You are on page 1of 7


Risk factors for endophthalmitis after

cataract surgery: Predictors for causative
organisms and visual outcomes
Mats Lundstr
om, MD, PhD, Emma Friling, MD, Per Montan, MD, PhD

PURPOSE: To investigate visual outcome, bacteriology, and time to diagnosis in groups identified
as being at risk for endophthalmitis following cataract surgery.
SETTING: Swedish National Cataract Register.
DESIGN: A retrospective review of postoperative endophthalmitis and control cases reported from
2002 to 2010.
METHODS: Three identified risk groups for endophthalmitis confirmed in previous multivariate
models were organized in such a way that the highest level of significance determined the allocation
of cases that belonged to more than one group. Control cases of the entire database were arranged
in the same manner.
RESULTS: Of the 244 endophthalmitis cases occurring in 692 786 surgeries, 148 did not belong to
any risk group, whereas the remaining cases were part of the following groups at risk: nontreatment
with intracameral antibiotic (n Z 22), communication with vitreous (n Z 18), and age 85 years or
more (n Z 56). Cefuroxime was the intracameral antibiotic used in 99% of treated cases. Cases
sustaining a communication with vitreous were found to have the worst visual prognosis. Among
causative organisms, Gram-positive bacteria were significantly more frequent in cases with a
communication with vitreous, whereas staphylococci and Gram-negative results were more
common in patients aged 85 years or more than in nonrisk patients.
CONCLUSION: Limiting the size of the risk groups by giving a prophylactic intracameral antibiotic to
every single patient and by intervening earlier in the course of cataract development appear to be
first steps in further reducing the endophthalmitis rate. Adjustments of the intracameral regimen
may be advantageous in some risk groups.
Financial Disclosure: None of the authors has any financial or propriety interest in any material or
method mentioned.
J Cataract Refract Surg 2015; 41:24102416 Q 2015 ASCRS and ESCRS

Postoperative endopthalmitis may totally negate the vitreous and patient age of 85 years or more were
expected vision improvement after cataract surgery. convincingly demonstrated as independent risk fac-
This is why endophthalmitis is still feared despite tors in both studies, whereas sex, intraocular lens ma-
better defined management1 and a declining inci- terial and incision location were not. In the present
dence, the latter demonstrated in large-scale pro- study we revisited the endophthalmitis cases from
spective studies published in the last 10-year these 2 papers to investigate bacterial etiology and
period.24 visual outcome in relation to the identified risk factors
In the 2 most recent publications on endophthalmi- with the aim of finding avenues to reduce the inci-
tis from the Swedish National Cataract Register dence of this complication, for example, by planning
(NCR), 244 cases were reported among 692 786 sur- the cataract extraction so as to avoid capsule complica-
geries collected over a time span of 9 years.2,3 Nonuse tions but also possibly by fine-tuning the prophylactic
of an intracameral antibiotic, communication with protocol.

2410 Q 2015 ASCRS and ESCRS

Published by Elsevier Inc. 0886-3350

PATIENTS AND METHODS S aureus and CoNS); enterococci; remaining Gram-positive

organisms (mainly streptococci); and Gram-negative spe-
The prospective reporting of incidents of endopthalmitis to
cies. Antibiotic resistance patterns for identified species
the NCR has been in place since 1998. Data collection and
were collected by the registry from 2008 and onward, but
results have been described in previous papers from our
the reporting was not consistent. Data in this respect
group.2,3,5,6 For the present study, endophthalmitis cases
were, however, available from St Erik Eye Hospital, a major
in the database of the NCR encompassing registrations
partner of the NCR, managing all endophthalmitis cases in
from 2002 through 2010 were grouped into the 3 risk co-
the greater Stockholm area, where 26.6% of all culture-
horts and ranked according to the level of statistical signif-
positive cases in the study period occurred. In this sample,
icance found in the 2 most recent epidemiologic surveys
resistance to cefuroxime, the most frequently used intra-
from the registry.2,3 This means that endophthalmitis case
cameral antibiotic, was assessed.
patients who were not given a prophylactic antibiotic intra-
The study received approval from the Ethics Committee,
camerally were allocated to a single group regardless of
Lund University, and the Swedish Data Inspection board
whether they had also had a perioperative communication
and was performed in accordance with the tenets of the
with vitreous or were aged 85 years or more. A second sub-
Declaration of Helsinki.
set included remaining case patients that had had a periop-
All statistical calculations were performed with the IBM
erative communication with vitreous, even if they were
SPSS (version 22 SPSS Inc., Chicago, IL). A Pearson c2 test
aged 85 years or more. A third cohort consisted of case pa-
was used when comparing proportions of the nonrisk cohort
tients aged 85 years or more after subtracting the above-
versus those of the 3 risk groups. A linear regression model
mentioned categories. A fourth subset was made up of cases
was used in the multivariate analysis. A Mann-Whitney U
not belonging to any of these risk groups. The entire data-
test was used for comparing days between the operation
base of all reported cataract surgeries was then structured
and the diagnostic procedure for the different groups. The
in the same fashion to obtain a proper denominator for
Bonferroni correction was applied for multiple comparisons.
determination of the absolute frequency of causative organ-
P values of less than 0.05 were considered statistically
isms. For the analysis of visual acuity of cases, evaluated
approximately 3 months following the infection, reported
ocular comorbidity as well as causative organisms were
entered into the model. Comorbidity was defined according
to the manual of the NCR as the preoperative presence of RESULTS
macular degeneration, diabetic retinopathy, glaucoma and Most endophthalmitis cases did not belong to any risk
other condition causing visual impairment during the
study period as a whole and from 2008 and onward also group (n Z 148; 65% of the entire sample). Second in
corneal guttata. In terms of culture results, the database size was the group aged 85 years or more, numbering
was stratified into virulent bacteria, ie, organism associated 56, whereas cases not receiving any intracameral anti-
with at least a 50% risk of a visual outcome less than 20/60, biotic or having a per-operative communication with
or conversely a favourable cause associated with at least a vitreous were fairly equal in size (n Z 22 and n Z 18,
50% chance of attaining a visual acuity of 20/60 or less, as
found in the 2 reanalyzed papers. According to this princi- respectively). The rate of endophthalmitis was strik-
ple, Staphylococcus aureus and coagulase-negative Staphylo- ingly different in all 3 populations at risk (P ! .001)
coccus (CoNS) and culture-negative cases constituted the versus that with no risk. Patients not receiving an
favorable group, whereas all other causes were labeled viru- intracameral antibiotic had a rate of 0.43% in 5119
lent. Causative organisms were described in terms of their surgeries; patients with communication with vitreous
relative and absolute frequencies in the respective risk
groups. To avoid too fragmented a grouping of the bacteria, had a rate of 0.14% in 13 231 operations; patients aged
the following classes were defined: staphylococci (including 85 years or more had a frequency of 0.05% in 106 002
operations; and patients in the nonrisk group had a
frequency of 0.026% in 568 223 operations. As for
patients receiving an intracameral antibiotic, 1 mg of
Submitted: January 13, 2015. cefuroxime was used in 99% of cases.
Final revision submitted: May 13, 2015. In the logistic regression model evaluating endoph-
Accepted: May 13, 2015.
thalmitis cases only, virulent bacteria, presence of any
From the Department of Clinical Sciences (Lundstrom), Ophthal- risk factor for endophthalmitis, and ocular comorbid-
mology, Faculty of Medicine, Lund University, Lund, Sweden; St ity were found to be independent determinants of
Erik Eye Hospital (Friling, Montan), Stockholm, Sweden. poorer visual outcomes, defined as acuity of less
than 20/60. In a post hoc analysis of the risk factors,
Financial support was provided by The Swedish Association of
Local Authorities and regions and through the regional agreement cases with a per-operative communication with vitre-
on Medical Training and Clinical Research (ALF) between the Stock- ous were found to have significantly worse outcomes
holm County Council and Karolinska Institutet (20120623). than nonrisk cases (Tables 1 and 2). More relevant,
however, is calculating the frequency of impaired
Presented in part at the XVIIIth Winter Meeting of ESCRS in Ljubl- vision following endophthalmitis, that is, taking into
jana, Slovenia, February 2014.
account the number of surgeries in the respective
Corresponding author: Per Montan, MD, PhD, St Erik Eye Hospital, risk categories (Table 3). This analysis clearly shows
112 82, Stockholm, Sweden. E-mail: an increased risk in absolute terms for visual loss after



Table 1. Logistic regression analysis within the endophthalmitis cohort (n Z 244) with visual outcome as the dependent variable, defined
as worse than 20/60 (0) or 20/60 and better (1).

Independent Factor B coefficient Standard Error P Exp of B 95% CI for Exp B

Virulent bacteria 1.630 0.306 !.001 5.106 2.804, 9.297

Any risk group for E 0.718 0.311 .021 2.050 1.115, 3.771
Comorbidity 0.639 0.302 .034 1.050 1.048, 3.428

CI Z confidence interval; Exp Z exponentiation

endophthalmitis in all risk groups compared to the period revealed that enterococci (n Z 22) were cefur-
nonrisk cohort. oxime resistant without exception (whereas all were
The distribution of causative organisms showed vancomycin susceptible). Of 18 CoNS, 15 were resis-
conspicuous dissimilarities within the risk categories. tant to cefuroxime, whereas only the S aureus strain
The rate of favorable causes of endophthalmitis, that was sensitive. Other Gram-positive bacteria (n Z 6),
is, culture negativity and staphylococci, was statisti- mainly streptococci, were all susceptible to cefurox-
cally significantly higher in the group not receiving ime, as were 4 of 7 Gram-negative strains.
an intracameral antibiotic (77%), as compared to a
the nonrisk group, in which the rate was 45.5%
(P Z .015, Bonferroni adjusted), whereas the corre- DISCUSSION
sponding proportion was 45% and 45.6% in the other The Swedish National Cataract Register (NCR) com-
risk populations (Figure 1). When calculating the fre- prises a unique nationwide registry of endophthalmitis
quency of causative organisms using the number of incidents following cataract surgery. Reported infected
cataract operations as the denominator, significant cases can be traced to the large database of the NCR,
differences in proportions were demonstrated in meaning that relevant patient background data and
the risk groups versus the nonrisk group (Table 4). information on operation technique are automatically
For example, patients not given an intracameral accessible for both cases and all controls. These assets
antibiotic and those with a peroperative capsule constitute the basis for the current investigation.
problem had significantly more endophthalmitis inci- Although infections from bacteria such as strepto-
dents caused by any Gram-positive organism. In the cocci and some Gram-negative organisms are known
patients aged 85 years or more, infections caused by to result in poor visual outcomes,2,3,7,8 only a few
staphylococci and Gram-negative species were signif- reports have been published concerning patient back-
icantly more frequent. ground and operation complications relative to visual
The median time span between the cataract opera- results after endophthalmitis following cataract sur-
tion and the diagnostic procedure was 4 days for pa- gery. In the present research, we demonstrated that
tients not given an intracameral antibiotic, 7 days for ocular comorbidity, virulent bacteria, and any risk cate-
patients aged 85 years or more and was 6.5 days for gory for endophthalmitis were all independently associ-
nonrisk cases. In contrast, in patients sustaining a ated with poor visual outcomes in a multivariate
communication with vitreous, the median time span analysis. It is of little surprise that ocular comorbidity,
was 15 days, but the difference did not achieve statis- defined as the presence of visually affecting glaucoma,
tical significance. age-related macular degeneration, diabetic retinopathy,
A scrutiny of resistance patterns of bacteria causing and/or any other sight-hindering condition, contributes
endophthalmitis in the Stockholm area in the study to worse vision once the infection has healed. A similar

Table 2. Visual acuity related to risk group for developing endophthalmitis.

Cause No Intracameral Antibiotic* Communication with Vitreous Age R85 y No Riskz

Visual acuity !20/60 8 (36.4%) 15 (83.3%) 35 (62.5%) 73 (49%)

Visual acuity R20/60 11 (50%) 3 (16.7%) 18 (32.1%) 66 (44.5%)
P, Bonferoni correction 1.0 .039 .51

Comparisons made between the nonrisk subset and the different risk groups
*Missing data for 3 cases

Missing data for 3 cases
Missing visual acuity data for 9 cases



Table 3. Absolute frequency of a less favorable visual result after endophthalmitis.

No Intracameral Communication with

Cause Antibiotics (n Z 5 119) Vitreous (n Z 13 231) Age R85 y (n Z 106 002) Nonrisk (n Z 568 223)

Visual acuity !20/60 0.16% 0.11% 0.03% 0.01%

P, Bonferoni correction !.001 !.001 !.001

Comparisons made between nonrisksubset and the different risk groups

finding was reported in the Endophthalmitis Vitrec- patients at risk. Our data clearly show that solely by
tomy Study program, in which diabetes was found to minimizing the risk populations, a rate of endoph-
have a negative impact on visual outcomes.9 In a rela- thalmitis as low as 0.025%, or 1 case among 4000
tively recent French study, the lack of intraoperative operations, is feasible. Moreover, a tailoring of the
capsule rupture and vitreous loss was a determinant prophylactic antibiotic protocol in some of the risk
of good visual prognosis.10 This, in essence, corrobo- subsets could possibly further limit the number of
rates the present finding of significantly worse VA for infections. The most obvious and immediate measure
patients sustaining communication with vitreous: in is to insist on surgeons' injecting an antibiotic intra-
all, 83% of affected patients ended up with visual acuity camerally in all cataract operations, given the striking
of less than 20/60, which indeed may be expected, as a increase in endophthalmitis incidence in the group
capsule rupture in itself jeopardizes the postoperative not receiving intracameral antibiotic in comparison
visual result.11 The actual impact of visual impairment with the nonrisk group. In light of the compelling
following endophthalmitis in our investigated cataract support for intracameral cefuroxime in the prospec-
populations is, however, evident only from the data pre- tive European Society of Cataract and Refractive Sur-
sented in Table 3. It demonstrates that despite the appar- geons (ESCRS) study on endophthalmitis prophylaxis,
ently benign incidents of endophthalmitis in the cohort in addition to the data from prospective, large-scale
not receiving an intracameral antibiotic, as seen in observational studies from our own registry,24 practi-
Table 2, the visual damage is still the highest in this cally no patient is presently denied an intracameral
very group, which could be explained by its overall antibiotic in Sweden (unpublished data from the
high numbers of infectious incidents. NCR). The recommendation can be extended to
Another purpose of this investigation was to high- include other antibiotics with reasonable safety and ef-
light the endophthalmitis frequency as well causative ficacy documentation with this mode of administra-
organisms in the respective structured risk subsets. tion, if surgeons are hesitant to use cefuroxime.12,13
This analysis, we believe, could help to outline an The elderly population can be limited by simply
action program to further reduce the number of intervening earlier in the course of the cataract

Figure 1. Distribution of bacteria

within each risk category. Note
the substantial differences in num-
ber of cases. When comparing the
group not receiving intracameral
antibiotics with the nonrisk group,
the proportion of favorable causes
of postoperative endophthalmitis,
that is, cases with staphylococcal
or culture-negative infections, was
statistically significantly higher in
the former group (P Z .015, Bonfer-
roni corrected).



Table 4. Numbers of endophthalmitis cases caused by a given agent.

No Intracameral Antibiotics Communication with

Cause of Endophthalmitis (n Z 5 119) Vitreous (n Z 13 231) R85 y* (n Z 106 002) Nonrisk (n Z 568 223)

Negative culture 4 (1: 1 300) 5 (1: 2 500) 5 (1: 21 000) 23 (1:25 000)
Staphylococci 13 (1: 400) 3 (1: 4 000) 21 (1: 5 000) 41 (1: 13 000)
P ! .001 P ! .001 P ! .001
Enterococci 3 (1: 1 700) 4 (1: 3 000) 14 (1: 7 500) 46 (1: 12 000)
P ! .001 P ! .018
Other Gram-positive organisms 2 (1: 2 500) 6 (1: 2 000) 5 (1: 21 000) 17 (1: 33 000)
P ! .001 P ! .001
Gram-negative organisms 0 0 10 (1: 10 000) 18 (1:32 000)
P Z .009

Rough ratios are given for didactic reasons

Comparisons are made between nonrisk subset and the different risk groups Significant P after Bonferoni corrections only are provided
*No culture taken (n Z 1)

No culture taken (n Z 3)

formation. The same approach would prevent the favor bacterial proliferation, which in itself may
development of small pupils, dense cataracts, and loose explain the higher rate of endophthalmitis in this
zonular fibers, features that are associated with a subgroup of patients. Whether the presence of vitre-
communication with vitreous.14 The latter was ous is particularly permissive of streptococcal growth
confirmed in an 8-year study from the NCR, in which remains to be proved, however. Another uncertainty
a decrease over time of the average patient age corre- relates to the failure of cefuroxime to prevent strepto-
lated well with a reduced rate of capsule complications, coccal infections in these cases. With the breaching of
as analyzed in more than 600 000 operations.15 A judi- the posterior capsule barrier, the compartment in
cious allocation of patients with the aforementioned which the prophylactic antibiotic is supposed to act
conditions to the best-suited surgeons should further is expanded. Consequently, the killing capacity of the
reduce the group of surgically treated patients sustain- intracameral antibiotic may be diluted, and perhaps
ing a communication with vitreous. for cefuroxime, with its time-dependent antibacterial
With regard to the possibility of adjusting the anti- activity, the otherwise apparently effective dose may
biotic prophylaxis in certain risk groups, the discus- not suffice. Raising the dose would then be one option,
sion will have to rely on the Stockholm cohort data but there are no published safety data to support such
in the absence of complete information on antibiotic an adjustment. In addition, vitreous strands caught in
resistance from the registry. It can safely be stated, the incision most certainly contribute to contamina-
however, that among the causative organisms, tion, and could act as routes for invading bacteria
enterococci are all cefuroxime resistant, as are the once the bioavailability of the antibiotic has waned.
vast majority of CoNS. In contrast, all of these strains This mechanism may explain the incidents of late-
were susceptible to vancomycin. The resistance data onset infections, mainly due to staphylococci, in this
of Gram-positive bacteria other than enterococci and group of patients, but may have been instrumental
coagulase-negative staphylococci, as well as those of also in the early-onset cases. Cataract surgeons should
the Gram-negative organisms in the Stockholm thus meticulously rinse the wound of any vitreous
cohort, can at best be considered indicative for the and, if needed, place a suture to make the incision wa-
entire study population. It could be expected, howev- ter tight, not only to avoid cystoid macular edema and
er, that various streptococci and also abiotropha retinal detachment but also to avert the risk of endoph-
and granulicatella species, the latter formally called thalmitis. One of the authors (P.M.) has personally
nutritionally variant streptococci, are sensitive to ce- seen 2 cases in which intraocular infection occurred
furoxime, as are the Gram-negative organisms Hae- in the face of vitreous remnants in the main incision.
mophilis influenzae, Escherichia coli, and Klebsiella Regarding the bacterial origin of endophthalmitis
spp as opposed to P. aeruginosa or S marscesens.16,17 in the elderly population, we found that the occurrence
With these data in mind, it is of interest that the spec- of Gram-negative results was almost exclusive to this
trum of bacteria in the communication with vitreous risk group but still as rare as one incident per 10 000
subset has a high proportion of streptococci that prob- operations. Harboring greater numbers of conjunctival
ably ought to have been eradicated by cefuroxime. bacteria including Gram-negatives seems to be a spe-
Exposure of the vitreous to contaminants is said to cific feature of the elderly cataract population.18 Given



the cefuroxime's broad spectrum of efficacy against or- and its associated endophthalmitis registry. We found
ganisms including some common Gram-negative spe- that operations with capsule complications were more
cies, it is not obvious that the alternative prophylactic prone to result in impaired vision. As for optimizing
regimens discussed below would perform better. prevention, giving an intracameral antibiotic should
Overall, the predominant causes of endophthalmitis, take care of the group with the highest incidence of en-
whether in populations at risk or not at risk, are entero- dophthalmitis, and intervening earlier in the course of
cocci and CoNS, all of which and most of which, respec- cataract development will help to the limit the number
tively, are cefuroxime resistant. A likely explanation is of patients aged 85 years or more and also hopefully
that these infections occur as a result of selection; the avoid cataract cases in which capsule complications
sensitive strains are killed by the prophylaxis, whereas more often occur, appear to be the most obvious mea-
the resistant strains remain and proliferate. Oddly sures. In terms of bacterial etiology, patients sustaining
enough, however, enterococcal infections were also a communication with vitreous and those aged 85 years
identified in cases in which cefuroxime was not admin- or more showed statistically significant differences in
istered, for which we have no explanation. Intracam- comparison with the nonrisk group, although all of
eral vancomycin would offer coverage of these these operations by definition were done with IC cefur-
organisms and most other Gram-positive bacteria.19 oxime. There is suspicion that contact between the vitre-
However, we strongly oppose widespread prophylac- ous phase and contaminants could favor infection even
tic use of an antibiotic that is indispensable for treat- with cefuroxime-sensitive strains after a communica-
ment of infections caused by multiresistant strains. In tion with vitreous, given the long latency (which by
addition, the broad effect on Gram-positive organisms far exceeds the half-life of any intracameral antibiotic)
by vancomycin would probably be somewhat balanced between the operation and infection in these cases.
by the emergence of Gram-negative infections. Most probably, a tightening of the wound with a suture
So what modifications of the present intracameral and elimination of the vitreous from the anterior cham-
treatment can be pursued? In fact, initiatives to that ber would prevent some incidents of endophthalmitis
effect are ongoing in Sweden that could present a basis in this risk situation. As for establishing a more efficient
for further improvement of endophthalmitis preven- intracameral antibiotic regimen aiming at protecting
tion. In some clinics, 0.2 mg moxifloxacin is injected. very aged patients and those sustaining a communica-
The rationale is that this fourth-generation quinolone tion with vitreous, further studies from the Swedish
has a broader spectrum of efficacy than cefuroxime. NCR may hopefully provide a point of departure.
For instance, there seems to be an intermediate effect
on enterococci, whereas Gram-negative bacteria WHAT WAS KNOWN
should be covered at least as well as by cefuroxime.20
 Previous prospective Swedish national registry data on
Preliminary and limited data from the NCR published
cataract surgery have shown that in operations in patients
in our 2013 paper, however, indicated no advantage
aged 85 years or more, peroperative communication with
with moxifloxacin as compared with cefuroxime.3 In
vitreou and, above all, nonuse of intracameral cefuroxime
an upcoming analysis of endophthalmitis including
are strongly associated with postoperative endophthalmitis.
more than 50 000 surgeries performed with intracam-
eral moxifloxacin, more robust results will be available.  Bacteria other than staphylococci, that is, virulent bacte-
In same day bilateral surgery, the practice is to add 100 ria, carry a poorer prognosis regarding visual outcome af-
mg ampicillin to the 1 mg cefuroxime dose. Ampicillin ter endophthalmitis.
boosts the intracameral regimen essentially by offering
potential activity against enterococci; this is to mini-
mize the risk of having bilateral endophthalmitis result WHAT THIS STUDY ADDS
from this aggressive organism. By now, well over 50  Ocular comorbidity, virulent bacteria, and belonging to any
000 bilateral operations have been registered in the risk group for endophthalmitis, in particular sustaining a
NCR database. This is considered a solid enough figure communication with vitreous, were all independently
to permit analysis of the possible impact on the rate of associated with a worse visual result after the infection.
endophthalmitis by the mere case selection for this
 Patients with a communication with vitreous were more
kind of surgery and by the added intracameral ampi-
prone to have endophthalmitis caused by any Gram-
cillin. This analysis is also planned in the year to come.
positive bacteria, and patients aged 85 years or more
In summary, this research aimed at defining various
were more likely to have staphylococci and Gram-
relevant characteristics of groups at risk for developing
negative bacteria as causative organsims than patients
endophthalmitis, drawing from the 2002 to 2010 data
not belonging to any risk group.
reporting to the Swedish National Cataract Register



REFERENCES 11. Sparrow JM, Taylor H, Qureshi K, Smith R, Birnie K,

Johnston RL; the UK EPR user group. The Cataract National
1. Endophthalmitis Vitrectomy Study Group. Results of the En-
Dataset electronic multi-centre audit of 55 567 operations: risk
dophthalmitis Vitrectomy Study; a randomized trial of immediate
indicators for monocular visual acuity outcomes. Eye 2012;
vitrectomy and of intravenous antibiotics for the treatment of
26:821826. Available at:
postoperative bacterial endophthalmitis. Arch Ophthalmol
v26/n6/pdf/eye201251a.pdf. Accessed September 13, 2015
1995; 113:14791496
m M, Wejde G, Stenevi U, Thorburn W, Montan P. En- 12. Arbisser LB. Safety of intracameral moxifloxacin for prophylaxis
2. Lundstro
of endophthalmitis after cataract surgery. J Cataract Refract
dophthalmitis after cataract surgery; a nationwide prospective
Surg 2008; 34:11141120
study evaluating incidence in relation to incision type and loca-
13. Arshinoff SA, Bastianelli PA. Incidence of postoperative endoph-
tion. Ophthalmology 2007; 114:866870
m M, Stenevi U, Montan P. Six-year incidence thalmitis after immediate sequential bilateral cataract surgery.
3. Friling E, Lundstro
J Cataract Refract Surg 2011; 37:21052114
of endophthalmitis after cataract surgery: Swedish national
14. Artzen D, Lundstrom M, Behndig A, Stenevi U, Lydahl E,
study. J Cataract Refract Surg 2013; 39:1521
Montan P. Capsule complication during cataract surgery:
4. Barry P, Seal DV, Gettinby G, Lees F, Peterson M, Revie CW;
case-control study of preoperative and intraoperative risk fac-
for the ESCRS Endophthalmitis Study Group. ESCRS study of
tors; Swedish Capsule Rupture Study Group report 2.
prophylaxis of postoperative endophthalmitis after cataract sur-
J Cataract Refract Surg 2009; 35:16881693
gery; preliminary report of principal results from a European
15. Lundstrom M, Behndig A, Kugelberg M, Montan P, Stenevi U,
multicenter study. J Cataract Refract Surg 2006; 32:407410;
Thorburn W. Decreasing rate of capsule complications in cata-
erratum, 709
m M, Stenevi U, Thorburn W. Endophthalmitis ract surgery; eight-year study of incidence, risk factors, and
5. Montan P, Lundstro
data validity by the Swedish National Cataract Register.
following cataract surgery in Sweden. The 1998 national prospec-
J Cataract Refract Surg 2011; 37:17621767
tive survey. Acta Ophthalmol Scand 2002; 80:258261. Available
16. Senn L, Entenza JM, Greub G, Jaton K, Wenger A, Bille J,
Calandra T, Prodhom G. Bloodstream and endovascular infec-
800305.x/pdf. Accessed September 13, 2015
m M, Stenevi U, Thorburn W. tions due to Abiotropha defectiva and Granulicatella species.
6. Wejde G, Montan P, Lundstro
BMC Infect Dis 2006; 6:9. Available at: http://www.ncbi.nlm.
Endophthalmitis following cataract surgery in Sweden: Na- Ac-
tional prospective survey 19992001. Acta Ophthalmol Scand
cessed September 13, 2015
2005; 83:710. Available at: http://www3.interscience.wiley.
17. Leder RD, Carson DS. Cefuroxime Axetil (Ceftin): a brief review.
com/cgi-bin/fulltext/120092610/PDFSTART. Accessed Sept-
Infect Dis Obstet Gynecol 1997; 5:211214. Available at: http://
ember 13, 2015 Accessed
7. Endophthalmitis Vitrectomy Study Group. Microbiologic factors
September 13, 2015
and visual outcome in the Endophthalmitis Vitrectomy Study.
18. Rubio EF. Influence of age on conjunctival bacteria of patients un-
Am J Ophthalmol 1996; 122:830846
dergoing cataract surgery. Eye 2006; 20:447454. Available at:
8. Barry P, Gardner S, Seal D, Gettinby G, Lees F, Peterson M, Ac-
Revie C; for the ESCRS Endophthalmitis Study Group. Clinical
cessed September 13, 2015
observations associated with proven and unproven cases in
19. Gimbel HV, Sun R, DeBrof BM. Prophylactic intracameral antibi-
the ESCRS study of prophylaxis of postoperative endophthalmi-
otics during cataract surgery: the incidence of endophthalmitis
tis after cataract surgery. J Cataract Refract Surg 2009;
and corneal endothelial loss. Eur J Implant Refract Surg 1994;
9. Doft BH, Wisniewski SR, Kelsey SF, Fitzgerald SG. Diabetes and
20. Matsuura K, Miyoshi T, Suto C, Akura J, Inoue Y. Efficacy and
postoperative endophthalmitis in the Endophthalmitis Vitrectomy
safety of prophylactic intracameral moxifloxacin injection in
Study; the Endophthalmitis Vitrectomy Study Group. Arch Ophthal-
Japan. J Cataract Refract Surg 2013; 39:17021706
mol 2001; 119:650656. Available at: http://archopht.jamanetwork.
com/article.aspx?articleidZ266225. Accessed September 13,
10. Combey de Lambert A, Campolmi N, Cornut P-L, Aptel F, Creu-
First author:
zot-Garcher C, Chiquet C; for the French Institutional Endoph-
thalmitis Study Group. Baseline factors predictive of visual
Mats Lundstr om, MD, PhD
prognosis in acute postoperative bacterial endophthalmitis in Department of Clinical Sciences,
patients undergoing cataract surgery. JAMA Ophthalmol 2013; Ophthalmology, Faculty of Medicine,
131:11591166. Available at: http://archopht.jamanetwork.
Lund University, Lund, Sweden
com/article.aspx?articleidZ1716433. Accessed September
13, 2015