Introduction

Diagnostic vitrectomy is a safe and efficient procedure for diagnosing intraocular inflammation of the posterior segment associated with infectious, neoplastic and autoimmune diseases [1, 2].

In 1981, Carrol and Franklin reported the convenience of diagnostic vitrectomy for uveitis of unknown etiology [3]. Diagnostic vitrectomy can be performed alone with limited removal of vitreous body (biopsy) or as a preliminary part of more complex surgical procedure (therapeutic vitrectomy) that may include specific maneuvers for different purposes (i.e., retinal detachment repair).

Innovations of laboratory methods in the past decades have widened the indications of diagnostic vitrectomy. Cytological and cultural analyses are considered the main methods of processing the vitreous samples, and their effectiveness has been well demonstrated [4,5,6]. In addition, polymerase chain reaction (PCR) methods may be very useful in establishing the pathogen when only modest quantities of ocular fluids are available for examination [7].

Several techniques have been described to sample vitreous fluid, including 20-, 23- and 25-gauge pars plana vitrectomy. The recently introduced 27-gauge cutters represent a valid alternative for diagnostic vitrectomy [8].

One-port and three-port pars plana diagnostic vitrectomy have been described in the literature [9,10,11]. An ordinary three-port vitrectomy allows the surgeon to safely collect the vitreous sample and proceed with complete removal of vitreous body plus other surgical maneuvers. If it is needed to obtain a vitreous specimen for diagnostic purposes only, and a therapeutic approach is not required, then a one-port mechanized vitrectomy could be preferred.

Although both procedures may present pros and cons, the safety and efficacy of one-port diagnostic vitrectomy have never been compared to the three-port approach. The aim of this study was to compare the safety and efficacy of one-port to three-port vitrectomy in undiagnosed cases of intraocular inflammation.

Methods

After the approval of the Provincial Ethical Committee of Reggio Emilia, we retrospectively searched our electronic register for patients who, after signing the informed consent, underwent vitrectomy for diagnostic purposes from January 2011 to April 2017 in the tertiary referral center for uveitis at Santa Maria Nuova Hospital of Reggio Emilia, Italy [12]. Inclusion criteria were: ocular inflammation of unknown reason involving the vitreous, the pars plana and/or the retina; a preoperative diagnosis of posterior uveitis or panuveitis of unknown etiology; and laboratory analysis and imaging studies non-diagnostic for known systemic diseases. We excluded patients: with postoperative endophthalmitis; with systemic malignancies; and if the surgical time was not recorded.

The following data were collected from clinical charts: patient’s age and gender; laterality; preoperative best-corrected visual acuity (BCVA); BCVA at six months (± 1 month); last available postoperative BCVA; suspected preoperative diagnosis; final postoperative diagnosis; duration of surgical procedure; variants in vitreous surgery technique (size of trocar cannulas, type of anesthesia, associated concomitant cataract surgery, tamponades, sutures for sclerotomies); and intraoperative and postoperative complications. Preoperative and postoperative clinical examinations included measurement of BCVA, measurement of intraocular pressure (IOP), examination at slit lamp and indirect binocular ophthalmoscopy.

The consecutive procedures were performed by five surgeons using the Constellation Vision System (Alcon Laboratories Inc., Fort Worth, USA). The type of surgery (one-port or three-port vitrectomy) was selected by the operator taking into consideration personal preferences, the ocular and systemic condition of patients and, in particular, the presence of ocular comorbidities requiring complex vitreoretinal maneuvers.

Three-port pars plana transconjunctival vitrectomies were performed in standard fashion. After positioning the three trocar cannulas, infusion was not open and an undiluted vitreous sample (1–2 ml) was obtained with the vitrectomy cutter and manually aspirated into a connected syringe with the help of an assistant. At the same time, the surgeon's second hand was pushing on the sclera with the help of a q-tip in order to maintain an adequate IOP. The infusion was subsequently activated, and the surgical procedure was carried on with the removal of the remaining vitreous. According to the underlying retinal condition, the operation was then completed with the appropriate surgical maneuvers as necessary (endolaser, fluids exchanges, etc.).

One-port vitrectomies started with the positioning of a single trocar cannula in the pars plana 3.5 mm from limbus superotemporally (Fig. 1). The vitrectomy cutter connected to a syringe was carefully introduced in the vitreous cavity and held into view through the pupil. The cutter was then activated, and an undiluted vitreous sample (1–2 ml) was manually aspirated by an assistant. At the same time, the surgeon's second hand was pushing on the sclera with the help of a q-tip in order to maintain an adequate IOP. At the end of the vitreous biopsy, the cutter was carefully extracted, and balanced salt solution was gently injected into the eye throughout the trocar cannula in order to refill the globe. Retinal periphery was then checked for retinal breaks with the help of indirect ophthalmoscopy with scleral depression.

Fig. 1
figure 1

One-port vitrectomy in a phakic eye. The surgeon's right hand is holding a 27-gauge cutter that is visible behind the crystalline lens. The left hand is pushing on the sclera with a q-tip to maintain an adequate IOP

Full size image

Vitreous specimens were sent for laboratory analysis and processed for microbiological, cytological, molecular and immunohistochemical examination. Microbiological testing was made by cultural methods to identify the more common bacteria and fungi. PCR protocols targeting polymerase genes of VZV, HSV, CMV and toxoplasma were used. If the diagnosis of lymphoma or malignancy was suspected, the undiluted vitreous sample was utilized to prepare one air-dried Diff-Quick stained smear; the remaining part of the sample was processed by PCR, looking for T cell receptor and heavy-chain immunoglobulin rearrangement. We used a patient-tailored diagnostic approach depending on the suspected diagnosis; we considered confirmed the clinically suspected diagnosis hypothesized before surgery if the laboratory analysis resulted positive.

We evaluated the efficacy of the interventions, comparing the diagnostic pars plana vitrectomies performed with the three-port vs. the one-port approach. The primary outcome was surgical time. Secondary outcomes were: improvement of BCVA after diagnostic vitrectomy at six months and at the last follow-up; positivity of laboratory test; confirmation of the preoperative diagnostic hypothesis after laboratory tests; and postoperative complications.

BCVA was registered as Snellen visual acuity and converted into logarithm of minimum resolution (LogMAR) units for statistical analysis. The vision of finger counting was defined as 2.0 LogMAR, and hand movement was defined as 3.0 LogMAR according to Holladay [13]. Patients who presented with no light perception or light perception (LP) vision before surgery were not taken into account in the assessment of BCVA, as recommended by Holladay.

Values were expressed as mean ± standard deviation. Statistical analysis was conducted using STATA software (v.12.0, STATA Corp., USA). Two-tailed paired and unpaired Student’s t test were performed to compare continuous variables. Chi-square test was used to compare categorical variables. A p value < 0.05 was considered statistically significant.

Results

Diagnostic vitrectomy was performed on 80 eyes that met the inclusion criteria. Forty eyes underwent three-port diagnostic vitrectomy, and forty eyes underwent one-port diagnostic vitrectomy. Baseline characteristics were similar in both groups (Table 1).

Table 1 Demographic and baseline ocular characteristics of one-port and three-port vitrectomy groups
Full size table

Mean duration of surgery was significantly longer in the three-port vitrectomy group when compared to the one-port vitrectomy group (49 ± 30 vs. 15 ± 8 min; p = 0.0001).

In the one-port group, 23 vitrectomies were performed using a 25-gauge system (57%), 14 using a 23-gauge system (35%) and 3 using a 27-gauge system (8%); scleral wound was sutured in 37 eyes (93%); and most of the procedures (95%) were carried out under topical anesthesia with 0.4% benoxinate, the rest under retrobulbar anesthesia.

In the three-port group, 11 vitrectomies were performed using a 25-gauge system (28%) and 24 using a 23-gauge system (60%); intraocular tamponade was used in 11 eyes (28%; sulfur hexafluoride in all cases); scleral wounds were sutured in all eyes; and all surgeries in this group were completed under retrobulbar anesthesia with a 50/50 mixture of 2% lidocaine and 0.5% bupivacaine.

In the three-port group, additional surgical maneuvers were performed in 21 patients. Three eyes required retinal detachment repair, two sustained peeling of epiretinal membrane, five eyes necessitated endolaser retinopexy for peripheral retinal breaks, six eyes received intravitreal antibiotics or antifungals at the end of surgery, and five eyes required cataract surgery at the time of vitrectomy.

Following the criteria described above [13], 28 eyes for the one-port group and 34 eyes for three-port group were included in the evaluation of BCVA. Pre- or postoperative BCVA values were missing (n = 12) or were LP (n = 6) in the remaining cases.

Mean preoperative BCVA in the one-port diagnostic vitrectomy group was 1.31 ± 0.96 LogMAR (Snellen fraction 20/400; range: counting fingers to 20/20). Mean postoperative BCVA increased to 0.57 ± 0.59 LogMAR (Snellen fraction 20/80; range: counting fingers to 20/20), and the difference was statistically significant when compared to preoperative values (n = 28; p = 0.0009). After surgery, BCVA improved in 20 eyes (72%), with 16 eyes ameliorating by three or more ETDRS lines (0.3 LogMAR). Vision remained stable in two eyes (7%). Visual acuity decreased in six eyes (21%), with four eyes worsening by three or more ETDRS lines.

Mean preoperative BCVA in the three-port diagnostic vitrectomy group was 0.98 ± 0.76 LogMAR (Snellen fraction 20/200; range: counting fingers to 20/20). Mean postoperative BCVA improved to 0.51 ± 0.76 LogMAR (Snellen fraction 20/60; range: counting fingers to 20/20), and the difference was statistically significant when compared to preoperative values (n = 34; p = 0.0005). After surgery, BCVA improved in 27 eyes (79%), with 19 eyes ameliorating by three or more ETDRS lines (0.3 LogMAR). Vision remained stable in three eyes (9%). Visual acuity decreased in four eyes (12%), with three eyes worsening by three or more ETDRS lines.

In terms of postoperative mean BCVA, there were no significative differences between one-port and three-port vitrectomy groups (0.57 ± 0.59 vs. 0.51 ± 0.76 LogMAR; p = 0.72).

One-port vitrectomy leads to a precise etiologic diagnosis in 78% of the cases, while the three-port vitrectomy in 28%. Among cases with a non-diagnostic laboratory result, we diagnosed one idiopathic posterior uveitis in the one-port group and eight idiopathic posterior uveitis/panuveitis in the three-port group. Taking into account this idiopathic uveitis, a final diagnosis was reached in 80% and 48% of the cases, respectively.

The incidence of retinal detachment after surgery was the same in both groups (three cases in the one-port and three cases in the three-port group). Mild vitreous hemorrhage was recorded after surgery in one eye of the one-port group and spontaneously resolved within a month.

Discussion

A delay in diagnosing the etiology of posterior segment inflammation can cause permanent visual loss, as reported in several studies [14]. A prompt diagnosis is mandatory before irreversible damage is established. Diagnostic vitrectomy plays a crucial role in the assessment of intraocular inflammation of unknown origin and in the therapeutic decisions.

The aim of this study was to compare the efficacy and safety of one-port vs. three-port diagnostic vitrectomy. The primary outcome was to compare the mean operative time of the two procedures. Secondary outcomes were visual acuity improvement after surgery, incidence of postoperative complications and diagnostic success rate.

Mean operative time was considerably shorter in the one-port group when compared to the three-port one (49 ± 30 vs. 15 ± 8 min; p = 0.0001). It should be considered that all procedures in the one-port group were performed for diagnostic purposes only (vitreous biopsy), while in the three-port group diagnostic vitrectomy was performed in association with additional retinal procedures in most cases. Surgeons were free to perform one-port or three-port diagnostic vitrectomy according to personal preferences. This may have generated a bias, since it is possible that simpler cases were selected for one-port vitrectomy while three-port vitrectomies were reserved for more complex cases. Nevertheless, this study represents real-life conditions, where ophthalmologists are commonly called to choose between different surgical approaches based on patients' clinical conditions.

By one-port diagnostic vitrectomy, we were able to collect a vitreous specimen that led to a laboratory report with a precise etiology in 78% of the cases (i.e., lymphoma or infection). The three-port approach led to a laboratory diagnosis in 28% of eyes. Among eyes with a non-diagnostic test, with the help of other clinical details or attempted therapy, after surgery we diagnosed one case of idiopathic posterior uveitis in the one-port group and eight cases of idiopathic posterior uveitis/panuveitis in the three-port group. Considering these idiopathic forms of uveitis, a final diagnosis was reached in 80% and 48% of the cases, respectively. This may be due to a bias of selection, since cases with preoperative characteristics carrying a higher clinical suspicion for a certain etiology may have been selected for a one-port vitrectomy (i.e., vitreoretinal lymphoma).

After surgery, both groups had a similar improvement in mean BVCA from baseline. We may assume that the removal of central vitreous performed by one-port vitrectomy is as effective as the three-port one in removing vitreous haze and improving visual acuity.

Postoperative complications were similar in both groups. One-port vitrectomy did not increase the risk of retinal detachment since that adverse event occurred three times in both groups (7.5%). The overall incidence of postoperative retinal detachment is higher than reported in other general series on retinal detachment after vitrectomy [15], but this is possibly due to the severity of the underlying ocular conditions [16].

A one-port vitreous biopsy may be performed either by dry needle aspiration (“tap”) or using a mechanical cutter [17,18,19]. It has been proven that the cytological characteristics of specimens collected by needle aspiration or mechanical vitrectomy (with a maximum speed of 600 cuts per minute) are equivalent [20, 21]. Needle aspiration is a very simple technique that can be performed in the office under topical anesthesia, usually in cases of suspected postoperative endophthalmitis. It permits to collect a very small amount of vitreous sample and it does not improve vision since it does not remove an adequate amount of central vitreous. Needle tap biopsy demonstrated a similar diagnostic success of mechanized vitrectomy in cases of endophthalmitis [11]. Our study did not include cases of postoperative endophthalmitis but complex, undiagnosed cases of posterior segment inflammation. One of the benefits of one-port diagnostic vitrectomy, when compared to the tap, is the larger amount of undiluted vitreous available for laboratory analyses. Moreover, compared to the needle aspiration technique, one-port diagnostic vitrectomy often allows the removal of the cloudy central vitreous with benefit for vision.

This was a pilot study and has several limitations: retrospective design; bias of selection (simpler cases may have been selected for one-port vitrectomy); absence of a third study arm representing vitreous tap by needle aspiration; and the presence of multiple surgeons with slightly different techniques.

Our aim was to demonstrate that the three-port ordinary approach for diagnostic vitrectomy may be limited to those case where additional retinal maneuvers are needed. In simpler situations, one-port diagnostic vitrectomy may be as efficient and as safe as the three-port approach. One-port diagnostic vitrectomy warrants reduced surgical time and shorter surgical procedures benefit both the patients and the hospitals, with increased comfort and reduction of costs. It should be emphasized that almost all one-port vitrectomies were performed under topical anesthesia, setting aside the risks of peri/retrobulbar anesthesia.

Our work is a pilot study aiming to stimulate scientific discussion and has numerous limitations. Our purpose was to support the hypothesis that one-port diagnostic vitrectomy might be considered as the first step in unknown cases of posterior ocular inflammation or vitritis. The standard three-port approach could be reserved for complex cases where both diagnostic and therapeutic vitrectomies are needed. Large, multicenter randomized controlled trials are warranted to confirm the efficacy and safety of one-port diagnostic vitrectomy.