Pre-existing risk factor profiles in users and non-users of hormone replacement therapy: prospective cohort study in Gothenburg, Sweden
BMJ 1999; 319 doi: https://doi.org/10.1136/bmj.319.7214.890 (Published 02 October 1999) Cite this as: BMJ 1999;319:890
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Editor- Rödström et al in their report of a prospective cohort study
(1) found that women who would subsequently use hormone replacement
therapy (HRT users) already had a better cardiovascular profile before the
start of treatment than those women who would remain untreated (non-
users). Based on this finding the authors conclude that the beneficial
effect of HRT found in the available observational studies may be
attributable to a selection bias and that it is too early to recommend HRT
for prevention of coronary heart disease (CHD) before controlled
randomised studies have been performed.
It is widely recognised that women who choose to use HRT may also
make other choices that could be associated with reduced rates of heart
disease. The vast majority of the available observational studies have
taken this concern into account. The “healthier hormone user effect”
appears to apply to a large extent to studies in the general population.
To avoid the effect of these postulated differences, large epidemiological
cohort studies of more homogeneous populations have shown that after
analysis adjusting for many known coronary risk factors and lifestyle
related variables, results still have yielded strong inverse associations
between current HRT use and CHD incidence. More recently, other population
-based studies (2,3) have examined the CHD risk in relation to the dose,
duration and type of hormone used. Results have shown a reduced risk of
CHD for long-term medium-dose compared with low-dose or short-term
estrogen use, which cannot be explained by a self-selection or a healthy-
user bias.
Furthermore, a recent international clinical synthesis panel on HRT
organised by the European Institute of Oncology, Harvard School of Public
Health, the Istituto Superiore di Sanitá, and The Lancet (4), declared on
their evaluation of the risks and benefits of HRT on cardiovascular
disease that “although a cause and effect relation is not proved, evidence
that HRT lowers the risk of CHD in women without a history of the disease
is sufficiently strong to consider this potential benefit when deciding
whether to use HRT”.
We believe this issue will be addressed more directly in the next
decade when the results of ongoing clinical trials will be known. However,
the findings of a CHD risk rate reduction for HRT users shown by a large
number of observational studies in which serious attempts have been made
to control for multiple factors, including the better cardiovascular risk
profile of HRT users, cannot be neglected.
Lia Gutierrez
Cristina Varas
Alejandro Arana
Global Epidemiology
Novartis Pharmaceuticals
Gran Via de les Corts Catalanes 764
08013- Barcelona
Spain
The authors are employees of a manufacturer of hormone replacement
therapy products.
REFERENCES
1. Rödström K, Bengtsson C, Lissner L, Björkelund C. Pre-existing risk
factor profiles in users and non-users of hormone replacement therapy:
prospective cohort study in Gothenburg, Sweden. Br Med J 1999;319:890-893
2. Grodstein F, Stampfer MJ, Falkeborn M, Naessen T, Persson I.
Postmenopausal hormone therapy and risk of cardiovascular disease and hip
fracture in a cohort of Swedish women. Epidemiology 1999;5:476-480
3. Varas C, Garcia-Rodriguez LA, Perez-Gutthann S, Duque A. Hormone
replacement therapy and the incidence of acute myocardial infarction: a
population-based nested case-control study. Circulation (in press)
4. Hormone replacement therapy: Clinical synthesis conference. Lancet
1999;354:152-155
Competing interests: No competing interests
Response to Ross Lawrenson, 21st October, 1999
Sir,
We thank Dr Lawrenson for his comment on our paper "Preexisting risk factor profiles in
users and non-users of hormone replacement therapy: prospective cohort
study in Gothenburg, Sweden"(1).
To fully answer his question regarding diabetes, we would have to know
the exact dates of the start of medication as well as diabetes onset.
Unfortunately, with our data that is not possible at the moment, but we
can report on the association between HRT use and diabetes in our cohort,
in which 92 incident cases were observed. The odds ratio for having or
acquering diabetes during the 24 year period (1968/69-1992/93) in relation
to using HRT was 0,37 (CI 0.3-1.03),p=0.06. Further adjustment for body
mass index, systolic blood pressure and education weakens the
association(OR=0.56, 95%CI 0.19-1.59),p=0.27.
The observed shift of the association towards the null underscores the
importance of controlling, when possible, for factors that differed before
HRT started, when evaluating the "effect" of HRT on disease endpoints. We
observed similar attentuation patterns for HRT in relation to incident
myocardial infarction, after adjustment for the potential confounders
noted aabove. Thus, we reiterate our conclusion that pre-existing health
differences between users and non-osers of HRT may create spurious
associations with subsequent disease onset.
Ref. Rodstrom K, Bengtsson C, Lissner L, Bjorkelund C. Pre-existing
risk factor profiles in users and non-users of hormone replacement
therapy: prospective cohort study in Gothenburg, Sweden. BMJ 1999;319:890-
93
Yours sincerely
Kerstin Rodstrom
Cecilia Bjorkelund
Calle Bengtsson
Lauren Lissner
Competing interests: No competing interests
Editor,
There is considerable evidence that the post menopausal woman with
diabetes
is at marked increased risk of cardiovascular disease (CVD) (1). To
date there are few studies assessing the effects of specific treatments in
this high risk group. The recent paper by Rodstrom et al indicated
potential problems of using observational data to estimate the
cardiovascular benefits of hormone replacement therapy (HRT)(2). The
authors discussed the issue of diabetes in their methodology although
they presented no pertinent results.
We have shown that in the UK women with diabetes are less likely than
women
without diabetes to be prescribed HRT (OR 0.7 95% CI 0.6 - 0.8) (3).Our
group also carried out a case control study looking at women with diabetes
and comparing the use of HRT in women with CVD compared to women with
diabetes and no CVD and shown that they are 40% less likely to have been
prescribed HRT(4). The study also showed that women with a past history of
a hysterectomy had a doubling in this risk. It is likely women who have
had
a hysterectomy are also more likely to be prescribed HRT and thus past
hysterectomy should be considered an important potential confounding
variable. It would be of interest to know whether Rodstrom et al had
information on a history of past hysterectomy and how this related to HRT
prescribing. We agree with the authors that it is too early to judge
whether HRT should be prescribed prophylactically. However women with
diabetes may well
benefit from HRT in a number of ways and we believe that they should not
be discriminated by prescribers.
Yours sincerely
Ross Lawrenson
Director of Primary Care Research
University of Surrey,
Guildford
Michael Feher
Senior Lecturer in Clinical Pharmacology
Chelsea and Westminster Hospital,
Imperial College School of Medicine,
London
1) Grodstein F, Stampfer MJ, Manson J, Colditz G, Willett WC, Rosner
B et al. Postmenopausal estrogen and progestin use and the risk of
cardiovascular disease. New England Journal of Medicine 1996; 335 (7): 453
– 61
2) Rodstrom K, Bengtsson C, Lissner L, Bjorkelund C. Pre-existing
risk factor profiles in users and non-users of hormone replacement
therapy: prospective cohort study in Gothenburg, Sweden. BMJ 1999: 319:
890-3
3) Lawrenson RA, Newson RB, Feher MD. Do women with diabetes receive
hormone replacement therapy? Practical Diabetes International 1998;
15(3):71-72
4) Lawrenson RA, Leydon GM, Newson RB, Feher MD. Coronary heart
disease in women with diabetes: positive association with past
hysterectomy and possible benefits of hormone replacement therapy.
Diabetes Care 1999; 22(5): 856-7
Competing interests: No competing interests
Re: Differences in pre-existing cardiovascular profiles have already been taken into account
Response to Lia Gutierrez et al
Sir
We thank Senior Epidemiologist Lia Gutierrez et al for their comment on
our paper "Pre-existing risk factor profiles in users and non-users of
hormone replacement therapy: prospective cohort study in Gothenburg,
Sweden"(1)
Lia Gutierrez and co workers, all from Novartis, comment that differences
in pre-existing cardiovascular profiles have already been taken into
account in many studies, but that results of ongoing randomised controlled
trials are not yet available.
Our study strongly indicates that there has been a selection bias in
observational studies. It does not however, investigate whether or not HRT
has a beneficial effect on CHD in women. As in all epidemiological
research, observational studies should be considered together with
evidence from randomised controlled studies before general recommendations
are formulated. Our study is somewhat unique in that it has the
possibility of controlling for confounding risk factors that were present
prior to treatment. In this way it will be possible in future analyses to
adjust any risk estimates for covariates that may have influenced choice
of treatment. Our belif is that HRT may well influence in risk reduction,
but that 50% might be an overestimate of the true reduced risk.
Yours sincerely
Kerstin Rodstrom
Cecilia Bjorkelund
Calle Bengtsson
Lauren Lissner
Competing interests: No competing interests