Are breast cancer screening programmes increasing rates of mastectomy? Observational study
BMJ 2002; 325 doi: https://doi.org/10.1136/bmj.325.7361.418 (Published 24 August 2002) Cite this as: BMJ 2002;325:418All 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.
I have a major concern with this paper by Paci and colleagues.
As per the figure (Paci et al, BMJ 2002;325:418) the total (early
stage + late stage) incidence of breast cancer increased from about 2.7
(1.3+1.4)per 1000 in 1990 to 4.1 (2+2.1) per thousand in 1996.
So, if all the data was captured, I cannot understand what happened to
more than 70% of women being diagnosed with breast cancer in 1996. In 1990
there were 2.7 /1000 diagnoses and 2.2/1000 operations. In 1996, there
were 4.1 /1000 diagnoses but only 1.2 /1000 operations!? What happened to
the other 2.9 /1000 patients diagnosed with breast cancer? What
operations/treatments did they have?
Clearly a large amount of data about these patients' operations is
missing. No conclusion about operation rates can drawn from the data
presented.
Competing interests:
None declared
Competing interests: No competing interests
In response to some comments it is worth pointing out that:
• Gøtzsche and others have claimed that the introduction of mammographic
screening increases mastectomy rates.
• We have showed that this was not the case in Florence
However, Professor Baum’s irony about scars in the breasts of
Florentine women is at best provocative and aggressive.
In table 1 we present data of rounds 1 to 4 of the Florence
screening programme which started in 1990. It is a matter of regret to us
that some women attending for screening undergo surgical biopsy for a
condition which ultimately proves benign, but in total the number of women
with a negative excision biopsy was 190 in 10 years of duration of the
breast screening programme, with around 130,000 screens. Whatever your
opinion might be about aggressiveness of in situ carcinoma ( the issue of
overdiagnosis will be discussed in a forthcoming paper), the number is far
from epidemic. Women with a breast scar are in the large majority breast
cancer patients. In turn, the vast majority of these are invasive cancer
patients.
We suggest that is better to maintain the Botticelli's birth of Venus
as a symbol of the beauty of Florentine women.
Eugenio Paci
Stephen W. Duffy
Table 1 Service screening data from the Florence Breast Cancer Screening programme Round Women screened No of No of No of No of population referrals excision screen screen (50-69) for biopsies detected detected assessment invasive carcinoma cases in situ 1 28295 1284 278 211 21 2 33973 1420 224 158 30 3 35829 1397 215 167 12 4 35869 1458 222 133 17
Competing interests: Table 1Service screening data from the Florence Breast Cancer Screening programmeRound Women screened No of No of No of No of population referrals excision screen screen (50-69) for biopsies detected detected assessment invasive carcinoma cases in situ1 28295 1284 278 211 212 33973 1420 224 158 303 35829 1397 215 167 124 35869 1458 222 133 17
Sir,
With reference to the article ‘Are breast cancer
screening programmes increasing rates of
mastectomy ? Observational study’, Paci, E et al, BMJ
325, 24th August 2002, 418
Although there is value in raising the issue, that breast
conservation surgery and mastectomy rates may
change primarily as a result of the introduction of
screening, the conclusions as presented are internally
invalid and are not credible in the light of the evidence
presented.
There are many confounding and unaccounted for
variables which could be contributing to the observed
reduction in this population’s absolute mastectomy
rates which deserve identifiaction and discussion.
Screening Biasing Results:-
The advent of screening could bias the results towards
breast conserving surgery; no mention is made as to
whether the lumpectomies as described were purely
therapeutic or contained diagnostic breast biopsies as
a result of screening. Thus, the meaning of the terms
“breast conserving surgery” and “lumpectomy” requires
clarification in this context. Hidden within these results
of an increase in ‘breast conserving surgery’ could well
be what could have turned out to be unnecessary
biopsies for falsely positive mammograms, and it
would be dangerous to assume that all “lumpectomies’
were performed for a therapeutic intent only. This figure,
without such clarification, is potentially biased and
therefore internally invalid. It contains a potential bias in
favour of screening increasing the local excision rate
and relatively decreasing the mastectomy rate which in
reality could be hiding an increased rate of biopsies for
false positive disease.
Therefore a more meaningful figure would be the
observed rates of better defined surgical categories in
the screened versus the non screened population. As
such it can not be assumed that there was a purely
therapeutic intent in these patients, and the reported
mastectomy rate is purely an quantitative observation
from which no qualitative inference as to the reasons
underlying it can be made.
Neither is mention made made of the quantitative
contribution of the screened and non screened
populations towards the mastectomy rates, with an
absolute mastectomy rate only being presented. The
same problem applies to the breast conservation
surgery rate. If only 60% of invited women take up the
offer of screening, presumably the increase in breast
conservation surgery is attributed to this group,
however, the analysis in the paper refers to the
population of women invited as a whole rather that
those who accepted the screening offer. Herein lies an
unquantifiable bias.
Other Counfounding Factors:
Working Practices:-
No mention is made of whether changes in local
working practices have influenced the mastectomy
rates. Although it is appreciated that most of the results
date from 1990-1996, whether neoadjuvant
chemotherapy or primary hormonal therapy has been
used in this population has not been commented on. It
is exactly treatment biases such as these which can
influence mastectomy rates.
DCIS:-
It is also naive to include DCIS with early breast cancer.
In the paper Stage 0 Breast Cancer, DCIS, is classified
with early stage invasive breast cancer.As the two are
biologically different entities with potentially different
biological activities, prognosis and treatment, this will
introduce an unpredictable bias into the interpretation
of observed mastectomy rates. It almost leads to the
assumption that DCIS invariably leads to invasive
disease.
Observational Study:-
Being an observational study, there is no control group
to make a valid comparision, either temporally with
historical controls or any attempt to observe
concomitant mastectomy rates in screened versus non
screened populations.
Unfortunately this is essentially a retrospective study and
as such one can not expect the study design to have
been as appropriate as if it had been perormed as a
prospective study. However I do not feel that there is
enough internal or external validity within this paper to
reach the conclusions stated by the authors.
Competing interests: No competing interests
Sir-
The division of operation by Paci and colleagues( BMJ
2002;325:418)
between radical mastectomy and breast conserving operation
seems to equate
radical mastectomy (=total mastectomy + axillary clearance)
with simple
mastectomy ( without axillary clearance) in their argument
that breast
screening in Florence results in more frequent breast
conservation and less
frequent mastectomy. Maybe there was never need in Florence
between 1990 and
1996 to remove the whole breast without axillary clearance?
It also almost
looks from their figure that in 1991, 1992, and 1995 there
were more
operations than there were cases of cancer.
Whilst their figure might well show that between 1990
and 1996 the
proportion of early breast cancer cases treated by breast
conserving
operations did not change, and that the proportion of women
with tumours =/>
2 cm pathological diameter treated by axillary clearance as
well as total
mastectomy diminished, these trends might well result from
features not
directly related to tumour size caused by the introduction
of screening,
such as the undoubted changes of fashion for breast cancer
treatment during
the last twenty years.
Their inclusion of data from the years before 1990 (
breast screening
introduced in Florence) might have thrown light as to how
likely the
apparent reduction in radical mastectomy in favour of breast
conservation
might not have arisen from confounding variables.
Yours, Robert Hall
Hon research fellow
Competing interests: No competing interests
Language is very important and I am concerned buy the use of
language in many medical articles as I feel these are an attempt to
confuse and intimidate the reader. In this article the data is presented
in a potential flawed manner and misses a very basic issue and confounding
factor
There is observed fall in the number of mastectomies from 1990 after
the advent of breast screening and it is strongly suggested that trend
is related to breast screening programme. We are not shown, however, the
trend in mastectomy surgery prior to 1990 and whether a downward trend had
preceded screening mammography. It should be noted that from the mid 1980s
it was recognised that lumpectomy was as effective as mastectomy1 and
this change in surgical practice might explain the observed reduction in
mastectomy rates rather than breast screening.
There is a huge industry around breast screening and people want to
believe it works. We should recognise the harm of health care
interventions and believing is not enough.
1.Review confirms lumpectomy as safe as mastectomy Fiona Godlee
BMJ 1995; 311: 1451-1452.
Competing Interest "Pledge at nofreelunch.org"
Competing interests: No competing interests
The observational study by Eugenio Paci and colleagues builds on a
false assumption. The authors assert that if screening increases the
number of mastectomies, populations in which screening has been introduced
should see a subsequent increase (1). This is not correct. Since the
mastectomy rate has gone down steadily throughout many years, also in
countries without screening, it is only to be expected that Eugenio Paci
and colleagues also find a decrease in the mastectomy rate in the period
1990-1996 in Florence, when screening was introduced.
The relevant question is whether the decline in the mastectomy rate
is slower when women are invited to participate in screening programmes,
compared to when they are not invited. We found evidence from the
randomised trials of screening that this is the case (2). Furthermore, the
authors’ findings from Florence are contradicted by a far larger study
from the Southeast Netherlands, where screening was introduced in the same
time period. I calculated that the number of invasive cases increased by
78%, the numbers of women who underwent breast-conserving surgery
increased by 71%, and numbers of women who underwent mastectomy increased
by 84% (3). What is more, these authors did not include carcinoma in situ
which is rarely detected without screening but is frequently treated by
mastectomy (BASO Breast Audit 1999/2000, available from
http://www.cancerscreening.nhs.uk/breastscreen/publications.html)
There is much misinformation about screening. I therefore recommend
women who want to get a fair and balanced account of this issue to consult
an evidence-based consumer organisation, for example the National Breast
Cancer Coalition in USA (http://www.stopbreastcancer.org/bin/index.htm,
Positions, Facts and Analyses). At present, breast cancer screening is not
possible without overdiagnosis and overtreatment. This also applies to
mastectomies, and it should be remembered that breast conserving surgery
with radiotherapy is a pretty rough treatment as well that can lead to
decreased survival (4).
Competing interests: None.
1. Paci E, Duffy SW, Giorgi D, Zappa M, Crocetti E, Vezzosi V, et al.
Are breast cancer screening programmes increasing rates of mastectomy?
Observational study. BMJ 2002; 325: 418.
2. Olsen O, Gøtzsche PC. Systematic review of screening for breast
cancer with mammography
(http://image.thelancet.com/lancet/extra/fullreport.pdf).
3. Gøtzsche PC. Trends in breast-conserving surgery in the Southeast
Netherlands: comments on article by Ernst and colleagues. Eur J Cancer
2002;38:1288.
4. Early Breast Cancer Trialists' Collaborative Group. Favourable and
unfavourable effects on long-term survival of radiotherapy for early
breast cancer: An overview of the randomised trials. Lancet 2000;355:1757-
70.
Competing interests: No competing interests
Letter to BMJ August 2002
Editor,
Re: Are breast screening programmes increasing rates of mastectomy?
There are two welcome side effects of population screening programmes
for breast cancer. Firstly an increasing awareness of the disease which
encourages symptomatic women to present with smaller tumours and secondly
the establishment of specialist clinics set up to cope with the flow of
worried women resulting from the trawl of the asymptomatic population.
Specialist clinics are more likely to offer breast-conserving techniques.
It is therefore a great over-simplification for Paci on his colleagues to
assume that mammographic screening per se is responsible for the fall in
mastectomy rates in Florence ( BMJ 24th August 2002).
Furthermore it is an over-interpretation of the data to extrapolate the
findings from one city in Italy to National population based programmes
such as the NHSBSP.
In Florence the mastectomy rate is 33% for screen-detected cancer, which
compares well with our national statistics for 2000-2001 of 32% for DCIS
and 29% for invasive cancer.[1]
However in the UK there is an enormous range with for example 42% of women
in Wales experiencing mastectomy for DCIS (“ early breast cancer”!) and
38% of women in the Principality treated by mastectomy for screen detected
invasive disease.
I have been unable to trace the trends in mastectomy rates in the UK since
screening was introduced in 1988 but a natural experiment would be to
consider the rates in women under and over 50 years of age. If the trends
were the same then this could not be attributed to screening as the under
50s are not on the invitation list.
One final point, I notice that the absolute rate for breast cancer surgery
in Florence has risen from 2.26 to 2.49 per 1000 women a year between 1990
and 1996, which is a 10% rise. I wonder if this could be attributed to the
over diagnosis of Duct Carcinoma in Situ (DCIS). DCIS accounts for about
20% of “cancers” detected at screening [1] and it has been estimated that
at least half of these would fail to progress to an invasive phenotype if
left undiscovered [2].
The Uffizi Gallery in Florence is home to Botticelli’s “ birth of Venus”.
In this popular icon we see Aphrodite rising from a seashell demurely
covering her left breast with her hand. I always thought this was modesty
but maybe she’s hiding the scar of an unnecessary operation for screen
detected DCIS?
Yours Sincerely,
Michael Baum
References:
[1] NHS Breast Screening Programme and British Association of
Surgical Oncology Breast Group; An Audit of screen detected breast cancers
for the year of screening April 1999 to March 2000. Produced by the West
Midlands NHS and Cervical screening quality assurance reference centre:
Published by the NHS Breast Screening Programme April 2001
[2] Neilsen M, Thomsen JL, Primdahl S et al: Breast cancer and atypia
among young and middle aged- women : A study of 110 medico-legal
autopsies. British J Cancer 1987;56:814-819
Competing interests: No competing interests
The NHS Breast Screening Programme screened 1,409,790 women aged
mainly between 50-65 years in 1999/2000. [1] 9,525 cancers were detected
of which 2,009 (ca. 21%) were ductal carcinoma in situ (DCIS). This
represents overall 6.39 cancers detected per 1,000 women screened.
The `Florence` observational study [2] studied a total of 59,947
women aged 50-69 in the seven years from 1990-1996. Total numbers of
cancers detected, or of DCIS were not provided in the researchers` report,
but at a rate of, say, ca. 2 per thousand early stage cancer and, say,
0.75 per thousand late stage cancers (T2+) shown in their graph, we could
estimate, say 200 cancers at most were detected.
Bearing in mind the biases introduced by an average attendance in
Florence of only 60% of those invited, compared with 75.4% in the U.K.;
the cultural differences; and the difference in quality of evidence
obtained from a small number of participants in an observational study
compared with evidence from a large systematic review executed in
accordance within a rigorously designed Cochrane protocol [3] cited by the
authors, suggests to me that it is unwise to jump to the conclusions shown
in the media coverage in the UK today relating to the NHS Breast Screening
Programme. No mention was made of the psychological consequences to women
diagnosed through screening, receiving treatment by mastectomy or by
lumpectomy with or with radiation, (particularly for borderline cases), or
of other considerable psychological outcomes for which the researchers in
the Danish Review provided data, references and comment.
In the media coverage in the U.K. on day of publication, the
conclusion drawn by Paci et al that "the rate of breast conserving surgery
has increased significantly with the advent of screening, and the rate of
radical surgery has declined significantly" was presented with scarcely a
mention of the disparities in the weight, or place, of the findings. The
hasty endorsements in the media by public figures implying a refutation of
the UK findings that mammographic screening leads to over-diagnosis and
over-treatment is to be deplored, when the data from the UK indicate
otherwise.
Those who selectively present inadequate evidence to support a belief
will ultimately lose the trust of the public at large in the face of
mounting evidence that mammographic screening is not as effective as had
been predicted at reducing mortality from breast cancer; is certainly not
cost-effective; has numerous adverse effects for many women invited who
are not accurately diagnosed; and leads to mutilating treatment,
particularly of the 1 out of 5 borderline cases diagnosed through the
programme, three-quarters of whom will never progress to invasive cancer.
Those women would certainly agree that that surely is over-treatment?
(1,500 in the U.K. in 1999/2000 alone.)
Hazel Thornton.
Independent advocate for quality in research and healthcare.
References:
[1] NHS Breast Screening Programme Annual Review 2001.
[2] Eugenio Paci, Stephen W. Duffy, Daniela Georgi, Marco Zappa,
Emanuele Crocetti, et al. Are breast cancer screening programmes
increasing rates of mastectomy? Observational study. BMJ 2002; 325:418
[3] Olsen O. Goetzsche PC. Cochrane Review on screening for breast
cancer with mammography. Lancet 2001; 358: 1340-1342
Competing interests: No competing interests
Re: Incomplete data and wrong conclusion: what happened to 70% of women diagnosed in 1996?
I have realised that I looked at the wrong lines in the graph while
doing the above calculations- so my conclusions and comments are very
embarrassingly completely unfounded. I wish to withdraw the above rapid
response.
Competing interests:
I am the author of the earlier response and wish to delete that response
Competing interests: No competing interests