Cervicofacial tissue infarction in patients with acute invasive fungal sinusitis: prevalence and characteristic MR imaging findings. Neuroradiology. 2013;55:467-473. [PMID: 23377235 DOI: 10.1007/s00234-013-1147-8]

COVID-Associated Sinonasal Mucormycosis: Radiological Pathological Correlation and Prognostic Value of MR Imaging

Purpose  Our aim was to assess the sinonasal magnetic resonance imaging (MRI) features of acute invasive fungal rhinosinusitis (AIFRS) in coronavirus disease (COVID)-associated mucormycosis (CAM) and to correlate these with histopathology and patient outcome in terms of duration of hospital stay and survival at 10 weeks. Methods  Twenty patients with histopathologically confirmed sinonasal CAM underwent MRI (including postcontrast T1-weighted and diffusion-weighted imaging). Histopathological findings (presence of coagulative necrosis, granulomatous reaction, and fungal burden) were recorded and all patients were followed up at 6 and 10 weeks. Statistical analysis was done using chi-square test and Fischer's exact test. Results  Enhancement patterns seen in our subjects included homogeneous, heterogeneous, and lack of contrast enhancement (LOC), with LOC being the most common (65%). Diffusion restriction was found in 90% patients. Statistically significant correlation was found between LOC pattern and presence of coagulative necrosis ( p -value = 0.007), extent of fungal hyphae ( p -value = 0.047), and duration of hospital stay ( p -value = 0.004). Restricted diffusion was also seen to correlate with a high fungal load ( p -value = 0.007). Conclusion  Our study describes the MRI findings of AIFRS in CAM and highlights the imaging features which may be surrogate markers for coagulative necrosis and fungal burden.

Intracranial manifestations of rhinocerebral mucormycosis: a pictorial essay

Rhinocerebral mucormycosis has emerged as a common coinfection in coronavirus disease 2019 (COVID-19) patients during the convalescence period. Frequent spread of disease from sinonasal mucosa to bone, neck spaces, orbit, and brain occurs along the perivascular/perineural routes or through direct invasion. Brain involvement represents severe manifestation and is often associated with poor functional outcomes and high mortality rates. Magnetic resonance imaging (MRI) is the modality of choice for the intracranial assessment of disease severity in mucormycosis. Early and accurate identification of intracranial extension is imperative to improve survival rates. With this pictorial essay, we aim to familiarize the readers with the cross-sectional imaging features of intracranial complications of mucormycosis. The radiological details in this essay should serve as a broad checklist for radiologists and clinicians while dealing with this fulminant infection.

Magnetic Resonance Imaging Prognostic Findings for Visual and Mortality Outcomes in Acute Invasive Fungal Rhinosinusitis

PURPOSE

To identify initial, preintervention magnetic resonance imaging (MRI) findings that are predictive of visual and mortality outcomes in acute invasive fungal rhinosinusitis (AIFRS).

DESIGN

Retrospective cohort study.

PARTICIPANTS

Patients with histopathologically or microbiologically confirmed AIFRS cared for at a single, tertiary academic institution between January 2000 and February 2020.

METHODS

A retrospective review of MRI scans and clinical records of patients with confirmed diagnosis of AIFRS was performed. For each radiologic characteristic, a modified Poisson regression with robust standard errors was used to estimate the risk ratio for blindness. A multivariate Cox proportional hazards model was used to study AIFRS-specific risk factors associated with mortality.

MAIN OUTCOME MEASURE

Identification of initial, preintervention MRI findings associated with visual and mortality outcomes.

RESULTS

The study comprised 78 patients (93 orbits, 63 with unilateral disease and 15 with bilateral disease) with AIFRS. The leading causes of immunosuppression were hematologic malignancy (38%) and diabetes mellitus (36%). Mucormycota constituted 56% of infections, and Ascomycota constituted 37%. The overall death rate resulting from infection was 38%. Risk factors for poor visual acuity outcomes on initial MRI included involvement of the orbital apex (relative risk [RR], 2.0; 95% confidence interval [CI], 1.1-3.8; P = 0.026) and cerebral arteries (RR, 1.8; 95% CI, 1.3-2.5; P < 0.001). Increased mortality was associated with involvement of the facial soft tissues (hazard ratio [HR], 4.9; 95% CI, 1.3-18.2; P = 0.017), nasolacrimal drainage apparatus (HR, 5.0; 95% CI, 1.5-16.1; P = 0.008), and intracranial space (HR, 3.5; 95% CI, 1.4-8.6; P = 0.006). Orbital soft tissue involvement was associated with decreased mortality (HR, 0.3; 95% CI, 0.1-0.6; P = 0.001).

CONCLUSIONS

Extrasinonasal involvement in AIFRS typically signals advanced infection with the facial soft tissues most commonly affected. The initial, preintervention MRI is prognostic for a poor visual acuity outcome when orbital apex or cerebral arterial involvement, or both, are present. Facial soft tissues, nasolacrimal drainage apparatus, intracranial involvement, or a combination thereof is associated with increased mortality risk, whereas orbital soft tissue involvement is correlated with a reduced risk of mortality.

Post-COVID-19 acute invasive fungal rhinosinusitis: a systematic radiological approach in the light of clinico-surgical characteristics

Background

The objective is to describe the radiological imaging findings of post-COVID-19 acute invasive fungal rhinosinusitis (AIFRS), being familiar with the wide variety of imaging spectrum, thus raising the suspicion for early diagnosis. Methods: In this retrospective study, we reviewed the imaging findings in 54 patients with proven post-COVID-19 AIFRS who underwent endoscopic/surgical debridement from April 2020 to September 2021. Most of these patients presented with facial or orbital swelling or facial pain. Medical records with a special emphasis on radiological imaging (50 NCCT of the paranasal sinuses and 17 MRIs of the orbit) were reviewed regarding the degree of mucosal disease of sinuses, nasal cavity, and nasopharynx, extra sinus soft tissue infiltration, especially orbital and cerebral extension (parenchymal, cavernous sinus, vascular or neuro-invasion).

Results

We reported findings in 54 patients with post-COVID-19 AIFRS, of whom 30 were men and 24 were women with a mean age of 48.06. Unexpectedly, infiltration of pterygopalatine fossa was found to precede mucosal opacification of sinuses nasal cavity and affection of nasopharynx. Out of 54 patients, 49 showed inflammatory changes involving pterygopalatine fossa, 29.6% of patients showed infiltration of orbital tissues, 22 patients suffered from a fungal invasion of the cavernous sinus and 3 patients had carotid artery involvement.

Conclusions

Imaging findings of AIFRS significantly vary from subtle mucosal thickening of paranasal sinuses, up to orbital and intracranial extension with vascular thrombosis and neuroinvasion. The hallmark inflammatory tissue infiltration into the pterygopalatine fossa and facial soft tissue may precede mucosal disease.

Magnetic Resonance Imaging in Coronavirus Disease - 2019 Associated Rhino-Orbital-Cerebral Mucormycosis (CA-ROCM) - Imaging Analysis of 50 Consecutive Patients

BACKGROUND AND PURPOSE

Rhino-orbital-cerebral mucormycosis has emerged as a major opportunistic infection in patients with COVID-19. High clinical suspicion and prompt imaging are crucial for early diagnosis and management. Our study evaluates imaging characteristics of patients with COVID-19 associated Rhino-orbital-cerebral Mucormycosis (CA-ROCM) in a tertiary care hospital in India.

MATERIALS AND METHODS

A retrospective analysis of clinical and imaging data of patients with CA-ROCM who presented between December 2020 to June 2021 was performed. All patients had microbiologically or histologically proven sino-nasal mucormycosis along with documented SARS-CoV-2 positive RT-PCR test and/or classical lung imaging features of COVID-19 infection. The extent of sinus involvement, bony erosions, extra-sinus soft tissue extension, orbital-intracranial invasion, perineural spread, and vascular complications were assessed.

RESULTS

Fifty patients were included for the final analysis. Diabetes was the most common associated comorbidity. Seven patients presented with stage I disease, 18 patients with stage II, and 25 patients with stage III disease. The stage of disease showed a positive statistical correlation with HbA1c levels using Pearson's correlation. The common imaging features were "Black turbinate sign" and nonenhancing sino-nasal mucosa (82%), orbital involvement (76%), and diffusion restriction in the optic nerve (24%). Intracranial involvement was seen as perineural extension into the brain (42%), cerebritis (30%), and internal carotid artery involvement (16%).

CONCLUSIONS

CA-ROCM is an acute invasive fungal sinusitis with an aggressive clinical course. Black-turbinate sign and peri-antral soft tissue infiltration are early features, whereas extra-nasal tissue infarction, optic nerve diffusion restriction, and vascular invasion are seen with advanced disease.

Trismus originating from rare fungal myositis in pterygoid muscles: A case report

BACKGROUND

Trismus is a common problem with various causes. Any abnormal conditions of relevant anatomic structures that disturb the free movement of the jaw might provoke trismus. Trismus has a detrimental effect on the quality of life. The outcome of this abnormality is critically dependent on timely diagnosis and treatment, and it is difficult to identify the true origin in some cases. We present a rare case of trismus due to fungal myositis in the pterygoid muscle, excluding any other possible pathogenesis.

CASE SUMMARY

The patient presented with a 2-mo history of restricted mouth opening. Computed tomography showed obvious enlargement of the left pterygoid muscles. Furthermore, the patient had trismus without obvious predisposing causes. The primary diagnosis was pterygoid myosarcoma. Consequently, lesionectomy of the left pterygoid muscle was performed. Intraoperative frozen biopsy implied the possibility of an uncommon infection. Postoperative pathologic examination confirmed myositis and necrosis in the pterygoid muscle. Fungi were detected in both muscle tissue and surrounding necrotic tissue. The patient recovered well with antifungal therapy and mouth opening exercises. The rarity of fungal myositis may be responsible for the misdiagnosis. Although the origin of pathogenic fungi is still unknown, we believe that both hematogenous spread and local invasion could be the most likely sources. To the best of our knowledge, this is the first case in the literature that reported fungal myositis in pterygoid muscles as the only reason that results in trismus.

CONCLUSION

Surgeons should remain vigilant to the possibility of trismus originating from fungal myositis.

Imaging of COVID-19-associated craniofacial mucormycosis: a black and white review of the "black fungus"

A subset of diabetic COVID-19 patients treated with steroids, oxygen, and/or prolonged intensive care admission develop rhino-orbito-cerebral mucormycosis. Radiologists must have a high index of suspicion for early diagnosis, which prompts immediate institution of antifungal therapy that limits morbidity and mortality. Assessment of disease extent by imaging is crucial for planning surgical debridement. Complete debridement of necrotic tissue improves survival. Imaging features reflect the angioinvasive behaviour of fungal hyphae from the Mucoraceae family, which cause necrotising vasculitis and thrombosis resulting in extensive tissue infarction. Contrast-enhanced magnetic resonance imaging (MRI) is the imaging technique of choice. The classic “black turbinate” on contrast-enhanced imaging represents localised invasive fungal rhinosinusitis (IFRS). A striking radiological feature of disseminated craniofacial disease is non-enhancing devitalised and necrotic soft tissue at the orbits and central skull base. Sinonasal and extrasinonasal non-enhancing lesions in IFRS are secondary to coagulative necrosis induced by fungal elements. Multicompartmental and extrasinonasal tissue infarction is possible without overt bone involvement and caused by the propensity of fungal elements to disseminate from the nasal cavity via perineural and perivascular routes. Fungal vasculitis can result in internal carotid artery occlusion and cerebral infarction. Remnant non-enhancing lesions after surgical debridement portend a poor prognosis. Assessment for the non-enhancing MRI lesion is crucial, as it is a sole independent prognostic factor for IFRS-specific mortality. In this review, we describe common and uncommon imaging presentations of biopsy-proven rhino-orbito-cerebral mucormycosis in a cohort of nearly 40 COVID-19 patients.

Chronic invasive fungal rhinosinusitis vs sinonasal squamous cell carcinoma: the differentiating value of MRI

OBJECTIVES

To investigate MRI features in discriminating chronic invasive fungal rhinosinusitis (CIFRS) from sinonasal squamous cell carcinomas (SNSCC).

METHODS

MRI findings of 33 patients with CIFRS and 47 patients with SNSCC were retrospectively reviewed and compared. Multivariate logistic regression analysis was performed to identify significant imaging features in distinguishing between CIFRS and SNSCC. The ROC curves and the AUC were used to evaluate diagnostic performance.

RESULTS

There were significant differences in cavernous sinus involvement (p < 0.001), sphenoid sinus involvement (p < 0.001), meningeal involvement (p = 0.024), T2 signal intensity (p = 0.006), and enhancement pattern (p < 0.001) between CIFRS and SNSCC. Multivariate logistic regression analysis identified cavernous sinus involvement (odds ratio [OR] = 0.06, 95% confidence interval [95% CI] = 0.02-0.20) and sphenoid sinus involvement (OR = 0.14, 95% CI = 0.05-0.45) as significant indicators for CIFRS and T2 isointensity to gray matter (OR = 4.44, 95% CI = 1.22-16.22) was a significant indicator for SNSCC. ROC curve analysis showed the AUC from a combination of three imaging features was 0.95 in differentiating CIFRS and SNSCC.

CONCLUSIONS

MRI showed significant differences between CIFRS and SNSCC features. In immunocompromised patients, a sinonasal hypointense mass on T2WI with septal enhancement or loss of contrast enhancement, and involvement of cavernous sinus, sphenoid sinus, and meninges strongly suggest CIFRS.

KEY POINTS

• Chronic invasive fungal rhinosinusitis (CIFRS) is often difficult to distinguish from sinonasal squamous cell carcinomas (SNSCC) in clinical practice. • Cavernous sinus and sphenoid sinus involvement appear to be significant indicators for CIFRS. T2 isointensity to gray matter appears to be a significant indicator for SNSCC. • Loss of contrast enhancement and septal enhancement can be used to distinguish CIFRS from SNSCC with a high degree of specificity.

Treatment outcomes in acute invasive fungal rhinosinusitis extending to the extrasinonasal area

Acute invasive fungal rhinosinusitis (AIFRS) can spread beyond the sinonasal cavity. It is necessary to analyze the association between the specific site involved in the extrasinonasal area and the survival rate to predict patient prognosis. We investigated 50 patients who had extrasinonasal lesions on preoperative gadolinium (Gd)-enhanced magnetic resonance imaging (MRI) scan and underwent wide surgical resection of AIFRS. The specific sites with loss of contrast enhancement (LoCE) on Gd-enhanced MRI were analyzed for AIFRS-specific survival rate. The most common underlying disease was diabetes mellitus followed by hematological malignancy. The most common symptoms were headache and facial pain. Seven patients (14.0%) expired because of AIFRS progression. Poor prognosis was independently associated with LoCE at the skull base on preoperative MRI (HR = 35.846, P = 0.004). In patients with AIFRS extending to the extrasinonasal area, LoCE at the skull base was an independent poor prognostic factor.

The Black Turbinate Sign, A Potential Diagnostic Pitfall: Evaluation of the Normal Enhancement Patterns of the Nasal Turbinates

BACKGROUND AND PURPOSE

Lack of enhancement of a nasal turbinate on MR imaging, known as the black turbinate, has been reported as a finding highly suggestive of invasive fungal rhinosinusitis in immunocompromised patients. Our purpose was to investigate the normal enhancement pattern of nasal turbinates and to determine whether a black turbinate occurs in patients without invasive fungal rhinosinusitis and, if so, to distinguish differentiating features from pathologic enhancement.

MATERIALS AND METHODS

We examined patient medical records and available MR imaging, which included most nasal turbinates in more than 1 sequence. Imaging was performed with contrast, either with or without fat saturation, in patients without invasive fungal rhinosinusitis. All MR images were evaluated for a turbinate enhancement pattern and the presence of nonenhancing areas.

RESULTS

After we applied the exclusion criteria, 75 MR images from each group, with and without fat saturation, were included. Overall, the frequency of observed nonenhancing portions of turbinates, ie, black turbinates, was 30%. Most observed black turbinates were located in the posterior portion of inferior turbinates. Tiny areas of nonenhancement within turbinates were also a common finding.

CONCLUSIONS

A black turbinate is not uncommon on MR images in immunocompetent patients who have no chance of having invasive fungal rhinosinusitis. Benign turbinate nonenhancement improves over subsequent series, has preserved thin peripheral enhancement, and often has thin internal septa. Knowledge of this normal pattern of nasal turbinate enhancement can aid in the differentiation of the benign black turbinate in immunocompetent patients from the pathologic black turbinate seen in patients with invasive fungal rhinosinusitis.

Using IFN-γ antibodies to identify the pathogens of fungal rhinosinusitis: A novel immunohistochemical approach

Fungal rhinosinusitis (FRS) is commonly caused by various Aspergillus species (spp) and Mucorales fungi, and the treatment and prognosis of cases differ depending on the causative fungus. The present study describes a novel immunohistochemical method that has high sensitivity and specificity for distinguishing between these two types of fungi in patients with FRS. Three groups were included in the study. Group A included formalin-fixed paraffin-embedded blocks of 51 nasal tissue specimens of patients with FRS (27 Aspergillus spp and 24 Mucorales) that were continuously obtained from the Department of Pathology of Tongren Hospital in Beijing as the experimental group and 34 cultures (26 Aspergillus spp and 8 Mucorales) of FRS that were randomly selected from the bacterial laboratory of Tongren Hospital in Beijing to verify the staining results of the paraffin-embedded blocks. Formalin-fixed paraffin-embedded blocks of 10 esophageal cancer specimens were included in Group B as the positive control group. All specimens in Groups A and B were stained with interferon-γ (IFN-γ) antibody. Group C consisted of the same specimens as described in Group A, however, when performing the immunohistochemical assay, IFN-γ antibody was replaced by PBS and this served as the negative control group. The differences in IFN-γ immunohistochemical staining between Aspergillus spp and Mucorales were analyzed. Staining of IFN-γ in paraffin-embedded samples was positive in 92.6% (25/27) of specimens in which Aspergillus spp were the causative pathogen, which was significantly higher compared with specimens in which Mucorales was causative (P<0.001), with only 4.2% (1/24) of specimens staining positive for IFN-γ. Immunohistochemical staining of cell cultures was 100% positive for Aspergillus spp, whereas all Mucorales were negative. Thus, the results of the current study indicated that IFN-γ antibody immunohistochemical staining may be used as a novel diagnostic tool to distinguish between Aspergillus spp and Mucorales when identifying the causative agent in FRS, providing a useful supplementary test to the current immunohistochemical methods in the clinical diagnosis of FRS.

Extrasinonasal infiltrative process associated with a sinonasal fungus ball: does it mean invasive fungal sinusitis?

PURPOSE

Invasive fungal sinusitis (IFS) has rarely been reported to develop from non-IFS. The purpose of this study was to disclose the nature of the extrasinonasal infiltrative process in the presence of a sinonasal fungus ball (FB).

METHODS

We retrospectively reviewed the medical records, computed tomography, magnetic resonance images of 13 patients with sinonasal FB and the extrasinonasal infiltrative process. Based on histology and clinical course, we divided the extrasinonasal infiltrative process into IFS and the nonfungal inflammatory/infectious process (NFIP). The images were analyzed with particular attention to the presence of cervicofacial tissue infarction (CFTI).

RESULTS

Of the 13 patients, IFS was confirmed in only one, while the remaining 12 were diagnosed to have presumed NFIP. One patient with IFS died shortly after diagnosis. In contrast, all 12 patients with presumed NFIP, except one, survived during a mean follow-up of 17 months. FB was located in the maxillary sinus (n=4), sphenoid sinus (n=8), and both sinuses (n=1). Bone defect was found in five patients, of whom four had a defect in the sphenoid sinus. Various sites were involved in the extrasinonasal infiltrative process, including the orbit (n=10), intracranial cavity (n=9), and soft tissues of the face and neck (n=7). CFTI was recognized only in one patient with IFS.

CONCLUSION

In most cases, the extrasinonasal infiltrative process in the presence of sinonasal FB did not seem to be caused by IFS but probably by NFIP. In our study, there were more cases of invasive changes with the sphenoid than with the maxillary FB.

Distinguishing the dominant species of pathogen in maxillary sinusitis by sequencing DNA dataset analysis

This study determined the microbial composition in maxillary sinusitis and determined the predominant fungal and bacterial species. Samples were obtained from 11 patients with fungal sinusitis and 2 healthy patients. The 13 samples were sequenced using an Illumina Hi-Seq 2000, and mapping to the human, bacterial, and the fungal genomes were based on several steps. Bioinformatics and statistical analyses were then performed for bacterial and fungal expressions in the samples. All samples were cultures for fungal growth. The sequencing data revealed that Aspergillus flavus and Aspergillus oryzae were the dominant fungal strains in the rhinosinusitis samples, and Bacillus cereus and Bacillus thuringiensis were the dominant bacterial strains. Indeed, the fungi and bacteria were associated with the development of fungal rhinosinusitis. Furthermore, B. cereus and B. thuringiensis may cooperate with Pseudomonas aeruginosa to inhibit the growth of fungal mycelia. Knowledge of the microbial composition can provide a diagnostic reference for patients with maxillary sinusitis.

Prognostic Factors for Survival in Patients with Acute Invasive Fungal Rhinosinusitis

Background

Acute invasive fungal rhinosinusitis (AIFR) is an aggressive opportunistic infection with a high mortality rate. There are few reports that demonstrate an improvement in the overall prognosis. Furthermore, definite prognostic factors related to patient survival remain unclear.

Objective

The objective of this study was to evaluate the various clinical factors related to survival of patients with AIFR.

Methods

This is a retrospective case series with patients treated for AIFR between 1997 and 2013. A total of 45 patients with AIFR were enrolled for analysis. We evaluated demographics, clinical characteristics, and disease course.

Results

Mean age of patients was 59.6 years. AIFR developed most commonly in patients with immunocompromised host, such as diabetes (n = 23) or hematologic malignancy (n = 17). There were two main genera of fungus, Aspergillus (n = 30) and Mucor (n = 14). Headache, cranial neuropathy, visual loss, and orbital pain were the most common presenting symptoms. Overall survival was 53%. Underlying hematologic malignancy and diabetes were significantly associated with overall survival, and accompanying severe neutropenia and elevated C-reactive protein (CRP) were also related to poor prognosis. Initial presentation with facial swelling, involvement of nasal septum, or shorter symptom duration was also associated with survival reduction. Multivariate analysis revealed that CRP more than 5.50 mg/dL (Hazard ratio [HR], 9.04; p = 003) was an independent prognostic factor in patients with AIFR.

Conclusions

Overall survival rate remained approximately 50% in patients with AIFR. The prognosis of AIFR is significantly influenced by underlying diseases, accompanying neutropenia, CRP levels, symptom duration, involvement of septum, and the presence of facial swelling. Elevation of CRP, in particular, was an independent predictor of poor outcomes and should be monitored appropriately.

A novel fungus concentration-dependent rat model for acute invasive fungal rhinosinusitis: an experimental study

Background

Acute invasive fungal rhinosinusitis is a lethal infectious process afflicting immunocompromised individuals. Knowledge about this disease is still limited due to the scarcity of animal models designed to study the pathogenesis of this infection. Mast cells are tissue-resident immune cells that participate in a variety of allergic and inflammatory conditions. Limited attention has been given to the role of mast cells in acute invasive fungal rhinosinusitis. Therefore, the objectives of this study were to create a rat model of acute invasive fungal rhinosinusitis based on analyzing the impact of different fungal concentrations on establishing infection, and to observe the changes of mast cells in rats with this disease.

Methods

Sprague–Dawley rats were divided randomly into four groups, three of which were experimental and received different concentrations of Aspergillus fumigatus inoculations, and one was a control group (D). The inoculated Aspergillus fumigatus concentrations were 5 × 10⁷ conidia/ml in group A, 10⁷ conidia/ml in group B, and 10⁶ conidia/ml in group C. Before fungal inoculation, rats were immunosuppressed using cyclophosphamide and cortisone acetate, and had Merocel sponges inserted into the right nares. Hematology and histopathology investigations were then performed.

Results

An acute invasive fungal rhinosinusitis rat model was established successfully with an incidence rate of 90% in group A, 50% in group B and 10% in group C. Aspergillus fumigatus invasion was observed in 20% of the lungs in group A, but was not seen in the remaining groups. In addition, no fungi invaded the orbital tissue, brains, livers, spleens or kidneys of any rat. Compared with the control set, the total number of mast cells in the experimental groups was not significantly increased, but mast cell degranulation, on the other hand, was only found in infected nasal cavities.

Conclusions

This investigation illustrates that various fungal concentrations have different effects on the incidence of acute invasive fungal rhinosinusitis, and it also demonstrates the feasibility of using this model to study the process of fungal rhinosinusoidal invasion. In addition, the results suggest that mast cells may play a role in the protection of sinuses against acute Aspergillus fumigatus infection and in the clearance of established hyphal masses.

Electronic supplementary material

The online version of this article (doi:10.1186/s12879-014-0713-y) contains supplementary material, which is available to authorized users.