitamin-Kdependent plasma proteins (which we measured to formally rule out vitaminK deficiency). *Walter A Wuillemin, Franziska Demarmels Biasiutti, Heinrich P Mattle, Bernhard Lämmle Central Haematology Laboratory and Department of Neurology, Inselspital, University of Bern, Switzerland; and *Division of Haematology, Department of Internal Medicine, Kantonsspital, 6000 Lucerne 16, Switzerland 1
2
3
Vasse M, Guegan-Massardier E, Borg J-Y, Woimant F, Soria C. Frequency of protein Z deficiency in patients with ischaemic stroke. Lancet 2001; 357: 933–34. Kobelt K, Demarmels Biasiutti F, Mattle HP, Lämmle B, Wuillemin WA. Protein Z in ischaemic stroke. Br J Haematol 2001; 119: 169–73. Ravi S, Mauron T, Lämmle B, Wuillemin WA. Protein Z in healthy human and in patients with a bleeding tendency. Br J Haematol 1998; 192: 1219–23.
Authors’ reply Sir—Wuillemin and colleagues note no higher prevalence of protein Z deficiency in 157 patients with a history of ischaemic stroke than in healthy controls. By contrast, they conclude that high protein Z concentration is a risk factor for stroke. They also suggest that the high frequency of protein Z deficiency that we report could be due to vitamin K deficiency. For all the patients included in our study, prothrombin time was within the normal range (>75%), and factor VII and protein C, two vitamin K sensitive factors, were measured, respectively, in the Paris and Rouen groups, and both were higher than 70%. Wuillemin and colleagues slightly
modified the commercial assay to obtain a better reproducibility. We followed the manufacturer’s instructions and used an automated workstation. Under these conditions, the between-assay coefficient was 8·2% and similar to that obtained by Wuillemin and colleagues using their modified method. Furthemore, a technical error in our study seems unlikely, since a low concentration of protein Z was systematically controlled on another ELISA plate. In addition, protein Z deficiency was confirmed on new samples in 12 patients. Wuillemin and colleagues also modified the assay calibration by use of an normal plasma pool, calibrated with purified protein Z. Since distribution of protein Z concentration in healthy controls is very wide, a high number of donors is required to obtain a representative pool. This necessity could perhaps explain why the mean plasma concentration of protein Z in their control group was 1·5 mg/L, which is much lower than that obtained by Miletich and Broze1 in plasma in edetic acid from 455 blood donors (2·9 mg/L) or that of our series in citrated plasma (2·29 mg/L). The low concentration of protein Z in the Swiss control group could also represent a genetic selection of the control population studied. Consequently, Wuillemin and colleagues’ normal range being substantially lower than normal could explain why they did not detect protein Z deficiency. The stroke patients analysed by Wuillemin and colleagues differ from ours. They did not exclude patients with hyperlipidaemia and diabetes, common causes of ischaemic stroke. However, vitamin K dependent proteins are strikingly increased in patients with hyperor hypertriglycholesterolaemia2 ceridaemia.3 This effect could, perhaps, explain why Wuillemin and colleagues saw raised protein Z concentrations in patients. We suggest that analysis of the effects of hypertriglyceridaemia and hypercholesterolaemia on plasma protein Z concentration could be useful. *Marc Vasse, Evelyne Guegain Massardier, Jeanne-Yvonne Borg, France Woimont, Claudine Soria *Laboratoire d’Hématologie and Département de Neurologie, Hôpital Charles Nicolle, 76031 Rouen, France; and Service de Neurologie and Laboratoire d’Hématologie, Hôpital Lariboisière, Paris, France 1
Miletich JP, Broze GJ Jr. Human plasma protein Z antigen: range in normal subjects and effects of warfarin therapy. Blood 1987; 69: 1580–86.
THE LANCET • Vol 358 • September 8, 2001
For personal use. Only reproduce with permission from The Lancet Publishing Group.