Senile Cataract

The Case Against Light.

R. A. WEALE

Abstract: A photochemical theory of the etiology of senile cataract is de-

veloped on the assumption that observed prevalences are representable by
cumulants of tolerances, the latter being distributed normally and depending
in a quantitatively defined way on the amount of short-wavelength radiation
absorbed during life. The theoretical risk-curves are consistent with the ob-
served ones. The variation of the predicted and observed values has a slope
not differing significantly from unity, but the intercept differs from zero. It is
concluded tentatively that the photochemical hypothesis accounts for the ob-
servations better and better as age advances, and more so where the radiation
is more intense. [Key words: photic radiation, quantitative analysis, senile
cataract.] Ophthalmology 90:420-423, 1983

There is increasing evidence to show that ordinary
daylight may have a cumulatively toxic effect on our
eyes. Short-wavelength, if not ultraviolet, radiations ap-
pear to be noxious. For example, climatic droplet ker-
atopathyl and pterygium,2-3 both grave corneal diseases,
are associated with high irradiances, and the latter con-
dition seems to be prevented even by ordinary specta-
cles. 3 Insofar as cataract is concerned, the effects of ra-
diations may be direct, following photic absorption in
the lens, 4-6 or indirect, resulting from modifications of
the aqueous humor. 7
Senile cataract, increases with age, as its name implies.
We face the question whether this is a manifestation of
a gerontologic programS or whether one is dealing with
an example of pseudosenility. Old crystalline lenses are
not only aged, but, like old retinae,9 have also been ex-
posed to radiation for longer than is true of young ones.
Moreover, in the South, cataract is not only rifer, but
also begins earlier. lo Other conditions may, of course,
be wholly responsible, eg, diarrhea has been mentioned II
as a powerful potential etiological factor.
One of the difficulties associated with cataract epi-
demiology is the absence of quantitative hypotheses. 12
The following is an attempt to remedy this. A general-
ized photochemical theory of cataract formation is ad-
From the Department of Visual Science, Institute of Ophthalmology,
London.
Reprint requests to R. A. Weale, Department of Visual Science, Institute
of Ophthalmology, Judd Street, London, England WC1 H 905.
vanced on the basis of available clinical and experi-
mental data. The assumption is made that senile cata-
racts in general are caused by radiation and that their
specific form is governed by idiosyncratic, and possibly
genetic, factors. The theoretical result is applied to a
simple analysis of tolerance to toxic effects of short-
wavelength radiation assumed to follow a normal dis-
tribution. Certain predictions are made relating cataract
prevalence and ambient illumination. Seven prevalence
studies are used to test them.
THEORY
Assume that the absorption of every quantum of
(short-wavelength) radiation leads to unreversed damage
of a lenticular molecule, fiber, or more generally, struc-
tural entity of the lens. The number of entities damaged
is related to their number, v; to the irradiance per unit
area, I; to the exposure time, t; to the pupillary area A;
to the spectral absorbance, a (A) where A is the wave-
length of the radiation; to the quantum efficiency, 'Y (ie,
the number of molecules altered per quantum ab-
sorbed); and the specific tissue sensitivity, which is a
constant of proportionality. Thus,
dv = -vas'YIAT (equation 1)
'Y is a very low number since the half-time of senile
cataract is some 64 months (as compared eg, with the
few seconds for visual pigments). Both a and A vary
with age. A number of studies suggest that, at a wave-

420 0161-6420/83/0400/0420/$1.00 American Academy of Ophthalmology

 WEALE SENILE CATARACTS
0.1
0.01
/ / / /
0.001
fill
Ii ~
e
/ / lee
0.0001
o No, and u is the standard deviation. It follows from
0.00001 .........:I._.a...._.l....---.l._.....I.._.l....---.l._ _ _ _ _......J
-u -I~
109,
Fig 1. Ordinate: Risk. Abscissa: logarithm of the deviate, i.e. tolerance.
The continuous curve represents 'the cumulant best fitting UK
data. '8.'9 The dashed curves are horizontal displacements of the base
curve, representing changes in the standard deviation of 0.1 log units
or 26%. Symbols: 0 = refs,'8.19 CD etc,22 ~,23 e,24 V,25 ~,26 ().27 A
qualitative description of this figure may be of assistance. If it be as-
sumed that the sensitivity to photically induced cataracts follows a
Gaussian distribution, then all persons having a given threshold and
all those having a lower one will be affected by the corresponding
threshold exposure. Hence the use of cumulants (or integrals of the
Gaussian curve). If the exposure is high then the total integral will be
reached over a narrow range much as a high dosage of a toxin acts
more quickly on a susceptible population than does a low one. Nor-
malized Gaussian curves are distinguished only by their width (the
u of the data): similarly the sigmoid cumulants intersect at their 50%
value and splay out with u. Although they deviate in theory from
-OCJ to +OCJ for values of -0.5 and 0.5 respectively, about the common
point of intersection, in practice -4.9 and +4.9 form acceptable limits.
Consequently if a number of related cumulants are locked at either
of these points instead of at their mid-point then their separation will
increase linearly with u. If they differ owing to the variation of a single
variable then their logged value will exhibit parallel displacements as
shown in the above Figure. Steep curves have smaller displacements
than do shallow ones. The single variable hypothecized is irradiation.
length of 400 nm l3 and also in the UVA region,14 the
age variation of the absorbance a is given by
a = k(I.OS + 0.01922t)
where k is a constant. The pupil area of the very light-
adapted eye9 is given approximately by
A = 6.223 exp( -0.00391 t)
According to Said and Sawires,15 it is significantly
smaller (by '- 38%) among indigenous Egyptians; this
is likely to be the case in the South in general. Even
smaller pupil sizes have been reported for the Philip-
pines. 16 When these two expressions are substituted in
equation I and this is integrated by parts
log., vivo = -{ I - (l + 0.00322t)
X exp(-0.0039lt)}{ks'YIA}. (equation 2)
In order to develop quantitative hypothesis, we define
a cataract as being present when a fixed fraction of len-
ticular entities is degraded. If this is SO%, the LHS
= -0.693 so that
r = lis = k''YIA[(l - (l + 0.00322t)
X exp(-0.039lt)] (equation 3)
where k' is another constant; r is the reciprocal of the
sensitivity s, and, therefore, the tolerance. In a random
population of N persons, r is assumed to be distributed
normally so that
NINo = exp -([r - I]laY (equation 4)
where I is the mean tolerance for the total population
equation 3 that u is a linear function of 'YIA. This is to
say that, the greater the ambient illumination the nar-
rower the spread of the age spectrum of the patients.
The cumulant of equation 4 gives the fraction of the
total population having a tolerance equal to r or less and
can, therefore, be compared with clinical prevalence
data valid for randomly selected cases. Standard data
for cumulants have been tabulated. 17 They are expressed
as a function, eg, of (r - I)lu (equation 4), ie, in terms
of the ratio of the deviate to the standard deviation. It
follows that the latter and its submultiples are merely
scaling constants. Its value was determined as follows.
It was assumed that the experimental prevalences can
be represented by cumulants. Consequently a deviate,
D(th) could be read off the table for each risk value.
Each risk value being associated with an age group, a
value for r was determined from equation 2. If all the
risks are very low the average of the observed r-values
and the I-value corresponding to a risk R = O.S cannot
be the same. The experimental deviates D(ex) were ob-
tained by determining I-value (=0.187), which mini-
mized the variance of p, the ratio D(th)/D(ex) for the
data obtained in the United Kingdom. 18 ,19 It will be re-
called that the cOhstancy of this ratio would be expected
if the data obey a normal distribution. The calculated
value of p = 26.S92 1.366 and this is the value whereby
the tabulated deviates and, therefore, the values r (=4.9-
deviate) have to be divided in order to yield the best-
fitting theoretical curve. This is shown in Figure I.
If the cumulants are plotted, not against r, but against
log r, it follows from equation 3 that the abscissa can
be calibrated in log irradiance units. This is to say, that,
if we use the best-fitting curve as a base line (Fig I), a
displacement of the curve to the left predicts prevalences
at higher intensities, but a rightward one at reduced ones.
421
 OPHTHALMOLOGY APRIL 1983 VOLUME 90 NUMBER 4

40
TESTS

Figure 1 shows that the majority of the data are well

described by equation 4 but, in itself, this does not pro-
vide a sufficiently detailed test for the role of light. For
this to be done, estimates for annual azimuthal sun-
shine-hours averages have to be usedY Note that sun-
shine hours on their own inform merely on the presence
or absence of cloud cover. 20 The azimuthal light of the en
.J:J
o

.
sun also enters the picture. The values used are shown a:
in Table 1.
It will be noted that what is used is sky luminance
rather than solar intensity. This means not only that
there is no need for eyes to be exposed to direct sunlight
for the hypothesis to operate, but also that the way is ., .~
left open for cataracts other than nuclear ones to be
examined in this context.

A prevalence is predicted theoretically as follows. We
assume the calculated curve to represent a base-line.
Since there is more doubt about its course at 30 years Rtheor
and below than elsewhere, the lowest age considered is
- 32 years. The observed data run mostly approximately Fig 2. Ordinate: Weighted observed risk. Abscissa: Weighted theoret-
parallel to the base curve. Their displacement at all ages ical risk. The dashed line is the expected, the continuous one the
is, therefore, given by one constant. The hypothesis pos- derived regression. For details see text. A qualitative description of this
figure may be of assistance. It would be relatively easy to test the
tulates that this is given by the appropriate value in the validity of a hypothesis if the experimental data were found to aggre-
last column of Table 1, with the dashed curves in Figure gate at two extremes: a high correlation is easy to establish in such
1 representing 0.1 log radiation units. One can therefore circumstances except that, to be valid, the data have to follow a normal
calculate the prevalence to be expected in a given photic distribution, a condition clearly not satisfied by a "dumb-bell" pop-
environment in terms of the smoothed UK data, and ulation. One way of dealing with this problem is by weighting the data
to compare the expected with the observed data. It is, with the sample size: this minimizes the effect of "out-liers". The data
however, necessary to weight the observations for sam- in this figure therefore do not represent simply observed vs. predicted
ple size. There is a variety of ways of doing this: the one values, but are weighted, thereby making significance tests both more
adopted here was to divide both the observed datum severe and more objective.
and the appropriate prediction by the standard deviation
of the former. The points in Figure 2, therefore, repre-
sent VRN/( 1 - R) where R is the risk fraction and N observed data are highly correlated (r = 0.908). Note,
the size of the sample. however, that the test is much more severe than one
Analysis of Figure 2 shows that the predicted and testing the significance or otherwise of a correlation:
what is tested is the prediction that the observed points
do not depart significantly from the dashed line. It is
Table 1. Annual Azimuthal Sunshine Hours found that they do. To be specific, the slope of the regres-
sion does not differ significantly from the predicted value
Near Sunshine Mean Sky of unity (t = 1.583, 2P > 0.1); however, the intercept
Hours Luminance differs significantly from zero (P < 0.001). The observed
Country 20 (pa) 21 (pa) ReI. Log. data are, in general, greater than predicted. It may be,
therefore, that the prediction accounts only for part of
UK18.19 1501 1982.5 1 0 what passes as senile cataract. It is also possible that
Australia 22 zone 1 2037 2646 1.811 0.258* more detailed attention to senile miosis, lenticular ab-
zone 2 2504 2813 2.367 0.374*
zone 3 2792 3146 2.952 0.470*
sorption, not to mention the prevalence data themselves
zone 5 3077 3413 3.529 0.548* may help to support or reject the hypothesis.
Denmark 23 1603 1765.5 0.951 -0.022 Since the slope of the regression does not differ sig-
Greenland 24 2443 1015 0.833 -0.079 nificantly from unity, one asks oneself what is the mag-
India 25 3285 3055 3.373 0.528*
Korea 26 2423 2585 2.105 0.323
Mali27 3341 3378 3.793 0.579* Table 2. Percentage of Cataracts Attributable to Light
The entries marked * are to be reduced by 0.139 to allow for a Age (years) 42 65
"southern" pupil area. The tabulated data were obtained on the as- Denmark 17 76
sumption that Western pupil diameters and senile miosis apply eve rye India 61 85
where. The data reduced by the above factor were used in Figure 2.

422
 WEALE SENILE CATARACTS
nitude of the intercept on the y(observed)-axis if unity
is assumed. The coordinates of the mean values are
7.2092,8.9217 so that the intercept is 1.7125. This rep-
resents observed cataracts not accounted for in terms of
the photic hypothesis. As might be expected, there is a
convergence at higher ages of the extent to which light
acts as a factor. This is illustrated in Table 2, which
shows the percentage of cataracts so accounted for at
two ages in very different photic environments.
One28 authority questions to what extent interna-
tional comparisons of prevalences are feasible, and
quotes the fact that frequency (of operations) in the USA
is three times greater than is true of the UK. Although
other explanations may spring to mind, the overall better
climate in the United States is consistent with the pres-
ent hypothesis if the frequencies of operations and of
incidence are comparable. Figure 1 shows that a mere
20 pc more light could create the greater risk. The ob-
served variation between relative prevalence and (non-
azimuthal) sunshine hours in the US 29 is predicted by
equation 4. A report on whether it is explained quan-
titatively better by the recently mentioned incidence of
diarrhea30 is still awaited.
CONCLUSIONS
Simplifying assumptions notwithstanding (a correla-
tion has been reported6 between geographical latitude
and only brunsescent cataract), the above hypothesis
describes at least three sets of observation provided it
is valid. In the first place it states that, in addition to
biological tolerance variations, senile cataract depends
on light exposure, ie, absorbed intensity multiplied by
exposure time. Secondly, it places the vastly higher prev-
alence in southern countries within the same framework.
Thirdly, it indicates the involvement not just of direct
sunlight but illumination from the whole sky: this would
explain why coolie hats appear to delay the onset of
cataract in at least one controlled situation. The quan-
titative agreement may, of course, be partly or wholly
accidental. Dietetic factors may playa notable and cu-
mulative role, leaving light with nothing to do. Cultural
and social factors, eg in the shape of a siesta, may render
the above quantification harder. Nevertheless, radiation
has long been speculated on as a cataractogenic agent,
and the above analysis provides a crude but quantitative
test.
It is easily seen that, if the photic hypothesis holds,
eye protection or even suitable head-gear may reduce
prevalences everywhere by an order of magnitude or
more.
ACKNOWLEDGMENTS
The author thanks Dr. V. M. Reading of this Institute and
Mr. M. R. Weale, Clare College, Cambridge, for helpful dis-
cussions.
REFERENCES
1. Freedman A. Climatic droplet keratopathy. I. Clinical aspects. Arch
Ophthalmol 1973; 89:193-7.
2. Clear AS, Chirambo MC, Hutt MSR. Solar keratosis, pterygium, and
squamous cell carcinoma of the conjunctiva in Malawi. Br J Ophthal-
mol 1979; 63:102-9.
3. Cameron ME. Pterygium Throughout the World. Springfield IL:
Charles C Thomas, 1965.
4. Cremer-Bartels G. A light-sensitive, fluorescent substance in bovine
and rabbit lenses. Exp Eye Res 1962; 1:443-8.
5. Lerman S, Borkman R. Spectroscopic evaluation and classification
of the normal, aging, and cataractous lens. Ophthalmic Res 1976;
8:335-53.
6. Zigman S, Datiles M, Torczynski E. Sunlight and human cataracts.
Invest Ophthalmol Vis Sci 1979; 18:462-7.
7. Ringvold A. Aqueous humour and ultraviolet radiation. Acta Ophthal-
mol 1980; 58:69-82.
8. Orgel LE. Ageing of clones of mammalian cells. Nature 1973;
243:441-5.
9. Weale RA. A Biography of the Eye: Development, Growth Age. Lon-
don: HK Lewis, 1982.
10. Weale RA. The age variation of "senile" cataract in various parts of
the world. Br J Ophthalmol 1982; 66:31-4.
11. Harding JJ, Rixon KC. Carbamylation of lens proteins: a possible
factor in cataractogenesis in some tropical countries. Exp Eye Res
1980; 31 :567 -71.
12. Anon. Epidemiology of cataract. Lancet 1982; 1:1392-3.
13. Wemer JS. Development of scotopic sensitivity and the absorption
spectrum of the human ocular media. J Opt Soc Am 1982; 72:247-
58.
14. Boettner EA, Wolter JR. Transmission of the ocular media. Invest
Ophthalmol1962; 1:776-83.
15. Said FS, Sawires WS. Age dependence of changes in pupil diameter
in the dark. Optica Acta 1972; 19:359-61.
16. Fajardo RV, Talusan ED. The Filipino pupil pattem. Philipp J Ophthal-
mol 1976; 8:38-40.
17. Fisher RA, Yates F. Statistical Tables. Edinburgh: Oliver and Boyd,
1963.
18. Sorsby A. Cataract: some statistical and genetic aspects. Exp Eye
Res 1962; 1:296-9.
19. Caird FI, Hutchinson M, Pirie A. Cataract extraction in an English
population. Br J Prev Soc Med 1965; 19:80-4.
20. Climatological normals 1931-1960. World Meteorol Org, 1971.
21. Walsh JWT. The Science of Daylight. London: Mac;:donald, 1961.
22. Hollows F, Moran D. Cataract-the ultraviolet risk factor. Lancet
1981; 1:1249-50.
23. Braendstrup P. Senile cataract; account of cataract extractions per-
formed in an urbanized population during the third quarter of the
present century. Acta OphthalmoI1977; 55:337-46.
24. Klauber A. Frequency of cataract extraction in Greenlanders. Acta
Ophthalmol 1981: 59:532-8.
25. Chatterjee A. Cataract in Punjab. In: The Human Lens-in Relation
to Cataract. Ciba Foundation Symposium 19 (new series). Amster-
dam: Associated Scientific Publishers, 1973: 265-71.
26. Koo BS, Chai BS. Survey of causes of blindness on hospital patients
in Korea. J Korean Ophthalmol Soc 1974; 15:22-35.
27. Anon. Preliminary results of a survey on loss of vision in a Sahel area
in West Africa. WHO-PBL/AG/81.7 1981.
28. Sommer A. Cataracts as an epidemiologic problem. Am J Ophthalmol
1977; 83:334-9.
29. Hiller R, Giacometti L, Yuen K. Sunlight and cataract: an epidemio-
logical investigation. Am J Epidemiol 1977; 105:450-9.
30. Harding JJ. Cataract: sanitation or sunglasses? Lancet 1982; 1:39.
423