Visual outcome in 2000 eyes following microscopic transsphenoidal surgery for pituitary adenomas: Protracted blindness should not be a deterrent
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.193829
Source of Support: None, Conflict of Interest: None
Objective: To study the visual outcome after surgery for pituitary adenomas with visual deficits.
Keywords: Optic atrophy; pituitary adenoma; surgery; visual acuity; visual field
Pituitary adenomas can produce visual loss by compression of the optic chiasm or nerves. An extension of >10 mm above the seller diaphragm is necessary to compress the anterior visual system., Typically, pituitary adenomas cause temporal field defects, explained by anatomy of the visual pathways in the chiasm: Crossing inferonasal fibers lie at the anterior part of the chiasma and are therefore compressed first. This causes an upper temporal visual field (VF) defect progressing in an anticlockwise direction in the left eye and in a clockwise direction in the right eye.,,
Large adenomas produce an early loss of central visual acuity (VA) and VF, and induce optic disc pallor. Asymptomatic visual impairment is more frequent than clinically apparent vision loss. In the past few decades, trans-sphenoidal surgery (TSS) has replaced craniotomy as the preferred operative approach. Visual disturbance is one of the common presenting manifestations of pituitary macroadenomas both in functioning and nonfunctioning pituitary adenomas (NFPA). The recovery of vision occurs in a triphasic manner. There is no dearth of literature regarding visual outcome in patients undergoing TSS.,,, However, they are small series, and in many studies, the evaluation was done at a variable time frame. Similarly, most studies have considered VA and VF as a single unit, although these parameters are nonlinear.,,, Our study included a large cohort of patients, and the visual outcome was measured at discharge and at 3 and 12 months following surgery, and VA and VF were measured separately at each visit.
Patients who underwent surgery for pituitary adenoma at the Postgraduate Institute of Medical Education and Research, Chandigarh, from 2003 to 2014 were selected. From 2003 to 2010, patients were selected retrospectively from hospital records and computerized pituitary clinic database; and, from 2011 to 2014, patients were recruited prospectively. All patients included in the study had histopathologically confirmed pituitary disease, with documented evidence of preoperative visual dysfunction. A total of 1513 consecutive patients with sellar pathology were seen during this period out of which 1019 were having pituitary adenomas. In the initial years, the microscopic transphenoidal approach was sublabial rhinoseptal. This was followed by transnasal transseptal approach. For more than a decade, the senior author (KKM) uses a direct endonasal approach to the junction of vomer and sphenoid bone using microscope. For final analysis, 1001 consecutive patients with pituitary adenomas were included [Figure 1]. These patients underwent transsphenoidal (TSS)/transcranial surgery and had at least three visual assessments on follow-up visits (at discharge, and at 3 and 12 months after surgery). The exclusion criteria were: (i) Visual impairment due to a cause other than chiasmal compression; and, (ii) patients having an incomplete data.
A contrast-enhanced magnetic resonance imaging (MRI) of the sella was performed in all these patients preoperatively. The adenoma volume was calculated by the De Chiro and Nelson formula [volume = (sagittal × coronal × axial diameters) × π/6]. Modified Hardy's  classification was used for staging (extension) and grading (degree of sellar destruction) of the pituitary adenomas. We also used Knosp grading to document the parasellar extension. Based on clinical features, hormone profile, and immunohistochemistry (IHC) of the tumor tissue, pituitary tumors were classified as: (i) Nonfunctional, or (ii) functional: prolactinoma, somatotropinoma, corticotropinoma, and thyrotropinoma.
Ophthalmological assessment for color vision, VA, VFs, and fundus was performed by various neuro-ophthalmology technicians over a period of 11 years with the same machine, at presentation, immediately after surgery, and at 3, 6, and 12 months postoperatively. All ophthalmological records were analyzed by a single neuro-opthalmologist (RB) who had no knowledge of the radiology and the endocrinal profile. Color vision was assessed with standard Ishihara charts and American optical color plates. An Arden grating was used to search for elevations in combined thresholds. Other accessory visual examinations like intraocular tension and slit lamp evaluations were also performed. Both eyes were considered separately for the degree of involvement. VA was determined by the Snellen's chart, and VF testing was performed by Humphrey automated computerized perimeter, C76 Panel (Carl Zeiss, Germany). In patients with finger counting, hand movement, and only perception of light, the assessment of VF was done manually using confrontation test by trained neuro-ophthalmology technicians, as Freiberg quantitative test for such patients is not available in our hospital. The results were recorded as follows: (i) No change, (ii) improved, and (iii) worsened. Significant improvement/worsening was defined as any grade improvement or deterioration in the VA and VF, based upon a 30% change to avoid inter- and intra-individual variation, according to John Thomas Smith's rule of the one-third. Blindness was defined as absence of perception of light. Visual evoked potential was not performed routinely.
Data were explored for any outliers and missing values. Quantitative data were described as mean and standard deviation with their 95% confidence intervals. Categorical data were shown as frequencies and proportions. Frequency distributions of each study characteristics were shown separately for the preoperative and postoperative patient groups. Comparison of the preoperative and the postoperative patient group was carried out for various categorical variables using chi-square test of association to determine any statistical association between them. The mean values of age, duration of visual symptoms, and preoperative and postoperative visual status and other quantitative variables such as tumor height, tumor diameter, and tumor extension were compared within the group using chi-square test. A multivariate analysis for features such as age, diabetes, hypertension, and duration of visual impairment was done in patients only in the prospective group, as the data for presence of diabetes and hypertension were not recorded in the patients in the retrospective group. A P < 0.05 was taken as significant.
A total of 1007 consecutive patients with pituitary adenomas were recruited in the study. Twelve eyes (six patients) had dense cataract, proliferative diabetic retinopathy, age-related macular degeneration, and nonregmatogenous retinal detachment. 1001 patients (2002 eyes) were available for the final analysis [Figure 1]. The prospective data were available for 296 (29.6%) patients, and retrospective data were obtained for 705 (70.4%) patients. The mean age of the study population light (PL) was 41.60 ± 13.33 years (5–80 years), and of these patients, 58.3% (584) were male and 41.7% (417), female. Nonfunctional pituitary tumor (NFPT) was present in 64.8% of patients, while the rest had functional pituitary tumors, somatotropinoma being the commonest (23.9%) followed by corticotropinoma (8.6%), prolactinoma (2.6%), and thyrotropinoma (0.2%).
Headache and visual impairment were the most common symptoms and were present in 875 (87.5%) and 873 (87.3%) patients, respectively. Of the 2002 eyes, 1569 (78.3%) were involved. Clinical pituitary apoplexy (5.5%), seizures (3.8%), diplopia (1.3%), focal neurological deficits (2.1%), and altered sensorium (1.8%) were the other symptoms present in these patients. Endocrinological manifestations were present in 56% (561) of patients.
Visual symptoms were present for <6 months in 47%, for 6 months to <1 year in 10%, for 1–2 years in 12%, and for >2 years in 31% of patients. Preoperative impairment in VF and VA is shown in [Table 1], [Table 2] and [Figure 2], respectively. Fifteen percent (10.8% left, 4.8% right) of eyes had no perception of light. Bitemporal hemianopia was the commonest type of field defect (47.6%), followed by blindness, 8.9% in the right eye and 9.5% in the left. The frequency of upper temporal and lower temporal quadrantic defects did not show a significant difference. Ten percent of patients were not aware of their visual defect, which could be diagnosed only on routine ophthalmological workup for the first time.
Optic disc pallor was present in the right eye in 480 (48.0%) and in the left eye in 476 (47.6%) patients [Table 2]. Optic atrophy was seen in the right eye in 147 (14.7%) and in the left eye in 182 (18.2%) patients. There was significant positive correlation between the occurrence of optic atrophy and the duration of visual symptoms [Figure 3]; optic atrophy was present in only 0.2% of patients in the right eye and 0.9% in the left eye in those with duration of visual symptoms <6 months, while optic atrophy was present in 33.9% in the right eye and 41.3% in the left eye in those with duration of symptoms >1 year (P < 0.001).
Majority of patients (90%) had a macroadenoma, including a giant (≥4 cm) adenoma, and 10% of patients had a microadenoma. None of the patients with a microadenoma had visual abnormalities. Suprasellar extension of the tumor was present in 892 (89.2%) patients. Trans-sphenoidal adenomectomy was performed in 969 (96.9%) patients, while 31 (3.1%) patients underwent transcranial surgery.
In the immediate postoperative period, VA and VFs improved in 163 (16.3%) and deteriorated in 36 (3.6%) patients, and there was no change in either of the parameters in 801 (80.1%) patients. The deterioration of vision could be attributable to optic nerve injury, tumor bed hematoma or surgicel ® (ethicon) placement. Most of the visual improvement following surgery occurred after 6 months of surgery. After 1 year of surgery, 93.2% of patients had improvement in VA and VFs, and only 1.3% of patients had deterioration of the pre-existing VA. In 5.2% of patients, the visual parameters remained static as compared with the baseline status [Table 3].
Postoperative visual outcome was directly proportional to the preoperative VA; patients with better preoperative VA had a better visual outcome. A total of 82% of patients with VA 6/6 to 6/24 prior to surgery had improvement in visual status postoperatively, whereas only 10% of patients with perception of light, and 5%, without perception of light had improvement in visual status postoperatively [Table 4] and [Table 5].
There was no difference in visual outcome in the prospective and retrospective groups. Age, diabetes, and hypertension were not affecting the visual outcome in the prospective group.
There was a significant correlation between the postoperative visual outcome and preoperative duration of visual symptoms (P < 0.05). More number of patients had an earlier recovery of vision if the duration of symptoms was shorter (r = 0.64, P = 0.01). However, the ultimate outcome was good at the end of 1 year irrespective of the duration of symptoms. In the immediate postoperative period, 33.2% of patients with visual symptoms for <6 months had improvement in VA, whereas only 1.0% of patients with visual symptoms for 6 months to 1 year, and 1.4% of patients with visual symptoms for >1 year showed improvement. While most of the patients improved after 6 months of surgery, those who underwent surgery at least 1 year after the onset of visual symptoms showed improvement only after 1 year. One of the striking features of our study was that 8 of 156 patients who were PL negative got back serviceable vision. All these patients had giant adenomas, were symptomatic for ≥2 years, and the recovery occurred gradually from 3 months, with maximum amount occurring at 6 months to 1 year [Table 4]b. The improvement in visual outcome was independent of tumor size and optic disc pallor. There was no difference in the visual recovery in patients with a giant adenoma or a multicompartmental disease (P = 0.4). Pre- and postoperative VA and VF were linearly correlated in the majority of instances.
On subgroup analysis, 79 patients had apoplexy (55 clinical, 24 radiological). Of them, 76% had an abnormal vision and 26.5% had no serviceable vision. Similar to the group as a whole, bitemporal hemianopia was the commonest field defect (36.8%) in patients with apoplexy. The mean time lag between clinical presentation and surgery was 4.8 ± 5.7 days (median, 2 days). Following surgery, 73.4% of patients had improvement in vision (P < 0.001) and 31% had normalization of vision. Of the 10 eyes that were not having perception of light, 9 improved, and the improvement continued till the end of 1 year.
Visual symptom is one of the major presenting manifestations of a pituitary macroadenoma causing considerable burden to patients and their families., Bitemporal hemianopia was the commonest type of VF defect. Both TSS and transcranial surgery are effective in decompressing the visual pathway. Long-standing symptoms (for >1 year) and complete optic atrophy indicate a poor visual outcome.
Before the 1970s, VF defects were reported in up to 90% of patients with pituitary adenomas of all types. Owing to increasing awareness and availability of modern imaging techniques, the incidence is coming down in the Western literature. In a single-center study by Elgamal et al., the incidence of VF defect was 44%. However, in our study, we found it in 70.5% of patients. Upward-growing pituitary tumors supposedly impinge on the anterior notch of the chiasm at its lowest-lying aspect and produce upper-temporal bitemporal quadrantanopia progressing to hemianopia. Bitemporal hemianopia is considered the “hallmark” of pituitary tumors. In our study too, majority of the subjects (47.6%) presented with bitemporal hemianopia as the commonest VF defect, followed by complete blindness in the left eye in 8.9%, and the right eye in 9.5% of patients. However, the absence of significant difference in the frequency of upper temporal and lower temporal quadrant defects in such a large series, indicates the compression in many patients not to be from the undersurface. This kind of similar frequency of upper and lower quadrant defects has been reported in craniopharyngiomas also. The presence of all-four-quadrant defects was the rarest type (0.6%) of VF impairment.
The preoperative VA was highly predictive of outcome. Cohen et al., in his study on visual recovery after trans-sphenoidal removal of a pituitary adenoma, showed that of the eyes that had preoperative VA better than 20/100, 89% were normal or had improved postoperatively, whereas only 62% of the eyes with VA equal to or worse than 20/200 showed improvement. Our study has also shown similar result. Patients with better preoperative VA had a better visual outcome. Eighty-two percent of patients with an acuity of 6/12 to 6/24 had improvement in visual status, whereas only 10% of patients with no perception of light improved.
The fundoscopic sign of long-standing chiasmal compression from a pituitary macroadenoma is primary optic atrophy secondary to retrograde axonal degeneration. Long-standing compression by a pituitary macroadenoma leads to optic atrophy. In a study by Dhasmana et al., optic atrophy was seen clearly in 17% of patients with pituitary adenomas, and all of them had significantly affected vision (VA, 20/100 or worse). In our study too, similar percentage of patients presented with optic atrophy, and most of the patients had a poor VA ranging from 6/36- 6/60 to no light perception.
As described previously, visual improvement occurred in three phases, with the earliest improvement taking place in the immediate postoperative period, followed by improvement over weeks after surgery., Most of the improvement in our patients occurred after 6 months of surgery. It has been postulated that the initial improvement in vision is the result of recovery of nerve conduction. After 1 year, there was not much improvement. It could be due to remyelination of decompressed optic pathways. An inverse correlation was found between the preoperative duration of symptoms and rapidity of postoperative VF recovery, signifying the importance of early surgical intervention.,,, While around 90% of patients who underwent an early surgery started improving after 6 months, 88% of those who had symptom onset-surgery interval longer than a year showed improvement only after an year. This suggests that future studies on visual outcome should not have a follow-up of less than 6 months; in fact, it is preferable to have a follow up of at least an year. Similar to the study by Miller et al., we observed that an increasing age has a bearing on the final visual outcome when all other parameters, such as the type and size of tumor and duration of symptoms are matched, and those subjects with <6 months of visual symptoms showed maximum improvement. This could be due to decreased capability of neuronal regeneration with advancing age. The other factors influencing poor visual outcome were size of the tumor, degree of suprasellar extension (tumor height, >2 cm), and severity of involvement at the time of presentation. All, except one patient, with absent perception of light had a giant pituitary adenoma, and the duration of symptoms was >1 year. Visual loss is not commonly associated with functional pituitary tumors because they present earlier with symptoms of pituitary hyperfunctioning. We, however, encountered 23% of patients with acromegaly who presented for the first time with visual loss due to a large pituitary adenoma. The reason for such a phenomenon is ignorance of the disease and late reporting in our social setup.
In patients with apoplexy, 90% of blind eyes showed a remarkable improvement in vision if surgical decompression of the optic apparatus was undertaken early. This is similar to the findings of the previous studies.,, Impairment of neurological function appears irreversible when central nervous system is compressed to the point of total loss of function. Optic chiasma may be one exception. Therefore, an increased awareness and a timely neurosurgical intervention can lead to a rewarding visual outcome, and the vision loss should not be treated expectantly.
The relevance of microscopic transphenoidal surgery in the current endoscopic era lies not only on the weight of experience but also on the evidence that it may be more than sufficient for the majority of pituitary adenomas without cavernous sinus extension. Though there were limitations of our study in eliciting differences in visual outcome especially with respect to the extent of resection, mere decompression of the tight compartment might have led to visual recovery. The speed of recovery proportional to the duration of visual deficit noted in our study need not be overemphasized. While the visual deficit present for 6 months takes 6 months to improve, the deficit present for 1 year takes at least an year to show improvement.
To conclude, final visual outcome appears to be dependent on the suprasellar extension of the tumor, but it was good in the long run irrespective of the duration of symptoms. The speed of visual recovery was proportional to the duration of visual deficits. While improvement is to be expected gradually by 6 months, clinicians should be even more patient when dealing with subjects with a longer onset-surgery interval. However, long-standing blindness lasting for even up to an year, should not lead any one towards surgical nihilism; in fact, this subgroup of patients should also be subjected to surgery as early as is feasible.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5]