Randomised controlled trial assessing the impact of a nurse delivered, flow monitored protocol for optimisation of circulatory status after cardiac surgery
BMJ 2004; 329 doi: https://doi.org/10.1136/bmj.38156.767118.7C (Published 29 July 2004) Cite this as: BMJ 2004;329:258
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.
Sir
I was most interested to read the article by McKendry and colleagues
suggesting that optimizing stroke volume in the first four hours after
cardiac surgery lead to an 18% saving in hospital bed days (1). It is
unique that a single intervention lasting only four hours has such a
dramatic effect on hospital bed occupation, which probably reflects why
the BMJ gave their paper front page status. I however have some
reservations about the study.
The choice of hospital stay as the primary outcome is debatable.
Length of stay as an outcome after cardiac surgery is at best a surrogate
measure which does not necessarily bear relationship to success of
surgery. Factors unrelated to outcome also influence hospital stay
including physician preference and local practice. Indeed the hospital
stay the authors report is longer than experienced in most UK cardiac
units (the National registry reports 80% of patients were discharged
within 10 days of surgery in 2000 and 2001 (2)). Although they list
complications as a secondary outcome, some bear no conceivable association
to their intervention (such as aortic regurgitation) and some relevant
outcomes, such as low cardiac output, are not presented. The accuracy of
their complication data is also questionable. The 6% incidence of atrial
fibrillation in the protocol group is well below that experienced in
cardiac surgical practice (where up to a third of operations are
complicated by atrial fibrillation). No intervention introduced to reduce
atrial fibrillation after cardiac surgery has achieved such low incidence
of atrial fibrillation (3). The low incidence may represent under-
reporting or biased reporting of complication data.
Blanket application of any haemodynamic intervention to all cardiac
surgical patients is not practical. The pathophysiology, and hence the
approach to post-operative haemodynamic care, differs depending on
pathology, cardiac function, and operative procedure (transplantation,
coronary bypass, valvar stenosis, valvar regurgitation, congenital etc).
Even with a recording of low stroke volume, fluid administration may not
necessarily be the appropriate intervention. Indeed in certain post-
operative conditions, such as cardiac tamponade, fluid administration is
potentially harmful. Fluid administration to an already distended and
failing ventricle (which will have a low stroke volume) can be
catastrophic. The authors have not discussed problems that may arise from
errors in stroke volume measurement or interpretation. Although the
oesophageal doppler is well validated, it is subject to measurement and
interpretation errors which, although minimal in the specialist and
research setting, will be expected to increase as usage shifts from
specialist and research users to the routine clinical setting. It would
therefore be ill advised, as the paper seems to suggest, to direct
clinical management based on stroke volume alone. Management decisions
should depend on summation of pre-operative and operative history,
clinical examination, various haemodynamic measures (including cardiac
output), urine output and metabolic data.
The authors have given no biologically plausible explanations as to
how their intervention reduces hospital stay. The assumption that this is
due to fewer complications is not borne by their data. The authors did
not record any complication for majority (70%) of patients in the control
group, yet half still spent 9 or more days in hospital; the occurrence of
complications therefore did not correlate to hospital stay. It is unclear
why they restricted the intervention to 4 hours as for most patients the
period of critical haemodynamic management lasts for at least the first 12
hours; the protocol group would therefore have reverted to standard care
for the rest of this critical period. The authors did not offer any
explanation as to why the first 4 hours alone would have bearing on the
rest of the post-operative period independent of the subsequent
management.
Two final issues deserve mention – acceptability by surgeons and
safety. The authors recruited 5 patients a month in an institution that
performed over 50 operations a month in the same period (2). Given their
liberal inclusion criteria, only a minority of eligible patients was
recruited. It is not stated what factors limited recruitment. There is no
statement on the acceptance of their intervention by local surgeons and
intensivists; indeed none of their surgeons have been acknowledged in the
study. Additionally, despite a higher death rate in the intervention
group, the authors do not discuss the possibility that their intervention
could cause harm. They dismissed the four deaths in the intervention group
as not directly attributable to the early post-operative care, yet all the
patients sustained major haemodynamic complications, and in at least 3,
the deaths were related to low-cardiac output states.
Without further study, one cannot be certain that their intervention
is indeed beneficial or that it can be extrapolated to the general cardiac
surgical community.
References
1) McKendry M, McGloin H, Saberi D, et al. Randomised controlled
trial assessing the impact of a nurse delivered, flow monitored protocol
for optimisation of circulatory status after cardiac surgery. BMJ
2004;329:258-61.
2) The Society of Cardiothoracic Surgeons of Great Britain and
Ireland. National Adult Cardiac Database Report 2000-2001. Henley-on-
Thames: Dendrite, 2003
3) Crystal E, Connolly SJ, Sleik K, Ginger TJ, Yusuf S.
Interventions on prevention of postoperative atrial fibrillation in
patients undergoing heart surgery: a meta-analysis. Circulation.
2002;106:75-80.
Competing interests:
None declared
Competing interests: No competing interests
We express concern at the conclusions of this study.
We contest, firstly, that this is an invasive intervention with
little demonstrable impact on the primary outcomes cited in the study i.e.
intensive care stay and hospital stay. In fact, both of these figures
exceed the national averages. Average length of stay for all cardiac
surgical patients admitted to the Western Infirmary, Glasgow for example
was 7.2 days for 2004 and has varied little since 1994. A more pertinent
question might be to ask why length of stay is so prolonged in this
group’s experience.
Secondly, important and irrefutable outcomes such as death were
casually dismissed by the authors despite the fact that mortality was
higher in the protocol group than the control group (4 vs. 2). Indeed, one
is forced to go to the web to ascertain the precise circumstances of
death. In that regard, we believe that the term ‘optimisation of
circulatory status’ used throughout the study is presumptive especially
given that multiorgan failure, low cardiac output and cardiac arrest were
cited as causes of death in the protocol patients and that these patients
all arrived in good condition in their intensive care unit (see exclusion
criteria). In the cited morbidities, chest infection and wound infection
were grouped together despite their separate aetiologies and the main
reduction was in the incidence of AF which is not easily explained by the
study and may be a spurious result.
Finally, only 179 patients were recruited from this institution into
the study over almost three years. We are, therefore, drawn to conclude
that this study is representative of only a small and highly selected
group of patients and that its results cannot be widely applicable.
In conclusion, we purport that this study adds nothing to the
scientific literature and that the technique has little to offer patients
undergoing cardiac surgery.
Competing interests:
None declared
Competing interests: No competing interests
McKendry et al have made a valuable attempt to tackle the difficult
problem of post-operative circulatory status. (1) They claim that outcomes
after cardiac surgery may be improved by a nurse-delivered protocol
optimising stroke volume index in the immediate post-operative period.
However, we have some queries regarding the study.
The study was powered to detect a reduction of 3 days in mean
hospital stay. Even by the standards of April 2000 when the study began
recruitment, their mean stay of 13 days seems long. Other UK centres had
reported average hospital stays of 8 days back in 1995. (2) Even in this
study which excluded any unstable patients the mean length of stay was
over 11 days. Why is the average length of stay so long in UCL?
Despite the input of the study nurses, nearly 40% of protocol
patients failed to meet the stroke volume index target of 35ml/m2, and
none of them received inotropes as per protocol. This latter was due to
time-constraints and reticence of nurses. What hope then for the control
patients if even the protocol patients failed to receive prescribed
treatment, despite the extra input of 4 research sisters.
The main finding of this study was the difficulty ensuring adequate
resuscitation post-operatively. The authors state that standard monitoring
consists of arterial and central venous pressures, urine output and
arterial base deficit. None of these data are presented. Could the control
patients have been underfilled by these usual criteria yet not acted on,
again due to time-constraints and reticence?
It would be interesting to know if blood lactate concentrations were
available for all control and protocol patients. Most blood gas analysers
now routinely measure lactate. Blood lactate concentration was a target of
Polonen’s study of goal directed therapy (3), and hyperlactataemia can
identify those at higher risk of complications and mortality. (4) Perhaps
this measure allied to the standard monitoring could help guide and
optimise circulatory status as effectively as a doppler probe.
Finally a criticism aimed squarely at the BMJ. For a journal that
trumpets ethical research so loudly, it publishes a print version of this
study with not a single mention of informed consent from the study
participants. Surely this is an important aspect of any study and should
be included even in abridged versions.
1. McKendry M, McGloin H, Saberi D, et al. Randomised controlled
trial assessing the impact of a nurse delivered, flow monitored protocol
for optimisation of circulatory status after cardiac surgery. BMJ
2004;329:258-61.
2. Birdi I, Izzat MB, Angelini GD, Bryan AJ. In-hospital audit
underestimates early postoperative morbidity after cardiac surgery. Br
Heart J 1995;74(5):556-8.
3. Polonen P, Ruokonen E, Hippelainem M, Poyhonen M, Takala J. A
prospective randomised study of goal oriented hemodynamic therapy in
cardiac surgical patients. Anaesth Analg 2000;90:1052-9.
4. Maillet JM, Le Besnerais P, Cantoni M, et al. Frequency, risk factors,
and outcome of hyperlactatemia after cardiac surgery. Chest.
2003;123(5):1361-6.
Competing interests:
None declared
Competing interests: No competing interests
McKendry et al raise a number of important points in their study,
which need expanding or clarifying.
They mention that base excess, urine output central and arterial pressures
are the corner stones of ITU management in cardiac surgery, and this is
true. However they fail to mention arterial line swing, the importance of
a negative fluid balance, and the central and peripheral temperature
profile of the patient. These are all important pieces of clinical
information that are evaluated by the on call cardiac surgical registrar
in ITU, differentiating them from someone who is just “chart or guideline
trained”.
The study purely demonstrated that more fluid is required than the usual
ITU nurse care provides. This is evident on a daily basis to anyone who
works in the ITU post cardiac surgery. It is common to find someone who
has a CVP of +2cm, no base excess and adequate urine output with a
negative balance of –500mL, NOT being given fluid by the nursing staff
“because they are alright on the chart”. Practical experience shows they
need more fluid.
Cardiac tamponade is an easily missed complication of cardiac surgery. It
does not always present in a classical way, and a normal CVP and adequate
blood pressure on inotropes is not an uncommon finding. The flow chart
presented will result in a delayed or a missed diagnosis.
Excessive bleeding, and high dose inotrope usage is relatively common post
cardiac surgery, so any trial that eliminates these patients has to be
treated with caution as the results are unlikely to be applicable to the
real clinical situation.
Despite all patients being done on bypass, the lumping together of valves
and / or coronarys, redo surgery and “other” reduces the power of the
study considerably as they involve completely different risk profiles,
hence the changing study groups in the cardiothoracic publications the
Annals of Thoracic Surgery and The Journal of Cardiothoracic Surgery.
With regard to risk profiles the use of Parsonnet scoring has been proven
on numerous occasions, including the Society of Cardiothoracic Surgeons of
England and Ireland, to be an inadequate way of risk assessing and
comparing patients pre cardiac surgery, hence the evolution of the
Euroscore additive and logistic.
The duration of the study was 3 years, meaning about one case per week for
the study at maximum. Did this mean only one surgeon was involved? This
may have important implications as different surgeons use different bypass
temperatures and methods of cardioplegia, which can all affect
postoperative care.
Liverpool is one of the largest cardiac surgical units in the UK. Our
average ITU stay and hospital stay are significantly shorter than those in
the study. The vast majority of our patients leave the ITU on day 2 and go
home by day 8, so it is difficult to understand how their average stay was
so high, nearly 1.5 times as much.
In summary any competent cardiac surgical registrar, who are cognisant
with oesophageal doppler, who have daily exposure to ITU post cardiac
surgery will already knows that clinical examination of the patient and
assessment of arterial line swing, fluid balance, and the central and
peripheral temperature profile of the patient are far more helpful than
blind flow chart driven oesophageal doppler probe use. The importance of
an intelligent, clinically alert ITU nurse with previous experience should
also not be underestimated.
1. Moira McKendry, Helen McGloin, Debbie Saberi, Libby Caudwell,
Anthony R Brady, Mervyn Singer. Randomised controlled trial assessing the
impact of a nurse delivered, flow monitored protocol for optimisation of
circulatory status after cardiac surgery. BMJ 2004;329:258
Competing interests:
None declared
Competing interests: No competing interests
This study (1)is very similar to that performed in the same
institution by anaesthetists (2) but the nurses' results are not nearly as
good as the anaesthetists. One reason for this apparent difference might
be that anaesthetists in the institution have learned from the first study
and are more attentive to volume resuscitation and that operative
mortality in controls was lower in this study and the room for improvement
less. The more likely reason would seem to be that current management is
conceptually flawed and will continue to be associated with unexpected
outcomes until thinking is changed.
The nurses protocol did not improve outcome when measured in terms of
complications and deaths. Twenty six controls, of the 174 patients having
cardiac surgery studied by the nurses, had postoperative complications
(two deaths) compared with 17 (four deaths) protocol patients (P = 0.08).
In reality the mortality doubled with the nurses' interventions but the
difference was not significant. If, however, a Fisher's exact test were
done comparing the reslts of the two studies I suspect the difference in
mortality might be highly significant.
Every death and most complications in the present study as in the
past study may be attributed to schaemia/reperfusion injury of gut mucosa
but as gastric intramucosal pH was not measured that cannot be established
as in the anaesthetists study. It is an important omission for in its
absence management cannot be rational. Indeed even when anaesthetists were
optimising stroke volume according to protocol an intramucosal acidosis
persisted in some 12% of the patients. Deaths and complications will
continue to occur unpredictably after all forms of surgery until this
residual is eliminated by managing tissue energetics intelligently and
effectively. And don't expect any miracles from "real-time feedback-
controlled ultraprecise resuscitation", be it driven by intramucosal pCO2
or perfusion measured with a laser Doppler imager, or from nuclear
magnetic resonance spectroscopy (3).
The issues are complex (4) and even anaesthetists do not appear to
appreciate just how complex they are. Anaesthetists are, however, far
better qualified to appreciate, identify, address and resolve them than
nurses. As good as some nurse anaesthetists undoubtedly are what medical
professional would ever agree to be a subject in a prospective study
comparing a anesthetist's performance with a nurse's? Having said that
making anaesthesia more of a science than the art it is might make it
possible for nurse anaethetists to compete much more favourably with
anaesthetists.
1. Moira McKendry, Helen McGloin, Debbie Saberi, Libby Caudwell,
Anthony R Brady, and Mervyn Singer
Randomised controlled trial assessing the impact of a nurse delivered,
flow monitored protocol for optimisation of circulatory status after
cardiac surgery
BMJ 2004; 329: 258-0
2. Mythen MG, Webb AR. Perioperative plasma volume expansion reduces
the incidence of gut mucosal hypoperfusion during cardiac surgery.
Arch Surg. 1995 Apr;130(4):423-9.
3. Real-time feedback-controlled ultraprecise resuscitation
Richard G Fiddian-Green
bmj.com, 27 Jul 2004 Rapid response to: Mark Ansermino and Carolyn Hemsley
Intensive care management and control of infection
BMJ 2004; 329: 220-223
4. Fiddian-Green RG. eLetters re: David Roy Dantzker
Monitoring Tissue Oxygenation : The Quest Continues
Chest 2001; 120: 701-702
Competing interests:
None declared
Competing interests: No competing interests
Senior author's reply
Having just read the five rapid responses posted, I feel that some
comment is needed to correct some of the misinterpretations
stated.
Fiddian-Green makes unnecessarily demeaning remarks about
nurses when he should perhaps direct his ire elsewhere. The
previous study he cited from our institution was performed INTRA-
operatively and the approximate 40% reduction in hospital stay
achieved in cardiac surgical patients through intravascular volume
optimization was indeed impressive. Notwithstanding this result,
any such protocolised haemodynamic optimisation has failed to
become a standard of care and a significant morbidity persists.
Thus the present study was designed to assess whether a belated
POST-operative attempt to 'catch up' with the tissue oxygen debt
incurred from inadequate intraoperative resuscitation could still
have some impact in improving outcomes. The answer was yes,
though was only half as effective as the similar fluid-loading
protocol followed intraoperatively. The critical care nurses (not
nurse anaesthetists as erroneously stated) were simply attempting
to redress the damage induced by a synergistic combination of
cytokine-generating trauma and tissue hypoxia. To draw
comparison with 'skilled anaesthetists' is thus spurious in the
extreme. However, we are in full agreement over the need to
optimise tissue energetics - alas, the problem is that no
appropriate monitoring device, including the one Fiddian-Green
himself pioneered (gastric tonometry), has become a standard of
care. The anaesthetic and surgical communities need to ask
themselves why not.
Poullis places the on-call cardiac surgical registrar on a lofty
pedestal which is certainly not borne out in my experience. If only
they were all so haemodynamically astute! It is easy to throw barbs
at the ICU nurse but surely postoperative care mandates a multi-
disciplinary team approach. If it is so apparent that "more fluid is
required than the usual ITU nurse care provides. This is evident on
a daily basis to anyone who works in the ITU post cardiac surgery"
then why has he and his ilk not sought to improve this suboptimal
care being provided to his patients, be it through education or
positioning themselves at the bedside and providing hands-on
care? Our study does indeed highlight the imperfections in current
management and shows that "someone who is just chart or
guideline trained” can indeed improve upon standard care being
offered by surgeons, anaesthetists, intensivists .. and nurses.
Incidentally, our need for high-dose inotropes either intra- or post-
cardiac surgery is uncommon - thankfully, this is one injurious and
misbegotten 'therapy' that is generally avoided in our institution. At
the time the protocol was designed, Parsonnet was the 'in' score
but we accept that it has been superceded by other scores. We
recruited cases performed by four surgeons and the slow
recruitment was due in part to other competing studies and in part
to pregnancy 'affecting' three of the four study nurses who worked
consecutively rather than together. With regard to bed stay, I would
point out the disproportionately high proportion of valve
procedures among our study group.
In answer to Campbell et al, please see the above comment on
length of stay. I would suggest they audit the performance in their
own postoperative unit to see what amounts of fluid are given in
the first four postoperative hours -it may be revealing! Word count
restrictions prevented publication of all the haemodynamic data
but these can be supplied on request. Standard triggers for fluid
administration such as low blood pressure, low urine output,
increasing base deficit etc.. were routinely observed and acted
upon in the control patients. The use of blood lactate levels will be
confounded by any concurrent use of lactate-containing solutions
such as Hartmann's and the accelerated glycolytic effect seen with
adrenaline. I am also unaware of studies showing clear superiority
of lactate over base deficit in prognosticating outcomes.
Sutherland also queries the length of stay and slow recruitment -
again, please refer to above. We did not 'casually dismiss' the
deaths - indeed, their management was carefully reviewed and
they did not receive large amounts of fluid (or inotropes) in the first
4 hours to provide a cause attributable to the protocol for their
subsequent demise. Chest and wound infection were grouped
together - a strong argument could be made for their similar
aetiology - namely, poor tissue perfusion leading to poor wound
healing, increased lethargy, reduced mobilisation, weakened
cough and chest infection. It is also a well known fact that one of
the commonest causes of post-operative atrial fibrillation is
hypovolaemia, particular when inotropes are being given - thus
the improvement in the protocol group is expected rather than
spurious. They conclude that our study adds nothing to the
scientific literature and the technique has little to offer patients
undergoing cardiac surgery. They are obviously entitled to their
opinion. However, there are now two orthopaedic studies, two
abdominal surgery studies and one other cardiac study all
showing outcome improvements with similar protocols using the
same oesophageal Doppler device. There's only so much sand
one can stick one's head in!
Dr Anyanwu should also refer to the host of peroperative studies
using Doppler, PA catheters and the like whereby short-term
protocolised intervention lasting hours made a huge difference to
subsequent outcomes. With regard to hospital stay, we would
argue that it is an irrefutable endpoint and the randomisation
technique takes into account local practice, surgeon preference,
etc.. The low rate of AF is fully commensurate with adequate fluid
loading and avoidance of inotropes - I would commend this
approach to him. He suggests that blanket application of any
haemodynamic intervention to all cardiac surgical patients is
impractical - I disagree entirely. The principles of post-op care are
identical, namely pain relief, adequate gas exchange, adequate
filling and maintenance of an oxygen delivery sufficent to maintain
tissue perfusion. This obviously requires tailoring to the individual
but the principles remain. The use of a Starling fluid challenge
avoids excessive fluid loading and aids detection of conditions
such as tamponade before catastrophes ensue. I fully agree that
correct use of any device is mandatory - a good education
program with regular refreshing is essential. The biologically
plausible explanation for the improvement seen is the reduction in
tissue hypoperfusion - this translates into less lethargy, less ileus,
and a greater ability and will to mobilise. We did acknowledge the
medical staff (this does include the surgeons!) though not by
name. The technique has been in common use on the ICU for
over 10 years thus general acceptance of its utility is not an issue.
This study was, however, the first by which its use was
protocolised for post-operative management. We too would
encourage further study to confirm or refute the generalisability of
our results.
Competing interests:
senior author of the study
Competing interests: No competing interests