I say — (almost) all septic patients need surgery, ventilatory and
inotropic support, volume loading, and antibiotics. Accept it. Rivers
data suggest that EARLY circulatory resuscitation is probably good for
you. So, I teach the residents to ventilate septic patients early,
accept that you will porbably need 40-60% oxygen and 5-12 PEEP during
the peak of the septic illness, start inotropic support at once, aim for
normotension, get good stable monitoring of MAP by art.line, HR from
ECG, SPO2 from a finger and look at the ventilator peak pressures and
the tidal volumes (we usually use pressure-regulated volume control on
the Servo300) and then assess the ongoing effects of aliqouts of volume
on MAP, HR, SpO2 signal/noise ratio (pulse height), respiratory “swing”
of MAP, urine volumes and peak airway pressures. Wean inotrope
(noradrenaline) as tolerated. Accept that most septic patients need
0.5-2 mg/hr of noradrenaline. Just get over it. Accept that some
patients will need drainage of pleural fluid collections (not many). Be
aware of intra-abdominal hypertension and watch out for it. I have
attached a snippet of my lecture on sepsis to our residents which talks
about the extremely practical details. I know that some of this will
fill “toyboys” (and girls) who need PA caths, TOEs, PiCCOs and Svo2
before making decisions with horror. However, as I keep saying, your
mileage may differ — mine certainly does.
I take issue with your emphasis about sepsis and renal failure — Most
patients in our unit do not need dialysis — we take the view that
“renal failure is a capital offence” and try and reverse it aggressively
at the beginning. However, if polyuria does not develop and oliguria
supervenes we don’t try extraordinary measures (“saltwater drowning”,
mannitol, frusemide, rain-dances etc) to get urine out of rocks. We just
Sometimes in extremely co-morbid patients with sepsis the development of
dialysis-dependent renal failure is an end-point (i.e. we would have
decided that we would not dialyse these patients) but that is not very
common. Here are some data I presented in Mumbai early this year by way
of illustration — in 2002 we admitted 145 patients with severe sepsis,
median age 60, 80% had co-morbidity, 26% had “serious co-morbidity”
(Prior Health Status C or D — Knaus et al APACHE-I), 16/145 (11%) had
therapies withdrawn, 6/16 had been “admitted with reservations”, 14/16
had prior “intention to withhold”, 11/16 had had therapies withheld
before being withdrawn and 16/16 died in ICU 13 – 431, median 66 hours
after admission. Another 9 patients died without withdrawal of therapy
— total 25/145 deaths or 18%. Only 12 of the 145 septic patients were
dialysed (10 of whom survived).
The recent (rather methodologically poor) studies of fluid restriction
are to my mind largely showing that unncessary “salt-water drowning”
perioperatively (including in the five days post-operatively) is bad for
you (surprise). Since we don’t give resuscitation fluids (Na+140-154/l)
as maintenance fluids anyway (we give 1ml/kg/hr of D5W) then we were not
altered in our practice by the Ann Surg Nov 2003 trial. We did
participate (very heavy recruiting centre — 439 patients randomised) in
the SAFE trial and that involved a lot of SAFE fluid (especially the
“thin fluid” and especially in septic patients). As your can see from
our mortality data on the previous jpg — survival was the same (~82%)
during the SAFE trial as before and after it.
I loved your comment about “advanced care” — it is very American. 😉
Joan Cassell made the same comment when she came to work with us. I told
her what I have said for many years — it is not technology that makes
intensive care work — it is people who intensively care — and she
liked it and included it in the paper. (Cassell J, Buchman TG, Streat S,
Stewart RM. Surgeons, Intensivists and the covenant of care –
administrative models and values affecting care at the end of life –
Updated. Crit Care Med 2003;31(5):1551-1559). Let me be blunt — IMAO
much of the invasive technotoys have been foisted unproven (they do
measure something — I’m not disputing that, however — does it make a
difference — that’s not proven) on a gullible and easily seduced
population of technoleaning ‘crats (no offence, really no offence) and
have been slickly marketed as the “best thing since sliced bread which
everyone else using and which you simply must have in order to stay up
with the pack”. No $600 toilet seats here in NZ. However, we can and do
use a lot of high-tech at times (but very selectively, accepting that
the marginal benefits of it are small and will be confined to a small
number of patients). We just don’t think that everyone needs a 30 foot
gold plated cadillac limo to get them down to the corner store for a
pepsi. Sometimes, it would be perfectly good to walk. I know that NASA
is still smarting from that awful “metric versus imperial” (American
empire that is) screwup on Mars a few years back but I agree with their
faster, cheaper, better idea. Faster and cheaper can also mean “with
better outcomes” in health care — it just takes application and
innovation — and — and this is utterly crucial for the US of A who
utterly refuse to face this — SELECTIVITY. This includes another one of
those awful commie words Leo — RATIONING. Now I know that you (and many
others) lament the waste of your time and everyones money that goes into
non-selective ICU practice in many parts of the USA. However, this is
what truly must be faced up to sooner (or in the standard US manner —
later — much later — as later as you and your politicians can get away
with). It is just silly to plumb every dying patient with every
high-tech toy and probe and line and pump and just continue to accept
this as the standard modus operandi. You know it Leo from your own hard
working experience. You gotta go easy on this techie stuff and spend a
great deal more time dealing with the “soft stuff” like setting limits
on treatments and monitoring modes and toys and fancy pharmaceuticals.
Easy for me to say. 😉
Enough non-EBM opinion from me — Gord will be stressing the walls of
his cerebral vasculature. However, some of this stuff really SHOULD be
subjected to the kind of (difficult) studies that PACMAN and our own
experience leads you to think about.
Finally, lactate is a useful marker of badness — no question — (there
is even quite good evidence for this) and rising or falling lactate
during best clinical treatment gives a warning or a promise of things to
come. However, there are few occasions when the course of clinical
treatment with respect to the nature of CVS support is altered by
knowing a lactate — or even serial lactates. We (that’s you too Leo I
bet) try and resuscitate the patient as best we can and we watch the
lactate. If it rises we worry (and we make sure that we had in fact done
everything we should have) but we don’t usually change treatment.
Similarly, if it falls we feel pleased with ourselevs and maybe we relax
a bit — but we don’t withdraw all the CVS support we have put in place
which seems to be a “good thing”.
Immediately on admission ensure reliable standard monitoring (ECG, SpO2, MAP transduced from arterial line), place at least one large peripheral line for colloid and at least a quadruple lumen central line. Some patients need six central line lumens.
Use standard curarisation/sedation regimen Pancuronium/Morphine infusions, diazepam 5-10mg 6-12 hourly. Front-load the sedation perhaps 20-40 mg diazepam (before complete curarisation if possible) to ensure no awareness under curarisation. Use Pressure Regulated Volume Control (PRVC) on Servo 300) to start with and continue this mode if respiratory failure is severe (FIO2>0.6 or PEEP>10). Start with 100% oxygen and 5 cm PEEP. Set VT to 10 12 ml/kg, adjust rate (10 14) for normocarbia (35-45 mm Hg), and increase PEEP if necessary to control frothing or for unsatisfactory SpO2 despite high FIO2 (0.6 or more). Adjust FIO2 and PEEP to obtain SpO2 95-98%. Keep peak airway pressure under 30 cm H2O Use standard I:E ratio (30%) for Servo 300 ventilator (see Servo Ventilation Charts) unless severe ARDS. Increased I:E ratio (e.g. 1:1) ventilation may help oxygenation but has risks, call Intensivist and discuss if you think this necessary. Be aware of PEEP effects on cardiac function, renal and hepatic blood flow and pulmonary (hyper)inflation..
Find out if the patient responds to volume give a rapid colloid infusion (500ml over 5 10 minutes) and watch the patient closely! Look at peripheral temperature, urine output, heart rate on ECG, transduced MAP, peak airway pressures on the ventilator and continuous SpO2 (having set FIO2 and PEEP so that SpO2 is about 95%, not 100%) and answer two questions: i) Were there favourable changes in haemodynamics and clinical effects in keeping with a rise in cardiac output and oxygen delivery? (E.g. falling HR, rising MAP, oximeter begins to work, rise in SpO2, rise in evident peripheral warmth or colour, increase in urine volume). ii) Were there adverse effects on pulmonary function? (E.g. rising airway pressure, fall in oxygen saturation, frothing). If volume is beneficial and further benefit is desirable on clinical grounds then repeat the volume challenge. At some stage it will be clear that further volume merely worsens oxygenation and lung compliance. Many patients may need pleural fluid drainage at some stage and this may reveal further volume responsiveness. Lusitropic drugs (E.g. beta agonists, milrinone) and vasodilators (E.g. beta-agonists, milrinone, alpha-blockers) may need further volume loading for good effect.
This is near-universal in severe sepsis1 and as such when used intelligently with DCCM standard monitoring most patients do not need pulmonary artery catheterisation. Aim to keep MAP 90 110 and HR 90 120 in SR for optimal cardiac function. Combinations of inotropes are often used depending on haemodynamics so be prepared to change as the patients circulation changes. Start with noradrenaline first (>80% of patients with severe sepsis have noradrenaline), increase to 1 2 mg/hr if required for MAP. Remember NA is a mixed a,b-agonist. Some patients may benefit from the beta effect of additional dopamine with an increase in rate (and sometimes contractility) but beware AF or SVT are common. Renal dopamine in sepsis is a myth5. Some patients require high doses of NA (5 10 mg/hr) and some of these may benefit from vasopressin or terlipressin. Dobutamine is used rarely (usually with noradrenaline) when BP will stand it and may help to increase output by chrono/lusitropic effect in selected patients. Rare patients may benefit from adrenaline or milrinone. Give all patients digoxin after K+ repletion (for both inotropy and rhythm control) unless conduction disturbances contraindicate it. Be very cautious with amiodarone (heart block, vasodilatation, inotropic depression) in septic patients with abnormal hearts.
Monitoring of CVS function
Clinical assessment is most important. No CVP measures in DCCM patients except in transplant protocols or rarely at specialist instruction (e.g. caval surgery). Pulmonary artery catheterisation may occur rarely (DCCM specialist decision) when diagnosis is obscure.
No specific therapy if pHa>7.2 and circulation is stable and improving use serial pHa as sign of progress. If pHa is <7.2 and circulation is failing give THAM (1 2 mmol/kg via CVL over 20 minutes) and consider continuous THAM infusion 0.25 2 mmol/kg/hr when maximum CVS support will not keep pHa >7.2.
1. Thomas MG, Streat SJ. Infections in Intensive Care Patients. Chapter 44 (pp 564-576) in Antibiotic and Chemotherapy. 8th Edition. Eds Finch R,Greenwood D, Norrby R, Whitley R. Churchill Livingstone. London 2003
2. Bone, R.C., Balk, R.A., Cerra, F.B., Dellinger, R.P., Fein, A.M., Knaus, W.A., Schein, R.M., Sibbald, W.J.: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee. Chest 101:1644, 1992
3. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, Peterson E, Tomlanovich M. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med. 2001 Nov 8;345(19):1368-77.
4. Clark, M.A., Plank, L.D., Connolly, A.B., Streat, S.J., Hill, A.A., Gupta, R., Monk, D.N., Shenkin, A., Hill, G.L.: Effect of a chimeric antibody to tumor necrosis factor–a on cytokine and physiologic responses in patients with severe sepsis – a randomized clinical trial. Crit. Care Med. 26:1650, 1998
5. Bellomo R, Chapman M, Finfer S, Hickling K, Myburgh J. Low-dose dopamine in patients with early renal dysfunction: a placebo-controlled randomised trial. Australian and New Zealand Intensive Care Society (ANZICS) Clinical Trials Group. Lancet. 2000; 356(9248): 2139-43.
6. Nasraway SA, Rackow EC, Astiz ME, Karras G, Weil MH. Inotropic response to digoxin and dopamine in patients with severe sepsis, cardiac failure, and systemic hypoperfusion. Chest. 1989 Mar;95(3):612-5.
7. Streat SJ, Plank LD, Hill GL. An overview of modern management of patients with critical injury and severe sepsis. World J Surg 2000 Jun;24(6):655-663
8. Streat SJ. Abdominal surgical catastrophes. In Ohs Intensive Care Manual 5th Edition. Eds Soni N, Bersten A. Butterworth-Heinemann, 2003
Here’s another approach.
Unfortunately 50% of critically ill patients do not respond to fluid loading
by increasing their cardiac output (CO) and oxygen delivery. Loading these
patients aggressively is therefore harmful as their lungs, gut and all other
organs/tissues will just get more edematous. One has therefore to use those
parameters that PREDICT the response to fluid loading before taking the
decision to actually do it. The CVP and PCWP have been shown again and again
and again to be poor predictors of fluid responsiveness. Volumetric
parameters, like the EDA (TEE) and the ITBV are somewhat better. Functional
hemodynamic parameters like the Systolic Pressure Variation, Pulse Pressure
Variation and Stroke Volume Variation are the best predictors, but only in
patients on fully controlled mechanical ventilation. The least one should do
in these patients is monitor their CO, or its surrogate like the ScvO2. The measurement of CO is important since it will tell you, first and foremost, if the CO is low(!), as well as if, and by how much, does the CO respond to fluid loading. However, the CO value by itself, like the ScvO2 or SvO2 values, is not enough, since you still would not know why is it low. Many septic patients have significant myocardial depression, with or without hypovolemia and with or without hypotension. The Rivers data are impressive, but the success was achieved because the protocol was applied in the very early sepsis phase only, a phase in which most patients may indeed be hypovolemic. In later sepsis a more advanced hemodynamic monitoring is necessary in order to more accurately assess the cardiovascular status and prevent unnecessary and potentially harmful fluid loading.
I think that it is somewhat more than that. We accept that there is/will be a certain level of inotropic support (we use noradrenaline) required in patients with sepsis (that is mainly what we are talking about here — other general ICU patietns can be managed even easier). We establish what that “minimum dose” is, early in the course of the treatment, during intensive volume challenges (“Is the haematocrit down to 0.28 or so”?) and then give further volume, including sometimes as an infusion, according to the clinical circumstances — knowing some of the obligatory fluid losses. For example — if the patient has faecal peritonitis (our most common cause of severe sepsis, not community-acquired pneumonia which is second most common), then we understand that there will be “weeping” from the inflammed peritoneum — remember 1mm of exudate across the peritoneal surface equals about a litre of fluid. Similarly, there may well be increased luminal fluid — some of which will come up the NG tube and some of which will remain intraluminal. There may also be plasma volume loss into the interstitium — we accept that all of these things may amount to several hundred mls per hour for 12-24 hours, depending on the surgical situation. We are happy to give this and watch all the parameters I referred to before. At some stage (6-12 hours, 12-24 hours ..) we stop the volume and watch. We define the end of the resuscitation period as ‘haemodynamic stability’ as having been reached without a need for either escalating inotrope or further volume load. Usually this point is reached within 12 hours, sometimes 24. All this stuff requires a fair bit of ‘close clinical observation of the patient’. Our nurse:patient ratios in NZ and Australia for critically ill patients (such as severe sepsis) are 1:1 and this is fantastic. We watch the patients like hawks and teach the registrars (senior residents – we don’t have junior residents making decisions in ICU) to do likewise. I know this stuff might seem strange — but it works in our hands. Remember we put in hundreds of PA caths over 25 years and found over and over that they helped us less and less. Outcomes continue to improve at the same rate during their disappearance and afterward …