REVIEW
SA JOURNAL OF DIABETES & VASCULAR DISEASE
52
VOLUME 9 NUMBER 2 • JUNE 2012
Correspondence to: Dr Gregory Hough
Specialist physician/endocrinologist/diabetologist
Greenacres Hospital, Greenacres, Port Elizabeth
Tel: +27 0(41) 363-5249
Fax: 086 648 9407
e-mail: drhough@telkomsa.net
S Afr J Diabetes Vasc Dis
2012;
9
: 52–54
T
his is an important but often neglected topic that is frequently
poorly managed. Too often peri-operative care is left in the
hands of nursing staff in the ward, following the anaesthetist’s
peri- and postoperative instructions. These instructions usually take
the form of a crude generic sliding scale adapted to fit comfortably
into the usual six-hourly patient monitoring that occurs in most
wards. This form of care is wrong in every possible way and should
be actively discouraged by everyone involved in patient care, from
the nursing staff to the surgeon and anaesthetist, and even the
attending physician, if involved.
‘Why bother with good glycaemic control; surely a few days of
erratic or high sugar levels cannot do any real harm?’ This seems to
be the pervasive attitude that has crept into much of routine medical
care in the peri-operative period. A similar attitude prevailed, until
a few years ago, with regard to the care needed to prevent venous
thrombo-embolism in the peri-operative period. Through education
and new guidelines, our attitudes and practices to prevent venous
thrombo-embolism have changed dramatically. A similar change
is now needed with regard to what is considered to be routine
practice in the care of diabetes in the peri-operative period.
There is a large body of observational evidence linking in-hospital
hyperglycaemia to poorer outcomes. There is also a growing body
of evidence in the form of cohort studies and even some early
randomised, controlled trials showing that intensive treatment of
hyperglycaemia in hospital improves outcomes.
1,2,3
The most famous study analysing tight glycaemic control in the
peri-operative period was a single-centre study which reported a
42% reduction in ICU mortality.
2
Recent multi-centre studies in
both medical and surgical patients have failed to demonstrate the
same benefits; in fact the outcomes in the intensive groups were
slightly worse.
4
However closer examination of the largest of these
trials reveals that the difference between intensive and standard care
was very small; blood sugar levels of 6.4 versus 8%. These trials do
not suggest that intensive target glucose control is not important or
causes harm but rather that the target chosen for tight control may
have been a little too ambitiously low, with the resultant frequent
hypoglycaemia contributing to the increase in adverse outcomes.
A recent meta-analysis of 26 trials
5
assessed care of hospitalised
diabetic patients, with intensive glucose control (glucose target
4.5–6 mmol/l) versus conventional control (glucose target 7.8–10
mmol/l). The relative risk (RR) of death was 0.93, favouring
conventional control. About 50% of the trial participants reported
hypoglycaemia, with a pooled RR of 6 for hypoglycaemia in the
intensive-control group.
Peri-operative management of patients with diabetes
GREGORY HOUGH
This suggests that the reason for the negative outcome of the
trials of intensive treatment in hospitalised diabetics may be due to
the inevitable increase in hypoglycaemia. This trend has also been
seen in outpatient-care trials of intensive versus standard care in
diabetics. The ADVANCE, ACCORD and VADT trials of intensive
versus standard care in diabetics with a high risk of cardiovascular
disease all showed an increase in mortality in the intensive-control
groups. Subsequent meta-analyses have shown that in these
groups, this incidence of hypoglycaemia was significantly higher
than in the standard-care groups, suggesting this was the possible
cause for the negative outcomes.
It may be reasonable therefore to say that hypoglycaemia should
be avoided in these patients at the expense of tight glycaemic
control. It seems that high-risk patients in both in- and out-patient
settings are more vulnerable to the severe adverse outcomes of
hypoglycaemia.
Interestingly, both the meta-analysis mentioned above,
5
and
randomised, controlled trials
2,6,7
show that critically ill surgical
patients may actually benefit from tighter glycaemic control,
whereas critically ill medical patients do not. This may be a reflection
of the pre-hospital health of surgical and medical patients where, by
their very nature, surgical patients are likely to have had better pre-
admission health status and therefore be better able to tolerate the
inevitable hypoglycaemia associated with tight glycaemic control.
How do we incorporate this information into
guidelines for peri-operative care of diabetics?
See Tables 1 and 2 for definitions of hyper- and hypoglycaemia.
Where hyperglycaemia is discovered incidentally on routine
testing of patients who have not previously been diagnosed with
dysglycaemia or diabetes, these patients should be monitored and
managed as if they were diabetic for the duration of the admission.
Upon discharge, a formal plan for follow up of the impaired glucose
metabolism should be made to assess if this was just a transient
problem or one which will need further treatment and follow up.
Table 2.
Hypoglycaemia.
Blood glucose (mmol/l)
Hypoglycaemia
< 3.9
Mild to moderate hypoglycaemia
2.2–3.9
Severe hypoglycaemia
< 2.2
Note: these are laboratory values. Finger-prick values should guide treatment
but if there is doubt, confirm with a formal laboratory test.
Table 1.
Hyperglycaemia.
Hyperglycaemia Diagnostic of diabetes
Fasting glucose level
> 5.9 mmol/l
> 7 mmol/l
Random glucose level
> 7.8 mmol/l
> 11.1 mmol/l
HbA
1c
level
> 6%
> 6.5%
Note: these are laboratory values. Finger-prick values should prompt a formal
laboratory test.