Rapid Responses to:
|
|
Rapid Responses: Rapid Responses published:
-
![[Read Rapid Response]](/icons/misc/sectionDOWN.gif)
Primary end point undefined and confounders uncontrolled for
- Luca Puccetti (18 January 2008)
-
Adjusting for current smoking status
- Doina Kulick (19 January 2008)
-
Oxidative stress, vascular calcification and coronary events in advanced age
- Katrin Aasvee (20 January 2008)
-
Misrepresentation of study data
- Wendy S Biggs (23 January 2008)
-
The calcium hypothesis, once again
- José Pedro L. Nunes (25 January 2008)
-
Invalid Result
- Yasser K. Alotaibi (30 January 2008)
-
Some limitations in Bolland's study
- Jian-Lie Zhou (31 January 2008)
-
Role of Magnesium in Calcium Metabolism
- David P. Robinson (1 February 2008)
-
The need for vitamin K in calcium supplements
- Leon J Schurgers, Cees Vermeer (1 February 2008)
-
Vascular events and calcium supplementation
- Joan M. Lappe, Robert P. Heaney (4 February 2008)
-
CONSORT statement forgotten by BMJ
- Klaus H Witt, Marjukka Mäkelä (5 February 2008)
-
calcium is toxic to vessels
- Jayendra Gohil (5 February 2008)
-
Omission of important vascular events
- Phillip Moradi (6 February 2008)
-
Vitamin D
- MJ Wilkinson (6 February 2008)
-
Calcium supplementaion and vascular events
- Jinelle N. Ramlackhansingh (6 February 2008)
-
RE: Forgotten CONSORT
- Amy L Davis, John Fletcher, Clinical Epidemiologist BMJ (7 February 2008)
-
Questionable physical chemistry and statistics
- Christopher Nordin, Andrew Metcalfe, Department of Mathematics, University of Adelaide (8 February 2008)
-
Bad science
- Ivor Cammack (11 February 2008)
-
Response to Letters to BMJ re Bolland et al
- Ian R Reid, Mark J Bolland, Gregory D Gamble, Andrew Grey (19 February 2008)
-
Possible side effect of Calcium and/or Vitamin D supplements
- Audrey Lorraine Aspeling (4 May 2008)
|
|
|||
|
Luca Puccetti, MD, President of Promed Galileo Medical Society Pisa, 56011
Send response to journal:
|
Bolland et. al did not clearly state which is the primary end point of the study. Is the primary end point the composite index (MI, CV sudden death?) or any other single event? This is a crucial matter since the primary end point guides the assumptions about the sample size and the statistical power of study. Moreover the study did not control for the vitamin D levels during the study, but it was only stated that the women with serum 25- hydroxyvitamin D levels less than 25 nmol/l were exluded from the study. This implies that serum levels of serum vitamin D were not been monitored during the study. This is an important issue since, according to the results of the Framingham Offspring Study, low serum levels of vitamin D are strictly associated with an increase in the vascular risk (1)and may impair the calcium absorption itself. The baseline descriptive and biochemical characteristics between the calcium and placebo groups appear to be umbalanced from the point of view of the global cardiovascular risk. The calcium group has a little bit more smokers, total and LDL cholesterol levels are a little bit higher, as well as it includes some more patients with previous stroke, transient ischaemic attack, hyperthension, dyslipidaemia and previous ischaemic heart disease. Albeit Bolland and coll. controlled for each single confounder I wonder if they controlled for the "summa" of these differences in the baseline characteristics by referring, for instance, to one of the various cardiovascular risk scores. Is is quite surprising that all, but few, baseline differences are unidirectionally directed, generally favouring the placebo group and it is also surprising that this umbalance does not affect the global cardiovascular risk. The point is that if one seeks the effect of calcium supplementation on cardiovascular events it is important to rule-out that the baseline cardiovascular global risk is different between the groups. Mostly important the control for two important confounding factors is lacking: NSAIDs/analgesics consumption and previous hormonal replacement therapy. These two potent confounders may completely reverse the results and they are also interacting each other in regard with the effect on cardiovascular risk (2,3,4). Finally it is quite surprising to consider separately the self reported events from the ones unreported but added from the national database of hospital admissions. By referring to the sum of the self- reported and unreported events the differences between calcium and placebo group are no more statistically significant. This is the result of the study, the analysis of the self reported events alone looks like just window dressing. best regards Luca Puccetti References 1) Wang TJ, Pencina MJ, Booth SL et al. Vitamin D Deficiency and Risk of Cardiovascular Disease. Circulation Jan 2008 DOI 10.1161/CIRCULATIONAHA.107.706127 2) PLoS Med 4(5): e157 10.1371/journal.pmed.0040157 3) The Lancet - Vol. 368, Issue 9549, 18 November 2006, Pages 1771-1781 4) N Engl J Med. 2005;352:1092-102 Competing interests: None declared |
|||
|
|
|||
|
Doina Kulick, Assistant Professor University of Nevada School of Medicine, Reno, 89502Ths s
Send response to journal:
|
With its findings of increased cardiovascular morbidity and mortality this study continues the controversy about calcium intake and health in postmenopausal women. The authors reported no changes in results when adjusting for age and glomerular filtration rate (above or below median for each); history of ischaemic heart disease, stroke, hypertension, dyslipidaemia, diabetes;and compliance with the study drug. Nonetheless no adjustment was made in regard to current smoking status. In the Table 1, one can see that the calcium group has more smokers than the placebo group, and the difference borders statistical significance. Thus a baseline variable that has pathophysiological relation to the measured outcome, and which seems to be imbalanced between the two groups, was not taken into consideration by the authors into their Poisson regressing analysis. Competing interests: None declared |
|||
|
|
|||
|
Katrin Aasvee, Senior Researcher, MD, PhD National Institute for Health Development, 42 Hiiu Str, 11619, Tallinn, Estonia
Send response to journal:
|
The result shown by Bolland et al. seems not surprising, if to consider the possible process mentioned by the way also by authors “calcium supplements can accelerate vascular calcification”. It is found that calcification of coronary arteries and simultaneous loss of minerals from bones occurs in elderly, the process named paradox of arterial calcification in osteoporotic patients (1). Different mechanisms are considered to be associated with this paradox, the most relevant of these is basing on reciprocal effects of specific oxidized lipids on vascular calcification and on bone mineral formation (2, 3). Oxidative stress is considered to be a “biological marker of ageing” (4). In elderly, when the balance between oxidized compounds and antioxidant parameters is changed (5), the calcium supplementation can favour calcification of coronary arteries and also to increase the risk of plaque rupture (6). Thus, the level of traditional risk factors of atherosclerosis and also the level of oxidative stress can be quite similar in postmenopausal women taking calcium or placebo, but vascular calcification can be considerably accelerated in the group with calcium supplementation. The results by Bolland et al. can say to clinicians that calcium supplementation in advanced age must be taken with certain caution. 1. Boukhris R, Becker KL. JAMA. 1972;219:1307-1311. 2. Parhami F, Morrow AD, Balucan J et al. Arterioscler Thromb Vasc Biol. 1997;17:680- 687. 3. Parhami F, Basseri JH, Hwang J et al. Circ Res. 2002;91:570-576. 4. Nakamura YK, Omaye ST. Life Sci. 2004;74: 1265-1275. 5. Andriollo- Sanchez M, Hininger-Favier I, Meunier N et al. Eur J Clin Nutr. 2005;59:558-562. 6. Li ZY, Howarth S, Tang T et al. Cerebrovasc Dis. 2007;24:452-459. Competing interests: None declared |
|||
|
|
|||
|
Wendy S Biggs, Associate Editor JWWH Midland Family Medicine Residency, 4005 Orchard Dr, Midland MI 48670
Send response to journal:
|
To the Editor of BMJ, Your boxed “What This Study Adds” accompanying the article by Bolland et al1 misrepresents the study data. The first bullet states that “Healthy older women randomised to calcium supplementation showed increased rates of myocardial infarction.” Although the data show that women reported more events, analysis of New Zealand hospital records erased any statistically significant difference in myocardial infarction rates between the calcium and placebo groups. Your bulleted statement is not evidence-based: the adjudicated results showed no statistically significant increase in cardiovascular end-points among women taking calcium. The authors characterized the between-group differences in event rates for myocardial infarction and the composite endpoint as being of “borderline significance”. It cannot be concluded that the women showed “increased rates of myocardial infarction”, because a “borderline” P value of 0.058 still means the events could occur by chance. In addition, the authors’ statements in the body of the article do not fully support the conclusion they present in the abstract. In their discussion, the authors state that the study “does not unequivocally show an adverse cardiovascular effect of calcium...” The double negative of “not unequivocally” should simply say “not”: it is a leap to suggest that an adverse effect may occur when the data show no statistical difference between groups. Although the authors go on to note that “The present data do not permit definitive conclusions to be reached [regarding calcium and cardiovascular events]”, they highlight “upward trends in cardiovascular event rates” and a “potentially detrimental effect” of calcium supplementation in the abstract’s conclusion. Many people, both clinicians and laypersons, might only read the abstract of this paper; thus, it is unfortunate that your peer review process allowed a stronger statement than the data support. An “upward trend” does not connote statistical significance, so no comment can be made regarding a detrimental effect of calcium. If we are to practice evidence-based medicine, than we need to remain true to the evidence. Wendy S. Biggs, MD
1. Bolland et al. Vascular events in healthy older women receiving calcium supplementation: randomized controlled trial. BMJ Online First 2008 Jan 15 (doi:10.1136/bmj.39440.525752.BE). Competing interests: None declared |
|||
|
|
|||
|
José Pedro L. Nunes, Associate professor Faculdade de Medicina da Universidade do Porto, 4200-319 Porto, Portugal
Send response to journal:
|
In 1960, Briggs et al. published an autopsy study in patients with peptic ulcer treated with high-milk diets 1. An increased incidence of myocardial infarction was noted, using control groups of peptic ulcer disease patients not under such treatment and patients without peptic ulcer 1. Briggs et al. speculated that “something in the diet associated with medication increases the incidence of infarcts” 1. The data from peptic ulcer patients were later interpreted by Seely as providing a basis for a possible connection between milk and coronary heart disease – “the calcium hypothesis” 2. Further important data came from studies on renal failure patients – patients with increased incidence of cardiovascular disease. The importance of calcium and phosphorus has been well documented in such patients, in which increased vascular calcification may be associated to the use of calcium-containing oral phosphate binders 3. In what concerns atherosclerosis, the situation is less clear in non- renal failure patients, in which conflicting data exists concerning calcium/phosphorus metabolism and cardiovascular disease 4-6. A case could be made that “patients with significant atherosclerotic disease should be informed that the ingestion of milk and calcium supplementation have neither been conclusively proven to be safe, nor the opposite” 4. It is against this background that the most interesting paper by Bolland et al. sheds some light. After reviewing previous evidence pointing in a different direction, the authors show that calcium supplementation in healthy postmenopausal women was found to be associated to upward trends in cardiovascular event rates 6. If confirmed by further studies, we may be in the presence of a landmark paper, which may help in the elucidation of which groups may benefit from calcium supplementation and which may possibly be harmed. References 1. Briggs RD, Rubenberg ML, O´Neal RM, Thomas WA, Hartroft WS. Myocardial infarction in patients treated with Sippy and other high-milk diets: an autopsy study of fifteen hospitals in the U.S.A. and Great Britain. Circulation 1960; 21: 538-542. 2. Seely S. Possible connection between milk and coronary heart disease: The calcium hypothesis. Med Hypotheses 2000; 54: 701-703. 3. Goldsmith D, Ritz E, Covic A. Vascular calcification: a stiff challenge for the nephrologist. Kidney Int 66: 1315-1333, 2004. 4. Nunes JPL. The case for dietary calcium restriction in patients with atherosclerosis. Med Hypotheses 2005; 65: 521-524. 5. Nunes JPL, Faria MS, Garcia JMM, Gonçalves FR. Glomerular filtration rate and coronary artery disease burden in patients with acute coronary syndrome. Clin Cardiol 2007; 30: 464-468. 6. Bolland MJ, Vascular events in healthy older women receiving calcium supplementation: randomised controlled trial. BMJ, doi:10.1136/bmj.39440.525752.BE (published 15 January 2008). José Pedro L. Nunes Associate professor at Faculdade de Medicina da Universidade do Porto, 4200-319 Porto, Portugal. jplnunes@med.up.pt Competing interests: None declared |
|||
|
|
|||
|
Yasser K. Alotaibi, Family physician Riyadh Military Hospital, Kingdom of Saudi Arabia PC:11159
Send response to journal:
|
The results of this study are very questionable as it was plagued by many CRITICAL weaknesses.
A) Confounders (non-adjusted):
B) Drop-outs:
C) Intention to treat analysis:
Competing interests: None declared |
|||
|
|
|||
|
Jian-Lie Zhou, Medical Advisor Rm.2603-2609, CITIC Square, 1168 Nanjing Road(W), Shanghai 200041, P.R. China
Send response to journal:
|
To the Editor of BMJ, I would like to point out some limitations not mentioned by the authors. 1.With comparison of the baseline (mean; standard deviation) in the two groups, the number in calcium supplement group was a few more than the number in placebo group, e.g., number with previous hypertension (220; 30.1 vs. 207; 28), number of current smokers (25; 3.4 vs.19; 2.6); number of former smokers (295; 40.3 vs. 275; 37.2), number with previous ischaemic heart disease (59; 8.1 vs. 54; 7.3), number with previous dyslipidemia (67; 9.2 vs. 56; 7.6), previous stroke or transient attack (12; 1.6 vs. 7; 1.0). The few difference is very important that the number of cardiovascular events were a few difference. 2 Vitamin D supplementation is critical to calcium absorption from the gastrointestinal track as well as for the uptake and maintenance of calcium mineralization in the bone. Although the study excluded women with serum 25OHD less than 25 nmol/l, 75 nmol/l are essential to maintain optimal calcium absorption and bone mineralization. Inadequate vitamin D supplementation in elderly women may have an acceleration of the vascular calcification. 3 Total calcium intake in calcium supplement group was 861¡À390 mg/d + 1000 mg/d. It may be too high to effect its results of cardiovascular events which are not suitable to Chinese healthy older women. Chinese AI of calcium is 1000 mg/d and UL is 2000 mg/d for healthy older women. However, dietary calcium (853; 381 mg/d) in placebo group was suitable to Chinese healthy older women. Best regards, Jian-Lie Zhou, MD, PhD Competing interests: None declared |
|||
|
|
|||
|
David P. Robinson, Retired Department of Biomedical Sciences, Curtin University of Technology, Perth 6102, Western Australia
Send response to journal:
|
The following is the slightly amended text of letter sent to Professor Reid. Dear Professor Reid, Reading your paper on the above subject reminded me of an incidental finding in 1980 of calcium deposits on the pericardium of Balb C mice. The mice were germfree and being used for the principal study of the ultrastructural effects of Parainfluenza virus on the mouse respiratory system. Because the mice were germfree they had to have a special diet made to allow for dietary deficiencies due to a lack of bacterial synthesis of nutrients, also it had to withstand the sterilising effects of gamma irradiation. Their water supply was distilled. The image shown is a x3 macro photo showing the ventral surface of the worst affected mouse of a batch of about 12 in all. Superficially it resembled fat but gave a characteristic rasping sound when scraped with the edge of a scalpel blade. Two day old mice also had visible white deposits on their hearts. Had the mice been encouraged to exercise on a mouse wheel they probably wouldn’t have lasted more than a few rotations before death intervened! On consultation with our nutritional biochemist I took blood samples and the dietary cube samples and had them analysed for Ca++, P+ and Mg++. The Mg++ levels in both types of samples were considerably reduced. The assay of the diet for Magnesium was found to be 1.16 gm/kg and that of Calcium, 6.04 gm/kg. A Ratio 1:5 (Mg:Ca) Subsequent histochemical and qualitative x-ray (autoprobe) showed that the heart deposits were calcium. No lipids were evident. Histological and ultrastructural sections showed the calcium to be penetrating into the sub pericardial muscle fibres. When the diet suppliers were advised of the problem the magnesium deficiency was corrected and no further mice were affected. Remaining mice that had been affected were put on the new diet and no further calcium was found on their hearts. The problem of calcium deposits had been previously reported on several other strains of mice only in the kidneys and liver, but not on Balb C types. The Balb C mice I used had “clean” livers, spleens and kidneys. The Balb C mice may well be an exquisitely sensitive model for Ca++ P+ Mg++ ratios. I observe that calcium supplements are now recommended to be taken in a 2:1 ratio with magnesium. On researching the literature I found very many Public Health references to water supplies in many countries of the world in which a lack of, or low levels of magnesium was implicated in high levels of population cardiac disease. At one time a general practitioner in Wollongong complained about the lack of magnesium in their water supply being responsible for the high incidence of heart disease in the local BHP foundry workers. However causality has to be proved – as per tobacco and respiratory disease? A paper on this subject was presented by me at the Perth Calcified Tissue Group Symposium in the early 80s. The Symposium was initiated by Richard Prince, Department of Medicine, Sir Charles Gardiner Hospital. I was a member of the group for many years whilst I was a lecturer in Histopathology, Department of Biomedical Sciences, Curtin University of Technology, Perth, Western Australia. I am now retired. Kind Regards,
Competing interests: None declared |
|||
|
|
|||
|
Leon J Schurgers, Senior Scientist University Maastricht, VitaK BV universiteitssingel 50, 6200 MD Maastricht, the Netherlands, Cees Vermeer
Send response to journal:
|
Calcium supplements and myocardial infarction risk Bolland and colleages reported that intake of supplemental calcium resulted in improved bone strength but that the increased calcium consumption was paralleled by an over 2-fold increase of adverse cardiovascular events (1). Here we wish to put forward that during recent years it has become clear that two vitamins are involved in the regulation of calcium metabolism: whereas vitamin D stimulates intestinal uptake of calcium, vitamin K is required for the synthesis of two regulatory proteins: osteocalcin in bone and matrix Gla-protein (MGP) in the vasculature. Vitamin K is required for the biological activity of both proteins (2). It has been demonstrated that the majority of young healthy adults not receiving vitamin K supplements are vitamin K-insufficient, resulting in a substantial fraction (about 30%) of osteocalcin and MGP being synthesized as inactive proteins. This fraction increases with age (3). Various research groups have reported an association between poor vitamin K status and the rate of bone loss and cardiovascular calcification. MGP is an inhibitor of calcium salt precipitation and it is the strongest inhibition of soft tissue calcification presently known. Moreover, it is the only known defence system of the vasculature against mineralization. This defence system is compromised in non-supplemented subject since the typical “western” diet contains insufficient vitamin K to activate all MGP. Under such conditions it is only to be expected that if the calcium load is increased by administering a calcium supplement, the risk of vascular calcification will increase. We have recently shown that aortic calcification, myocardial infarction risk and cardiovascular mortality are inversely associated with the intake of vitamin K, notably K2 (4). Moreover, in a 3-year intervention study age-related vascular stiffening was completely blocked by high vitamin K intake (5). It is to be expected, therefore, that the increased infarction risk associated with calcium supplements is counteracted by the concomitant intake of vitamin K2. We therefore strongly recommend that calcium supplements (which in most cases also contain vitamin D) will be enriched with a low dose of vitamin K2. Preliminary data from our lab suggest that a dose between 20 and 200 micrograms per day may be effective in this respect. 1. Bolland MJ, Barber, PA, Doughty RN, Mason B, Horne A, Ames R, Gamble GD, Grey A, Reid IR. Vascular events in healthy older women receiving calcium supplementation: randomised controlled trial. Brit Med J published online 15 Jan 2008. 2. Berkner KL, Runge KW. The physiology of vitamin K nutrititure and vitamin K-dependent protein function in atherosclerosis. J Thromb Haemostas 2004; 2: 2118-2132. 3. Schurgers LJ, Teunissen KJF, Hamulyák K, Knapen MHJ, Vik H, Vermeer C. Vitamin K-containing dietary supplements: comparison of synthetic vitamin K1 and natto-derived menaquinone-7. Blood 2007; 109: 3279-3283. 4. Geleijnse JM, Vermeer C, Grobbee DE, Pols HAP, Schurgers LJ, Knapen MHJ, van der Meer IM, Hofman A, Witteman JCM. Dietary intake of vitamin K-2 reduces the risk of cardiac events and aortic atherosclerosis: The Rotterdam Study. J. Nutr. 2004; 134: 3100-3105. 5. Braam LAJLM, Hoeks APG, Brouns F, Hamulyák K, Gerichhausen MJW, Vermeer C. Beneficial effects of vitamin K on the elastic properties of the vessel wall in postmenopausal women: a follow-up study. Thromb Haemostas 2004; 91: 373-380. Competing interests: None declared |
|||
|
|
|||
|
Joan M. Lappe, Professor of Nursing and Medicine Creighton University, Omaha, NE 68178, Robert P. Heaney
Send response to journal:
|
To the Editor: Recently Bolland et al. (1) published data showing an increased risk of vascular events, particularly myocardial infarction, in healthy postmenopausal women receiving calcium supplements. As the trial had been carefully performed and the adverse events validated, this report elicited considerable interest, and to some extent, concern, with respect to the safety of calcium supplementation. The authors noted that several other studies had not shown comparably significant increases in vascular event risk and that, specifically, the much larger Women’s Health Initiative (2) found no increase in myocardial infarction in the calcium-supplemented group. Because we had recently completed a study very similar to the one described by Bolland et al., also with a primary skeletal endpoint (3,4), we thought it would be useful to query our database with respect to the principal adverse events reported by Bolland et al. (myocardial infarction/coronary disease; stroke; and sudden death). In our study there were 1,179 women followed for four years, very similar to the 1,471 women in the Bolland study, followed for five years. The calcium supplement dose in our study was 1400–1500 mg/d, nearly 50% greater than in Bolland et al., and hence more likely to produce adverse consequences if there were an effect of calcium on risk of vascular events; moreover, the principal source of that calcium was the same product (Citracal; Mission Pharmacal, San Antonio, TX) used by Bolland et al. Adverse events were recorded at 6- monthly intervals, verified with the participant’s personal physician, and then recorded in the database using the corresponding ICD-9 code. Unlike Bolland et al., our study had three treatment arms, one consisting of a double placebo, one of calcium plus vitamin D, and one of calcium and a vitamin D placebo. The total number of individuals treated with calcium (with or without additional vitamin D) was 892, and the person years of exposure to calcium was 3568, virtually identical to the figure in Bolland (3660). The vascular event rate across all groups was 5.3/1000 persons/yr, less than half the event rate calculated from the data of Bolland et al. (i.e., 12.2/1000) This difference may be due to the fact that the participants in Bolland et al. were ~7 years older, on average. More to the point, there was no excess occurrence in our study of myocardial infarction nor of either or the other vascular events in either calcium treatment group, relative to placebo, despite the higher calcium supplement dose. If the two calcium treatment groups in our study are combined, they exhibit a vascular event rate of 4.76 per 1000 persons per year while the placebo group exhibited a rate of 6.94 events per 1000 persons per year. This difference, while not statistically significant, is in a direction opposite to that found by Bolland et al. Thus we conclude that the excess of events in the Bolland study, largely comprised of myocardial infarction, was either a chance occurrence or at least not generalizable to other populations. In the boxed section of the Bolland article “What this study adds”, the authors note that the increase in myocardial infarction risk could “outweigh any benefits on bone from the calcium supplements”. That might be correct if the increased risk were real. The failure to find such a risk in other studies, such as WHI, and now in our own, very similar study, suggests that there is no reason at the moment to back away from current calcium intake recommendations. Joan M. Lappe, Ph.D.
References 1. Bolland MJ, Barber PA, Doughty RN, Mason B, Horne A, Ames R, Gamble GD, Grey A, Reid IR. Vascular events in healthy older women receiving calcium supplementation: randomized controlled trial. BMJ (online) 2008. 2. Jackson RD, LaCroix AZ, Gass M, Wallace RB, et al. Calcium plus vitamin D supplementation and risk of fractures. N Engl J Med 354:669-683, 2006. 3. Lappe JM, Davies KM, Travers-Gustafson D, Heaney RP. Vitamin D status in a rural postmenopausal female population. J Am Coll Nutr 25(5):395-402, 2006. 4. Lappe JM, Travers-Gustafson D, Davies KM, Recker RR, Heaney RP. Vitamin D and calcium supplementation reduces cancer risk. Am J Clin Nutr 85:1586-1591, 2007. Competing interests: None declared |
|||
|
|
|||
|
Klaus H Witt, Associate professor Department of General Practice, Copenhagen University, DK1014 Copenhagen K, Denmark, Marjukka Mäkelä
Send response to journal:
|
In a critical appraisal course for PhD students at Copenhagen University, we used the article by Bolland et al. (1). It provided several major discussion points for teaching, including the question whether the BMJ requires authors to adhere to the CONSORT statement (2). Going through a checklist of study quality, our students noted that a number of items in the CONSORT checklist were not covered. Most problematic was the lack of specification of the main objectives and outcome measures: we could choose from seven individual and two composite outcomes, often measured in two different ways. As our students pointed out, this amounts to fishing for results. Four of the 22 CONSORT items were not covered at all. Answers to five others could be found in a reference of this paper (3) or a more remote publication from the project (4). The original study question was whether one gram of elemental calcium per day (as compared to placebo) would decrease the rate of fractures in elderly, postmenopausal women (3). The study was powered (80%, with α = 0.05) to detect a 40% decrease in fracture rate over five years. A reduction in bone loss and turnover was observed, but the effect on fractures remained uncertain. The report presents a flow chart of the participants and lists important adverse events, both missing in the BMJ paper. Entirely missing items include enrolment practices and the dates defining the recruitment period. Details of randomisation were only available in the secondary reference (4). In the BMJ article, the authors present a pre-planned secondary analysis of data collected for another purpose, so the results are valid as hypotheses only. The material has also been used to report on the effects of calcium supplementation on body weight and blood pressure (4), serum lipids (5), and CRP levels (6). The CONSORT statement is specifically mentioning the need to address multiplicity by reporting any other analyses performed. Some of the confusion might have been avoided if the BMJ would have adhered to the CONSORT recommendations. As others have already pointed out, the messages in the abstract are overstatements. 1. Bolland MJ, Barber PA, Doughty RN, et al. Vascular events in healthy older women receiving calcium supplementation: randomized controlled trial. BMJ Online First 2008 Jan 15 (doi:10.1136/bmj.39440.525752.BE). 2. Moher D, Schultz KF, Altman DG for the CONSORT Group: The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomised trials. Lancet 2001;357:1191-4. 3. Reid IR, Mason B, Horne A, et al. RCT of Calcium in healthy older women. Am J Med 2006;119:777-785 (ref 15 of 1) 4. Reid IR, Horne A, Mason B, et al. Effects of calcium supplementation on body weight and blood pressure in normal older women: A randomised controlled trial. J Clin Endocr Metab 2005; 90:3824-9. 5. Reid IR, Mason B, Horne A, et al. Effects of calcium supplementation on serum lipid concentrations in normal older women: a randomized controlled trial. Am J Med 2002;112:343-7. 6. Grey A, Gamble G, Ames R, Horne A, Mason B, Reid IR. Calcium supplementation does not affect CRP levels in postmenopausal women--a randomized controlled trial. Osteoporos Int 2006;17:1141-5. Competing interests: None declared |
|||
|
|
|||
|
Jayendra Gohil, faculty medical college bhavnagar
Send response to journal:
|
Calcium channel blockers are widely used and beneficial in cardiac & vascular diseases so in an old lady/ person its no wonder calcium is detrimental to vessels and heart. Competing interests: None declared |
|||
|
|
|||
|
Phillip Moradi, SpR Ophthalmology Institute of Ophthalmology EC1V 9EL
Send response to journal:
|
Bolland et al. 1 reported significant increases in the rates of vascular events in women allocated to calcium supplementation. The similarities between these findings and those from the dialysis literature were noted. However , an important omission in the suggestions for future research involves a complete lack of evidence regarding ocular vascular events and calcium supplementation. Central retinal artery occlusion, branched retinal artery occlusion and macroaneurysms are leading causes of visual impairment and blindness in ageing populations. Furthermore, calcium supplementation is regularly prescribed to post menopausal women, with increased risk of developing these ocular conditions, and patients with chronic inflammatory eye diseases requiring prolonged oral steroid treatment. The paucity of data regarding this important group of costly , blinding cardiovascular events should be highlighted to future investigators. References (1)Bolland MJ, Barber PA, Doughty RN, Mason B , Horne A et al. Vascular events in healthy older women receiving calcium supplementation: randomised controlled trial BMJ 2008; 336: 262-266 Competing interests: None declared |
|||
|
|
|||
|
MJ Wilkinson, General Practitioner B34 7BP
Send response to journal:
|
Boland reports a study with many unanswered questions. This is an interesting trawl of data oringinally collected for another purpose and not designed to pick up this end point, so there may be power problems here. There does appear to be a difference in the baseline of the groups in terms of smoking and vascular events, and the drop outs or intention to treat analysis is ignored. Most people prescribe calcium supplements with vitamin D rather than calcium alone, so this paper does not apply to my routine practice and I will not be changing as a result. Further studies are needed that are set up to control for baseline predictors of vascular events. The editorial went too far by suggesting we stop using calcium supplementation now. Yours sincerely, Dr Martin Wilkinson 1. Mark J Bolland, P Alan Barber, Robert N Doughty, Barbara Mason, Anne Horne, Ruth Ames, Gregory D Gamble, Andrew Grey, and Ian R Reid Vascular events in healthy older women receiving calcium supplementation: randomised controlled trial BMJ 2008; 336: 262-266 Competing interests: None declared |
|||
|
|
|||
|
Jinelle N. Ramlackhansingh, SPR Public Health Manchester PCT, Southmoor House, M239LT
Send response to journal:
|
To the Editor, This article by Bolland et al was reviewed at our Public Health Departments monthly Journal club(1). There were some issues highlighted concerning the outcomes used, the method of allocation of the women, and the presented results. 1. The use of composite end point (Myocardial infarction, stroke or
sudden death) creates difficulties in determining the real impact of
calcium supplementation on the relative risk of myocardial infarction. It
appeared as though these three outcomes were combined in order to create a
significant p value.
It is commendable however that this RCT engaged with an often neglected group-the elderly and called for further research into effects of calcium and elderly women. Reference List (1) Bolland MJ, Barbar PA, Dougty RN. Vascular events in healthy older women receiving calcium supplementation:randomised controlled trial. BMJ 2008; 336:262-266. (2) Reid IR, Mason B, Horne A. Effects of calcium supplementation on serum lipid concentrations in normal older women: a randomized control trial. Am J Med 2002; 112:343-347. Competing interests: None declared |
|||
|
|
|||
|
Amy L Davis, Editorial registrar BMJ Editorial, BMA House London WC1H 9JR, John Fletcher, Clinical Epidemiologist BMJ
Send response to journal:
|
We made a mistake in not insisting the title of this paper was clearer. We do insist that the study design is included in the title of all our research and this was indeed a randomised controlled trial. However this paper was not the primary report of the trial. It would have been clearer to readers that this was a secondary analysis focusing on adverse events (item 19 of the CONSORT statement) if this had been in the title in some way. We ask editors to submit a CONSORT checklist to help in the peer review process and we often refer to this when reading the paper. We don’t necessarily insist all items are included in the BMJ publication, though, and wouldn’t necessarily refuse publication if some less important items were missing. We publish the flow diagram for trials where the BMJ publication is the primary report but we don’t publish the checklist. We didn’t publish a flow diagram for this trial as it was a secondary analysis. Harms are often poorly reported in trials but this trial is an exception. The trial protocol specifically addresses the possibility of cardiovascular harms and specifies the individual and composite endpoints that would be reported. It is clear from the first sentence of the methods in this paper that this is a secondary analysis. The first sentence under “Cardiovascular assessment” describes the composite endpoint specified in the protocol. The authors added another composite endpoint and have described how and why they did this. We agree the strength of evidence in this study is weakened because it is from a secondary analysis and uses composite endpoints. But this is much better than a purely observational cohort as the intervention was allocated at random and the analysis and outcomes were described beforehand in a protocol lodged on a trial register. It would be a bit much to expect investigators to steer through an ethics committee an RCT where the primary outcome is a harmful side effect of a drug. Competing interests: None declared |
|||
|
|
|||
|
Christopher Nordin, Consultant Physician Endocrine and Metabolic Unit, Royal Adelaide Hospital, Adelaide 5000, South Australia, Andrew Metcalfe, Department of Mathematics, University of Adelaide
Send response to journal:
|
The implication of the paper by Bolland et al is that calcium supplementation raises the serum (ionised) calcium sufficiently to precipitate calcium phosphate in the coronary vessels, thus causing coronary occlusion. The analogy with vascular calcification from large calcium supplements in renal failure is false; it is the high serum phosphate in renal disease that causes the trouble. The blood and tissue fluids are normally supersaturated, but not oversaturated with calcium phosphate (CaHPO4) i.e. the product is in the metastable region. This means that precipitation of the salt does not occur in soft tissue unless it is damaged or dead and can act as a nucleus or seed. This is known as dystrophic calcification and is why atheromatous plaques calcify. The calcified tissue is a useful marker of the location and severity of atheroma but is not the cause of the vascular occlusion which is due to plaque rupture and the formation of thrombi. Calcification from oversaturation of the tissue fluids with CaHPO4 - as in renal failure - is called metastatic calcification and occurs when the serum product CaXP (each expressed in mmol/L) rises above about 4.5 i.e. when serum phosphate exceeds about 2 mmol/L; it rarely causes coronary occlusion. Hyperphosphataemia is not a feature of normal ageing and would not be present in the vast majority of postmenopausal and elderly women receiving calcium supplements. Moreover, because of the nature of the calcium absorption process, only 5-10 percent of supplementary calcium is absorbed [1]; even with calcium citrate, the maximum rise in serum calcium on a 500 mg tablet is only 5 percent at 4 hours [2]. This can only have a negligible effect on the CaxP product and is extremely unlikely to cause or accelerate vascular calcification. The hypothesis of Bolland et al is therefore implausible on physico-chemical grounds. As for the statistics, the authors state that the number of women needed to treat for 5 years to cause one myocardial infarction was 44, without stating which analysis led to this result. We calculate the NNT from the data in Table 2 as 1/(31/732-14/739), which equals 43. Similar calculations based on the data in Table 3 and Table 4 give 66 and 72 respectively. We therefore guess that the reported figure of 44 is based on Table 2. However, Table 2 takes no account of the covariates, a point which other correspondents have already made in the context of statistical significance. A less misleading calculation could be based on the Poisson regression Model 3, reported in Table 5, but even this ignores the fact that there were more smokers in the calcium group. Also, the estimated probabilities, derived from the Poisson regression, would not be significantly different at the 5% level (P=0.08). It seems unwise to compare the benefits of increased bone density attributed to calcium with such an imprecise statistic as the reciprocal of a non significant difference. We also calculated the NNT to cause one stroke from Tables 2- 4 and obtained values of 60, 79 and 79 respectively. None equals the claimed 56. 1. Nordin B.E.C. Calcium and osteoporosis. Nutrition 1997;13:664- 686. 2. Heller HJ, Stewart A, Haynes S., Pak CYC. Pharmacokinetics of calcium absoprtion from two commercial calcium suppplements. J Clin Pharmacol 1999;39:1151-1154. Christopher Nordin, Visiting Professor, Department of Medicine, University of Adelaide (email christopher.nordin@imvs.sa.gov.au) and Senior Lecturer Andrew Metcalfe, Department of Mathematics, University of Adelaide, South Australia 5005. Competing interests: None declared |
|||
|
|
|||
|
Ivor Cammack, House Officer. Taranaki Base Hospital, New Zealand
Send response to journal:
|
Dear Editor, The presentation of this paper and its results are flawed. The title and abstract of this paper fail to disclose that this article is merely a rehash of data used in an older randomised trial with a different primary endpoint. As many readers will only read the abstract, this is misleading. Tables 2 and 3 describe the incidence of vascular events as reported by the patients themselves. The inaccuracy of this method of collating data is demonstrated by the authors themselves in table 4 which shows the actual events as documented in hospital records. Surely this is a more accurate method of data collection, therefore tables 2 and 3, which are inaccurate, should not have been reported or described in the abstract. There are no statistically significant findings in table 4 which looks at the rates of vascular events as obtained from hospital records. Therefore this is a negative study and the discussion should clarify this. The only positive finding is that patients who take calcium are more likely to 'report' that they have had a myocardial infarction. Instead, the discussion is heavily biased towards cautioning the prescribing of calcium supplements, as if there were a meaningful positive finding in this study. The findings of this study were presented at an American Society of Bone and Mineral Research conference before publication. The New Zealand Listener, a weekly magazine, ran an article in an October 2007 issue describing the results... "There were fewer broken bones and a 75 percent reduction in loss of bone density in the group taking supplements, but further analysis of the research data showed a worrying trend. There were 36 heart attacks among the group taking calcium supplements, and 22 among those taking a placebo. The chances of having a heart attack or a stroke increased by up to 50 percent in the group taking supplements. Professor Ian Reid, a professor of medicine at Auckland University, says the results were unexpected and very significant. “The message seems to be that very high doses of calcium supplements can help bone density, but this may come at too high a cost.” " 1. Clever use of relative risk to accentuate the findings but the results described by this magazine do not agree with those in this article. Professor Reid suggests that the risk of calcium supplements may outweight the benefits, but this is not at all clear in this published article. Finally, in the original trial publication, there was a high dropout rate and the abstract concludes that ..."Poor long-term compliance limits its effectiveness." 2. If drop-out rates and poor compliance were significant, these should also have been documented in this article. Overall, this is a poor example of searching for publishable material in old data. The results have been misrepresented and the NZ lay public have been informed before publication, which makes it difficult for NZ general practitioners to advise their patients should their calcium supplementation be questioned. 1. "Crunch Time", Linley Boniface, NZ Listener. October 13-19 2007 Vol 210 No 3518 http://www.listener.co.nz/issue/3518/columnists/9789/crunch_time.html 2. Reid et al. Randomized Controlled Trial of Calcium in Healthy Older Women. The American Journal of Medicine Volume 119, Issue 9, September 2006, Pages 777-785 Competing interests: None declared |
|||
|
|
|||
|
Ian R Reid, professor University of Auckland, Private Bag 92019, Auckland, New Zealand, Mark J Bolland, Gregory D Gamble, Andrew Grey
Send response to journal:
|
We note with interest the data supplied by Drs Lappe and Heaney. While it is somewhat reassuring that they did not observe an increase in vascular event rate in subjects randomised to calcium, the small number of events means that this finding should be interpreted with caution. These data will be an important addition to a future meta-analysis. We do not agree that the data from the Women’s Health Initiative (WHI) are completely reassuring. Participants in WHI were younger, had a higher baseline calcium intake and a higher BMI than the subjects in our study. If the tertiles of each of these variables are chosen that most closely match the characteristics of the subjects in our own study, then the calculated relative risk of coronary heart disease associated with calcium use in WHI is 1.36. Other correspondents raise methodological issues. It is made clear in the manuscript that vascular events were specified in the original protocol for this study as secondary events, the hypothesis being one of benefit. This protocol and a subsequent expansion of it were supplied to the BMJ, along with a full Consort Statement (also supplied to the American Journal of Medicine as part of the primary publication from the study). The original protocol did not make provision for adjudication of events, nor did it provide for searching of the National Hospital Admissions Database. Therefore, the originally specified data are those provided in Table 2, which clearly demonstrate significant effects of calcium supplementation on the risk of myocardial infarction. The data in Tables 3 and 4 provide further scrutiny of the primary data. There is still a statistically significant effect of calcium supplementation on vascular endpoints in the most conservative dataset, provided in Table 4. We believe that we were appropriately conservative in interpreting these data in our manuscript. In general, adjustment for multiple statistical tests is not made when assessing adverse events, in order to maximise sensitivity for such negative outcomes. Adjustment for multiple testing is even less appropriate in the present context, since the adverse effects studied were pre-specified in the original protocol. The two groups in this study were well balanced with respect to cardiovascular risk factors, P values for the factors presented in Table 1 ranging from 0.22 to 0.99. Prior HRT use was not different between treatment arms (calcium 8%, placebo 6%, P=0.13). We believe our paper presents important, unexpected findings from the secondary analysis of a carefully conducted, randomised, controlled trial. They do not definitively establish that calcium supplementation increases vascular events, but they mandate that this possibility is further investigated as a matter of urgency. Ian R Reid
Competing interests: IRR has received research support from and acted as a consultant for Fonterra and Mission Pharmacal. |
|||
|
|
|||
|
Audrey Lorraine Aspeling, Division 1 Registered Nurse Maroondah Hospital, Ringwood East, Victoria Australia 3135
Send response to journal:
|
As Calcium Channel Blockers counteract ischemia is it possible that supplemental Calcium or Vitamin D could cause ischemia in susceptible patients? Competing interests: None declared |
|||