Information from drug companies and opinion leadersBMJ 2003; 326 doi: https://doi.org/10.1136/bmj.326.7400.1156 (Published 29 May 2003) Cite this as: BMJ 2003;326:1156
All rapid responses
The responses to our editorial call for some clarifications. Overall,
we believe that our alert or the risk of double standard in information
between peer reviewed, non peer reviewed journals, conferences and opinion
leaders remain unchallenged. We tried to support our alert with two
examples: the document of the European Federation of Pharmaceutical
Companies (EFPIA) and the debate that followed the publication of the
ALLHAT study on hypertension management. Nobody has so far challenged our
remarks about the EFPIA’s document poor quality and misleading contents
and we are concerned that Dr Baicus (BMJ Rapid Responses June 2, 2003)
suggested that the situation may be even worse than the one we alluded to.
We are glad that Prof.’s Heckman’s (BMJ Rapid Responses May 30, 2003),
Ambrosioni’s (BMJ Rapid Responses July 12, 2003) and Agabiti Rosei’s (BMJ
Rapid Responses July 12, 2003) provide us with the opportunity to clarify
further some points that for space reasons could not be adequately
discussed in our editorial.
Before going into the details, however, we want to draw readers’attention
to the fact that Prof Ambrosioni confirmed what he said in the concluding
message of the article we quoted (1) “it is unacceptable the transfer of
ALLHAT results to Europeans, and in particular to Italian clinical
practice where the percentage of black people (those who benefit of
diuretic treatment) where the percentage of black people is low and very
far from the 36% of the population in the study”. All those caring for
hypertensive patients can judge this statement by themselves and compare
it with articles and commentaries that followed the publication of ALLHAT
(2) including those published in the Journal of Hypertension and referred
to in Prof Agabiti Rosei’s response as well as in our Editorial
Specific points were raised against ALLHAT in the two articles (1, 3)
and in the Rapid Response by Heckman GA (BMJ Rapid response May 30, 2003)
we referred to in our editorial:
the absence of a washout period in ALLHAT,
the lack of a proper validation of heart failure events,
the better control of hypertension achieved with thiazides is responsible
of the higher incidence of some end-points in the early stopped doxazosin
arm and more recently with amlodipine and lisinopril
1 The fear of a harmful effect or the potential negative effect of
switching (randomising) a patient with left ventricular disfunction
already on thiazide treatment to one of the other treatments (amlodipine,
lisinopril or doxazosin) is a far-fetched argument: are we really saying
that a hypertensive patient should not be abruptly switched from a
diuretic to another drug? Are prescribers to be informed that a patient
cannot be switched so easily? We admit to be puzzled by such a statement.
2 Lack of adequate validation of heart failure and the fact it was not a
pre-specified end-point. The endpoint was specified in the original
protocol and the validation, as reported in the discussion in JAMA, was
extensive. In their reply to critiques in fact Wright and co-workers (2)
stated: “Furthermore, although heart failure events were not designed to
be centrally reviewed in the original protocol, when large differences
were noted between treatments groups, a staged review of the hospitalised
and fatal cases, approximately 1773 of 2188 cases or 81% of heart failure
cases, was initiated. … The clinical presentation, diagnostic criteria,
and high-case fatality (2-year mortality rate of 20%) after the diagnosis
of the heart failure observed in ALLHAT was certainly more than “ankle
The better control of hypertension with thiazides is true in comparison
with lisinopril and doxazosin whereas it is not so for amlodipine though
the difference in heart failure is increased with all three. The fact that
everything is due to lowering blood pressure per se seems to be
contradicted by these results. Surprisingly, no comment is made by
Ambrosioni and Agabiti Rosei to the fact that amlodipine - a calcium-
channel blockers - is still the top-selling antihypertensive drug in Italy
Finally, we envy Prof Agabiti Rosei’s assertive confidence that his
judgment has never been influenced by relationship with drug companies.
Many articles included in the BMJ special issue and illustrating the
problems emerging when “dancing with porcupines” would -in our opinion -
suggest a humbler and more prudent attitude.
Alessandro Liberati MD,
University of Modena and Reggio Emilia, Modena Italy
Nicola Magrini, MD,
Centre for Evaluation of Effectiveness of Health Care (Ce.V.E.A.S.),
Azienda USL Modena Modena, Italy
Valutare l’effectiveness pesando tutti gli effetti. Troppa distanza dalla
Il Sole 24 Ore Sanità 2003, Marzo 4 pag 21
Wright JT, Davis BR, Cutler JA
Long term cardiovascular consequences of diuretics vs. calcium channel
blockers vs angiotensin-converting enzymes inhibitors (Reply to)
Agabiti Rosei E
Quali nuovi insegnamenti dallo studio ALLHAT per il trattamento della
Ipertensione e Prevenzione Cardiovascolare 2002; 9:133-135 (available in
English at http://www.unibs.it/castell/IPCenglish.pdf)
Ministero della Salute.
Osservatorio Nazionale dei Medicinali (OSMED) “L’uso dei farmaci in
Italia. Rapporto Nazionale 2002”
Roma, Il Pensiero Scientifico 2003
Competing interests: No competing interests
In a recent Editorial on BMJ (1), Liberati and Magrini referred to a “predefined strategy aimed at challenging unwelcome results” from the ALLHAT study. Such a strategy would be a kind of conspiracy between “pharmaceutical companies and some opinion leaders”, aimed at denigrating ALLHAT on national journals read by practitioners or policy makers, while international journals “substantially praised the strengths” of that study.
The Editorial includes some wrong statements that must be brought to the readers’ attention because they may give them an idea of the methodology used by the authors in dealing with this matter.
First, Drs. Liberati and Magrini accused me, among others, of adopting a “double standard” of information. In fact, I have expressed my view on ALLHAT in one article only (how can this be “double standard”?), i.e. the Editorial on Ipertensione e Prevenzione Cardiovascolare(2) of which I have now provided the English version at http://www.unibs.it/~castell/IPCenglish.pdf, so that the readers of BMJ may actually read what I have written and not what Liberati e Magrini wanted them to read.
Second, Liberati and Magrini write that in the Journal of Hypertension ALLHAT received only favourable reports. This is false because highly critical commentaries on ALLHAT were published both in this and in other European and American Journals, including BMJ(3).
For those interested, I have provided a link (http://www.unibs.it/~castell/IPCenglish.pdf) with a list of them. I hope Liberati and Magrini will consult it and avoid in the future the sin they accuse other people of, i.e. selective suppression of information.
Third, Liberati and Magrini were wrong in stating that in ALLHAT 55% of the patients were controlled by monotherapy; in fact, it has been clarified that 63% of the patients required two or more drugs(4). In addition, they were also clearly wrong in stating that heart failure was “a totally validated end-point”; in fact, there was no independent Event Committee and a “soft” end-point such as heart failure was checked post-hoc in only a small sample of hospitalized cases.
I fully agree that doctors must not be influenced in any biased or unfair way by drug companies and that patients should be given the best possible treatment for their disease, based on scientific evidence. I’m also convinced, however, that bias may also come from health providers and that this type of conflict of interest, namely that of Drs. Liberati and Magrini, cannot be sublimated just by defining it as a “struggle for truly independent information”.
1. Liberati A, Magrini N. Information from drug companies and opinion leaders. BMJ 2003;326: 1156-1157. (31 May.)
2. Agabiti Rosei E. Quali nuovi insegnamenti dallo studio ALLHAT per il trattamento dell’ipertensione arteriosa. Ipertensione e Prevenzione Cardiovascolare 2002; 9:133-135. or [English translation: http://www.unibs.it/~castell/IPCenglish.pdf]
3. Williams B. Drug treatment of hypertension. BMJ 2003;326: 61-62. (11 January)
4. Cushman W.C. et al – Success and Predictors of Blood Pressure Control in Diverse North American Settings: The Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT). J Clin Hyp 2002; 4: 393-404
EAR has received research grants, fees for consulting and reimbursements for attending symposiums from many different pharmaceutical companies as well from public institutions. He believes that this fact does not influence his view of scientific data
Competing interests: No competing interests
I refer to the editorial “Information from drug companies and opinion
leaders. -Double standards in information for medical journal and
practitioners should go “by Alessandro Liberati and Nicola Magrini BMJ
As an example of this double standard which “represents a predefined
strategy aimed at challenging unwelcome results” they quoted my article on
ALLHAT trial printed in an Italian newspaper (1).
In their editorial on BMJ the authors not only did not furnish any proof
of their offensive affirmations but they operated a full distortion of my
writing by omitting substantial parts and misquoting other ones. They did
not even mentioned the content of the first part of my article: in this
part I wrote “that ALLHAT, the largest study ever performed on treatment
of hypertension, has properly found large space on the press. A minor part
of these articles was reserved to the important informations obtained by
the study and great emphasis was given to a partial interpretation of the
results suggested by the authors” And this was of particular relevance
because the results on primary end point were conditioned by those on
secondary end points on which the conclusions were based.
In this completely ignored first part of my article I reported the
results of ALLHAT on primary end points as they have been printed on the
ALLHAT main report: chlortalidone, amlodipine and lisinopril gave the same
protection on coronary mortality and non fatal myocardial infarction. I do
not see any possibility of a different interpretation of these results and
any chance for those who wrote the editorial to evaluate my position as
“double standard or a prespecified strategy aimed….” if not by ignoring
the first part of my article.
Following this part I continued: “the frequency of some secondary end
points, stroke and combined cardiovascular diseases, was higher in the
lisinopril group than in chlortalidone group”; between the two groups
there was a significant difference in systolic blood pressure levels. Then
I considered the possible causes of these differences in blood pressure
and their consequences.
The authors of the editorial beside ignoring this substantial part of my
article misquoted my sentences in many ways. Let us compare what the
authors of editorial attributed to me and what I wrote.
In the editorial: Liberati and Magrini wrote that I saw as a major
flaw in ALLHAT “the ineffectiveness of monotherapy”.
What I wrote: In “more than 60% of the patients it was necessary to
add a second drug to control blood pressure and some one of these drugs
may increase the effect of diuretic and not that of ACE inhibitor”, which
is substantially different.
In the editorial: “the fact that 90% of the population was already
treated with an antihypertensive drugs (doesn’t this happen in almost all
What I wrote: “it is difficult to evaluate the real incidence of
some cardiovascular events (i.e. heart failure)” “in absence of a period
of washout when 90% of the patients were on treatment “ (mainly with
In the editorial: “heart failure was not a prespecified end point (it
was, and totally validated too)”.
What I wrote: “nevertheless the high number of patients, hospitals
and physicians participating into the study, only 10% of diagnosis have
been randomily evaluated”. On the main report of ALLHAT (2) page 2983 is
stated: more detailed information was collected on a random (10%) subset
of CHD and stroke events to validate the procedure of using physician
In the editorial: “and the follow up was too short to show
differences that could became important with long term treatments (ALLHAT
has one of the longest follow up among hypertension trials)”.
What I wrote: “the diuretic used in ALLHAT study caused negative
metabolic effects: the increase of incidence of diabetes and of
cholesterol levels represents a serious problem. This problem can not be
solved by saying, as the authors of ALLHAT did, that in the 2-4 years of
follow up these metabolic differences did not cause any increase in
cardiovascular events or in all cause mortality”. I never questioned the
duration of the ALLHAT trial but I only discussed the fact that metabolic
changes started in the course of the trial may causes detectable
cardiovascular events before the conclusion of the trial.
In the editorial: “in conclusions where the article says that
ALLHAT’s results should not be applied to Italian or European clinical
practice, given the countries’ low percentage of black patients (who
benefit most from a thiazide diuretic treatment). Interestingly in ALLHAT
main report a sub group analysis indicated that results did not change
according to race”.
For the first time they did not misquote me but challenged my
interpretation of data.
But if one examines the subgroup analysis reported in fig.6 page 2993 of
the ALLHAT’s main report it can observe that the higher incidence (about +
40%) of stroke in the group receiving ACEi is in black patients. No
difference between diuretic and chlortalidone could be detected in non
Since the population(s), like intervention (s) and primary end point
represent the basis on which one has to measure the results of a trial,
I am still convinced that what I concluded does correspond to the results.
It is even a paradox that expressing different opinions or drawing
different conclusions on a clinical trial (an essential part of any
scientific evaluation) could represent a definite demonstration of a
partecipation to “a predefined strategy aimed at challenging unwelcome
And it is absolutely unacceptable that in order to charge other opinion
leaders with these accusations the authors of the editorial have
misquoted many sentences of my article and have given misleading
interpretations of its content.
1)Ambrosioni E. Valutare l'effectiveness pesando tutti gli
effetti.Troppa distanza dalla pratica clinica. SOLE 24 ORE SANITA'2003 4
2)ALLHAT Officers and Coordinators for the ALLHAT Collaborative Research
Group. Major outcomes in high risk hypertensive patients randomised to
angiotensin-converting enzyme inhibitor or calcium channel blocker vs
diuretic: the antihypertensive and lipid lowering treatment to prevent
heart attack trial (ALLHAT). JAMA 2002;288:2981-97
EA received personal compensation for articles or lectures from following industries during the past 5 years: ABBOTT, ASTRA-ZENECA, AVENTIS PHARMA, BAYER, BOEHRINGER, BRISTOL MYERS SQUIBB, CHIESI FARMACEUTICI, GLAXOSMITHKLINE, GUIDOTTI, LUSOFARMACO, MENARINI, MERCK SHARP & DOHME, NOVARTIS, PFIZER ITALIA, PHARMACIA, RECORDATI, SANOFI-SYNTHELABO, SERVIER ITALIA, SOLVAY, SIGMA-TAU
Competing interests: No competing interests
Romania is a poor country, and here the facts are even worse.
Poor people are recommended to buy expansive drugs, whose benefit is not
formidable in comparison with older and much cheaper ones.
The wealthy pharmaceutical industry is not interested in promoting
the local clinical research – a lot of their money is going here in:
- drug symposiums in luxury locations, which finish always with good
meals; these symposiums may last a few hours or a whole weekend – the firm
pays everything, including accommodation in expansive resorts; at these
symposiums there are the leading specialist in domain who are paid to
conference (“third party” method);
- “trips” for teams of 20-30 leading specialists to congresses in their
specialty, in the neighborhood (Europe) or at greater distance (United
States, Asia, Australia); for remote destinations, the trips are generally
of two weeks;
- sponsorship of local congresses, which are in compensation full of drug
symposiums and where, generally, you are not welcome to say anything
against any of their drugs;
- fourth phase studies (with marketing purpose);
- sometimes, the firm may pay a commission for every prescription of their
And probably some other destinations if we take into account the presence
of a few drugs whose efficacy was never proved, on the list of reimbursed
(And where do the funds of pharmaceutical industry never go? In original
local research! In evidence-based medicine education!)
Up to date medical information is sparse here; the majority of
doctors take it from the pharmaceutical industry brochures or abstracts. A
minority who enjoys the Internet can read BMJ, which is free for anybody,
and the specialty journals of BMJ Group and NEJM, which are free for
undeveloped countries. You cannot find recent journals at University
libraries – the last subscriptions were made a decade ago.
ALLHAT study is a good example for Romania, too. The titles of
articles in non-professional journals and symposiums sponsored by the
producer of amlodipine were: “ALLHAT demonstrated that amlodipine was
efficient in lowering blood pressure”. This was true – but it wasn’t the
point. Diuretics, which are 100 times cheaper, lowered blood pressure as
well and with lesser side effects.
A cardiology “third party” was driven the whole country in the car of the
producer of rofecoxib, in order to demonstrate that this drug didn’t do
any harm to the coronary patients. And this "play" culminated with a
presentation at 2003 Congress of Internal Medicine, where another “third
party” claimed that celecoxib, the other blockbuster among the COX2
inhibitors, might be even protective for the heart.
Sincerely, I am not amazed these things happen here; I am amazed by
the fact that the same happens in the super-civilized world!
For example, the European Respiratory Society still recommends the chronic
inhalatory corticosteroid treatment in COPD, although 3 RCT were negative,
and only one, of uncertain validity, showed a reduction of exacerbations
from 1,2 per year to 0,9 per year.
Competing interests: No competing interests
I agree with the authors that reports from so-called opinion leaders
are overused by the pharmaceutical industry, whereas more relevant and
quality evidence is rarely, if ever, discussed. However, I take issue with
the seeming dismissal of some of the criticism levelled at ALLHAT as being
biased and thus less than credible.
First, I disagree that the absence of washout was not significant.
Davis et al (BMJ 2002;325:1156) reported a very high prevalence(22.1%) of
left ventricular (LV) systolic dysfunction in a community cohort of high
cardiovascular risk patients, half of whom were asymptomatic. Men were
more likely to have systolic dysfunction. By design, ALLHAT participants
were at high-risk for cardiovascular events, and the prevalence of
systolic dysfunction in ALLHAT was undoubtably important. Abrupt
withdrawal of stable diuretic therapy certainly had the potential to
unmask asymptomatic LV systolic dysfunction in some of these patients. The
early difference in heart failure incidence between the lisinopril and
chlorthalidone groups is entirely consistent with this possibility.
Furthermore, the heart failure curves started merging towards the end of
the trial, suggesting that perhaps longer follow-up would have shown at
least equivalence, if not superiority, of lisinopril in heart failure
Second, the authors suggest that the heart failure end-point was
"totally validated". In ALLHAT, the heart failure diagnosis was left up to
the local investigator. A single sample of 39 heart failure
hospitalizations was reviewed by an ALLHAT subcommittee, who only agreed
with 85% of heart failure diagnoses. In contrast, blinded endpoint
committees found no difference in heart failure incidence in STOP-
Hypertension 2 (Hansson et al, Lancet 1999;354:1751) or in ANBP2 (Wing et
al, NEJM 2003; 348:583). In the latter study, men in particular appeared
to benefit more from ACE inhibitors than diuretics, perhaps a reflection
of the findings by Davis et al.
Third, the authors suggest that in ALLHAT "results did not change
according to race". In the ALLHAT report (JAMA 2002; 288:2981), there do
appear to be race-based differences between the lisinopril and
chlorthalidone group with respect to the risks of stroke, and possibly
combined coronary heart disease, combined cardiovascular disease, and
indeed heart failure. Unfortunately, the statistical significance of these
apparent interactions is not provided.
While the tone of the entire ALLHAT debate has been somewhat
strident, the arguments for and against the study results should not be
dismissed lightly, for we may neglect to pursue potentially important
racial and gender differences.
George A. Heckman MD FRCPC
Heart and Stroke Fellow
Geriatrics and Internal Medicine
As a geriatric medicine resident, I received funding from Astra Zeneca to attend the 1999 AHA meeting in Atlanta.
Competing interests: No competing interests