Introduction

Acute postoperative endophthalmitis (POE) is still one of the most serious complications of cataract surgery.1 Indeed, bacteria are sometimes postoperatively present within the aqueous humour without inducing this condition.2 Precisely why bacteria induce POE after cataract surgery is not entirely understood. The risk of its development may be influenced by several factors, above all by bacterial adherence to the intraocular lens (IOLs).2, 3, 4 The ability of an organism to adhere to the IOL surface is believed to be associated with a risk of infection at the implantation site.5, 6, 7, 8 Obviously, bacterial adhesion to an IOL will be influenced by the properties of the biomaterial employed, such as its chemical composition, surface roughness, hydrophilicity, and surface electric charge.2, 4, 9, 10 A causal relationship between bacterial adhesion to IOLs and POE was first demonstrated for the material polypropylene.

Epidemiological data relating to this supposed relationship are scarce and have provided conflicting results for some IOL materials.11, 12, 13, 14, 15, 16, 17, 18 The present retrospective cohort study was conducted with a view to throwing further light on the association between IOL material and type and the risk of developing POE.

Materials and methods

All patients who had undergone extracapsular cataract extraction or phacoemulsification with IOL implantation at a single centre (Croix-Rousse Hospital, University of Lyon, France) between 1st January 1994 and 31st December 2004 were included in this cohort. The IOL type implanted was classified according to its biomaterial composition into one of the five categories: polymethylmethacrylate (PMMA), heparinized PMMA, silicone, hydrophilic acrylic, and hydrophobic acrylic. Fluorine PMMA was not used. Rigid (PMMA and heparinized PMMA) IOLs with an optic diameter of 5.5–6.5 were used as the IOLs of choice in cases without complications before 2000. Small-incision cataract surgery using foldable IOLs has been performed as standard of care since 2000.

Our standard technique for phacoemulsification consists of a three-step incision at the limbal sclera (sclerocorneal phacoemulsification) associated with a clear corneal paracenteses. Scleral tunnels are always constructed in the superior quadrant. Rarely, clear corneal incisions (corneal phacoemulsification) have to be performed, with the incisions also always in the superior quadrant. In all cases, the wounds are sutured using a 10-0 nylon suture, to guarantee perfect closure of the wound. All patients are given a single dose of oral ofloxacine 2 h before surgery, if no contraindication is found. Use of topical 5% povidone-iodine and careful draping of the eyelids are systematically performed in any case, and additionally, a subconjunctival aminoglycoside injection was placed at the end of surgery until 2001. Intracameral or systemic antibiotics were never applied prior to or during surgery. All patients receive topical antibiotic and dexamethasone after surgery during 1 month.

Cases of POE are usually identified at the Lyon Croix-Rousse University Hospital in a prospective manner by the Department of Infectious Diseases. Patients presenting with acute POE within 6 weeks of cataract surgery were diagnosed according to criteria established by the Endophthalmitis Vitrectomy Study (EVS) Group.19 POE was defined either as an impairment of visual acuity (6/12 or worse) or as a manifestation of pain, hypopion, or significant clouding of the anterior chamber or vitreous after cataract surgery. The medical characteristics of the patients with acute POE were documented, with particular attention being paid to recognized risk factors (ie, diabetes mellitus, chronic obstructive-airway disease, immunosuppressive therapy, long-term steroid treatment, glaucoma treatment, lid-margin disease, previous intraocular surgery, disruption of the capsular barrier, loss of vitreous, and postoperative wound abnormalities).11, 16 After 1997 (but not before), aliquots of aqueous humour or vitreous were routinely withdrawn from patients with POE for laboratory testing, and the condition was managed according to guidelines established by the EVS Group.19

All statistical analyses were carried out using SPSS version 12.0 (SPSS Inc., Chicago, IL, USA) or SAS version 8.2 (SAS Institute Inc., Cary, NC, USA) packages. The incidence of acute POE and the exact 95% confidence interval were estimated according to a Poisson distribution. When applicable, odds ratios and 95% confidence intervals were used to assess the influence of each IOL material on the occurrence of POE. Fisher's exact test was used to compare the incidence rates associated with the different IOL materials, with the time of surgery, and the surgical technique (extracapsular cataract extraction vs phacoemulsification). Bonferoni corrections were made for the multiple comparisons before concluding that differences were statistically significant.

Results

Between 1st January 1994 and 31st December 2004 (11 years), 5837 IOLs were implanted during cataract surgery in our department, 4907 during phacoemulsification, and 930 during extracapsular cataract extraction. The medical characteristics from patients developing POE may be depicted from Table 1. The IOLs, which were manufactured by various companies (Table 2), composed of five different biomaterials: 5% of silicone, 14% of PMMA, 20% of hydrophilic acrylic, 20% of hydrophobic acrylic, and 41% of heparinized PMMA.

Table 1 Medical characteristics of patients with POE
Full size table
Table 2 Numerical distribution of the five different IOL materials implanted in relation to the incidences of POE
Full size table

Eight cases (two females and six males) of acute POE were registered within 6 weeks of cataract surgery (Tables 1 and 2), which represented an overall incidence of 0.14% (95% confidence interval: 0.06–0.27%) during the 11-year study period. The median age of the POE patients was 62.5 years (range: 47–88 years). The median time elapsing between cataract surgery and the diagnosis of POE was 8 days (range: 4–16 days). Seven of the eight patients underwent phacoemulsification (7/4907; 0.14%) and one other extracapsular cataract extraction (1/930; 0.11%; Table 1). Two patients had a history of prostate adenocarcinoma and one had undergone topical treatment for glaucoma. None of these POE patients presented with either diabetes mellitus, chronic obstructive-airway disease, lid-margin disease, disruption of the capsular barrier, loss of vitreous, or postoperative wound abnormalities, and none of them had undergone either immunosuppressive therapy, long-term steroid treatment, or previous intraocular surgery. All patients were given a single dose of oral ofloxacine 2 h before surgery. Only four of the eight patients had been subjected to tapping of the anterior chamber or vitreous. POE had been triggered predominantly by Gram-positive bacteria (Table 1). A final visual acuity ≥20/40 was achieved in five of the eight patients.

In the eight POE patients, seven of the IOLs were composed of PMMA and the other one of heparinized PMMA (Table 2). The impact of IOL material on acute POE was significant (global Fisher's exact test: P=0.00007). Patients with PMMA IOLs carried a statistically higher risk of developing POE than did those implanted with either hydrophilic acrylic, hydrophobic acrylic, or heparinized PMMA IOLs (odds ratio: 0.05; 95% confidence interval: 0.01–0.41) (Table 2). The cataract surgery technique (ie, phacoemulsification or extracapsular cataract extraction) had no significant influence on the development of POE (P=1). Between 1994 and 1997, extracapsular cataract extraction was performed more frequently (79%) than phacoemulsification (Figure 1); hence, rigid IOLs were implanted more often (96%) than foldable ones (Table 3) and thus rigid IOL implantation was not due to the presence of surgical complications in this series. Between 1998 and 2004, phacoemulsification was performed more commonly (88%) than extracapsular cataract extraction; hence, foldable IOLs were implanted more frequently (57%) than rigid ones. However, the incidences of POE occurring within these two periods did not differ significantly from each other and actually depended of the nature of the IOL implanted (Fisher's exact test including Bonferoni's correction: P=0.018; but Fisher's exact test after adjusting for the IOL type implanted: P=0.74).

Figure 1
figure 1

Numerical distribution of the five different IOL materials implanted between 1994 and 2004.

Full size image
Table 3 Numerical distribution and incidences of POE, of the five different intraocular lens materials, and of the two different types of cataract-extraction technique between the periods 1994–1997 and 1998–2004
Full size table

Discussion

The present retrospective study affords evidence that the IOL material is correlated to the risk of acute POE after cataract surgery. Since POE is a rare complication of cataract surgery, large reference samples are necessary to establish reliable results. Hence, epidemiological data relating to the impact of IOL materials on the development of POE are scarce, and available data have provided contradictory results regarding the contribution of IOL materials.11, 12, 13, 14, 15, 16, 17, 18

Nevertheless, there is a general agreement on a qualitative level that the IOL material and its surface composition are linked to bacterial adhesion and the risk to develop POE (Table 4). There is some evidence that polypropylene haptics support bacterial adherence to the IOL. Dilly and Sellors5 and Raskin et al20 have shown that bacteria are more prone to adhere to polypropylene haptics than either to PMMA haptics20 or to the PMMA optical component,5, 20 both in vitro5, 20 and in explants derived from patients with recurrent episodes of intraocular inflammation.5 And in an earlier epidemiological study, Raskin et al20 demonstrated polypropylene haptics to be a significant risk factor in the development of POE (risk ratio: 4.5; P<0.007). Lastly, a recent epidemiological retrospective case–control study demonstrated that silicone IOLs with polypropylene haptics could be a significant risk factor for the development of POE (relative risk: 21; P=0.013) in comparison to silicone IOLs with PMMA haptics.14 Adding to this, we observed that patients with PMMA IOLs carried a higher risk of developing POE than did those implanted with either heparinized PMMA, hydrophilic acrylic, or hydrophobic acrylic IOLs. Silicone IOLs were implanted in only 5% of the patient population; hence, data relating to this material lacked statistical power. Consistent to our results, Montan et al11 have demonstrated that patients implanted with heparinized PMMA IOLs are better protected against POE than are those with either PMMA (P=0.002) or silicone IOLs (P=0.04). However, in a more recent study,12 the same authors demonstrated the risk of developing POE to be lower in patients with acrylic IOLs than in those with either silicone (P=0.014), PMMA (P=0.001), or hydrogel IOLs (P=0.001), whereas the risk difference between acrylic IOLs and heparinized PMMA IOLs was not significant (P=0.06). The lower risk of POE associated with heparinized PMMA IOLs was attributed by Montan et al11 to the biocompatible, hydrophilic surface properties of the heparin coating. This is supported by in vitro data (Table 4).3, 28, 29, 30, 31, 32, 33, 34

Table 4 Factors with influence on the risk to develop POE after cataract surgery
Full size table

Many of the discrepant results published so far may be derived from the retrospective nature of most clinical studies. A quantitative analysis of any single study might therefore misweight the study findings. In order to provide an overview over the existing studies and to give the reader space for an own weighting of the pertaining results, we have therefore displayed the most important findings from published clinical (Table 5a) and in vitro studies (Table 5b) in a qualitative manner instead of discussing the single studies.

Table 5 Overview over the most relevant findings in clinical settings reporting on the risks of POE and factors possibly influencing incidences or outcomes
Full size table
Table 6 Overview over some findings from in vitro cultural studies reporting on risk modification for POE
Full size table

In the current study, the incidence of endophthalmitis in the PMMA group (0.84% or 7/833 cases) is rather high compared to other reports (0.16% or 15/9453 cases and 0.06% or 5/7873 cases).12, 17 This is probably explained by the difference in the number of patients included. Nevertheless, even if the current study involved less patients, it is noteworthy that the number of patients included per group, that is, with PMMA (833 cases), hydrophilic acrylic (1186 cases), and hydrophobic acrylic (1140 cases), was comparable. Therefore, our study is probably strong enough to allow to read out on a qualitative basis a tendency that heparinized PMMA IOLs have a lower risk, that is, 20-fold lower risk of endophthalmitis (0.04% or 1/2362 cases), compared to PMMA (0.84%) and actually a similarly low risk compared to acrylic lenses.

Also in the context of our results, the findings of Nagaki et al15 may be worth mentioning: cataract patients were randomized into three groups. In the first, a hydrophobic acrylic IOL was implanted through a temporal corneal incision; in the second, the same type of IOL was implanted via a superior sclerocorneal incision; and in the third, a silicone IOL was implanted via the latter route. The incidence of POE was higher in the first group than in the other two. This finding indicates that a temporal corneal incision may heighten the risk of POE developing, but that hydrophobic acrylic and silicone IOLs silicone do not have a differential impact on its manifestation. In the present series of patients, the incision was systematically made at the 12 o’clock position, with the conjunctiva covering the wound postoperatively. Hence, we can exclude the influence of incision site (Table 1). Moreover, the cataract-extraction technique probably had no influence on the incidence of POE; hence, several studies have reported a reduction in the incidence of POE following the change from intra- to extracapsular extraction,11 but no further reduction in incidence has been reported following the change from extracapsular extraction to phacoemulsification (Table 4).21, 22, 23

Rigid IOLs require a wound enlargement to 5.5–6.5 mm, which could possibly result in a higher chance of wound leakage and thereby increase the risk of POE. Moreover, Mayer et al24 reported that injectable IOLs were less likely to cause POE because rigid lenses may be contaminated by contact with conjunctiva upon insertion.24 Nevertheless, among the rigid IOLs group, we demonstrated a significant higher risk of POE with PMMA than with heparinized PMMA, proving that wound's size is not the main aetiological factor in this series and that biomaterial seems on the contrary to be the most important one.

We are aware of the limitations of our study, which namely lie in its retrospective nature and the relative limited number of patients included. Therefore, the conclusions drawn from our results have to be interpreted with care. On the other hand, we feel that reporting these data may be worthwhile because the incidence of POE is so low that at a later point, a pooling of published data could virtually contribute to understand the contradictory results and to find a more reliable answer in which direction to focus with future large multicentre studies. The strength of this retrospective investigation lies in its single-centre design, which guaranties tight control and uniformity of diagnostic criteria and surgical techniques.

In summary, we gathered further evidence that the risk to develop an acute POE after cataract surgery is associated to the lens material and type used and may be increased with the use of PMMA IOLs. Several factors possibly add to this increased risk, with the IOL material obviously being according to published and own findings beyond the more relevant ones. Nevertheless, a substantial number of operating technique-dependent variables and systemic risk factors will have to be addressed in a prospective multivariate analysis before the actual role of IOL material in POE development can definitively be assessed.