Patient Safety and Risk Management in an Accumulation of Postoperative Endophthalmitis Cases after Vitrectomy in a University Eye Clinic

Abstract

Purpose To describe the risk management at a university eye hospital after two outbreaks of nosocomial endophthalmitis cases after pars plana vitrectomy.

Methods In two series of postoperative endophthalmitis cases after in-house vitrectomy, the basic workflows in direct patient care were evaluated with regard to patient safety. Hygienic microbiological environmental examinations were performed on relevant materials and surfaces. In particular, the direct surgical utensils were inspected with regard to possible bacterial colonisation.

Results Pathogens (Staphylococcus aureus) were detected in 2 of 7 endophthalmitis cases. The S. aureus strains showed no clonality. The procedures were 23 G and 25 G vitrectomies for retinal detachment (3× rhegmatogenous, 1× PVR), subretinal macular hemorrhage (1×) and vitreous haemorrhage for proliferative retinopathy (2×). The duration of surgery was between 20 min and 65 min; the time between initial vitrectomy and the surgery for endophthalmitis was between 2 and 5 days (mean 3.6 days). A silicone oil filling was instilled once during the first operation and otherwise the eye was tamponaded with gas (4×) or air (2×). The surgical teams were heterogeneous; n = 5 surgeons were involved and the initial procedures took place in n = 4 different operating theatres. In all cases, general anaesthesia was applied (6× laryngeal mask, 1× endotracheal intubation). No definitive source of infection was found. The interventions with regard to patient safety were therefore aimed at strengthening compliance with existing measures for preventing infection and adapting work processes. In the acute phase, antibiotics were instilled intraoperatively into the anterior chamber after vitrectomy, contrary to the usual in-house procedure. Other types of intraocular surgery were not affected.

Conclusion The accumulation of in-house endophthalmitis cases is a catastrophic event in an eye clinic and stringent risk management is required to identify the causes. Openness and transparency are essential factors for an adequate workup. This manuscript shows what the individual steps could look like and how the results can be dealt with. The problem of not having found a clear point source for the infections is discussed.

Keywords

postoperative endophthalmitis - risk management - patientsʼsafety - nosocomial infection - vitrectomy

Introduction

Postoperative endophthalmitis is a rare and severe complication after intraocular procedures [1]. Cataract surgery and intravitreal operative drug application (IVOM) are typically affected, not least because of the large number of operations [2], [3]. A cluster of cases of endophthalmitis after independently performed uncomplicated operations is a serious situation, especially in an ophthalmological clinic, which may give rise to concerns that hygienic deficiencies could have a decisive effect here. In particular, if only vitrectomies are affected, but not other operations such as cataract surgery, the whole perioperative setup must be thoroughly and promptly reassessed. Our ophthalmology clinic experienced two clusters of this kind. In 2023, there were 4 such cases of endophthalmitis in rapid succession; in the following year (2024) and a significant time after the first episode (8 months), there was another cluster of 3 cases.

The aim of this work is to describe the risk management we carried out to identify the cause of these cases and to describe the measures taken to improve patient safety. A swift, coordinated and comprehensive approach is essential, since of course emergency care of patients in a maximum medical care hospital must also be guaranteed; however, the continued operation of the surgical theatre should be urgently called into question in the circumstances described. Discussion of the possible cause of these case clusters is, lastly, the major focus of this manuscript.

Patients and Methods

In close cooperation with the inhouse unit for Medical Process and Patient Safety and the Hospital Hygiene department of the Institute for Medical Microbiology and Hospital Hygiene, a comprehensive catalogue of examinations and the preparation of appropriate measures was drawn up after the first cases were identified. The fundamental patient processes from admission to discharge following inpatient surgery were investigated, and hygienic microbiological assessment of the hospital environment was carried out on relevant materials and surfaces both in the operating theatre and on the ward. Particular attention was paid to direct surgical utensils, since it is assumed that the pathogen is “flushed” into the eye intraoperatively and thus may also be introduced via sterile material or contaminated single-use items of certain batches. In this respect, our Medical Device Processing Unit (AEMP); (Central Sterile Supply Department) was also involved in the investigation, and an examination of the vitrectomy single-use packs of the corresponding batches was commissioned. The idea that pathogens enter the eye via irrigation fluid was explicitly considered, but no cataract operations performed with the same irrigation fluid were affected. The risk for endophthalmitis after cataract surgery is significantly greater than after vitrectomies [4].

Upon initial presentation or onset, patients with endophthalmitis and its typical symptoms were routinely readied on the ward for immediate vitrectomy under general anaesthesia. The operation was performed by experienced surgeons and included anterior chamber (with irrigation) and posterior segment (with complete vitrectomy) sampling and antibiotic administration (ceftazidime/vancomycin). In addition to standard topical therapy, Decortin H 250 mg/day was administered systemically by infusion. Microbiological tests were performed on each sample to detect the pathogen.

Results

Across 2023 and 2024, the ophthalmological clinic performed approximately 1,500 vitrectomies and approximately 1,000 cataract operations. The cluster of 4 cases of endophthalmitis within only 1 week after in-house vitrectomy was uncommon and triggered the cascade of measures outlined below for the purpose of significant risk management. The main features of the patients and their operations are listed in [Table 1]. These were 23 G and 25 G vitrectomies for ablatio retinae (3× rhegmatogenous, 1× PVR), subretinal macular haemorrhage (1×) and vitreous haemorrhage in proliferative retinopathy (2×). The operating times were between 20 min and 65 min, the interval between the initial operation and the endophthalmitis surgery was between 2 and 5 days (on average 3.6 days); 1× silicone oil was instilled on initial operation and a gas (4×) or air (2×) tamponade was used ([Table 1]). The surgical team was heterogeneous, n = 5 surgeons were involved in the cases and the initial operations took place in n = 4 different operating theatres. The operations were all performed under general anaesthesia (6× laryngeal mask, 1× endotracheal intubation). Postoperative topical drip anaesthesia in the recovery room in case of pain was required in only 2 of the 7 patients.

Table 1 Age, sex, diagnosis, type of surgery for endophthalmitis-inducing surgery (OP1), surgery duration, time to endophthalmitis surgery (in days), and possible re-operations.

	Age/sex	Diagnosis	OP 1	Surgery duration	Days to re-OP	Re-OPs
f: female; m: male; PDR: proliferative diabetic retinopathy; ppV: pars plana vitrectomy; GK bleeding: vitreous haemorrhage; PVR: Proliferative vitroretinopathy
Pat-1	57 f	Retinal detachment	ppV 25 G + gas	20 minutes	5
Pat-2	56 m	PDR/vitreous hemorrhage	ppV 23 G+Cat-OP + air	23 minutes	2
Pat-3	63 m	Retinal detachment	ppV 23 G + oil	63 minutes	4	Yes
Pat-4	83 m	PR/vitreous hemorrhage in ischaemia syndrome	ppV 25 G + Cat-OP + air	48 minutes	3
Pat-5	61 m	Retinal detachment	ppV 23 G + Cat-OP + Gas	23 minutes	3
Pat-6	72 m	Retinal detachment	ppV 25 G + gas	65 minutes	3	Yes
Pat-7	80 f	Macular haemorrhage	ppV 25 G + gas	17 minutes	5

These facts were reported in-house to the Patient Safety and Hospital Hygiene unit and a root cause analysis was initiated immediately with regard to patient processes and the specificities of the procedure itself. In addition, the AEMP, through which the ophthalmological clinic obtains its sterile goods, was informed. No irregularities were found here after appropriate investigation. In addition, the single-use packs for vitrectomy (D-VIT PAK 25 G with valves) were sent to the manufacturer for investigation due to a suspected serious incident related to a medical device that was notified to the Federal Institute for Drugs and Medical Devices (BfArM).

There had been no anonymous reports to the Medical Process and Patient Safety Unit of near-misses or incidents on this topic.

Investigation

With regard to in-house measures, it was agreed that the following investigations would be carried out in the operating theatre and on the ward:

• Creation of a timeline and depiction of the epidemiological situation

• Securing the microbiological isolates from patient specimens

• Mapping the patient journey and the work processes carried out on patients (pre-, peri-/intra- and postoperative)

• Workflow observations in the framework of direct patient care

• Assessment of the hospital environment and training in line with the results of the workflow observations

• Microbiological-hygienic investigations of, for example, eye drops, ointments and other medicinal products applied to the eye

• Close collaboration between the treatment team and the patient safety and hospital hygiene units

• Regular feedback as new results/findings emerge

Microbiological examinations of the anterior chamber and vitreous body punctures revealed positive findings in only 2 of the first 4 cases (and also across all 7 cases). Here, Staphylococcus aureus could be isolated, whereby the molecular biological investigation by means of whole-genomic investigation (including Multilocus Sequence Typing [MLST], Core Genome MLST [cgMLST], spa-type determination [typing of the gene of Staphylococcus aureus-protein-A]) did not demonstrate identical strains. Thus, the strains were a different sequence type (ST) in the MLST (45 and 72) and also different SPA types (t1510 and t4087).

The hygienic-microbiological examinations of the eye ointments, eye drops and NaCl 0.9% used on the ward showed the following results:

• All ointments and drops remained sterile.

• NaCl was contaminated:

  • NaCl 0.9% (sample 1): 2CFU (colony-forming units)/100 ml: 1 CFU Micrococcus luteus and 1 CFU nonfermenter (not further differentiable)

  • NaCl 0.9% (sample 2): 2CFU/100 ml: both CFU coagulase-negative staphylococci

  • NaCl 0.9% (sample 3): 4CFU/100 ml: 3 CFU Moraxella osloensis and 1 CFU nonfermenter (not further differentiable)

In this context, the department received intensive training with particular focus on the aseptic handling of medicinal products. Special attention was paid to hand disinfection before aseptic activities and to the handling of medicinal products such as NaCl (strict use as a single-dose container), ideally to prevent contamination and transmission.

Overall, no evidence of a point source of endophthalmitis clusters was found with the corresponding investigations/workflow observations. Investigations of the packs by the manufacturer after an incident report to the BfArM also revealed no anomalies. The BfArMʼs final report of 18 April 2024 is worded as follows:

Reporting of a suspected serious event or incident (for IVD) related to the device

D-VIT PAK 25 G with valves

Serial/batch number: XXXXXX/XXXXXX

BfArM case number(s): XXXXX/XX

Thank you for your report on the above-mentioned medical device.

The Federal Institute for Drugs and Medical Devices (BfArM) carried out a risk assessment on this notification in accordance with Section 71(2) of the Medizinprodukterecht-Durchführungsgesetz (Medical Device Law Implementing Act) in conjunction with Articles 89(3) of Regulation (EU) 2017/745 and 84(3) of Regulation (EU) 2017/746. The responsible medical device manufacturer was obliged to cooperate in this regard under § 72 Medizinprodukterecht-Durchführungsgesetz in conjunction with Art. 89(1) Regulation (EU) 2017/745 and Art. 84(1) Regulation (EU) 2017/746 and was obliged to carry out the necessary investigations.

These showed that the verification of the history of the equipment and the sterilisation batches did not reveal any anomalies. No sample had the reported malfunction. There were 4 similar complaints in the same hospital and two more in another hospital. For the endophthalmitis problem, the cause was determined as external – handling by the user. Custom Paks are sterilised with ethylene oxide (ETO). The sterilisation cycle was reviewed and released in accordance with the validated specifications. The BfArM currently sees no need for further action and considers the process to be completed. If we become aware of similar situations in the future, your report will be taken into account in the necessary risk assessment procedures.

It therefore appeared likely that during this period either patient-related factors or personnel/process-related factors or the combination of all these factors could have led to a cluster of cases.

Measures

Following an appropriately swift examination of the personnel/process-related factors, the following measures were recommended and largely established (see also [Tables 2] and [3]):

Table 2 Recommendations for improving asepsis in perioperative management of surgical patients.

Date	Recommendation	Status
15 November 2023	Enhanced preoperative skin and mucosal disinfection	Implemented
15 November 2023	Use of nasal ointment (antiseptic or mupirocin) for all vitrectomies on the day of surgery and for the two days after (3/day, possibly the day before surgery if already hospitalised)
15 November 2023	If necessary, antiseptic facial wash preoperatively on the ward
15 November 2023	Enhanced hygienic handling of eye drops and eye ointments (preferably patient-related single-use products)	Implemented
15 November 2023	Improvements in the handling of medical devices and medicinal products used intraoperatively	Implemented
15 November 2023	Training in basic hygiene for ward and operating theatre personnel	Implemented
15 November 2023	Enhanced operating theatre preparation:	Implemented
15 November 2023	Enhanced hand disinfection	Implemented
17 November 2023	Improving the handling of antibiotic solutions used intraocularly in surgery (such as ceftazidime or cefuroxime)
17 November 2023	Enhanced hygienic handling of eye drops and eye ointments
17 November 2023	Improving the handling of intraocular gases	Implemented
17 November 2023	Basic hygiene training in the operating theatre	Implemented
17 November 2023	Basic hygiene training on the ward	Implemented

Table 3 Type and timing of measures to evaluate a possible source of the hospital-acquired endophthalmitis cases.

Date	Type of action	Results
14 November 2023	Screening of the nasopharynx for S. aureus in 3 of the patients	No evidence of S. aureus in the nose
15 November 2023	Workflow observation in the operating theatre
15 November 2023	Workflow observation outpatient eye clinic
15 November 2023	Assessment of the hospital environment
20 November 2023	Result of the molecular biological investigation of the 2 S. aureus strains of 2 of the patients	These are distinct strains
21 November 2023	Personnel training on ward 47
23 November 2023	Personnel training outpatient eye clinic
27 November 2023	Personnel training surgical care and anaesthesia ophthalmological operating theatre
30 November 2023	Personnel training on ward 47
07 December 2023	Personnel training for doctors
15 December 2023	Rounds in ward 47
15 December 2023	Workflow observation in the operating theatre
12 February 2024	Rounds in ward 47
06 March 2024	Observation of hand disinfection in the operating theatre	Compliance: 29%
09 September 2024	Workflow observation in the operating theatre
12 September 2024	Workflow observation in wards 47 and 37 (Same-Day-Surgery (SDS))
13 September 2024	Workflow observation in the operating theatre
25 September 2024	Assessment of the hospital environment	Increased colony count on the processed microscope in the operating theatre.
26 September 2024	Assessment of the hospital environment	S. aureus detected on the handpiece of the intraocular pressure gauge (on ward 47).
07 October 2024	Examination of eye drops and creams on ward and in the operating theatre	No pathogen detection
06 November 2024	Workflow observation ward 47

• Enhanced preoperative skin and mucosal disinfection

• Application concentration of Braunol 1 : 3 to 1 : 6

• Extension of the exposure time according to the manufacturerʼs instructions

• Use of nasal ointment (antiseptic or mupirocin [antibiotic to eliminate staphylococci in the nasal mucosa]) in all patients before pars-plana vitrectomy on the day of operation and the 2 days after (3 times a day, possibly the day before surgery if already hospitalised); (recommended but not implemented)

• If necessary, antiseptic facial wash preoperatively on the ward (recommended but not implemented)

• Enhanced hygienic handling of eye drops and eye ointments (preferably patient-related single-use products)

• Improvements in the handling of medical devices and medicinal products used intra- and postoperatively

  • Medicinal products:

    • Antibiotic prophylaxis of infection at the end of surgery (ceftazidime)

    • Eye drops

    • Eye ointments: Application, for example with spatula/paper strip

    • Hand disinfection before aseptic activity (for example, before application of eye drops/eye ointments)

    • Goal: Preferably, use of patient-related single-dose ampoules. Indeed, use single-dose ampoules only once and discard any remaining contents immediately. Also discard NaCl 0.9% bottles immediately after the single patient-related withdrawal of the liquid (for example for rinses during dressing change); change to smaller containers if necessary. Use a strictly aseptic, contamination-free procedure when using multiple dose containers.

  • Medical devices:

    • Use of single-use products (for example, do not reuse sterile stoppers)

    • Goal: Correctly seal gases after removal according to the manufacturerʼs instructions. Currently, the gas cylinders are closed with red combination stoppers, which are used several times. However, multiple use of the combi stoppers is not permitted. Prompt clarification and definition of the procedure by the ophthalmological clinic.

  • Training in basic hygiene for ward and operating theatre personnel

  • Snood and face covering completely cover the scalp and beard hair

  • Hand disinfection (HD) before aseptic activities, such as handing over sterile goods and before using the gas cylinders

  • Non-sterile personnel must maintain an adequate distance from sterile surfaces and sterile staff

  • Enhanced operating theatre preparation: Wipe disinfection of the floor under the operating table and the hand contact surfaces.

  • Enhanced hand disinfection on the ward and in the operating theatre:

  • Examples:

    • before aseptic activity → application of drops/ointments, dressing change

    • before contact with patients: Palpation of the eyelids/examinations in the periocular region

  • Handling of antibiotic solutions used intraocularly in the operating theatre (for example, ceftazidime or cefuroxime)

  • Goal: The antibiotics should be freshly reconstituted under hygienic conditions for each individual patient and administered immediately (at the latest 1 h after reconstitution). Any remaining product is discarded immediately.

  • Structured information and reporting procedures

  • Goal: The information is distributed quickly and securely through clearly established information channels: any infection is immediately reported to the clinic director or the deputy clinic director; from there, the information is forwarded to the departmental hygiene officers, the management of hospital hygiene and the management of the Medical Process and Patient Safety Unit to initiate initial measures and secure evidence. In addition, the Medical Director is informed in parallel. These steps were implemented accordingly.

In addition, it was decided to establish the practice of administering an intracameral antibiotic after vitrectomies (this was previously not routine in the clinic, either for vitrectomies or after cataract operations). In cooperation with the in-house pharmacy and taking into account the requirements of hospital hygiene and the current feasibility, it was found that a sterilised individual preparation of cefuroxime is not viable for various reasons and violates requirements (including hygiene/legal). Thus, it was decided to standardise the approved and individualised administration doses of Aprokam 50 mg in sufficient quantities after each vitrectomy. This was carried out for a period of 3 months and then – in the absence of further hospital-acquired endophthalmitis cases – suspended again.

After the first 4 cases of endophthalmitis, 3 more cases occurred 8 months later (Patients 5, 6, 7 from [Table 1]) over a longer period of 6 weeks. Detection of the pathogen failed in all three cases. Since the detailed risk management described above had not previously revealed a point source, no other direct changes were made to management procedures in the operating theatre; however, the application observations and hygienic-microbiological assessment of the hospital environment were intensified again and, for example, it was now additionally discovered that an increased number of bacteria were detectable on the operating microscope and Staphylococcus aureus was also present on the handpiece of the tonometer on the ward. This prompted renewed emphasis on basic hygiene measures (such as hand disinfection and surface disinfection) for medical personnel. Furthermore, the option of antiseptic preoperative facial washing, the decolonisation of the nose, for example by means of mupirocin before vitrectomies, a declared 10-minute exposure time of the antiseptic, the improvement of basic hygiene with complete covering of scalp hair and beard hair with a snood and face protection, the maintenance of the exposure time during surgical hand disinfection and the option of wearing 2 pairs of sterile gloves was discussed so that one pair could be removed after covering and “double-gloving” to be able to detect possible contamination more quickly and more reliably. In particular, compliance with basic hygiene measures should be improved as to date this was not considered sufficient. However, as discussed below, there was no overall, definite starting point for solving the problem.

Discussion

The cluster of in-house endophthalmitis cases is a catastrophic event in an ophthalmic clinic, and stringent risk management is needed to identify the cause in the interest of patient safety. Openness and transparency are essential factors in creating appropriate confidence in the correct handling. The aim of this work was to describe in detail the root cause analysis procedure and thus possibly also to create instructions on how facilities can deal with such incidents in similar cases. As the BfArM report showed, there were apparently also 2 similar reports from another hospital at the same time.

Basic hygiene

In the case of postoperative bacterial infections, the focus is on all peri-operative infection prevention measures. It was therefore essential to closely examine the entire patient process from patient admission, any preliminary examinations, asepsis in the operating theatre and postoperative drug administration. Improving hygienic hand disinfection through direct observation with feedback and training was a primary focus, since this was regularly not carried out, as for example during intraoperative drawing-up of the gas or during postoperative medication administration (our in-house observation revealed a compliance rate of only 29% for regular hand disinfection in these situations; [Table 3]). Careful observation is therefore essential and it is known that hand disinfection is often insufficiently carried out in everyday clinical practice [5], [6], [7]. It was also important to change dropper bottles to single-use eye drop ampoules to prevent possible contamination of the eye surface, especially in intra- and postoperative settings. The detection of pathogens on the operating microscope after operating theatre preparation showed that the responsibilities for the preparation of the individual surfaces and apparatus were not clear. The cleaner is responsible for cleaning the operating theatre floor and the wall surfaces, but it became clear that there was no clear responsibility for cleaning the microscope itself and that it was obviously neither a medical nor a nursing task. Here, as in the ward, where it was not a nursing task, for example to clean the slit lamps and examination equipment (in the outpatient clinic, it is the task of the medical assistant), an interface conflict has been observed, but it has only been identified because of this risk management. The problem here, as may be the case in other hospitals, may be that the hospital management only has technical authority to issue instructions for the maintenance of the operating theatre and the ward. However, preparation can be delegated to the cleaning personnel in compliance with the appropriate conditions, provided that they have been instructed in the preparation of the medical devices.

All the appropriate tightened hygiene measures were extremely important in the context of this cluster of cases of endophthalmitis. They have once again raised the awareness of each and every individual involved in the process. However, in the first 4 patients who developed endophthalmitis within an extremely short period of only 1 week, no definitive point source of the cause could be identified. Although this can be seen as positive in principle, all hygiene measures should continue to be closely monitored in the long term.

Detection of pathogens

Certainly, the detection of pathogens in only 2 of 7 patients was definitely too low. It is known that such proof can often not be provided [8], [9], [10]; here, of course, the aim must be to achieve a higher detection rate. Perhaps the lower rate – unlike perhaps after cataract surgery and IVOM – can be explained by the fact that in vitrectomy of endophthalmitis after vitrectomy, there is of course no more or little vitreous body where the bacteria could settle, so that isolation may be more difficult. Furthermore, the optimisation of pre-analysis with, for example, timely transport and examination of the microbiological samples in the laboratory could improve the detection rate. Special examination kits such as those in the Freiburg University Hospital [8] could also be helpful across the board to achieve high detection rates. However, the value of these investigations remains the subject of debate when potent antibiotics such as vancomycin – to which the pathogens are sensitive in over 98% of cases [11] – are used routinely by the ophthalmological clinic. In our case, the higher detection rate would have been desirable, since it would have enabled a more robust and evidence-based identification of the bacterial strains involved. The two positive samples demonstrated that the strains of staphylococci were genetically distinct, thus excluding a common external root cause. The microbiological examinations of the vitrectomy single-use packs also showed no possible contamination of the batches, as clearly shown above.

Thus, the fact remains that even after intensive investigation, no definitive root cause for any of the listed cases could be identified in our hospital. The patients were treated in different operating theatres, with different operating machines, by different surgeons and different nursing personnel. Therefore, a “random” cluster must be assumed in this case.

Surgical instruments and intraoperative antibiotics

However, it is now known that cases of endophthalmitis after vitrectomy are rare [4], [12], [13]. Thus, in the face of such a cluster of cases, it is difficult for the willing observer to believe in a coincidence. But then what can be the reason for this? A variety of debates on this topic in and outside the clinic showed that important surface disinfection and aseptic activities may play a rather subordinate role in our issues (“in principle, we have all been doing everything the same”), but that the potential pathogen must be actively introduced into the eye, which means that the focus is primarily on the direct instruments – such as the sterile single-use cassette with infusion tube and cutter, possibly the light source and/or also the air/gas tamponade. Since the irrigation solutions are the same for cataract surgery and vitrectomy, these seem to be out of the question as a potentially causative agent. This is an important point because endophthalmitis is much more common after cataract surgery [14], but not a single case has occurred in our setting and we do not routinely administer intracameral antibiotics after cataract surgery. The added value of this antibiotic is controversial [15]. The randomised prospective study of European Society of Cataract & Refractive Surgeons (ESCRS) demonstrated a significantly lower endophthalmitis rate after cataract surgery with antibiosis than without antibiosis [16], however, in this analysis, the incidence of endophthalmitis was also significantly higher in the control group than in other studies, which was discussed accordingly [17]. For this aspect, the Freiburg Eye Clinic was also able to show, based on a very large internal patient cohort, that even without intracameral antibiotics after cataract surgery, there was no increased incidence of endophthalmitis. In this respect, the recommendation was to dispense with this surgical step in order to avoid possible sensitisation [17]. Also in [11] it could be shown that intracameral antibiotics may have a subordinate role and that surgical aspects are more important. However, recent meta-analyses confirm the clear trend of fewer cases of endophthalmitis with prophylactic antibiotics in recent years (18). The fact that no cataract surgery was involved in our clinical setting supports our approach to operate without prophylactic antibiotics, and it also argues against the possibility that the irrigation fluid used may have been contaminated.

The pump systems of the vitrectomy machines [19] [209] [21] used to be open systems, so theoretically and practically pathogens could be passed from one to the next patient. Nowadays, only closed systems are available with the single-use packs, so that a transfer of pathogens is not possible here. In this respect, it can also be assumed that the sterilisation process of single batches is safe nowadays. This was confirmed in our case by the BfArM investigation.

Regardless, the immune status of the patient may also affect the development of endophthalmitis. Predisposing factors such as alcohol or nicotine abuse were both present in one of our patients. Another patient had diabetes mellitus with impaired wound healing as a further predisposing factor. The other patients were supposedly healthy in terms of their immunocompetence.

Trocar-guided vitrectomy with higher incidence of endophthalmitis?

Following the results of the current POET study, which investigates the incidence of endophthalmitis in a Germany-wide registry, we see that initial evaluations show that in over 40% of cases IVOM and cataract surgery are the cause of endophthalmitis. Interestingly, vitrectomy is also the cause in more than 8% of cases (22). This may be surprising, since obviously 1 in 11 cases of endophthalmitis – and this is very high proportion – occurs nowadays after vitrectomy and thus may prompt considerations as to whether trocar-guided vitrectomy, unlike 20 G vitrectomy, may be a contributing trigger. Roughly speaking, in classic 20 G vitrectomy with open sclerostomy and potentially high flow volumes, there is more irrigation, and impurities and pathogens are better flushed out. Last but not least, some surgeons prefer the “open” approach to endophthalmitis management to potentially generate a significantly higher outflow from the eye than is possible or performed with valve trocars in 23 G/25 G vitrectomies [23]. What is certain is that with the predominantly trocar-guided surgical technique (20 G packs are sometimes no longer available from the manufacturers) the flow rate is significantly reduced and pathogens may be able to settle more easily intraocularly.

A further factor could be the fact that, when the infusion trocar is inserted, the lumen is immediately routed through peripheral vitreous bodies and partially remains there. As a result, an increased initial pressure is often required to generate the desired flow. This phenomenon can be observed in vitrectomies for silicone oil removal, where initially work must be carried out with higher machine-side infusion pressure against “vitreous” resistance (in this case oil) to open the lumen. In some cases, however, it can also be observed in the non-vitrectomised state, where initially, occasionally, increased infusion pressure is required to act against the resistance of the vitreous in the trocar. If the trocar is then pulled at the end of surgery after primary vitrectomy, then the remaining peripheral vitreous body is “used” virtually together with the air/gas filling as a tamponade of the sclerostomies. If, on the other hand, open 20 G vitrectomy typically involves trimming the applied sclerostomies with the cutter, the trocar-guided technique leaves vitreous bodies in the sclerostomy area. In the sense of a wicking effect, this could act as a potential entry port for pathogens. This is not the case after cutting free and suturing. Some surgeons therefore prefer, even with trocar-guided surgical techniques, to place a final suture in all access sites. Last but not least, this would also reduce the higher hypotension risk in trocar-guided vitrectomy without a final suture, which could also be a trigger for endophthalmitis. If such considerations are correct and if this explains an increased risk of endophthalmitis – as identified in POET [22] – a rise in incidence would be plausible. Nevertheless, the clustered timing of the endophthalmitis cases is significant and it is essential to carry out the risk management described here with transparent disclosure and assessment of all steps. In our case, all operations described were trocar-guided 23 G/25 G vitrectomies. It is, of course, not desirable to suture the trocar-guided sclerostomies closed again. Due to the small absolute number of problem cases, this does not seem to make sense and undermines the smart approach of trocar-guided operating procedure methodology. There is also a lack of consensus in the literature on whether trocar-guided vitrectomy generally increases the risk of endophthalmitis [4], [13] or not [12], [24]. Our own observations from the case series give reason to consider this possible risk increase and therefore to “pay attention” in our own operating theatre settings and to work towards a very high level of hygiene compliance.

Influence of anaesthetic mask ventilation?

Since no clear cause could be identified in the direct surgical procedure, the factor of anaesthesia was finally discussed again, since laryngeal masks were used in almost all cases. It is known that oral and pharyngeal pathogens can have a major impact in endophthalmitis [25], [26], [27]. Not least for this reason, the surgeon is required to wear an effective face covering. Considering the anaesthesia procedures performed here (6× laryngeal mask and 1× endotracheal intubation), this could also be a possible connection. Experience with vitrectomies shows that “secondary air” can occur more frequently precisely in the nasal portion of the adhesive due to a suboptimal masking of the eye – partially due to the design of the drape and its adhesive film, which may not adequately or properly seal the lid margin in that area. As a result, the routine use of a BIOM leads to fogging of the periocular auxiliary lens (which means the procedure cannot be continued), which can generally only be reduced or eliminated either by firmly re-sealing the drape or, more commonly, by cutting the drape free at the nasal canthus. Since laryngeal masks are prone to leakage, contamination of the surgical area by the patientʼs own oropharyngeal bacterial flora, in which staphylococcal species are usually also detectable, would be conceivable here by this route. In the event of leakage of a laryngeal mask, in addition to correcting its position and repositioning the head – which is difficult to do during an ongoing operation – as a last resort the ventilation pressure may also be increased, which could further complicate this approach. In this case, the patientʼs own (contaminated) exhaled air would also be actively directed over the sterile operating field with force. The additional customary use of valve ports in vitrectomy with significantly lower flushing rates than in open vitrectomies has already been pointed out above. Possibly, the combination of port-guided vitrectomy, leaks at the adhesive margins of the surgical drape, and use of a laryngeal mask is a convergence of unfortunate circumstances with unexpected collateral damage. An examination of this aspect seems appropriate, since the laryngeal mask plays a role in 6 of 7 affected cases. Only endotracheal intubation does not fit this pattern. In addition, this point of discussion would not apply in the same way to cataract surgeries, since, due to the operation, high-volume irrigation through the handpieces is maintained throughout almost the entire procedure.

Summary

Overall, there are several possibilities for an increased incidence of endophthalmitis after vitrectomies with standard hygiene. However, a clear cause with a definite solution to the problem did not emerge here. Nonetheless, as we were able to show plenty of room for improvement in our setting, particularly with regard to aseptic measures, which must be utilised appropriately. Furthermore, our experience continues to encourage us to forgo intracameral antibiotic administration at the end of surgery (cataract surgeries were not affected!), but on the other hand to examine to what extent a possible patient-derived, aerosol-carrying respiratory gas flow could transverse the surgical field.

In conclusion, it may be noted that it is extremely important for every surgical institution, particularly hospitals providing emergency care, to foster and uphold an adequate error and safety culture and to question possible extreme situations such as the very serious cluster of hospital-acquired endophthalmitis cases and to conduct root cause analysis in the interest of patient safety. Failure to take these necessary measures would be considered grossly negligent and must be avoided as a matter of urgency.

Conflict of Interest

The authors declare that they have no conflict of interest.

• References/Literatur

• 1 Bisorca-Gassendorf L, Boden KT, Szurman P. et al. [Postoperative endophthalmitis-a review of literature]. Ophthalmologe 2021; 118: 210-218
• 2 Chronopoulos A, Schwarz P, Fornoff L. et al. [Vitrectomy for endophthalmitis: Incubation time and prognosis following intravitreal injection vs. cataract surgery]. Ophthalmologe 2019; 116: 746-752
• 3 Michel T, Häberle H, Pham DT. [Follow-up of postoperative endophthalmitis after cataract surgery and intravitreal injection of drugs – an analysis since 2009]. Klin Monbl Augenheilkd 2015; 232: 676-681
• 4 Chen G, Tzekov R, Li W. et al. Incidence of Endophthalmitis after Vitrectomy: A Systematic Review and Meta-analysis. Retina 2019; 39: 844-852
• 5 Lotfinejad N, Peters A, Tartari E. et al. Hand hygiene in health care: 20 years of ongoing advances and perspectives. Lancet Infect Dis 2021; 21: e209-e221
• 6 Tanner J, Dumville JC, Norman G. et al. Surgical hand antisepsis to reduce surgical site infection. Cochrane Database Syst Rev 2016; (2016) CD004288
• 7 Ceparano M, Covelli A, Baccolini V. et al. Prevalence and predictors of hand hygiene compliance in clinical, surgical and intensive care unit wards: results of a second cross-sectional study at the Umberto I teaching hospital of Rome. Ann Ig 2025; 37: 25-36
• 8 Ness T, Pelz K. [Endophthalmitis: improvement of culture results]. Ophthalmologe 2000; 97: 33-37
• 9 Ness T, Serr A. [Diagnostics for endophthalmitis]. Klin Monbl Augenheilkd 2008; 225: 44-49
• 10 Seal D, Reischl U, Behr A. et al. ESCRS Endophthalmitis Study Group. Laboratory diagnosis of endophthalmitis: comparison of microbiology and molecular methods in the European Society of Cataract & Refractive Surgeons multicenter study and susceptibility testing. J Cataract Refract Surg 2008; 34: 1439-1450
• 11 de Geus SJR, Hopman J, Brüggemann RJ. et al. Acute Endophthalmitis after Cataract Surgery: Clinical Characteristics and the Role of Intracameral Antibiotic Prophylaxis. Ophthalmol Retina 2021; 5: 503-510
• 12 Bhende M, Raman R, Jain M. et al. Vitreoretinal Study Group (SNVR-Study Group). Incidence, microbiology, and outcomes of endophthalmitis after 111,876 pars plana vitrectomies at a single, tertiary eye care hospital. PLoS One 2018; 13: e0191173
• 13 Tabatabaei SA, Soleimani M, Vakili H. et al. The rate of endophthalmitis after pars plana vitrectomy and its risk factors. Int Ophthalmol 2019; 39: 1299-1305
• 14 Sun J, Guo Z, Li H. et al. Acute Infectious Endophthalmitis After Cataract Surgery: Epidemiological Characteristics, Risk Factors and Incidence Trends, 2008–2019. Infect Drug Resist 2021; 14: 1231-1238
• 15 Garg P, Roy A, Sharma S. Endophthalmitis after cataract surgery: epidemiology, risk factors, and evidence on protection. Curr Opin Ophthalmol 2017; 28: 67-72
• 16 Endophthalmitis Study Group, European Society of Cataract & Refractive Surgeons. Prophylaxis of postoperative endophthalmitis following cataract surgery: results of the ESCRS multicenter study and identification of risk factors. J Cataract Refract Surg 2007; 33: 978-988
• 17 Ness T, Kern WV, Frank U. et al. Postoperative nosocomial endophthalmitis: is perioperative antibiotic prophylaxis advisable? A single centreʼs experience. J Hosp Infect 2011; 78: 138-142
• 18 Shi SL, Yu XN, Cui YL. et al. Incidence of endophthalmitis after phacoemulsification cataract surgery: a Meta-analysis. Int J Ophthalmol 2022; 15: 327-335
• 19 Dhaliwal DK, Kowalski RP, Donahue SP. et al. Evaluation of vitrectomy machines as a source of false-positive culture contamination in endophthalmitis. Am J Ophthalmol 1995; 119: 62-64
• 20 Janknecht P, Kappstein I. Bacterial contamination of the pressure receiver of a vitrectomy machine. Ophthalmic Surg Lasers 1998; 29: 345-347
• 21 Kappstein I, Schneider CM, Grundmann H. et al. Long-lasting contamination of a vitrectomy apparatus with Serratia marcescens. Infect Control Hosp Epidemiol 1999; 20: 192-195
• 22 Hattenbach LO, Lommatzsch A, Liegl R. et al. Postoperative Endophthalmitis-Therapie (PO.E.T.) – multizentrische retrospektive Retina.net Registerstudie zu aktuellen Behandlungsstrategien. Abstract Jahrestagung der Deutschen Ophthalmologischen Gesellschaft (DOG) vom 10.–13.10.2024 in Berlin.
• 23 Framme C, Sachs HG, Wachtlin J. et al. Main Principles of Vitrectomy Using Intraocular Tamponades – A Basic Course in Surgery. Klin Monbl Augenheilkd 2022; 239: 1337-1353
• 24 Bahrani HM, Fazelat AA, Thomas M. et al. Endophthalmitis in the era of small gauge transconjunctival sutureless vitrectomy–meta analysis and review of literature. Semin Ophthalmol 2010; 25: 275-282
• 25 Doshi RR, Leng T, Fung AE. Reducing oral flora contamination of intravitreal injections with face mask or silence. Retina 2012; 32: 473-476
• 26 Shimada H, Hattori T, Mori R. et al. Minimizing the endophthalmitis rate following intravitreal injections using 0.25 % povidone-iodine irrigation and surgical mask. Graefes Arch Clin Exp Ophthalmol 2013; 251: 1885-1890
• 27 Brynskov T, Kemp H, Sørensen TL. No cases of endophthalmitis after 20,293 intravitreal injections in an operating room setting. Retina 2014; 34: 951-957

Publikationsverlauf

Eingereicht: 16. Mai 2025

Angenommen: 22. Juli 2025

Artikel online veröffentlicht:
01. September 2025