Clinical Characteristics and Prognosis of Adult Orbital Cellulitis in a Tertiary General Hospital

Purpose Adult orbital cellulitis (OC) occurs relatively rarely, and comprehensive studies that retrospectively evaluate OC are lacking. Here, we aimed to examine the clinical characteristics and prognosis of OC in a tertiary general hospital. Methods Between October 2010 and May 2019, patients presenting with clinical symptoms of OC in a tertiary general hospital were analyzed in this retrospective study. Twenty-six cases were identified for a detailed review. In these cases, 16 males and 10 females were diagnosed with orbital cellulitis by clinical characteristics and multimodal examinations. We divided patients into three groups: (1) patients secondary to rhinosinusitis, (2) patients secondary to endogenous infection(s) without endophthalmitis, and (3) patients secondary to endophthalmitis. For each group, age, gender, eye type, combined systemic diseases, clinical presentation, leukocyte count, blood culture, diagnostic imaging, therapeutic methods, length of stay, time of postoperation, and patient prognosis were analyzed in detail. Results There were no significant differences regarding age, gender, preoperative leukocyte count, exophthalmia, blood culture, treatment, or visual changes within the three groups (P < 0.05). There were significant differences, however, in postoperative leukocyte count and ophthalmoplegia between the three groups (P < 0.05). The preoperative and postoperative logarithms of the minimum resolution angle scored by the best-corrected visual acuity (LogMAR BCVA) of group 3 were statistically significant compared to group 1 and group 2 (P < 0.05). Conclusion We confirmed that the prognosis of OC patients combined with systemic diseases was poor. Patients with OC secondary to endophthalmitis infrequently experience ophthalmoplegia; however, these patients still have poor visual outcomes. Patients cultured positive for Klebsiella pneumoniae infection may not be associated with liver abscess. The level of leukocytes may indicate the condition of the disease.


Introduction
Orbital cellulitis (OC) is an urgent and deadly inflammation of the orbital soft tissue. OC is more common in children and relatively rare in adults [1]. Patients often present with blepharedema, ptosis, exophthalmia, ophthalmoplegia, ocular pain, and visual impairment [2]. OC can spread to the adjacent tissue or the cavernous sinus, leading to cavernous sinus thrombophlebitis, and endanger the patient's life [3]. Computed tomography (CT) scans or magnetic resonance imaging (MRI) are the standard tools for diagnosing OC, defining the extent of the infection [4] and identifying the size and location of the orbital abscess [2].
Several reports suggested that sinus inflammation, endophthalmitis, or endogenous infection metastasis could cause OC [5][6][7]. To the best of our knowledge, however, the available retrospective and comprehensive studies regarding the clinical characteristics and prognosis of adult OC in the literature are lacking. is article aims to describe the detailed clinical features and prognosis of adult OC in a tertiary general hospital over a 10-year period.

Study Design.
is study was designed as a retrospective review of the medical records of 26 adult patients diagnosed with OC. e patients were admitted to a tertiary general hospital between October 2010 and May 2019. Children patients with OC were excluded from the review. Age, gender, eye type, combined systemic diseases, clinical presentation, leukocyte count, blood culture, CT or MRI, therapeutic methods, length of stay, time of postoperation, and prognosis are all described in detail.
is study was designed and conducted in accordance with the Declaration of Helsinki and was submitted to the appropriate review board. All participating patients provided informed consent for this study.

Statistical Analysis.
We collected data from the patients' clinical records, and they were processed with SPSS (Statistical Package for Social Sciences, version 19.0, IBM). For categorical variables, Pearson's χ 2 or Fisher's exact test was used; for continuous variables, the Kruskal-Wallis test was used. Two-tailed tests of significance were performed, and P values <0.05 were regarded as statistically significant.

Demographics.
e adult patients, 16 males and 10 females that were diagnosed with OC, had a median age of 48 years (range 32.75-66.50). All patients presented with OC in only one eye, with 14 cases that occurred in the right eye and 12 cases that occurred in the left eye. e patients were divided into three groups according to etiology: (1) patients secondary to rhinosinusitis, (2) patients secondary to endogenous infection(s) without endophthalmitis, and (3) patients secondary to endophthalmitis. In the six cases comprising group 1, the instances of OC were split evenly between the right and left eyes. Group 2 (n � 3) included two right eye cases and one left eye case. Finally, in group 3 (n � 17), nine and eight cases presented in the right and left eyes, respectively (Tables 1-3). ere were no significant differences in age or gender among the three groups (P < 0.05).

Clinical
Presentation. Group 1 and group 2 patients presented with blepharedema and moderately reduced visual acuity. Group 1 had five cases with exophthalmia and five cases with ophthalmoplegia, while group 2 had two cases with exophthalmia and two cases with ophthalmoplegia, with one patient presenting with both (Tables 1 and 2). e patients in group 3 presented with blepharedema and severe visual impairment: 11 cases with exophthalmia and 5 cases with ophthalmoplegia. e best-corrected visual acuity (BCVA) of 15 cases presented with no light perception (NLP), 1 case had LP, and 1 case had BCVA of 2/20 (Table 3).

Combined Systemic Diseases.
OC does not always present alone. Two cases were complicated with diabetes in group 1, and 1 case was secondary to T-cell lymphoma. In group 2, two cases were complicated with iron-deficiency anemia or secondary to acute lymphocytic leukemia. Five cases were caused by trauma in group 3. Furthermore, 3 cases were complicated with liver or cervical abscesses, and 10 cases were complicated with other systemic diseases including sepsis, diabetes, hypertension, organ failure, or hematopathy (Tables 1-3).

Leukocyte Counts.
Routine blood examinations were performed on all patients, and leukocyte counts were analyzed.
e normal value of leukocyte counts was 4.0-10.0 × 1000 uL. e leukocyte counts of 6 patients, 3 patients, and 16 patients were abnormal before the treatment in groups 1, 2, and 3, respectively. While after treatment, the leukocyte counts of 5 patients and 16 patients returned to normal in groups 1 and 3, respectively. However, the total leukocyte counts of 3 patients in group 2 remained abnormal consistently (Tables 1-3).

Blood
Cultures. Blood cultures were obtained from all patients; the bacterial culture and fungal culture were both ordered in the blood culture, and three of them were positive. Two cases tested positive for Klebsiella pneumoniae, and one case was positive for Pseudomonas aeruginosa. e fungal cultures were all negative (Tables 2 and 3).

CT or MRI.
Each patient received CT initially. MRI was performed in ten patients (Tables 1-3). CT scans revealed swollen orbital soft tissue, increased fat density, and opacification of the involved sinuses in the patients with sinusitis. Two cases presented bony destruction, and thickening of the associated muscles was present in eight cases. Optic nerve involvement was demonstrated in one case. In addition, we observed one case combined with cavernous sinus thrombophlebitis.
MRI revealed ill-defined infiltration of orbital fat. Only four cases demonstrated an extremely hyperintense signal representing orbital abscess formation when analyzed by diffusion-weighted imaging (DWI) and the corresponding hypointense signal on the apparent diffusion coefficient (ADC).

Length of Stay.
In group 1, the mean hospital stay was 17.17 ± 9.87 days, and the mean time to discharge after operation was 12.33 ± 9.45 days. For group 2, the mean hospital stay was 19.33 ± 3.06 days, and the mean time to discharge after operation was 2.00 ± 3.46 days. Finally, in group 3, the mean hospital stay was 11.94 ± 5.73 days with a mean time to discharge after operation of 5.71 ± 3.80 days (Tables 1-3).

erapeutic Methods and Prognosis.
All patients diagnosed with OC were treated with active treatments once admitted to the hospital, including systemic and topical antibiotics, and further therapeutic plans were made according to the patients' conditions.    ree cases in group 1 were treated with endoscopic sinus surgery, while another three cases were treated conservatively. Only one patient could not visualize hand motion at the time of hospitalization and had NLP at the time of discharge. In this case, we considered that inflammation spread to the optic nerve. One case in group 2 was treated with abscess incision, and the other two cases were treated with conservative treatment.
Six cases in group 3 were treated with intravitreal injection of vancomycin and cefazolin. Five and two cases were treated with the evisceration of eye contents and enucleation, respectively. Abscess incision was done in two cases, and three cases were treated with conservative treatment. Unfortunately, all patients had a BCVA of NLP (Tables 1-3).
Microbial cultures of the drained abscesses from three patients were performed, and Klebsiella pneumoniae was identified and cultured in one patient. In two patients, there was no growth reported.

Group Comparisons.
ere were no significant differences in preoperative leukocyte count, exophthalmia, blood culture, treatment, and visual changes among the three groups (P < 0.05).
ere were significant differences in postoperative leukocyte count and ophthalmoplegia between the three groups (P < 0.05). e preoperative and postoperative minimum resolution angles in logarithm bestcorrected visual acuity (LogMAR BCVA) of group 3 were statistically significant compared with group 1 and group 2 (P < 0.05) (Tables 4 and 5).

Discussion
OC can be divided into anterior orbital septum cellulitis and posterior orbital septum cellulitis [8]. A variety of reasons cause OC. Firstly, inflammation of ethmoid, maxillary, frontal, or sphenoid sinuses can lead to OC [5]. Secondly, reports indicate that endophthalmitis commonly causes OC [6]. irdly, some OC patients had secondary endogenous infection metastasis [7]. In immunocompromised patients or in patients with systemic diseases, such as leukemia, T-cell lymphoma, sepsis, and diabetes, OC more easily leads to endogenous and secondary infections [9].
OC resulting from rhinosinusitis occurs most frequently in children and rarely in adults [2]. Contradicting this finding, however, Siedek et al. found that 75% of rhinosinusitis-caused OC patients in their study were adults [5]. e anatomy of the sinus lies close to the orbit. As a result, sinus inflammation may spread directly to the orbit through the bone or indirectly through the valveless venous plexus surrounding the orbit and the sinuses [10]. Spread through the ethmoid sinusitis was the most common because of the thin ethmoid bone [2]. According to Stammberger's classification, OC is the most serious ocular complication, secondary to rhinosinusitis (stage IV) [11]. Reports suggested that stages I and II can be treated conservatively, while stages III and IV were recommended to receive surgery [5]. In our group 1 of all adult patients, five patients belonged to stage IV, 60% (n � 3) of which were treated with endoscopic sinus surgery. Of those who received surgery, 66.7% (n � 2) recovered, while one patient with T-cell lymphoma deteriorated.
OC also can be caused by endophthalmitis or panophthalmitis. While the underlying mechanism remains unclear, it is generally assumed that intraocular inflammation can spread to the periocular tissues, resulting in orbital cellulitis [6,12]. In traumatic endophthalmitis, the wound may implant the pathogenic bacteria directly in the periocular or orbital tissues. By a similar mechanism, cataract, strabismus, or retinal detachment surgery can also lead to OC [6,13,14]. Endogenous endophthalmitis is especially prone to occur in immunocompromised patients or patients with severe systemic diseases. In these cases, the primary origins of the infection cannot be found [9].
Endophthalmitis patients with mild symptoms can be treated conservatively, but severe cases can be treated with intravitreal injection. In these instances, vancomycin and cefazolin have proven effective in controlling the inflammation [15]. If the disease still progresses, enucleation or evisceration surgery should be performed [16]. In our study, six cases in group 3 were treated with intravitreal injection (vancomycin and cefazolin), five cases were treated with evisceration of the eye contents, two cases were treated with enucleation, and three cases were treated conservatively. In total, 17 cases (100%) of OC with secondary endophthalmitis were effectively controlled. We considered that OC secondary to endophthalmitis could be controlled if active treatments are given.
K. pneumoniae is a small, packaged, Gram-negative bacterium [17], and secondary endophthalmitis complicated with a liver abscess is considered a clinical syndrome closely associated with diabetes in Asians [18]. In our study, two patients cultured positive for K. pneumoniae infection by blood culture. ey were found with cervical abscess and sepsis, however, which were not associated with liver abscess. On the contrary, the blood cultures of the two liver abscess patients were negative for pathogenic bacteria.
OC in only three patients was caused by endogenous infection without endophthalmitis. In these instances, one case was complicated with acute lymphocytic leukemia, and one situation was complicated with iron-deficiency anemia and ulcerative colitis. In the first patient, the blood culture was positive for P. aeruginosa, and the patient had complications with sepsis during the hospitalization period. e condition deteriorated eventually even after timely treatment.
Routine blood tests showed evidence of an increase in leukocyte count in most OC patients [2]. In our study, comparisons among the three groups illustrated that postoperative leukocyte counts of patients in group 2 changed more dramatically than the other two groups. We also noticed that the incidence of ophthalmoplegia in group 3 was less than that of other groups, indicating that OC with secondary endophthalmitis rarely invaded extraocular muscles (29.4%), which have not been reported in the literature. Furthermore, in comparing the preoperative and postoperative LogMAR BCVA among the three groups, we concluded that patients in group 3 had poorer visual   Journal of Ophthalmology outcomes than any other group because all patients ended up with the BCVA of NLP. Endophthalmitis is a severe disease that can result in serious visual impairment and even blindness. erefore, we concluded that the poorer vision in group 3 is attributable to endophthalmitis. In our study, bacterial and fungal blood cultures were performed. ree patients were positive for bacteria (11.5%), while the other patients were negative. Fungal infections are common in adult OC patients with diabetes [10]. However, in the present study, the results of fungal cultures were all negative, even in the patients combined with systemic diseases, such as diabetes, leukemia, T-cell lymphoma, and sepsis. Similarly, in a study by Hsu J et al., blood cultures were the most frequently collected but were the least likely to be positive (17.6%) [19]. We speculated that the results were negative due to the antibiotic treatment prior to blood collection. In addition, leukocyte counts are still a suitable parameter for diagnosing infection. However, the relationship between the prognosis of infection and leukocyte has been rarely described [20]. In our study, the routine blood tests showed evidence of elevated leukocyte count in all the patients except for 2, one complicated with acute lymphocytic leukemia and the other with T-cell lymphoma. e leukocyte counts returned to normal when the patients recovered. However, the counts of the two patients with acute lymphocytic leukemia and T-cell lymphoma remained abnormal. is finding is consistent with the deterioration of their disease, so we concluded that the leukocyte counts might indicate the recovery or deterioration of the disease. ere were several limitations to this study. Firstly, this retrospective study had a small sample size. Secondly, the patient had been treated with antibiotics at the time of blood culture collection. As a result, the positive rate is relatively low. irdly, the patients included were all adults, so this study is unable to describe the differences in clinical characteristics and prognosis between adults and children with OC.

Conclusions
We confirmed that OC could be well controlled if timely treatments are given. e prognosis of orbital cellulitis patients combined with severe systemic diseases, however, was generally poor. While patients with orbital cellulitis secondary to endophthalmitis infrequently experience ophthalmoplegia, these patients still have poor visual outcomes. Patients cultured positive for Klebsiella pneumoniae infection may not be associated with liver abscess. Furthermore, the leukocyte levels may indicate the condition of the disease. More data by multiple centers are required to understand the clinical characteristics and prognosis of OC better.