To evaluate the long-term results of autologous submandibular gland transplantation in eyes with cicatrizing conjunctivitis and to determine biomechanical and biochemical features of the resulting salivary tear film.
Prospective, observational case series.
Fifteen eyes with cicatrizing conjunctivitis with a viable autologous submandibular gland transplantation were compared with 10 eyes with cicatrizing conjunctivitis and a failed submandibular gland transplantation or no submandibular gland transplantation. Best-corrected visual acuity, frequency of tear substitute instillation, severity of dry eye discomfort, lid margin erythema, conjunctival hyperemia, corneal epithelial edema, tear film break-up time, Schirmer test results, and corneal fluorescein and conjunctival Rose Bengal staining were evaluated. In a subgroup central corneal thickness and sensitivity, corneal epithelial barrier function, conjunctival and lid margin flora, and conjunctival impression cytologic analysis results were evaluated. In 3 patients, preoperative and postoperative tear samples were analyzed for viscosity, surface tension, and presence of mucins.
Submandibular gland autotransplantation resulted in long-term improvement of subjective, objective, and some ocular surface parameters. Salivary mucins were detectable in salivary tears after submandibular gland transplantation. The viscosity of salivary tears was more similar to normal saliva and the surface tension was intermediate between the 2 original secretions.
Submandibular gland autotransplantation provides long-term relief from pain and reduces the need for frequent installation of lubricants.
Dry eye syndrome refers to a spectrum of ocular surface diseases with diverse and multiple causes. Although in most cases it is the result of excessive loss, anomalous composition, or maldistribution of the tear film caused by ineffective lid blinking and closure, in some patients, insufficient tear supply also can be the predominant pathomechanism. Pharmaceutical tear substitutes can control the symptoms of most patients and can prevent ocular surface damage. When tear deficiency is caused by a cicatrizing conjunctivitis, as in autoimmunologic disorders such as Stevens-Johnson syndrome or ocular cicatricial mucous membrane pemphigoid, the tear deficiency can be absolute and can lead to blindness.
An autologous submandibular gland can be used to provide a more continuous and seromucinous, and hence more tear-like, tear replacement. In this surgical procedure, a submandibular gland is transplanted into the temporal fossa and anastomosed to the superficial temporal artery and vein, and its excretory duct is implanted into the upper conjunctival fornix. In 1998, we reported significant relief of subjective symptoms, reduction of tear application, and an increase of baseline lubrication over the first postoperative year in 26 eyes with a submandibular gland transplant. A detailed analysis of the resulting ocular lubrication confirmed its salivary character with a specific low osmolality likely to be responsible for the corneal epithelial microcystic edema observed in a number of patients with excessive salivary tearing. Although in vitro tests showed less toxicity of submandibular gland saliva, using a cell culture model of primary human corneal epithelial cells rather than unpreserved pharmaceutical tear substitutes, the long-term effect of submandibular gland transplantation on the signs and symptoms of absolute dry eyes remained unclear. Herein, we present the long-term results of this procedure in absolute dry eyes secondary to cicatrizing conjunctivitis.
Between 1994 and the beginning of 2005, 44 eyes (35 patients) underwent an autologous submandibular gland transplantation. All patients had severe to absolute aqueous tear deficiency. Three months after surgery and during the subsequent follow-up, 35 of the 44 transplants were classified as viable. Viability was judged either on the basis of an increase of Schirmer test results or by means of technetium-99m–pertechnetate scintillography. Loss of graft viability was thought to be predominantly the result of problems with the vascular anastomosis. However in 2 patients, the secretory duct became obstructed, leading to secondary graft atrophy, and in 1 patient with a systemic polyautoimmune endocrinopathy, antibodies against salivary gland tissue developed. No difference in the rate of graft survival were noted between patients undergoing surgery early or later in the study.
From this total number of submandibular gland transplantation patients, we prospectively followed up a clinical cohort of 13 patients with 15 viable submandibular gland transplants who underwent surgery for severe aqueous tear deficiency resulting from cicatrizing conjunctivitis ( Table 1 ). This transplantation group was compared with a control group of 9 patients (10 eyes) in whom either submandibular gland transplantation had been performed for study purposes in patients with bilateral symmetric disease only in 1 eye (n = 7) or in whom the graft had failed after 3 months (n = 3). The reason for graft failure could not be identified, although it is likely (but difficult to prove) to be the result of ischemia in the early postoperative phase. The mean follow-up in the transplantation group was of 41 ± 9.5 months (range, 7 to 68 months) and that in the control group was 45.7 ± 14.5 months (20 to 61 months); the preoperative medical history of severe dry eye symptoms spanned from 4 to 24 years. Both lachrymal puncti of all eyes in the transplantation and in the control groups were occluded either surgically or as a result of the underlying disease. During this period, all patients had applied at least 4 different preparations of pharmaceutical tear substitutes, and among those, at least 1 preparation was preservative free. The patients in the transplantation group lost 3 eyes because of severe ocular surface disease, including corneal perforation, before they underwent a submandibular gland transplantation in the fellow eye. Before transplantation, 14 of the 15 eyes in the transplantation group and 5 of the 10 eyes in the control group had undergone surgery to treat keratoconjunctivitis sicca, to reconstruct the corneal or conjunctival surface, and to correct any obvious eyelid malposition and lid margin pathologic features ( Table 1 ).
|Transplantation Group||Control Group|
|Average age ± SD (minimum to maximum), yrs||43 ± 20.1 (13 to 84)||37 ± 13.2 (17 to 57)|
|No. of patients (male:female)||13 (6:7)||9 (5:4)|
|No. of eyes included||15||10|
|Unicus (only 1 morphologic eye remaining) a||3||0|
|Blind eye b||7||4|
|Ocular cicatricial pemphigoid||2||—|
|Toxic epidermal necrolysis syndrome||3||2|
|Mean duration of keratoconjunctivitis sicca ± SD (minimum to maximum), yrs||11 ± 6 (5 to 24)||13.7 ± 6.5 (4 to 24)|
|Median no. of previous surgical procedures in the evaluated eye (minimum to maximum)||2 (0 to 6)||6 (1 to 11)|
|Mean follow-up ± SD (minimum to maximum), mos||41 ± 19.5 (7 to 68)||45.7 ± 14.5 (20 to 61)|
Surgery was performed by a maxillofacial surgeon (P.S., N.H.) in cooperation with an ophthalmologist (G.G., J.R.O.C.) in 2 different clinical institutions (n = 9 at the University Hospital Lübeck and subsequently n = 6 viable grafts at Moorfields Eye Hospital) under the guidance of 1 senior surgeon (P.S.). All postoperative follow-up examinations were performed by 1 examiner (G.G.). For technical reasons, the full range of examinations and assays could be performed only at Moorfields Eye Hospital. The surgical details have been described previously.
Preoperative and Postoperative in Vivo Assessments
All patients enrolled in the study underwent evaluation of subjective parameters, such as symptoms of discomfort (self-graded on a ordinal scale from 0 = none to 4 = very severe), best-corrected visual acuity, and the following objective parameters: slit-lamp examination to grade lid erythema, conjunctival injection, conjunctival chemosis, inflammatory infiltrate, corneal epithelial edema, corneal pachymetry, Schirmer test results without anesthesia, tear film break-up time, fluorescein staining, Rose Bengal staining, corneal esthesiometry, Schirmer test results with anesthesia, conjunctival and lid margin smear, fluorometry, and impression cytologic analysis. The full examination was carried out before transplantation surgery and was repeated at 1 week (n = 15 eyes), 3 months (n = 15 eyes), 1 year (n = 14 eyes), 2.5 years (n = 11 eyes), and 5 years (n = 9 eyes) after surgery.
The submandibular gland is completely denervated during transplantation and therefore no reflex secretion is expected. However, the flow of salivary tears can be increased by physical activity or local hyperthermia, which suggests that the basal secretion rate is stimulated by an increased vascular perfusion. To evaluate this hypothesis, the Schirmer test after topical anesthesia was repeated in a subset of 6 patients immediately after the increased physical activity of climbing 4 flights of stairs as fast as possible.
In another subset of 7 patients in each group, the peripheral corneal thickness at the 6-o’clock position and near the tear meniscus, the central corneal sensitivity, the barrier function of the corneal epithelium, and the type and quantity of the conjunctival and lid margin bacterial flora were determined before surgery and at 1 year after surgery. Pachymetry was measured with an optical pachymeter (Haag-Streit AG, Koeniz, Switzerland). Corneal sensitivity was measured with a Cochet-Bonnet aesthesiometer (Western Ophthalmics Corporation, Lynnwood, Washington, USA). Corneal epithelial permeability to fluorescein was determined with a fluorometer. At each time point, first corneal autofluorescence was determined by averaging the data of 4 fluorometric measurements before and 30 minutes after a 3-minute eye bath in 0.25% sodium fluorescein at room temperature, and the average value minus the autofluorescent background was calculated.
Ex Vivo Assessments
In a subset of patients, microbiologic (n = 7 in the transplantation and in the control groups) and histopathologic (transplantation group, n = 10; control group, n = 4) examinations were performed. A smear was obtained from the lower conjunctival fornix and the lid margin, and the number of species growing from these and the severity of growth were graded from 0 (none) to 3 (heavy). Conjunctival impression cytologic analysis was performed after topical anesthesia with mounted sheets of cellulose acetate filter paper (Millicell CM; 0.4 μm; Millipore Corporate, Billerica, Massachusetts, USA) in the 12-o’clock position 3 to 5 mm away from the limbus, stained with periodic acid–Schiff reagent and graded according to a published scoring system.
Samples of normal tears (n = 3), stimulated (by climbing 4 flights of stairs as fast as possible) salivary tears (n = 3), and stimulated normal submandibular glad saliva (n = 6, sublingual lemon juice application) were collected from patients before and after transplantation by means of a micropipette and were stored frozen at −70 C until they were analyzed for viscosity, surface tension, and presence of mucins. Viscosity was measured using the Contraves Low Shear 30 rheometer (Contraves AG, Zürich, Switzerland) over the range of shear rates of 1.7 to 128.5 sec −1 , according to the method published by Tiffany. Surface tension was measured in a 1-μL sample at room temperature by a glass-capillary microtechnique using a sensitive electronic manometer (model MP6KMD; Air Instrument Resources Ltd, Chalgrove, Oxfordshire, United Kingdom). Mucin detection was by 2 different techniques: size exclusion chromatography (TSK-3000 gel) and ultraviolet detection at 254 nm with comparison with standard bovine submaxillary mucin and electrophoresis on precast 4% to 12% gradient Bis Tris gels and periodic acid–Schiff staining similarly to a the description from a previously described study.
Differences in the results obtained within the same group of patients during the follow-up were evaluated using the Wilcoxon test for related samples. The Mann–Whitney U test for unrelated samples was used to compare the results of the transplantation group with those of the control group at each single observation point; P values ≤ .05 were considered significant.
Primary Surgical Success and Postoperative Complications
Early and late postoperative complications are listed in Table 2 . One patient (2 eyes) died in the second postoperative year of cardiovascular disease.
|Early Complications||Late Complications|
|Graft necrosis (because of vascular complication or infection)||6||—|
|Disinsertion of the excretory duct||1||—|
|Stenosis of the excretory duct||1||—|
|Temporal limitation of jaw movements||1||—|
|Partial facial palsy||1||—|
|Obstruction of the excretory duct||—||1|
|Infection of the excretory duct ( Staphylococcus aureus )||—||1|
|Microbial keratitis ( Pseudomonas aeruginosa )||—||1|
|Salivary lubrication resulted in burning sensation requiring transplant removal||—||2|
|Microcystic edema resulting from excessive salivary epiphora requiring transplant reduction||—||6|
Before surgery, no difference between the transplantation group and the control group in visual acuity ( P = .9), the frequency of tear substitute instillation ( P = .4), and the severity of symptoms ( P = .13) was found.
The best-corrected visual acuity in 7 of the 15 eyes in the transplantation group and in 4 of the 10 eyes in the control group was 20/400 or worse on the logarithmic Snellen acuity chart. Mean logarithm of the minimal angle of resolution visual acuity was 1.5 ± 1.1 (equivalent to 20/2000) in the transplantation group and 1.3 ± 0.8 (equivalent to 20/400) in the control eyes. No statistically significant difference between the groups was found until the last follow-up (minimum P = .2; Figure 1 ).
The transplantation group’s frequency of pharmaceutical tear substitute application was reduced significantly 1 year after transplantation from 59 ± 47 applications in 14 hours to 2 ± 4 applications ( P = .008), and this improvement was maintained to the end of the follow-up. Tear substitute application was discontinued completely in 5 eyes and was reduced in all, with the median dropping from 56 applications per day before surgery to 1 daily after surgery. In the control group, there was also some reduction, from 60 ± 69 applications before surgery to 44 ± 26 applications (median, 44 applications) at 5 years, but the reduction was not statistically significant ( P = .5; Figure 1 ).
The severity of symptoms in the transplantation group improved from a preoperative median of very severe, to mild at 3 months to 2.5 years and moderate at 5 years, whereas it remained severe throughout the period of observation in the control group. The mean severity score remained unchanged for the control group, whereas the transplantation group improved progressively and significantly until the end of the follow up (before surgery vs year 5, P = .02). The transplantation group results were significantly better than those of the control group (year 1, P = .03; year 2.5, P = .01; year 5, P = .016; Figure 1 ).
The findings for the objective parameters were slightly more severe before surgery in the transplantation group than in the control group, but this was only significant for fluorescein-positive epitheliopathy ( P = .046).
The average preoperative Schirmer test without topical anesthesia was 1 ± 1.8 mm for the transplantation group and 1.2 ± 2.1 mm for the control group ( Figure 2 ). Ocular lubrication improved significantly in the transplantation group from 1 week through to year 5 (before surgery vs year 5, P = .02) and was significantly better than the control group from 3 months after surgery to the end of the follow-up. No patient reported gustatory reflex tearing. The basal tear secretion (Schirmer test with topical anesthesia) showed similar results ( Figure 2 ). Salivary tear production was significantly stimulated by physical activity (Schirmer test on activity) at 1 year from 12 ± 12 mm to 43 ± 26 mm ( P = .017; Figure 2 ). The fluorescein tear film break-up time showed a lasting improvement in the transplantation group over the follow-up from 2.7 ± 4.9 seconds before surgery to 16 ± 11.2 seconds 5 years after surgery. The improvement in the transplantation group was statistically significant from 1 week to 5 years after surgery ( P = .016), whereas the control group remained unchanged throughout the follow-up ( Figure 2 ).
Fluorescein staining of the corneal epithelium was significantly reduced in the transplantation group at 1 and 2.5 years after surgery (year 1, P = .02; year 2.5, P = .03), but was found to have deteriorated again at the 5-year follow-up ( P = .042). The staining score in the control group remained unchanged throughout the follow-up period. A significant difference between the 2 groups was detected only before surgery ( Figure 3 ). Peripheral corneal pachymetry as well as central corneal sensitivity in the transplantation group remained unchanged between the preoperative and postoperative evaluation at 1 year.
Temporary but severe microcystic corneal epithelial edema requiring reduction of the salivary gland transplant was observed in 6 of 15 eyes in the transplantation group. This superficial corneal edema was diffuse and was noted by the patients as a temporary reduction of visual acuity. It disappeared if salivary tearing was reduced either spontaneously, by means of periglandular botulinum toxin injection, or by surgically reducing the graft volume. Despite a significant continuous increase in surface lubrication, the difference in corneal epithelial barrier function remained unchanged in the transplantation or the control group from 3 months to 1 year after surgery ( Table 3 ).
|Before Surgery||1 Year after Surgery|
|Transplantation group||768 ± 83||792 ± 57|
|Control group||752 ± 106||808 ± 35|
|Transplantation group||5.2 ± 1.1||4.9 ± 1.1|
|Control group||5.1 ± 1.2||4.5 ± 1.5|
|Transplantation group||No data||0.58 ± 0.28|
|Control group||No data||0.28 ± 0.13|