The Prevalence of Physiological Anisocoria and its Clinical Significance - A Neurosurgical Perspective
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.273623
Source of Support: None, Conflict of Interest: None
Keywords: Anisocoria, neurosurgery, physiological, simple
Anisocoria has been defined as a difference in pupillary diameter of more than 0.3 mm. The occurrence of anisocoria without any underlying pathology is called physiological or simple anisocoria. Lam et al. in 1987 reported the prevalence of a physiological anisocoria to be 19%, while examining subjects at different times of the day.
Assessment of the pupils is part of the routine neurological assessment. In patients with traumatic brain injury, large mass lesions, and malignant infarcts, the finding of anisocoria may be the first sign of lateral transtentorial herniation of the brain and subsequent permanent brainstem injury. This is especially true in patients who are being sedated and ventilated since the Glasgow Coma Scale (GCS) score can no longer be used to monitor their neurological status. Decisions to repeat computerized tomography (CT) or magnetic resonance imaging (MRI) scans and occasionally even to operate may be made depending on pupil size. Keeping in mind that roughly a fifth of the population may have simple anisocoria at any given point in time, the degree of importance a neurosurgeon gives to the clinical finding of anisocoria may need to be questioned. We, therefore, aimed to study the prevalence of physiological anisocoria in the Indian population and also to evaluate the accuracy of clinical assessment of anisocoria employed in routine bedside examination.
This prospective study was approved by the Ethics committee of the Institutional Review Board (IRB no: 7935; 12th August, 2013). A total of 708 men and women between the ages of 20–69 years who did not have history of ophthalmologic or neurological disease other than refractive error were included after obtaining informed consent. They were current or retired staff and students of the tertiary care center, where this study was conducted, and relatives of patients admitted in the hospital. Those with a history of head or eye trauma including surgery, recent use of medications that affect pupillary size, and those on examination who were found to have corneal opacities, mature cataracts, colobomas, oculomotor nerve palsies, Horner's syndrome, or Adie's pupil were excluded from the study.
Photography and measurements
In a closed room with uniform ambient lighting of 28 foot candles/301 lux (measured at the site of examination and photography) designed to roughly match that of the neurosurgical intensive care unit (ICU) of the hospital, the subjects' pupils were examined clinically, and the presence or absence of anisocoria was recorded. Flash photography was used with a Canon Power Shot SD1100 IS digital camera fixed on a tripod at a distance of 25 cm from the subject. The duration of the flash of the camera was 1.4 ms, with a shutter speed of 16.7 ms. This allowed photography well before the onset of pupillary constriction, which has a physiological delay of approximately 250 ms. After removing their spectacles (if any), the subjects rested their chin and forehead in the frame and were asked to look into the distance when the photograph was taken to avoid accommodation. A graduated ruler was placed horizontally 1 cm below the lower eyelid for calibration [Figure 1]. Images were uploaded onto a computer, and the area of each pupil was calculated in pixels using GNU image manipulation program (GIMP Version 2.8 ©2001–2015). The diameter was then derived and converted to mm using the calibration provided by the ruler. A difference in pupillary size of 0.4 mm or more was considered anisocoria. Two authors (APA - Ananth P Abraham, ASG - Anjali Sarah George) examined subjects. The first 423 subjects were examined by APA, and the next 285 subjects were examined by ASG. Thirty random subjects were later examined independently by both researchers to assess the inter-rater reliability, and they were found to have 100% agreement – 29 had isocoria and one had anisocoria. Both the concerned authors had sufficient experience working in the neurosurgical ICU and were well versed in examining pupils.
Statistical analysis was done using SPSS (Version 21.0. Armonk, NY: IBM Corp.). Proportions were used for prevalence. Sensitivity and specificity were calculated for the clinical assessment of anisocoria using photographic measurement of pupillary diameter as the gold standard. Percentage agreement was used to compare inter-rater reliability.
Of the 708 subjects (age range 20–69 years), 361 (51%) were males. The study population included 65 diabetics, 60 hypertensives, and 159 subjects with refractive errors using spectacles. The average pupillary diameter among the males and females was 4.99 ± 0.97 mm and 4.98 ± 0.89 mm, respectively. The mean diameter of the pupil decreased with age as shown in [Table 1]. We found that of those with measured anisocoria, the left pupil was larger in 58 (60%).
Ninety-seven subjects (13.7%) had measured anisocoria, 54 of whom were males (56%). [Figure 2] shows the distribution of pupillary size differences among the subjects. It is evident from the bar graph that there is a sharp fall in the number of subjects with pupillary size differences of ≥0.4 mm. Measured anisocoria increased with age among males less than 60 years, whereas no such difference was found among the females. Eight each of the diabetic (12.3%), and hypertensive subjects (13.3%) had measured anisocoria, whereas it was found in only 9 (5.7%) of the subjects with refractive error.
Interestingly, the prevalence of anisocoria on clinical examination was also 13.7%. However, of the 97 subjects detected to have anisocoria on clinical examination, only 45 actually had measured anisocoria. Fifty-two subjects had measured anisocoria that was not detected clinically. This difference in clinical and measured anisocoria is depicted in [Figure 3]. The clinical measurement of anisocoria showed a specificity of 0.91 and a sensitivity of 0.46. With a prevalence of anisocoria of 13.7%, the positive predictive value (PPV) was 0.46, and the negative predictive value (NPV) was 0.91. The sensitivity of clinical assessment in the diagnosis of anisocoria increased with the degree of anisocoria, whereas the specificity remained almost the same [Table 2].
Prevalence of physiological anisocoria
The prevalence of physiological anisocoria in our study population, using the traditional definition of a pupillary size difference of ≥0.4 mm was 13.7%. This is lower than that reported in the western population (19–21%)., However, it must be remembered that while the pupils of our subjects were assessed with ambient lighting matching that of the neurosurgical ICU, previous authors had photographed pupils in dim light., A study by Lam et al. has reported that the prevalence of physiological anisocoria decreases under brighter lighting conditions, and it is therefore possible that the lower prevalence in our study could be because of the difference in ambient lighting. This was intentionally done by us in order to simulate the background lighting in our ICU because from a neurosurgical perspective it is this prevalence which is clinically relevant and should be kept in mind while assessing patients' pupils at the bedside.
There was an obvious trend of decreasing pupillary size with age [Table 1] as expected. There was no significant difference in the prevalence of anisocoria between the sexes or between various age groups. There was no increase in anisocoria among subjects with refractory error, diabetes, or hypertension. An interesting finding, although of doubtful clinical significance, was that 60% of those with anisocoria had a larger left pupil.
Accuracy of clinical assessment of anisocoria
The clinical accuracy of diagnosing anisocoria showed a large margin of error with a low sensitivity of 0.46 and a relatively high specificity of 0.91. The implication of this is that several patients who have anisocoria will not be detected on clinical examination, but anyone found to have anisocoria on clinical examination is quite likely to truly have anisocoria. This makes bedside pupillary assessment a poor screening tool to detect early anisocoria, but the clinical finding of unequal pupils should be given due importance. The inaccuracy of clinical assessment in diagnosing anisocoria becomes especially problematic in the context of sedated and ventilated patients in whom assessment of GCS score is not possible. In such patients, anisocoria can be the first sign of impending uncal herniation, and it is possible that this finding can be missed on routine clinical examination of pupils. The resultant delay in appropriate management may result in severe and irreversible brain damage.
It is essential to keep in mind that in the background of a low pretest probability of anisocoria, such as inpatients with normal brain imaging and no intra-orbital pathology; the PPV may further decrease and the NPV likely increase. The inverse will occur in patients with a high pretest probability of anisocoria such as those patients found in neurosurgical ICUs, and hence, the presence of anisocoria on bedside clinical examination should always be taken seriously. Our finding of a lower than previously reported prevalence of physiological anisocoria reinforces the need to assume any observation of anisocoria in the ICU to be of significance.
Increasing asymmetry of the pupil made the clinical detection more sensitive with almost unchanged specificity as seen in [Table 2]. For example, if a difference in pupillary size of ≥0.6 mm is employed as the cut-off for anisocoria, the sensitivity of clinical detection of anisocoria would improve to 0.62 with no remarkable reduction in specificity. In addition, as is evident in the same table, the prevalence of physiological anisocoria becomes less with larger degrees of pupillary asymmetry. This implies that while large degrees of anisocoria may be picked up more easily, early anisocoria can often be overlooked.
Another important consideration is the occurrence of hippus, which is the physiological oscillation of the pupil seen in most individuals. It could possibly affect the accuracy of pupillary size measurements, but because it is a normally a synchronous phenomenon in both pupils, it is unlikely to affect the determination of anisocoria, which is depends on on difference in pupillary size.
Given the unreliability of clinical assessment at detecting early anisocoria, the use of a hand-held pupillometer must be considered. Recent studies have shown that automated quantitative pupillometry in the ICU is user-friendly and provides a less subjective assessment of pupils at the bedside; thus, enabling earlier detection of anisocoria for more timely patient treatment.,
Physiological anisocoria was present in 13.7% of subjects in this study, which is less than the previously described prevalence. The sensitivity of clinical examination in detecting early anisocoria is poor. This delay could lead to adverse outcomes in neurosurgical patients who are at a high risk of uncal herniation, especially those who are being sedated and ventilated. Such patients may benefit from routine bedside pupillary assessment with a portable device such as the automated pupillometer.
The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2]