SHOULD WE START TREATMENT ?
"When does one start treatment ?" is one of the most
common and persistent questions asked at conferences and
symposia about glaucoma. The experts are always pleased to
oblige with advice. You start treatment for glaucoma, they
tell us, when the intraocular pressure reaches a level of
30, plus or minus 4 mm. Hg., provided that the disc appears
healthy. If the disc is damaged, treatment should be
started at a lower pressure. The advice is simple and
obvious. We have heard it often and yet the answer does not
dispose of the question. Confronted with a patient who has
early glaucoma, one frequently asks oneself in perplexity,
"Should I start treatment?" If the question is to be
considered once more, perhaps one should seek not to provide
yet another answer, but to try to understand what is wrong
with the conventional answer, and perhaps also, what is
wrong with the conventional question.
The conventional question "When should you begin
treatment for glaucoma?" fails, in the first place, to
acknowledge that glaucoma is not one but many diseases.
Although it is seldom phrased explicitly, I have always
assumed that the conventional question addressed itself, if
to a real disease at all, then only to primary open angle
glaucoma, leaving a set of cognate questions concerning the
optimal time for initiating treatment of other glaucomas not
only unanswered but also unasked. In view of the diversity
of these other kinds of glaucoma, the question about open
angle glaucoma represents only the small fragment of a much
broader issue; and if notwithstanding this arbitrary
limitation the narrower issue does not appear soluble,
perhaps reviewing it in a broader frame of reference might
be of help.
To the extent that the conventional question "When do
you start treatment for glaucoma?" addresses itself to an
actual disease rather than an abstract conception, it
ignores the fact that primary open angle glaucoma can often
not be unequivocally identified. The diagnosis of open
angle glaucoma is in large measure a diagnosis of exclusion.
We refer to it as primary because examination of the eye,
and particularly of the angle, shows nothing that might
explain its origin. Yet primary open angle glaucoma differs
among patients in many ways. One patient develops primary
open angle glaucoma against the background of a strong
family history; another is the first in his family to
acquire the disease. In some patients, the disease appears
to be unmasked by corticosteroids, while others are
demonstrably unaffected by this class of drugs. In some
patients, the disease develops and progresses at relatively
low intraocular pressures, while yet other patients are able
to tolerate markedly elevated tensions for many years before
the discs finally show the characteristic stigmata of
glaucomatous excavation. Then there are eyes that have been
subjected to blunt trauma that develop a disease
indistinguishable from open angle glaucoma without showing
any disruption of the trabecular meshwork, while other
patients with significant angle recession seem to have
normal pressures, normal outflows, and no evidence of
progressive disc excavation. There are relatively many
patients in whom open angle glaucoma and angle closure
glaucoma appear to coexist, where tension elevation at one
stage of the disease is attributable to the open angle
component, at another stage to reversible angle closure, and
at yet a third stage to the formation of permanent anterior
peripheral synechia. There are, furthermore, the well-
identified patterns of pigment dispersion syndrome and
pseudoexfoliation of the lens which are sometimes
accompanied by glaucoma and sometimes not. In all of these
situations, where open angle glaucoma coexists with other
pathology, the indications for therapy will be
correspondingly different, and the conventional answer may
be inapposite.
It is remarkable, in view of the diversity of the
glaucomas and in view also of the frequent difficulty of
identifying primary open angle glaucoma as such, that the
question "When do we start treatment of glaucoma?" is
nonetheless asked - and answered - in so general a form.
And if those who ask and those who answer the question agree
that it is meaningful that may be because at its root there
is an assumption about glaucoma which should be made
explicit. This assumption is that although there are many
kinds of glaucoma and many different causes of elevated
pressure and even though it may be possible to recognize
subspecies even of open angle glaucoma, yet one may
meaningfully speak of the glaucomas as a group because of
the process that they have in common, namely the excavation
of the optic nerve and the ensuing field loss brought about
by an intraocular pressure of which, over a period of time,
the eye proves intolerant. In this perspective the question
is logical, for if tension elevation is the unique mechanism
by which nerve excavation and field loss are brought about,
then the modality by which that pressure is elevated to
intolerable levels is irrelevant to its destructive effects.
Thus it appears that the conventional question "When do
we institute treatment for glaucoma?" addresses itself not
to an actual disease, but to a schematic representation of
that disease, in other words, to a model. If the
conventional answer to the question is unsatisfactory that
may be because it is not models of disease that we are
constrained to treat, but actual pathologic processes in
unique human beings. This conundrum about the proper time
to begin the treatment of glaucoma illustrates the hazard of
uncritical reliance on an abstraction. The model is a
generalization that disregards the mechanism by which the
pressure is elevated. It disregards also the particular
circumstances surrounding the disease, the patient's age,
sex, race. It disregards the history of glaucoma in his
family. It ignores the concomitant existence of other
pathology. It ignores the patient's acceptance or rejection
of treatment. It ignores the severity or mildness of any
side effects that the patient might experience from the
therapy, and finally it ignores the relative effectiveness
or ineffectiveness of the treatment in actually lowering the
pressure. It addresses itself solely to the phenomenon that
a given intraocular pressure may over a period of time cause
excavation of the optic disc and ensuing field loss, that
increasing the pressure will accelerate, and diminishing
the pressure, will retard this process. From this model one
infers intuitively that when, in an eye with a normal disc,
the intraocular pressure reaches some stated value such as
30 mm Hg, treatment to lower the pressure should be begun.
Implicit in this conclusion is the assumption that for
pressures under the threshold value, the probability of
field loss within a forseeable period of time is relatively
small, and perhaps also, although it is never made explicit,
the assumption that the elevated pressure might subside
spontaneously before it had caused any damage to the field.
For values above the threshold pressure, on the other hand,
the probability that the patient will develop field loss in
a forseeable span of time is deemed to be substantial and
the possibility that the pressure would remit spontaneously,
relatively remote.
It is useful to make explicit the reasoning by which
such conclusions are reached. By definition, a healthy disc
subjected to a normal pressure should not deteriorate. It
is also a matter, unfortunately, of common experience that a
disc subjected to a pressure of 45 mm Hg usually begins in a
matter of months to show evidence of excavation, which, if
the pressure is unrelieved, will soon progress to field loss
and then to blindness. For intraocular pressures less that
45, for pressures of 42, 39, 36, 33, 30, for example, the
rate of disc deterioration will become progressively slower
until at at pressure of 15 it will have ceased altogether.
Between 45 and 15 there must be some value of the pressure
at which the rate of disc deterioration has become
sufficiently retarded that although disc damage progresses,
the rate is so slow that the patient will experience no
functional disability during his lifetime. When the experts
postulate a tension of under 30 as requiring no treatment,
that is what they have in mind. Nor is it unconditionally
necessary that the loss of vision be postponed for the whole
of the patient's life. As a practical matter, if loss of
vision is delayed for perhaps a matter of five or ten years,
the advice to withold treatment may, rightly or otherwise,
appear to have been correct. If the patient is followed
carefully, excavation of the disc would be detected before
it had advanced too far; treatment could then be instituted
before field loss supervened, although it is necessary to
point out in this context that under those circumstances
where excavation has already occurred, a much lower pressure
level appears to be required if the progression of
excavation is to be stopped.
The observation that the conventional answer refers not
to an actual patient but to a model of a disease is borne
out by its implicit assumption of an elevated intraocular
pressure fixed at a specified level. The pressure may
indeed be 30 today; it may have been 42 yesterday and it may
be 22 tomorrow. The examiner who measures the patient's
intraocular pressure is after all, only sampling that
pressure in the statistical sense. And it would be rare
indeed for an examiner to obtain only a single pressure
measurement on a suspected glaucoma patient. More likely,
there would be a series of measurements over a period of
weeks or months. Under these circumstances, linking the
decision to start treatment to a single tension reading is
clearly erroneous. It is a set of numbers which we must
interpret. It is faulty to extract from this set at random,
or even worse, with some ill-defined preconceptions of its
significance, a number that strikes our fancy, and interpret
it as being of decisive value. A statistician could suggest
a better approach.
A statistician would wish to know, in order to obtain a
clearer impression of pressure behavior of the eye, how many
measurements had been taken, and over what period of time.
He would calculate the intervals between tonometric
readings, and the more closely they were spaced, the more
confidently he would interpret them. He might calculate
also the arithmetic mean of the pressures and the standard
deviation from that mean. He might also determine whether
the pressures tended to rise or fall by fitting individual
measurements to a curve by the least squares approximation,
and thereby project values for the tension to any arbitrary
point in the future. Such mathematical techniques would aid
the interpretation of tonometry and would make decisions
based on tonometric data more rational and more persuasive.
The present methods of interpreting tonometric values in the
treatment of glaucoma are adumbrated with the aura of number
magic. Perhaps as computers become less expensive we will
become accustomed to more rigorous analyses of the
tonometric data on which the decision to begin treatment is
based.
More difficult and less accessible to rationalization
than tonometry is the interpretation of the disc. The rules
of thumb that we have been taught are qualified with the
provision: "If the disc appears normal ..." but it is often
very difficult to identify a normal disc. Consequently we
classify the discs that we observe into three groups. The
first of these groups consists of those discs which are
obviously pathologically excavated, and in the presence of
elevated tension one has no choice but to attribute such
excavation to the elevated intraocular pressure. The second
group consists of those discs which appear to be normal. Of
such discs once cannot with certainty say that they have not
been damaged by tension elevation. All that one can say is
that such glaucomatous damage, if any, is not visible
through the ophthalmoscope. The third group, and it is by
no means small, contains all the rest, all those discs of
which one cannot say with certainty whether or not the
morphology of the disc has been altered by elevated
intraocular pressure. To be sure, for the experienced
examiner that number will be smaller than for the novice,
and with experience ones uncertainty in this judgment
diminishes. But there are many discs about which not even
the most experienced of examiners can say whether or not
they exhibit glaucomatous damage. In a surprisingly large
number of patients who develop no other signs of glaucoma
the discs show some measure of excavation, or some degree of
asymmetry, some pallor and/or some nasal displacement of the
retinal vessels which though far from diagnostic for
glaucomatous change would not be incompatible with such
damage in its early stages. When one measures an elevated
pressure in an eye that has a disc of this third class, then
the simple rule that one should begin to treat patients when
the pressure is over 30 is potentially very misleading.
Now we know why the question "When do we start treating
glaucoma?" is asked over and over again, and why the
answers to that question, however elegantly they may be
presented, do not seem to help us very much when we are
faced with practical problems. At the heart of the
conventional answer there is a model of glaucoma, but it is
not a very useful model because it is not articulated
logically or mathematically, and its features therefore
necessarily remain subjective to the expert.
The best way that I know of to broach the question of
whether or not we should institute treatment is to outline
for ourselves as best we can the answers to the questions:
what will happen to him if we treat the patient, and, on the
other hand, what will happen to him if we withhold
treatment. And if, at any given juncture, we must admit to
ourselves that we do not really know, then it is important
to take that uncertainty into account when we make plans for
our patient. That we must take action in the face of
uncertainty is integral to our responsibility. To
acknowledge that uncertainty will give to our decisions a
measure of humility and a measure of dignity to our
mistakes. It will discourage us from presuming to know the
future, and it will leave us with a mind open to new
observations. Perhaps the best answer to the question, when
do we start treatment, is that it is not the time at which
but the manner in which treatment is instituted or withheld
which is critical. Whether the decision is for treatment or
against it, the fact that the issue ripened to the point of
decision means that we must thereafter follow our patient
faithfully, on the lookout for new evidence that might
compel us to reverse whichever course we had taken.
Here are two illustrations of the pitfalls of treating
patients by rule rather than by observation:
A physician sees a patient in his early thirties and
notes an intraocular pressure of 28. The disc appears
normal. He tells his patient that he has ocular
hypertension and instructs him to return in half a year.
The next morning at 6 a.m. however, unbeknownst to the
physician, the pressure is 39, the following afternoon it is
18, and so it fluctuates from day to day and from week to
week. At the time of the six months' examination, the
physician still finds no disc changes and records a pressure
of 20. He tells his patient that there is nothing wrong
with him and tells him that no further examinations are
required. By the time the patient returns, he is almost
blind. Consider, on the other hand, a patient with large
and somewhat asymmetric central excavations who was found by
his physician to have a tension of 31. He was immediately
given medication and subsequent pressure readings ranged
between 15 and 19. It was assumed that he had glaucoma,
that the excavations were the result of tension elevation,
and that the medication had brought the tension to a safe
range which prevented further damage. The patient is told
that he must take medication for the rest of his life.
What the physician did not know was that perhaps
because of transient extraocular muscle tension, or for some
other ill-defined reason, this patient's tension elevation
was not reproduceable. Subsequent tonometries, even without
medication would have given values in the mid to high teens.
If the physician had obtained a few more pressure readings
at intervals of perhaps six days, no need to wait six
months, he would have found the tensions much lower. It was
only when the patient switched physicians and was taken off
all medications to assess the effect of the drops that he
was receiving, that it was found that his pressures remained
normal even without medication, in fact that he "no longer"
had glaucoma.
Both of these errors could have been avoided. The
first one, if the patient with a tension of 28 had been seen
for at least a few times at one month's rather than at six
months' intervals, and if he had not been dismissed on the
basis of only a single "normal" tonometry. The second error
might have been avoided if the physician had waited a few
days before instituting therapy while additional pressure
measurements were obtained, or at least if, after six or
twelve months of presumably essential therapy, he had taken
his patient off medication for a few days to determine what
the pressures would be without treatment.
* * * * *
Back
Next
Glaucoma Letter Index
Website Index
Copyright 2006, Ernst Jochen Meyer